INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
INTER-AGENCY REVIEW DRAFT EPA/####
DO NOT CITE OR QUOTE DATE
Review Draft
www.epa.gov/ncea
E. ANALYSIS OF LIVER AND CO-
EXPOSURE ISSUES FOR THE TCE
TOXICOLOGICAL REVIEW
NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally
released by the U.S. Environmental Protection Agency and should not at this
stage be construed to represent Agency policy. It is being circulated for comment
on its technical accuracy and policy implications.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
02/12/09
E-i
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
DISCLAIMER
This document is an external review draft for review purposes only and does not
constitute U.S. Environmental Protection Agency policy. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
02/12/09
E-ii
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
CONTENTS—Appendix E: Analysis of Liver and Co-exposure Issues for the TCE
Toxicological Review
1. BASIC PHYSIOLOGY AND FUNCTION OF THE LIVER - A STORY OR
HETEROGENEITY E-l
1.1. HETEROGENEITY OF HEPATOCYTES AND ZONAL
DIFFERENCES IN FUNCTION AND PLOIDY E-l
1.2. EFFECTS OF ENVIRONMENT AND AGE; VARIABILITY OF
RESPONSE E-6
2. CHARACTERIZATION OF HAZARD FROM TCE STUDIES E-8
2.1. ACUTE TOXICITY STUDIES E-9
2.1.1. Soni etal. 1998 E-9
2.1.2. Soni etal. 1999 11-12
2.1.3. Okino etal. 1991 11-12
2.1.4. Nunes etal. 2001 1 >13
2.1.5. Tao etal. 2000 11-14
2.1.6. Tucker etal. 1982 11-14
2.1.7. Goldsworthy and Popp 1987 E-16
2.1.8. Elcombe etal 1985 E-17
2.1.9. Dees and Travis 1993 E-27
2.1.10. Nakajima et al. 2000 E-31
2.1.11. Berman et al 1995 E-33
2.1.12. Melnick etal. 1987 11-3 5
2.1.13. Laughter et al. 2004 E-37
2.1.14. Ramdhan et al. 2008 E-41
2.2. SI BCI IRONIC AND CHRONIC STUDIES OF TCE I>47
2.2.1. Merrick et al. 1989 E-47
2.2.2. Goel etal. 1992 1>51
2.2.3. Kj ell strand et al. 1981 1 >52
2.2.4. Woolhiser et al. 2006 1>55
2.2.5. Kj ell strand et al 1983(a) E-56
2.2.6. Kj ell strand et al 1983(b) I >60
2.2.7. Buben and O'Flaherty 1985 E-64
2.2.8. Channel et al. 1998 1>66
2.2.9. Dorfmueller et al. 1979 E-69
2.2.10. Kumar et al. 2001 E-70
2.2.11. Kawamoto et al. 1988 E-71
2.2.12. NTP 1990 1>72
2.2.12.1. 13-week studies E-72
2.2.12.2. 2-year studies E-74
2.2.13. NTP 1988 1>75
2.2.14. Fukuda etal. 1983 1 >77
2.2.15. Henschler et al. 1980 1>78
2.2.16. Maltoni et al 1986 E-80
2.2.17. Maltoni et al. 1988 1>84
2.2.18. Van Duuren et al. 1979 E-84
02/12/09
E-iii
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
CONTENTS (continued)
2.2.19. NCI 1976 I >85
2.2.20. Herren-Freund et al. 1987 E-89
2.2.21. Anna et al. 1994 E-89
2.2.22. Bull et al. 2002 1>90
2.3. MODES OF ACTION: RELATIVE CONTRIBUTION OF TCE
METABOLITES I >92
2.3.1. Acute studies of DCA/TCAs E-93
2.3.1.1. Sanchez and Bull 1990 E-93
2.3.1.2. Nelson et al. 1989 E-101
2.3.1.3. Styles et al. 1991 E-103
2.3.1.4. Carter etal. 1995 E-104
2.3.1.5. DeAngelo et al. 1989 E-109
2.3.2. Subchronic and Chronic Studies of DCA and TCA E-l 15
2.3.2.1. Snyder etal. 1995 E-115
2.3.2.2. Mather etal. 1990 E-l 16
2.3.2.3. Parrish et al. 1996 E-l 17
2.3.2.4. Bull etal. 1990 E-l 19
2.3.2.5. Nelson etal. 1990 E-122
2.3.2.6. DeAngelo et al. 1999 11-124
2.3.2.7. Carter et al. 2003 E-129
2.3.2.8. Stauber and Bull 1997 E-132
2.3.2.9. Pereira 1996 E-133
2.3.2.10. Pereira and Phelps 1996 E-137
2.3.2.11. Ferreira-Gonzalez et al. 1995 E-138
2.3.2.12. Pereira et al. 2004 E-139
2.3.2.13. DeAngelo et al. 2008 E-141
2.3.2.14. DeAngelo et al. 1997 E-163
2.3.2.15. DeAngelo et al. 1996 11-164
2.3.2.16. Richmond et al. 1995 E-165
2.4. SUMMARIES AND COMPARISONS BETWEEN TCE, DCA, AND
TCA STUDIES E-l66
2..4.1. Summary of Results for Short-term Effects of TCE E-l 67
2..4.2. Summary of Results for Short-term Effects of DCA and TCA:
Comparisons with TCE E-172
2..4.3. Summary of Results for Subchronic and Chronic Studies of TCE E-194
2..4.4. Summary of Results for Subchronic and Chronic Studies of DCA
and TCA: Comparisons with TCE E-204
2.5. STUDIES OI CHLORAL HYDRATE I>225
2.6. SERUM BILE ACID ASSAYS I>230
3. STATE OF SCIENCE OF LIVER CANCER MOAS I >232
3.1. STATE OF THE SCIENCE FOR CANCER AND SPECIFICALLY
HUMAN LIVER CANCER I>234
3.. 1.1. Epigenetics and Diseases States (Transgenerational Effects, Effects
of Aging, and Background Changes E-234
02/12/09
E-iv
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
CONTENTS (continued)
3..1.2. Emerging Technologies, DNA and siRNA, miRNA Microarrays-
promise and limitations for MOA E-241
3.. 1.3. Etiology, Incidence and Risk Factors for HCC E-248
3.. 1.4. Issues Associated with Target Cell Identification E-251
3..1.5. Status of Mechanisms of Action for Human HCC E-255
3.. 1.6. Pathway and Genetic Disruption Associated with HCC and
Relationship to Other Forms of Neoplasia E-257
3.. 1.7. Epigenetic Alterations in HCC E-258
3..1.8. Heterogeneity of Preneoplastic and HCC Phenotypes E-260
3.2. ANIMAL MODELS OF LIVER CANCER I>266
3..2.1. Similarities with Human and Animal Transgenic Models E-269
3 .3. HYPOTHESIZED KEY EVENTS IN HCC USING ANIMAL MODELS.. .E-272
3..3.1. Changes in Ploidy E-272
3..3.2. Hepatocellular Proliferation and Increased DNA Synthesis E-277
3..3.3. Nonparenchymal Cell Involvement in Disease States Including
Cancer E-280
3.3.3.1. Epithelial cell control of liver size and cancer-
angiogenesis E-280
3.3.3.2. Kupffer cell control of proliferation and cell signals, role
in early and late effects E-281
3.3.3.3. Nf-KB and TNF-a - context, timing, and source of cell
signaling molecules E-281
3. .3.4. Gender Influences on Susceptibility E-286
3..3.5. Epigenetic Modification E-288
3.4. SPECIFIC HYPOTHESIS FOR MOA OF TCE
HEPATOCARCINOGENECITY IN RODENTS I>290
3..4.1. PPARa Agonism as the MOA for Liver Tumor Induction - the
State of the Hypothesis E-290
3.4.1.1. Heterogeneity of PPARa agonist effects and inadequacy
of WY-14,643 paradigm as prototype for class E-291
3.4.1.2. New Information on causality and sufficiency for PPARa
receptor activation E-297
3.4.1.3. Use ofPPAR-/- knockout and humanized mouse E-299
3.4.1.4. NF-kB activation E-311
3.4.1.5. Phenotype as an indicator of a PPARa MOA E-312
3.4.1.6. Human relevance E-315
3. .4.2. Other TCE Metabolite Effects That May Contribute to its
Hepatocarcinogeneicity E-316
3.4.2.1.. DCA effects and glycogen accumulation correlations
with cancer E-316
3.4.2.2.. Genetic profiling data for TCE: gene expression and
methylation status studies E-318
3.4.2.3.. Oxidative stress E-322
4. EFFECTS OF CO-EXPOSURE ON MOA-INTERNAL AND EXTERNAL
EXPOSURES TO MIXTURES INCLUDING ALCOHOL E-325
02/12/09 E-v
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
CONTENTS (continued)
4.1. INTERNAL CO-EXPOSURES TO TCE METABOLITES:
MODULATION OF TOXICITY AND IMPLICATIONS FOR TCE MOA.E-327
4.2. INITIATION STUDIES AS CO-EXPOSURES E-328
4.2.1. Herren-Freund et al 1987 E-328
4.2.2. Parnell et al. 1986 E-329
4.2.3. Pereira and Phelps 1996 E-330
4.2.4. Tao et al. 2000 E-334
4.2.5. Lantendresse and Pereira 1997 E-334
4.2..6. Pereira et al. 1997 E-336
4.2.7. Tao etal. 1998 E-338
4.2.8. Stauber etal. 1998 E-339
4.3. COEXPOSURES OF HALO ACETATES AND OTHER SOLVENS E-341
4.2.1. Carbon tetrachloride, DC A, TCA: implications for MOA from Co-
exposures E-341
4.2.2. Chloroform, DCA, and TCA Co-exposures: Changes in
Methylation Status E-343
4.2.3. Co-exposures to Brominated Haloacetates: Implications for
Common MOAs and Background Additivity to Toxicity E-345
4.2.4. Co-exposures to Ethanol: Common Targets and MOAs E-347
4.2.5. Co-exposure Effects on Pharamacokinetics: Predictions Using
Physiologically Based Pharmacokinetic (PBPK) Models E-349
5. POTENTIALLY SUSCEPTIBLE LIFE STAGES AND CONDITIONS THAT
MAY ALTER RISK OF LIVER TOXICITY AND CANCER 11-3 51
6. UNCERTAINTY AND VARIABILITY E-352
7. REFERENCES E-353
02/12/09
E-vi
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
FOREWORD
The purpose of this Appendix is to provide scientific support and rationale for the hazard
and dose-response sections regarding liver effects and those of co-exposure of the Toxicological
Review of Trichloroethylene (TCE). It is not intended to be a comprehensive treatise on the
chemical or toxicological nature of TCE. Please refer to the Toxicological Review of TCE for
characterized of EPA's overall confidence in the quantitative and qualitative aspects of hazard
and dose response for TCE-induce liver effects. Matters considered in this appendix include
knowledge gaps, uncertainties, quality of data, and scientific controversies. This characterization
is presented in an effort to make apparent the scientific issues regarding the data and MOA
considerations for experimental animal data for liver effects in the TCE assessment.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to U.S. EPA's IRIS Hotline at 301-345-2870.
02/12/09
E-vii
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
AUTHORS, CONTRIBUTORS, AND REVIEWERS
The National Center for Environmental Assessment (NCEA), within U.S. EPA's Office
of Research and Development, was responsible for the preparation of the IRIS toxicological
review and summary documents for TCE. The MTBE team would like to thank the NCEA
management team for their comments and support, and XXXX for technical editing support.
Chemical Manger
Weihsueh Chiu, U.S. EPA, NCEA-Washington
Principal author
Jane Caldwell, U.S. EPA, NCEA-Washington
Contributors
Weihsueh Chiu, U.S. EPA, NCEA-Washington
Marina V. Evans, U.S. EPA, NHEERL-RTP on detail to NCEA-Washington
Kate Guyton, U.S. EPA, NCEA-Washington
NCEA Management team
Paul White, U.S. EPA, NCEA-Washington
Bob Sonawane, U.S. EPA, NCEA-Washington
Dave Bussard, U.S. EPA, NCEA-Washington
John Vandenberg, U.S. EPA, NCEA-IO
Peter Preuss, U.S. EPA, NCEA-IO
External Consultant Reviewer
External Reviewers
02/12/09
E-viii
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1. Basic physiology and function of the liver - a story of
heterogeneity
The liver is a complex organ whose normal function and heterogeneity are key to
understanding and putting into context perturbations by TCE, cancer biology, and variations in
response observed and anticipated for susceptible life stages and background conditions.
1.1 Heterogeneity of Hepatocytes and Zonal Differences in Function and
Ploidy
Malarkey et al 2005 state that (1) the liver transcriptome (i.e., genes expressed as
measured by mRNA) is believed only second to the brain in its complexity and includes about
25-40% of the approximately 50,000 mammalian genes, (2) during disease states the
transcriptome can double or triple and its increased complexity is due not only to differential
gene expression (up- and down-regulation of genes) but also to the mRNA contributions from
the heterogeneous cell populations in the liver and (3) when one considers that over a dozen cell
types comprise the liver in varying proportions, particularly in disease states, knowledge about
the cell types and cell-specific gene expression profiles help unravel the complex genomic and
protenomic data sets. Gradients of gene and protein activity varying from the periportal region
to the centrilobular region also exist for sinusoidal endothelial cells, Kuffper cells, hepatic
stellate cells, and the matrix in the space of Disse. Malarkey et al (2005) also estimate that
hepatocytes constitute 60%, sinusoidal endothelial cells 20%, Kupffer cells 15%, and stellate
cells 5% of liver cells. Therefore in experimental paradigms where liver homogenates are used
for the determination of "changes in liver" gene expression or other parameters the individual
changes from cells residing in differing zones and by differing cell type is lost. Malarkey et al
define the need to better characterize the histological cellular components of the tissues from
which mRNA and protein is extracted and referred to "phenotypic anchoring" and cite
acetaminophen as a "model hepatotoxicant under study to assess the strengths and weaknesses of
genomics and proteinomics technologies" as well as "a good example for understanding and
utilizing phenotypic anchoring to better understand genomics data." After acetaminophen
exposure "there is an unexplained and striking inter and intralobular variability in acute hepatic
necrosis with some regions having massive necrosis and adjacent areas within the same lobe or
other lobes showing no injury at all." Malarkey et al. (2005) go on to cite similar lobular
variability in response for "copper distribution, iron and phosphorous, chemical and spontaneous
carcinogenesis, cirrhosis and regeneration" and suggest that although uncertain "factors such as
portal streamlining of blood to the liver, redistribution of blood to core of the liver secondary to
nerve stimulation, and exposures during fetal development and possibly lobular gradients are
important." Hepatic interlobe differences exist for initiating agents in terms of DNA alkylation
and cell replication. In the rat, DEN alkylation has been reported to occur preferentially in the
left and right median lobes, while cell replication was higher in the right median and right
anterior lobes (Richardson et al 1986). Richardson et al (1986) reported that exposure to DEN
induced a 100% incidence of hepatocellular carcinoma in the left, caudate, left median and right
02/12/09
E-l
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
median lobes of the liver by 20 weeks versus only 30% in the right anterior and right posterior
hepatic lobes. There was a reported interlobe difference in adduct formation, cell proliferation,
liver lobe weight gain, number and size of GGT+ foci, and carbon 14 labeling from a single dose
of DEN. Richardson et al. (1986) suggest that many growth-selection studies utilizing the liver
to evaluate the carcinogenic potential of a chemical often focus on only one or two of the hepatic
lobes, which is especially true for partial hepatectomy, and that for DEN and possibly other
chemicals this procedure removes the lobes most likely to get tumors. Thus, the "distribution of
toxic insult may not be correctly assessed with random sampling of the liver tissue for
microarray gene expression analysis" (Malarkey et al., 2005) and certainly any such
distributional differences are lost in studies of whole liver homogenates.
The liver is normally quiescent with few hepatocytes undergoing mitosis and, as
described below, normally occurring in the periportal areas of the liver. Mitosis is observed only
in approximately one in every 20,000 hepatocytes in adult liver (Columbano and Ledda-
Columbano, 2003). The studies of Schwartz-Arad et al. (1989), Zajicek et al. (1991), Zajicek
and Schwartz-Arad (1990), and Zajicek et al. (1989) have specifically examined the birth, death
and relationship to zone of hepatocytes as the "hepatic streaming theory." They report that
hepatocytes and littoral cells continuously steam from the portal tract toward the terminal hepatic
vein and that the hepatocyte differentiates as it goes with biological age closely related to cell
differentiation. In other words, the acinus may be represented by a tube with two orifices: for
cell inflow situated at the portal tract rim and other for cell outflow, at the terminal hepatic vein
with hepatocytes streaming through the tube in an orderly fashion. In normal liver, cell
proliferation is suggested as the only driving force of this flow with each mitosis associated with
displacement of the cells by one cell location and the greater the cell production, the faster the
flow and visa versa (Zajicek et al., 1991). Thus, the microscopical section of the liver "displays
an instantaneous image of a tissue in flux"(Schwartz-Arad et al., 1989). Schwartz-Arad et al
further suggest that "throughout its life the hepatocyte traverses three acinus zones; in each it is
engaged in different metabolic activity. When young it performs among other functions
gluconeogenesis, which is found in zone 1 hepatocytes (i.e. periportal), and when old it turns into
a zone 3 cell (i.e., pericentral), with a pronounced glycolitic make up. The three zones thus
represent differentiation stages of the hepatocyte, and since they differ by their distance from the
origin, e.g. zone 2 (i.e., midzonal) is more distant than zone 1, again, hepatocyte differentiation is
proportional to its distance." Chen et al. (1995) report that "Hepatocytes are a heterogeneous
population that are composed of cells expressing different patterns of genes. For example,
gamma-glutamyl transpeptidase and genes related to gluconeogenesis are expressed preferential
in periportal hepatocytes, whereas enzymes related to glycolysis are more abundant in the
centrilobular area. Glutamine synthetase is expressed in a small number of hepatocytes
surrounding the central veins. Most cytochrome p450 enzymes are expressed or induced
preferentially in centrilobular hepatocytes relative to periportal hepatocytes." Along with
changes in metabolic function, Vielhauer et al. (2001) reported that there is evidence of zonal
differences in carcinogen DNA effects and also chemical specific differences for DNA repair
enzyme and that enhanced DNA repair is a general feature of many carcinogenic states including
the enzymes that repair alkylating agents but also oxidative repair. As part of this process of
02/12/09
E-2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
differentiation and as livers age, the hepatocyte changes and increases its ploidy with polyploid
cells predominant in zone 2 of the acinus (Schwartz-Arad et al., 1989). The reported decrease in
DNA absorbance in zone 3 may be due to 1) a decline in chromatin affinity to the dye, 2) cell
death, and 3) DNA exit from intact cells and Zajicek and Schwartz-Arad (1990) suggest that the
fewer metabolic demands in Zone 3, under normal conditions, causes the cell to "deamplify" its
genes and for DNA excess to leak out cells adjacent to the terminal hepatic vein or to be
eliminated by apoptosis reflecting cell death. Thus, the three acinus zones represent
differentiation states of one and the same hepatocyte which increase ploidy as functional
demands change. Zajicek and Schwartz-Arad, (1990) also report that nuclear size is generally
proportional to DNA content and that as DNA accumulates, the nucleus enlarges. This has
import for histopathological descriptions of hepatocellular hypertrophy and attendant nuclear
changes after toxic insult as well.
The gene amplification associated with polyploidy is manifested by DNA accumulation
that involves the entire genome (Zajicek and Schwartz-Arad, 1990). Polyploidization is always
attended by the intensification of the transcription and translation and in rat liver the amino acid
label and activity of many enzymes increases proportionately to their ploidy. "Individual
chromosomes of a tetraploid genome of a hepatocyte reduplicate in the same sequence as in a
diploid one. In this case the properties of the chromosomes evidently remain unchanged and
polyploidy only means doubling the indexes of the diploid genome" (Brodsky and Uryvaeva
1977). Polyploidy will be manifest in the liver by either increases in the number of
chromosomes per nucleus in an individual cell, or by the appearance of two nuclei in a single
cell. Most cell polyploidization occurs in youth with mitotic polyploidization occurring
predominantly from 2 to 3 weeks post-natally and increases with age in mice (Brodsky and
Uryvaeva, 1977). Hepatocytes progress through a modified or polyploidizing cell cycle which
contains gaps and S-phases, but proceeds without cytokinesis. The result is the formation of the
first polyploidy cell which is binucleated with diploid nuclei and has increased cell ploidy but
not cell number. The subsequent proliferation of bi-nucleated hepatocytes occurs with a fusion
of mitotic nuclei during metaphase that gives rise to mononucleated cells with higher levels of
ploidy. Thus, during normal liver ontogenesis, a polyploidizing cell cycle without cytokinesis
alternates with a mitotic cycle of binucleated cells and results in progressive and irreversible
increases in either cell or nuclear ploidy (Brodsky and Uryvaeva 1977).
Polyploidization of the liver occurs during maturation in rodents and therefore
experimental paradigms that treat or examine rodent liver during that period should take into
consideration the normally changing baseline of polyploidy in the liver. The development of
polyploidy has been correlated in rodents to correspond with maturation. Brodsky and Uryvaeva
(1977) report it is cells with diploid nuclei that proliferate in young mice, but that among the
newly formed cells, the percentage of those with tetraploid nuclei is high. By 1 month most
mice (CBA/C57BL mice) already have a polyploid parenchyma, but binucleate cells with diploid
nuclei predominate. In adult mice the ploidy class with the highest percentage of hepatocytes
was the 4n X 2 class. The intensive proliferation of diploid hepatocytes occurs only in baby
02/12/09
E-3
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mice during the first 2 weeks of life and then toward 1 month, the diploid cells cease to maintain
themselves and transform into polyploid cells. In aged animals the parenchyma retains only 0.02
percent of the diploid cells of the newborn animal. While the weight of the liver increases
almost 30 times within 2 years, the number of cells increases much less than the weight or mean
ploidy. Hence the postnatal growth of the liver parenchyma is due to cell polyploidization
(Brodsky and Uryvaeva, 1977). In male Wistar rats fetal hepatocytes (22 days gestation) were
reported to be 85.3% diploid (2n) and 7.4% polyploid (4n + 8n) cells with 7.3% of cells in S-
phase (SI and S2). By one month of age (25- day old suckling rats) there were 92.9 % diploid
and 2.5% polyploid, at 2 months 47.5% diploid and 50.9% polyploid, at 6 months 29.1 % diploid
and 69.6% polyploid, and by 8 months 11.1 % diploid and 87.3% polyploidy (Sanz et al., 1996).
However, mouse and rat differ in their polyploidization. "In the mouse, which has a higher
degree of polyploidy than the rats, the scheme of polyploidization differs in that each cell class,
including mononucleate cells, forms from the preceding one without being supplemented by self-
maintenance. Each cell class is regarded as the cell clone and it is implied that the cells of each
class have the same mitotic history and originate from diploid initiator cells with similar
properties. In this model 1 reproduction would give a 2n X 2 cell, the second reproduction a 4n
cell, and third reproduction a 4n X 2 cell all coming from an originator diploid cell" (Brodsky
and Uryvaeva, 1977).
The cell polyploidy is most extensive in mouse liver, but also common for rat and
humans livers. The livers of young and aged mice differ considerably in the ploidy of the
parenchymal cells, but still perform fundamentally the same functions. In some mammals, such
as the mouse, rats, dog and human, the liver is formed of polyploid hepatocytes. In others, for
example, guinea pig and cats, the same functions are performed by diploid cells (Brodsky and
Uryvaeva, 1977). One obvious consequence of polyploidization is enlargement of the cells. The
volume of the nucleus and cytoplasm usually increases proportionately to the increased in the
number of chromosome sets with polyploidy reducing the surface/volume ratio. The labeling of
tritium doubles with the doubling of the number of chromosomes in the hepatocyte nucleus
(Brodsky and Uryvaeva, 1977). Kudryavtsev et al. (1993) have reported that the average levels
of cell and nuclear ploidy are relatively lower in humans than in rodent but the pattern of
hepatocyte polyploidization is similar and at maturity and especially during aging, the rate of
hepatocyte polyploidization increases with elderly individuals having binucleated and polyploid
hepatocytes constituting about one-half of liver parenchyma. Gramantieri et al. (1996) report
that in adult human liver a certain degree of polyploidization is physiological; the polyploidy
compartment (average 33% of the total hepatocytes) includes both mononucleated (28%) and
binucleated {12%) cells and the average percentage of binucleated cells in the total hepatocyte
population is 24% (Melchiorri et al., 1994). Historically, aging in human liver has been
characterized by fewer and larger hepatocytes, increased nuclear polyploidy and a higher index
of binucleate hepatocytes (Popper, 1986) but Schmucker (2005) notes that data concerning the
effect of aging on hepatocyte volume in rodent and humans are in conflict with some showing
increases volume to be unchanged and to increase by 25% by age 60 by others in humans. The
irreversibility of hepatocyte polyploidy has been used in efforts to identify the origin of tumor
02/12/09
E-4
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
progenitor cells (diploid vs. polyploidy) (see Section 3.1.8, below). The associations with
polyploidy and disease have been an active area of study in cancer MOA studies (see Sections
3.1.4. and 3.3.1, below).
Not only are polyploid cells most abundant in zone 2 of the liver acinus and increase in
number with age, but polyploid cells have been reported to be more abundant following a
number of toxic insults and exposure to chemical carcinogens. Wanson et al. (1980) reported
that one of the earliest lesions obtained in the liver after N-nitrosomorpholine (NNM) treatment
development of hypertrophic parenchymal cells presenting a high degree of ploidy. Gupta
(2000) reports hepatic polyploidy is often encountered in the presence of liver disease and that
for animals and people, polyploidy is observed during advancement of liver injury due to
cirrhosis or other chronic liver disease (often described as large-cell dysplasia referring to
nuclear and cytoplasmic enlargement, nuclear pleomorphisms and multinucleation and probably
representing increased prevalence of polyploidy cells) and in old animals with toxic liver injury
and impaired recovery. Gorla et al. (2001) report that weaning and commencement of feeding,
compensatory liver hypertrophy following partial hepatectomy, toxin and drug-induced liver
disease, and administration of specific growth factors and hormones may induce hepatic
polyploidy. They go on to state that "although liver growth control has long been studied,
whether the replication potential of polyploidy hepatocytes is altered remains unresolved, in part,
owing to difficulties in distinguishing between cellular DNA synthesis and generation of
daughter cells." Following CCL4 intoxication, the liver ploidy rises and more cells become
binucleate (Zajicek et al., 1989). Minamishima et al. (2002) report that in 8-12 week old female
mice before partial hepatectomy there were 78.6% 2C, 19.1% 4C, and 2.3% 8C cells but 7 days
after there were 42.0% 2C, 49.1% 4C, and 9.0% 8C. Zajicek et al. (1991) describe how
hepatocyte streaming is affected after the rapid hepatocyte DNA synthesis that occurs after the
mitogenic stimulus of a partial hepatectomy. These data are of relevance to findings of increased
DNA synthesis and liver weight gain following toxic insults and disease states. Zajicek et al.
(1991) suggest that following a mitogenic stimulus, not all DNA synthesizing cells do divide but
accumulate newly formed DNA and turn polyploid (i.e., during the first 3 days after partial
hepatectomy in rats 50% of synthesized DNA was accumulated) and that since the acinus
increased 15% and cell density declined 10%, overall cell mass increased 5%. However, cell
influx rose 1300%. "In order to accommodate all these cells, the 'acinus-tube' ought to swell 13-
fold, while in reality it increased only 5%" and that on day 3 "the liver remnant did not even
double in its size". Zajicek et al. conclude that apparently "cells were eliminated very rapidly,
and may have even been sloughed off, since the number of apoptotic bodies was very low" and
therefore "partial hepatectomy triggers two processes: an acute process lasting about a week
marked by massive and rapid cell turnover during which most newly formed hepatocytes are
eliminated, probably sloughed off into the sinusoids; and a second more protracted process
which served for liver mass restoration mainly by forming new acini." Thus, a mitogenic
stimulus may induce increased ploidy and increased cell number as a result of increased DNA
synthesis, and many of the rapidly expanding number of cells resulting from such stimulation are
purged and therefore do not participate in subsequent disease states of the liver.
02/12/09
E-5
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Zajicek et al. (1989) note that the accumulation of DNA rather than proliferation of
hepatocytes "should be considered when evaluating the labeling index of hepatocytes labeled
with tritiated thymidine" as the labeling index, defined as the proportion of labeled cells, can
serve as a proliferation estimate only if it is assumed that a synthesizing cell will ultimately
divide. In tissues, such as the liver, "where cells also accumulate DNA, proliferation estimates
based on this index may fail" (Zajicek et al., 1989). The tendency to accumulate DNA is also
accompanied by a decreasing probability of a cell to proliferate, since young hepatocytes
generally divide after synthesizing DNA while older cells prefer instead to accumulate DNA.
However, polyploidy per se does not preclude cells from dividing (Zajicek et al., 1989). The
ploidy level achieved by the cell, no matter how high, does not in itself prevent it from going
through the next mitotic cycle and the reproduction of hepatocytes in the ploidy classes of 8n and
8n X2 is common phenomenon (Brodsky and Uryvaeva, 1977). However, along with a reduced
capacity to proliferate, Sigal et al. (1999) report that the onset of polyploidy increases the
probability of cell death. The proliferative potentials of hepatocytes not only depend on their
ploidy, but also on the age of the animals with liver restoration occurring more slowly in aged
animals after partial hepatectomy (Brodsky and Uryvaeva, 1977). Species differences in the
ability of hepatocytes to proliferate and respond to a mitogenic stimulus have also been
documented (see Section 3.4.2, below). The importance of the issues of cellular proliferation vs.
DNA accumulation and the differences in ability to respond to a mitogenic stimulus becomes
apparent as identification of the cellular targets of toxicity (i.e., diploid vs. polyploidy) and the
role of proliferation in proposed modes of action (MO A) are brought forth. Polyploidization , as
discussed above, has been associated with a number of types of toxic injury, disease states, and
carcinogenesis by a variety of agents.
1.2. Effects of Environment and Age: Variability of Response
The extent of polyploidization of the liver not only changes with age, but structural and
functional changes, as well as environmental factors (e.g., polypharmacy), affect the
vulnerability of the liver to toxic insult. In a recent review by Schmucker (2005), several of
these factors are discussed. Schmucker reports that approximately 13% of the population of the
United States is over the age of 65 years, that the number will increase substantially over the next
50 years, and that increased age is associated with an overall decline in health and vitality
contributing to the consumption of nearly 40% of all drugs by the elderly. Schmucker estimates
that 65%) of this population is medicated and many are on polypharmacy regimes with a major
consequence of a marked increase in the incidence of adverse drug reactions (ADRs) (i.e., males
and females exhibit 3- and 4-fold increases in ADRs, respectively, when 20- and 60-year old
groups are compared). The percentage of deaths attributed to liver diseases dramatically
increases in humans beyond the age of 45 years with data from California demonstrating a 4-fold
increase in liver disease-related mortality in both men and women between the ages of 45 and 85
years (Seigel and Kasmin, 1997). Furthermore, Schmucker cites statistics from the United Stated
Department of Health and Human Services to illustrate a loss in potential lifespan prior to 75
years of age due to liver disease (i.e., liver disease reduced lifespan to a greater extent than
colorectal and prostatic cancers, to a similar extent as chronic obstructive pulmonary disease, and
02/12/09
E-6
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
nearly as much as HIV). Thus, the elderly are predisposed to liver disease.
As stated above, the presence of high polyploidy cell in normal adults, nuclear
polyploidization with age, and increase in the mean nuclear volume have been reported in
people. Wantanabe et al. (1978) reported the results from a cytophotometrical analysis of 35
cases of sudden death including 22 persons over 60 years of age, that revealed that although the
nuclear size of most hepatocytes in a senile liver remains unchanged, there was an increase in
cells with larger nuclei. Variations in both cellular area and nucleocytoplasmic ratio were also
analyzed in the study, but the binuclearity of hepatocytes was not considered. No cases with a
clinical history of liver disease were included. Common changes in senile liver were reported to
include atrophy, fatty metamorphosis of hepatocytes, and occasional collapse of cellular cords in
the centrilobular area, slight cellular infiltration and proliferation of Kupffer cells in sinusoids,
and elongation of Glisson's triads with slight to moderate fibrosis in association with round cell
infiltration. Furthermore, cells with giant nuclei, with each containing two or more prominent
nucleoli, and binuclear cell. There was a decrease in diploid populations with age and an
increase in tetraploid population and a tendency of polyploidy cells with higher values than
hexaploids with age. Cells with greater nuclear size and cellular sizes were observed in livers
with greater degrees of atrophy.
Schmucker notes that one of the most documented age-related changes in the liver is a
decline in organ volume but also cites a decrease in functional hepatocytes and that other studies
have suggested that the size or volume of the liver lobule increases as a function of increasing
age. Data is cited for rats suggesting sinusoidal perfusion rate in the rat liver remains stable
throughout the lifespan (Vollmar et al., 2002) but evidence in humans shows age-related shifts in
the hepatic microcirculation attributable to changes in the sinusoidal endothelium (McLean et al.,
2003) (i.e., a 60% thickening of the endothelial cell lining and an 80% decline in the number of
endothelial cell fenestrations, or pores, with increasing age in humans) that are similar in baboon
liver (Cogger et al., 2003). Such changes could impair sinusoidal blood flow and hepatic
perfusion, and the uptake of macromolecules such as lipoproteins from the blood. Schmucker
reports that there is a consensus that hepatic volume and blood flow decline with increasing age
in humans but that the effects of aging on hepatocyte structure are less clear. In rats the volume
of individual hepatocytes was reported to increase by 60% during development and maturation,
but subsequently decline during senescence yielding hepatocytes of equivalent volumes in
senescent and very young animals (Schmucker, 2005). The smooth surfaced endoplasmic
reticulum (SER), which is the site of a variety of enzymes involved in steroid, xenobiotic, lipid
and carbohydrate metabolism, also demonstrated a marked age-related decline rat hepatocytes
(Schumucker et al., 1977: Schmucker et al., 1978). Schmucker also notes that several studies
have reported that the older rodents have less effective protection against oxidative injury in
comparison to the young animals, age-related decline in DNA base excision repair, and increases
in the level of oxidatively damaged DNA in the livers of senescent animals in comparison to
young animals. Age-related increases in the expression an activity of stress-induced
transcription factors (i.e., increased NF-kB binding activity but not expression) were also noted,
02/12/09
E-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
but that the importance of changes in gene expression to the role of oxidative stress in the aging
process remains unsolved. An age-related decline in the proliferative response of rat hepatocytes
to growth factors following partial hepatectomy was noted, but despite a slower rate of hepatic
regeneration, older livers eventually achieved their original volume with the mechanism
responsible for the age-related decline in the post-hepatectomy hepatocyte proliferative response
unidentified. As with other tissues, telomere length has been identified as a critical factor in
cellular aging with the sequential shortening of telomeres to be a normal process that occurs
during cell replication (see Sections 3.1.1 and 3.1.7, below). An association in telomere length
and strain susceptibility for carcinogenesis in mice has been raised. Herrera et al., (1999)
examined susceptibility to disease with telomere shortening in mice. However, this study only
cites shorter telomeres for C57BL6 mice in comparison to mixed C57BL6/129sv mice. The
actual data are not in this paper and no other strains are cited. Of the differing cell types
examined, Takubo and Kaminishi (2001) report that hepatocytes exhibited the next fastest rate of
telomere shortening despite being relatively long-lived cells raising the question of whether or
not there are correlations between age, hepatocyte telomere length and the incidence of liver
disease (Schmucker 2005). Aikata et al. (2000) and Takubo et al. (2000) report that the mean
telomere length in healthy livers is approximately 10 kilobase pairs at 80 years of age and these
hepatocytes retain their proliferative capacity but that in diseased livers of elderly subjects was
approximately 5 kb pairs. Thus, short telomere length may compromise hepatic regeneration and
contribute to a poor prognosis in liver disease or as a donor liver (Schmucker, 2005).
Schmucker (2005) reports that interindividual variability in Phase I drug metabolism was
so large in human liver microsomes, particularly among older subjects, that the determination of
any statistically significant age or gender-related differences were precluded. In fact Schmucker
(2001) notes that "the most remarkable characteristic of liver function in the elderly is the
increase in interindividual variability, a feature that may obscure age-related differences".
Schumer notes that The National Institute on Aging estimates that only 15% of individuals aged
over 65 years exhibit no disease or disability with this percentage diminishing to 11% and 5% for
men and women respectively over 80 years. Thus, the large variability in response and the
presence of age-related increases in pharmacological exposures and disease processes are
important considerations in predicting potential risk from environmental exposures.
2. Characterization of Hazard From TCE Studies
The 2001 Draft assessment of the health risk assessment of TCE (U.S. EPA, 2001)
extensively cited the review article by Bull (2000) to describe the liver toxicity associated with
TCE exposure in rodent models. Most of the attention has been paid to the study of TCE
metabolites, rather than the parent compound, and the review of the TCE studies by Bull (2000)
was cursory. In addition, gavage exposure to TCE has been associated with a significant
occurrence of gavage-related accidental deaths and vehicle effects, and TCE exposure through
drinking water has been reported to decrease palatability and drinking water consumption, and to
have significant loss of TCE through volatilization, thus, further limiting the TCE database. In
02/12/09
E-8
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
its review of the draft assessment, EPA's Science Advisory regarding this topic suggested that in
its revision, the studies of TCE should be more fully described and characterized, especially
those studies considered to be key for the hazard assessment of TCE. Although the database for
studies of the parent compound is somewhat limited, a careful review of the rodent studies
involving TCE can bring to bring to light the consistency of observations across these studies,
and help inform many of the questions regarding potential MO As of TCE toxicity in the liver.
Such information can inform current MOA hypothesis (e.g., such as PPARa activation) as well.
Accordingly the primary acute, subchronic and chronic studies of TCE will be described and
examined in detail below and with comments on consistency, major conclusions and the
limitations and uncertainties that their design and conduct. Since all chronic studies were
conducted primarily with the goal of ascertaining carcinogenicity, their descriptions focus on that
endpoint, however, any noncancer endpoints described by the studies are described as well. For
details regarding evidence of hepatotoxicity in humans and associations with increased risk of
hepatocellular carcinoma, please refer to Section XXX of the 2008 revised assessment. Given
that some of the earlier studies with TCE were contaminated with epichlorhydrin, only the ones
without such contamination are examined below.
2.1 Acute Toxicity Studies.
A number of acute studies have been undertaken to describe the early changes in the liver
after TCE administration with the majority using the oral gavage route of administration. Some
have been detailed examinations while others have reported primarily liver weight changes as a
marker of TCE-response. The matching and recording of age but especially initial and final
body weight for control and treatment groups is of particular importance for studies using liver
weight gain as a measure of TCE-response as difference in these parameter affect TCE-induced
liver weight gain. Most data are for exposures of at least 10 days.
2.1.1. Soni et al. 1998
Soni et al. (1998) administered TCE in corn oil to male Sprague-Dawley rats (200-250 g,
8-10 weeks old) intraperitoneally at exposure levels of 250, 500, 1250, and 2500 mg/kg. Groups
(4-6 animals per group) were sacrificed at 0, 6, 12, 24, 36, 48, 72, and 96 hours after
administration of TCE or corn oil. Using this paradigm only 50% of rats survived the 2400
mg/kg ip TCE administration with all deaths occurring between days 1 and 3 after TCE
administration. Tritiated thymidine was also administered i.p. to rats 2 hours prior to euthanasia.
Light microscopic sections of the central lobe in 3-4 sections examined for each animal. The
grading scheme reported by the authors was: 0 , no necrosis; +1 minimal, defined as only
occasional necrotic cells in any lobule; +2, mild, defined as less than one-third of the lobular
structure affected; +3, moderate, defined as between one-third and two-thirds of the lobular
structure affected; +4 severe, defined as greater than two-thirds of the lobular structure affected.
At the 2500 mg/kg dose histopathology data were obtained for the surviving rats (50%).
Lethality studies were done separately in groups of 10 rats. The survival in the groups of rats
02/12/09
E-9
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
administered TCE and sacrificed from 0 to 96 hours was given as 30% mortality at 48 hours and
50% mortality by 72 hours.
The authors report that controls and 0-hour groups did not show sign of tissue injury or
abnormality. The authors only report a single number with one significant figure for each group
of animals with no means or standard deviations provided. In terms of the extent of necrosis
there is no difference between the 250 and 500 mg/kg/treated dose groups though 96 hours with
a single +1 given as the maximal amount of hepatocellular necrosis (minimal as defined by
occasional necrotic cells in any lobule). At the 1250 mg/kg dose the maximal score was
achieved 24 hours after TCE administration and was reported as simply +2 (mild, defined as less
than one-third of lobular structure affected). The level of necrosis was reported to diminish to a
score of 0 72 hours after 250 mg/kg TCE with no decrease at 500 mg/kg. At 1250 mg/kg the
extent of necrosis was reported to diminish from +2 to +1 by 72 hours after administration. At
the 2500 mg/kg dose (LD50 for this route) by 48 hours the surviving rats were reported to have a
score of +4 (severe as defined by greater than two thirds of the lobular structure affected). The
authors report that "The necrosed cells were concentrated mostly in the midzonal areas and the
cells around central vein area were unaffected. Extensive necrosis was observed between 24 and
48 hours for both 1250 and 2500 mg/kg groups. Injury was maximal in the group receiving 2500
mg/kg between 36 and 48 hours as evidenced by severe midzonal necrosis, vacuolization, and
congestion. Infiltration of polymorphonuclear cell was evident at this time as a mechanism for
cleaning dead cells and tissue debris from the lobules. At the highest dose, the injury also started
to spread toward the centrilobular areas. At highest dose 30 and 50% lethality was observed at 48
and 72 h, respectively. After 48 h, the number of necrotic cells decreased and the number of
mitotic cells increased. The groups receiving 500 and 1250 mg/kg TCE showed relatively higher
mitotic activity as evidenced by cells in metaphase compared to other groups." The authors do
not give a quantitative estimate or indication as to the magnitude of the number of cells going
through mitosis. Although there was variability in the number of animals dying at 1250 mg/kg
TCE exposure though this route of exposure, no indication of variability in response within these
treatment groups is given by the author in regard to extent of histopathological changes. The
authors do not comment on the manner of death using this paradigm or of the effects of i.p.
administration regarding potential peritonitis and inflammation.
TCE hepatotoxicity was "assessed by measuring plasma" sorbitol dehydrogenase (SDH)
and alanine aminotransferase (ALT) after TCE administration with vehicle treated control groups
reported to induce no increases in these enzymes. Plasma SDH levels were reported to increase
in a linear fashion after 250, 500 and 1250 mg TCE/kg ip administration by 6 hours (i.e., ~ 3-
fold, 10.5-fold, 22-fold, and 24.5-fold in comparison to controls from 250, 500, 1250, and 2500
mg/kg TCE, respectively) with little difference between the 1250 mg/kg and 250 mg/kg dose.
By 12 hours the 250, 500, and 1250 levels has diminished to levels similar to that of the 250
mg/kg dose at 6 hours. The 2500 mg/kg levels was somewhat diminished from its 6 hour level.
By 24 hours after TCE administration by the i.p. route of administration all doses were similar to
that of 250 mg/kg TCE 6 hour level. This pattern was reported to be similar for5 36, 48, 72, and
96 hour time points as well. The results presented were the means and SE for four rats per
02/12/09
E-10
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
group. The authors did not indicate which rats were selected for these results from the 4-6 that
were exposed in each group. Thus, only SDH levels showed dose dependence in results at the 6
hour time point and such increases did not parallel the patterns reported for hepatocellular
necrosis from histopathological examination of liver tissues.
For ALT, the pattern of plasma concentrations after i.p. TCE administration differed both
from that of SDH but also from liver histopathology. Plasma ALT levels were reported to
increase in a nonlinear fashion and to a much smaller extent that SDH (i.e., ~ 2.7-fold, 1.9-fold,
2.1-fold, and 4.0-fold of controls from 250, 500, 1250, and 2500 mg/kg TCE, respectively). The
patterns for 12, 24, 36, 48, 72, and 96 hours were similar to that of the 6 hour exposure and did
not show a dose-response. The authors injected carbon tetrachloride (2.5.ml/kg) into a separate
group of rats and then incubated the resulting plasma with unbuffered TCA (0, 200, 600, or 600
nmol and go decreases in enzyme activity in vitro at the two higher concentrations. It is not
clear whether in vitro unbuffered TCE concentrations of this magnitude, which could precipitate
proteins and render the enzymes inactive, are relevant to the patterns observed in the in vivo
data. The extent of extinguishing of SDH and ALT activity at the two highest TCA levels in
vitro were the same, suggestive of the generalized in vitro pH effect. However, the enzyme
activity levels after TCE exposure had different patterns, and thus suggesting that in vitro TCA
results are not representative of the in vivo TCE results. Neither ALT nor SDH levels
corresponded to time course or dose-response reported for the histopathology of the liver
presented in this study.
Tritiated thymidine results from isolated nuclei in the liver did not show a pattern
consistent with either the histopathology or enzyme results. These results were for whole liver
homogenates and not separated by nuclear size or cell origin. Tritiated thymidine incorporation
was assumed by the authors to represent liver regeneration. There was no difference between
treated and control animals at 6 hours after i.p. TCE exposure and only a decrease (~ 50%
decrease) in thymidine incorporation after 12 hours of the 2500 mg/kg TCE exposure level. By
24 hours, there as 5.6-fold and 2.8- fold tritiated thymidine incorporation at the 500 and 1250
mg/kg TCE levels with the 250 and 2500 mg/kg levels similar to controls. For 36, 48, and 72
hours after i.p. TCE exposure there continued to be no dose-response and no consistent pattern
with enzyme or histopathological lesion patterns. The authors presented "area under the curve"
data for tritiated thymidine incorporation for 0 to 95 hours which did not include control values.
There was a slight elevation at 500 mg/kg TCE and slight decrease at 2500 mg/kg from the 250
mg/kg TCE levels. Again, these data did not fit either histopathology or enzyme patterns and
also can include the contribution of nonparenchymal cell nuclei as well as changes in ploidy.
The use of an i.p route of administration is difficult to compare to oral and inhalation
routes of exposure given that peritonitis and direct contact with TCE and corn oil with liver
surfaces may alter results. Whereas Soni et al. (1998) report the LD50 to be 2500 mg/kg TCE
via i.p administration, both Elcombe et al. (1985) and Melnick et al. (1987) do not report
lethality from TCE administered for 10 days at 1500 mg/kg in corn oil, or up to 4800 mg/kg/day
for 10-days in encapsulated feed. Also TCE administered via gavage or oral administration
02/12/09
E-ll
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
through feed will enter the liver through the circulation with periportal areas of the liver the first
areas exposed with the entire liver exposed in a fashion dependent on blood concentrations
levels. However, with i.p administration, the absorption and distribution pattern of TCE will
differ. The lack of concordance with measures of liver toxicity from this study and the lack
concordance of patterns and dose-response relationships of toxicity reported from other more
environmentally and physiologically relevant routes of exposure make the relevance of these
results questionable.
2.1.2. Soni et al. 1999.
A similar paradigm and the same results were reported for Soni et al (1999), in which
hepatocellular necrosis, tritiated thymidine incorporation, and in vitro inhibition of SDH and
ALT data were presented along with dose-response studies with ally alcohol and a mixture of
TCE, Thioacetamine, allyl alcohol, and chloroform. The same issues with interpretation present
for Soni et al. (1998) also apply to this study as well.
2.1.3. Okino et al. 1991
This study treated adult Wistar male rats (8 weeks of age) with TCE after being on a
liquid diet for 3 weeks and either untreated or pretreated with phenobarbital or ethanol. TCE
exposure was at 8000 ppm for 2 hours, 2000 or 8000 ppm for 2 hours, and 500 or 2000 ppm for
8 hours. Each group contained 5 rats. Livers from rats that were not pre-treated with either
ethanol or phenobarbital were reported to show only a few necrotic hepatocytes around the
central vein at 6 and 22 hours after 2 hours of 8000 ppm TCE exposure. At increased lengths
and/or concentrations of TCE exposure, the frequencies of necrotic hepatocytes in the
centrilobular area were reported to be increased but the number of necrotic hepatocytes was still
relatively low (out of-150 hepatocytes the percentages of necrotic pericentral hepatocytes were
0.2 ± 0.4, 0.3 ± 0.4, 2.7 ± 1.0, 0.2 ± 0.4, and 3.5 ± 0.4 % for control, 2000 ppm TCE for 2 hours,
8000 ppm TCE for 2 hours, 500 ppm TCE for 8 hours, and 2000 ppm TCE for 8 hours,
respectively). "Ballooned" hepatocytes were reported to be zero for controls and all TCE
treatments with the exception of 0.3 ± 0.6 % ballooned midzonal hepatocytes after 8000 ppm
TCE for 2 hours exposure. Microsomal protein (mg/g/liver) was increased with TCE exposure
concentration and duration, but not reported to be statistically significant (mg/g/liver microsomal
protein was 21.2 ± 4.3, 22.0 ± 1.5, 25.9 ± 1.3, 23.3 ± 0.8, and 24.1 ±1.0 for control, 2000 ppm
TCE for 2 hours, 8000 ppm TCE for 2 hours, 500 ppm TCE for 8 hours, and 2000 ppm TCE for
8 hours, respectively). The metabolic rate of TCE was reported to be increased after exposures
over 2000 ppm TCE (metabolic rate of TCE in nmol/g/liver/min was 29.5 ± 5.7, 51.3 ± 6.0, 63.1
± 16.0, 37.3 ± 3.3, and 69.5 ± 4.3 for control, 2000 ppm TCE for 2 hours, 8000 ppm TCE for 2
hours, 500 ppm TCE for 8 hours, and 2000 ppm TCE for 8 hours, respectively). However, the
cytochrome P450 content of the liver was not reported to increase with TCE exposure
concentration or duration. The liver/body weight ratios were reported to increase with all TCE
exposures except 500 ppm for 8 hours (the liver/ body weight ratio was 3.18 ± 0.15, 3.35 ± 0.10,
3.39 ± 0.20, 3.15 ± 0.10, and 3.57 ± 0.14 for control, 2000 ppm TCE for 2 hours, 8000 ppm TCE
02/12/09
E-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
for 2 hours, 500 ppm TCE for 8 hours, and 2000 ppm TCE for 8 hours, respectively). These
values represent 1.05-fold, 0.99-fold, 1.06-fold, and 1.12-fold of control in the 2000 ppm TCE
for 2 hours, 8000 ppm TCE for 2 hours, 500 ppm TCE for 8 hours, and 2000 ppm TCE for 8
hours treatment groups, respectively, with a statistically significant difference observed after 8
hours of 2000 ppm TCE exposure. Initial body weights and those 22 hours after cessation of
exposure were not reported, which may have affected liver weight gain. However, these data
suggest that TCE-related increases in metabolism and liver weight occurred as early as 22 hours
after exposures of this magnitude from 2 to 8 hours of TCE with little concurrent hepatic
necrosis.
Ethanol and phenobarbital pretreatment were reported to enhance TCE toxicity. In
ethanol-treated rats a few necrotic hepatocytes were reported to be around the central vein along
with hepatocellular swelling without pyknotic nuclei at 6 hours after TCE exposure with no
pathological findings in the midzonal or periportal areas. At 22 hours centrilobular hepatocytes
were reported to have a few necrotic hepatocytes and cell infiltrations around the central vein but
midzonal areas were reported to have ballooned hepatocytes with pyknotic nuclei frequently
accompanied by cell infiltrations. In phenobarbital treated rats 6 hours after TCE exposure,
centrilobular hepatocytes showed prenecrotic changes with no pathological changes reported to
be observed in the periportal areas. By 22 hours zonal necrosis was reported in centrilobular
areas or in the transition zone between centrilobular and periportal areas. Treatment with
phenobarbital or ethanol induced hepatocellular necrosis primarily in centrilobular areas with
phenobarbital having a greater effect (89.1 % ± 8.5% centrilobular necrosis) at the higher dose
and shorter exposure duration (8000 ppm x 2 hours) and ethanol having a greater effect (16.8 ±
5.3% centrilobular necrosis) at the lower concentration and longer duration of exposure (2000
ppm x 8 hours).
2.1.4. Nunes et al. 2001
This study was focused on the effects of TCE and lead co-exposure but treated male 75-
day old Sprague Dawley rats with 2000 mg/kg TCE for 7 days via corn-oil gavage (n= 10). The
rats ranged in weight from 293 to 330 g (-12%) at the beginning of treatment and were
pretreated with corn oil for 9 days prior to TCE exposure. TCE was reported to be 99.9% pure.
Although the methods section states that rats were exposed to TCE for 7 days, Table 1 of the
study reports that TCE exposure was for 9 days. The beginning body weights were not reported
specifically for control and treatment groups, but the body weights at the end of exposure were
reported to be 342 ± 18 g for control rats and 323 ± 3 g for TCE exposed rats, and that difference
(-6%) to be statistically significant. Because beginning body weights were not reported, it is
difficult to distinguish whether differences in body weight after TCE treatment were treatment
related or reflected differences in initial body weights. The liver weights were reported to be
12.7 ± 1.0 g in control rats and 14.0 ± 0.8 g for TCE treated rats with the % liver/body weight
ratios of 3.7% and 4.3 %, respectively. The increase in % liver/body weight ratio represents
1.16-fold of control and was reported to be statistically significant. However, difference in
initial body weight could have affected the magnitude of difference in liver weight between
02/12/09
E-13
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
control and treatment groups. The authors report no gross pathological changes in rats gavaged
with corn oil or with corn oil plus TCE but observed that one animal in each group had slightly
discolored brown kidneys. Histological examinations of "selected tissues" were reported to
show an increased incidence of chronic inflammation in the arterial wall of lungs from TCE-
dosed animals. There were no descriptions of liver histology given in this report for TCE-
exposed animals or corn-oil controls.
2.1.5. Tao et al. 2000
The focus of this study was to assess the affects of methionine on methylation and
expression of c-Jun and C-Myc in mouse liver after 5 days of exposure to TCE (1000 mg/kg in
corn oil) and its metabolites. Female 8-week old B6C3F1 mice (n= 4-6) were administered TCE
("molecular biology or HPLC grade") for 5 days with and without methionine (300 mg/kg i.p).
Data regarding % liver/body weight was presented as a figure. Of note is the decrease in
liver/body weight ratio by methionine treatment alone (~ 4.6 % liver/body weight for control and
~ 4.0 % liver/body weight for control mice with methionine or ~ 13% difference between these
groups). Neither initial body weights or body weights after exposure were reported by the
authors so that the reported effects of treatment could have reflected differences in initial body
weights of the mice. TCE exposure was reported to increase the % liver/body weight ratio to ~
5.8 % without methionine and to increase % liver/body weight ratio to ~ 5.7 % with methionine
treatment. These values represent 1.26-fold of control levels from TCE exposure without
methionine and 1.43-fold of control from TCE exposure with methionine. The number of
animals examined was reported to be 4-6 per group. The authors reported the differences
between TCE treated animals and their respective controls to be statistically significant but did
not examine the differences between controls with and without methionine. There were no
descriptions of liver histology given in this report for TCE-exposed animals or corn-oil controls.
2.1.6. Tucker et al. 1982
This study describes acute LD50, and 5- and 14-days studies of TCE in a 10% emulphor
solution administered by gavage. Screening level subchronic drinking water experiments with
TCE dissolved in 1% emulphor in mice were also conducted but with little detail reported. The
authors did describe the strains used (CD-I and ICR outbred albino) and that they are "weanling
mice," but the ages of the mice and their weights were not given. The TCE was described as
containing 0.004% diisopropylamine as the preservative and that the stabilizer had not been
found carcinogenic or overtly toxic. The authors report that "the highest concentration a mouse
would receive during these studies is 0.03 mg/kg/day". The main results are basically an LD50
study and a short term study with limited reporting for 4 and 6-month studies of TCE.
Importantly, the authors documented the loss of TCE from drinking water solutions (less than
20% of the TCE was lost during the 3 or 4 days in the water bottles at 1.0, 2.5 and 5.0 mg/ml
concentrations, but in the case of 0.1 mg/ml, up to 45% was lost over a 4-day period). The
authors also report that high doses of TCE in drinking water reduced palatability to such an
extent that water consumption by the mice was significantly decreased.
02/12/09
E-14
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The LD50 with 95% confidence were reported to be 2443 mg/kg (1839 to 3779) for
female mice and 2402 mg/kg (2065 to 2771) for male mice. However, the number of mice used
in each dosing group was not given by the authors. The deaths occurred within 24 hours of TCE
administration and no animals recovering from the initial anesthetic effect of TCE died during
the 14-day observation period. The authors reported that the only gross pathology observed was
hyperemia of the stomach of mice dying form lethal doses of TCE, and that mice killed at 14
days showed not gross pathology. In a separate experiment, male CD-I mice were exposed to
TCE by daily gavage for 14 days at 240 and 24 mg/kg. These two doses did not cause treatment
related deaths and body weight and "most" organ weights were reported by the authors to not be
significantly affected but the data was not shown. The only effect noted was an increased liver
weight which appeared to be dose dependent but was reported to be significant only at the higher
dose. The only significant difference found in hematology was s 5% lower hematocrit in the
higher dose group. The number of animals tested in this experiment was not give by the authors.
Male CD-I mice (n=l 1) were given TCE via gavage for 5 days (0.73 g/kg TCE twice on day 0,
1.46 g/kg twice on day 1, 2.91 g/kg twice on day 3 and 1.46 g/kg TCE on days 4 and 5) with
only 4 of 11 mice treated with TCE surviving.
In a subchronic study, male and female CD-I mice received TCE in drinking water at
concentrations of 0, 0.1, 1.0, 2.5 and 5 mg/ml in 1% emulphor, and a naive group received
deionized water. There were 140 animals of each sex in the naive group and in each treatment
group, except for 260 mice in the vehicle groups. Thirty mice of each sex and treatment were
selected for recording body weights for 6 months. The method of "selection" was not given by
the authors. These mice were weighed twice weekly and fluid consumption was measured by
weighing the six corresponding water bottles. The authors reported that male mice at the two
highest doses of TCE consumed 41 and 66 ml/kg/day less fluid over the 6 months of the study
than mice consuming vehicle only and that this same decreased consumption was also seen in the
high dose (5 mg/ml) females. They report that weight gain was not affected except at the high
dose (5mg/ml) and even though the weight gain for both sexes was lower than the vehicle
control group, it was not statistically significant but this data was not shown. The authors report
that gross pathological examinations performed on mice killed at 4 and 6 months were
unremarkable and that a number of mice from all the dosing regimens had liver abnormalities,
such as pale, spotty, or granular livers. They report that 2 of 58 males at 4 months, and 11 of 59
mice at 6 months had granular livers and obvious fatty infiltration, and that mice of both sexes
were affected. Animals in the naive and vehicle groups were reported to infrequently have pale
or spotty livers, but exhibit no other observable abnormalities. No quantitation or more detailed
descriptions of the incidence of or severity of effects were given in this report.
The average body weight of male mice receiving the highest dose of TCE was reported to
be 10% lower at 4 months and 11% lower at 6 months with body weights of female mice at the
highest dose also significantly lower. Enlarged livers (as percentage of body weight) were
observed after both durations of exposure in males at the three highest doses, and in females at
the highest dose. In the 4 month study, brain weights of treated females were significantly
increased when compared to vehicle control. However, the authors state that "this increase is
02/12/09
E-15
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
apparently because the values for the vehicle group were low, because the naive group was also
significantly increased when compared to vehicle control." "A significant increase in kidney
weight occurred at the highest dose in males at 6 months and in females, after both 4 and 6
months of TCE exposure. Urinalysis indicated elevated protein and ketone levels in high-dose
females and the two highest dose males after 6 months of exposure (data not shown)." The
authors describe differences in hematology to include "a decreased erythrocyte count in the high
dose males at 4 and 6 months (13% and 16%, respectively); decreased leukocyte counts,
particularly in the females at 4 months and altered coagulation values consisting of increased
fibrinogen in males at both times and shortened prothrombin time in females at 6 months (data
not shown). No treatment related effects were detected on the types of white cells in peripheral
blood." It must be noted that effects reported from this study may have also been related to
decreased water consumption, this study did included any light microscopic evaluation, and that
most of the results described are for data not shown. However, this study does illustrate the
difficulties involved in trying to conduct studies of TCE in drinking water, that the LD50s for
TCE are relatively high, and that liver weight increases were observed with TCE exposure as
early as few weeks and increased liver weight were sustained through the 6 month study period.
2.1.7. Goldsworthy and Popp 1987
The focus of this study was peroxisomal proliferation activity after exposure to a number
of chlorinated solvents. In this study 1000 mg/kg TCE (99+ % epoxide stabilizer free) was
administered to male F-344 rats (170-200 g or ~ 10% difference) and B6C3F1 (20-25 g or ~ 20%
difference) mice for 10 days in corn oil via gavage. The ages of the animals were not given. The
TCE-exposed animals were studied in two experiments (Experiments #1 and #3). In experiment
#2 corn oil and methyl cellulose vehicles were compared. Animals were killed 24 hours after the
last exposure. The authors did not show data on body weight but stated that the administration of
test agents (except WY-14,643 to rats which demonstrated no body weight gain) to rats and mice
for 10 days "had little or no effect on body weight gain." Thus, differences in initial body
weight between treatment and control groups, which could have affected the magnitude of TCE-
induced liver weight gain, were not reported. The liver/body weight ratios in corn oil gavaged
rats were reported to be 3.68 ± 0.06% and 4.52 ± 0.08 % after TCE treatment which represented
1.22-fold of control (n = 5). CN-insenstive palmitoyl CoA oxidation (PCO) was reported to be
1.8-fold increased after TCE treatment in this same group. In B6C3F1 mice the liver/body
weight ratio in corn oil gavaged mice was reported to be 4.55 ± 0.13 % and 6.83 ± 0.13 % after
TCE treatment which represented 1.50-fold of control (n = 7). CN-insensitive PCO activity was
reported to be 6.25-fold of control after TCE treatment in this same group. The authors report no
effect of vehicle on PCO activity but do not show the data nor discuss any effects of vehicle on
liver weight gain. Similarly the results for experiment #3 were not shown nor liver weight
discussed with the exception of PCO activity reported to be 2.39-fold of control in rat liver and
6.25-fold of control for mouse liver after TCE exposure. The number of animals examined in
experiment #3 was not given by the authors or the variation between enzyme activities.
However, there appeared to be a difference in PCO activity experiments #1 and #3 in rats.
02/12/09
E-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
There were no descriptions of liver histology given in this report for TCE-exposed animals or
corn-oil controls.
2.1.8. Elcombe et al. 1985
In this study, preservative free TCE was given via gavage to rats and mice for 10
consecutive days with a focus on changes in liver weight, structure, and hepatocellular
proliferation induced by TCE. Male Alderly Park rats (Wistar derived) (180-230 g), male
Obsborne-Mendel rats (240-280 g), and male B6C3F1 or male Alderly Park Mice (Swiss)
weighing 30 to 35 g were administered 99.9% pure TCE dissolved in corn oil via gavage. The
ages of the animals were not given by the authors. The animals were exposed to 0, 500, 1000 or
1500 mg/kg body wt TCE for 10 consecutive days. The number of mice and rats varied widely
between experiments and treatment groups and between various analyses. In some experiments
animals were injected with tritiated thymidine approximately 24 hours following the final dose
of TCE and killed one hour later. The number of hepatocytes undergoing mitosis was identified
in 25 random high-power fields (X40) for each animal with 5000 hepatocyte per animal
examined. There was no indication by the authors that zonal differences in mitotic index were
analyzed. Sections of the liver were examined by light and electron microscopy by conventional
staining techniques. Tissues selected for electron microscopy included central vein and portal
tract so that zonal differences could be elucidated. Morphometric analysis of peroxisomes was
performed "according to general principles of Weibel et al (1964) on electronphotomicrographs
from pericentral hepatocytes." DNA content of samples and peroxisomal enzyme activities were
determined in homogenized liver (catalase and PCO activity).
The authors reported that TCE treatment had no significant effect on body weight gain
either strain of rat or mouse during the 10 days exposure period. However, marked increases
(up to 175% of control value) in the % liver/body weight ratio were observed in TCE treated
mice. Smaller increases (up to 130% of control) in relative liver weight were observed in TCE-
treated rats. No significant effects of TCE on hepatic water content were seen so that the liver
weight did not represent increased water retention.
An interesting feature of this study was that it was conducted in treatment blocks at
separate times with separate control groups of mice for each experimental block. Therefore,
there were 3 control groups of B6C3F1 mice (n= 10 for each control group) and 3 control groups
for Alderly Park (n= 9 to 10 for each control group) mice that were studied concurrently with
each TCE treatment group. However, the % liver/body weight ratios were not the same between
the respective control groups. There was no indication from the authors as to how controls were
selected or matched with their respective experimental groups. The authors did not give liver
weights for the animals so the actual changes in liver weights are not given. The body weights
of the control and treated animals were also not given by the authors. Therefore, if there were
differences in body weight between the control groups or treatment groups, the liver/body weight
ratios could also have been affected by such differences. The percentage increase over control
could also have been affected by what control group each treatment group was compared to.
02/12/09
E-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
There was a difference in the mean % liver/body weight ratio in the control groups which ranged
from 4.32 to 4.59% in the B6C3F1 mice (~ 6% difference) and from 5.12 to 5.44 % in the
Alderly Park mice (~ 6% difference). The difference in average % liver/body weight ratio for
untreated mice between the two strains was ~ 16%. Because the ages of the mice were not
given, the apparent differences between strains may have been due to both age or to strain. After
TCE exposure, the mean % liver/body weight ratios were reported to be 5.53% for 500 mg/kg,
6.50% for 1000 mg/kg and 6.74% for 1500 mg/kg TCE exposed B6C3F1 mice. This resulted in
1.20-fold, 1.50-fold and 1.47-fold values of control in % liver weight/body weight for B6C3F1
mice. For Alderly Park mice the % liver/body weight ratios were reported to be 7.31, 8.50 and
9.54 % for 500, 1000, and 1500 mg/kg TCE treatment, respectively. This resulted in 1.43-fold,
1.56-fold and 1.75-fold of control values. Thus, there appeared to be more of a consistent dose-
related increase in liver/body weight ratios in the Alderly Park mice than the B6C3F1 mice after
TCE treatment. However, the variability in control values may have distorted the dose-response
relationship in the B6C3F1 mice. The Standard deviations for liver/body weight ratio were as
much as 0.52 % for the treated B6C3F1 mice and 0.91% for the Alderly Park treated mice. In
regard to the correspondence of the magnitude of the TCE-induced increases in % liver/body
weight with the magnitude of difference in TCE exposure concentrations, in the B6C3F1 mice
the increases were similar (~ 2-fold) between the 500 mg/kg and 1000 mg/k TCE exposure
groups. For the Alderly Park mice, the increases in TCE exposure concentrations were slightly
less than the magnitude of increases in % liver/body ratios between all of the concentrations (i.e.,
-1.3 fold of control vs. 2-fold for 500 and 1000 mg/kg TCE dose and 1.3-fold of control vs. 1.5-
fold for the 1000 and 1500 mg/kg TCE dose).
The DNA content of the liver varied greatly between control animal groups. For B6C3F1
mice it ranged from 2.71 to 2.91 mg/g liver. For Alderly Park mice it ranged from 1.57 to 2.76
mg/g liver. The authors do not discuss this large variability in baseline levels of DNA content.
The DNA content in B6C3F1 mice was mildly depressed by TCE treatment in a non-dose
dependent manner. DNA concentration decrease from control ranged from 20-25% between all
three TCE exposure levels in B6C3F1 mice. For Alderly Park mice there was also non-dose
related decrease in DNA content from controls that ranged from 34% to 18%. Thus, the extent
of decrease in DNA content of the liver from TCE treatment in B6C3F1 mice was similar to the
variability between control groups. The lack of dose-response in apparent treatment related
effect in B6C3F1 mice and especially in the Alderly Park mice was confounded by the large
variability in the control animals. The changes in liver weight after TCE exposure for the AP
mice did not correlate with changes in DNA content further, raising doubt about the validity of
the DNA content measures. However, a small difference in DNA content due to TCE treatment
in all groups was reported for both strains and this is consistent with hepatocellular hypertrophy.
The reported results for incorporation of tritiated thymidine in liver DNA showed large
variation in control groups and standard deviations that were especially evident in the Alderly
Park mice. For B6C3F1 mice, mean control levels were reported to range from 5,559 to 7,767
dpm/mg DNA with standard deviations ranging from 1,268 to 1,645 dpm/mg DNA. In Alderly
Park mice mean control levels were reported to range from 6,680 to 10,460 dpm/mg DNA with
02/12/09
E-18
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
standard deviations ranging from 308 to 5,235 dpm/mg DNA. For B6C3F1 mice, TCE treatment
was reported to induce an increase in tritiated thymidine incorporation with a very large standard
deviation, indicating large variation between animals. For 500 mg/kg TCE treatment group the
values were reported as 12,334 ± 4,038, for 1000 mg/kg TCE treatment group 21,909 ± 13,386,
and for 1500 mg/kg treatment TCE group 26,583 ± 10,797 dpm/mg DNA. In Alderly Park mice
TCE treatment was reported to give an increase in tritiated thymidine incorporation also with a
very large standard deviation. For 500 mg/kg TCE the values were reported as 19,315 ± 12,280,
for 1000 mg/kg TCE 21,197 ± 8,126 and for 1500 mg/kg TCE 38,370 ± 13,961. As a percentage
of concurrent control, the increase in tritiated thymidine was reported to be 2.11-fold, 2.82-fold,
and 4.78-fold of control in B6C3F1 mice, and 2.09-fold, 2.03-fold, and 5.74-fold of control in
Alderly Park mice. Accordingly, the change in tritiated thymidine incorporation did show a
treatment related increase but not a dose-response. Similar to the DNA content of the liver, the
large variability in measurements between control groups and variability between animals limit
quantitative interpretation of this data. The increase in tritiated thymidine, seen most
consistently only at the highest exposure level in both strains of mice, could have resulted from
either a change in ploidy of the hepatocytes or cell number. However, the large change in
volume in the liver (75%) in the Alderly Park mice, could not have resulted from only a 4-fold of
control in cell proliferation even if all tritiated thymidine incorporation had resulted from
changes in hepatocellular proliferation. As mentioned in Section 1.1 above, the baseline level of
hepatocellular proliferation in mature control mice is very low and represents a very small
percentage of hepatocytes.
In the experiments with male rats, the same issues discussed above, associated with the
experimental design, applied to the rat experiments with the additional concern that the numbers
of animals examined varied greatly (i.e., 6 to 10) between the treatment groups. In Obsborne-
Mendel rats the control liver/body weight ratio was reported to vary from 4.26 to 4.36% with the
standard deviations varying between 0.22 to 0.21%. For the Alderly Park rats the liver/body
weight ratios were reported to vary between 4.76 and 4.96% (in control groups with standard
deviations varying between 0.24 to 0.47%. TCE treatment was reported to induce a dose-related
increase in liver/body weight ratio in Obsborne-Mendel rats with mean values of 5.16, 5.35 and
5.53%) in 500, 1000 and 1500 mg/kg TCE treated groups, respectively. This resulted in 1.18-
fold, 1.26-fold, and 1.30-fold values of control. In Alderly Park rats, TCE treatment was
reported to resulted increased liver weights of 5.45, 5.83 and 5.65 % for 500, 1000, and 1500
mg/kg TCE respectively. This resulted in 1.14-fold, 1.17-fold, and 1.17-fold values of control.
Again, the variability in control values may have distorted the nature of the dose-response
relationships in Alderly Park rats. TCE treatment was reported to result in standard deviations
that ranged from 0.31 to 0.48% for OM rats and 0.24 to 0.38% for Alderly Park rats. What is
clear from these experiments is that TCE exposure was associated with increased liver/body
weight in rats.
The reported mean hepatic DNA concentrations and standard deviations varied greatly in
control rat liver as it did in mice. The variation in DNA concentration in the liver varied more
between control groups than the changes induced by TCE treatment. For Obsborne-Mendel rats
02/12/09
E-19
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the mean control levels of mg DNA/g liver were reported to range from 1.99 to 2.63 mg
DNA/liver with standard deviations varying from 0.17 to 0.33 mg DNA/g. For Alderly Park rats
the mean control levels of mg DNA/ g liver were reported to be 2.12 to 3.16 mg DNA/g with
standard deviation ranging from 0.06 to 1.04 mg DNA/g. TCE treatment decreased the liver
DNA concentration in all treatment groups. For Obsborne-Mendel rats the decrease ranged from
8 to 13% from concurrent control values and for Alderly Park rats the decrease ranged from 8 to
17%. There was no apparent dose-response in the decreases in DNA content with all TCE
treatment levels giving a similar decrease from controls and the same limitations discussed above
for the mouse data apply here. The magnitude of increases in liver/body ratios shown by TCE
treatment were not correlated with the changes in DNA content. However, as with the mouse
data, the small differences in DNA content due to TCE treatment in all groups and in both strains
was consistent with hepatocellular hypertrophy.
Incorporation of tritiated thymidine was reported to be even more variable between
control groups of rats than it was for mice and was reported to be especially variable between
control groups (i.e., 2.7-fold difference between control groups within strain) and differed
between the strains (average of 2.5 fold between strains). For Obsborne-Mendel rats the mean
control levels were reported to range from 13,315 to 33,125 dpm/mg DNA, while for Alderly
Park rats tritiated thymidine incorporation ranged from 26,613 to 69,331 dpm/mg DNA for
controls. The standard deviations were also very large (i.e., for control groups of Obsborne-
Mendel rats they were reported to range from 8,159 to 13,581 dpm/mg DNA, while for Alderly
Park rats they ranged from 9,992 to 45,789 dpm/mg DNA). TCE treatment was reported to
induce increases over controls of 110, 118, and 106% for 500, 1000, and 1500 mg/kg TCE
exposed groups, respectively, in Obsborne-Mendel rats with large standard deviations for these
treatment groups as well. In Alderly Park rats the increases over controls were reported to be
206, 140 and 105% for 500, 1000, and 1500 mg/kg TCE, respectively. In general, these data do
indicate that TCE treatment appeared to give a mild increase in tritiated thymidine incorporation
but the lack of dose-response can be attributable to the highly variable measurements of tritiated
thymidine incorporation in control animal groups. The variation in the number of animals
examined between groups and small numbers of animals examined additionally decrease the
likelihood of being able to discern the magnitude of difference between species- or strain-related
effects for this parameter. Again, given the very low level of hepatocyte turnover in control rats,
this does not represent a large population of cells in the liver that may be undergoing
proliferation and cannot be separated from changes in ploidy.
The authors report that the reversibility of these phenomena was examined after the
administration of TCE to Alderly Park mice for 10 consecutive days. Effects upon liver weight,
DNA concentration, and tritiated thymidine incorporation 24 and 48 hours after the last dose of
TCE were reported to still be apparent. However, 6 days following the last dose of TCE, all of
these parameters were reported to return to control values with the authors not showing the data
to support this assertion. Thus, cessation of TCE exposure would have resulted in a 75%
reduction in liver weight by one week in mice exposed to the highest TCE concentration.
02/12/09
E-20
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Analyses of hepatic peroxisomal enzyme activities were reported for catalase and P-
oxidation (PCO activity) following administration of TCE to B6C3F1 mice and Alderly Park rats
exposed to 1000 mg/kg TCE for 10 days. The authors only used 5 control and 5 exposed
animals for these tests. An 8-fold of control value for PCO activity and a 1.5-fold of control
value for catalase activity were reported for B6C3F1 mice exposed to 1000 mg/kg TCE. In the
Alderly Park rats no significant changed occurred. It is unclear which mice or rats were selected
from the previous experiments for these analyses and what role selection bias may have played
in these results. The reduced number of animals chosen for this analysis also reduces the power
of the analysis to detect a change. In rats there was a reported 13% increase in PCO, however
the variation between the TCE treated rats was more than double that of the control animals in
this group and the other limitations described above limit the ability to detect a response. There
was no discussion given by the authors as to why only one dose was tested in half of the animals
exposed to TCE or why the strain with the lowest liver weight change due to TCE exposure was
chosen as the strain to test for peroxisomal proliferative activity.
The authors provided a description of the histopathology at the light microscropy level in
B6C3F1 mice, Alderly Park mice, Osborne-Mendel rats, and Alderly Park rats, but did not
provide a quantitative analysis or specific information regarding the variability of response
between animals within groups. There appeared to be 20 animals examined in the 1000 mg/kg
TCE exposed group of B6C3F1 mice but no explanation as to why there were only 10 animals
examined in analyses for liver weight changes, DNA concentration and tritiated thymidine
incorporation. There was no indication by the authors regarding how many rats were examined
by light microscopy.
Apart from a few inflammatory foci in occasional animals, H&E section from B6C3F1
control mice were reported to show no abnormalities. The authors suggest that this is a normal
finding in the livers of mice kept under "non-SPF conditions". A stain for neutral lipid was
reported to not be included routinely in these studies, but subsequent electron microscopic
examination of lipid to show increases in the livers of corn-oil treated control animals. The
individual fat droplets were described as "generally extremely fine and are not therefore
detectable in conventionally process H&E stained sections, since both glycogen and lipid are
removed during this procedure." Thus, this study documents effects of using corn oil gavage in
background levels of lipid accumulation in the liver.
The finding of little evidence of gross hepatotoxicity in TCE-treated mice was reported,
even at a dose of 1500 mg/kg. Specifically, "Of 19 animals examined receiving 1500 mg/kg
body weight TCE, only 6 showed any evidence of hepatocyte necrosis, and this pathology was
restricted to single small foci or isolated single cells, frequently occurring in a subcapsular
location. Examination of 20 animals receiving 1000 mg/kg body wt TCE demonstrated no
hepatocyte necrosis. Of 20 animals examined receiving 500 mg/kg body wt TCE, 1 showed
necrosis of single isolated hepatocytes; however, this change was not a treatment-related
finding." TCE-treated mice were reported to show "a change in staining characteristic of the
hepatocytes immediately adjacent to the central vein of the hepatocyte lobules, giving rise to a
02/12/09
E-21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
marked 'patchiness' of the liver sections. Often this change consisted of increased eosinophilia
of the central cells. There was some evidence of cell hypertrophy in the centrilobular regions.
These changes were evident in most of the TCE treated animals, but there was a dose-related
trend, relatively few of the 500 mg/kg animals being affected, while the majority of the 1500
mg/kg animals showed central change. No other significant abnormalities were seen in the liver
of TCE treated mice compared to controls apart from occasional mitotic figures and the
appearance of isolated nuclei with an unusual chromatin pattern. This pattern generally
consisted of a course granular appearance with a prominent rim of chromatin around the
periphery of the nucleus. These nuclei may have been in the very early stages of mitosis.
Similar changes were not seen in control mice." The authors briefly commented on the findings
in the Alderly Park mice stating that "H& E sections from Alderly Park mice gave similar results
as for B6C3F1 mice. No evidence of hepatotoxicity was seen at a dose of 500 mg/kg body wt
TCE. However, a few animals at the higher doses showed some necrosis and other degenerative
changes. This change was very mild in nature, being restricted to isolated necrotic cells or small
foci, frequently in subcapsular position. Hypertrophy and increased eosinophilia were also
noticed in the centrilobular regions at higher doses." Thus, from the brief description given by
the authors, the centrilobular region is identified as the location of hepatocellular hypertrophy
due to TCE exposure in mice, and for it to be dose-related with little evidence of accompanying
hepatotoxicity.
The description of histopathology for rats was even more abbreviated than for the mouse.
H& E sections from Osborne-Mendel rats showed that "livers from control rats contained large
quantities of glycogen and isolated inflammatory foci, but were otherwise normal. The majority
of rats receiving 1500 mg/kg body weight TCE showed slight changes in centrilobular
hepatocytes. The hepatocytes were more eosinophilic and contained little glycogen. At lower
doses these effects were less marked and were restricted to fewer animals. No evidence of
treatment-related hepatotoxicity (as exemplified by single cell or focal necrosis) was seen in any
rat receiving TCE. H& E sections from Alderly Park Rats showed no signs of treatment-related
hepatotoxicity after administration of TCE. However, some signs of dose-related increase in
centrilobular eosinophilia were noted." Thus both mice and rats exhibited pericentral
hypertrophy and eosinophilia as noted from the histopathological examination.
The study did report a quantitative analysis of the effects of TCE on the number of
mitotic figures in livers of mice. Few if any control mice exhibited mitotic figures. But the
authors report "a considerable increase in both the numbers of figures per section was noted after
administration of TCE". The numbers of animals examined for mitotic figures ranged from 75
(all control groups were pooled for mice) to 9 in mice, and ranged from 15 animals in control rat
groups to as low as 5 animals in the TCE treatment groups. The range of mitotic figures found in
25 high-power fields was reported and is equivalent to the number of mitotic figures per 5000
hepatocytes examined in random fields". Thus the predominance of mitotic figures in any zone
of the liver cannot be ascertained.
02/12/09
E-22
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For B6C3F1 mice the number of animals with mitotic figures was reported to be 0/75,
3/20, 7/20 and 5/20 for control, 500, 1000, and 1500 mg/kg TCE exposed mice, respectively.
The range of the number of mitotic figures seen in 5000 hepatocytes was reported to be 0, 0-1, 0-
5, 0-5 for those same groups with group means of 0, 0.15 ± 0.36, 0.6 ± 1.1 and 0.5 ± 1.2. These
results demonstrate a very small and highly variable response due to TCE treatment in B6C3F1
mice in regard to mitosis. Thus, the highest percentage of cells undergoing mitosis within the
window of observation would be on average 0.012% with a standard deviation twice that value.
The data presented for mitotic figures also indicated no differences in results between 1000 and
1500 mg/kg treated B6C3F1 mice in regard to mitotic figure detection. However, the tritiated
thymidine incorporation data indicated that thymidine incorporation was ~2-fold greater at 1500
than 1000 mg/kg TCE in B6C3F1 mice. For Alderly Park mice, the number of animals with
mitotic figures was reported to be 1/15, 0/9, 4/9 and 2/9 for control, 500, 1000, and 1500 mg/kg
TCE exposed mice. The range of the number of mitotic figures seen in 5000 hepatocytes was 0-
1, 0, 0-2, 0-1 for those same groups with group means of 0.06 ± 0.25, 0. 0.7 ± 0.9, and 0.2 ± 0.4.
These results reveal the detection of at the most 2 mitotic figure in 5000 hepatocytes for any
mouse an any treatment group and no dose-related increased after TCE treatment in Alderly Park
mice. Thus, the highest percentage of cells with a mitotic figure would be on average 0.014%
with a standard deviation twice that value. The small number of animals examined reduces the
power of the experiment to draw any conclusions as to a dose-response. Similar to the B6C3F1
mice, there did not appear to be concordance between mitotic figure detection and thymidine
incorporation for Alderly park mice. Thymidine incorporation showed a 2- fold increase over
control for 500 and 1000 mg/kg TCE and a 5.7- fold increase for 1500 mg/kg TCE treated
animals. However, in regard to mitotic figure detection, there were fewer mitotic figures in 500
mg/kg TCE treated mice than controls, and fewer animals with mitotic figures and fewer
numbers of figures in the 1500 mg/kg dose than the 1000 mg/kg exposed group. The
inconsistencies between mitotic index data and thymidine incorporation data in both strains of
mice suggests that either thymidine incorporation is representative of only DNA synthesis and
not mitosis, an indication of changes in ploidy rather than proliferation, or that this experimental
design is incapable of discerning the magnitude of these changes accurately. Data from both
mouse strains show very little if any hepatocyte proliferation due to TCE exposure with the
mitotic figure index data having that advantage of being specific for hepatocytes and to not to
also include nonparenchymal cells or inflammatory cells in the liver.
The results for rats were similar to those for mice and even more limited by the varying
and low number of animals examined. For Osborne-Mendal rats the number of animals with
mitotic figures were reported to be 8/15, 2/9, 0/7 and 0/6 for control, 500, 1000, and 1500 mg/kg
TCE exposed rats groups, respectively, with the range of the number of mitotic figures seen in
5000 hepatocytes to be 0-8, 0-3, 0, and 0. The group mean was 1.5 ± 2.0, 0.4 ± 1.0, 0 and 0 for
these groups. It would appear from these results that there are fewer mitotic figures after TCE
treatment with the highest percentage of cells undergoing mitosis to be on average 0.03% in
control rats. However, thymidine incorporation studies show a modest increase at all treatment
levels over controls in Osborne Mendel rats rather than a decrease from controls. For Alderly
Park rats the number of animals with mitotic figures was reported to be 13/15, 5/9, 9/9 and 4/9
02/12/09
E-23
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
for control, 500, 1000, and 1500 mg/kg TCE exposed rat groups with the range of the number of
mitotic figures seen in 5000 hepatocytes to be 0-26, 0-5, 1-7, and 0-9. The group mean was 7.2
± 4.7, 1.6 ± 4.3, 3.8 ± 3.4 and 1.8 ± 2.9 for these groups. It would appear that there are fewer
mitotic figures after TCE treatment with the highest percentage of cells to an average of 0.14%
in control rats. However, thymidine incorporation studies show 2-fold greater level at 500 mg/kg
TCE than for control animals and a 40% and 5% increase at 1000 mg/kg and 1500 mg/kg TCE
exposure groups, respectively. Similar to the results reported in mice, results in both rat strains
show an inconsistency in mitotic index and thymidine incorporation. The control rats appear to
have a much greater mitotic index than any of the mouse groups (treated or untreated) or the
TCE treatment groups. However, it is the mice that were exhibiting the largest increased in liver
weight after TCE exposure. By either thymidine incorporation or mitosis, these data do provide
a consistent result that at 10 days of exposure very little sustained hepatocellular proliferation is
occurring in either mouse or rat and neither is correlated well with the concurrent changes in
liver weight observed from TCE exposure.
This study provided a qualitative discussion and quantitative analysis of structural
changes using electron microscopy. The qualitative discussion was limited and included
statements about increased observances without quantitative data shown other than the
morphometric analysis. The authors reported that "the ultrastructure of control mouse liver was
essentially normal, although mild dilatation of RER and SER was a frequent finding. Lipid
droplets were also usually present in the cell cytoplasm. The ultrastructural changes seen in
mouse liver following administration of up to 1500 mg/kg body wt TCE for 10 days were
essentially similar in the B6C3F1 mouse and the Alderly Park mouse. The most notable change
in both strains of mouse was a dramatic increase in the number of peroxisomes. This change was
only apparent in the cells immediately surrounding the central veins. Peroxisome proliferation
was not noticeable in periportal cells. The induced peroxisomes were generally small and very
electron dense and frequently lacked the characteristic nucleoid core found in peroxisomes of
control livers." The authors conclude that "morphometric analysis showed evidence of a dose-
related response, peroxisomal induction appearing to reach a maximum at 1000 mg/kg in
B6C3F1 mice." "Lipid was increased in the livers of treated mice at all doses and was present
both as free droplets in the cytoplasm and as liposomes (small lipid droplets in ER cisternae).
The centrilobular cell which showed the greatest increase in numbers of peroxisomes showed no
evidence of this lipid accumulation: fatty change was more prominent in those cells away from
the central vein (i.e., zone 2 of the liver acinus). Accumulation of lipid, particularly in
liposomes, was less marked in Alderly Park mouse than in B6C3F1 mouse. Mild proliferation of
smooth endoplasmic reticulum was seen in both strains and both rough and smooth endoplasmic
reticulum was generally more dilated than in control mice." Electron microscopic results for rat
liver were reported "to show similar changes in Osborne-Mendel and Alderly Park rat treated
with TCE." "Rats receiving either 1000 or 1500 mg/kg TCE for 10 days generally showed mild
proliferation of SER in centrilobular hepatocytes. The cisternae of RER were frequently dilated,
giving rise to a rather disorganized appearance in contrast to the parallel stacks seen in control
livers, although no detachment of ribosomes was evident. The SER was also dilated. In contrast
to mice, peroxisomes were only very slightly and not significantly, increased in the liver of TCE
02/12/09
E-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
-treated rats. Morphometric analysis confirmed this observation, with the volume density of
peroxisomes in the cytoplasm of centrilobular hepatocytes being only slightly increased in rats of
both strains receiving 1000 or 1500 mg/kg body wt TCE." "Lipid droplets were occasionally
increased in some livers obtained from rats receiving TCE, but the degree of fatty change
generally appeared similar to that found in control rats receiving corn oil. There were no
changes in membrane -bound liposomes, other organelles, or Golgi condensing vesicles.
Centrilobular glycogen was somewhat depleted in male rats receiving 1500 mg/kg TCE.
Periportal cells were ultrstructually normal in all rats."
For the morphometric analysis, the number of mice examined ranged from 7 in the
control group to 8 in the 1500 mg/kg TCE exposed group. The authors did not indicate which
control animals were used for the morphometric analysis from the 75 animals examined for
mitotic index, the 20 examined by light microscopy, or the 30 mice used as concurrent controls
in the liver weight, DNA concentration and tritiated thymidine incorporation studies. The
authors stated that morphometry was performed on three randomly selected photomicrographs
from each of three randomly selected pericentral hepatocytes for each animal (i.e., nine
photomicrographs per animal). A mean value representing the exposure group was reported with
the variability between photomitographs per animal or the variation between animals unclear.
The morphometric analysis did not examine all treatment groups (e.g., only the control and 500
mg/kg TCE group were examined in Alderly Park mice). The % cytoplasmic volume of the
peroxisomal compartment (x ± SD) was reported to be 0.6 ± 0.6 % for controls, 4.8 ±3.3 % for
500 mg/kg TCE, 6.7 ± 1.9 % for 1000 mg/kg TCE, and 6.4 ± 2.5% for 1500 mg/kg TCE in
B6C3F1 mice. In Alderly Park mice, only 12 control and 12 500 mg/kg TCE exposed mice
were examined and, similarly, their selection criteria was not given. The % cytoplasmic volume
of the peroxisomal compartment was 1.2 ± 0.4% for control and 4.7 ± 2.8 % for 500 mg/kg TCE
exposed mice. For Osborne-Mendel rats control rats were reported to have a % cytoplasmic
volume of the peroxisomal compartment for control rats (n=9) of 1.8 ± 0.4 %, 1000 mg/kg TCE
(n= 5) 2.3 ±1.6 %, and for 1500 mg/kg exposed rats (n= 7) 2.3 ± 2.0%. For Alderly Park rats
only two groups were examined (control and 1000 mg/kg TCE exposure). The % cytoplasmic
volume of the peroxisomal compartment for control rats (n= 15) was reported to be 1.8 ± 0.8 %
and for 1000 mg/kg TCE (n= 16) to be 2.4 ± 1.2 %. The varying numbers of animals examined,
the varying and inconsistent number of treatment groups examined, the limited number of
photomitographs per animal, and the potential selection bias for animals examined make
quantitative conclusions regarding this analysis difficult. Although control levels differed by a
factor of 2 between the two strains of mice examined, as well as the number of control animals
examined (7 vs. 12), it appears that the 500 mg/kg TCE exposed B6C3F1 and Alderly Park mice
had similar percentages of peroxisomal compartment in the pericentral cells examined (~ 4.8 %).
There also appeared to be little difference between 1000 mg/kg TCE treated Osborne-Mendel
and Alderly Park rats for this parameter (-2.4%). Although few animals were examined, there
was little difference reported between 500, 1000 and 1500 mg/kg TCE exposure groups in regard
to percentages of peroxisomal compartment in B6C3F1 mice (4.8 - 6.1%). For the few rats of
the Osborne-Mendel strain examined, there also did not appear to be a difference between 1000
and 1500 mg/kg TCE exposure for this parameter (2.3%).
02/12/09
E-25
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Based on peroxisome compartment volume data, one would expect there to be little
difference between TCE exposure groups in mice or rats in regard to enzyme activity or other
"associated events." However such comparisons are difficult due to limited power to detect
differences and the possibility of bias in selection of animals in differing assays. For the
B6C3F1 mice, only 5 animals per group were examined for enzyme analysis, 7 to 8 for
morphometric analysis, 75 animals in control and 20 animals in 1000 mg/kg TCE exposed
groups for mitotic figure identification, and 10 animals per group for thymidine incorporation.
Since only a few animals were tested for enzyme activity the comparison between peroxisomal
compartment volume and that parameter is very limited. There was a reported 47% increase in
catalase activity between control (n = 5) and 1000 mg/kg TCE exposed B6C3F1 mice (n = 5) and
7.8- fold increase in PCO activity. The % peroxisome compartment was reported to be 10.6-fold
greater (0.6 vs. 6.4%). However, the B6C3F1 control % volume of peroxisomal compartment
was reported to be half that of the AP mouse control. An accurate determination of the
quantitative differences in peroxisomal proliferation would be dependent on an accurate and
stable control value. For Alderly Park rats there was an 8% decrease in catalase activity between
control (n = 5) and 1000 mg/kg TCE exposed rats (n = 5), and a 13% increase in PCO activity.
The % peroxisome compartment was reported to be 33% greater in the TCE-exposed than
control group. Thus, for the very limited data that was available to compare peroxisomal
compartment volume with enzyme activity, there was consistency in result.
However, were such increases in peroxisomes associated with other events reported in
this study? Mouse peroxisome proliferation associated enzyme activities in B6C3F1 mice at
1000 mg/kg TCE were reported to be 8-fold over control values in mice after 10 days of
treatment. However, this increase in activity was not accompanied by a similar increase in
thymidine incorporation (2.8- fold of control) or concordant with increases in mitotic figures
(7/20 mice having any mitotic figures at all with a range of 0 - 5 and a mean of 0.014% of cells
undergoing mitosis for 1000 mg/kg TCE vs. 0 for control). Although results reported in the rat
showed discordance between thymidine incorporation and detection of mitotic figures, there was
also discordance with these indices and those for peroxisomal proliferation. In comparison to
controls, there was a reported 13% increase in PCO activity in Alderly park rats exposed to 1000
mg/kg TCE, a group mean of mitotic figures half that in the TCE treated animals vs. controls,
and increase in thymidine incorporation of 40%. Thus, these results are not consistent with TCE
induction of peroxisome enzyme activity to be correlated with hepatocellular proliferation by
either mitotic index or thymidine incorporation. Thymidine incorporation in liver DNA seen
with TCE exposure also did not correlate with mitotic index activity in hepatocytes and suggests
that this parameter may be a reflection of polyploidization rather than hepatocyte proliferation.
More importantly, these data show that hepatocyte proliferation, indicated by either measure, is
confined to a very small population of cells in the liver after 10 days of TCE exposure.
Hepatocellular hypertrophy in the centrilobular region appears to be responsible for the liver
weight gains seen in both rats and mice rather than increases in cell number. These results at 10
days do not preclude the possibility that a greater level of hepatocyte proliferation did not occur
earlier and then had subsided by 10 days, as is characteristic of many mitogens. Thymidine
incorporation represents the status of the liver at one time point rather than over a period of
02/12/09 E-26
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
whole week and thus would not capture the earlier bouts of proliferation. However, there is no
evidence of a sustained proliferative response, as measured at the 10-day time period, in
hepatocytes in response to TCE indicated from this data.
In regards to weight gain, although the volume of the peroxisomal compartment was
reported to be similar at 500 mg/kg TCE in B6C3F1 and Alderly Park mice (4.3%), the liver
weight./body weight gain in comparison to control was 20% higher in B6C3F1 mice vs. 43%
higher in Alderly Park mice after 10 days of exposure. The liver/body weight ratio was 5.53 %
in the B6C3F1 mice and 7.31% in the Alderly Park mice at 500 mg/kg TCE for 10 days.
Similarly, although the peroxisomal compartment was similar at 1000 mg/kg TCE in Osborne-
Mendel (2.3%) and Alderly Park rats (2.4%), the liver weight/body weight gain was 26% in
Osborne-Mendel rats but 17% in Alderly Park rats at this level of TCE exposure. The liver/body
weight ratio was 5.35 % in the Osborne-Mendel rats and 5.83% in the Alderly Park mice at 1000
mg/kg TCE for 10 days. Although there are several limitations regarding the quantitative
interpretation of the data, as discussed above, the data suggest that liver weight and weight gain
after TCE treatment was not just a function of peroxisome proliferation. This study does clearly
demonstrate TCE-induced changes at the lowest level tested in several parameters without
toxicity and without evidence of regenerative hyperplasia or sustained hepatocellular
proliferation. In regards to susceptibility to liver cancer induction in more susceptible (B6C3F1)
vs. less susceptible (Alderly Park/Swiss) strains of mice (Maltoni et al., 1988), there was a
greater baseline level of liver weight/body weight ratio change, a greater baseline level of
thymidine incorporation as well as greater responses for those endpoints due to TCE exposure in
the "less susceptible" strain. However, both strains showed a hepatocarcinogenic response to
TCE induction and the limitations of being able to make quantitative conclusions regarding
species and strain susceptibility TCE toxicity from this study have been described in detail
above.
2.1.9. Dees and Travis 1993
The focus of this study was to evaluate the nature of DNA synthesis induced by TCE
exposure in mice. The mitotic rate of liver cells was extrapolated using tritiated thymidine
uptake into DNA of male and female mice treated with HPLC grade (99+pure) TCE. Male and
female hybrid B6C3F1 mice 8 weeks of age (male mice weighed 24-27 g (~ 12% difference)
and females weighing 18-21 g (~ 14% difference) were dosed orally by gavage for 10 days with
100, 250, 500, and 1000 mg/kg body weight TCE in corn oil (n=4 per treatment group). 16
hours after the last daily dose of TCE, mice received tritiated thymidine and were sacrificed 6
hours later. Hepatic DNA was extracted form whole liver and standard histopathology was also
performed. Hepatic DNA content and cellular distributions were also determined for thymidine
uptake using autoradiography of tissue sections. Tritiated thymidine incorporation into DNA
was determined by microscopic observations of autoradiography slides and reported as positive
cells per 100 (200X power) fields.
02/12/09
E-27
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Changes in the treatment groups were reported to "include an increase in eosinophilic
cytoplasmic staining of hepatocytes located near central veins, accompanied by loss of
cytoplasmic vacuolization. Intermediate zones appeared normal and no changes were noted in
portal triad areas. Male and female mice given 1000 mg/kg body weight TCE exhibited
apoptosis located near central veins. No evidence of cellular proliferation was seen in the portal
areas. No evidence of increased lipofuscin was seen in liver sections from male and female mice
treated with TCE. Evaluation of cell death in male and female mice receiving TCE was
performed by enumerating apoptoses." The apoptoses "did not appear to be in proportion to the
applied TCE dose given to male or female mice." The mean number of apopotosis per 100 (400
X) fields in each group of 4 animals (male mice) was 0, 0, 0, 1 and 8 for control, 100, 250, 500,
and 1000 mg/kg TCE treated groups, respectively. Variations in number of apoptoses between
mice were not given by the authors. Feulgen stain was <1 for all doses except for 9 at 1000
mg/kg.
Mitotic figure were reported to be "frequently seen in liver sections from both male and
female mice treated with TCE. Dividing cells were most often found in the intermediate zone
and resembled mature hepatocytes. Incorporation of the radiolabel into cells located near the
portal triad areas was rare. In general mitotic figures were very rare, but when found they were
usually located in the intermediate zone. Little or no incorporation of label was seen in areas
near the bile duct epithelia or in areas close to the portal triad." No quantitative description of
mitotic index was reported by the authors but this description is consistent with there being
replication of mature hepatocytes induced by TCE.
The distribution of tritiated thymidine was given for specific cell types in the livers of 5
animals per treatment group and radiolabel was reported to be predominantly associated with
perisinusoidal cell in control mice. The authors state that the label was more often found in cells
resembling mature hepatocytes. The mean number of labeled cells in autoradiographs per 100
(200X power) fields was reported to be ~ 125 and -150 labeled perisinusoidal cells in controls
male and female mice, respectively. The authors do not give any standard deviations for the
female perisinusoidal data except for the 1000 mg/kg exposure group. For mature hepatocytes,
the mean baseline level of cell labeling for control male and female mice were reported to be ~
65 and ~ 90 labeled cells, respectively. Although the baseline levels of hepatocyte labeling were
reported to differ between male and female mice, the mean peak level of labeling was similar at
~ 250 labeled cells for male and female mice treated with TCE. Thus, in male mouse liver, the
number of labeled cells increased ~ 2-fold of control levels after 500 and 1000 mg/kg TCE and
in female mouse liver increased ~ 4-fold of control levels after 250, 500 and 1000 mg/kg TCE in
female mouse liver hepatocytes over their respective control levels.
Incorporation of tritiated thymidine into DNA extracted from whole liver in male and
female mice was reported to be significantly elevated after TCE treatment but, unlike the
autoradiographic data, there was no difference between genders and the mean peak level of
tritiated thymidine incorporation occurred at 250 mg/kg TCE treatment and remained constant
for the 500 and 1000 mg/kg treated groups. Increased thymidine incorporation into DNA
02/12/09
E-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
extracted from liver of male and female mice were reported to show a very large standard
deviation with TCE treatment (e.g., at 100 mg/kg TCE exposure, male mice had a mean of-130
dpm tritiated thymidine/ microgram DNA with the upper bound of the standard deviation to be
225 dpm). The increased thymidine incorporation peaked at a level that was a little less than 2-
fold of control level. Thus, for both male and female mice both autoradiographs and total
hepatic DNA were reported to show that male and female mice had similar peaks of increased
thymidine incorporation after TCE exposure that reached a plateau at the 250 mg/kg TCE
exposure level and did not increase with increasing exposure concentration. These data also
indicate a very small population undergoing mitosis due to TCE exposure after 10 days of
exposure. If higher levels of hepatocyte replication had occurred earlier, such levels were not
sustained by 10 days of TCE exposure. More importantly, these data suggest that tritiated
thymidine levels were targeted to mature hepatocytes and in areas of the liver where greater
levels of polyploidization. The ages and weights of the mice were described by these authors,
unlike Elcombe et al, and a different strain was used. However, these results are consistent with
those of Elcombe in regard to the magnitude of thymidine incorporation induced by TCE
treatment and the lack of a dose response once a relative low level of exposure has been
exceeded.
The total liver DNA content of male and female mice treated with TCE were also
determined with the total micrograms DNA/g liver reported to be ~ 4 microgram/g for female
control mice and ~ 2 micrograms/g for male control mice. Although not statistically significant,
the total DNA concentration dropped from ~ 4 to ~ 3 at 100 mg/kg through 1000 mg/kg exposure
to TCE in female mice. For male mice the total DNA rose slightly in the 250 and 500 mg/kg
groups to ~ 3 micrograms/gram and was similar to control levels at the 100 and 1000 mg/kg TCE
treatment groups. The standard deviation in male mice was very large and the number of
animals small making quantitative judgments regarding this parameter difficult. The slight
decrease reported for female mice would be consistent with the results of Elcombe et al (1985)
who describe a slight decrease in hepatic DNA in male mice. However, the reported slight
increase in hepatic DNA in male mice in this study is not consistent. Given the small number of
animals and the large deviations for female and male mice in the TCE treated groups, this study
may not have had the sensitivity to detect slight decreases reported by Elcombe et al.
In regard to clinical evaluation and weight analyses, both male and female mice given
TCE were reported "to appear clinically ill. These mice showed reduced activity and failed to
groom. Control mice showed no adverse effects. Female mice were markedly more affected by
TCE than their male counterparts. Several deaths of female mice occurred during the course of
the TCE treatment regimen." The authors do not give cause of deaths but state that 2 female
mice died in the group receiving 250 mg/kg TCE and one in the group receiving 1000 mg/kg
during the gavage regimen of the female mice. This appears to be similar gavage error or
"accidental death" reported in NTP studies chronic studies of TCE (see below).
The authors report "no significant difference in the absolute body weight of male and
female mice were noted in control groups. Body weight gain in female and males mice treated
02/12/09
E-29
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
with TCE was not significantly different from that of control mice. Liver weights in male mice
given 500 or 1000 mg/kg and corrected for total body weight were significantly elevated. The
corrected liver weights of female mice increase proportionally with the applied dose of TCE."
For male mice, liver weights were reported to be 1.40 ± 0.16, 1.38 ± 1.23, 1.48 ±0.09, 1.61 ±
0.07, and 1.63 ± 0.11 g for control, 100, 250, 500, and 1000 mg/kg TCE in male mice (n=5),
respectively. Body weights were smaller for the 100 mg/kg TCE treatment group although not
statistically significant. The liver weights after treatment had a much larger reported standard
deviation (1.23 g for 100 mg/kg group vs. <0.16 for all other groups). The % liver/ body weight
ratios were reported to be 5.40, 5.41, 5.42, 5.71, and 6.34 % for the same groups in male mice,
this represents 1.06-fold and 1.17-fold of control at the 500 and 1000 mg/kg dose. The authors
report a statistically significant increase in % liver/body weight ratio only for the 500 mg/kg (i.e.,
1.06-fold of control) and 1000 mg/kg (i.e., 1.17-fold of control) TCE exposure groups. The
results for female mice liver weights were reported in Table III of the paper, which was
mistakenly labeled as for male mice. The reported values for liver weight were 1.03 ± 0.07, 1.05
± 0.10, 1.15 ± 0.98, 1.21 ± 0.18 and 1.34 ± 0.08 g for control, 100, 250, 500, and 1000 mg/kg
TCE in female mice (n=5, except for 250 mg/kg and 1000 mg/kg groups), respectively. The %
liver/body weight ratios were 5.26, 5.44, 5.68, 6.24, and 6.57% for the same groups. These
values represent 1.03-fold, 1.08-fold, 1.19-fold, and 1.25-fold of controls in % liver/body weight.
The magnitude of increase in TCE-induced %liver/ body weight ratio in female mice is reflective
of the magnitude of the difference in dose up to 1000 mg/kg where it is slightly lower. The
female mice were reported to have statistically significant increases in % liver/body ratios at the
lowest dose tested (100 mg/kg TCE) after 10 days of TCE exposure that also increased
proportionately with dose. Male mice were not reported to have a significant increase in %
liver/body weight until 500 mg/kg TCE but a statistically significant increase in liver weight at
250 mg/kg TCE. Male mice had a much larger variation in initial body weight than did female
mice (range of means of 24.86 to 27.84 g between groups for males or ~11% difference and
range of means of 19.48 to 20.27 g for females or -4%) which may contribute to an apparent
lack of effect for a parameter that is dependent on body weight. Only 5 mice were used in each
group so the power to detect a change was relatively small.
The results from this experiment are consistent with those of Elcombe et al (1985) in
showing a slight increase in thymidine incorporation (~2-fold of control) and mitotic figures that
are rare after TCE exposure. This study also records a lack of apoptosis with TCE treatment
except at the highest exposure level (i.e., 1000 mg/kg). The increases in liver weight induced by
TCE were reported to be dose-related, especially in female mice where baseline body weights
were more consistent. However, the incorporation of tritiated thymidine reached a plateau at 250
mg/kg TCE in the DNA of both genders of mice. This study specifically identified where
thymidine incorporation and mitotic figures were occurring in TCE-treated livers and noted that
the mature hepatocyte that appeared to be primarily affected, as well as in the portion of the liver
where mature hepatocytes with higher ploidy are found. The authors note that the "lack of
thymidine incorporation in the periportal area, where the liver stem cells are reside," suggesting
that the mature hepatocyte is the target of TCE effects on DNA synthesis. This finding is
consistent with a change in ploidy accompanying hepatocellular hypertrophy and not just cell
02/12/09
E-30
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
proliferation after 10 days of TCE exposure. Like Elcombe et al (1985), these data represent "a
snapshot in time" which does not show whether increased cell proliferation may have happened
at an earlier time point and then subsided by 10 days. However, like Elcombe et al (1985) it
suggests that sustained proliferation is not a feature of TCE exposure and that the level of DNA
synthesis (which is very low in quiescent control liver) is increased in a small population of
hepatocytes due to TCE exposure that is not dose-dependent (only 2- fold increase over control
in animals exposed from 250 to 1000 mg/kg TCE. In regards to toxicity, no evidence of
increased lipid peroxidation in TCE-treated animals was reported using histopathologic sections
stained to enhance observation of lipofuscin. No necrosis is noted by these authors and the
deaths in female mice are likely due to gavage error.
2.1.10. Nakajima et al. 2000
This study focused on the effect of TCE treatment on PPARa-null mice in terms of
peroxisome proliferation but also included information on differences in liver weight between
null and wild-type mice, as well as gender-related effects. SV129 wild-type and PPARa-null
mice (10 weeks of age) were treated with corn oil or 750 mg/kg TCE in corn oil daily for 2
weeks via gavage (n= 6 per group). A small portion of the liver was removed for
histopathological examination but the lobe used was not specified by the authors. Liver
peroxisome proliferation was reported to be evaluated morphologically using 3,3'-
diaminobenzidine (DAB) staining of sections and electron photomicroscopy to detect the volume
density of peroxisomes (% of cytoplasm) in 15 micrographs of the pericentral area per liver. A
number of P-oxidation enzymes and P450s were analyzed by immunoblot of liver homogenates.
The final body weights, liver weights and % liver/body weight ratios were reported for all
treatment groups. For male mice, vehicle treated PPARa-null mice had slightly lower mean
body weights (24.5 ± 1.8 g vs. 25.4 ± 1.9 g [SD]), slightly larger liver weights (1.14 ± 0.13g vs.
1.05 ± 0.15g or ~ 9%), and slightly higher % liver/body weight ratios (4.12 ± 0.32 % vs. 4.10 ±
0.37%) than wild-type mice. The mean values for final body weights of the groups of mice in
this study were reported and were similar which, as demonstrated by the inhalation studies by
Kjellstrand et al (1983a) (see Section 2.2.4), is particularly important for determining the effects
of TCE treatment on % liver/body weight ratios. For both groups of male mice, 2 weeks of TCE
treatment significantly increased both liver weight and % liver/body weight ratios. For male
wild type mice the increase in % liver/body weight was 1.50-fold of vehicle control and for male
PPARa-null mice the increase was 1.26-fold of control after 2 weeks of TCE treatment. For
female mice, vehicle treated PPARa-null mice had slightly higher mean body weights (22.7 ±
2.1 g vs. 22.4 ± 2.0 g ), slightly larger liver weights (0.98 ± 0.15g vs. 0.95 ± 0.14g or ~ 3%), and
slightly higher % liver/body weight ratios (4.32 ± 0.35 % vs. 4.24 ± 0.41%) than wild-type mice.
For both groups of female mice, 2 weeks of TCE treatment significantly increased % liver/body
weight ratios. For liver weights there was a reporting error for PPARa-null female treated with
TCE so that liver weight changes due to TCE treatment cannot be determined for this group. For
female wild type mice the increase in % liver/body weight was 1.24-fold of vehicle control and
for female PPARa-null mice the increase was 1.26-fold of control after 2 weeks of TCE
02/12/09
E-31
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
treatment. Thus for both wild-type and PPARa-null mice, TCE exposure resulted in increased %
liver/body weight over controls that was statistically significant after 2 weeks of oral gavage
exposure using corn oil as the vehicle. For male mice there was a greater TCE-induced increase
in % liver/body weight in wild type than PPARa-null mice (1.50 -fold vs. 1.26-fold of control)
that was statistically significant, but for female mice the induction of increased liver weight was
statistically increased but the same in wild-type and PPARa-null mice (i.e., both were -1.25 -
fold of control). These date indicate that TCE-induced increases in mouse liver weight were not
dependent on a functional PPARa receptor in female mice and suggest that some portion may be
in male mice.
In regard to light and electron microscopic results, the numbers of peroxisomes in
hepatocytes of wild-type mice were reported to be increased, especially in the pericentral area of
the hepatic lobule, to a similar extent in both males and females (15 micrographs, n= 4 mice).
TCE exposure was reported to increase the volume density of peroxisomes 2-fold of control in
the pericentral area with no evident change in peroxisomes in the periportal areas, but data was
not shown for that area of the liver lobule. In contrast, no increase in peroxisomes was reported
to be observed in PPARa-null mice. Therefore, increases in liver weight observed in PPARa-
null mice after TCE treatment did not result from peroxisome proliferation. Similarly, the small
2-fold increase in peroxisome volume from 2% to 4% of cytoplasmic volume in the pericentral
area of the liver lobule in wild-type mice could not have been responsible for the 50% increase
liver weight observed in male wild-type mice.
Although no difference was reported between male and female wildtype mice in regard to
TCE-induced peroxisome proliferation in wild type mice, the levels of hepatic enzymes
associated with peroxisomes [acyl-CoA (AOX), peroxisomal bifunctional protein (PH),
peroxisomal thiolase (PT), very long chain acyl-CoA synthetase (VLACS), and D-type
peroxisomal bifunctional protein (DBF), cytosolic enzyme [cytosolic thioesterase II (CTEII)],
mitochondrial enzymes [mitochondrial trifunctional protein a subunits a and P(TPa and TPP)],
and microsomal enzymes [Cytochrome P450 4A1 (CYP4A1)] as measured by immunoblot
analysis were significantly elevated in male wild type mice (n= 4) by a factor of ~ 2-3, but
except for a slight elevation in PH and PT, were reported to not be elevated in female wild-type
mice (n= 4). The magnitude of increase in peroxisomal enzymes was similar to that of
peroxisomal volume in male mice. No TCE-induced increases in any of these enzymes were
reported in male or female PPARa-null mice by the authors. For CYP4A1, an enzyme reported
to be induced by peroxisomal proliferators, TCE exposure resulted in a much lower amount in
female than male wild-type mice (i.e., 2% of the level induced by TCE in males). However, the
expression of catalase was reported to be "nearly constant in all samples" (at most ~ 30%
change) which the authors suggested resulted from induction by TCE that was independent of
PPARa. The basis for selection of 4 mice for this comparison out of the 6 studied per group was
not given by the authors. A comparison of control wild-type and PPARa-null mice showed that
in males background levels of the enzymes examined were generally similar except for DBF in
which the null mice had values ~ 50% of the wild type controls. A similar decrease was reported
for female PPARa-null mice. With regard to gender differences in wild-type mice, females had
02/12/09
E-32
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
similar values as males with the exceptions of TPa, TPp, and CYP2E1 which were in untreated
female wild-type mice at a 3.06-fold, 2.38-fold and 1.63-fold for 1 TPa, TPp, and CYP2E1 levels
over males, respectively. Female PPARa-null mice had increases of 2.50-fold, 1.54-fold, and
2.07-fold over male wild-type mice.
With regard to the induction of TCE metabolizing enzymes (CYP1A2, CYP2E1 and
ALDH), CYP1A2 was reported to be decreased by TCE treatment of both male and female wild-
type mice but liver CYP2E1 reported to be increased in male mice and constant in female mice
which resulted in similar expression level in both genders after TCE treatment. There was no
gender difference in ALDH activity reported after TCE exposure and activity was reported to be
independent of PPARa. The authors concluded that TCE metabolizing abilities of the liver of
male and female mice were similar and therefore poor induction of peroxisomal related enzymes
was not due to gender-related differences in TCE metabolism.
To investigate whether the a gender-related difference peroxisomal enzymes after TCE
exposure was due to a lower levels of PPARa and RXRa receptors, western blotting was
employed (n= 3). The level of PPARa protein was reported to be increased in both male wild-
type mice with less induction in females (control vs. TCE, 1.00 ± 0.20 vs. 2.17 ± 0.24 in males
and 0.95 ± 0.25 vs. 1.44 ± 0.09 in females) after TCE treatment. The hepatic level of RXRa was
also reported to be increased in the same manner as PPARa (control vs. TCE, 1.00 ± 0.33 vs.
1.92 ± 0.04 in males 0.81 ± 0.16 vs. 1.14 ± 0.10 in females). Northern blot analysis of hepatic
PPARa mRNA was reported to show greater TCE induction in male (2.6-fold of control) than in
female (1.5-fold of control) wild-type mice. Thus, males appeared to have higher induction of
the two receptor proteins as well as a greater response in peroxisomal enzymes and CYP4A1,
even though TCE-induced increases in peroxisomal volume was similar between male and
female mice. The increased response in males for induction of the two receptor proteins is
consistent with liver weight data that shows some portion of the induction of increased liver
weight response in male mice using this paradigm may be due to gender-specific differences in
PPARa response. However, as noted below (see Sections 2.2), corn oil vehicle has liver effects
alone, especially in the male liver, that have also been associated with PPARa responses.
2.1.11. Berman et al. 1995
This study included TCE in a suite of compounds used to compare endpoints for
toxicological screening methods. Female Fischer 344 rats of 77 days of age (n= 8 per group)
were administered TCE in corn oil for 1 day (0, 150, 500, 1500, or 5000 mg/kg/day) or for 14
days (0, 50, 150, 500, or 1500 mg/kg/day). Blood samples were taken 24 hours after the last
dose and livers were weighed and H&E sections were examined for evidence of parenchymal
cell degeneration, necrosis, or hypertrophy. No details were provided by the authors for the
extent or severity of the liver affects by histopathological examination. The serum chemistry
analysis included LDH, Alkaline phosphatase, ALT, AST, total bilirubin, creatine, and BUN.
The starting and ending body weights of the animals or the absolute liver weights were not
reported by the authors.
02/12/09
E-33
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The results of a multivariate analysis were reported to show a lowest effective dose of
1500 mg/kg after 1 day of TCE exposure and 150 mg/kg after 14 days of TCE exposure that was
statistically significant. Liver weight and liver weight changes were not reported by the authors
but the % liver to body weight ratios were. For the two control groups there was a difference in
% liver/body weight of - 8% (3.43 ± 0.74 % for the 1-day control group and 3.16 ± 0.41 % for
the 14-day control group, mean ± SEM). For the 1-day groups only the 5000 mg/kg group was
reported to show a statistically significant difference in % liver/body weight between control and
TCE treatment (i.e., ~ 1.08 fold increase). Hepatocellular necrosis was noted to occur in the
1500 and 5000 mg/kg groups in 6/7 and 6/8 female rats, respectively but not to occur in lower
doses. The extent of necrosis was not noted by the authors for the two groups exhibiting a
response after 1 day of exposure. Serum enzymes indicative of liver necrosis were not presented
and because only positive results were presented in the paper, presumed to be negative.
Therefore, the extent of necrosis was not of a magnitude to affect serum enzyme markers of
cellular leakage.
After 14-days of TCE exposure, there was a dose-related increase reported for %
liver/body weight ratios that was statistically significant at all TCE dose levels although the
multivariate analysis indicated the lowest effective dose to be 150 mg/kg. The % liver/body
weight ratio was 3.16 ± 0.41%, 3.38 ± 0.56%, 3.49 ± 0.69%, 3.82 ± 0.76%, and 4.47 ± 0.66 %
for control, 50 mg/kg, 150 mg/kg, 500 mg/kg and 1500 mg/kg TCE exposure levels, respectively
after 14 days of exposure. No hepatocellular necrosis was reported at any dose and
hepatocellular hypertrophy was reported only at the 1500 mg/kg dose and in all rats. These rat
liver weights are 1.07-fold, 1.10-fold, 1.21-fold, and 1.41-fold of controls for the 50, 150, 500
and 1500 mg/kg TCE dose groups, respectively. The 7% increase in liver weight at the 50
mg/kg dose is approximately the same difference between the two control groups for days 1 and
14 treatments. Without the data for starting and final body weights and an examination of
whether the control animals had similar body weight, it is impossible to discern whether the
reported effects at the low dose of TCE was also reflected differences between the control
groups. No serum enzyme levels changes were reported after 14 days of exposure to TCE for
any group.
The authors not that their study provided evidence of liver effects at lower levels than
other studies citing Elcombe et al (1985) and Goldsworthy and Popp (1987). They suggest that
the differences in sensitivity to TCE between their results and those of these two studies may
reflect differences in strain or gender of the rats examined. However, they did not study male rats
of this strain concurrently so that differences in gender may have reflected differences between
experiments. The increase in liver weight without reporting increases in hepatocellular
hypertrophy as well as the lack of necrosis as low doses is consistent with the results of Melnick
et al (1987) in male Fischer rats given TCE orally (see Section 2.1.11 below).
02/12/09
E-34
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.1.12. Melnick et al. 1987
The focus of this study was to assess microencapsulation as a way to expose rodents to
substances such as TCE that have issues related to volatilization in drinking water or apparent
gavage-related deaths. In this study, liver weight changes, extent of focalized necrosis, and
indicators of peroxisome proliferation were reported as metrics of TCE toxicity. TCE (99+ %)
was encapsulated in gelatin-sorbitol microcapsuls and was 44.1% TCE w/w. The TCE
microcapsules were administered to male Fischer 344 rats (6-week old and weighing between 89
and 92 g or -3% difference) in the diet (0, 0.55, 1.10, 2.21 and 4.42% TCE in the diet) for 14
days. The number of animals in each group was 10. A parallel group of animals was
administered TCE in corn oil gavage for 14 consecutive days (corn oil control, 0.6, 1.2, and 2.8
g/kg/day TCE). The dosage levels of TCE in the gavage study were reported to be "adjusted 5
times during the 14-day" treatment period to be similar to the dosage levels of TCE in the feed
study. The time-weighted average dosage levels of TCE in the feed study were reported to be
0.6, 1.3, 2.2, and 4.8 g/kg/day.
There was less food consumption reported in the 2.2 and 4.8 g/kg/day dose feed groups
which the authors attribute to either palatability or toxicity. There were no deaths in any of the
groups treated with microencapsulated TCE while, similar to many other gavage studies of TCE
reported in the literature, there were 4 deaths in the high-dose gavage group. Mean body weight
gains of the two highest dose groups of the feed study and of the highest dose group of the
gavage study were reported to be significantly lower than the mean body weight gains of the
respective control groups (i.e., ~ 22 % and ~ 35% reduction at 2.2 and 4.8 g/kg/day in the feed
study, respectively, and ~ 33% reduction at 2.8 g/kg/day TCE in the gavage study). After 14
days of treatment, liver weights were reported to be 8.1 ± 0.8, 8.4 ± 0.8, 9.5 ± 0.5, 10.1 ± 1.2,
8.9 ± 1.3, and 7.4 ± 0.5 g for untreated control, placebo control, 0.6, 1.3, 2.2, and 4.8 g/kg TCE
exposed feed groups, respectively. The corresponding % liver/body weight ratios were reported
to be 5.2 ± 0.3, 5.3 ± 0.2, 6.0 ± 0.3, 6.5 ± 0.5, 7.0 ± 0.9, and 7.1 ± 0.5 % for untreated control,
placebo control, 0.6, 1.3, 2.2, and 4.8 g/kg TCE exposed groups, respectively. The increased %
liver/body weight ratio represents 1.13-fold, 1.23-fold, 1.32-fold, and 1.34-fold of placebo
controls, respectively. For the gavage experiment, after 14 days of treatment liver weights were
reported to be 7.1 ± 1.3, 9.3 ± 1.2, 9.1 ± 0.9, and 7.7 ± 0.4 g for corn oil control, 0.6, 1.2, and 2.8
g/kg TCE exposed groups, respectively. The corresponding % liver/body weight ratios were
reported to be 5.0 ± 0.4, 6.0 ± 0.4, 6.1 ± 0.3, and 7.3 ± 0.5 % for corn oil control, 0.6, 1.2, and
2.8 g/kg TCE exposed groups, respectively. The % liver/body weight ratios represent 1.20-fold,
1.22-fold, and 1.46-fold of corn oil controls, respectively. The 2.8 g/kg TCE gavage results are
reflective of the 6 surviving animals in the group rather than 10 animals in the rest of the groups.
There was no explanation given by the authors for the lower liver weights in the control gavage
group than the placebo control in the feed group (i.e., 20% difference) although the initial and
final body weights appeared to be similar. The decreased body weights in the feed and gavage
study are reflective if TCE systemic toxicity and appeared to affect the TCE-induced liver weight
increases in those groups.
02/12/09
E-35
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The authors reported that the only treatment-related lesion observed microscopically in
rats from either dosed-feed or gavage groups was individual cell necrosis of the liver with the
frequency and severity of this lesion similar at each dosage levels of TCE administered
microencapsulated in the feed or in corn oil. Using a scale of minimal = 1-3 necrotic
hepatocytes/10 microscopic 200X fields, mild = 4-7 necrotic necrotic hepatocytes/10
microscopic 200X fields, and moderate = 8-12 necrotic hepatocytes/10 microscopic 200X fields,
the frequency of lesion was 0-1/10 for controls, 2/10 for 0.6 and 1.3 g/kg and 9/10 for 2.2 and 4.8
g/kg feed groups. The mean severity was reported to be 0.0-0.1 for controls, 0.3-0.4 for 0.6 and
1.3 g/kg, and 2.0 -2.5 for 2.2 and 4.8 g/kg feed groups. For the corn oil gavage study, the corn
oil control and 0.6 g/kg groups were reported to have a frequency of 0 lesions/10 animals, the
1.2 g/kg group a frequency of 1/10 animals, while the 2.8 g/kg group to have a frequency of 5/6
animals. The mean severity score was reported to be 0 for the control and 0.6 g/kg groups, 0.1
for the 1.2 g/kg groups, and 1.8 for the remaining 6 animals in the 2.8 g/kg group. The
individual cell necrosis was reported to be randomly distributed throughout the liver lobule with
the change to not be accompanied by an inflammatory response. The authors also report that
there was no histologic evidence of cellular hypertrophy or edema in hepatic parenchymal cells.
Thus, although there appeared to be TCE-treatment related increases in focal necrosis after 14
days of exposure, the extent was even at the highest doses mild and involved few hepatocytes.
Microsomal NADPH cytochrome c-reductase was reported to be elevated in the 2.2 and
4.8 g/kg feed groups and in the 1.2 and 2.8 g/kg gavage groups. Cytochrome P450 levels were
reported to be elevated only in the two highest dose groups of the feed study. The authors
reported a dose-related increase in peroxisome PCO and catalase activities in liver homogenates
from rats treated with TCE microcapsules or by gavage and that treatment with corn oil alone,
but not placebo capsules, caused a slight increase in PCO activity. After 14 days of treatment,
PCO activities were reported to be 270 ± 12, 242 ± 17, 298 ± 64, 424 ± 55, 651 ± 148, and 999 ±
266 nmole H202 produced/ min/g liver for untreated control, placebo control, 0.6, 1.3, 2.2, and
4.8 g/kg TCE exposed feed groups, respectively. This represents 1.23-fold, 1.75-fold, 2.69-fold,
and 4.13-fold of placebo controls, respectively. After 14 days of treatment, catalase activities
were reported to be 8.49 ± 0.81, 7.98 ± 1.62, 8.49 ± 1.92, 8.59 ±1.31, 13.03 ± 2.01, and 15.76 ±
1.11 nmole H202 produced/ min/g liver for untreated control, placebo control, 0.6, 1.3, 2.2, and
4.8 g/kg TCE exposed groups, respectively. This represents 1.06-fold, 1.07-fold, 1.63-fold, and
1.97-fold of placebo controls, respectively. Thus, although reported to be dose related, only the
two highest exposure levels of TCE increased catalase activity and to a smaller extent than PCO
activity in microencapsulated TCE fed rats. For the gavage experiment, after 14 days of
treatment PCO activities were reported to be 318 ± 27, 369 ± 26, 413 ± 40, and 1002 ±271
nmole H202 produced/ min/g liver for corn oil control, 0.6, 1.2, and 2.8 g/kg TCE exposed
groups, respectively. This represents 1.16-fold, 1.29-fold, and 3.15-fold of corn oil controls.
After 14 days of treatment, catalase activities were reported to be 8.59 ± 0.91, 10.10 ± 1.82,
12.83 ± 3.43, and 13.54 ± 2.32 nmole H202 produced/ min/g liver for corn oil control, 0.6, 1.2,
and 2.8 g/kg TCE exposed groups, respectively. This represents 1.18-fold, 1.49-fold, and 1.58-
fold of corn oil controls. As stated by the authors the corn oil vehicle appeared to elevate
catalase activities and PCO activities.
02/12/09
E-36
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
In regard to dose-response, liver and body weight were affected by decreased body
weight gain in the higher dosed animals in this experiment (i.e., 2.2 g/kg/day TCE exposure and
above) and by gavage related deaths in the highest-dosed group. The lower liver weight in the
gavage control group also may have affected the determination of the magnitude of TCE-related
liver weight gain at that dose. At the 2 doses, below which body weight gain was affected, there
appeared to be an approximately 20% increase in % liver/body weight ratio in the gavage study
and a 13% and 23 % weight increase in the feed study. The extent of PCO activity appeared to
increase more steeply with dose in the feed study than did liver weight gain (i.e., a 1.23-fold of
liver/body weight ratio at 1.3 g/kg/day corresponded with a 1.75-fold PCO activity over control).
At the two highest doses in the feed study the increase in PCO activity was 2.69-fold and 4.13-
fold of control but the increase in liver weight was not more than 34%. For the gavage study
there was also a steeper increase in PCO activity than liver weight gain. For catalase activity, the
increase was slightly less than that of liver/body weight ratio % for the two doses that did not
decrease body weight gain in the feed study. In the gavage study they were about the same. In
regard to what the cause of liver weight gain was, the authors report that there was no histologic
evidence of cellular hypertrophy or edema in hepatic parenchymal cells and do not describe
indicators of hepatocellular proliferation or increased polyploidy. Accordingly, the cause of liver
weight gain after TCE exposure in this paradigm is not readily apparent.
2.1.13. Laughter et al. 2004
Although the focus of the study was an exploration of potential MO As for TCE effects
through macroarray transcript profiling (see Section 3.1.2. for discussions of limitations of this
approach and especially the need for phenotypic anchoring, Section 3.4.1.3 for use of PPARa
knockout mice, and Section 3.4.2.2 for discussion of genetic profiling data for TCE), information
was reported regarding changes in the liver weight of PPARa-null mouse and their background
strains. SV129 wild type and PPARa-null male mice (9 ± 1.5 weeks of age) were treated with 3
daily doses of TCE in 0.1% methyl cellulose for either 3 days or 3 weeks (n= 4-5/group). Thus,
this paradigm does not use corn oil, which has been noted to affect toxicity (see Section 2.2
below), but is not comparable to other paradigms that administer the total dose in one daily
gavage administration rather than to give the same cumulative dose but in 3 daily doses of lower
concentration. The initial or final body weights of the mice were not reported. Thus, the effects
of systemic toxicity from TCE exposure on body weight and the influence of differences in
initial body weight on % liver/body weight determinations cannot be made. For the 3-day study,
mice were administered 1500 mg/kg TCE or vehicle control. For the 3-week study mice were
administered 0, 10, 50, 125, 500, 1000, or 1500 mg/kg TCE 5 days a week except for 4 day/
week on the last week of the experiment. In a separate study mice were given TCA or DCA at
0.25, 0.5, 1, or 2 g/1 (pH ~ 7) in the drinking water for 7 days. For each animal a block of the
left, anterior right, and median liver lobes was reported to be fixed in formalin with 5 sections
stained for H&E and examined by light microscopy. The remaining liver samples were
combined and used as homogenates for transcript arrays. In the 3-week study, BrdU was
administered via miniosmotic pump on day one of week 3 and sections of the liver assessed for
BrdU incorporation in at least 1000 cells per animal in 10 - 15 fields.
02/12/09 E-37
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Although initial body weights, final body weights, and the liver weights were not
reported, the % liver/ body ratios were. In the 3-day study, control wild type and PPARa-null
mice were reported to have similar % liver/body weight ratios of ~ 4.5%. These animals were ~
10 weeks of age upon sacrifice. However at the end of the 3-week experiment the % liver/body
weight ratios were increased in the PPARa-null male mice and were 5.1%. There was also a
slight difference in the % liver/body weight ratios in the 1-week study (4.3 ± 0.4 % vs. 4.6 ± 0.2
% for wild-type and PPARa-null mice, respectively). These results are consistent with an
increasing baseline of hepatic steatosis with age in the PPARa-null mice and increase in liver
weight. In the 3-day study, the mean reported the % liver/body ratio was 1.4-fold of the animals
tested with TCE in comparison to the control level. In the PPARa-null mice there was a 1.07-
fold of control level reported by the authors to not be statistically significant. However, given
the low number of animals tested (the authors give only that 4-5 animals were tested per group
without identification as to which groups has 4 animals and which had 5), the ability of this study
to discern a statistically significant difference is limited. In the 3-week study, wild-type mice
exposed to various concentrations of TCE had % liver/body weights that were within ~ 2% of
control values except for the 1000 mg/kg and 1500 mg/kg groups that were ~ 1.18 - fold and
1.30 - fold of control levels, respectively. For the PPARa-null mice exposed to TCE for 3
weeks, the variability in % liver/body weight was greater than that of the wild-type mice in most
of the groups. The baseline level % liver/body weight was 1.16-fold in the PPARa-null mice in
comparison to wild-type mice. At the 1500 mg/kg TCE exposure level % liver/body weights
were not recorded because of the death of the null mice at this level. The authors reported that at
the 1500 mg/kg level all PPARa-null mice were moribund and had to be removed from the
study. However, at 1000 mg/kg TCE exposure level there was a 1.10-fold of control %
liver/body weight value that was reported to not be statistically significant. However, as noted
above, the power of the study was limited due to low numbers of animals and increased
variability in the null mice groups. The % liver/body weight reported in this study was actually
greater in the null mice than the wild-type male mice at the 1000 mg/kg TCE exposure level (5.6
± 0.4 % vs. 5.2 ± 0.5%), for null and wild-type mice, respectively). Thus, at 1-week and at 3-
weeks, TCE appeared to induce increases in liver weight in PPARa-null mice, although not
reaching statistical significance in this study, with concurrent background of increased liver
weight reported in the knockout mice. At 1000 mg/kg TCE exposure for 3 weeks, % liver/body
weight was reported to be 1.18-fold in wild-type and 1.10-fold in null mice of control values. As
discussed above, Nakajima et al. (2000) reported statistically significant increased liver weight in
both wild-type and PPARa-null mice after 2 weeks of exposure with less TCE-induced liver
weight increases in the knockout mice (see Section 2.1.10). They also used more mice, carefully
matched to weights of their mice, and used a single dose of TCE each day with corn oil gavage.
The authors noted that inspection of the livers and kidneys of the moribund null mice,
who were removed from the 3-week study, "did not reveal any overt signs of toxicity in this dose
group that would lead to morbidity" but did not show the data and did not indicate when the
animals were affected and removed. For the wild-type mice exposed to the same concentration
(1500 mg/kg) but whose survival was not affected by TCE exposure, the authors reported that at
the 1500 mg/kg dose these mice exhibited mild granuloma formation with calcification or mild
02/12/09
E-38
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
hepatocyte degeneration but gave not other details or quantitative information as to the extent of
the lesions or what parts of the liver lobule were affected. The authors noted that "wild-type
mice administered 1000 and 1500 mg/kg exhibited centrilobular hypertrophy" and that "the mice
in the other groups did not exhibit any gross pathological changes after TCE exposure." Thus,
the hepatocellular hypertrophy reported in this study for TCE appeared to be correlated with
increases in % liver/body weight in wild-type mice. In regard to the PPARa-null mice, the
authors stated that "differences in the liver to body weights in the control PPARa-null mice
between Study 1 and 2 [sic the 3-day and 3-week studies] were noted and may be due to
differences in the degree of steatosis that commonly occurs in this strain." Further mention of
the background pathology due to knockout of the PPARa was not discussed. The increased %
liver/body weight reported between control and 1000 mg/kg TCE exposed mice (5.1% vs. 5.6%)
was not accompanied by any discussion of pathological changes that could have accounted for
the change.
Direct comparisons of the effects of TCE, DC A, and TCA cannot be made from this
study as they were not studied for similar durations of exposure. However, while TCE induced
increased in % liver/body weight ratios after 3 days and 3 weeks of exposure in wild-type mice
at the highest dose levels, for TCA exposure % liver/body weight after 1 week exposure in
drinking water was slightly elevated at all dose levels with no dose-response (-10% increase),
and for DCA exposure in drinking water a similar elevation in % liver/body weight was also
reported for the 0.25, 0.5, and 1.0 g/1 dose levels (~ 11%) and that was increased at the 2.0 g/1
level by -25% reaching statistical significance. The authors interpret this data to show no TCA-
related changes in wild-type mice but the limited power of the study makes quantitative
conclusions difficult. For PPARa-null mice all there was a slight decrease in % liver/body
weight between control and TCA treated mice at the doses tested (-2%). For DCA treated mice,
all treatment levels of DCA were reported to induce a higher % liver/body weight ratio of at least
- 5% with a 13%) increase at the 2.0 g/1 level. Again the limited power of the study and the lack
of data for TCE at similar durations of exposure as those studied for TCA and DCA makes
quantitative conclusions difficult and comparisons between the chemicals difficult. However,
the pattern of increased % liver/body weight appears to be more similar between TCE and DCA
than TCA in both wild-type and PPARa-null mice. In terms of histological description of
effects, the authors note that "livers from the 2 g/1 DCA-treated wild-type and PPARa-null mice
had hepatocyte cytoplasmic rarefication probably due to an increase in glycogen accumulation."
However, no special procedures are staining were performed to validate the assumption in this
experiment. No other pathological descriptions of the DCA treatment groups were provided. In
regard to TCA, the authors noted that "the livers from wild-type but not PPARa-null mice
exposed to 2.0g/l TCA exhibited centrilobular hepatocyte hypertrophy." No quantitative
estimate of this effect was given and although the extent of increase of % liver/body weight was
similar for all dose levels of TCA, there is no indication from the study that lower concentrations
of TCA also increased hepatocellular hypertrophy or why there was no concurrent increase in
liver weight at the highest dose of TCA in which hepatocellular hypertrophy was reported.
Thus, reports of hepatocellular hypertrophy for DCA and TCA in the 1-week study were not
correlated with changes in % liver/body weight.
02/12/09
E-39
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For control animals, BrdU incorporation in the last week of the 3-week study was
reported to be at a higher baseline level in PPARa-null mice than wild-type mice (~ 2.5 -fold).
For wild type mice the authors reported a statistically significant increase at 500 and 1000 mg/kg
TCE at levels of-1% and ~ 4.5% hepatocytes incorporating the label after 5 days of BrdU
incorporation. Whether this measure of DNA synthesis is representative of cellular proliferation
or of polyploidization was not examined by the authors. Even at 1000 mg/kg TCE the % of cells
that had incorporated BrdU was less than 5% of hepatocytes in wild type mice. The magnitude
% liver/body weight ratio change at this exposure level was 4-fold greater than that of
hepatocytes undergoing DNA synthesis (16% increase in % liver/body weight ratio vs. 4%
increase in DNA synthesis). The ~ 1% of hepatocytes undergoing DNA synthesis at the 500
mg/kg TCE level, reported to be statistically significant by the authors, was not correlated with a
concurrent increase in % liver/body weight ratio. Thus, TCE-induced changes in liver weight
were not correlated with increases in DNA synthesis in wild-type mice after 3 weeks of TCE
exposure. For PPARa-null mice, there was a ~ 3-fold of control value for the % of hepatocytes
undergoing DNA synthesis at the 1000 mg/kg TCE exposure level. The higher baseline level in
the null mouse, large variability in response at this exposure level, and low power of this
experimental design limited the ability to detect statistical significance of this effect although the
level was greater than that reported for the 500 mg/kg TCE exposure in wild-type mice that was
statistically significant. Thus, TCE appeared to induce an increase in DNA synthesis in PPARa-
null mice, albeit at a lower level than wild-type mice. However, the ~ 2% increase in % of
hepatocytes undergoing DNA synthesis during the 3rd week of a 3-week exposure to 1000 mg/kg
TCE in PPARa-null mice was insufficient to account for the —10% observed increase in liver
weight. For wild-type and PPARa-null mice, the magnitude of TCE-induced increases in liver
weight were 4-5 fold higher than that of increases in DNA-synthesis under this paradigm and in
both types of mice, a relatively small portion of hepatocytes were undergoing DNA synthesis
during the last week of a 3-week exposure duration. Whether the increases in liver weight could
have resulted from and early burst of DNA synthesis as well as whether the DNA synthesis
results reported here represents either proliferation or polyploidization, cannot be determined
from this experiment. Because of the differences in exposure protocol (i.e. use of 3 daily doses
in methylcellulose rather than one dose in corn oil) the time course of the transient increase in
DNA synthesis reported cannot be assumed to be the same for this experiment and others.
Not only were PPARa-null mice different than wild-type mice in terms of background
levels of liver weights, and hepatic steatosis, but this study reported that background levels of
PCO activity to be highly variable and in some instances different between wild-type and null
mice. There was reported to be ~ 6-fold PCO activity in PPARa-null control mice in comparison
to wild-type control mice in the 1-week DCA/TCA experiment (~ 0.15 vs. 0.85 units of
activity/g protein). However, in the same figure a second set of data is reported for control mice
for comparison to WY-14,643 treatment in which PCO activity was slightly decreased in
PPARa-null control mice vs. wild-type controls (~ 0.40 vs. 0.65 units of activity/g protein). In
the experimental design description of the paper, WY-14,643 treatment and a separate control
were not described as part of the 1 -week DCA/TCA experiment. For the only experiment in
which PCO activity was compared between wild-type and PPARa-null mice exposed to TCE
02/12/09
E-40
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(i.e., 3-day exposure study), there was a reported increased over the control value of - 2.5 fold
that was reported to be statistically significant at 1500 mg/kg TCE (1.5 vs. 0.60 units of
activity/g protein). For control mice in the 3-day TCE experiment there was an increase in this
activity in PPARa-null mice in comparison to wild-type mice (~ 0.60 vs. 0.35 units of activity/g
protein). While not statistically significant, there appeared to be a slight increase in PCO
activity after 1500 mg/kg TCE exposure for 3 days in PPARa-null mice of - 30%. However, as
noted above the background levels of this enzyme activity varied widely between the
experiments with not only values for control animals varying as much as 6-fold (i.e. for PPARa-
null mice) but also for WY-14,643 administration. There was a 6.6-fold difference in PCO
results for WY-14,643 in PPARa-null mice at the same concentration of WY-14,643 in the 3-day
and 1 week experiment, and a 1.44-fold difference in results in wild-type mice in these two data
sets.
2.1.14. Ramdhan et al. 2008
Ramdhan et al (2008) examined the role of CYP 2E1 in TCE-induced hepatotoxicity,
using cyp2El+/+ (wild-type) and cyp2El-/- (null) Sv/129 male mice (6/group) which were
exposed for 7 days to 0, 1000, or 2000 ppm TCE by inhalation for 8 hours/day (Ramdhan et al.
2008). The exposure concentrations are noted by the authors to be much higher than
occupational exposures and to have increased liver toxicity after 8 hours of exposure as
measured by plasma AST levels. To put this exposure concentration into perspective, the
Kjellstrand et al (1983a; 1983 b) inhalation studies for 30 days showed that these levels were
well above the 150 ppm exposure levels in male mice that induced systemic toxicity. Nunes also
reported hepatic necrosis up to 4% in rats at 2000 ppm for just 8 hours not 7 days. AST and
ALT were measured at sacrifice. Histological changes were scored using a qualitative scale of 0
= no necrosis, 1 = minimal as defined as only occasional necrotic cells in any lobule, 2 = mild as
defined as less than one-third of the lobule structure affected, 3= moderate as defined as between
one-third and two-thirds of the lobule structure affected and 4 = severe defined as greater than
two-thirds of the lobule structure affected. Real-time PCR was reported for mRNA encoding a
number of receptors and proteins. Total RNA and Western Blot analysis was obtained from
whole liver homogenates The changes in mRNA expression were reported as means for 6 mice
per group after normalization to a level of P-actin mRNA expression and were shown relative to
the control level in the CYP2E1 wild type mice.
The deletion of the CYP2E1 gene in the null mouse had profound effects on liver weight.
The body were was significantly increased in control CYP2E1 -/- mice in comparison to wild
type controls (24.48 g ± 1.44 g for null mice vs. 23.66 ± 2.44 g , m ± SD). This represents a
3.5% increase over wild type mice. However, the liver weight was reported in the CYP2E1 -/-
mice to be 1.32-fold of that of CYP2E1+/+mice (1.45 ± 0.10 g vs. 1.10±0.14g). The %
liver/body weight ratio was 5.47% vs. 4.63 % or 1.18-fold of wild type control for the null mice.
The authors report that 1000 ppm and 2000 ppm TCE treatment did induce a statistically
significant change body weight for null or wild type mice. However there was an increase in
body weight in the wild type mice (i.e., 23.66 ± 2.44, 24.52 ± 1.17, and 24.99 ± 1.78 for control,
02/12/09
E-41
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1000 ppm, and 2000 ppm groups, respectively) and an increase in the variability in response in
the null mice (i.e., 24.48 ± 1.44, 24.55 ± 2.26, and 24.99 ± 4.05, for control, 1000 ppm, and 2000
ppm exposure groups, respectively). The % liver/body weight was 5.47 ± 0.23, 5.51 ± 0.27, and
5.58 ± 0.70 % for control, 1000 ppm and 2000 ppm the CYP2E -/- mice, respectively. The %
liver/body weight was 4.63 ± 0.13, 6.62 ± 0.40, and 7.24 ± 0.84 % for control, 1000 ppm, and
2000 ppm wild type mice, respectively. Therefore, while there appeared to be little difference in
the TCE and control exposures for % liver/body weights in the CYP2E1 -/- mice (2%) there was
a 1.56-fold of control level after 2000 ppm in the wild type mice after 7 days of inhalation
exposure.
The authors reported that "in general, the urinary TCE level in CYP2E1 -/- mice was less
than half that in CYP2E1 +/+ mice: urinary TCA levels in the former were about one-fourth
those in the latter." Of note is the large variability in urinary TCE detected in the 2000 ppm TCE
exposed wild type mice , especially after day 4, and that in general the amount of TCE in the
urine appeared to be greatest after the 1st day of exposure and steadily declined between 1 and 7
days (i.e., ~ 45% decline at 2000 ppm and a -70% decline at 1000 ppm) in the wild type mice.
The amount of TCE in the urine was proportional to the difference in dose at days 1 and 5 (i.e., a
2-fold difference in dose resulted in a 2-fold difference in TCE detected in the urine). As the
detection of TCE in the urine declined with time, the amount of TCA was reported to steadily
increase between days 1 and 7 (e.g., from ~ 3 mg TCA after the 1st day to ~ 5.5 mg after 7 days
after 2000 ppm exposure in wild-type mice). However, unlike TCE, there was a with much
smaller differences in response between the two TCE exposure levels (i.e., a 12 - 44% or 1.12-
fold to 1.44-fold difference in TCA levels in the urine at days 1-7 for exposure concentrations
that differ by a factor of 2). This could be indicative of saturation in metabolism and TCA
clearance into urine at these high concentrations levels. The authors note that their results
suggest that the metabolism of TCE in both null and wild-type mice may have reached saturation
at 1000 ppm TCE.
For ALT and AST activities in CYP2E1 -/- or CYP2E1 +/+ mice, both liver enzymes
were significantly elevated only at the 2000 ppm level in CYP2E1 +/+ mice. Although the
increases in excreted TCA in the urine differed by only ~ 33% between the 1000 and 2000 ppm
levels, liver enzyme levels in plasma differed by a much greater extent after 7 days exposure
between the 1000 and 2000 ppm groups of CYP2E1 +/+ mice (i.e., 1.26-fold and 1.83-fold of
control (ALT) and 1.40-fold and 2.20-fold of control (AST) for 1000 ppm and 2000 ppm TCE
exposure levels, respectively). The authors reported a correlation between plasma ALT and both
TCE (r= 0.7331) and TCA (r= 0.8169) levels but do not report details of what data were included
in the correlation (i.e., were data from CYP2E1 +/+ mice combined with those of the CYP2E1 -/-
mice and were control values included with treated values?).
The authors show photomicrograph of a section of liver from control CYP2E1 +/+ and
CYP2E1 -/- mice and describe the histological structure of the liver to appear normal. This
raises the question as to the cause of the hepatomegaly for the CYP2E1 mice in which the liver
weight was increased by a third. The qualitative scoring for each of the 6 animals per group
02/12/09
E-42
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
showed that none of the CYP2E1 -/- control or treated mice showed evidence of necrosis. For
the CYP2E1 +/+ mice there was no necrosis reported in the control mice and in 3/6 mice treated
with 1000 ppm TCE. Of the 3 mice that were reported to have necrosis, the score was reported
as 1-2 for 2 mice and 1 for the third. It is not clear what a score of 1-2 represented given the
criteria for each score given by the authors which defined a score of 1 as minimal and one of 2 as
mild. For the 2000 ppm TCE-exposed mice, all mice were reported to have at least minimal
necrosis (i.e., 4 mice were reported to have scores of 1-2, one mouse a score of 3 and one mouse
a score of 1). What is clear from the histopathology data is that there appeared to be great
heterogeneity of response between the 6 animals in each TCE-exposure group in CYP2E1 +/+
mice and that there was a greater necrotic response in the 2000 ppm exposed mice than the 1000
ppm mice. These results are consistent with the liver enzyme data but not consistent with the
small difference between the 1000 ppm and 2000 ppm exposure groups for TCA content in urine
and by analogy, metabolism of TCE to TCA. A strength of this study is that it reports the
histological data for each animal so that the heterogeneity of liver response can be observed (e.g.,
the extent of liver necrosis was reported to range from only occasional necrotic cells in any
lobule to between one-third and two-thirds of the lobular structure affected after 2000 ppm TCE
exposure for 7 days). Immunohistochemical analysis was reported to show that CYP2E1 was
expressed mainly around the centrilobular area in CYP2E1 +/+ mice where necrotic changes
were observed after TCE treatment.
Given the large variability in response within the liver after TCE exposure in CYP2E1
mice, phenotypic anchoring becomes especially important for the interpretation of mRNA
expression studies (see Sections 1.1 and 3.1.2. for macroarray transcript profiling limitations and
the need for phenotypic anchoring). However, the data for mRNA expression of PPARa,
peroxisomal bifunctional protein (hydratase+3-hydroxyacyl-CoA dehydrogenase),very long
chain acyl-CoA dehydrogenase (VLCAD), CYP4A10, NFkB (p65, P50, P52), and IicBa was
reported at the means ± SD for 6 mice per group and represented total liver homogenates. A
strength of the study was that they did not pool their RNA and can show means and standard
deviations between treatment groups. The low numbers of animals tested however limits the
ability to detect statistically significance of the response. By reporting the means, differences in
the responses within dose groups was limited and reflected differential response and involvement
for different portions of the liver lobule and for the responses of the heterogeneous group of
liver cells populating the liver. The authors reported that they normalized values to the level of
P-actin mRNA in same preparation with a value of 1 assigned as the mean from each control
group. The values for mRNA and protein expression reported in the figures appeared to have all
been normalized to the control values for the CYP2E1-/- mice. Although all of the CYP2E1 -/-
control values were reported as a value of 1, the control values for the CYP2E1+/+ mice differed
with the greatest difference being presented for the CYP4A10-mRNA (i.e., the control level of
CYP4A10 mRNA was ~ 3-fold higher in the CYP2E1+/+ mice than the CYP2E1 -/- mice).
Further characterization of the CYP2E1 mouse model was not provided by the authors.
The mean expression of PPARa mRNA was reported slightly reduced after TCE
treatment in CYP2E1-/- mice (i.e., 0.72-fold and 0.78-fold of control after 1000 and 2000 ppm
02/12/09
E-43
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
TCE exposure, respectively). The CYP2E1 -/- mice had a higher baseline of PPARa mRNA
expression than the CYP2E1+/+ mice (i.e., the control level of the CYP2E1 -/- mice was 1.5-fold
of the CYP2E1+/+ mice). After TCE exposure, the CYP2E1 +/+ had a similar increase in
PPARa mRNA (~ 2.3-fold) at both 1000 ppm and 2000 ppm TCE. Thus, without the presence
of CYP2E1 there did not appear to be increased PPARa mRNA expression. For PPARa protein
expression, there was a similar pattern with ~ 1.6-fold of control levels of protein in the CYP2E1
-/- mice after both 1000 ppm and 2000 ppm TCE exposures. In the CYP2E1 +/+ mice the
control level of PPARa protein was reported to be ~ 1.5-fold of the CYP2E1-/- control level.
Thus, while the mRNA expression was less, the protein level was greater. After TCE treatment
there was a 2.9-fold of control level of protein at 1000 ppm TCE and a 3.1-fold of control level
of protein at 2000 ppm. Thus, the magnitude of mRNA increase was similar to that of protein
expression for PPARa in CYP2E1 +/+ mice. The magnitude of either was 3-fold or less over
control after TCE exposure. This pattern was similar to that of TCA concentration formed in the
liver where there was very little difference between the 1000 and 2000 ppm exposure groups in
CYP2E1 +/+ mice. However, this pattern was not consistent with the liver enzyme and
histopathology of the liver that showed a much greater response after 2000 ppm exposure than
1000 ppm TCE. In addition, where the mean enzyme markers of liver injury and individual
animals displayed marked heterogeneity in response to TCE exposure, there was a much smaller
degree of variability in the mean mRNA expression and protein levels of PPARa
For peroxisomal bifunctional protein there was a greater increase after 1000 ppm TCE-
treated exposure than after 2000 ppm TCE-treatment for both the CYP2E1 -/- and CYP2E1 +/+
mice (i.e., there was a 2:1 ratio of mRNA expression in the 1000 ppm vs. 2000 ppm exposed
groups). The CYP2E1 +/+ mice had a much greater response than the CYP2E1-/- mice (i.e., the
CYP2E1-/- mice had a 2-fold of control and the CYP2E1+/+ mice had a 7.8-fold of control level
after 1000 ppm TCE treatment). For peroxisomal bifunctional protein expression the magnitude
of protein induction after TCE exposure was much greater than the magnitude of increase in
mRNA expression. In the CYP2E1-/- mice 1000 ppm TCE exposure resulted in a 6.9-fold of
control level of protein while the 2000 ppm TCE group had a 2.3-fold level. CYP2E1 +/+ mice
had a ~ 50% higher control level than CYP2E1 mice and after TCE exposure the level of
peroxisomal bifunctional protein expression was 44-fold of control at 1000 ppm TCE and 40-
fold of control at 2000 ppm. Thus, CYP2E1-/- mice were reported to have less mRNA
expression and peroxisomal bifunctional protein formed than CYP2E1+/+ mice after TCE
exposure. However, there appeared to be more mRNA expression after 1000 ppm than 2000
ppm TCE in both groups and protein expression in the CYP2E1-/- mice. After 2000 ppm TCE,
there was similar peroxisomal bifunctional protein expression between the 1000 ppm and 2000
ppm TCE treated CYP2E1+/+ mice. Again this pattern was more similar to that of TCA
detection in the urine but not that of liver injury.
For VLCAD the expression of mRNA was similar between control and treated CYP2E1-
/- mice. For CYP2E1+/+ mice the control level of VLCAD mRNA expression was half that of
the CYP2E1-/- mice. After 1000 ppm TCE the mRNA level was 3.7-fold of control and after
2000 ppm TCE the mRNA level was 3.1-fold of control. For VLCAD protein expression was
02/12/09
E-44
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1.8-fold of control after 1000 ppm and 1.6-fold of control after 2000 ppm in CYP2E1-/- mice.
The control level of VLCAD protein in CYP2E1+/+ mice appeared to be 1.2-fold control
CYP2E1-/- mice After 1000 ppm TCE treatment the CYP2E1 -/- mice were reported to have
3.8-fold of control VLCAD protein levels and after 2000 ppm TCE treatment to have 3.9-fold of
control protein levels. Thus, although showing no increase in mRNA there was an increase in
VLCAD protein levels that was similar between the two TCE exposure groups in CYP2E1-/-
mice. Both VLCAD mRNA and protein levels were greater in CYP2E1 +/+ mice than CYP2E1-
/- mice after TCE exposure. This was not the case for peroxisomal bifunctional protein. The
magnitudes of TCE-induced increases in mRNA and protein increases were similar between the
1000 ppm and 2000 ppm TCE exposure concentrations, a pattern more similar to TCA detection
in the urine but not that of liver injury.
Finally, for CYP4A10 mRNA expression, there was an increase in expression after TCE
treatment of 3-fold for 1000 ppm and 5-fold after 2000 ppm in CYP2E1-/- mice. Thus, although
the enzyme assumed to be primarily responsible for TCE metabolism to TCA was missing, there
was still a response for the mRNA of this enzyme commonly associated with PPARa activation.
Of note is that urinary concentrations of TCA were not zero after TCE exposure in CYP2E1-/-
mice. Both 1000 and 2000 ppm TCE exposure resulted in ~ 0.44 mg TCA after 1 day or about
15 - 22% of that observed in CYP2E1+/+ mice. Thus, some metabolism of TCE to TCA is
taking place in the null mice, albeit at a reduced rate. For CYP2E1+/+ mice, 1000 ppm TCE
resulted in an 8.3-fold of control level of CYP4A10 mRNA and 2000 ppm TCE resulted in a 9.3-
fold of control level. The authors did not perform an analysis of CYP4A10 protein. The authors
state that "in particular, the mRNA levels of microsomal enzyme CYP4A10 significantly
increased in CYP2E1+/+ mice after TCE exposure in a dose-dependent manner." However, the
2-fold difference in TCE exposure concentrations did not result in a similar difference in
response as shown above. Both resulted in ~ 9-fold of control response in CYP2E1+/+ mice. As
with PPARa, peroxisomal bifunctional protein, and VLCAD, the response was more similar to
that of TCA detection in the urine and not measured of hepatic toxicity. These data are CYP2E1
metabolism of TCE is important in the manifestation of TCE liver toxicity, however it also
suggests that effects other than TCA concentration and indicators of PPARa are responsible for
acute hepatotoxicity resulting from very high concentrations of TCE.
The NFkB family and IicBa was also examined for mRNA and protein expression. These
cell signaling molecules are involved in inflammation and carcinogenesis and are discussed in
Section 3.3.3.3. and 3.4.1.4. Given that presence of hepatocellular necrosis in some of the
CYP2E1+/+ mice to varying degrees, inflammatory cytokines and cell signaling pathways would
be expected to be activated. The authors reported that "overall, TCE exposure did not
significantly increase the expression of p65 and p50 mRNAs in either CYP2E1+/+ or CYP2E1 -
/- mice... However, p52 mRNA expression significantly increased in the 2000 ppm group of
CYP2E1+/+ mice, and correlation analysis showed that a significant positive relationship existed
between the expression of NFkB p52 mRNA and plasma ALT activity.., while no correlation
was seen between NFkB p64 or p50 and ALT activity (data not shown)." The authors also note
02/12/09
E-45
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
that TCE treatments "did not increase the expression of TNFR1 and TNFR2 mRNA in
CYP2E1+/+ and CYP2E1 -/- mice (data not shown)."
A more detailed examination of the data reveals that there was a similar increases in p65,
p50, and p52 mRNA expression increases with TCE treatment in CYP2E1+/+ mice at both TCE
exposure levels. However, only p52 levels for the 2000 ppm-exposed mice were reported to be
statistically significant (see comment above about the statistical power of the experimental
design and variability between animals). For 1000 ppm TCE exposure the levels of p65, p50,
and p52 mRNA expression were 1.5-fold, 1.8-fold, and 2.0-fold of control. For 2000 ppm TCE
the levels of p65, p50, and p52 mRNA expression were 1.8-fold, 1.8-fold, and 2.1-fold of
control. Thus, there was generally a similar response in all of these indicators of NFkB mRNA
expression in CYP2E1 +/+ mice that was mild with little to no difference between the 1000 ppm
and 2000 ppm TCE exposure levels. For IicBa mRNA expression there was not difference
between control and treatment groups for either type of mice. For CYP2E1-/- mice there
appeared to be a ~ 50% decrease in P52 mRNA expression in mice treated with both exposure
concentrations of TCE. The authors plotted the relationship between p52 mRNA and plasma
ALT concentration for both CYP2E1 -/- and CYP2E1 +/+ mice together and claimed the
correlation coefficient (r= 0.5075) was significant. However, of note is that none of the CYP2E1
-/- mice were reported to have either hepatic necrosis or significant increases in ALT detection.
For protein expression the authors showed results for p50 and p42 proteins. The control
CYP2E1-/- mice appeared to have a slightly lower level of p50 protein expression (~ 30%) with
a much larger increase in p52 protein expression (i.e., 2.1-fold) than CYP2E1+/+ mice. There
appeared to be a 2-fold increase in p50 protein expression after both 1000 ppm and 2000 ppm
TCE exposures in the CYP2E1 +/+ mice and a similar increase in p52 protein levels (i.e., 1.9-
fold and 2.5-fold of control for 1000 ppm and 2000 ppm TCE exposures, respectively). Thus,
the magnitude of mRNA and protein levels were similar for p50 and p52 in CYP2E1 +/+ mice
and there was no difference between the 1000 ppm and 2000 ppm treatments. For the CYP2E1 -
/- mice there was a modest increase in p50 protein after TCE exposure (1.1-fold and 1.3-fold of
control for 1000 ppm and 2000 ppm respectively) and a slight decrease in p52 protein (0.76-fold
and 0.79-fold of control). There was little evidence that the patterns of either expression or
protein production of NFkB family and IicBa corresponded to the markers of hepatic toxicity or
that they exhibited a dose-response. The authors note that although he expression of p50 protein
increased in CYP2E1+/+ mice, "the relationship between p50 protein and ALT levels was not
significant (data not shown)." For TNFR1 there appeared to be less protein expression in the
CYP2E1 +/+ mice than the CYP2E1-/- mice (i.e., the null mice levels were 1.8-fold of the wild-
type mice levels). Treatment with TCE resulted in mild decrease of protein levels in the
CYP2E1 -/- mice and a 1.4-fold and 1.7-fold of control level in the CYP2E1+/+ mice for 1000
ppm and 2000 ppm levels, respectively. For p65, although TCE treatment-related effects were
reported, of note the levels of protein were 2.4 higher in the CYP2E1+/+ mice than the CYP2E1-
/- mice. Thus, protein levels of the NFkB family appeared to have been altered in the knockout
mice. Also, as noted in Section 3.4.1.4, the origin of theNF-KB is crucial as to its effect in the
liver and the results of this report are for whole liver homogenates that contain parenchymal as
02/12/09
E-46
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
well as nonparenchymal cell and have been drawn from liver that are heterogeneous in the
magnitude of hepatic necrosis. The authors suggest that "TCA may act as a defense against
hepatotoxicity cause by TCE-delivered reactive metabolite(s) via PPARa in CYP2E1+/+ mice."
However, the data from this do not support such an assertion.
2.2. Subchronic and Chronic Studies of TCE
For the purposes of this discussion, studies of duration of 4 weeks or more are considered
subchronic. Like those of shorter duration, there is variation in the depth of study of liver
changes induced by TCE with many of the longer duration studies focused on the induction of
liver cancer. Many subchronic studies were conducted a high doses of TCE that caused toxicity
with limited reporting of effects. Similar to acute studies some of the subchronic and chronic
studies have detailed examinations of the TCE-induced liver effects while others have reported
primarily liver weight changes as a marker of TCE-response. Similar issues also arise with the
impact of differences in initial and final body weights between control and treatment groups on
the interpretation of liver weight gain as a measure of TCE-response. For many of the
subchronic inhalation studies, issues associated with whole body exposures make determination
of dose levels difficult. For gavage experiments, death from gavage dosing, especially at higher
TCE exposures, is a recurring problem and, unlike inhalation exposures, the effects of vehicle
can also be at issue for background liver effects. Concerns regarding effects of oil vehicles,
especially corn oil, have been raised with Kim et al. (1990) noting that a large oil bolus will not
only produce physiological effects, but alter the absorption, target organ dose, and toxicity of
VOCs. Charbonneau et al. (1991) reported that corn oil potentiates liver toxicity from acetone
administration that is not related to differences in acetone concentration. Several oral studies in
particular document that use of corn oil gavage induces a different pattern of toxicity, especially
in male rodents (see Merrick et al., 1989, Section 2.2.1. below). Several studies listed below
report the effects of hepatocellular DNA synthesis and indices of lipid peroxidation (i.e., Channel
et al., 1998) are especially subject to background vehicle effects. Rusyn et al (1999) report that a
single dose of dietary corn oil increases hepatocyte DNA synthesis 24 hours after treatment by ~
3.5 fold, activation of NF-kB to a similar extent ~ 2 hours after treatment almost exclusively in
Kupffer cells, a ~ 3-4 fold increase in hepatocytes after 8 hours, and increased in TNFa mRNA
between 8 and 24 hours after a single dose in female rats. In regard to studies that have used the
i.p. route of administration, as noted by Kawamoto et al (1988) (see Section 2.2.10 below),
injection of TCE may result in paralytic ileus and peritonitis and that subcutaneous treatment
paradigm will result in TCE not immediately being metabolized but retained in the fatty tissue.
Wang and Stacey (1990) state that "intraperitoneal injection is not particularly relevant to
humans" and that intestinal interactions require consideration in responses such as increase
serum bile acid (see Section 2.3.5 below).
2.2.1. Merrick et al. 1989
The focus of this study was the examination of potential differences in toxicity or orally
gavaged TCE administered in corn oil an aqueous vehicle in B6C3F1 mice. As reported by
02/12/09
E-47
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Melnick et al. (1987) above, corn oil administration appeared to have an effect on peroxisomal
enzyme induction. TCE (99.5% purity) was administered in corn oil or an aqueous solution of
20% Emulphor to 14-17 week old mice (n = 12/group) at 0, 600, 1200 and 2400 mg/kg/day
(males) and 0, 450, 900, and 1800 mg/kg/day (females) 5 times a week for 4 weeks. The authors
state that due to "varying lethality in the study, 10 animals per dose group were randomly
selected (where possible) among survivors for histological analysis." Hepatocellular lesions
were characterized "as a collection of approximately 3-5 necrotic hepatocytes surrounded by
macrophages and polymorphonuclear cells and histopathological grading was reported as based
on the number of necrotic lesions observed in the tissue sections: 0 = normal; 1 = isolated lesions
scattered throughout the section; 2 = one to five scattered clusters of necrotic lesions; 3 = more
than five scattered clusters of necrotic lesions; and 4 = clusters of necrotic lesions observed
throughout the entire section." The authors described lipid scoring of each histological section
as "0 = no Oil-Red O staining present; 1 = less than 10% staining; 2 = 10-25%) staining; 3 = 25-
30%o staining; and 4 = greater than 50% staining."
The authors reported dose-related increases in lethality in both males and females
exposed to TCE in Emulphor with all male animals dying at 2400 mg/kg/day with 8/12 females
dying at 1800 mg/kg/day. In both males and females, 2/12 animals also died at the next highest
dose as well with no unscheduled deaths in control or lowest dose animals. For corn oil gavaged
mice, there were 1-2 animals in each TCE treatment groups of male mice that died while there
were no unscheduled deaths in female mice. The authors state that lethality occurred within the
first week after chemical exposure. The authors present data for final body weight and
liver/body weight values for 4 weeks of exposure and list the number of animals per group to be
10-12 for corn oil gavaged animals and the reduced number of animals in the Emulphor gavaged
animals reflective of lethality and limiting the usefulness of this measure at the highest doses
(i.e., 1800 mg/kg/day for female mice). In mice treated with TCE in Emulphor gavage, the final
body weight of control male animals appeared to be lower than those which were treated with
TCE while for female mice the final body weights were similar between treated and control
groups. For male mice treated with Emulphor, body weights were 22.8 ± 0.8, 25.3 ± 0.5, and
24.3 ± 0.4 g for control, 600 mg/kg/day, and 1200 mg/kg/day and for female mice body weights
were 20.7 ± 0.4, 21.4 ± 0.3, and 20.5 ± 0.3 g for control, 450 mg/kg/day, and 900 mg/kg/day of
TCE.
For % liver/body weight ratios, male mice were reported to have 5.6 ± 0.2, 6.6 ±0.1, and
7.2 ± 0.2 % for control, 600 mg/kg/day, and 1200 mg/kg/day and for female mice were 5.1 ± 0.1,
5.8 ± 0.1, and 6.5 ± 0.2 % for control, 450 mg/kg/day, and 900 mg/kg/day of TCE. These values
represent 1.11-fold and 1.07-fold of control for final body weight in males exposed to 600 and
1200 mg/kg/day and 1.18-fold and 1.29-fold of control for %> liver/body weight, respectively.
For females, they represent 1.04-fold and 0.99-fold of control for final body weights in female
exposed to 450mg/kg/day and 900 mg/kg/day and 1.14-fold and 1.27-fold of control for %>
liver/body weight, respectively.
02/12/09
E-48
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
In mice treated with corn oil gavage the final body weight of control male mice was
similar to the TCE treatment groups and higher than the control value for male mice given
Emulphor vehicle (i.e., 22.8 ± 0.8 g for Emulphor control vs. 24.3 ± 0.6 g for corn oil gavage
controls or a difference of ~ 7%). The final body weights of female mice were reported to be
similar between the vehicles and TCE treatment groups. The baseline % liver/body weight was
also lower for the corn oil gavage control male mice (i.e., 5.6 % for Emulphor vs. 4.7 % for corn
oil gavage or a difference of-19% that was statistically significant). Although the final body
weights were similar in the female control groups, the % liver/body weight was greater in the
Emulphor vehicle group (5.1 ± 0.1 % in Emulphor vehicle group vs. 4.7 ±0.1% for corn oil
gavage or a difference of ~ 9 % that was statistically significant). For male mice treated with
corn oil, final body weights were 24.3 ± 0.6, 24.3 ± 0.4, 25.2 ± 0.6 and 25.4 ± 0.5 g for control,
600 mg/kg/day, 1200 mg/kg/day and 2400 mg/kg/day, and for female mice body weights were
20.2 ± 0.3, 20.8 ± 0.5, 21.8 ± 0.3 g, and 22.6 ± 0.3 g for control, 450 mg/kg/day, 900 mg/kg/day,
and 1800 mg/kg/day of TCE. For % liver/body weight ratios, male mice were reported to have
4.7 ± 0.1, 6.4 ± 0.1, 7.7 ±0.1 % , and 8.5 ± 0.2 % for control, 600 mg/kg/day, 1200 mg/kg/day,
and 2400 mg/kg/day and for female mice were 4.7 ± 0.1, 5.5 ± 0.1, 6.0 ± 0.2 %, and 7.2 ±0.1 %
for control, 450 mg/kg/day, 900 mg/kg/day, and 1800 mg/kg/day of TCE. These values
represent 1.0-fold, 1.04-fold, and 1.04-fold of control for final body weight in males exposed to
600, 1200 mg/kg/day and 2400 mg/kg/day TCE and 1.36-fold, 1.64-fold, and 1.81-fold of
control for % liver/body weight, respectively. For females they represent 1.03-fold, 1.08-fold,
and 1.12-fold of control for body weight in female exposed to 450 mg/kg/day, 900 mg/kg/day,
and 1800 mg/kg/day and 1.17-fold, 1.28-fold, and 1.53-fold of control for % liver/body weight,
respectively.
Because of premature mortality, the difference in TCE treatment between the highest
doses that are vehicle-related cannot be determined. The decreased final body weight and
increased % liver/body weight ratios in the Emulphor control animals make comparisons of the
exact magnitude of change in these parameters due to TCE exposure difficult to determine as
well as differences between the vehicles. The authors did not present data for age-matched
controls which did not receive vehicle so that the effects of the vehicles cannot be determined
(i.e., which vehicle control values were most similar to untreated controls given that there was a
difference between the vehicle controls). A comparison the % liver/body weight ratios at
comparable doses between the two vehicles shows little difference in TCE-induced liver weight
increases in female mice. However, the corn oil vehicle group was reported to have a greater
increase in comparison to controls for male mice treated with TCE at the two lower dosage
groups. Given that the control values were approximately 19% higher for the Emulphor group,
the apparent differences in TCE-dose response may have reflected the differences in the control
values rather than TCE exposure. Because controls without vehicle were not examined, it cannot
be determined whether the difference in control values was due to vehicle administration or
whether a smaller or younger group of animals was studied on one of the control groups. The
body weight of the animals was also not reported by the authors at the beginning of the study so
that the impact of initial differences between groups vs. treatment cannot be accurately
determined.
02/12/09
E-49
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Serum enzyme activities for ALT, AST and LDH (markers of liver toxicity) showed that
there was no difference between vehicle groups at comparable TCE exposure levels for male or
female mice. Enzyme levels appeared to be elevated in male mice at the higher doses (i.e., 1200
and 2400 mg/kg/day for ALT and 2400 mg/kg/day for AST) with corn oil gavage inducing
similar increases in LDH levels at 600 mg/kg/day, 1200 mg/kg/day and 2400 mg/kg/day TCE.
For ALT and AST there appeared to be a dose-related increase in male mice with the 2400
mg/kg treatment group having much greater levels than the 1200 mg/kg group. In Emulphor
treatment groups there was a similar increase in ALT levels in males treated with 1200 mg/kg
TCE as with those treated with corn oil and those increases were significantly elevated over
control levels. For LDH levels there were similar increase at 1200 mg/kg TCE for male mice
treated using either Emulphor or corn oil. The authors report that visible necrosis was observed
in 30-40% of male mice administered TCE in corn oil but not that there did not appear to be a
dose-response (i.e., the score for severity of necrosis was reported to be 0, 4, 3, and 4 for corn oil
control, 600 mg/kg/day, 1200 mg/kg/day, and 2400 mg/kg/day treatment groups from 10 male
mice in each group). No information in regard to variation between animals was given by the
authors. For male mice given Emulphor gavage the extent of necrosis was reported to be 0, 0,
and 1 for 0, 600 mg/kg/day and 1200 mg/kg/day TCE exposure, respectively. For female mice,
the extent of necrosis was reported to be 0 for all control and TCE treatment groups using either
vehicle. Thus, except for LDH levels in male mice exposed to TCE in corn oil there was not a
correlation with the extent of necrosis and the increases in ALT and AST enzyme levels.
Similarly there was an increase in ALT levels in male mice treated with 1200 mg/kg/day
exposure to TCE in Emulphor that did not correspond to increased necrosis. For Oil-Red O
staining there was a score of 2 in the Emulphor treated control male and female mice while 600
mg/kg/day TCE exposure in Emulphor gavaged male mice and 900 mg/kg/day TCE in corn oil
gavaged female mice had a score of 0, along with the corn oil gavage controls in male mice. For
female control mice treated with corn oil gavage, the staining was reported to have a score of 3.
Thus, there did not appear to be a dose-response in Oil-Red oil staining although the authors
claimed there appeared to be a dose-related increase with TCE exposure. The authors described
lesions produced by TCE exposure as "focal and were surrounded by normal parenchymal tissue.
Necrotic areas were not localized in any particular regions of the lobule. Lesions consisted of
central necrotic cells encompassed by hepatocytes with dark eosinophilic staining cytoplasm
which progressed to normal-appearing cells. Areas of necrosis were accompanied by localized
inflammation consisting of macrophages and polymorphonuclear cells." No specific
descriptions of histopathology of mice given Emulphor were provided in terms of effects of the
vehicle or TCE treatment. The scores for necrosis was reported to be only a 1 for the 1200
mg/kg concentration of TCE in male mice gavaged with Emulphor but 3 for male mice given the
same concentration of TCE in corn oil. However enzyme levels of ALT, AST, and LDH were
similarly elevated in both treatment groups.
These results do indicate that administration of TCE for 4 weeks via gavage using
Emulphor resulted in mortality of all of the male mice and most of the female mice at a dose in
corn oil that resulted in few deaths. Not only was there a difference in mortality, but vehicle also
affected the extent of necrosis and enzyme release in the liver (i.e, Emulphor vehicle caused
02/12/09
E-50
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mortality as the highest dose of TCE in male and female mice that was not apparent from corn
oil gavage, but Emulphor and TCE exposure induced little if any focal necrosis in males at
concentrations of TCE in corn oil gavage that caused significant focal necrosis). In regard to
liver weight and body weight changes, TCE exposure in both vehicles at non-lethal doses
induced increased % liver/body weight changes male and female mice that increased with TCE
exposure level. The difference in baseline control levels between the two vehicle groups
(especially in males) make a determination of the quantitative difference vehicle had on liver
weight gain problematic although the extent of liver weight increase appeared to be similar
between male and female mice given TCE via Emulphor and female mice given TCE via corn
oil. In general, enzymatic markers of liver toxicity and results for focal hepatocellular necrosis
were not consistent and did not reflect dose-responses in liver weight increases. The extent of
necrosis did not correlate with liver weight increases and was not elevated by TCE treatment in
female mice treated with TCE in either vehicle, or in male mice treated with Emulphor. There
was a reported difference in the extent of necrosis in male mice given TCE via corn oil and
female mice given TCE via corn oil but the necrosis did not appear to have a dose-response in
male mice. Female mice given corn oil and male and female mice given TCE in Emulphor had
no to negligible necrosis although they had increased liver weight from TCE exposure.
2.2.2. Goel et al. 1992
The focus of this study was the description of TCE exposure related changes in mice after
28 days of exposure with regard to TCE-induced pathological and liver weight change. Male
Swiss mice (20-22 g body weight or 9% difference) were exposed to 0, 500, 1000 or 2000
mg/kg/day TCE (BDH analytical grade) by gavage in groundnut oil (n= 6 per group) 5 days a
week for 28 days. The ages of the mice were not given by the authors. Livers were examined
for "free -SH contents", total proteins, catalase activity, acid phosphatase activity, and "protein
specific for peroxisomal origin of approx, 80 kd." The authors report no statistically significant
change in body weight with TCE treatment but a significant increase in liver weight. Body
weight (mean ± SE) was reported to be 32.67 ± 1.54, 31.67 ± 0.61, 33.00 ± 1.48, and 27.80 ±
1.65 g from exposure to oil control, 500, 1000, and 2000 mg/kg/day TCE, respectively. There
was a 15% decrease in body weight at the highest exposure concentration of TCE that was not
statistically significant, but the low number of animals examined limits the power to detect a
significant change. The % relative liver/ body weight was reported to be 5.29 ± 0.48, 7.00 ±
0.36, 7.40 ± 0.39, and 7.30 ± 0.48 % from exposure to oil control, 500, 1000, and 2000
mg/kg/day TCE, respectively. This represents 1.32-fold, 1.41-fold, and 1.38-fold of control in %
liver/body weight for 500, 1000, and 2000 mg/kg/day TCE, respectively. The "free -SH
content" in [jmole -SH/g tissue was reported to be 5.47 ± 0.17, 7.46 ± 0.21, 7.84 ± 0.34, and
7.10 ± 0.34 from exposure to oil control, 500, 1000, and 2000 mg/kg/day TCE, respectively.
This represents 1.37-fold, 1.44-fold, and 1.30-fold of control in -SH/g tissue weight for 500,
1000, and 2000 mg/kg/day TCE, respectively. Total protein content in the liver in mg/g tissue
was reported to be 170 ± 3, 183 ± 5, 192 ± 7, and 188 ± 3 from exposure to oil control, 500,
1000, and 2000 mg/kg/day TCE, respectively. This represents 1.08-fold, 1.13-fold, and 1.11-
fold of control in total protein content for 500, 1000, and 2000 mg/kg/day TCE, respectively.
02/12/09
E-51
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Thus, the increases in liver weight, "free -SH content" and increase protein content were
generally parallel and all suggest that liver weight increases had reached a plateau at the 1000
mg/kg/day exposure concentration perhaps reflecting toxicity at the highest dose as demonstrated
by decreased body weight in this study.
The enzyme activities of 8-ALA dehydrogenase ("a key enzyme in heme biosynthesis"),
catalase, and acid phosphatase were assayed in liver homogenates. Treatment with TCE
decreased 8-ALA dehydrogenase activity to a similar extent at all exposure levels (32 - 35%
reduction). For catalase the activity as units of catalase/mg protein was reported to be 25.01 ±
1.81, 32.46 ± 2.59, 41.11 ± 5.37, and 33.96 ± 3.00 from exposure to oil control, 500, 1000, and
2000 mg/kg/day TCE, respectively. This represents 1.30-fold, 1.64-fold, and 1.36-fold in
catalase activity for 500, 1000, and 2000 mg/kg/day TCE, respectively. The increasing
variability in response with TCE exposure concentration is readily apparent from this data as is
the decrease at the highest dose, perhaps reflective of toxicity. For acid phosphatase activity in
the liver there was a slight increase (5-11%) with TCE exposure that did not appear to be dose-
related.
The authors report that histologically, "the liver exhibits swelling, vacuolization,
widespread degeneration/necrosis of hepatocytes as well as marked proliferation of endothelial
cells of hepatic sinusoids at 1000 and 2000 mg/kg TCE doses." Only one figure is given at the
light microscopic level in which it is impossible to distinguish endothelial cells from Kupffer
cells and no quantitative measures or proliferation were examined or reported to support the
conclusion that endothelial cells are proliferating in response to TCE treatment. Similarly, no
quantitation regarding the extent or location of hepatocellular necrosis is given. The presence or
absence of inflammatory cells is not noted by the authors as well. In terms of white blood cell
count, the authors note that it is slightly increased at 500 mg/kg/day but decreased at 1000 and
2000 mg/kg/day TCE, perhaps indicating macrophage recruitment from blood to liver and
kidney, which was also noted to have pathology at these concentrations of TCE.
2.2.3. Kjellstrand et al. 1981
This study was conducted in mice, rats, and gerbils and focused on the effects of 150 ppm
TCE exposure via inhalation on body and organ weight. No other endpoints other than organ
weights were examined in this study and the design of the study is such that quantitative
determinations of the magnitude of TCE response are very limited. NMRI mice (weighing ~ 30
g with age not given), Sprague-Dawley rats (weighing ~ 200 g with age not given, and
Mongolian gerbils (weighing ~ 60 g with age not given) were exposed to 150 ppm TCE
continuously. Mice were exposed for 2, 5, 9, 16, and 30 days with the number of exposed
animals and controls in the 2, 5, 9 and 16 days groups being 10. For 30-day treatments there
were 2 groups of mice containing 20 mice per group and one group containing 12 mice per
group. In addition there was a group of mice (n= 15) exposed to TCE for 30 days and then
examined 5 days after cessation of exposure and another group (n= 20) exposed to TCE for 30
days and then examined 30 days after cessation of exposure. For rats there were 3 groups
02/12/09
E-52
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
exposed to TCE for 30 days which contained 24, 12 and 10 animals per group. For gerbils there
were 3 groups exposed to TCE for 30 days which contained 24, 8 and 8 animals per group. The
groups were reported to consist of equal numbers of males and female but for the mice exposed
to TCE for 30 days and then examined 5 days later, the number was 10 males and 5 females.
Body weights were reported to be recorded before and after the exposure period. However the
authors state "for technical reasons the animals within a group were not individually identified,
i.e., we did not know which initial weight in the group corresponded to which final one." They
authors state that this design presented problems in assessing the precision of the estimate. They
go on to state that rats and gerbils were partially identifiable as the animals were housed 3 to a
cage and cage averages could be estimated. Not only were mice in one group housed together
but "even worse: at the start of the experiment, the mice in M2 [sic, group exposed for 2 days]
and M9 [sic group exposed for 9 days] were housed together, and similarly M5 [group exposed
for 5 days] and M16 [sic group exposed for 16 days]. Thus, we had, e.g., 10 initial weights for
exposed female mice in M2 and M9 where we could not identify those 5 which were M2
weights. Owing to this bad design (forced upon us by the lack of exposure units), we could not
study weight gains for mice and so we had to make do with an analysis of final weights." The
problems with the design of this study are obvious from the description given by the authors
themselves. The authors state that they assumed that the larger the animal the larger the weight
of its organs so that all organ weights were converted into relative weights as percentage of body
weight. The fallacy of this assumption is obvious, especially if there was toxicity that decreased
body weight and body fat but at the same time caused increased liver weight as has been
observed in many studies at higher doses of TCE. In fact, Kjellstrand et al. (1983b) report that a
150 ppm TCE exposure for 30 days does significantly decreases body weight while elevating
liver weight in a group of 10 male NMRI mice. Thus, the body weight estimates from this study
are inappropriate for comparison to those in studies where body weights were actually measured.
The liver/body weight ratios that would be derived from such estimates of body weights would
be meaningless. The group averages for body weight reported for female mice at the beginning
of the 30 - day exposure varied significantly and ranged from 23.2 to 30.2 g (~ 24%). For
males the group averages ranged from 27.3 to 31.4 g (~ 14%). For male mice there was no
weight estimate for the animals that were exposed for 30 days and then examined 30 days after
cessation of exposure.
The authors only report relative organ weight at the end of the experiment rather than the
liver weights for individual animals. Thus, these values represent extrapolations based on to
what body weight may have been. For mice that were exposed to TCE for 30 days and the
examined after 30 days of exposure, male mice were reported to have "relative organ weight" for
liver of 4.70 ± 0.10 % vs. 4.27 ±0.13 % for controls. However, there were no initial body
weights reported for these male mice and the body weights are extrapolated values. Female mice
exposed for 30 days and then examined 30 days after cessation of exposure were reported to
have "relative organ weights" for liver of 4.42 ± 0.11 % vs. 3.62 ± 0.09 %. The group average of
initial body weights for this group was reported by the authors. Although the initial body weight
for female control mice as a group average was reported to be similar between the female group
exposed to 30 days of TCE and sacrificed 30 days later and those exposed for 30 days and
02/12/09
E-53
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
sacrificed 5 days later (30.0 g vs. 30.8 g), the liver/body weight ratio varied significantly in
these controls (4.25 ± 0.19 vs. 3.62 ± 0.09) as did the number of animals studied (5 female mice
in the animals sacrificed after 5 days exposure vs. 10 female mice in the group sacrificed after 30
days exposure). In addition, although there were differences between the 3 groups of mice
exposed to TCE for 30 days and then sacrificed immediately, the authors present the data for
extrapolated liver/body weight as pooled results between the 3 groups. In comparison to control
values, the authors report 1.14-fold, 1.35-fold, 1.58-fold, 1.47-fold, and 1.75-fold of control for
% liver/body weight using body weight extrapolated values in male mice at 2, 5, 9, 16, and 30
days of TCE exposure, respectively. For females they report 1.27-fold, 1.28-fold, 1.49-fold,
1.41-fold, and 1.74-fold of control at 2, 5, 9, 16, and 30 days of TCE, respectively. Although the
authors combine female and male relative increases in liver weight in a figure, assign error bars
around these data point, and attempt to draw assign a time-response curve to it, it is clear from
this data, especially for female mice, do not display time-dependent increase in liver/body weight
from 5 to 16 days of exposure and that a comparison of results between 5 animals and 26 is very
limited in interpretation. Of note is the wide variation in the control values for relative
liver/body weight. For male mice there did not seem to be a consistent pattern with increasing
duration of the experiment with values at 4.61, 5.15, 5.05, 4.93, and 4.04 % for 2, 5, 9, 16, and
30 day exposure groups. This represented a difference of - 27%. For female mice the relative
liver/body weight was 4.14, 4.58, 4.61, 4.70, and 3.99 % for 2, 5, 9, 16, and 30 day exposure
groups. Thus, it appears that the average relative liver/body weight % was higher in the 5, 9, and
16 day treatment group for both genders than that to the 30 day group and was consistent
between these days. There is no apparent reason for there to be such large difference between 16
day and 30-day treatment groups due to increasing age of the animals. Of note is that for the
control groups pared with animals treated for 30 days and then examined 30 days later, the male
mice had increases in relative liver/body weight (4.27% vs. 4.04%) but that the females had a
decrease (3.62% vs. 3.99%). Such variation between controls does not appear to be age and size
related but to variations in measure or extrapolations, which can affect comparisons between
treated and untreated groups and add more uncertainty to the estimates.
The number of mice in the groups exposed to 2 though 16 days were only 5 animals for
each gender in each group while the number of animals reported in the 30 -day exposure group
numbered 26 for each gender.
For animals exposed to 30 days and then examined after 5 or 30 days, male mice were
reported to have % liver/body weight 1.26-fold and 1.10-fold of control after 5 and 30 days
cessation of exposure while female mice were reported to have values of 1.14-fold and 1.22-fold
of control after 5 and 30 days cessation of exposure, respectively. Again, the male mice exposed
for 30 days and then examined after 30 days of cessation of exposure did not have reported
initial body weights giving this value a great deal of uncertainty. Thus while liver weights
appeared to increase during 30 days of exposure to TCE and decreased after cessation of
exposure in both genders of mice, the magnitudes of the increases and decrease cannot be
determined from this experimental design. Of note is that liver weights appeared to still be
elevated after 30 days of cessation exposure.
02/12/09
E-54
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
In regard to initial weights, the authors report that the initial weight of the rats were
different in the 3 experiments they conducted with them and state that "in those 2 where
differences were found in females, their initial weights were about 200 g and 220 g, respectively,
while the corresponding weights were only about 160 g in that experiment where no differences
were found." The differences in initial body weight of the rat groups were significant. In
females group averages were 198, 158, and 224 g, for groups 1, 2, and 3, respectively and for
males group averages were 222, 166, and 248 g for groups 1,2, and 3 respectively. This
represents as much as a 50% difference in initial body weights between these TCE treatment
groups. Control values varied as well with group averages for controls ranging froml67g for
group 2 to 246 g for group 3 at the start of exposure. For female rats control groups ranged from
158 to 219 g at the start of the experiment. The number of animals in each group varied greatly
as well making quantitative comparison even more difficult with the numbers varying between 5
and 12 for each gender in rats exposed for 30 days to TCE. The authors pooled the results for
these very disparate groups of rats in their reporting of relative organ weights. They reported
1.26-fold and 1.21-fold of control in male and female rat % relative liver/body weight after 30
days of TCE exposure. However, as stated above, these estimates are limited in their ability to
provide a quantitative estimate of liver weight increase due to TCE.
There were evidently differences between the groups of gerbils in response to TCE with
one group reported to have larger weight gain than control and the other 2 groups reported to not
show a difference by the authors. Of the 3 groups of gerbils, group 1 contained 12 animals per
gender but groups 2 and 3 only 4 animals per gender. As with the rat experiments, the initial
average weights for the groups varied significantly (30% in females and males). The authors
pooled the results for these very disparate groups of gerbils in their reporting of relative organ
weights as well. They reported a nearly identical increase in relative liver/body weight increase
for gerbil (1.22-fold of control value in males and 1.25-fold in females) as for the rat after 30
days of TCE exposure. However, similar caveats should be applied in the confidence in this
experimental design to determine the magnitudes of response to TCE exposure.
2.2.4. Woolhiser et al. 2006.
An unpublished report by Woolhiser et al (2006) was received by the EPA to fill the
"priority data needed" (PDN) for the immunotoxicity of TCE as identified by the Agency for
Toxic Substances and Disease Registry (ATSDR) and designed to satisfy U.S. EPA OPPTS
870.7800 Immunotoxicity Test Guidelines. The study was conducted on behalf of the
Halogenated Solvents Industry Alliance and has been submitted to the EPA but not published.
Although conducted as an immunotoxicity study, it does contain information regarding liver
weight increases in female Sprague Dawley (CD) female rats exposed to 0, 100 ppm, 300 ppm,
and 1000 ppm TCE for 6 hours/day, 5 days/week for 4 weeks. The rats were 7 weeks of age at
the start of the study. The report gives data for body weight and food weight for 16 animals per
exposure group and the mean body weights ranged between 181.8g to 185.5 g on the first day of
the experiment. Animals were weight pre-exposure, twice during the first week, and then "at
02/12/09
E-55
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
least weekly throughout the study." All rats were immunized with a single i.v. injection of
SRBC via the tail vein at day 25. Liver weights were taken and samples of liver retained "should
histopathological examinination have been deemed necessary." But histopathological analysis
was not conducted on the liver.
The effect on body weight gain by TCE inhalation exposure was shown by 5 days and
continued for 10 days of exposure in the 300 ppm and 1000 ppm exposed groups. By day 28, the
mean body weight for the control group was reported to be 245.7 g but 234.4 g, 232.4 g, and
232.4 g for the 100 ppm, 300 ppm, and 1000 ppm exposure groups, respectively. Food
consumption was reported to be decreased in the day 1-5 measurement period for the 300 ppm
and 1000 ppm exposure groups and in the 5-10 day measurement period for the 100 ppm group.
Although body weight and food consumption data are available for 16 animals per exposure
group, for organ and organ/body weight summary data, the report gives information for only 8
rats per group. The report gives individual animal data in its appendix so that the data for the 8
animals in each group examined for organ weight changes could be examined separately. The
final body weights were reported to be 217.2 g, 212.4 g, 203.9 g, and 206.9 g for the control, 100
ppm, 300 ppm, and 1000 ppm exposure groups containing only 8 animals. For the 8-animal
exposure groups, the mean initial body weights were 186.6 g, 183.7 g, 181.6 g, and 181.9 g for
the control, 100 ppm, 300 ppm, and 1000 ppm exposure groups. Thus, there was a difference
from the initial and final body weight values given for the groups containing 16 rats and those
containing 8 rats. The ranges of initial body weights for the eight animals were 169.8 g - 204.3
g, 162.0 g - 191.2 g, 169.Og - 201.5 g, and 168.2 g - 193.7 g for the control, 100 ppm, 300 ppm,
and 1000 ppm groups. Thus, the control group began with a larger mean value and large range
of values (20% difference between highest and lowest weight rat) than the other groups.
In terms of the % liver/body weight ratios, an increase due to TCE exposure is reported in
female rats, although body weights were larger in the control group and the two higher exposure
groups did not gain body weight to the same extent as controls. The mean % liver/body weight
ratios were 3.23 %, 3.39 %, 3.44 %, and 3.65%, respectively for the control, 100 ppm, 300 ppm,
and 1000 ppm exposure groups. This represented 1.05-fold, 1.07-fold, and 1.13-fold of control
% liver/body weight changes in the 100 ppm, 300 ppm and 1000 ppm-exposure groups.
However, the small number of animals and the variation in initial animal weight limit the ability
of this study to determine statistically significant increases and the authors report that only the
1000 ppm group had statistically significant increased liver weight increases.
2.2.5. Kjellstrand et al. 1983(a)
This study examined seven strains of mice (wild, C57BL, DBA, B6CBA, A/sn, NZB and
NMRI) after continuous inhalation exposure to 150 ppm TCE for 30 days. "Wild" mice were
reported to be composed of "three different strains: 1. Hairless (HR) from the original strain, 2.
Swiss (outbred), and 3. Furtype Black Pelage (of unknown strain) " The authors do not state the
age of the animals prior to TCE exposure but state that weight-matched controls were exposed to
air only chambers. The authors state that "the exposure methods" have been described earlier
02/12/09
E-56
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(Kjellstrand et al., 1980) but only reference Kjellstrand et al. (1981). In both of this and the 1981
study, animals were continuously exposed with only a few hours of cessation of exposure noted a
week for change of food and bedding. Under this paradigm, there is the possibility of additional
oral exposure to TCE due to grooming and consumption of TCE on food in the chamber. The
study was reported to be composed of two independent experiments with the exception of strain
NMRI which had been studied in Kjellstrand et al 1981 & 1983b. The number of animals
examined in this study ranged from 3-6 in each treatment group. The authors reported
"significant difference between the animals intended for TCE-exposure and the matched controls
intended for air-exposure were seen in four cases (table 1.)", and stated that the grouping effects
developed during the 7-day adaptation period. Premature mortality was attributed to an accident
for one TCE-exposed DBA male and fighting to the deaths of two TCE-exposed NZB females
and one B6CBA male in each air exposed chamber. Given the small number of animals
examined in this study in each group, such losses significantly decrease the power of the study to
detect TCE-induced changes. The range of initial body weights between the groups of male
mice for all strains was between 18 g (as mean value for the A/sn strain) and 32 g (as mean value
for the B6CBA strain) or ~ 44%. For females the range of initial body weights between groups
for all strains was 15g (as mean value for the A/sn strain) and 24 g (as mean value for the DBA
strain) or -38%..
Rather than reporting % liver/body weight ratios or an extrapolated value, as was done in
Kjellstrand et al. (1981), this study only reported actual liver weights for treated and exposed
groups at the end of 30 days of exposure. The authors report final body weight changes in
comparison to matched control groups at the end of the exposure periods but not the changes in
body weight for individual animals. They report the results from statistical analyses of the
difference in values between TCE and air-exposed groups. A statistically significant decrease in
body weight was reported between TCE exposed and control mice in experiment 1 of the C57BL
male mice (~ 20% reduction in body weight due to TCE exposure). This group also had a slight
but statistically significant difference in body weight at the beginning of exposure with the
control group having a -5% difference in starting weight. There was also a statistically
significant decrease in body weight of 20% reported after TCE exposure in one group of male
B6CBA mice that did not have a difference in body weight at the beginning of the experiment
between treatment and control groups. One group of female and both groups of male A/sn mice
had statistically significant decreases in body weight after TCE exposure (10% for the females,
and 22 and 26% decreases in the two male groups) in comparison to untreated mice of the same
strain. The magnitude of body weight decrease in this strain after TCE treatment also reflects
differences in initial body weight as there were also differences in initial body weight between
the two groups of both treated and untreated A/sn males that were statistically significant, 17%
and 10%) respectively. One group of male NZB mice had a significant increase in body weight
after TCE exposure of 14% compared to untreated animals. A female group from the same
strain treated with TCE was reported to have a nonsignificant but 7% increase in final body
weight in comparison to its untreated group. The one group of male NMRI mice (n= 10) in this
study was reported to have a statistically significant 12% decrease in body weight compared to
controls.
02/12/09
E-57
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For the groups of animals with reported TCE exposure-related changes in final body
weight compared to untreated animals, such body weight changes may also have affected the
liver weights changes reported. The authors do not explicitly state that they did not record liver
and body weights specifically for each animal, and thus would be unable to determine liver/body
weight ratios for each, however they do state that he animals were housed 4-6 in each cage and
placed in exposure chambers together. The authors only present data for body and liver weights
as the means for a cage group in the reporting of their results. While this approach lends more
certainty in their measurements than the approach taken by Kjellstrand et al (1981) as described
above, the relative liver/body weights cannot be determined for individual animals. It appears
that the authors have tried to carefully match the body weights of the control and exposed mice
at the beginning of the experiment to minimize the effects of initial body weight differences and
distinguish the effects of treatment on body weight and liver weight. However, there is no ability
to determine liver/body weight ratios and adjust for difference in initial body weight from
changes due to TCE exposure. For the groups in which there was no change in body weight after
TCE treatment and in which there was no difference in initial body weight between controls and
TCE-exposed groups, the reporting of liver weight changes due to TCE exposure is a clearer
reflection of TCE-induced effects and the magnitude of such effects. Nevertheless the small
number of animals examined in each group is still a limitation on the ability to determine the
magnitude of such responses and there statistical significance.
In "wild type" mice there were no reported significant differences in the initial and final
body weight of male or female mice before or after 30 days of TCE exposure. For these groups
there was 1.76-fold and 1.80-fold of control values for liver weight in groups 1 and 2 for female
mice, and for males 1.84-fold and 1.62-fold of control values for groups 1 and 2, respectively.
For DBA mice there were no reported significant differences in the initial and final body weight
of male or female mice before or after 30 days of TCE exposure. For DBA mice there was 1.87-
fold and 1.88-fold of control for liver weight in groups 1 and 2 for female mice, and for males
1.45-fold and 2.00-fold of control for groups 1 and 2, respectively. These groups represent the
most accurate data for TCE-induced changes in liver weight not affected by initial differences in
body weight or systemic effects of TCE which resulted in decreased body weight gain. These
results suggest that there is more variability in TCE-induced liver weight gain between groups of
male than female mice.
The C57BL, B6CBA, NZB, and NMRI groups all had at least one group of male mice
with changes in body weight due to TCE exposure. The A/sn group not only had both male
groups with decreased body weight after TCE exposure (along with differences between exposed
and control groups at the initiation of exposure) but also a decrease in body weight in one of the
female groups. Thus, the results for TCE-induced liver weight change in these male groups also
reflect changes in body weight. These results suggest a strain-related increased sensitivity to
TCE toxicity as reflected by decreased body weight. For C57BL mice there was 1.65-fold and
1.60-fold of control for liver weight after TCE exposure was reported in groups 1 and 2 for
female mice, and for males 1.28-fold (the group with decreased body weight) and 1.82-fold of
02/12/09
E-58
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
control values for groups 1 and 2, respectively. For B6CBA mice there was 1.70-fold and 1.69-
fold of controls values for liver weight after TCE exposure in groups 1 and 2 for female mice,
and for males 1.21-fold (the group with decreased body weight) and 1.47-fold of control values
reported for groups 1 and 2, respectively. For the NZB mice there was 2.09-fold (n=3) and 2.08-
fold of control values for liver weight after TCE exposure in groups 1 and 2 for female mice and
for males 2.34-fold and 3.57-fold (the group with increased body weight) of control values
reported for groups 1 and 2, respectively. For the NMRI mice, whose results were reported for
one group with 10 mice, there was 1.66-fold of control value for liver weight after TCE exposure
for female mice and for males 1.68-fold of control value reported (a group with decreased body
weight). Finally, for the A/sn strain that had decreased body weight in all groups but one after
TCE exposure and significantly smaller body weights in the control groups before TCE exposure
in both male groups, the results still show TCE-related liver weight increases. For the As/n mice
there was 1.56-fold and 1.72-fold (a group with decreased body weight) of control value for liver
weight in groups 1 and 2 for female mice and for males 1.62-fold (a group with decreased body
weight) and 1.58-fold (a group with decreased body weight) of control values reported for
groups 1 and 2, respectively.
The consistency between groups of female mice of the same strain for TCE-induced liver
weight gain, regardless of strain examined, is striking. The largest difference within female
strain groups occurred in the only strain in which there was a decrease in TCE-induced body
weight. For males, even in strains that did not show TCE-related changes in body weight, there
was greater variation between groups than in females. For strains in which one group had TCE-
related changes in body weight and another did not, the group with the body weight decrease
always had a lower liver weight as well. Groups that had increased body weight after TCE
exposure also had an increased liver weight in comparison to the groups without a body weight
change. These results demonstrate the importance of carefully matching control animals to
treated animals and the importance of the effect of systemic toxicity, as measured by body
weight decreases, on the determination of the magnitude of liver weight gain induced by TCE
exposure. These results also show the increased variation in TCE-induced liver weight gain
between groups of male mice and an increase incidence of body weight changes due to TCE
exposure in comparison to females, regardless of strain.
In terms of strain sensitivity, it is important not only to take into account differing effects
on body weight changes due to TCE exposure but also to compare animals of the same age or
beginning weight as these parameters may also affect liver weight gain or toxicity induced by
TCE exposure. The authors do not state the age of the animals at the beginning of exposure and
report, as stated above, a range of initial body weights between the groups as much as 44 % for
males and 38% for females. These differences can be due to strain and age. The differences in
final body weight between the groups of controls, when all animals would have been 30 days
older and more mature, was still as much as 48% for males and 44% for females. The data for
female mice, in which body weight was decreased by TCE exposure only in on group in one
strain, suggest that the magnitude of TCE-induced liver weight increase was correlated with
body weight of the animals at the beginning of the experiment. For the C57BL and As/n strains,
02/12/09
E-59
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
female mice starting weights were averaged 17.5 g and 15.5 g, respectively, while the average
liver weights were 1.63-fold and 1.64-fold of control after TCE exposure, respectively. For the
B6CBA, "wild" type, DBA, and NZB female groups the starting body weights averaged 22.5 g,
21.0 g, 23.0 g, and 21.0 g, respectively, while the average liver weight increases were 1.70-fold,
1.78-fold, 1.88-fold, and 2.09-fold of control after TCE exposure. Thus, groups of female mice
with higher body weights, regardless of strain, generally had higher increases in TCE-induced
liver weight increases. The NMRI group of female mice, did not follow this general pattern and
had the highest initial body weight for the single group of 10 mice reported (i.e., 27 g) associated
with a 1.66-fold of control value for liver weight. It is probable that the data for these mice had
been collected from another study. In fact, the starting weights reported for these groups of 10
mice are identical to the starting weights reported for 26 mice examined in Kjellstrand et al
(1981). However, while this study reports a 1.66-fold of control value for liver weight after 30
days of TCE exposure, the extrapolated % liver/body weight given in the 1981 study for 30 days
of TCE exposure was 1.74-fold of control in female NMRI mice. In the Kjellstrand et al (1983b)
study, discussed below, 10 female mice were reported to have a 1.66-fold of control value for
liver weight after 30 days exposure to 150 ppm TCE with an initial starting weight of 26.7 g.
Thus, these data appear to be from that study. Thus, differences in study design, variation
between experiments, and strain differences may account for the differences results reported in
Kjellstrand et al (1983a) for NMRI mice and the other strains in regard to the relationship to
initial body weight and TCE response of liver weight gain.
These data suggest that initial body weight is a factor in the magnitude of TCE-induced
liver weight induction rather than just strain. For male mice, there appeared to be a difference
between strains in TCE-induced body weight reduction, which in turn affects liver weight. The
DBA and "wild" type mice appeared to be the most resistant to this effect (with no groups
affected), while the C57BL, B6CBA, and NZB strains appearing to have at least one group
affected, and the A/sn strain having both groups of males affected. Only one group of NMRI
mice were reported in this study and that group had TCE-induced decreases in body weight. As
stated above there appeared to be much greater differences between groups of males within the
same strain in regard to liver weight increases than for females and that the increases appeared to
be affected by concurrent body weight changes. In general the strains and groups within strain,
that had TCE-induced body weight decreases, had the smallest increases in liver weight, while
those with no TCE-induced changes in body weight in comparison to untreated animals (i.e.,
"wild type" and DBA) or had an actual increase in body weight (one group of NZB mice) had
the greatest TCE-induced increase in liver weight. Therefore, only examining liver weight in
males rather than % liver/ body weight ratios would not be an accurate predictor of strain
sensitivity at this dose due to differences in initial body weight and TCE-induced body weight
changes.
2.2.6. Kjellstrand et al. 1983(b)
This study was conducted in male and female NMRI mice with a similar design as
Kjellstrand et al. (1983a). The ages of the mice were not given by the authors. Animals were
02/12/09
E-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
housed 10 animals per cage and exposed from 30 to 120 days at concentrations ranging from 37
to 3600 ppm TCE. TCE was stabilized with 0.01% thymol and 0.03% diisopropylene. Animals
were exposed continuously with exposure chambers being opened twice a week for change of
bedding food and water resulting in a drop in TCE concentration of ~ 1 hour. A group of mice
was exposed intermittently with TCE at night for 16 hours. This paradigm results not only in
inhalation exposure but also oral exposure from TCE adsorption to food and grooming
behaviour. The authors state that "the different methodological aspects linked to statistical
treatment of body and organ weights have been discussed earlier (Kjellstrand et al., 1981). The
same air-exposed control was used in three cases." The design of the experiment, in terms of
measurement of individual organ and body weights and the inability to assign a % liver/body
weight for each animal, and limitations are similar to that of Kjellstrand et al. (1983b). The
exposure design was for groups of male and female mice to be exposed to 37 ppm, 75 ppm, 150
ppm, and 300 ppm TCE continuously for 30 days (n = 10 per gender and group except for the 37
ppm exposure groups) and then for liver weight and body weight to be determined. Additional
groups of animals were exposed for 150 ppm continuously for 120 days (n = 10). Intermittent
exposure of 4 hours/day for 7 days a week were conducted for 120 days at 900 ppm and
examined immediately or 30 days after cessation of exposure (n = 10). Intermittent exposures of
16 hours/day at 255 ppm group (n = 10), 8 hours/day at 450 ppm, 4 hours/day at 900 ppm, 2
hours/day at 1800 ppm, and 1 hour/day at 3600 ppm 7 days/week for 30 days were also
conducted (n = 10 per group).
As in Kjellstrand et al. (1983a), body weights for individual animals were not recorded in
a way that the initial and final body weights could be compared. The approach taken by the
authors was to match the control group at the initiation of exposure and compare control and
treated average values. At the beginning of the experiment only one group began the experiment
with a statistically significant change in body weight between treated and control animals
(female mice exposed 16 hours a day for 30 days). In regard to final body weight, which would
indicate systemic TCE toxicity, 5 groups had significantly decreased body weight (i.e., males
exposed to 150 ppm continuously for 30 or 120 days, males and females exposed continuously to
300 ppm for 30 days) and 2 groups significantly increased body weight (i.e., males exposed to
1800 ppm for 2 hours/day and 3600 ppm for 1 hour/day for 30 days) after TCE exposure. Thus,
the accuracy of determining the effect of TCE on liver weight changes, reported by the authors in
this study for groups in which body weight were also affected by TCE exposure, would be
affected by similar issues as for data presented by Kjellstand et al. (1983a). In addition,
comparison in results between the 37 ppm exposure groups and those of the other groups would
be affected by difference in number of animals examined (10 vs. 20). As with Kjellstrand et al
(1983 a), the ages of the animals in this study are not given by the author. Difference in initial
body weight (which can be affected by age and strain) reported by Kjellstrand et al (1983a)
appeared to be correlated with the degree of TCE-induced change in liver weight. Although each
exposed group was matched to a control group with a similar average weight, the average initial
body weights in this study varied between groups (i.e., as much as 14% in female control, 16%
in TCE-exposed female mice, 12% in male control, and 16% in male exposed mice).
02/12/09
E-61
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For female mice exposed from 37 ppm to 300 ppm TCE continuously for 30 days, only
the 300 pm group experienced a 16% decrease in body weight between control and exposed
animals. Thus, liver weight increased reported by this study after TCE exposure were not
affected by changes in body weight for exposures below 300 ppm in female mice. Initial body
weights in the TCE-exposed female mice were similar in each of these groups (i.e., range of 29.2
- 31.6 g, or 8%), with the exception of the females exposed to 150 ppm TCE for 30 days (i.e.,
initial body weight of 27.3 g), reducing the effects of differences in initial body weight on TCE-
induced liver weight induction. Exposure to TCE continuously for 30 days resulted in a dose-
dependent change in liver weight in female mice with 1.06-fold, 1.27-fold, 1.66-fold, and 2.14-
fold of control values reported for liver weight at 37 ppm, 75 ppm, 150 ppm, and 300 ppm TCE,
respectively. The increase at 300 ppm in females was accompanied by statistically significant
decreased body weight in the TCE exposed groups compared to control (~ 16%). Thus, the
response in liver weight gain at that exposure is in the presence of toxicity. However, the TCE-
induced increases in liver weight consistently increased with dose of TCE in a linear fashion.
For male mice exposed to 37 to 300 ppm TCE continuously for 30 days, both the 150
ppm and 300 ppm exposed groups experienced a 10% and 18% decrease in body weight after
TCE exposure, respectively. The 37 ppm and 75 ppm groups did not have decreased body
weight due to TCE exposure but varied by 12% in initial body weight. Thus, there are more
factors affecting reported liver weight increases from TCE exposure in the male than female
mice, most importantly toxicity. Exposure to TCE continuously for 30 days resulted in liver
weights of 1.15-fold, 1.50-fold, 1.69-fold, and 1.90-fold of control for 37 ppm, 75 ppm, 150
ppm, and 300 ppm, respectively. The flattening of the dose-response curve for liver weight in
the male mice is consistent with the effects of toxicity at the two highest doses and thus the
magnitude of response at these doses should be viewed with caution. Consistent with Kjellstrand
et al. (1983a) results, male mice in this study appeared to have a higher incidence of TCE-
induced body weight changes than female mice.
The effects of extended exposure, lower durations of exposure but at higher
concentrations, and of cessation of exposure were examined for 150 ppm and higher doses of
TCE. Mice exposed to TCE at 150 ppm continuously for 120 days were reported to have
increased liver weight (i.e., 1.57-fold of control for females and 1.49-fold of control for males),
but in the case of male mice, also to have a significant decrease in body weight of 17% in
comparison to control groups. Increasing the exposure concentration to 900 ppm TCE and
reducing exposure time to 4 hours/day for 120 days also resulted in increased liver weight (i.e.,
1.35-fold of control for females and 1.49-fold of controls for males) but with a significant
decrease in body weight in females of 7% in comparison to control groups. For mice that were
exposed to 150 ppm TCE for 30 days and then examined 120 days after the cessation of
exposure, liver weights were 1.09-fold of control for female mice and the same as controls for
male mice. With the exception of 1800 ppm and 3600 ppm TCE groups exposed at 2 and 1 hour,
respectively, exposure from 225 ppm, 450 and 900 ppm at 16 hours, 8 hours, and 4 hours,
respectively for 30 days did not result in decreased body weight in males or female mice. These
exposures did result in increased liver weights in relation to control groups and for female mice
02/12/09
E-62
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the magnitude of increase was similar (i.e., 1.50-fold, 1.54-fold, and 1.51-fold of control for liver
weight after exposure to 225 ppm TCE 16 hours/day, 450 ppm TCE 8 hours/day, and 900 ppm
TCE 4 hours/day, respectively). For these groups, initial body weights varied by 13% in females
and 14% in males. Thus, under circumstances without body weight changes due to TCE toxicity,
liver weight appeared to have a consistent relationship with the product of duration and
concentration of exposure in female mice. For male mice, the increases in TCE-induced liver
weight were more variable (i.e., 1.94-fold, 1.74-fold, and 1.61-fold of control for liver weight
after exposure to 225 ppm TCE 16 hours/day, 450 ppm TCE 8 hours/day, and 900 ppm TCE 4
hours/day, respectively) with the product of exposure duration and concentration did not result in
a consistent response in males (e.g., a lower dose for a longer duration of exposure resulted in a
greater response than a larger dose at a shorter duration of exposure).
Kjellstrand et al. (1983b) reported light microscopic findings from this study and report
that "after 150 ppm exposure for 30 days, the normal trabecular arrangement of the liver cells
remained. However, the liver cells were generally larger and often displayed a fine vacuolization
of the cytoplasm. The nucleoli varied slightly to moderately in size and shape and had a finer,
granular chromatin with a varying basophilic staining intensity. The Kupffer cells of the
sinusoid were increased in cellular and nuclear size. The intralobular connective tissue was
infiltrated by inflammatory cells. There was not sign of bile stasis. Exposure to TCE in higher
or lower concentrations during the 30 days produced a similar morphologic picture. After
intermittent exposure for 30 days to a time weighted average concentration of 150 ppm or
continuous exposure for 120 days, the trabecular cellular arrangement was less well preserved.
The cells had increased in size and the variations in size and shape of the cells were much
greater. The nuclei also displayed a greater variation in basophilic staining intensity, and often
had one or two enlarged nucleoli. Mitosis were also more frequent in the groups exposed for
longer intervals. The vacuolization of the cytoplasm was also much more pronounced.
Inflammatory cell infiltration in the interlobular connective tissue was more prominent. After
exposure to 150 ppm for 30 days, followed by 120 days of rehabilitation, the morphological
picture was similar to that of the air-exposure controls except for changes in cellular and nuclear
sizes." Although not reporting comparisons between changes in male and female mice in the
results section of the paper, the authors state in the discussion section that "However, liver mass
increase and the changes in liver cell morphology were similar in TCE-exposed male and female
mice."
The authors do not present any quantitative data on the lesions they describe, especially
in terms of dose-response. Most of the qualitative description is for the 150 ppm exposure level,
in which there are consistent reports of TCE induced body weight decreases in male mice. The
authors suggest that lower concentrations of TCE give a similar pathology as those at the 150
ppm level, but do not present data to support that conclusion. Although stating that Kupffer cells
were increased in cellular and nuclear size, no differential staining was applied light microscopy
sections distinguish Kupffer from endothelial cells lining the hepatic sinusoid in this study.
Without differential staining such a determination is difficult at the light microscopic level.
Indeed, Goel et al. (1992) describe proliferation of sinusoidal endothelial cells after 1000
02/12/09
E-63
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mg/kg/day and 2000 mg/kg/day TCE exposure for 28 days in male Swiss mice. However, the
described inflammatory cell infiltrates in the Kjellstrand et al. (1983b) study are consistent with
invasion of macrophages and well as polymorphonuclear cells into the liver, which could
activate resident Kupffer cells. Although not specifically describing the changes as consistent
with increased polyploidization of hepatocytes, the changes in cell size and especially the
continued change in cell size and nuclear staining characteristics after 120 days of cessation of
exposure are consistent with changes in polyploidization induced by TCE. Of note is that in the
histological description provided by the authors, although vacuolization is reported and
consistent with hepatotoxicity or lipid accumulation which is lost during routine histological
slide preparation, there is no mention of focal necrosis or apoptosis resulting from these
exposures to TCE.
2.2.7. Buben and O'Flaherty 1985
This study was conducted with older mice than those generally used in chronic exposure
assays (Male Swiss-Cox outbred mice between 3 and 5 months of age) with a weight range
reported between 34 to 45 g. The mice were administered distilled TCE in corn oil by gavage 5
times a week for 6 weeks at exposure concentrations of either 0, 100, 200, 400, 800, 1600, 2400,
or 3200 mg TCE /kg/day. While 12-15 mice were used in most exposure groups, the 100- and
3200- mg/kg groups contained 4-6 mice and the two control groups consisted of 24 and 26 mice.
Liver toxicity was determined by "liver weight increases, decreases in liver glucose-6-phosphate
(G6P) activity, increases in liver triglycerides, and increases in serum glutamate-pyruvate
transaminase (SGPT) activity". Livers were perfused with cold saline prior to testing for weight
and enzyme activity and hepatic DNA was measured.
The authors reported the mice to tolerate the 6-week exposed with TCE with few deaths
occurring except at the highest dose and that such deaths were related to central nervous system
depression. Mice in all dose groups were reported to continue to gain weight throughout the 6-
week dosing period. However, TCE exposure caused "dose-related increases in liver weight to
body weight ratio and since body weight of mice were generally unaffected by treatment, the
increases represent true liver weight increases." Exposure concentrations, as low as 100
mg/kg/day, were reported to be "sufficient to cause statistically significant increase in the liver
weight/body weight ratio," and the increases in liver size to be "attributable to hypertrophy of the
liver cells, as revealed by histological examination and by a decrease in the DNA concentration
in the livers." Mice in the highest dose group were reported to display liver weight/body weight
ratios that were about -75% greater than those of controls and even at the lowest dose there was
a statistically significant increase (i.e., control liver/ body weight % was reported to be 5.22 ±
0.09 % vs. 5.85 ± 0.20 % in 100 mg/kg/day exposed mice). The % liver/body ratios were 5.22 ±
0.09, 5.84 ±0.20, 5.99 ±0.13, 6.51 ±0.12, 7.12 ±0.12, 8.51 ±0.20, 8.82 ±0.15 and 9.12 ±0.15
% for control (n= 24), 100 (n=5), 200 (n=12), 400 (n=12), 800 (n= 12), 1600 (n=12), 2400 (n=
12) and 3200 (n= 4) mg/kg/day TCE. This represents 1.12-fold, 1.15-fold, 1.25-fold, 1.36-fold,
1.63-fold, 1.69-fold, and 1.75-fold of control for these doses. All dose groups of TCE induced a
statistically significant increase in liver/body weight ratios. For the 200 through 1600 mg/kg
02/12/09
E-64
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
exposure levels, the magnitudes of the increases in TCE exposure concentrations were similar to
the magnitudes of TCE-induced increases in % liver/body weight ratios (i.e., a ~ 2-fold increase
in TCE dose resulted in ~ 1.7 fold increase change in % liver/body weight).
TCE exposure was reported to induce a dose-related trend towards increased
triglycerides (i.e., control values of 3.08 ± 0.29 vs. 6.89 ± 1.40 at 2400 mg/kg TCE) with
variation of response increased with TCE exposure. For liver triglycerides the reported values in
mg/g liver were 3.08 ± 0.29 (n= 24), 3.12 ± 0.49 (n= 5), 4.41 ± 0.76 (n= 12), 4.53 ± 1.05 (n=12),
5.76 ± 0.85 (n=12), 5.82 ± 0.93 (n=12), 6.89 ± 1.40 (n= 12), and 7.02 ± 0.69 (n= 4) for control,
100, 200, 400, 800, 1600, 2400, and 3200 mg/kg/day dose groups, respectively.
For G6P the values in jag phosphate/mg protein/20 minutes were 125.5 ±3.2 (n=12),
117.8 ± 6.0 (n=5), 116.4 ± 2.8 (n=9), 117.3 ± 4.6 (n=9), 111.7 ±3.3. (n=9), 89.9 ± 1.7 (n= 9),
83.8 ± 2.1 (n=8), and 83.0 ± 7.0 (n=3) for the same dose groups. Only the 2400 mg/kg/day
dosing group was reported to be statistically significantly increased for triglycerides after TCE
exposure although there appeared to be a dose-response. For decreases in G6P the 800
mg/kg/day and above doses were statistically significant. The numbers of animals varied
between groups in this study but in particular only a subset of the animals were tested for G6P
with the authors providing no rationale for the selection of animals for this assay. The
differences in the number of animals per group and small number of animals per group affected
the ability to determine a statistically significant change in these parameters but the changes in
liver weights were robust enough and variation small enough between groups that all TCE-
induced changes were described as statistically significant. The livers of TCE treated mice,
although enlarged, were reported to appear normal. A dose-related decrease in glucose-6-
phophatase activity was reported with similar small decreases (~ 10%) observed in the TCE
exposed groups that did not reach statistical significance until the dose reached 800 mg/kg TCE
exposure. SGPT activity was not observed to be increased in TCE-treated mice except at the two
highest doses and even at the 2400 mg/kg dose half of the mice had normal values. The large
variability in SGPT activity was indicative of heterogeneity of this response between mice at the
higher exposure levels for this indicator of liver toxicity. However, the results of this study also
demonstrate that hepatomegaly was a robust response that was observed at the lowest dose
tested, was dose-related, and was not accompanied by toxicity.
Liver histopathology and DNA content were determined only in control, 400, and 1600
mg/kg TCE exposure groups. DNA content was reported to be significantly decreased from 2.83
± 0.17 mg/g liver in controls to 2.57 ± 0.14 in 400 mg/kg TCE treated group, and to 2.15 ± 0.08
mg/kg liver in the 1600 mg/kg exposed group. This result was consistent with a decreased
number of nuclei per gram of liver and hepatocellular hypertrophy. Liver degeneration was
reported as swollen hepatocytes and to be common with treatment. "Cells had indistinct borders;
their cytoplasm was clumped and a vesicular pattern was apparent. The swelling was not simply
due to edema, as wet weight/dry weight ratios did not increase." Karyorhexis (the disintegration
of the nucleus) was reported to be present in nearly all specimens and suggestive of impending
cell death. A qualitative scale of negative, 1, 2, 3, or 4 was given by the authors to rate their
02/12/09 E-65
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
findings without further definition or criterion given for the ratings. "No Karyorhexis, necrosis,
or polyploidy was reported in controls, but a score of 1 for Karyorhexis was given for 400 mg/kg
TCE and 2 for 1600 mg/kg TCE." Central lobular necrosis reported to be present only at the
1600 mg/kg TCE exposure level and as a score of 1. "Polyploidy was also characteristic in the
central lobular region" with a score of 1 for both 400 mg/kg and 1600 mg/kg TCE. The authors
reported that "hepatic cells had two or more nuclei or had enlarged nuclei containing increased
amounts of chromatin, suggesting that a regenerative process was ongoing" and that there were
no fine lipid droplets in TCE exposed animals. The finding of "no polyploidy" in control mouse
liver is unexpected given that binucleate and polyploid hepatocytes are a common finding in the
mature mouse liver. It is possible that the authors were referring to unusually high instances of
"polyploidy" in comparison to what would be expected for the mature mouse. The score given
by the authors for polyploidy did not indicate a difference between the two TCE exposure
treatments and that it was of the lowest level of severity or occurrence. No score was given for
centrolobular hypertrophy although the DNA content and liver weight changes suggested a dose
response. The "Karyorhexis" described in this study could have been a sign of cell death
associated with increased liver cell number or dying of maturing hepatocytes associated with the
increased ploidy, and suggests that TCE treatment was inducing polyploidization. Consistent
with enzyme analyses, centrilobular necrosis was only seen at the highest dose and with the
lowest qualitative score, indicating that even at the highest dose there was little toxicity.
Thus, the results of this study of TCE exposure for 6 weeks, is consistent with acute
studies and show that the region of the liver affected by TCE is the centralobular region, that
hepatocellular hypertrophy is observed in that region, and that increased liver weight is induced
at the lowest exposure level tested and much lower than those inducing overt toxicity. These
authors suggest polyploidization is occurring as a result of TCE exposure although a quantitative
dose-response cannot be determined from these data.
2.2.8. Channel et al. 1998
This study was performed in male hybrid B6C3Fl/CrlBR mice (13 weeks-old, 25-30
grams) and focused on indicators of oxidative stress. TCE was administered by oral gavage 5
days a week in corn oil for up to 55 days for some groups. Although the study design indicated
that water controls, corn oil controls, and exposure levels of 400, 800, and 1200 mg/kg day TCE
in corn oil, results were not presented for water controls for some parameters measured. Initial
body weights and those recorded during the course of the study were not reported for individual
treatment groups. Liver samples were collected on study days 2, 3, 6, 10, 14, 21, 28, 35, 42, 49,
and 56. Histopathology was studied from a single section taken from the median lobe.
Thiorbarbiturate acid-reactive substances (TBARS) were determined from whole liver
homogenates. Nuclei were isolated from whole liver homogenates and DNA assayed for 8-
hydroxydeoxyguanosine (8-OHdG). There was no indication that parenchymal cell and
nonparenchymal cells were distinguished in the assay. Free radical electron paramagnetic
resonance (EPR) for total radicals was analyzed in whole liver homogenates. For peroxisome
detection and analysis, livers from 3 mice from the 1200 mg/kg TCE and control (oil and water)
02/12/09
E-66
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
groups were analyzed via electron microscopy. Only centrilobular regions, the area stated by the
authors to be the primary site of peroxisome proliferation, were examined. For each animal, 7
micrographs of randomly chosen hepatocytes immediately adjacent to the central vein were
examined with peroxisomal area to cytoplasmic area, the number of peroxisomes per unit area of
cytoplasm, and average peroxisomal size quantified. Proliferation cell nuclear antigen (PCNA),
described as a marker of cell cycle except GO, was examined in histological sections for a
minimum of 18 fields per liver section. The authors did not indicate what areas of the liver
lobule were examined for PCNA. Apoptosis was detected on liver sections using a apoptosis kit
using a single liver section from the median lobe and based on the number of positively labeled
cells per 10 mm2 in combination with the morphological criteria for apoptosis of Columbano et
al. (1985). However, the authors did not indicate what areas of the liver lobule were specifically
examined.
The authors reported that body weight gain was not adversely affected by TCE dosing of
the time course of the study but did not show the data. No gross lesions were reported to be
observed in any group. For TBARS no water control data was reported by the authors. Data was
presented for 6 animals per group for the corn oil control group and the 1200 mg/kg group (error
bars representing the SE). No data was presented without corn oil so that the effects of corn oil
on the first day of the study (day 2 of dosing) could not be determined. After 2 and 3 days of
dosing the corn oil and 1200 mg/kg TCE groups appeared to have similar levels of TBAR
detected in whole liver as nmol TBARS/mg protein. However, by day 6 the corn oil treated
control had a decrease in TBAR that continued until day 15 where the level was ~ 50% of that
reported on days 2 and 3. The variation between animals as measured by SE was reported to be
large on day 10. By day 20 there was a slight increase in variation that declined by day 35 and
stayed the same through day 55. For the TCE exposed group the TBARs remained relatively
consistent and began to decline by about day 20 to a level that similar to the corn oil declines by
day 35. Therefore, corn oil alone had a significant effect on TBAR detection inducing a decline
by 6 days of administration that persisted thought 55 days. TCE administration at the 1200
mg/kg dose in corn oil appeared to have a delayed decline in TBARS. The authors interpreted
this pattern to show that lipid peroxidation was elevated in the 1200 mg/kg TCE group at day 6
over corn oil. However, corn oil alone induced a decrease in TBARs. At no time was TBARS in
TCE treatment groups reported to be greater than the initial levels at days 2 and 3, a time in
which TCE and corn oil treatment groups had similar levels. Rather than inducing increasing
TBARS over the time course of the study TCE, at the 1200 mg/kg dose, appeared to delay the
corn oil induced suppression of TBARs detection. Because the authors did not present data for
aqueous control animals, the time course of TBARS detection in the absence of corn oil, cannot
be established.
For the 800 and 400 mg/kg TCE data the authors presented a figure, without standard
error information, for up to 35 days that shows little difference between 400 mg/kg TCE
treatment and corn oil suppression of TBAR induction. There was little difference between the
patterns of TBAR detection for 800 and 400 mg/kg TCE, indicating that both delayed TBAR
suppression by corn oil to a similar extent and did not induce greater TBAR than corn oil alone.
02/12/09
E-67
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For 8-OHdG levels, the authors report that elevations were modest with the greatest
increase noted in the 1200 mg/kg day TCE treatment group of 196% of oil controls on day 56.
Levels fluctuated throughout the study with most of the time points that were elevated showing
129% of control for the 1200 mg/kg/day group. Statistically significant elevations were noted on
days 2, 10, 28, 49 and 56 with depression on day 3. On all other days (i.e., days 6, 14, 21, 35 and
42) the 8-OHdG values were similar to those of corn oil controls. No statistically significant
effects were reported to be observed at lower doses. The figure presented by the authors shows
the % of controls by TCE treatment at 1200 mg/kg/day but not the control values themselves.
The pattern by corn oil is not shown and neither is the standard error of the data. As a percent of
control values the variations were very large for many of the data points and largest for the data
given at day 55 in which the authors report the largest difference between control and TCE
treatment. There was no apparent pattern of elevation in 8-OHdG when the data was presented
in this manner. Because the data for the corn oil control was not given, as well as no data given
for aqueous controls, the effects of corn oil alone cannot be discerned.
Given that for TBARS corn oil had a significant effect and showed a pattern of decline
after 6 days, with TCE showing a delayed decline, it is especially important to discern the effects
of corn oil and to see the pattern of the data. At time points when TBARS levels were reported
to be the same between corn oil and TCE (days 42, 49 and 56) the pattern of 8-OHdG was quite
different with a lower level at day 42 a slightly increased level at day 49 and the highest
difference reported at day 56 between corn oil control and TCE treated animals. The authors
report that the pattern of "lipid peroxidation" to be similar between the 1200 and 800 mg/kg
doses of TCE but for there to be no significant difference between 800 mg/kg TCE and corn oil
controls. Thus the pattern of TBARS as a measure of lipid peroxidation and 8-OHdG level in
nuclear DNA did not match.
In regard to total free radical levels as measured by EPR, results were reported for the
1200 mg/kg TCE as a signal that was subtracted from control values with the authors stating that
only this dose level induced an elevation significantly different from controls. Again, aqueous
control values were not presented to discern the effects of corn oil or the pattern that may have
arisen with time of corn oil administration. The pattern of total free radical level appeared to
differ from that of lipid peroxidation and for that of 8-OHdG DNA levels with no changes at
days 2, 3, a peak level at day 6, a rapid drop at day 10, mild elevation at day 20, and significant
decrease at day 49. The percentage differences between control and treated values reported at
day 6 and 20 by the authors was not proportional to the fold-difference in signal indicating that
there was not a consistent level for control values over the time course of the experiment. While
differences in lipid peroxidation detection between 1200 mg/kg TCE and corn oil control were
greatest at day 14, total free radicals showed their biggest change between corn oil controls and
TCE exposure on day 6, time points in which 8-OHdG levels were similar between TCE
treatment and corn oil controls. Again, there was no reported difference between corn oil control
and the 800 mg/kg TCE exposed group in total free radical formation but for lipid peroxidation
the 800 mg/kg TCE exposed group had a similar pattern as that of 1200 mg/kg TCE.
02/12/09
E-68
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Only the 1200 mg/kg group was evaluated for peroxisomal proliferation at days 6, 10,
and 14. Thus, correlations with peroxisome proliferation and other parameters in the report at
differing times and TCE exposure concentrations could not be made. The authors report that
there was a treatment and time effect for percent peroxisomal area, a "treatment only "effect for
number of peroxisome and no effect for peroxisomal size. They also report that hepatocytes
examined from corn oil control rats were no different that those from water control rats for all
peroxisomal parameter, thus discounting a vehicle effect. However, there was an effect on
peroxisomal size between corn oil control and water with corn oil decreasing the peroxisomal
size in comparison to water on all days tested. The highest TCE-induced percent peroxisomal
area and number occurred on day 10 of the 3 time points measured for this dose and the fold
increase was -4.5 fold and -3.1 fold increase, respectively. The day-10 peak in peroxisomal
area and number does not correlate with the reported pattern of free radical or 8-OHdG
generation.
For cell proliferation and apoptosis, data were given for days 2, 6, 10, 14, and 21 in a
figure. PCNA cells, a measure of cells that have undergone DNA synthesis, was elevated only
on day 10 and only in the 1200 mg/kg/day TCE exposed group with a mean of - 60 positive
nuclei per 1000 nuclei for 6 mice (~ 6%). Given that there was little difference in PCNA
positive cells at the other TCE doses or time points studied, the small number of affected cells in
the liver could not account for the increase in liver size reported in other experimental paradigms
at these doses. The PCNA positive cells as well as "mitotic figures" were reported to be present
in centrilobular, midzonal, and periportal regions with no observed predilection for a particular
lobular distribution. No data was shown regarding any quantitative estimates of mitotic figures
and whether they correlated with PCNA results. Thus, whether the DNA synthesis phases of the
cell cycle indicated by PCNA staining were indentifying polyploidization or increased cell
number cannot be determined. The authors reported that there was no cytotoxicity manifested as
hepatocellular necrosis in any dose group and that there was no significant difference in
apoptosis between treatment and control groups with data not shown. The extent of apoptosis in
any of the treatment groups, or which groups and timepoints were studied for this effect cannot
be determined. No liver weight or body weight data was provided in this study.
These results confirm that as a vehicle corn oil is not neutral in its affects in the liver. The
TBARS results indicate a reduction in detection of TBARS in the liver with increasing time of
exposure to corn oil alone. Although control animals "treated with water" gavage were studied,
only the results for peroxisome proliferation were presented by the study so that the effects of
corn oil gavage were not easy to discern. In addition, the data was presented in such a way for 8-
OHdG and total free radical changes that the pattern of corn oil administration was obscured. It
is not apparent from this study that TCE exposure induces oxidative damage.
2.2.9. Dorfmueller et al. 1979
The focus of this study was the evaluation of "teratogenicity and behavioral toxicity with
inhalation exposure of maternal rats" to TCE. Female Long-Evans hooded rats (n=12) of - 210
02/12/09
E-69
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
g weight were treated with 1800 ± 200 ppm TCE for 6 hours/day 5 days/week for 22 ± 6 days
(until pregnancy confirmation) continuing through day 20 of gestation. Control animals were
exposed 22 ± 3 days before pregnancy confirmation. The TCE used in this study contained 0.2%
epichlorhydrin. Body weights were monitored as well as maternal liver weight at the end of
exposure. Other than organ weight, no other observations regarding the liver were reported in
this study. The initial weights of the dams were 212 ± 39 g (mean ± SD) and 204 ± 35 g for
treated and control groups, respectively. The final weights were 362 ± 32 g and 337 ± 48 g for
treated and control groups, respectively. There was no indication of maternal toxicity by body
weight determinations as a result of TCE exposure in this experiment and there was also no
significant difference in absolute or relative % liver/body weight between control and treated
female rats in this study.
2.2.10. Kumar et al. 2001
In this study, adult male Wistar rats (130 ± 10 g body weight) were exposed to 376 ±
1.76 ppm TCE ("AnalaR grade") for 8 , 12, and 24 weeks for 4 hours/day 5 days/week. The
ages of the rats were not given by the authors. Each group contained 6 rats. The animals were
exposed in whole body chambers and thus additional oral exposure was probable. Along with
histopathology of light microscopic sections, enzymatic activities of alkaline phosphatase and
acid phosphatase, glutamic oxoacetate transaminase, glutamic pyruvate transaminase, reduced
glutathione and "total sulphydryl" were assayed in whole liver homogenates as well as total
protein. The authors state that "the size and weight of the liver were significantly increased after
8, 12, and 24 weeks of TCE exposure." However, the authors do not report the final body weight
of the rats after treatment nor do they give quantitative data of liver weight changes. In regard to
histopathology, the authors state "After 8 weeks of exposure enlarged hepatocytes, with uniform
presence of fat vacuoles were found in all of the hepatocytes affecting the periportal, midzonal,
and centrilobular areas, and fat vacuoles pushing the pyknosed nuclei to one side of hepatocytes.
Moreover congestion was not significant. After exposure of 12 and 24 weeks, the fatty changes
became more progressive with marked necrosis, uniformly distributed in the entire organ." No
other description of pathology was provided in this report. In regard to the description of fatty
change, the authors only do conventional H&E staining of sections with no precautions to
preserve or stain lipids in their sections. The authors provide a table with histological scoring of
simply + or - for minimal, mild or moderate effects and do not define the criteria for that
scoring. There is also no quantitative information given as to the extent, nature, or location of
hepatocellular necrosis. The authors report "no change was observed in GOT and GPT levels of
liver in all the three groups. The GSH level was significantly decreased while TSH level was
significantly increased during 8, 12, and 24 weeks of TCE exposure. The acid and alkaline
phosphatases were significantly increased during 8, 12, and 24 weeks of TCE exposure." The
authors present a series of figures that are poor in quality to demonstrate histopathological TCE-
induced changes. No mortality was observed from TCE exposure in any group despite the
presence of liver necrosis.
02/12/09
E-70
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.2.11. Kawamoto et al. 1988
The focus of this study was the long-term effects of TCE treatment on induction of
metabolic enzymes in male adult Wistar rats. The authors reported that 8 rats weighing 200 g
were treated with 2.0 g/kg TCE in olive oil administered subcutaneously twice a week for 15
weeks with 7 rats serving as olive oil controls. In a separate experiment, 5 rats were injected
with 1.0 g/kg TCE in olive oil i.p. once a day for 5 continuous days. For comparative purposes
groups of 5 rats each were administered 3-methylcholanthrene (20 mg/kg in olive oil i.p),
Phenobarbital (80 mg/kg in saline i.p) for 4 days as well as ethanol administered in drinking
water containing 10% ethanol for 14 days. Microsomes were prepared one week after the last
exposure from rats administered TCE for 15 weeks and 24 hours after the last exposure for the
other treatments.
Body weights were reported to be slightly less for the TCE treated group than for controls
with the initial weights, shown in a figure, to be similar for the first weeks of exposure. At 15
weeks there appeared to be ~ 7.5% difference in mean body weights between control and TCE
treated rats which the authors reported to not be significantly different. Organ weights at the
termination of the experiment were reported to only be different for the liver with a 1.21-fold of
control value reported as a percentage of body weight with TCE treatment. The authors report
their increase in liver weights in male rats from subcutaneous exposure to TCE in olive oil (2.0
g/ kg) to be consistent with the range of liver weight gain in rats reported by Kjellstrand et al
(1981) for 150 ppm TCE inhalation exposure (see comments on that study above). The 5-day
i.p. treatment with TCE was also reported to only produce increased liver weight but the data
was not shown and the magnitude of the percentage increase was not given by the authors. No
liver pathology results were studied or reported as well.
Along with an increase in liver weight, 15-week treatment with TCE was reported to
cause a significant increase of microsomal protein/g liver of -20% (10.64 ± 0.88 vs. 12.58 ±
0.71 mg/g liver for olive oil controls and TCE treatment, respectively). Microsomal cytochrome
P450 content was reported to show a mild increase that was not statistically significant of 1.08-
fold (1.342 ± 0.205 vs. 1.456 ±0.159 nmol/mg protein for olive oil controls and TCE treatment,
respectively) of control. However, Cytochrome P-450 content showed 1.28-fold of control value
(14.28 ± 2.41 vs. 18.34 ± 2.31 nmol/g liver for olive oil controls and TCE treatment,
respectively) in terms of g/liver. Chronic treatment of TCE was also reported to cause a
significant increase in cytochrome b-5 level (~1.35-fold of control) and NADPH-cytochrome c
reductase activity (~ 1.50-fold of control) in g/liver.
The 5-day TCE treatment via the i.p. route of administration was reported to cause a
significant increase in microsomal protein (~ 20%), induce cytochrome P450 (~ 50% increase
g/liver and 22% increase in microsomal protein), but to also increase cytochrome b-5 and
NADPH-cytochrome c reductase activity by 50% and 70% in g/liver, respectively. Although
weaker, 5-day i.p treatment with TCE induced an enzyme pattern more similar to that of
Phenobarbital and ethanol rather methylcholanthrene (i.e., increased cytochrome P-450 but not
02/12/09
E-71
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
microsomal protein and NADPH-cytochrome c reductase). Direct quantitative comparisons of
vehicle effects and potential impact on response to TCE treatments for 15 weeks subcutaneous
exposure and 5-day i.p. exposure could not be made as baseline levels of all enzyme and protein
levels changed as a function of age. Of note is that, in the discussion section of the paper, the
authors disclose that injection of TCE 2.0 or 3.0 g/kg i.p. for 5 days resulted in paralytic ileus
from TCE exposure as unpublished observations. They note that the rationale for injecting TCE
subcutaneously was not only that it did not require an inhalation chamber but also guarded
against peritonitis that sometimes occurs following repeated i.p. injection. In terms of
comparison with inhalation or oral results, the authors note that the subcutaneous treatment
paradigm will result in TCE not immediately being metabolized but retained in the fatty tissue
and that after cessation of exposure TCE metabolites continued to be excreted into the urine for
more than 2 weeks.
2.2.12 NTP 1990
2.2.12.1. 13-week studies
The National Toxicology Program (NTP) conducted a 13 weeks study of 7 week old
F344/N rats (10 rats per group) that received doses of 125 to 2000 mg/kg (males) [0, 125, 250,
500, 1000, or 2000 mg/kg] and 62.5 to 1000 mg/kg (females) [0, 62.5, 125, 250, 500, or 1000
mg/kg] TCE via corn oil gavage 5 days per week. For 7-week old B6C3Flmice (n= 10 per
group), the dose levels were reported to be 375 to 6000 mg/kg TCE [0, 375, 750, 1500, 3000, or
6000 mg/kg]. Animals were exposed via corn oil gavage to TCE that was epichlorhydrin free.
All rats were reported to survive the 13 week study, but males receiving 2000 mg/kg exhibited a
24% difference in final body weight. However, there was great variation in initial weights
between the dose groups with mean initial weights at the beginning of the study reported to 87,
88, 92, 95, 101, and 83 grams for the control, 125, 250, 500, 1000, and 2000 mg/kg dose groups
in male rats, respectively. This represents a 22 % difference between the highest and lowest
initial weights between groups. Thus, changes in final body weight after TCE treatment also
reflect differences in starting weights between the groups which in the case of the 500, and 1000
mg/kg groups would results in an lower than expected change in weight due to TCE exposure.
For female rats the mean initial starting weights were reported to be 81, 72, 74, 75, 73, and 76 g,
respectively for the control, 62.5, 125, 250, 500, and 1000 mg/kg dose groups. This represents a
~ 13% difference between initial weights. In the case of female rats the larger mean initial weight
in the control group would tend to exaggerate the effects of TCE exposure on final body weight.
The authors did not report the variation in initial or final body weights within the dose groups.
At the lowest doses for male and female rats body mean weights were reported to be decreased
by 6 and 7 % in male and female rats, respectively. Organ weight changes were not reported for
rats.
For male mice, mean initial body weights ranged from 19 to 22 g (—16% difference) and
for female mice ranged between 18 and 15 g (20% difference), and thus similar to rats, the final
body weights in the groups dose with TCE reflect not only the effects of the compound but also
02/12/09
E-72
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
differences in initial weights. For male mice the mean final body weights were reported to be 3
to 17% less than controls for the 375 to 3000 mg/kg dose. For female mice the % difference in
final body weight was reported to be the same except for the 6000 mg/kg dose group but this
lack of difference between controls and treated female mice reflected no change in mice that
started at differing weights. Male mice started to exhibit mortality at 1500 mg/kg with 8/10
surviving the 1500 mg/kg dose, 3/10 surviving the 3000 mg/kg dose, and none surviving the
6000 mg/kg dose of TCE till the end of the study. For females 1 animals out of 10 died in the
750, 1500, and 3000 mg/kg dose groups and one surviving the 6000 mg/kg group. In general,
the magnitude of increase in TCE exposure concentration was similar to the magnitude of
increase in % liver/body weight for the 750 and 1500 mg/kg TCE exposure groups in male
B6C3F1 mice and for the 750 to 3000 mg/kg TCE exposure groups in female mice (i.e., a 2-fold
increase in TCE exposure resulted in ~ 2-fold increase in % liver/body weight.
The descriptions of pathology in rats and mice given by this study were not very detailed.
For rats only control and high dose rats were examined histologically. For mice only controls
and the two highest dose groups were examined histologically. Only mean liver weights were
reported with no statistical analyses provided to ascertain quantitative differences between study
groups.
Pathological results were reported to reveal that 6/10 males and 6/10 female rats had
pulmonary vasculitis at the highest concentration of TCE. This change was also reported to have
occurred in 1/10 control male and female rats. Most of those animals were also reported to have
had mild interstitial pneumonitis. The authors report that viral titers were positive during this
study for Sendai virus.
In mice, liver weights (both absolute and as a percent of body weight) were reported to
increase with TCE -exposure level. Liver weights were reported to have increased by more than
10% relative to controls for males receiving 750 mg/kg or more and for females receiving 1,500
mg/kg or more. The most prominent hepatic lesions detected in the mice were reported to be
centrilobular necrosis, observed in 6/10 males and 1/10 females administered 6000 mg/kg.
"Although centrilobular necrosis was not seen in either males or females administered 3000
mg/kg, 2/10 males had multifocal areas of calcifications scattered throughout their livers. These
areas of calcification were considered to be evidence of earlier hepatocellular necrosis.
Multifocal calcification was also seen in the liver of a single female mouse that survived the
6000 mg/kg dosage regime. One female mouse administered 3000 mg/kg also had a
hepatocellular adenoma, an extremely rare lesion in female mice of this age (20 weeks)." There
appeared to be consistent decrease in liver weight at the lowest dose in both female and male
mice after 13 weeks of TCE exposure. Liver weight was increased at exposure concentrations in
which there was not increased mortality due to TCE exposure at 13 weeks of TCE exposure.
Mice data for 13 weeks: mean body and liver weights
Dose survival body weight liver weight % liver Wt/body Wt
02/12/09
E-73
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(mg/kg TCE) (mean in g) (mean final in g) (fold change vs.
control)
Male initial final
0
10/10
21
36
2.10
5.8
375
10/10
20
35
1.74
5.0(0.86)
750
10/10
21
32
2.14
6.8 (1.17)
1,500
8/10
19
29
2.27
7.6(1.31)
3,000
3/10
20
30
2.78
8.5 (1.46)
6,000
0/10
22
-
-
-
Female
0
10/10
18
26
1.40
5.5
375
10/10
17
26
1.31
5.0(0.91)
750
9/10
17
26
1.55
5.8 (1.05)
1,500
9/10
17
26
1.80
6.5 (1.18)
3,000
9/10
15
26
2.06
7.8 (1.42)
6,000
1/10
15
27
2.67
9.5 (1.73)
2.2.12.2. 2-year studies
In the 2-year phase of the NTP study, TCE was administered by corn oil gavage to
groups of 50 male and 50 female F344/N rats, and B6C3F1 mice. Dosage levels were 500 and
1000 mg/kg for rats and 1000 mg/kg for mice. TCE was administered 5 times a week for 103
weeks and surviving animals were killed between weeks 103 and 107. The same number of
animals receiving corn oil gavage served as controls. The animals were 8 weeks old at the
beginning of exposure. The focus of this study was to determine if there was a carcinogenic
response due to TCE exposure so there was little reporting of non-neoplastic pathology or
toxicity. There was no report of liver weight at termination of the study, only body weight.
The authors reported that there was no increase in necrosis in the liver from TCE
exposure in comparison to control mice. In control male mice, the incidence of hepatocellular
carcinoma (tumors with markedly abnormal cytology and architecture) was reported to be 8/48
in controls, and 31/50 in TCE-exposed male mice. For females control mice hepatocellular
carcinomas were reported in 2/48 of controls and 13/49 of TCE-exposed female mice.
Specifically, the authors described liver pathology in mice as follows: "Microscopically the
hepatocellular adenomas were circumscribed areas of distinctive hepatic parenchymal cells with
a perimeter of normal appearing parenchyma in which there were areas that appeared to be
undergoing compression from expansion of the tumor. Mitotic figures were sparse or absent but
the tumors lacked typical lobular organization. The hepatocellular carcinomas had markedly
abnormal cytology and architecture. Abnormalities in cytology included increased cell size,
decreased cell size, cytoplasmic eosinophilia, cytoplasmic basophilia, cytoplasmic vacuolization,
02/12/09
E-74
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
cytoplasmic hyaline bodies and variations in nuclear appearance. In many instance, several or all
of the abnormalities were present in different areas of the tumor. There were also variations in
architecture with some of the hepatocellular carcinomas having areas of trabecular organization.
Mitosis was variable in amount and location." The authors report that the non-neoplastic lesion
in male mice differing from controls was focal necrosis in 4 vs. 1 animal in the dosed group (8%
vs. 2%). There was no fatty metamorphosis in treated male mice vs. 2 animals in control. In
female mice there was focal inflammation in 29% vs. 19% of animals (dosed vs. control) and no
other changes. Therefore the reported pathological results of this study did not show that the
liver was showing signs of toxicity after two years of TCE exposure except for neoplasia.
For hepatocellular adenomas the incidence was reported to be "7/48 control vs. 14/50
dosed in males and 4/48 in control vs. 16/49 dosed female mice." The administration of TCE to
mice was reported to cause increased incidences of hepatocellular carcinomas in males (control,
8/48; dosed 31/50: P, 0.001) and in females (control 2/48; dosed 13/49; P<0.005).
Hepatocellular carcinomas were reported to metastasize to the lungs in five dosed male mice and
one control male mouse, while none were observed in females. The incidences of hepatocellular
adenomas were reported to be increased in male mice (control 7/48; dosed 14/50) and in female
mice (control 4/48; dosed 16/49; P<0.05). The survival of both low and high dose male rats and
dosed male mice was reported to be less than that of vehicle controls with body weight decreases
dose dependent. Female mice body weights were comparable to controls. The authors report
adjusted rates of 20.6% for control vs. 53.1 % for dosed males for adenoma, 22.1 % control, and
92.9 % for carcinoma in males, and liver carcinoma or adenoma adjusted rates of 100%. For
female mice the adjusted rates were reported to be 12.5% adenoma for control vs. 55.6% for
dosed, and 6.2% control carcinoma vs. 43.9% dosed, with liver carcinoma or adenoma adjusted
rates of 18.7% for control vs. 69.1% for dosed. All of the liver results for male and female
mice were reported to be statistically significant. The administration of TCE was reported to
cause earlier expression of tumors as the first animals with carcinomas were 57 weeks for TCE-
exposed animals and 75 weeks for control male mice.
In male rats there was no reported treatment related non-neoplastic liver lesions. In
female rats a decrease in basophilic cytological change was reported to be of note in TCE treated
rats (~ 50% in controls but ~ 5% in TCE treatment groups). However, the authors reported that
"the results in male F344/N rats were considered equivocal for detecting a carcinogenic response
because both groups receiving TCE showed significantly reduced survival compared to vehicle
controls (35/70, 70%; 20/50, 40%; 16/50, 32%) and because 20% of the animals in the high-dose
group were killed accidently by gavage error." Specifically 1 male control, 3 low-dose males, 10
high-dose males, 2 female controls, 5-low dose females and 5 high-dose female rats were killed
by gavage error.
2.2.13. NTP 1988
The studies described in the NTP 1988 TCE report were conducted "to compare the
sensitivities of four strains of rats to diisopropylamine-stabilized TCE." However, the authors
02/12/09
E-75
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
conclude "that because of chemically induced toxicity, reduced survival, and incomplete
documentation of experimental data, the studies are considered inadequate for either comparing
or assessing TCE-induced carcinogenesis in these strains of rats. TCE (more than 99% pure,
stabilized with 8ppm diisopropylamine) was administered via corn oil gavage at exposure
concentrations of 0, 500 or 1000 mg/kg per day, 5 days per week, for 103 weeks to 50 male and
female rats of each strain. The survival of "high-dose male Marshal rats was reduced by a large
number of accidental deaths (25 animals were accidentally killed)." However, the report notes
survival was decreased at both exposure levels of TCE because of mortality that occurred during
the administration of the chemical. The number of animals accidently killed were reported to be:
11 male ACI rats at 500 mg/kg, 18 male ACI rats at 1000 mg/kg, 2 vehicle control female ACI
rats, 14 female ACI rats at 500 mg/kg , 12 male ACI rats at 1000 mg/kg, 6 vehicle control male
August rats, 12 male August rats at 500 mg/kg, 11 male August rats at 1000 mg/kg, 1 vehicle
control female August rats, 6 female August rats at 500 mg/kg, 13 male August rats at 1000
mg/kg, 2 vehicle control male Marshal rats, 12 male Marshal rats at 500 mg/kg, 25 male Marshal
rats at 1000 mg/kg, 3 vehicle control female Marshal rats, 14 female Marshal rats at 500 mg/kg,
18 female Marshal rats at 1000 mg/kg, 1 vehicle control male Osborne-Mendel rat, 6 male
Osborne-Mendel rats at 500 mg/kg, 7 male Osborne-Mendel rats at 1000 mg/kg, 8 vehicle
control female Osborne-Mendel rats, 6 female Osborne-Mendel rats at 500 mg/kg, and 6 female
Osborne-Mendel rats at 1000 mg/kg. The age of the rats "when placed on the study" were
reported to differ and were for ACI rats (6.5 weeks), August rats (8 weeks), Marshal rats (7
weeks), and Osborne-Mendel rats (8 weeks). The ages of sacrifice also varied and were 17-18
weeks for the ACI and August rats, and 110-111 weeks for the Marshal rats.
Results from a 13-week study were briefly mentioned in the report. For the 13-week
duration of exposure, groups of 10 male ACI and August rats were administered 0,125, 250, 500,
1,000 or 2,000 mg/kg TCE in corn oil gavage. Groups of 10 female ACI and August rats were
administered 0, 62.5, 125, 250, 500 or 1,000 mg/kg TCE. Groups of 10 male Marshal rats
received 0, 268, 308, 495, 932, or 1,834 mg/kg and groups of female Marshal rats were given 0,
134, 153, 248, 466, or 918 mg/kg TCE. With the exception of 3 male August rats receiving 2000
mg/kg TCE, all animals survived to the end of the 13-week experimental period. "The
administration of the chemical for 13 weeks was not associated with histopathological changes."
In the two-year study the report noted that there "was no evidence of liver toxicity
described as non-neoplastic changes in male ACI rats due to TCE exposure with 4% or less
incidence of any lesion in control or treated animals. For female ACI rats, the incidence of fatty
metamorphosis was 6% in control vehicle, 9% in low dose TCE, and 13% in high dose TCE
groups. There was also a 2%, 11%, and 8% incidence of clear cell change, respectively. A 6%
incidence of hepatocytomegaly was reported in vehicle control and 15% incidence in the high
dose group." All other desciptors had reported incidences of less than 4%. For August rats there
was also little evidence of liver toxicity. In male August rats there was a reported incidence of
8%>, 4% and 10% focal necrosis in vehicle control, low dose, and high dose, respectively. Fatty
metamorphosis was reported to be 8% in control, and 2% and 4% in low and high dose. All
other descriptors were reported to be less than 4%. In female August rats, all desciptors of
02/12/09
E-76
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
pathology were reported to have a 4% or less incidence except for hepatomegaly which was 10%
for vehicle control, 6% for the low dose and 2% for high dose TCE. For male Marshal rats there
was a reported 63% incidence of Inflammation, NOS in vehicle control, 12% in low dose and
values not recorded at the high dose. There was a reported 6% and 14% incidence of fatty
metamorphosis in control and low dose male rats. Clear cell change was 8% in vehicle with all
other values 4% or less. For female Marshal rats all values were 4% or less except for fatty
metamorphosis in 6% of vehicle controls. For male Osborne-Mendel rats, there was a reported
4%, 10%), and 4% incidence of focal necrosis in vehicle control, low and high dose respectively.
For "cytoplasmic change/ NOS," there were reported incidences of 26%, 32% and 27% in
vehicle, low dose, and high dose animals, respectively. All other descriptors were reported to be
4% or less. In female Osborne-Mendel rats there was a reported incidence of 10% of focal
necrosis at the low dose with all other descriptors reported at 4% or less.
Obviously the negative results in this bioassay are confounded by the killing of a large
portion of the animals accidently by experimental error. Still, these large exposure
concentrations of TCE did not seem to be causing overt liver toxicity in the rat. Organ weights
were not reported in this study, which would have been hard to interpret if they had been
reported because of the mortality.
2.2.14. Fukuda et al. 1983
In this 104-week bioassay designed primarily to determine a carcinogenic response,
female non-inbred Cij :CD-1 (ICR) mice and female Crj :CD (SD) rats 7 weeks of age were
exposed to "reagent grade" TCE at 0, 50, 150, and 450 ppm for 7 hours a day, 5 days a week.
During the 2-year duration of the experiment inhalation concentrations were reported to be
within 2% of target values. The numbers of animals per group were reported to be 49-50 mice
and 49-51 rats at the beginning of the experiment. The impurities in the TCE were reported to be
0.128 % carbon tetrachloride benzene, 0.019% epichlorohydrin and 0.019% 1,1,2-
trichloroethane. After 107 weeks from commencement of the exposure, surviving animals were
reported to be killed and completely necropsied. "Tumors and abnormal organs as well as other
major organs were excised and prepared for examination in H&E sections." No other details of
the methodologies used for pathological examination of tissues were given including what areas
of the liver and number of sections examined by light microscopy.
Body weights were not given but the authors reported that "body weight changes of the
mice and rats were normal with a normal range of standard deviation." It was also reported that
there were no significant differences in average body weight of animals at specified times during
the experiments and no significant difference in mortality between the groups of mice. The
report includes a figure showing, that for the first 60 weeks of the experiment, there was a
difference in cumulative mortality at the 450 ppm dose in ICR mice and the other groups. The
authors reported that significantly increased mortalities in the control group of rats compared to
the other dosed groups were observed at 85 weeks and after 100 weeks reflecting many deaths
during the 81-85 week and 96-100 week periods for control rats. No significant comparable
02/12/09
E-77
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
clinical observations were reported to be noted in each group but that major symptoms such as
bloody nasal discharge (in rats), local alopecia (in mice and rats), hunching appearance (in mice)
and respiratory disorders (in mice and rats) were observed in some animals mostly after 1 year.
The authors report that "the numbers of different types of tumors were counted and only
malignant tumors were counted when both malignant and benign tumors were observed within
one organ." They also reported that "all animals were included in the effective numbers except
for a few that were killed accidently, severely autolyzed or cannibalized, and died before the first
appearance of tumors among the groups." In mice the first tumors were observed at 286 days as
thymic lymphoma and most of the malignant tumors appearing later were described as
lymphomas or lymphatic leukemias. The incidences of mice with tumors were 37%, 36%, 54%
and 52 % in the control, 50 ppm, 150 ppm and 450 ppm groups, respectively, by the end of the
experiment. "Tumors of the ovary, uterus, subcutaneous tissue, stomach and liver were observed
in the dose groups at low incidences {2-1%) but not in the controls." For the liver, the control,
50 ppm, and 150 ppm groups were all reported to have no liver tumors with one animal (2%)
having an adenoma at the 450 ppm dose. For rats the first tumor was reported to be observed at
410 days and for the incidences of animals with tumors to be 64%, 78%, 66%, and 63% for
control, 50 ppm, 150 ppm, and 450 ppm TCE, respectively, by the end of the experiment. Most
tumors were distributed in the pituitary gland and mammary gland with other tumors reported at
a low incidence of 2-4% with none in the controls. For the liver there were no liver tumors in the
control or 150 ppm groups but 1 animal (2%) had a cystic cholangioma in 50 ppm group and one
animal (2%) had a hepatocellular carcinoma in the 450 ppm group of rats. No details concerning
the pathology of the liver or organ weight changes were given by the authors, including any
incidences of hepatomegaly or preneoplastic foci. On note is that in these strains, there were no
background liver tumors in either strain, indicative of the relative insensitivity of these strains to
hepatocarcinogenicity. However, the carcinogenic potential of TCE was reflected by a number
of other tumor sites in this paradigm.
2.2.15. Henschler et al. 1980
This report focused on the potential carcinogenic response of TCE in mice (NMRI
random bred), rats (WIST random bred) and hamsters (Syrian random bred) exposed to 0, 100
and 500 ppm TCE for 6 hours/day 5 days/week for 18 months. The TCE used in the experiment
was reported to be pure with the exception of trace amounts of chlorinated hydrocarbons,
epoxides and triethanolamines (< 0.000025%) w/w) and stabilized with 0.0015%) triethanolamine.
The number of animals in each group was 30 and the ages and initial and final body weights of
the animals were not provided in the report. For the period of exposure (8 am - 2 pm), animals
were deprived of food and water. The exposure period was for 18 months with mice and
hamsters sacrificed after 30 months and rats after 36 months. "Deceased animals" were reported
to be autopsied, spleen, liver, kidneys, lungs and heart weighed, and these organs, as well as
stomach, CNS, and tumorous tissues, examined in H&E sections.
02/12/09
E-78
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Body weight gain was reported to be normal in all species with no noticeable differences
between control and exposed groups but data was not shown. However, a "clearly dose-
dependent decrease in the survival rate for both male and female mice" was reported to be
statistically significant in both sexes and concentrations of TCE with no other significant
differences reported in other species. The increase in mortality was more pronounced in male
mice, especially after 50 weeks of exposure. Hence the opportunity for tumor development was
diminished due to decreased survival in TCE treated groups. No organ weights were provided
for the study due to the design, in which at considerable period of time occurred between the
cessation of exposure and the sacrifice of the animals and liver weights changes due to TCE may
have been diminished with time. For the 30 autopsied male mice in the control group, 1
hepatocellular adenoma and 1 hepatocellular carcinoma was reported. Whether they occurred in
the same animal cannot be determined from the data presentation. In the 29 animals in thelOO
ppm TCE exposure group 2 hepatocellular adenomas and 1 mesenchymal liver tumor were
reported but no hepatocellular carcinomas also without a determination as whether they occurred
in the same animal or not. In the 30 animals autopsied in the 500 ppm exposure group no liver
tumors were reported. In female mice, of the 29 animals autopsied in the control group, 30
animals autopsied in the 100 ppm group, and the 28 animals autopsied in the 500 ppm group,
there were also no liver tumors reported.
In both the 100 ppm and 500 ppm exposure groups, of male mice especially, low
numbers of animals studied, abbreviated TCE exposure duration, and lower numbers of animals
surviving to the end of the experiment, limit the power of this study to determine a treatment-
related difference in liver carcinogenicity. As discussed in section 2.3.2. below, the use of an
abbreviated exposure regime or study duration and low numbers of animals examined limits the
power of a study to detect a treatment-related response. The lack of any observed background
liver tumors in the female mice and a very low background level of 2 tumors in the male mice
are indicative of a low sensitivity to detect liver tumors in this paradigm which may have
occurred either through its design or a low sensitivity of mouse strain used for this endpoint.
However, the carcinogenic potential of TCE in mice was reflected by a number of other tumor
sites in this paradigm.
For rats and hamsters the authors reported "no dose-related accumulation of any kind of
tumor in either sex of these species." For male rats there was only 1 hepatocellular
adenoma reported at 100 ppm in the 30 animals autopsied and no carcinomas. For female rats
there were no liver tumors reported in control animals but, more significantly, at 100 ppm there
was 1 adenoma and 1 cholangiocarcinoma reported at 100 ppm and at 500 ppm 2
cholangioadenomas. Although not statistically significant, the occurrence of this relatively rare
biliary tumor was observed in both TCE dose groups in female rats. The difference in survival,
as reported in mice, did not affect the power to detect a response in rats, but the low numbers of
animals studied, abbreviated exposure duration and apparent low sensitivity to detect a
hepatocarcinogenic response suggest a study of low power. Nevertheless, the occurrence of
cholangioadenomas and 1 cholangiocarcinoma in female rats after TCE treatments is of concern,
especially given the relationship in origin and proximity of the bile and liver cells and the low
02/12/09
E-79
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
incidence of this tumor. For hamsters the low background rate of tumors of any kind suggests
that in this paradigm, the sensitivity for detection of this tumor is relatively low.
2.2.16. Maltoni et al. 1986
The report by Maltoni et al. (1986) included a series of "systematic and integrated
experiments (BT 301, 302, 303, 304, 304bis, 305, 306 bis) started in sequence, testing TCE by
inhalation and by ingestion." The first experiment (BT 301) was begun in 1976 and the last in
1983 with this report representing the complete report of the findings and results of project. The
focus of the study was detection of a neoplastic response with only a generalized description of
tumor pathology phenotype given and no reporting of liver weight changes induced by TCE
exposure.
In experiment BT 301, TCE was administered in male and female Sprague-Dawley rats
(13 weeks at start of experiment) via olive oil gavage at control, 50 mg/kg or 250 mg/kg
exposure levels for 52 weeks (4-5 days weekly). The animals (30 male, 30 female for each dose
group) were examined during their lifetime. In experiment BT 302, male and female Sprague-
Dawley rats (13 weeks old at start of the experiment) were exposed to TCE via inhalation at 0,
100, and 600 ppm, 7 hours a day, 5 days a week, for 8 weeks. The animals (90 animals in each
control group, 60 animals in each 100 ppm group, and 72 animals in each 600 ppm group) were
examined during their lifetime. In experiment BT 304, male and female Sprague Dawley rats
(12 weeks old at start of the experiment) were exposed TCE via inhalation at 0, 100, 300, and
600 ppm 7 hours a day, 5 days a week, for 104 weeks. The animals (95 male, 100 female rats
control groups, 90 animals in each 100 ppm group, 90 animals in each 300 ppm group, and 90
animals in each 600 ppm group) were examined during their lifetime. In experiment BT304bis,
male and female Sprague Dawley rats (12 weeks old at start of the experiment) were exposed to
TCE via inhalation at 0, 100, 300, and 600 ppm for 7 hours a day, 5 days a week, for 104 weeks.
The animals (40 male, 40 female rats control groups, 40 animals in each 100 ppm group, 40
animals in each 300 ppm group, and 40 animals in each 600 ppm group) were examined during
their lifetime.
In experiment BT 303, Swiss mice (11 weeks old at the start of the experiment) were
exposed to TCE via inhalation in for 8 weeks using the same exposure concentrations as for
experiment BT 302. The animals (100 animals in each control group, 60 animals in the 100 ppm
exposed group, and 72 animals in each 600 ppm group) were examined during their lifetime. In
experiment BT 305, Swiss mice (11 weeks old at the start of the experiment) were exposed to
TCE via inhalation in for 78 weeks, 7 hours a day, 5 days a week. The animals (90 animals in
each control group, 90 animals in the 100 ppm exposed group, 90 animals in the 300 ppm group,
and 90 animals in each 600 ppm group) were examined during their lifetime. In experiment BT
306, B6C3F1 mice (from NCI source) (12 weeks old at the start of the experiment) were exposed
to TCE via inhalation in for 78 weeks, 7 hours a day, 5 days a week. The animals (90 animals in
each control group, 90 animals in the 100 ppm exposed group, 90 animals in the 300 ppm group,
and 90 animals in each 600 ppm group) were examined during their lifetime. In experiment BT
02/12/09
E-80
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
306bis B6C3F1 mice (from Charles River Laboratory as source) (12 weeks old at the start of the
experiment) were exposed to TCE via inhalation for 78 weeks, 7 hours a day, 5 days a week. The
animals (90 animals in each control group, 90 animals in the 100 ppm exposed group, 90 animals
in the 300 ppm group, and 90 animals in each 600 ppm group) were examined during their
lifetime.
In all experiments, TCE was supplied tested and reported by the authors of the study to
be was highly purified and epoxide free with butyl-hydroxy-toluene at 20 ppm used as a
stabilizer. Extra-virgin olive oil was used as the carrier for ingestion experiments and was
reported to be free of pesticides. The authors describe the treatment of the animals and running
of the facility in detail and report that: "Animal rooms were cleaned every day and room
temperature varied from 19 degrees to 22 degrees and was checked 3 times daily. Bedding was
changed every two days and cages changes and washed once weekly. The animals were handled
very gently and, therefore, were neither aggressive nor nervous. Concentrations of TCE were
checked by continuous gas-chromatographic monitoring. Treatment was performed by the same
team. In particular, the same person carried out the gavage of the same animals. This is
important, since animals become accustomed to the same operators. The inhalation chambers
were maintained at 23 ± 2 degrees C and 50 ± 10% relative humidity. Ingestion from Monday to
Friday was usually performed early in the morning. The status and behavior of the animals were
examined at least three times daily and recorded. Every two weeks the animals were submitted
to an examination for the detection of the gross changes, which were registered in the
experimental records. The animals which were found moribund at the periodical daily inspection
were isolated in order to avoid cannibalism. The animals were weight every two weeks during
treatment and then every eight weeks. Animals were kept under observation until spontaneous
death. A complete necropsy was performed. Histological specimens were fixed in 70% ethyl
alcohol. A higher number of samples was taken when particular pathological lesions were seen.
All slides were screened by a junior pathologist and then reviewed by a senior pathologist. The
senior pathologist was the same throughout the entire project. Analysis of variance was used for
statistical evaluation of body weights. Results are expressed as means and standard deviations.
Survival time is evaluated using the Kruskal-Wallis test. For different survival rates between
groups, the incidence of lesions is evaluated by using the Log rank test. Non-neoplastic,
preneoplastic, and neoplastic lesions were evaluated using the Chi-square of Fisher' exact test.
The effect of different doses was evaluated using the Cochran-Armitage test for linear trends in
proportions and frequencies." The authors state that: "Although the BT project on TCE was
started in 1976 and most of the experiments were performed from the beginning of 1979, the
methodological protocol adopted substantially met the requirements of the Good Laboratory
Practices Act." Finally, it was reported that "the experiments ran smoothly with no accidents in
relation to the conduct of the experiment and the health of the animals, apart from an excess in
mortality in the male B6C3F1 mice of the experiment BT 306, due to aggressiveness and
fighting among the animals." This is in contrast to the description of the gavage studies
conducted by NTP (NTP 1990, 1988) in which gavage error resulted in significant loss of
experimental animals. Questions have been raised about the findings, experimental conditions,
and experimental paradigm of the European Ramazzini Foundation (ERF) from which the
02/12/09
E-81
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Maltoni et al (1986) experiments were conducted (EFSA, 2006). However, these concerns were
addressed by Caldwell et al. (2008a), who concluded that the ERF bioassay program produced
credible results that were generally consistent with those of NTP
In regards to effects of TCE exposure on survival, "a nonsignificant excess in mortality
correlated to TCE treatment was observed only in female rats (treated by ingestion with the
compound) and in male B6C3F1 mice. In B6C3F1 mice of the experiment BT 306 bis, the
excess in mortality in treated animals was higher (p < 0.05 after 40 weeks) but was not dose
correlated. No excess in mortality was observed in the other experiments." The authors reported
that "no definite effect of TCE on body weight was observed in any of the experiments, apart
from experiment BT 306 bis, in which a slight non-dose correlated decrease was found in
exposed animals."
In mice, "hepatoma" was the term used by the authors of these studies to describe all
malignant tumors of hepatic cells, of different sub-histotypes, and of various degrees of
malignancy. The authors reported that the hepatomas induced by exposure to TCE "may be
unique or multiple, and have different sizes (usually detected grossly at necropsy). Under
microscopic examination these tumors proved to be of the usual type observed in Swiss and
B6C3F1 mice, as well as in other mouse strains, either untreated or treated with
hepatocarcinogens. They frequently have medullary (solid), trabecular, and pleomorphic
(usually anaplastic) patterns. The hepatomas may produce distant metastases, more frequently in
the lungs."
In regard to the induction of "hepatomas" by TCE exposure, the authors report that in
Swiss mice exposed to TCE by inhalation for 8 weeks (BT303), the percentage of animals with
hepatomas was 1.0% in male mice and 1.0% in female mice in the control group (n= 100 for
each gender). For animals exposed to 100 ppm TCE, the percentage in female mice was 1.7%
and male mice 5.0% (n = 60 for each gender). For animals exposed to 600 ppm TCE, the
percentage in female mice was 0% and in male mice 5.5% (n= 72 for each gender). The
relatively larger number of animals used in this bioassay, in comparison to NTP standard assays,
allows for a greater power to detect a response. It is also apparent from these results that Swiss
mice in this experimental paradigm are a "less sensitive" strain in regard to spontaneous liver
cancer induction over the lifetime of the animals. These results suggest that 8 weeks of TCE
exposure via inhalation at 100 ppm or 600 ppm may have been associated with a small increase
in liver tumors in male mice in comparison to concurrent controls.
In Swiss mice exposed to TCE via inhalation for 78 weeks (BT 305), the percentage of
animals with hepatomas was reported to be 4.4% in male mice and 0% in female mice in the
control group (n= 90 for each gender). For animals exposed to 100 ppm TCE, the percentage in
female mice was reported to be 0% and male mice 2.2% (n = 90 for each gender). For animals
exposed to 300 ppm TCE, the percentage in female mice was reported to be 0% and in male
mice 8.9% (n= 90 for each gender). For animals exposed to 600 ppm TCE, the percentage in
female mice was reported to be 1.1% and in male mice 14.4%. As with experiment BT303, there
is a consistency in the relatively low background level of hepatomas reported for Swiss mice in
02/12/09
E-82
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
this paradigm. After 78 weeks of exposure there appears to be a dose-related increase in
hepatomas in male but not female Swiss mice via inhalation exposure.
In B6C3F1 mice exposed to TCE by inhalation for 78 weeks (BT306) the percentage of
animals with hepatomas was reported to be 1.1% in male mice and 3.3 % in female mice in the
control group (n= 90 for each gender). For animals exposed to 100 ppm TCE, the percentage in
female mice was reported to be 4.4% and in male mice 1.1% (n = 90 for each gender). For
animals exposed to 300 ppm TCE, the percentage in female mice was reported to be 3.3% and in
male mice 4.4% (n= 90 for each gender). For animals exposed to 600 ppm TCE, the percentage
in female mice was reported to be 10.0% and in male mice 6.7%. This was the experimental
group with excess mortality in the male group due to fighting. The excess mortality could have
affected the results. The authors do report that there was a difference in the percentage of males
bearing benign and malignant tumors that was due to early mortality among males in experiment
BT306. It is unexpected for the liver cancer incidence to be less in male mice than female mice
and not consistent with the results reported for the Swiss mice.
In B6C3F1 male mice exposed to TCE via inhalation (BT 306 bis) the percentage of
animals with hepatomas was reported to be 18.9% in male mice in the control group (n= 90).
For animals exposed to 100 ppm TCE, the percentage in male mice was reported to be 21.1% (n
= 90). For animals exposed to 300 ppm TCE, the percentage in male mice was reported to be
30.0 % (n= 90). For animals exposed to 600 ppm TCE, the percentage in male mice was
reported to be 23.3%. This experiment did not examine female mice. The authors do report a
decrease in survival in mice from this experiment that could have affected results. It is apparent
from the BT 306 and BT 306 bis experiments that the background level of liver cancer was
significantly different in male mice, although they were supposed to be of the same strain. The
finding of differences in response in animals of the same strain but from differing sources has
also been reported in other studies for other endpoints (see Section 3.1.2., below).
The authors reported 4 liver angiosarcomas: 1 in an untreated male rat (BT 304); 1 in a
male and 1 in a female rat exposed to 600 ppm TCE for 8 weeks (experiment BT302); and 1 in a
female rat exposed to 600 ppm TCE for 104 weeks (BT 304). The authors conclude that "the
tumors observed in the treated animals cannot be considered to be correlated to TCE treatment,
but are spontaneously arising. These findings are underlined because of the extreme rarity of this
tumor in control Sprague Dawley rats, untreated or treated with vehicle materials. The
morphology of these tumors is of the liver angiosarcoma type produced by vinyl chloride in this
strain of rats."
In rats treated for 104 weeks, TCE was reported to not affect the percentages of animals
bearing benign and malignant tumor and of animals bearing malignant tumors. Moreover, it did
not affect the number of total malignant tumors per 100 animals. This study did not report a
treatment related increase in liver cancer in rats. The report only explicitly described positive
findings so it is assumed that there were no increases in "hepatomas" in rat liver associated with
TCE treatment. The authors concluded that "under the tested experimental conditions, the
02/12/09
E-83
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
evidence of TCE (without epoxide stabilizer) carcinogenicity, gives the result of TCE treatment-
related hepatomas in male Swiss and B6C3F1 mice. A borderline increased frequency of
hepatomas was also seen after 8 weeks of exposure in male Swiss mice." Thus, the increase in
liver tumors in both strains of mice exposed to TCE via inhalation reported in this study is
consistent with the gavage results from the NTP 1990 study in B6C3F1 mice, where male mice
had a higher background level and greater response from TCE exposure than females.
2.2.17. Maltoni et al. 1988
This report was an abbreviated description of an earlier study (Maltoni et al 1986)
focusing on the identification of a carcinogenic response in rats and mice by chronic TCE
exposure.
2.2.18. Van Duuren et al. 1979
This study exposed male and female non-inbred HA:ICR Swiss mice at 6-8 weeks of age
to distilled TCE with no further descriptions of purity. Gavage feeding of TCE was once weekly
in 0.1 ml trioctanoin. Neither initial or final body weights were reported by the authors. The
authors reported that, at the termination of the experiments or at death, animals were completely
autopsied with specimens of all abnormal-appearing tissues and organs excised for
histopathologic diagnosis. Tissues from the stomachs, livers and kidneys were reported to be
taken routinely for the intragastric feeding experiments. Tissues were reported to be stained for
H&E for pathologic examination, but no further description of the lobe(s) of the liver examined
or the sections examined was provided by the authors. Results were as only reported the no of
mice with forestomach tumors 0.5 mg/ mouse of TCE treatment given once a week in 0.1 ml
trioctanoin. Mouse body weights were not given so the dose in mg/kg for the mice cannot be
ascertained. The protocol used in this experiment kept the mg/mouse constant with a 1 week
dosing schedule so that as the mice increased weight with age, the dose as a function of body
weight was decreased. The days on test were reported to be 622 for 30 male and female mice. 2
male and 1 female mice were reported as having forestomach tumors. For 30 mice treated with
trioctanoin alone the number of forestomach tumors was reported to be zero. For mice with no
TCE treatment, 5 of 100 male mice were reported to have forestomach tumors and of 8 of 60
female mice were reported to have forestomach tumors for 636 and 649 days on test. No results
for liver were presented by the authors by the intragastric route of administration including
background rates of the incidences of liver tumors or treatment results. The authors note that
except for repeated skin applications of certain chemicals, no significant difference between the
incidence of distant tumors in treated animals compared with no-treatment and vehicle control
groups was noted. Given the uncertainties in regard to dose, the once-a week dosing regime, the
low number of animals tested with resulting low power, and the lack of reporting of experimental
results, the ability to use the results from this experiment in regard to TCE carcinogenicity is
very limited.
02/12/09
E-84
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.2.19. NCI 1976
This bioassay was "initiated in 1972 according to the methods used and widely accepted
at that time" with the design of carcinogenesis bioassays having "evolved since then in some
respects and several improvements" having been developed. The most notable changes reported
in the foreward of the report are changes "pertaining to preliminary toxicity studies, numbers of
controls used, and extent of pathological examination." Industrial grade TCE was tested (99%
TCE, 0.19% 1,2,-epoxybutane, 0.04%v ethyl acetate, 0.09% epichlorhydrin, 0.02% N-methyl
pyrrole, and 0.03% diisobutylene) with rats and mice exposed via gavage in corn oil 5
times/week for 78 weeks using 50 animals per group at 2 doses with both sexes of Osborne-
Mendel rats and B6C3F1 mice. However for control groups only 20 of each sex and species
were used. Rats were killed after 110 weeks and mice after 90 weeks. Rats and mice were
initially 48 and 35 days of age, respectively, at the start of the experiment with control and
treated animals born within 6 days of each other. Initial weight ranges were reported as ranges
for treated and control animals of 168-229 g for male rats, 130-170 g for female rats, 11-22 g for
male mice, and 11-18 g for female mice. Animals were reported to be "randomly assigned to
treatment groups so that initially the average weight in each group was approximately the same."
Mice treated with TCE were reported to be "maintained in a room housing other mice being
treated with one of the following 17 compounds: 1,1,2-2-tetrachloroethane, chloroform, 3-
chloropropene, chloropicrin, 1,2-dibromochloropropane, 1,2, dibromoethane, ethylene
dichloride, 1,1-diochloroethane, 3-sulfolene, idoform, methyl chloroform, 1,1,2-trichloroethane,
tetrachloroethylene, hexachloroethane, carbon disulfide, trichlorofluoromethane, and carbon
tetrachloride. Nine groups of vehicle controls and 9 groups of untreated controls were also
housed in this same room." The authors note that "TCE-treated rats and their controls were
maintained in a room housing other rats being treated with one of the following compounds:
dibromochloropropane, ethylene di chloride, 1,1-dichloroethane, and carbon disulfide. Four
groups of vehicle-treated controls were in the same room." Thus, there was the potential of co-
exposure to a number of other chemicals, especially for the mice, resulting from exhalation in
treated animals housed in the same room, including the control groups, as noted by the authors.
The authors also noted that "samples of ambient air were not tested for presence of volatile
materials" but state that "although the room arrangement is not desirable as is stated in the
Guidelines for Carcinogen Bioassay in Small Rodents, there is not evidence the results would
have been different with a single compound in a room."
The initial doses of TCE for rats were reported to be 1300 and 650 mg/kg. However,
these levels were changed based on survival and body weight data "so that the time-weighted
average doses were 549 and 1097 mg/kg for both male and female rats." For mice, the initial
doses were reported to be 1000 and 2000 mg/kg for males and 700 and 1400 mg/kg for females.
The "doses were increased so that the time weighted average doses were 1169 mg/kg and 2339
mg/kg for male mice and 869 and 1739 mg/kg for female mice." The authors reported that signs
of toxicity, including reduction in weight, were evident in treated rats which, along with
increased mortality, "necessitated a reduction in doses during the test." In contrast "very little
evidence of toxicity was seen in mice, so doses were increased slightly during the study." Doses
02/12/09
E-85
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
were "changed for the rats after 7 and 16 weeks of treatment, and for the mice after 12 weeks."
At 7 weeks of age, male and female rats were dosed with 650mg/kg TCE, at 14 weeks they were
dosed with 750 mg/kg TCE, and at 23 weeks of age 500 mg/kg TCE. For the high exposure
level, the exposure concentrations were 1300 mg/kg, 1500 mg/kg, and 1000 mg/kg TCE,
respectively, for the same changes in dosing concentration. For rats the percentage of TCE in
corn oil remained constant at 60%. For female mice, the TCE exposure at the beginning of
dosing was 700 mg/kg TCE (10% in corn oil) at 5 weeks of age for the "lower dose" level. The
dose was increased to 900 mg/kg day (18% in corn oil) at 17 weeks of age and maintained until
83 weeks of age. For male mice, the TCE exposure at the beginning of dosing was 1000 mg/kg
TCE (15%) in corn oil) at 5 weeks of age for the "lower dose" level. At 11 weeks, the level of
TCE remained the same but the percentage of TCE in corn oil was reduced to 10%>. The dose
was increased to 1200 mg/kg day at 17 weeks of age (24% in corn oil) and maintained until 83
weeks of age. For the "higher dose," the TCE exposure at the beginning of dosing was 1400
mg/kg TCE (10%) in corn oil) at 5 weeks of age in female mice. At 11 weeks of age the exposure
level of TCE was kept the same but the percentage of TCE in corn oil increased to 20%. By 17
weeks of age the exposure concentration of TCE in corn oil was increased to 1800 mg/kg (18%
in corn oil) in female mice. For the "higher dose" in male mice, the TCE exposure at the
beginning of dosing was 2000 mg/kg (15%> in corn oil) which was maintained at 11 weeks in
regard to TCE administered but the % of TCE corn oil was increased to 20%. For male mice the
exposure concentration was increased to 2400 mg/kg (24% in corn oil). For all of the mice
treatment continued on a 5 days/week schedule of oral gavage dosing throughout the timecourse
of treatment (78 weeks of treatment). Thus, not only did the total dose administered to the
animals change, but the volumes of vehicle in which TCE was administered changed throughout
the experiment.
The authors stated that at 37 weeks of age, "To help assure survival until planned
termination the dosing schedule was changed for rats to a cycle of 1 week of no treatment
followed by 4 weeks of treatment." for male and female rats. Thus, the duration of exposure in
rats was also changed. All lobes of the liver were reported to be taken including the free margin
of each lobe with any nodule or mass represented in a block 10x5x3 mm cut from the liver and
fixed in a marked capsule.
Body weights (mean ± Std) were reported to be 193 ± 15.0 g (n=20), 193 ± 15.8 g
(n=50), and 195 ± 16.7 g (n=50) for control, low, and high dose male rats at initiation of the
experiment. By 1 year of exposure (50 weeks), 20/20 control male rats were still alive to be
weighed, 42/50 of the low dose rats were alive and 34/50 of high dose rats were still alive. The
body weights of those remaining were decreased by 6.2% and 17% in the low and high dose
animals in comparison with the controls. For female rats, the mean body weights were reported
to be 146 ± 11.4 g (n= 20), 144 ± 11.0 g (n= 50), and 144 ± 9.5 g (n=50) for control, low, and
high dose female rats at initiation of the experiment. By 1 year of exposure (50 weeks), 17/20
control female rats were still alive, 28/50 low dose and 39/50 of the high dose rats were alive.
The body weights of those remaining were decreased by 25% and 30% in the low and high dose
animals in comparison with the controls. For male mice the initial body weights were 17 ± 0.5 g
02/12/09
E-86
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(n=20), 17 ± 2.0 g (n=50), and 17 ± 1.1 g (n=50) for control, low and high doses. By 1 year of
exposure (50 weeks), 18/20 control male mice were still alive, 47/50 or the low dose, and 34/50
of the high dose groups were still alive. The body weights of those remaining were unchanged in
comparison to controls. For female mice the initial body weights were 14 ± 0.0 g (n=20), 14 ±
0.6 g (n=50), and 14 ± 0.7 g (n=50) for control, low and high doses. By 1 year of exposure (50
weeks), 18/20 control male mice were still alive, 45/50 or the low dose, and 41/50 of the high
dose groups were still alive. The body weights of those remaining were unchanged in
comparison to controls.
A high proportion of rats were reported to die during the experiment with 17/20 control,
42/50 low dose, and 47/50 high dose animals dying prior to scheduled termination. For female
rats, 12/20 control, 35/48 low dose, and 37/50 high dose animals were reported to die before
scheduled termination with two low dose females reported to be missing and not counted in the
denominator for that group. The authors reported that earlier death was associated with higher
TCE dose. A decrease in the percentage of tumor-bearing animals was reported to be lower in
treated animals and attributed by the authors to be likely related to the decrease in their survival.
A high percentage of respiratory disease was reported to be observed among the rats without any
apparent difference in the type, severity, or morbidity as to sex or group. The authors reported
that "no significant toxic hepatic changes were observed" but no other details regarding results in
the liver of rats. Carbon tetrachloride was administered to rats as a positive control. A low
incidence of both hepatocellular carcinoma and neoplastic nodule was reported to be found in
both colony controls (1/99 hepatocellular carcinoma and 0/99 neoplastic nodule in male rats and
0/98 hepatocellular carcinoma and 2/98 neoplastic nodules in female rats) and carbon-
tetrachloride-treated rats. Hepatic adenomas were included in the description of neoplastic
nodules in this study with the diagnosis of hepatocellular carcinoma to be "based on the presence
of less organized architecture and more variability in the cells comprising the neoplasms."
The authors reported that "increased mortality in treated male mice appears to be related
to the presence of liver tumors." For mice both male and female mice the incidences of
hepatocellular carcinoma were reported to be high from TCE treatment with 1/20 in age matched
controls, 26/50 in low dose and 31/48 in high dose males. Colony controls for male mice were
reported to be 5/77 for vehicle and 5/70 for untreated mice. For females mice hepatocellular
carcinomas were reported to be observed in 0/20 age matched controls, 4/50 low dose, and 11/47
high-dose female mice. Colony controls for female mice were reported to be 1/80 for vehicle
and 2/75 for untreated mice. In male mice, hepatocellular carcinomas were reported to be
observed early in the study with the first seen at 27 weeks. Hepatocellular carcinomas were not
observed so early in low dose male or female mice.
The diagnosis of hepatocellular carcinoma was reported to be based on histologic
appearance and the presence of metastasis especially to the lung with not other lesions
significantly elevated in treated mice. The tumors were reported to be "varied from those
composed of well differentiated hepatocytes in a relatively uniform trabecular arrangement to
rather anaplastic lesions in which mitotic figures occurred in cells which varied greatly in size
02/12/09
E-87
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and tinctorial characteristics. Many of the tumors were characterized by the formation of
relatively discrete areas of highly anaplastic cells within the tumor proper which were, in turn,
surrounded by relatively well differentiated neoplastic cells. In general, various arrangements of
the hepatocellular carcinoma occurred, as described in the literature, including those with an
orderly cord-like arrangement of neoplastic cells, those with a pseudoglandular pattern
resembling adenocarcinoma, and those composed of sheets of highly anaplastic cells with
minimal cord or gland-like arrangement. Multiple metaplastic lesions were observed in the
lung, including several neoplasms which were differentiated and relative benign in appearance."
The authors noted that almost all mice treated with carbon tetrachloride exhibited liver tumors
and that the "neoplasms occurring in treated [sic carbon tetrachloride treated] mice were similar
in appearance to those noted in the trichloroethylene-treated mice." Thus, phenotypically this
study reported that the liver tumors induced in mice by TCE were heterogeneous and typical of
those arising after carbon tetrachloride administration. The descriptions of liver tumors in this
study and the tendency of metastasis to the lung are similar to the descriptions provided by
Maltoni et al., (1986) for TCE-induced liver tumors in mice via inhalation.
In terms of non-cancer pathology of the liver, 1 control male rat was reported to display
fatty metamorphosis of the liver at 102 weeks. However for the low dose, 3 male rats were
reported to display fatty metamorphosis (90, 110, and 110 weeks), 2 rats to display cystic
inflammation (76, 110 weeks), and one rat to display general inflammation (110 weeks). At the
high dose, 6 rats were reported to display fatty metamorphosis (12, 35, 49, 52, 52, and 58
weeks), 1 rat was reported to display cytomegaly (42 weeks), 2 rats were reported to display
centrilobular degeneration (53 and 58 weeks), 1 rats to display diffuse inflammation (62 weeks),
1 rat to display congestion (week 12), and 5 rats to display angiectasis or abnormally enlarged
blood vessels which can be manifested by hyperproliferation of endothelial cells and dilatation of
sinusoidal spaces (35, 42, 52, 54, and 65 weeks). One control female rat was reported o display
fatty metaphorphosis of the liver at 110 weeks, and one control female rats to display
"inflammation" of the liver at 110 weeks. Of the TCE dosed female rats, only 1 high dose
female rat displayed fatty metaphorphosis at week 96. Thus for male rats, there was liver
pathology present in some rats due to TCE exposure examined from 12 weeks to a year at their
time of their premature death. For mice the liver pathology was dominated by the presence of
hepatocellular carcinoma with additional hyperplasia noted in 2 mice of the high dose male and
female groups and 1 or less mouse exhibiting hyperplasia in the control or low dose groups.
The authors note that "while the absence of a similar effect in rats appears most likely
attributable to a difference in sensitivity between the Osborne-Mendel rat and B6C3F1 mouse,
the early mortality of rats due to toxicity must also be considered." The conclude that "the test in
rats is inconclusive: large numbers of rats died prior to planned termination; in addition, the
response of this rat strain to the hepatocarcinogenicity of the positive control compound, carbon
tetrachloride, appeared relatively low." Finally, the authors note that "while the results obtained
in the present bioassay could possibly have been influenced by an impurity in the TCE used, the
extremely low amounts of impurities found make this improbable."
02/12/09
E-88
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.2.20. Herren-Freund et al. 1987
This study was gave results primarily in initiated male B6C3 F1 mice that were also
exposed to TCE metabolites in drinking water for 61 weeks. However in Table 1 of the report,
results were given for mice that received no initiator but were given 40 mg/1 TCE or 2 g/1 NaCL
as control. The mice were reported to be 28 days of age when placed on drinking water
containing TCE. The authors reported that concentrations of TCE fell by about V2 at the 40 mg/1
dose of TCE during the twice a week change in drinking water solution. For control animals
(n=22) body weight at termination was reported to be 32.93 ± 0.54 g, and liver weight was 1.80
± 0.05 g, % liver/body weight was 5.47 ± 0.16%. For TCE treated animals (n=32), body weight
at termination was reported to be 35.23 ± 0.66 g, and liver weight was 1.97 ± 0.10 g, %
liver/body weight was 5.57 ± 0.24%. Thus, hepatomegaly was not reported for this paradigm at
this time of exposure. The study reported that for 22 control animals, the prevalence of
adenomas was 2/22 animals (or 9%) with the mean number of adenomas per animal to be 0.09 ±
0.06 (SEM). The prevalence of carcinomas in the control group was reported to be 0/22. For 32 animals
exposed to 40 mg/1 TCE, the prevalence of adenomas was 3/32 animals (or 9%) with the mean number of
adenomas per animal tobe0.19±0.12 (SEM). The prevalence of animals with hepatocellular
carcinomas was 3/32 animals (or 9%) with the mean number of hepatocellular carcinomas to be 0.10 ±
0.05 (SEM). Thus, similar to the acute study of Tucker et al. (1982), significant loss of TCE is a
limitation for trying to evaluate TCE hazard in drinking water. However, despite difficulties in
establishing accurately the dose received, an increase in adenomas per animal and an increase in
the number of animals with hepatocellular carcinomas were reported to be associated with TCE
exposure after 61 weeks of exposure. Also of note is that the increase in tumors was reported
without significant increases in hepatomegaly at the end of exposure. The authors did not report
these increases in tumors as being significant but did not do a statistical test between TCE
exposed animals without initiation and control animals without initiation. The low numbers of
animal tested limits the statistical power to make such a determination. However for carcinomas,
there was none reported in controls but 9% of TCE-treated mice had hepatocellular carcinomas.
2.2.21. Anna et al. 1994:
The report focused on presenting incidence of cancer induction after exposure to TCE or
its metabolites and included a description of results for male B6C3F1 mice (8 weeks old at the
beginning of treatment) receiving 800 mg/kg/day TCE via gavage in corn oil, 5 days/week for 76
weeks. There was very limited reporting of results other than tumor incidence. There was no
reporting of liver weights at termination of the experiment. Although the methods section of the
report gives 800 mg/kg/day as the exposure level, Table 1 in the results section reports that TCE
was administered at 1700 mg/kg/day. This could be a typographical error in the table as a
transposition with the dose of "perc" administered to other animals in the same study. The
methods section of the report states that the authors based their dose in mice that used in the
1990 NTP study. The NTP study only used alOOO mg/kg/day in mice suggesting that the table is
mislabeled and suggests that the actual dose is 800 mg/kg/day in the Anna et al., 1994 study. All
treated mice were reported to be alive after 76 weeks of treatment. For control animals, 10
02/12/09
E-89
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
animals exposed to corn oil, and 10 untreated controls were killed in a 9-day period. The
remaining controls were killed at 96, 103, 134 weeks of treatment. Therefore, the control group
(all) contains a mixed group of animals that were sacrificed from 76-134 weeks and were not
comparable to the animals sacrificed at 76 weeks. At 76 weeks 3 of 10 the untreated and two of
the 10 corn oil treated controls were reported to have one small hepatocellular adenoma. None
of the controls examined at 76 weeks were reported to have any observed hepatocellular
carcinomas. The authors reported no cytotoxicity for TCE, corn oil, and untreated control group.
At 76 weeks, 75 mice treated with 800 mg/kg/day TCE were reported to have a prevalence of
50/75 animals having adenomas with the mean number of adenomas per animal to be 1.27 ± 0.14
(SEM). The prevalence of carcinomas in these same animals was reported to be 30/70 with the mean
number of hepatocellular carcinomas per animal to be 0.57 ± 0.10 (SEM). Although not comparable in
terms of time till tumor observation, Corn oil control animals examined at much later time points did
not have as great a tumor response as did those exposed to TCE. At 76-134 weeks 32 mice
treated with corn oil were reported to have a prevalence of 4/32 animals having adenomas with
the mean number of adenomas per animal to be 0.13 ± 0.06 (SEM). The prevalence of carcinomas in
these same animals was reported to be 4/32 with the mean number of hepatocellular carcinomas per
animal to be 0.12 ± 0.06 (SEM). Despite only examining one exposure level of TCE and the
limited reporting of findings other than incidence data, this study also reported that TCE
exposure in male B6C3F1 mice to be associated with increased induction of adenomas and
hepatocellular carcinoma, without concurrent cytotoxicity.
In terms of liver tumor phenotype, Anna et al. reported the % of H-ras codon 61
mutations in tumors from concurrent control animals (water and corn oil treatment groups
combined) examined in their study, historical controls in B6C3 Flmice, and in tumors from TCE
or DCA (0.5% in drinking water) treated animals. From their concurrent controls they reported
that H-ras codon 61 mutations in 17% (n=6) of adenomas and 100% (n=5) of carcinomas. For
historical controls (published and unpublished) they reported mutations in 73% (n=33) of
adenomas and mutations in 70% (n=30) of carcinomas. For tumors from TCE treated animals
they reported mutations in 35% (n=40) of adenomas and 69% (n=36) of carcinomas, while for
DCA treated animals they reported mutations in 54% (n=24) of adenomas and in 68% (n=40) of
carcinomas. The authors reported that "in this study, the H-ras codon 61 mutation frequency was
not statistically different in liver tumors from dichloroacetic acid and trichloroethylene-treated
mice and combined controls (62%, 51% and 69%, respectively). " In regard to mutation spectra
in H-ras oncogenes detected B6C3F1 mouse liver "tumors," the authors reported combined
results for concurrent and historical controls of 58% AAA, 27% CGA, and 14% CTA
substitutions for CAA at Codon 61 out of 58 mutations. For TCE "tumors" the substitution
pattern was reported to be 29% AAA, 24% CGA, and 40% CTA substitutions for CAA at Codon
61 out of 39 mutations and for DCA 28% AAA, 35% CGA, and 38% CTA substitutions for
CAA at Codon 61 out of 40 mutations.
2.2.22. Bull et al. 2002
This study primarily presented results from exposures to TCE, DCA, TCA and
02/12/09
E-90
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
combinations of DCA and TCA after 52 weeks of exposure with some animals examined at 87
weeks. It only examined and described results for liver. In a third experiment, 1000 mg/kg TCE
was administered once daily 7 days a week for 79 weeks in 5% alkamuls in distilled water to 40
B6C3F1 male mice (6 weeks old at the beginning of the experiment). At the time of euthanasia,
the livers were removed, tumors identified, and the tissues section of for examination by a
pathologist and immunostaining. Liver weights were not reported. For the TCE gavage
experiment there were 6 gavage-associated deaths during the course of this experiment among a
total of 10 animals that died with TCE treatment. No animals were lost in the control group.
The limitations of this experiment were discussed in Caldwell et al. (2008b). Specifically, for
the DCA and TCA exposed animals, the experiment was limited by low statistical power, a
relatively short duration of exposure, and uncertainty in reports of lesion prevalence and
multiplicity due to inappropriate lesions grouping (i.e., grouping of hyperplastic nodules,
adenomas, and carcinomas together as "tumors"), and incomplete histopatholology
determinations (i.e., random selection of gross lesions for histopathology examination). For the
reported TCE results, Bull et al., (2002) reported a high prevalence (23/36 B6C3F1 male mice)
of adenomas and hepatocellular carcinoma (7/36) and gave results of an examination of
approximately half of the lesions induced by TCE exposure. Tumor incidence data was provided
for only 15 control mice and reported as 2/15 (13%) having adenomas and 1/15 (7%)
carcinomas. Thus, this study presents results that are consistent with other studies of chronic
exposure that show TCE induction of hepatocellular carcinoma in male B6C3F1 mice.
For determinations of immunoreactivity to c-Jun as a marker of differences in "tumor"
phenotype, Bull et al. (2002) did include all lesions in most of their treatment groups, decreasing
the uncertainty of his findings. The exceptions were the absence of control lesions and inclusion
of only 16/27 and 38/72 lesions for 0.5g/L DCA + 0.05 g/L TCA and lg/kg/day TCE exposure
groups, respectively. Immunoreactivity results were reported for the group of hyperplastic
nodules, adenomas, and carcinomas. Thus, changes in c-Jun expression between the differing
types of lesions were not determined. Bull et al. (2002) reported lesion reactivity to c-Jun
antibody to be dependent on the proportion of the DCA and TCA administered after 52 weeks of
exposure. Given alone, DCA produced lesions in mouse liver for which approximately half
displayed a diffuse immunoreactivity to a c-Jun antibody, half did not, and none exhibited a
mixture of the two. After TCA exposure alone, no lesions were reported to be stained with this
antibody. When given in various combinations, DCA and TCA co-exposure induced a few
lesions that were only c-Jun+, many that were only c-Jun-, and a number with a mixed phenotype
whose frequency increased with the dose of DCA. For TCE exposure of 79 weeks, TCE-induced
lesions also had a mixture of phenotypes (42% c-Jun+, 34% c-Jun-, and 24% mixed and to be
most consistent with those resulting from DCA and TCA co-exposure but not either metabolite
alone.
Mutation frequency spectra for the H-ras codon 61 in mouse liver "tumors" induced by
TCE (n= 37 tumors examined) were reported to be significantly different than that for TCA (n=
41 tumors examined), with DCA-treated mice tumors giving an intermediate result (n= 64
tumors examined). In this experiment, TCA-induced "tumors" were reported to have more
02/12/09
E-91
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mutations in codon 61(44%) than those from TCE (21%) and DCA (33%). This frequency of
mutation in the H-ras codon 61 for TCA is the opposite pattern as that observed for a number of
peroxisome proliferators in which the mutation spectra in tumors has been reported to be much
lower than spontaneously arising tumors (see Section 3.4.1.5.). Bull et al. (2002) noted that the
mutation frequency for all TCE,TCA or DCA was lower in this experiment than for spontaneous
tumors reported in other studies (they had too few spontaneous tumors to analyze in this study),
but that this study utilized lower doses and was of shorter duration than that of Ferreira-Gonzalez
(1995). These are additional concerns along with the effects of lesion grouping in which a lower
stage of progression is group with more advanced stages. In a limited subset of tumor that were
both sequenced and characterized histologically, only 8 of 34 (24%) TCE-induced adenomas but
9/15 (60%) of TCE-induced carcinomas had mutated H-ras at codon 61, which the authors
suggest is evidence that this mutation is a late event.
The issues involving identification of MO A through tumor phenotype analysis are
discussed in detail below for the more general case of liver cancer as well as for specific
hypothesized MO As (see Sections 3.1.4., 3.1.8., 3.2.1., and 3.4.1.5). In an earlier paper, Bull
(2000) suggested that "the report by Anna et al (1994) indicated that TCE-induced tumors
possessed a different mutation spectra in codon 61 of the H-ras oncogene than those observed in
spontaneous tumors of control mice." Bull (2000) stated that "results of this type have been
interpreted as suggesting that a chemical is acting by a mutagenic mechanism" but went on to
suggest that it is not possible to a priori rule out a role for selection in this process and that
differences in mutation frequency and spectra in this gene provide some insight into the relative
contribution of different metabolites to TCE-induced liver tumors. Bull (2000) noted that data
from Anna et al. (1994), Ferreira-Gonzalez et al (1995), and Maronpot et al (1995) indicated that
mutation frequency in DCA-induced tumors did not differ significantly from that observed in
spontaneous tumors, that the mutation spectra found in DCA-induced tumors has a striking
similarity to that observed in TCE-induced tumors, and DCA-induced tumors were significantly
different than that of TCA-induced liver tumors. What is clear from these observations is the
phenotype of TCE-induced tumors appears to be more like DCA-induced tumors (which are
consistent with spontaneous tumors), or those resulting from a co-exposure to both DCA and
TCA, than from those induced by TCA. More importantly, these data suggest that using
measures other than dysplasticity and tincture indicate that mouse liver tumors induced by TCE
are heterogeneous in phenotype. The descriptions of tumors in mice reported by the NTP and
Maltoni et al studies are also consistent with phenotypic heterogeneity as well as consistency
with spontaneous tumor morphology.
2.3. Mode of Action: Relative Contribution of TCE Metabolites
Several metabolites of TCE have also been shown to induce liver cancer in rodents with
DCA and TCA having been the focus of study as potential active agent(s) of TCE liver toxicity
and/or carcinogenesis and both able to induce peroxisome proliferation (Caldwell and Keshava
2006). A variety of DCA effects from exposure have been noted that are consistent with
conditions that increase risk of liver cancer (e.g. effects on the cytosolic enzyme glutathione
02/12/09
E-92
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(GST)-S-transferase-zeta, diabetes, and glycogen storage disease), with the pathological changes
induced by DCA on whole liver consistent with changes observed in preneoplastic foci from a
variety of agents (Caldwell and Keshava, 2006). Chloral hydrate is one of the first metabolites
from oxidative metabolism of TCE with a large fraction of TCE metabolism appearing to go
through chloral hydrate and then subsequent metabolism to TCA and trichloroethanol (Chiu et
al., 2006). Similarities in toxicity may indicate that common downstream metabolites may be
toxicologically important, and differences may indicate the importance of other metabolic
pathways.
Although both induce liver tumors, DCA and TCA have distinctly different actions
(Keshava and Caldwell, 2006) and apparently differ in tumor phenotype (see discussion above in
Section 2.2.8) and many studies have been conducted to try to elucidate the nature of those
differences (Caldwell et al 2007b). Limitations of all of the available chronic studies of TCA
and most of the studies of DCA include less than lifetime exposures, varying and small numbers
of animals examined, and few exposure concentrations that were relatively high.
2.3.1 Acute studies of DCA/TCA
The studies in this section focus on studies of DCA and TCA that examine, to the extent
possible, similar endpoints using similar experimental designs as those of TCE examined above
and that give insight into proposed MO As for all three. Of note for any experiment involving
TCA, is whether exposure solutions were neutralized. Unbuffered TCA is commonly used as a
reagent to precipitate proteins so that any result from studies using unbuffered TCA could
potentially be confounded by the effects on pH.
2.3.1.1. Sanchez and Bull 1990
In this report TCA and DCA were administered to male B6C3F1 mice (9 weeks of age)
and male and female Swiss-Webster mice (9 weeks of age) for up to 14 days. At 2, 4, or 14 days,
mice were injected with tritiated thymidine. Experiments were replicated at least once but
results were pooled so that variation between experiments could not be determined. B6C3F1
male mice were given DCA or TCA at 0, 0.3 g/L, 1.0 g/L, 2.0 g/L in drinking water (n= 4 for
each group for 2 and 5 days, but n= 15 for control and n= 12 for treatment groups at day 14).
Swiss-Webster mice (N= 4) at were exposed to DCA only on Day 14 at 0, 1.0 or 2.0 g/L. Mice
were injected with tritiated thymidine 2 hours prior to sacrifice. The pH of the drinking water
was adjusted to 6.8 - 7.2 with sodium hydroxide. Concentrations of TCA and DCA were
reported to be stable for a minimum of 3 weeks. Hepatocyte diameters were reported to be
determined by randomly selecting 5 different high power fields (400X) in five different sections
per animals (total of 25 fields/ animal with "cells in and around areas of necrosis, close to the
edges of the section, or displaying mitotic figures were not included in the cell diameter
measurements." PAS staining was reported to be done for glycogen and lipofuscin determined
by autofluorescence. Tritiated thymidine was reported to be given to the animals 2 hours prior to
02/12/09
E-93
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
sacrifice. In 2 of 3 replications of the 14-day experiment, a portion of the liver was reported to
be set aside for DNA extraction with the remaining group examined autoradiographically for
tritiated thymidine incorporation into individual hepatocytes. Autoradiographs were also
reported to be examined in the highest dose of either DCA or TCA for the 2- and 5-day treatment
groups. Autoradiographs were reported to be analyzed in randomly selected fields (5 sections
per animal in 10 different fields) for a total of 50 fields/animal and reported as percentage of
cells in the fields which were labeled. There was no indication by the authors that they
characterized differing zones of the liver for preferential labeling. DNA thymidine incorporation
results were not examined in the same animals as those for individual hepatoctye incorporation
and also not examined at 2- or 5- day time periods. The only analyses reported for the Swiss-
Webster mice were of hepatic weight change and histopathology. Variations in results were
reported as standard error of the mean.
Liver weights were reported but not body weights so the relationship of liver/body weight
ratio could not be determined for the B6C3F1 mice. For liver weight, the numbers of animals
examined varied greatly between and within treatment groups. The number of control animals
examined were reported to be n = 4 on day 2, n=8 on day 5 and n=15 on day 14. There was also
a large variation between control groups in regard to liver weight. Control liver weights for day
2 were reported to be 1.3 ± 0.1, day 5 to be 1.5 ± 0.05 and for day 14 to be 1.3 ± 0.04 grams.
Liver weights in Day 5 control animals were much greater than those for day 2 and day 14
animals and thus the means varied by as much as 15%. For DCA, there was no reported change
in liver weights compared to controls values at any exposure level of DCA after 2 days of
exposure. After 5 days of exposure there was no difference in liver weight between controls and
0.3 g/1 exposed animals. However the animals exposed at 1.0 or 2.0 g/1 DCA had identical
increases in liver weight of 1.7 ± 0.13 and 1.7 ± 0.8 g, respectively. Due to the low power of the
experiment, only the 2.0 g/1 DCA result was identified by the authors as significantly different
from the control value. For TCA there was a slight decrease reported between control values and
the 0.3 g/1 treatment group (1.2 ± 0. lg vs. 1.3 ± 0.1) but the 1.0 g/1 and 2.0 g/1 treatment groups
had similar slight increases over control (for 1.0 g/1 liver weight was 1.5 ± 0.1 and for 2.0 g/1
liver weight was 1.4 ± 0.1 g). The same pattern was apparent for the 5-day treatment groups for
TCA as for the 2-day treatment groups.
For 14 days exposure periods the number of animals studied was increased tol2 for the
TCA and DCA treatment groups. After 14 days of DCA treatment, there was a reported dose-
related increase in liver weight that was statistically significant at the two highest doses (i.e., at
0.3 g/1 DCA liver weight was 1.4 ± 0.04, at 1.0 g/1 DCA liver weight was 1.7 ± 0.07 g, and at 2.0
g/1 DCA liver weight was 2.1 ± 0.08 g). This was 1.08-fold, 1.31-fold, and 1.62-fold of controls,
respectively. After 14 days of TCA exposure there was a dose-related increase in liver weight
that the authors reported to be statistically significant at all exposure levels (i.e., at 0.3 g/1 liver
weight was 1.5 ± 0.06, at 1.0 g/1 liver weight was 1.6 ± 0.07 g, and at 2.0 g/1 liver weight was 1.8
± 0.10 g). This represents 1.15-fold, 1.23-fold, and 1.38-fold of control. The authors note that at
14 days that DCA-associated increases in hepatic liver weight were greater than that of TCA.
What is apparent from this data is that while the magnitude of difference between the exposures
02/12/09 E-94
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
was ~ 6.7-fold between the lowest and highest dose, the differences between TCA exposure
groups for change in liver weight was -2.5. For DCA the slope of the dose-response curve for
liver weight increases appeared to be closer to the magnitude of difference in exposure
concentrations between the groups (i.e. a difference of 7.7-fold between the highest and lowest
dose for liver weight induction). Given that the control animal weights varied as much as 15%,
the small number of animals examined, and that body weights were also not reported, there are
limitations for making quantitative comparisons between TCA and DCA treatments. However,
after 14 days of treatment it is apparent that there was a dose-related increase in liver weight
after either DCA or TCA exposure at these exposure levels. For male and female Swiss-Webster
mice 1 g/1 and 2 g/1 DCA treatment (n= 4) was reported to also induce an increase in %
liver/body weight that was similar to the magnitude of exposure difference (see below).
Grossly, livers of B6C3F1 mice treated with DCA for 1 or 2g/L were reported to have
"pale streaks running on the surface" and occasionally, discrete, white, round areas were also
observed on the surface of these livers. Such areas were not observed in TCA-treated or control
B6C3F1 mice." Swiss-Webster mice were reported to have "dose-related increases in hepatic
weight and hepatic/body weight ratios were observed. DCA-associated increases in relative
hepatic weights in both sexes were comparable to those in B6C3F1 mice. Pale streaks on the
surface of the liver were not observed in Swiss-Webster mice. Again there was no significant
effect on total body or renal weights (data not shown)." The authors report liver weights for the
Swiss-Webster male mice (n= 4 for each group) to be 2.1 ± 0.1 g for controls, 2.1 ± 0.1 g for 1.0
g/1 DCA and 2.4 ± 0.2 g for 2.0 g/1 DCA 14-day treatment groups. The % liver/body weights for
these same groups were reported to be 6.4 ± 0.4 %, 6.9 ± 0.2 %, and 8.1 ± 0.3 %, respectively.
For female Swiss-Webster mice (n= 4 for each group) the liver weights were reported to be 1.1 ±
0.1 g for controls, 1.5 ± 0.1 g for 1.0 g/1 DCA and 1.7 ± 0.2 g for 2.0 g/1 DCA 14-day treatment
groups. The % liver/body weights for these same groups of Swiss mice were reported to be 4.8 ±
0.2 %, 6.0 ± 0.2 %, and 6.8 ± 0.4 %, respectively. Thus, while there was no significant
difference in "liver weight" between the control and the 1.0 g/1 DCA treatment group for male or
female Swiss-Webster mice, there was a statistically significant difference in liver/body weight
ratio reported by the authors. These data, illustrate the importance of reporting both measures
and the limitations of using small numbers of animals (n= 4 for the Swiss Webster vs. n= 12-14
for B6C3F1 14-days experiments). Relative liver weights were reported by the authors for male
B6C3F1 mice only for the 14-day groups, as a function of calculated mean water consumption,
as pooled data from the three experiments, and as a figure that was not comparable to the data
reported for Swiss-Webster mice. The liver weight data indicate that male mice of the same age
appeared to differ in liver weight between the two strains without treatment (i.e., male B6C3F1
mice had control liver weights at 14 days of 1.3 ± 0.04 g for 15 mice, while Swiss-Webster mice
had control values of 2.1 ±0.1 for 4 mice). While the authors report that results were
"comparable" between the B6C3F1 mice in regard to DCA-induced changes in liver weight, the
increase in % liver/body weight ratios were 1.27-fold of control for Swiss-Webster male mice
(n= 4) and 1.42-fold fold of control for female while the increase in liver weight for B6C3F1
male mice (n= 12-14) was 1.62-fold of controls after 14 days of exposure to 2 g/1 DCA.
02/12/09
E-95
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The concentration of DNA in the liver was reported as mg hepatic DNA/g of liver. This
measurement can be associated with hepatocellular hypertrophy when decreased, or increased
cellularity (of any cell type), increased DNA synthesis, and/or increased hepatocellular ploidy in
the liver when increased. The number of animals examined for this parameter varied. For
control animals there were 4 animals reported to be examined at 2 days, 8 animals examined at 5
days, and at 14 days 8 animals were examined. The mean DNA content in control livers were
not reported to vary greatly however and the variation between animals was relatively low in the
5- and 14-day control groups (i.e., 1.67 ± 0.27 mg DNA/g, 1.70 ± 0.05 mgDNA/g, and 1.69 mg
DNA/g, for 2 -, 5-, or 14-day control animals, respectively). For treatment groups the number of
animals reported to be examined appeared to be the same as the control animals. For DCA
treatment there did not appear to be a dose-response in hepatic DNA content with the 1 g/1
exposure level having the same reported value as control but the 0.3 g/1 and 2.0 g/1 values
reported to be lower (mean values of 1.49 and 1.32 mg DNA/g, respectively). After 5 days of
exposure, all treatment groups were reported to have a lower DNA content that the control value
(i.e., 1.44 ± 0.06 mg DNA/g, 1.47 ± mg DNA/g, and 1.30 ± 0.14 mg DNA/g, for 0.3 g/1, 1.0 g/1,
and 2.0 g/1 exposure levels of DCA, respectively). After 14 days of exposure, there was a
reported increase in hepatic DNA at the 0.3 g/1 exposure level but significant decreases at the 1.0
g/1 and 2.0 g/1 exposure levels (i.e., 1.94 ± 0.20 mg DNA/g, 1.44 ± 0.14 mg DNA/g, and 1.19 ±
0.16 mg DNA/g for the 0.3 g/1, 1.0 g/1, and 2.0 g/1 exposure levels of DCA, respectively).
Changes in DNA concentration in the liver were not correlated with the pattern of liver weight
increases after DCA treatment. For example, while there was a clear dose-related increase in
liver weight after 14 days of DCA treatment, the 0.3 g/1 DCA exposed group was reported to
have a higher rather than lower level of hepatic DNA than controls. After 2 or 5 days of DCA
treatment, liver weights were reported to be the same between the 1.0 g/1 and 2.0 g/1 treatment
groups but hepatic DNA was reported to be decreased.
For TCA there appeared to be a dose-related decrease in reported hepatic DNA after 2
days of treatment (i.e., 1.63 ± 0.07 mg DNA/g, 1.53 ± 0.08 mg DNA/g, and 1. 43 ± 0.04 mg
DNA/g for the 0.3 g/1, 1.0 g/1, and 2.0 g/1 exposure levels of TCA, respectively). After 5 days of
TCA exposure there was a reported decrease in hepatic DNA for all treatment groups that was
similar at the 1.0 g/1 and 2.0 g/1 exposure groups (i.e., 1.45 ± 0.17 mg DNA/g, 1.29 ± 0.18 mg
DNA/g, and 1. 26 ± 0.22 mg DNA/g for the 0.3 g/1, 1.0 g/1, and 2.0 g/1 exposure levels of TCA,
respectively). After 14 days of TCA treatment, there was a reported decrease in all treatment
groups in hepatic DNA content that did not appear to be dose-related (i.e., 1.31 ± 0.17 mg
DNA/g, 1.21 ± 0.17 mg DNA/g, and 1. 33 ± 0.18 mg DNA/g for the 0.3 g/1, 1.0 g/1, and 2.0 g/1
exposure levels of TCA, respectively). Thus, similar to the results reported for DCA, the
patterns of liver weight gain did not match those of hepatic DNA decrease for TCA treated
animals. For example, although there appeared to be a dose-related increase in liver weight gain
after 14 days of TCA exposure, there was a treatment but not dose-related decrease in hepatic
DNA content.
In regard to the ability to detect changes, the low number of animals examined after 2
days of exposure (n = 4) limited the ability to detect a significant change in liver weight and
02/12/09
E-96
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
hepatic DNA concentration. For hepatic DNA determinations, the larger number of animals
examined at 5 and 14 day time points and the similarity of values with relatively smaller standard
error of the mean reported in the control animals made quantitative differences in this parameter
easier to determine. However, animals varied in their response to treatment and this variability
exceeded that of the control groups. For DCA results reported at 14 days and those for TCA
reported at 5 and 14 days, the standard errors for treated animals showed a much greater
variability than those of the control animals (range of 0.04 - 0.05 mg DNA/g for control groups,
but ranges of 0.17 to 0.22 mg DNA/g for TCA at 5 days and 0.14 to 0.20 mg DNA/g for DCA or
TCA at 14 days). The authors stated that "the increases in hepatic weights were generally
accompanied by decreases in the concentration of DNA. However, the only clear changes were
in animals treated with DCA for 5 or 14 days where the ANOVAs were clearly significant
(P<0.020 and 0.005, respectively). While changes of similar magnitude were observed in other
groups, the much greater variation observed in the treated groups resulted in not significant
differences by ANOVA ( p = 0.41, 0.66. 0.26, 0.15 for DCA - 2 days, and TCA for 2,5, and 14
days, respectively)."
The size of hepatocytes is heterogeneous and correlated with its ploidy, zone, and age of
the animal (see Section 1.1 above). The authors do not indicate if there was predominance in
zone or ploidy for hepatocytes included in their analysis of average hepatocyte diameter in the
random selection of 25 fields per animal (n=3 to 7 animals). There appeared to be a dose-related
increase in cell diameter associated with DCA exposure and a treatment but not dose-related
increase with TCA treatment after 14 days of treatment. For control B6C3F1 male mice (n=7)
the hepatocyte diameter was reported to be 20.6 ± 0.4 microns. For mice exposed to DCA
hepatocyte diameter was reported to be 22.2 ± 0.2, 25.2 ± 0.6, and 26.0 ±1.0 microns for 0.3 g/1,
1.0 g/1, and 2.0 g/1 treated mice (n = 4 for each group), respectively. For mice exposed to TCA
hepatocyte diameter was reported to be 22.2 ± 0.2, 22.4 ± 0.6, and 23.2 ± 0.4 microns for 0.3 g/1,
1.0 g/1, and 2.0 g/1 treated mice (n = 4 for the 0.3 g/1 and 1.0 g/1 groups and n= 3 for the 2.0 g/1
group), respectively. The small number of animals examined limited the power of the
experiment to determine statistically significant differences with the authors reporting that only
the 1.0 g/1 DCA and 2.0 g/1 DCA and TCA treated groups statistically significant from control
values. The dose-related increases in reported cell diameter were consistent with the dose-
related increases in liver weight reported for DCA after 14 days of exposure. However, the
pattern for hepatic DNA content did not. For TCA, the dose-related increases in cell diameter
were also consistent with the dose-related increases in liver weight after 14 days of exposure.
Similar to DCA results, the changes in hepatic DNA content did not correlate with changes in
cell size. In regard to the magnitude of increases over control values, the 68% vs. 38% increase
in liver weight for DCA vs. TCA at 2.0 g/1, was less than the 26% and 13% increases in cell
diameter for the same groups, respectively. Therefore, for both DCA and TCA exposure there
appeared to be dose-related hepatomegaly and increased cell size after 14-days of exposure.
The authors reported PAS staining for glycogen content as an attempt to examine the
nature of increased cell size by DCA and TCA. However, they did not present any quantitative
data and only provided a brief discussion. The authors reported that "hepatic sections of DCA-
02/12/09
E-97
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
treated B6C3F1 mice (1 and 2 g/1) contained very large amounts of perilobular PAS-positive
material within hepatocytes. PAS stained hepatic sections from animals receiving the highest
concentration of TCA displayed a much less intense staining that was confined to periportal
areas. Amylase digesting confirmed the majority of the PAS-positive material to by glycogen.
Thus, increased hepatocellular size in groups receiving DCA appears to be related to increased
glycogen deposition. Similar increases in glycogen deposition were observed in Swiss-Webster
mice." There is no way to discern whether DCA-induced glycogen deposition was dose-related
and therefore correlated with increased liver weight and cell diameter. While the authors suggest
that Swiss-Webster mice displayed "similar increased in glycogen deposition" the authors did
not report a similar increase in liver weight gain after DCA exposure at 14 days (1.27-fold of
control % liver/body weight ratio in Swiss male mice and 1.42-fold in female Swiss-Webster
mice vs. 1.62-fold of control in B6C3F1 mice after 14 days of exposure to 2 g/1 DCA). Thus, the
contribution of glycogen deposition to DCA-induced hepatomegaly and the nature of increased
cell size induced by acute TCA exposure cannot be determined by this study. However, this
study does show that DCA and TCA differ in respect to their effects on glycogen deposition after
short-term exposure and the data suggest that.
The authors report that "localized areas of coagulative necrosis were observed
histologically in both B6C3F1 and Swiss-Webster mice treated with DCA at concentrations of 1
and 2 g/1 for 14 days. The necrotic areas corresponded to the pale streaked areas seen grossly.
These areas varied in size, shape and location within sections and occupied up to several mm2.
An acute inflammatory response characterized by thin rims of neutrophils was associated with
the necrosis, along with multiple mitotic figures. No such areas of necrosis were observed in
animals treated at lower concentrations of DCA, or in animals receiving the chemical for 2 or 5
days. Mice treated with 2 g/1 TCA for 14 days have some necrotic areas, but at such low
frequency that it was not possible to determine if it was treatment-related (2 lesions in a total of
20 sections examined). No necrosis was observed in animals treated at the lower concentrations
of TCA or at earlier time points." Again there were no quantitative estimates given of the size
of necrotic areas, variation between animals, variation between strain, or dose-response of
necrosis reported for DCA exposure by the authors. The lack of necrosis after 2 and 5 days of
exposure at all treatment levels and at the lower exposure level at 14 days of exposure is not
correlated with the increases in liver weight reported for these treatment groups.
Autoradiographs of randomly chosen high powered fields (400X) (50 fields/animal) were
reported as the percentage of cells in the fields that were labeled. There was significant variation
in the number of animals examined and in the reported mean % of labeled cells between control
groups. The number of control animals was not given for the 2-day group but for the 5-day and
14 day groups were reported to be n = 4 and n = 11, respectively. The mean % of labeling in
control animals was reported at 0.11 ± 0.03, 0.12 ± 0.04, and 0.46 ± 0.07% of hepatocytes for 2-
day, 5-day, and 14-day control groups, respectively. Only the 2.0 g/1 exposures of DCA and
TCA were examined at all 3 times of exposure while all groups were examined at 14 days.
However, the number of animals examined in all treatment groups appeared to be only 4 animals
in each group. There was not an increase over controls reported in the 2.0 g/1 DCA or TCA 2-
02/12/09
E-98
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and 5-day exposure groups in hepatocyte labeling with tritiated thymidine. After 14 days of
exposure, there was a statistically significant but very small dose-related increase over the
control value after DC A exposure (i.e., 0.46 ± 0.07, 0.64 ±0.15, 0.75 ± 0.22, and 0.94 ± 0.05 %
labeling of hepatocytes in control, 0.3 g/1, 1.0 g/1 and 2.0 g/1 DCA treatment groups,
respectively). For TCA, there was no change in hepatocyte labeling except for a 50% decrease
from control values at after 14 days of exposure to 2.0 g/1 TCA (i.e., 0.46 ± 0.07, 0.50 ± 0.14,
0.52 ± 0.26, and 0.26 ± 0.14 % labeling of hepatocytes in control, 0.3 g/1, 1.0 g/1 and 2.0 g/1 TCA
treatment groups, respectively). The authors report that "labeled cells were localized around
necrotic areas in these [sic DCA treated] groups. Since counts were made randomly, the local
increased in DCA-treated animals at concentrations of 1 and 2 g/1 are in fact much higher than
indicated by the data. Labeling indices in these areas of proliferation were as high as 30%.
Labeled hepatocytes in TCA-treated and the control animals were distributed uniformly
throughout the sections. There was an apparent decrease in the percentage of labeled cells in the
group of animals treated with the highest dose of TCA. This is because no labeled cells were
found in any of the fields examined for one animal." The data for control mice in this
experiment is consistent with others showing that the liver is quiescent in regard to
hepatocellular proliferation with few cells undergoing mitosis (see Section 1.1). For up to 14
days of exposure with either DCA or TCA, there is little increase in hepatocellular proliferation
except in instances and in close proximity to areas of proliferation. The increases in liver weight
reported for this study were not correlated with and cannot be a result of hepatocellular
proliferation as only a very small population of hepatocytes is undergoing DNA synthesis. For
TCA, there was no increase in DNA synthesis in hepatocytes, even at the highest dose, as shown
by autoradiographic data of tritiated thymidine incorporation in random fields.
Whole liver sections were examined for tritiated thymidine incorporation from DNA
extracts. The number of animals examined varied (i.e., n = 4 for the 2-day exposure groups and
n = 8 for 5- and 14-day exposure groups) but the number of control animals examined were the
same as the treated groups for this analysis. The levels of tritiated thymidine incorporation in
hepatic DNA (dpm/mg DNA expressed as mean x 103 ± SE of N animals) were reported to be
similar across control groups (i.e., 56 ± 11, 56 ± 6, and 56 ± 7 dpm/mg DNA, for 2-day, 5-day,
and 14-day treatment groups, respectively). After two days of DCA exposure, there appeared to
be a slight treatment-related but not dose-related increase in reported tritiated thymidine
incorporation into hepatic DNA (i.e., 72 ± 23, 80 ± 6, and 68 ± 7 dpm/mg DNA for 0.3, 1.0, or
2.0 g/1 DCA, respectively). After 5 days of DCA exposure, there appeared to be a dose-related
increase in reported tritiated thymidine incorporation into hepatic DNA (i.e., 68 ± 18, 110 ± 20,
and 130 ± 7 dpm/mg DNA for 0.3, 1.0, or 2.0 g/1 DCA, respectively). However, after 14 days of
DCA exposure, levels of tritiated thymidine incorporation were less than those reported at 5 days
and the level for the 0.3 g/1 exposure group was less than the control value (i.e., 33 ± 11, 77 ± 9,
and 81 ± 12 dpm/mg DNA for 0.3, 1.0, or 2.0 g/1 DCA, respectively). After two days of TCA
exposure there did not appear to be a treatment-related increase in tritiated thymidine
incorporation into hepatic DNA (i.e., 82 ± 16, 52 ± 7, and 54 ± 7 dpm/mg DNA for 0.3, 1.0, or
2.0 g/1 TCA, respectively). Similar to the reported results for DCA, after 5 days of TCA
exposure there appeared to be a dose-related increase in reported tritiated thymidine
02/12/09
E-99
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
incorporation into hepatic DNA (i.e., 79 ± 23, 86 ± 17, and 158 ± 33 dpm/mg DNA for 0.3, 1.0,
or 2.0 g/1 TCA, respectively). After 14 days of TCA exposure there were treatment related
increases but not a dose-related increase in reported tritiated thymidine incorporation into hepatic
DNA (i.e., 71 ± 10, 73 ± 14, and 103 ± 14 dpm/mg DNA for 0.3, 1.0, or 2.0 g/1 TCA,
respectively). It would appear that for both TCA and DCA the increase in tritiated thymidine
incorporation into hepatic DNA was dose related and peaked after 5 days of exposure. The
authors report that the decrease in incorporation into hepatic DNA observed after 14 days of
DCA treatment at 0.3 g/1 to be statistically significant as well as the increases after 5 and 14 days
of TCA exposure at the 2.0 g/1 level. The small numbers of animals examined, the varying
number of animals examined, and the degree of variation in treatment-related effects limits the
statistical power of this experiment to detect quantitative changes.
Given the limitations of this experiment, determination of an accurate measure of the
quantitative differences in tritiated thymidine incorporation into whole liver DNA or that
observed in hepatocytes are hard to determine. In general the results for tritiated thymidine
incorporation into hepatic DNA were consistent with those for tritiated thymidine incorporation
into hepatocytes in that they show that there were at most a small population of hepatocytes
undergoing DNA synthesis after up to 14 days of exposure at relative high levels of exposure to
DCA and TCA (i.e., the largest percentage of hepatocytes undergoing DNA synthesis for any
treatment group was less than 1% of hepatocytes). The highest increases over control levels for
hepatic DNA incorporation for the whole liver were reported at the highest exposure level of
TCA treatment after 5 days of treatment (3-fold of control) and after 14 days of TCA treatment
(2-fold of control). Although the authors report small areas of focal necrosis with concurrent
localized increases in hepatocyte proliferation in DCA treated animals exposed tol.O g/1 and 2.0
g/1 DCA, the levels of whole liver tritiated thymidine incorporation were only slightly elevated
over controls at these concentrations, and were decreased at the 0.3 g/1 exposure concentration
for which no focal necrosis was reported. The whole liver DNA incorporation of tritiated
thymidine was not consistent with the pattern of tritiated thymidine incorporation observed in
individual hepatocytes. The authors state that "at present, the mechanisms for increased tritiated
thymidine uptake in the absence of increased rates of cell replication with increasing doses of
TCA cannot be determined." The authors do not discuss the possibility that the difference in
hepatocyte labeling and whole liver DNA tritiated thymidine incorporation could have been due
to the labeling representing increased polyploidization rather than cell proliferation, as well as
increased numbers of proliferating nonparenchymal and inflammatory cells. The increased cell
size due from TCA exposure without concurrent increased glycogen deposition could have been
indicative of increased polyploidization. Finally, although both TCA- and DCA-induced
increases in liver weight were generally consistent with cell size increases, they were not
correlated with patterns of change in hepatic DNA content, incorporation of tritiated thymidine
in DNA extracts from whole liver, or incorporation of tritiated thymidine in hepatocytes. In
regard to cell size, although increased glycogen deposition with DCA exposure was noted by the
authors of this study, lack of quantitative analyses of that accumulation precludes comparison
with DCA-induced liver weight gain.
02/12/09
E-100
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.3.1.2. Nelson et al. 1989
Nelson and Bull (1988) administered TCE (0, 3.9, 11.4, 22.9, and 30.4 mmol/kg) in
tween 80 via gavage to male Sprague Dawley rats and male B6C3F1 mice, sacrificed them fours
hours after treatment (n= 4-7), and measured the rate of DNA unwinding under alkaline
conditions. They assumed that this assay represented increases in single-strand breaks. For rats
there was little change from controls up to 11.4 mmol/kg (1.5. g/kg TCE) but a significantly
increased rate of unwinding at 22.9 and 30.4 mmol/kg TCE (~ 2-fold greater at 30.4 mmol). For
mice there was a significantly increased level of DNA unwinding at 11.4 and 22.9 mmol.
Concentrations above 22.9 mmol/kg were reported to be lethal to the mice. In this same study,
TCE metabolites were administered in unbuffered solution using the same assay. DCA was
reported to be most potent in this assay with TCA being the lowest, while chloral hydrate (CH)
closely approximated the dose-response curve of TCE in the rat. In the mouse the most potent
metabolite in the assay was reported to be TCA followed by DCA with CH considerably less
potent.
The focus of the Nelson et al. (1989) study was to examine whether reported single strand
breaks in hepatic DNA induced by DCA and TCA (Nelson and Bull, 1988) were secondary to
peroxisome proliferation also reported to be induced by both. Male B6C3F1 mice (25-30 g but
no age reported) were given DCA (10 mg/kg or 500 mg/kg) or TCA (500 mg/kg) via gavage in
1% aqueous tween 80 with no pH adjustment. The animals were reported to be sacrificed 1, 2, 4,
or 8 hours after administration and livers examined for single strand breaks as a whole liver
homogenate. In a separate experiment (experiment #2) treatment was parallel to the first (500
mg/kg treatment of DCA or TCA) but levels of PCO activity were measured as an indication of
peroxisome proliferation and expressed as [j,mol/min/g liver. In a separate experiment
(experiment #3) mice were administered 500 mg/kg DCA or TCA for 10 days with Clofibrate
administered at a dose of 250 mg/kg as a positive control. 24 hours after the last dose, animals
were killed and liver examined by light microscopy and PCO activity. Finally, in an experiment
parallel in design to experiment #3, single strand breaks were measured in total hepatic DNA
after 500 mg/kg exposure to TCA (experiment #4). Electron microscopy was performed on 2
animals/group for vehicle, DCA or TCA treatment, with 6 randomly chosen micrographic fields
utilized for peroxisome profiles. These micrographs were analyzed without identification as to
what area of the liver lobules they were being taken from. Hence there is a question as to
whether the areas which are known to be peroxisome rich were assayed of not.
The data from all control groups were reported as pooled data in figures but statistical
comparisons were made between concurrent control and treated groups. The results for DNA
single strand breaks were reported for "13 control animals" and each experimental time point "as
at least 6 animals. DNA strand breaks were reported to be significantly increased over
concurrent control by a single exposure to 10 or 500 mg/kg DCA or 500 mg/kg TCA for 1, 2 or 4
hours after administration but not at 8 or 24 hours. There did not appear to be a difference in the
magnitude of response between the 3 treatments (the fraction of unwound DNA was ~ 2.5 times
that of control). PCO activity was reported to be not increased over control within 24 hours of
02/12/09
E-101
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
either DCA or TCA treatment, (n = 6 animals per group). The fraction of alkaline unwinding
rates as an indicator of single strand breaks were reported to not be significantly different from
controls and TCA-treated animals after 10 days of exposure (n= 5).
Relative to controls, body weights were reported to not be affected by exposures to DCA
or TCA for 10 days at 500 mg/kg (with data was not shown.) (n=6 per group). However, both
DCA and TCA were reported to significantly increase liver weight and liver/body weight ratios
(i.e., liver weights were 1.3 ± 0.05 g, 2.1 ± 0.10 g, and 1.7 ± 0.09 g for control, 500 mg/kg DCA
and 500 mg/kg TCA treatment groups, respectively while % liver/body weights were 4.9 ± 0.14
%, 7.5 ±0.18 %, and 5.7 ± 0.14 % for control, 500 mg/kg DCA and 500 mg/kg TCA treatment
groups, respectively). PCO activity (|imol/min/g liver) was reported to be significantly
increased by DCA (500 mg/kg), TCA (500 mg/kg), and Clofibrate (250 mg/kg) treatment (i.e.,
levels of oxidation were 0.63 ± 0.07, 1.03 ± 0.09, 1.70 ± 0.08, and 3.26 ± 0.05 for control, 500
mg/kg DCA, 500 mg/kg TCA and 250 mg/kg Clofibrate treatment groups, respectively). Thus,
the increases were ~ 1.63-fold, 2.7 -fold, and 5-fold of control for DCA, TCA and Clofibrate
treatments. Results from randomly selected electron photomicrographs from 2 animals (6 per
animal) were reported for DCA and TCA treatment and to show an increase in peroxisomes per
unit area that was reported to be statistically significant (i.e., 9.8 ± 1.2, 25.4 ± 2.9, and 23.6 ±1.8
for control, 500 mg/kg DCA and 500 mg/kg TCA, respectively). The 2.5-fold and 2.4-fold of
control values for DCA and TCA gave a different pattern than that of PCO activity. The small
number of animals examined limited the power of the experiment to quantitatively determine the
magnitude of peroxisome proliferation via electron microscopy. The enzyme analyses suggested
that both DCA and TCA were weaker inducers of peroxisome proliferation that Clofibrate.
The authors report that there was no evidence of gross hepatotoxicity in vehicle or TCA-
treated mice. Light microscopic sections from mice exposed to TCA or DCA for 10 days were
stained with H&E and PAS for glycogen. For TCA treatment, PAS staining "produced
approximately the same intensity of staining and amylase digesting revealed that the vast
majority of PAS-positive staining was glycogen." Hepatocytes were reported to be "slightly
larger in TCA-treated mice than hepatocytes from control animals throughout the liver section
with the architecture and tissue pattern of the liver intact. The histopathology after DCA
treatment was reported to be "markedly different than that observed with either vehicle or TCA
treatments" with the "most pronounced change in the size of hepatocytes." DCA was reported to
"produce marked cellular hypertrophy uniformly throughout the liver. The hepatocytes were
approximately 1.4 times larger in diameter than control liver cells. This hypertrophy was
accompanied by an increase in PAS staining; indicating greater glycogen deposition than in
TCA-treated and control liver tissue. Multiple white streaks were grossly visible on the surface
of the liver of DCA-treated mice. The white areas corresponded with subcapsular foci of
coagulative necrosis. These localized necrotic areas were not encapsulated and varied in size.
The largest necrotic foci occupied the area of a single lobule. These necrotic areas showed a
change in staining characteristics. Often this change consisted of increased eosinophilia. A
slight inflammatory response, characterized by neutrophil infiltration, was present. These
changed were evident in all DCA-treated mice."
02/12/09 E-102
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The results from this experiment cannot inform as to dose-response relationships for the
parameters tested with the exception of DNA single strand breaks where 2 concentrations of
DC A were examined (10 mg/kg and 500 mg/kg). For this parameter the 10 mg/kg exposure of
DCA was as effective as the 500 mg/kg dose where toxicity was observed. This effect on DNA
was also observed before evidence of induction of peroxisome proliferation. The authors did not
examine Clofibrate for effects on DNA so whether it too, would have produced this effect is
unclear. The results from this study are consistent with those of Sanchez and Bull (1990) for
induction of hepatomegaly by DCA and TCA, the lack of hepatotoxicity at this dose by TCA,
and the difference in glycogen deposition between DCA and TCA.
2.3.1.3. Styles et al. 1991
In this report a similar paradigm is used as Nelson et al (1989) for the determination of
repeating that work on single strand breakage and to study DNA synthesis and peroxisome
proliferation. In regard to the findings of single strand breaks, Styles et al (1991) reported for a
similar paradigm of 500 mg/kg neutralized TCA administered to male B6C3F1 mice (7-8 weeks
of age) examined at 1, 4, 8 and 24 hours after dosing, reported no increased unwinding of DNA 1
or 24 hours after TCA administration. In a separate experiment tritiated thymidine was
administered to mice 1 hour before sacrifice at 24, 36, 48, 72, and 96 hours after the first dose of
500 mg/kg TCA for 3 days via gavage (n= 5 animals per group).
The hepatic DNA uptake of tritiated thymidine was reported to be similar to control
levels up to 36 hours after the first dose and then to increase to a level ~ 6 fold greater than
controls by 72 hours after the first dose of TCA. By 96 hours the level of tritiated thymidine
incorporation had fallen to ~ 4 -fold greater than controls. The variation, reported by S.D. was
very large in treated animals (e.g., S.D. was equal to ~ ± 1.3 fold of control for 48 hour time
point). Individual hepatocytes were examined with the number of labeled hepatocytes /1000
cells reported for each animal. The control level was reported to be ~ 1 with a S.D. of similar
magnitude. The number of labeled hepatocytes was reported to decrease between 24 and 36
hours and then to rise slowly back to control levels at 48 hour and then to be significantly
increased 72 hours after the first dose of TCA (~ 9 cells/1000 with a S.D of 3.5 ) and then to
decrease to a level of ~ 5 cells/1000. Thus, it appears that increases in hepatic DNA tritiated
thymidine uptake preceded those of increased labeled hepatocytes and did not capture the
decrease in hepatocyte labeling at 36 hours. By either measure the population of cells
undergoing DNA synthesis was small with the peak level being less than 1% of the hepatocyte
population. The authors go on to report the zonal distribution of mean number of hepatocytes
incorporating tritiated thymidine but no variations between animals were reported. The decrease
in hepatocyte labeling at 36 hours was apparent at all zones. By 48 hours there appeared to be
slightly more perioportal than midzonal cells undergoing DNA synthesis with centrilobular cells
still below control levels. By 72 hours all zones of the liver were reported to have a similar
number of labeled cells. By 96 hours the midzonal and centrilobular regions have returned
almost to control levels while the periportal areas were still elevated. These results are consistent
with all hepatocytes showing a decrease in DNA synthesis by 36 hours and then a wave of DNA
02/12/09
E-103
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
synthesis occurring starting at the periportal zone and progressing through to the pericentral zone
until 72 hours and then the midzonal and pericentral hepatocytes completing their DNA
synthesis activity. Peroxisome proliferation was assessed via electron photomicrographs taken
in mice (4 controls and 4 treated animals) given 10 daily doses of 500 mg/kg TCA and killed 14
hours after the last dose. No details were given by the authors as to methodology for peroxisome
volume estimate (e.g., how many photos per animals were examined and whether they were
randomly chosen). The mean % cell volume occupied by peroxisome was reported to be 2.1 ±
0.386 and 3.9 ± 0.551 % for control and 500 mg/kg TCA, respectively. Given there were no
time points examined before 10 days for peroxisome proliferation, correlations with DNA
synthesis activity induced by TCA cannot be made from this experiment. However, it is clear
from this study that a wave of DNA synthesis occurs throughout the liver after treatment of TCA
at this exposure concentration and that it has peaked by 72 hours even with continuous exposure
to 96 hours. Whether the DNA synthesis represents polyploidization or cell proliferation cannot
be determined from this data as neither can a dose-response.
2.3.1.4. Carter et al. 1995
The aim of this study was to "use correlative biochemical, pathologic and morphometric
techniques to characterize and quantify the acute, short-term responses of hepatocytes in the
male B6C3F1 mouse to drinking water containing DC A." This report used tritiated thymidine
incorporation, DNA concentration, hepatocyte number per field (cellularity), nuclear size and
binuclearity (polyploidy) parameters to study 0, 0.5 and 5 g/1 neutralized DCA exposures up to
30 days. Male B6C3F1 mice were started on treatment at 28 days of age. Tritiated thymidine
was administered by miniosmotic pump 5 days prior to sacrifice. The experiment was conduced
in two phases which consisted of 5-15 days of treatment (phase I) and 20-30 days of treatment
(phase II) with 5 animals per group in groups sacrificed at 5-day intervals. Liver sections were
stained for H&E, PAS (for glycogen) or methyl green pryonin stain (for RNA). DNA was
extracted from liver homogenates and the amount of tritiated thymidine determined as dpm/[j,g
DNA. Autoradiography was performed with the number of hepatocyte nuclei scored in 1000
hepatocytes selected randomly to provide a labeling index of "number of labeled cells/1000 X
100%". Changes in cellularity, nuclear size and number of multinucleate cells were quantified
in H&E sections at 40 X power. Hepatocyte cellularity was determined by counting the number
of nuclei in 50 microscopic fields with multinucleate cells being counted as one cell and
nonparenchymal cells not counted. Nuclear size was also measured in 200 nuclei with the mean
area plus 2 S.D. was considered to be the largest possible single nucleus. Therefore, polyploid
diploid cells were identified by the authors but not cells that had undergone polyploidy with
increased DNA content in a single nucleus.
Mean body weights at the beginning of the experiment varied between 18.7 and 19.6 g in
the first 3 exposure groups of phase I of the study. Through 15 days of exposure there did not
appear to be a change in body weight in the 0.5 g/1 exposure groups but in the 5g/l exposure
group body weight was reduced at 5, 10 and 15 days with that reduction statistically significant
at 5 and. 15 days. Liver weights did not appear to be increased at day 5 but were increased at
02/12/09
E-104
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
days 10 and 15 in both treatment groups (i.e., means ± S.E.M. for day 10; 1.36 ± 0.03, 1.46 ±
0.03, and 1.59 ± 0.08 g for control, 0.5 g/1 and 5 g/1 DCA, respectively and for day 15; 1.51 ±
0.06, 1.72 ± 0.05, and 2.08 ± 0.11 g for control, 0.5 g/1 and 5 g/1 DCA, respectively). The %
liver/body weight followed a similar pattern with the exception that at day 5 the 5g/l exposure
group had a statistically significant increase over control (i.e., for day 10; 6.00 ± 0.10, 6.72 ±
0.17, and 8.21 ± 0.10 % for control, 0.5 g/1 and 5 g/1 DCA, respectively and for day 15; 6.22 ±
0.08, 6.99 ± 0.15, and 10.37 ± 0.27 % g for control, 0.5 g/1 and 5 g/1 DCA, respectively).
In phase II of the study, control body weights were smaller than phase I and varied
between 16.6 and 16.9 g in the first 3 exposure groups. Liver weights of controls were also
smaller making it difficult to quantitatively compare the two groups in terms of absolute liver
weights. However, the pattern of DCA-induced increases in liver weight and % liver/body
weight remained. The patterns of body weight reduction only in the 5g/l treatment groups and
increased liver weight with DCA treatment at both concentrations continued from 20 to 30 days
of exposure. For liver weight there was a slight but statistically significant increase in liver
weight for the 0.5 g/1 treatment groups over controls (i.e., for day 20; 1.02 ± 0.02, 1.18 ± 0.05,
and 1.98 ± 0.05 g for control, 0.5 g/1 and 5 g/1 DCA, respectively, for day 25; 1.15 ± 0.03, 1.34 ±
0.04, and 2.06 ± 0.12 g for control, 0.5 g/1 and 5 g/1 DCA, respectively, for day 30; 1.15 ± 0.03,
1.39 ± 0.08, and 1.90 ± 0.12 g for control, 0.5 g/1 and 5 g/1 DCA, respectively). For % liver/body
weight there was a small increase at 0.5 g/1 that was not statistically significant but all other
treatments induced increases in % liver/body weight that were statistically significant (i.e., for
day 20; 4.82 ± 0.07, 5.05 ± 0.09, and 9.71 ± 0.11 % for control, 0.5 g/1 and 5 g/1 DCA,
respectively, for day 25; 5.08 ± 0.04, 5.91 ± 0.09, and 10.38 ± 0.58 % for control, 0.5 g/1 and 5
g/1 DCA, respectively, for day 30; 5.17 ± 0.09, 6.01 ± 0.08, and 10.28 ± 0.28 % for control, 0.5
g/1 and 5 g/1 DCA, respectively). Of note is the dramatic decrease in water consumption in the 5
g/1 treatment groups which were consistently reduced by 64% in phase I and 46% in phase II.
The 0.5g/l treatment groups had no difference from controls in water consumption at any time in
the study. The effects of such water consumption decreases would affect body weight as well as
dose received. Given the differences in the size of the animals at the beginning of the study and
the concurrent differences in liver weights and % liver/body weight in control animals between
the two phases, the changes in these parameters through time from DCA treatments cannot be
accurately determined (e.g., control liver/body weights averaged 6.32% in phase I but 5.02% in
phase II). However, % liver/body weight increase were reported to be consistently increased
within and between both phases of the study for the 0.5 g/1 DCA treatment from 5 days of
treatment to 30 days of treatment (i.e., for phase I the average increase was 9.5% and for phase II
the average increased was 12.5% for 0.5 g/1 DCA treated groups). Although increase at 5 days
the nonsignificance of the change may be resultant from the small number of animals examined.
The difference in magnitude of dose and % liver/body weight increase is difficult to determine
given that the 5 g/1 dose of DCA reduced body weight and significantly reduced water
consumption by ~ 50% in both phases of the study. Of note is that the differences in DCA-
induced % liver/body weight were ~ 6-fold for the 15, 25, and 30-day data between the 0.5 and 5
g/1 DCA exposures rather than the 10-fold difference in exposure concentration in the drinking
water.
02/12/09
E-105
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The incorporation of tritiated thymidine into total hepatic DNA control treatment groups
was reported to be 73.34 ± 11.74 dpm/[j,g DNA at 5 days, 34 ± 4.12 dpm/[j,g DNA at 15 days,
and 28.48 ± 3.24 dpm/[j,g DNA at 20 days but was not reported for other treatments. The results
for 0.5g/l treatments were not reported quantitatively but the authors stated that the results
"showed similar trends of initial inhibition followed by enhancement of labeling, the changes
relative to controls were not statistically significant." For 5 g/1 treatment groups the 5-day
treated groups DNA tritiated thymidine incorporation was reported to be 42.8% of controls and
followed by a transient increase at 15 and 20 days (i.e., 2.65- and 2.45-fold of controls,
respectively) but after 25 and 30 days to not be significantly different from controls (data not
shown). Labeling indices of hepatocytes were reported as means but variations as either SEM or
S.D. were not reported. Control means were reported as 5.5, 4, 2, 2, 3.2, and 3.5 % of randomly
selected hepatocytes for 5, 10, 15, 20, 25, and 30 days, respectively, for 4 to 5 animals per group.
In contrast to the DNA incorporation results, no increase in labeling of hepatocytes was reported
to be observed in comparison to controls for any DCA treatment group from 5 to 30 days of
DCA exposure. The 5 g/1 treatment group showed an immediate decrease in hepatocyte labeling
from day 5 onwards that gradually increased ~ half of control levels by day 30 of exposure (i.e.,
<0.5 % LI at day 5, ~ 1% LI at day 10, -0.6% LI at day 20, 1% LI at day 25 and 2% LI at day
30). For the 0.5 g/1 treatment the labeling index was reported to not differ from controls from
days 5 though 15 but to be significantly decreased between days 20 and 30 to levels similar to
those observed for the 5 g/1 exposures. The relatively higher number of hepatocytes
incorporating label reported in this study than others can be reflection of the longer times of
exposure to tritiated thymidine. Here, incorporation was shown for 1 weeks worth of exposure
and reflects the % of cell undergoing synthesis during that time period. Also the higher labeling
index in control animals at the 5 and 10 day exposure periods is probably a reflection of the age
of the animals at the time of study. From the data reported by the authors, there was a
correlation between the patterns of total DNA incorporation of label and hepatocyte labeling
indices in control groups (i.e. higher level of labeling at 5 days than at 15 and 20). However, the
patterns of decreased thymidine labeling reported for hepatocytes were not correlated with a
transient increase in total DNA thymidine incorporation reported with DCA treatment, especially
at the 5 g/1 exposure level with a large decrease reported for the number of labeled hepatocytes at
the same time an increase in total DNA thymidine incorporation was reported. Although
reported to be transiently increased, the total hepatic DNA labeling still represented at most a 2.5
fold increase over control liver which represents a small population of cells. Given that the study
examined hepatocyte labeling in random fields and did not report quantitative zonal differences
in proliferation, a more accurate determination of what hepatocytes were undergoing
proliferation cannot be made from the LI results. Also although the authors report signs of
inflammatory cells for 5-day treatment there is no reference to any inflammatory changes that
may have been observed at later time periods when cellular degeneration and loss of nuclei were
apparent. Such an increase inflammatory infiltrates can increase the DNA synthesis
measurements in the liver. The difference in LI and total DNA synthesis could reflect
differences in nonparenchymal cell proliferation or ploidy changes vs. mitoses in hepatocytes.
02/12/09
E-106
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Clearly, the increases in liver weight that were reported as early as 5 days of exposure could not
have resulted from increased hepatocyte proliferation.
The H&E sections were reported to have been fixed in an aqueous solution that reduced
glycogen content. However, residual PAS positive material (assumed to be glycogen) was
reported to be present indicating that not all of the glycogen had been dissolved. The authors
report changes in pathology between 5 and 30 days in control animals that included straightening
of hepatocyte cording, decreased mitoses, less clarity and more fine granularity of pericentral
hepatocellular cytoplasm, increased numbers of larger nuclei that were not labeled, and reported
differences between animals in the amount of glycogen present (i.e., 2 or 3 animals out of the 5
had less glycogen than other members of the group with less glycogen in the central and
midzonal areas). These changes are consistent with increased polyploidization expected for
maturing mice (see Sections 1.1 and 1.2 above). After 5 days of treatment, 0.5 g/1 exposed
animals were reported to have livers with fewer mitoses and tritiated thymidine hepatocyte
labeling but by 10 days an increase in nuclear size. Labeling was reported to be predominantly in
small nuclei. Animals given 0.5 g/1 DCA for 15, 20, and 25 days were reported to have "focal
cells in the middle zone with less detectable or no cell membranes and loss of the coarse
granularity of the cytoplasm" with some cells not having nuclei or cells having a loss of nuclear
membrane and apparent karyolysis." "Cells without nuclei because the plane of the section did
not pass through the nuclei had the same type of nuclei. Cells without nuclei not related to plane
of section had a condensed cytoplasm." Livers from 20-day and later sacrifice groups treated
with 0.5 g/1 DCA were reported to have normal architecture. After 25 days of treatment
apoptotic bodies were reported to be observed with fewer nuclei around the central veins nuclei
that were larger in central and midzonal areas. In animals treated with 5 g/1 DCA the authors
report similar features as for 0.5 g/1 but in a zonal pattern. Inflammatory cells were reported to
not be observed and after 5 and 10 days a marked decrease in labeled nuclei. After 5 days of 5
g/1 DCA, nuclear depletion in the central and mid-zonal areas was reported. In methyl green
pyronin-stained slides a marked loss of cellular membranes was reported at 5 days with a loss of
nuclei and formation of "lakes of liver cell debris." After 15 days of treatment there was a
reported increase in labeling in comparison to animals sacrificed after 5 or 10 days. The cells
nearest to the triads were reported to have clearing of their cytoplasms and an increase in PAS
positivity. Hepatocytes of both 0.5 and 5 g/1 DCA treatment groups were reported to have
"enlarged, presumably polyploidy nuclei." Some of the nuclei were reported to be "labeled,
usually in hepatocytes in the mid-zonal area."
The morphometric analyses of liver sections were reported to reveal statistically
significant changes in cellularity, nuclear size (as measured by either nuclear area or mean
diameter of the nuclear area equivalent circle), and multinucleated cells during 30 days exposure
to DCA. The authors reported that the concentration of total DNA in the liver, reported as total
j_ig nuclear DNA/ g liver, ranged between 278.17 ± 16.88 and 707.00 ± 25.03 in the control
groups (i.e., 2-5 fold range). No 0.5 g/1 DCA treatment groups differed from their control group
in terms of liver DNA concentration. However, for 10 though 30 days of exposure hepatic DNA
concentrations were reported to be decreased in the 5 g/1 treatment groups (at 5 days there
02/12/09
E-107
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
appeared to be ~ 30% increase over control). The number of cells per field was reported to range
between 24.28 ± 1.94 and 43.81 ± 1.93 in control livers (i.e., 1.8-fold range). From 5 to 15 days
the number of cells/field decreased with 0.5 g/1 DCA treatment although only at day 15 was the
change statistically significant. From 20 to 30 days of treatment only the 30 day treatment
showed a slight decrease in cells/field and that change was statistically significant. After 5 days
of treatment, the number of cells/field was 1.6-fold of control, by 15 days reduced by -20%, and
for 20 to 30 days continued to be reduced by as much as 40%. Although the authors reported
that the changes in cellularity and DNA concentration to be closely correlated, the patterns in the
number of cells/field varied in their consistency with those of DNA concentration (i.e., for days
5, 20 and 25 there direction of change with dose was similar between the two parameters but for
days 10, 15 and 30 were not). If changes in liver weight were due to hepatocellular hypertrophy,
the increased liver size would be matched by a decrease in liver DNA concentration and by the
number of cells/field. The large increases in liver/body weight induced by 5 g/1 DCA were
matched by decreases in liver DNA concentration except for the 5 day exposure group. In
general, the small increases in liver/body weight consistently induced by 0.5 g/1 treatment from
day 5 through 30 were not correlated with DNA concentrations or cells/field. The small number
of animal examined for these parameters (i.e., 4-5) and the highly variable control values limit
the power to accurately detect changes. The apparent dehydration in the animals treated at 5 g/1
DCA was cited by the authors for the transient increase in cellularity and DNA concentration in
the 5-day exposure group. However, drinking water consumption was reported to be similarly
reduced at all treatment periods for 5 g/1 DCA-treated animals so that all groups would
experience the same degree of dehydration.
The percentage of mononucleated cells was reported as % of mononucleated hepatocytes
with results given as means but with no reports of variation within groups. The mean control
values were reported to range between 60% and 75% for phase I and between 58% and 71% for
phase II of the experiment (n= 4-5 animals per group). The % of mononucleated hepatocytes
was reported to be similar between control and DCA treatment groups at 5- and 10-day exposure.
At 15 days both DCA treatments were reported to give a similar increase in mononucleated
hepatocytes (~ 80 % vs. 60% in control) with only the 5 g/1 DCA group statistically significant.
The increase in mononucleated cells reported for DCA treatment is similar in size to the
variation between control values. For phase II of the study, DCA treatment was reported to
increase the number of mononucleated cells in at all concentrations and exposure time periods in
comparison to control values. However, only the increases for the 5 g/1 treatments at days 20
and 25, and the 0.5 g/1 treatment at day 30 were reported to be statistically significant. Again,
small numbers of animals limit the ability to accurately determine a change. However, the
consistent reporting of an increasing number of mononucleated cells between 15 and 30 days
could be associated with clearance of mature hepatocytes as suggested by the report of DCA-
induced loss of cell nuclei.
Mean nuclear area was reported to range between 45 and 54 \i2 in phase I and to range
between 41 and 48 \i2 in phase II of the experiment with no variation in measurements given by
the authors. The only statistically significant differences reported between control and treated
02/12/09
E-108
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
groups in phase I was a decrease from 54 to ~ 42 \i2 in the 0.5 g/1 DCA 10 day treatment group
and a small increase from 50 to ~ 52 \i2 15 day treatment group. Clearly the changes reported by
the authors as statistically significant did not show a dose-related pattern and were within the
range of variation reported between control groups. For phase II of the experiment both DCA
treatment concentrations were reported to induce a statistically significant increase the nuclear
area that was dose-related with the exception of day 30 in which the nuclear area was similar
between the 0.5 and 5 g/1 treatment groups. The largest increase in nuclear area was reported at
20 days for the 5 g/1 treatment group (~ 72 vs. 41 \i2 for control). The patterns of increases in
nuclear area were correlated with those of increased percentage of mononucleated cells in phase
II of the study (20 - 30 days of treatment) as well as the small changes seen in phase I of the
experiment. An increase in nuclear cell area is consistent with increase polyploidization without
mitosis as cells are induced towards polyploidization. A decrease in the numbers of binucleate
cells in favor of mononucleate cells is consistent with clearance of mature binucleate hepatocyte
as well induction of further polyploidization of diploid or tetraploid binucleate cell to tetraploid
or octoploid mononucleate cells. The authors suggested that the "large hyperchromatic
mononucleated hepatocytes are tetraploid" and suggest that such increases in tetraploid cells
have also been observed with non-genotoxic carcinogens and with DEHP. In terms of increased
cellular granularity observed by the authors with DCA treatment, this result is also consistent
with a more differentiated phenotype (Sigal et al., 1999). Thus, these results for DCA are
consistent with a DCA induced change in polyploidization of the cells without cell proliferation.
The pattern of consistent increase in % liver/body weight induced by 0.5 g/1 DCA treatment from
days 5 though 30 was not consistent with the increased numbers of mononucleate cells and
increase nuclear area reported from day 20 onward. The large differences in liver weight
induction between the 0.5 g/1 treatment group and the 5 g/1 treatment groups at all times studied
also did not correlate with changes in nuclear size and % of mononucleate cells. Thus, increased
liver weight was not a function of cellular proliferation, but probably included both aspects of
hypertrophy associated with polyploidization and increased glycogen deposition induced by
DCA. The similar changes reported after short-term exposure for both the 0.5 and 5 g/1 exposure
concentration were suggested by the authors to indicate that the carcinogenic mechanism at both
concentrations would be similar. Furthermore, they suggest that although there is evidence of
cytotoxicity (e.g., loss of cell membranes and apparent apoptosis), the present study does not
support that the mechanism of DCA-induced hepatocellular carcinogenesis is one of regenerative
hyperplasia following massive cell death nor peroxisome proliferation as the 0.5 g/1 exposure
concentration has been shown to increase hepatocellular lesions after 100 weeks of treatment
without concurrent peroxisome proliferation or cytotoxicity (DeAngelo et al., 1999).
2.3.1.5. DeAngelo et al. 1989
Various strains of rats and mice were exposed to TCA (12 and 31 mM) or DCA (16 and
39 mM) for 14 days with Sprague-Dawley rats and B6C3F1 mice exposed to an additional
concentration of 6 mM TCA and 8 mM DCA. Although noting that in a previous study that high
concentrations of chloracids, the authors did not measure drinking water consumption in this
study. This study exposed several strains of male rats and mice to TCA at two concentrations in
02/12/09
E-109
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
drinking water (12 mM and 3 ImM neutralized TCA) for 14 days. The conversion of mmols/1 or
mM TCA is 5 g/1 TCA, 2 g/1 TCA and 1 g/1 for 31 mM, 12 mM, and 6 mM TCA, respectively.
The conversion of mmols/1 of mM DCA is 5 g/1 DC A, 2 g/1 DCA, and 1 g/1 DCA for 39 mM, 16
mM and 8 mM DCA, respectively. The strains of mice tested were Swiss-Webster, B6C3F1,
C57BL/6, and C3H and for rats were Sprague Dawley, Osborne Mendel, and F344. For the
F344 rat and B6C3F1 mice data from two separate experiments were reported for each. The
number of animals in each group was reported to be 6 for most experiments with the exception
of the Sprague-Dawley rats (n=3 at the highest dose of TCA and n=4 or 5 for the control and the
lower TCA dose), one study in B6C3F1 mice (n= 4 or 5 for all groups), and one study in F344
rats (n=4 for all groups). The body weight of the controls was reported to range from 269 to 341
g in the differing strains of rats (1.27-fold) and 21 to 28 g in the differing strains of mice (1.33-
fold) ( age not reported). For % liver/body weight ratios the range was 4.4% to 5.6% in control
rats (1.27-fold) and 5.1 to 6.8% in control mice (1.33-fold).
As discussed in other studies, the determination of PCO activity appears to be highly
variable. This enzyme activity is often used as a proxy for peroxisome proliferation. For PCO
activity the range of activity in controls was much greater than for either body weight or %
liver/body weight. For rats there was a 2.8-fold difference in PCO control activity and in mice
there was a 4.6-fold difference in PCO activity. Between the two studies performed in the same
strain of rat (F344) there was a 2.83-fold difference in PCO activity between controls, and for the
two studies in the same strain of mouse (B6C3F1) there was a 3.14-fold difference in PCO
activity between controls. Not only were there differences between strains and experiments in
the same strain, but also differences in control values between species with a wider range of
values in the mice. The lowest level of PCO activity in control rats, expressed as nanomoles
NAD reduced/min/mg/protein, was 3.34 and for control mice was 1.40. The highest level
reported in control in rats was 9.46 and for control mice was 6.40.
These groups of rats and mice were exposed to 2 g/1 NaCl, 2 g/1 or 5 g/1 TCA in drinking
water for 14 days and their PCO activity assayed. These doses of TCA did not affect body
weight except for the Sprague-Dawley rats which lost ~ 16% of their body weight. This was also
the same group in which only 3 rats survived treatment. The Osborne-Mendel and F344 strains
did no exhibit loss of body weight or mortality due to TCA exposure. There was a large
variation in response to TCA exposure between the differing strains of rats and mice with a much
larger difference between the strains of mice. For the 3 rat strains tested there was a range
between 0% change and 2.38-fold of control for PCO activity at the 5 g/1 TCA exposure. For the
2 g/1 TCA exposure, there was a range of 0% change to 1.54-fold of control for PCO activity.
The Osborne-Mendel rats had 1.54-fold of control value for PCO activity at 2 g/1 TCA and 2.38-
fold of control value for PCO activity reported at 5 g/1, exhibiting the most consistent increase in
PCO with increased dose of TCA. Two experiments were reported for F344 rats with one
reporting a 1.63-fold of control and the other a 1.79-fold of control value for 5 g/1 TCA. Only
one of the F334 experiments also exposed rats to 2 g/1 TCA and reported no change from control
values.
02/12/09
E-110
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For the 4 strains of mice tested there was a range of 7.44- to 22.13-fold of control values
reported at the 5 g/1 TCA exposures and 3.76- to 25.92-fold of control values at the 2 g/1 TCA
exposures for PCO activity. For the C57BL/6 strain of mice there was little difference between
the 5 g/1 and 2 g/1 TCA exposures and a generally 3-fold higher induction of PCO activity by
TCA at the 5 g/1 TCA exposure level than for the other mouse strains. Although there was a 2.5-
fold difference between the 5 g/1 and 2 g/1 TCA exposure dose, the difference in magnitude of
PCO activity between these doses ranged from 0.85-fold to 2.23-fold for all strains of mice. For
the B6C3F1 mice there was a difference between reported increases of PCO activity in the text
(i.e., reported as 9.59-fold of control) for one of the experiments and that presented graphically
in Figure 2 (i.e., 8.70-fold of control). Nevertheless in the two studies of B6C3 F1 mice, 5 g/1
TCA was reported to induce 7.78-fold of control and 8.70-fold of control for PCO activity, and 2
g/1 TCA was reported to induce 5.56-fold of control and 4.70-fold of control for PCO activity.
For the two F344 rat studies in which ~ 200 mg/kg or 5 g/1 TCA was administered for 10 or 14
days, there was 1.63-fold of control and 1.79-fold of control values reported for PCO activity.
Thus, for experiments in which the same strain and dose of TCA were administered, there was
not as large a difference in PCO response than between strains and species.
Whether increases in % liver/body weight ratios were similar in magnitude to increased
PCO activity can be assessed by examination of the differences in magnitude of increase over
control for the 5 g/1 and 2 g/1 TCA treatments in the varying rat strains and mouse strains. The
relationship in exposure concentration was a 2.5:1 ratio for the 5 g/1 and 2 g/1 doses. For rats
treatment of 5 g/1 TCA to Sprague-Dawley rats resulted in a significant decrease in body weight
and therefore affected the magnitude of increase in % liver/body weight ratio for this group.
However, for the rest of the rat and mouse data, this dose was not reported to affect body weight
so that there is more confidence in the dose-response relationship. For the Sprague-Dawley rat
there was no change in the % liver/body weight ratio at 2 g/1 but a 10% decrease at 5 g/1 TCA
exposure with no change in PCO activity for either. However, for the Osborne-Mendel rats,
there was no change in % liver/body weight ratios for either exposure concentration of TCA, but
PCO activity was reported to be 1.54-fold of control at 2 g/1 and 2.38-fold of control at 5 g/1
TCA. Thus there was a ratio of 2.5-fold increase in PCO activity between the 5 g/1 and 2 g/1
treatment groups. For the F344 rats there was a 2-fold difference in liver weight increases (i.e.,
12% vs. 6% increase over control) between the two exposure concentrations but 1.6-fold of
control value for PCO activity at the 5 g/1 TCA exposure concentration and no increase in PCO
activity at the 2 g/1 level. Thus for the three strains of rats, there did not appear to be a consistent
correlation between liver weight induction by TCA and PCO activity.
For differing strains of mice, similar concentrations of TCA were reported to vary in the
induction of liver weight increases. The range of liver weight induction was 1.26-fold to 1.66-
fold of control values between the 4 strains of mice at 5 g/1 TCA and 1.16-fold to 1.63-fold at 2
g/1 TCA. In general, for mice the magnitudes of the difference in the increase in dose between
the 5 g/1 and 2 g/ TCA exposure concentration (2.5-fold) was generally higher than the increase
% liver/body weight ratios at these doses. The differences in liver weight induction between the
2g/l and 5g/l doses were ~ 40% for the Swiss-Webster, C3H, and for one of the B6C3F1 mouse
02/12/09
E-lll
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
experiments. For the C57BL/6 mouse there was no difference in liver weight induction between
the 2 g/1 and 5g/l TCA exposure groups. For the other B6C3F1 mouse experiments there was a
2.5-fold greater induction of liver weight increase for the 5g/l TCA group than for the 2 g/1
exposure group (1.39-fold of control vs. 1.16-fold of control for % liver/body weight,
respectively). For PCO activity the Swiss-Webster, C3H, and one of the B6C3F1 mouse
experiments were reported to have ~ 2-fold difference in the increase in PCO activity between
the two doses. For the other B6C3F1 mouse experiment there was only about a 50% increase
and for the C57BL/6 mouse data there was 15% less PCO activity induction reported at the 5 g/1
TCA dose that at the 2 g/1 dose. None of the difference in increases in liver weight or PCO
activity in mice from the 2 g/1 or 5g/1 TCA exposures were of the same magnitude as the
difference in TCA exposure concentration (i.e., 2.5-fold) except for liver weight from the one
experiment in B6C3F1 mice. This is also the data used fore comparisons with the Sprague-
Dawely rat discussed below.
In regard to strain differences for TCA response in mice, there did not appear to be
correlations of the magnitude of 5 g/1 TCA-induced changes in % liver/body weight ratio or PCO
activity, with the body weights reported for control mice for each strain. The control weights
between the 4 strains of mice varied from 21 to 28 grams. The strain with the greatest response
(C57B1/6) for TCA-induced changes in % liver/body weight ratio (i.e., 1.66-fold of control) and
PCO activity (22.13-fold of control) had a mean body weight reported to be 26 grams for
controls. At this dose, the range of % liver/body weight for the other strains was reported to be
1.26- to 1.39-fold of control and the range of PCO activity reported to be of 7.48- to 8.71-fold of
control.
Of note is that in the literature, this study has been cited as providing evidence of
differences between rats and mice for peroxisomal response to TCA and DCA. Generally the
PCO data from the Sprague Dawley rats and B6C3F1 mice at the highest dose of TCA and DCA
have been cited. However, the Sprague-Dawley strain was reported to have greater mortality
from TCA at this exposure than the other strains tested (i.e., only 3 rats survived and provided
PCO levels) and a lower PCO response (no change in PCO activity over control) that the other
two strains tested in this study (i.e., Osborne-Mendel rats was reported to have had 2.38-fold of
control and the F344-had a 1.63-tol.79-fold of control for PCO activity after exposure to 5 g/1
TCA with no mortality). The B6C3F1 mouse was reported to have a 7.78- or 8.71-fold of
control for PCO activity from 5 g/1 TCA exposure. Certainly the male mouse is more responsive
to TCA induction of PCO activity. However, as discussed above there are large variations in
control levels of PCO activity and in the magnitude and dose-response of TCA-induction of PCO
activity between rat and mouse strains and between species. If is not correct to state that the rat is
refractory to TCA-induction of peroxisome activity.
Unfortunately, the authors chose the Sprague-Dawley rat (i.e., the most unresponsive
strain for PCO activity and most sensitive to toxicity) for studies for comparative studies
between DCA and TCA effects. The authors also tested for carnitine acetyl CoA transferase
(CAT) activity as a marker of peroxisomal enzyme response and took morphometric analysis of
02/12/09
E-112
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
peroxisome # and cytoplasmic volume for one liver section for each of two B6C3F1 mice of
Sprague-Dawley rats from the 5 g/1 TCA and 5 g/1 DCA treatment groups. Only 6 electron
micrograph fields were analyzed from each section (12 fields total) were analyzed without
identification as to what area of the liver lobules they were being taken from. Hence there is a
question as to whether the areas which are known to be peroxisome rich were assayed of not.
Also as noted above, previous studies have indicate that such high concentration of DCA and
TCA inhibit drinking water consumption and therefore raising issues not only about toxicity but
also the dose which rats and mice received. The number of peroxisomes per 100 [j,m3 and
cytoplasmic volume of peroxisomes was reported to be 6.60 and 1.94%, respectively, for control
rats, and 6.89 and 0.61% for control mice, respectively. For 5 g/1 TCA and 5 g/1 DCA the
numbers of peroxisomes were reported to be increased to 7.14 and 16.75, respectively in treated
Sprague Dawley rats. Thus there was 2.5-fold and 1.08-fold of control reported in peroxisome #
for 5 g/1 DCA and TCA, respectively. The cytoplasmic volume of peroxisomes was reported to
be 2.80 % and 0.89% for 5 g/1 DCA and 5 g/1 TCA, respectively (i.e., a 1.44-fold of control and
~ 60% reduction for 5 g/1 DCA and 5 g/1 TCA, respectively). Thus, 5 g/1 TCA was reported to
slightly increase the number of peroxisomes and but decrease the % of the cytoplasmic volume
occupied by peroxisome by half. For DCA the reported pattern was for both to increase. PCO
activity was reported to increase by a similar magnitude as peroxisome # but not volume in the 5
g/1 TCA treated Sprague-Dawley rats. However, although peroxisomal volume was reported to
be cut nearly in half and for peroxisome number to be similar, 5 g/1 TCA treatment was not
reported to change PCO activity in the Sprague-Dawley rat.
For comparisons between DCA and TCA B6C3 F1 mice were examined at 1.0, 2.0, and
5.0 g/1 concentrations. DCA was reported to induce a higher % liver/body weight ratio that did
TCA at every concentration (i.e., 1.55-fold, 1.27-fold, 1.21-fold of control for DCA and 1.39-
fold, 1.16-fold, and 1.08-fold of control for TCA at 1.0, 2.0, and 5.0 g/1 concentrations,
respectively). As noted above, for other strains of mice tested and a second experiment with
B6C3F1 mice, there was 40% or less difference in % liver/body weight ratio between the 2.0 g/1
and 5.0 g/1 exposures to TCA but for this experiment there was a 2.5-fold difference. Thus, at 5
g/1 there was ~ 40% greater induction of liver weight for DCA than TCA. In the B6C3F1 mice,
5 g/1 TCA was reported to increase peroxisome number to 30.75 and cytoplasmic volume to
4.92% (i.e., 4.4-fold and 8.1-fold of control, respectively). For 5 g/1 DC A treatment, the
peroxisome number was reported to be 30.77 and 2.15% (i.e., 4.5-fold and 6.1-fold of control,
respectively). While there was no difference in peroxisome number and ~ 40% difference in
cytoplasmic volume at the 5.0 g/1 exposures of DCA and TCA, there was a greater difference in
the magnitude of PCO activity increase. The 5 g/1 TCA exposure was reported to induce 4.3-fold
of control for PCO activity while 5 g/1 DCA induced as 9.6 -fold of control PCO activity
(although a Figure in the report shows 8.7-fold of control) which is a ~ 2.5-fold difference
between DCA and TCA at this exposure concentration.. Thus for one of the B6C3F1 mouse
studies, 5 g/1 DCA and TCA treatments were reported to give a similar increase peroxisome #,
TCA to induce a 40% greater increase in peroxisomal cytoplasmic volume than DCA and a 2.5-
fold greater increase in PCO activity, but DCA to induce ~ 40% greater liver weight induction
than TCA.
02/12/09
E-113
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Not only were PCO activity, peroxisome number and cytoplasmic volume occupied by
peroxisomes analyzed but also CAT activity as a measure of peroxisome proliferation. For TCA
and DCA the results were opposite those reported for PCO activity. In Sprague-Dawley rats
control levels of CAT were reported to be 1.81 nmoles of carnitine transferred/min/mg/protein.
Exposure to 5 g/1 TCA was reported to increase CAT activity by 3.21-fold of control while 5 g/1
DCA was reported to induce CAT activity to 10.33-fold of control levels in Sprague-Dawley
rats. However, while PCO activity was reported to be the same as controls, and peroxisomal
volume decreased, 5 g/1 TCA increased CAT activity 3.21-fold of control in these rats. The level
of CAT induced by 5 g/1 DCA was over 10-fold of control in the rat while peroxisome # was
only 2.5-fold of control and cytoplasmic volume 1.4-fold of control. Thus, the fold increases for
these three measures were not the same for DCA treatment and for TCA in rats. Nevertheless for
CAT, DCA was a stronger inducer in rats than was TCA. In B6C3 F1 mice 5 g/1 TCA and 5 g/1
DCA induced CAT activity to a similar extent (4.50-fold and 5.61-fold of control, respectively).
The magnitude of CAT induction was similar to that of peroxisome # for both 5 g/1 DCA and 5
g/1 TCA and lower than PCO activity in DCA-treated mice and cytoplasmic volume in TCA-
treated mice by about half. Thus, using CAT as the marker of peroxisome proliferation, the rat
was more responsive than the mouse to DCA and nearly as responsive to TCA as the mouse at
this high dose in these two specific strains. This, data illustrates the difficulty of using only one
measure for peroxisome proliferation and shows that the magnitude of increased PCO activity is
not necessarily predictive of the peroxisome # or cytoplasmic volume or CAT activity. The
difficulty of interpretation of the data from so few animals and sections for the electron
microscopy analysis, and the low number of animals for PCO activity and CAT activity (n= 3 to
6), the high dose studied (5 g/1), and the selection of a rat strain that appears to be more resistant
to this activity but more susceptible to toxicity than the others tested, should be taken into
account before conclusions can be made about differences between these chemicals for
peroxisome activity between species.
Of note is that PCO activity was also shown to be increased by corn oil alone in F344 rats
and to potentiate the induction of PCO activity of TCA. After 10 days of exposure to either
water, corn oil, 200 mg/kg/day TCA in corn oil or 200 mg/kg TCA in water via gavage dosing,
there was 1.40-fold PCO activity from corn oil treatment alone in comparison to water, a 1.79-
fold PCO activity from TCA in water treatment in comparison to water, and a 3.14-fold PCO
activity from TCA in corn oil treatment in comparison to water.
The authors provided data for 3 concentrations of DCA and TCA for Sprague-Dawley
and for one experiment in the B6C3F1 mouse for examination of changes in body and %
liver/body weight ratios (1, 2, or 5 g/1 DCA or TCA) after 14 days of exposure. As noted above,
not only did the 5 g/1 exposure concentration of DCA result in mortality in the Sprague-Dawley
strain of rat, but the 5 g/1 and 2 g/1 concentrations of DCA were reported to decrease body weight
(~ 20 and 25%, respectively). The 5 g/1 dose of TCA was also reported to induce a statistically
significant decrease in body weight in the Sprague-Dawley rat. There were no differences in
final body weight in any of the mice exposed to TCA or DCA. As noted above no TCA or DCA
exposure group of Sprague-Dawley rats was reported to have a statistically significant increase
02/12/09
E-114
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
in % liver/body weight ratio over control. For the B6C3F1 male mice, the % liver/body weight
ratio was 1.22-fold, 1.27-fold and 1.55 fold of control after exposure to 1 g/1, 2 g/1 and 5 g/1
DCA, respectively, and 1.08-fold, 1.16-fold, and 1.39-fold of control after exposure to 1 g/1, 2
g/1, and 5 g/1 TCA, respectively. Thus, for DCA there was only a 20% increase in liver weight
corresponding to the 2-fold increase between the 1 and 2 g/1 exposure levels of DCA. Between
the 2 g/1 and 5 g/1 exposure concentrations of DCA there was a 2-fold increase in liver weight
corresponding to a 2.5-fold increase in exposure concentration. For TCA, the magnitude of
increase in dose was reported to be proportional to the magnitude of increase in % liver/body
weight ratio in the B6C3 F1 male mouse. As stated above, the correspondence between
magnitude of dose and % liver weight for TCA exposure in this experiment differed from the
other experiment reported for this strain of mouse and also differed from the other 3 strains of
mice examined in this study where the magnitude in liver weight gain was much less than
exposure concentration.
2.3.2. Subchronic and Chronic Studies of DCA and TCA
Several experiments have been conducted with exposure to DCA and TCA, generally at
very high levels with a limited dose range, for less periods of time than standard carcinogenicity
bioassays, and with very limited information on any endpoints other than the liver tumor
induction. Caldwell and Keshava (2006) and Caldwell et al. (2008b) have examined these
studies for inferences of modes of action for TCE. Key studies are briefly described below for
comparative purposes of results reported in TCE studies.
2.3.2.1. Snyder et al. 1995
Studies of TCE have reported either no change or a slight increase in apoptosis only after
a relatively high exposure level (Dees and Travis, 1993; Channel et al., 1998). Inhibition of
apoptosis, which has been suggested to prevent removal of "initiated" cells from the liver and
lead to increased survival of precancerous cells, has been proposed as part of the MOA for
peroxisome proliferators (see Section 3.4). The focus of this study was to examine whether
DCA, which has been shown to inhibit DNA synthesis after an initial transient increase (see
Section 2.3.3 below), also alters the frequency of spontaneous apoptosis in mice. This study
exposed 28-day old male B6C3F1 male mice (n=5) to 0, 0.5 or 5.0 g/1 buffered DCA in drinking
water for up to 30 days (phase 1 = 5-15 days exposure and phase II = 20-30 days treatment).
Portions of the left lobe of the liver were prepared for histological examination after H&E
staining. Hepatocyte number was determined by counting nuclei in 50 fields with
nonparenchymal cell nuclei excluded on the basis of nuclear size. Multinucleate cells were
counted as one cell. Apoptotic cells were visualized by in situ TDT nick end-labeling (TUNEL)
assay from 2-4 different liver sections from each control or treated animal. The average number
of apoptotic cells was then determined for each animal in each group. The authors reported that
in none of the tissues examined were necrotic foci observed, there was no any indication of
lymphocyte or neutrophil infiltration indicative of an inflammatory response, and suggested that
no necrotic cells contributed to the responses in their analysis.
02/12/09
E-115
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Control animals were reported to exhibit apoptotic frequencies ranging from ~ 0.04 to
0.085% and that over the 30-day period the frequency rate declined. The authors suggested that
this result is consistent with reports of the livers of these young animals undergoing rapid
changes in cell death and proliferation. They note that animals receiving 0.5 g/1 DCA also had a
similar trend of decreasing apoptosis with age, supportive of the decrease being a physiological
phenomenon. The 0.5 g/1 exposure level of DCA was reported to decrease the percentage of
apoptotic hepatocytes as the earliest time point studied and to remain statistically significantly
decreased from controls from 5 to 30 days of exposure. The rate of apoptosis ranged from ~
0.025 to 0.060% after 0.5 g/1 DCA exposure during the 30-day period (i.e. and ~ 30-40 %
reduction). Animals receiving the 5.0 g/1 DCA dose exhibited a significant reduction at the
earliest time point that was sustained at a similar level and statistically significant throughout the
time-course of the experiment (% apoptosis ranged from 0.015 - 0.030%). The results of this
study not only provides a baseline of apoptosis in the mouse liver, which is very low, but also to
show the importance of taking into account the effects of age on such determinations. The
authors reported that the for rat liver the estimated frequency of spontaneous apoptosis to be ~
0.1% and therefore greater than that of the mouse. The significance of the DCA-induced
reduction in apoptosis, of a level that is already inherently low in the mouse, for the MOA for
induction of cancer is difficult to discern.
2.3.2.2. Mather et al. 1990
This 90-day study in male Sprague-Dawley rats examined the body and organ weight
changes, liver enzyme levels, and PCO activity in livers from rats treated with estimated
concentrations of 3.9, 35.5, 345 mg/kg day DCA or 4.1, 36.5, or 355 mg/kg/day TCA from
drinking water exposures (i.e., 0, 50, 500 and 5000 ppm or 0.05, 0.5, or 5.0 g/1 DCA or TCA in
the drinking water). All dose levels of DCA and TCA were reported to result in a dose-
dependent decrease in fluid intake at 2 months of exposure. The rats were 9 (DCA) or 10 (TCA)
weeks old at the beginning of the study (n= 10/group). Animals with body weights that varied
more than 20% of mean weights were discarded from the study. The DCA and TCA solutions
were neutralized. The mean values for initial weights of the animals in each test group varied
less than 3%. DCA treatment induced a dose-related decrease in body weight that was
statistically significant at the two highest levels (i.e., a 6%, 9.5%, and 17% decrease from
control). TCA treatment also resulted in lower body weights that were not statistically
significant (i.e., 2.1%, 4.4%, and 5.9%). DCA treatments were reported to result in a dose-
related increase in absolute liver weights (1.01-fold, 1.13-fold, and 1.36-fold of control that were
significantly different at the highest level) and % liver/body weight ratios (1.07-fold, 1.24-fold,
and 1.69-fold of control that were significant at the two highest dose levels). TCA treatments
were reported to not result in changes in either absolute liver weights or % liver/body weight
ratios with the exception of statistically significant increase in % liver/body weight ratios at the
highest level of treatment (1.02-fold of control). Total serum protein levels were reported to be
significantly depressed in all animals treated with DCA with animals in the two highest dose
groups also exhibiting elevations of alkaline phosphatase. Alanine-amino transferase levels were
reported to be elevated only in the highest treatment group. No consistent treatment-related
02/12/09
E-116
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
effect on serum chemistry was reported to be observed for the TCA-treated animals with data not
shown. In terms of PCO activity, there was only a mild increase at the highest dose of 15% for
TCA and a 2.5-fold level of control for DC A treatment that were statistically significant. The
difference in PCO activity between control groups for the DCA and TCA experiments was
reported to be 33%. No treatment affect was reported to be apparent for hepatic microsomal
enzymes, or measures of immunotoxicity for either DCA or TCA but data were not shown.
Focal areas of hepatocellular enlargement in both DCA- and TCA-treated rats were reported to
be present with intracellular swelling more severe with the highest dose of DCA treatment.
Livers from DCA treated rats were reported to stain positively for PAS, indicating significant
amounts of glycogen with TCA treated rats reported to display "less evidence of glycogen
accumulation." Of note is that, in this study of rats, DCA was reported to induce a greater level
of PCO activity than did TCA.
2.3.2.3. Parrish etal. 1996
Parrish et al. (1996) exposed male B6C3F1 mice (8 weeks old and 20-22 g upon
purchase) to TCA or DCA (0, 0.01, 0.5 and 2.0 g/1) for 3 or 10 weeks (n= 6). Livers were
excised and nuclei isolated for examination of 8-OHdG and homogenates examined for cyanide
insensitive Acyl-CoA oxidase (PCO) and laurate hydroxylase activity. The authors noted that
control values between experiments varied as much as a factor of 2-fold for PCO activity and
that data were presented as percent of concurrent controls. Initial body weights for treatment
groups were not presented and thus differences in mean values between the groups cannot be
ascertained.
Final body weights were reported to not be statistically significantly changed by DCA or
TCA treatments at 21 days or 71 days of treatment (all were within ~ 8% of controls). The mean
% liver/body ratios were reported to be 5.4%, 5.3%, 6.1%, and 7.2% for control, 0.1 g/1, 0.5 g/1,
and 2.0 g/1 TCA, respectively and 5.4%, 5.5%, 6.7%, and 7.9% for control, 0.1 g/1, 0.5 g/1 and
2.0 g/1 DCA, respectively after 21 days of exposure. This represents 0.98-fold, 1.13-fold, 1.33-
fold of control levels with these exposure levels of TCA and 1.02-fold, 1.24-fold, and 1.46-fold
of control levels with DCA after 21 days of exposure. For 71 days of exposure the mean %
liver/body ratios were reported to be 5.1%, 4.6%, 5.8%, and 6.9% for control, 0.1 g/1, 0.5 g/1,
and 2.0 g/1 TCA, respectively and 5.1%, 5.1%, 5.9%, and 8.5% for control, 0.1 g/1, 0.5 g/1 and
2.0 g/1 DCA, respectively. This represents 0.90-fold, 1.14-fold, 1.35-fold of control with TCA
exposure and 1.0-fold, 1.15-fold, and 1.67-fold of control with DCA exposure after 71 days of
exposure. The magnitude of difference between the 0.1 and 0.5 g/1 TCA doses is 5 and 0.5 and
2.0 g/1 doses is 4-fold. For the 21-day and 71-day exposures the magnitudes of the increases in
% liver/body weight over control values were greater for DCA than TCA exposure at same
concentration with the exception of 0.5 g/1 doses at 71 days in which both TCA and DCA
induced similar increases. For TCA, the 0.01 g/1 dose produces a similar 10% decrease in %
liver/body weight. Although there was a 4-fold increase in magnitude between the 0.5 and 2.0
g/1 TCA exposure concentrations, the magnitude of increase for % liver/body weight increase
was 2.5-fold between them at both 21 and 71 days of exposure. For DCA, the 0.1 g/1 dose was
02/12/09
E-117
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
reported to have a similar value as control for % liver/body weight ratio. Although there was a
4-fold difference in dose between the 0.5 and 2.0 g/1 DCA exposure concentrations, there was a
~ 2-fold increase in % liver/body weight increase at 21 days and ~ 4.5-fold increase at 71 days.
As a percentage of control values, TCA was reported to induce a dose-related increase in
PCO activity at 21 days (~ 1.5-fold, 2.2-fold, and ~ 4.1-fold of control, for 0.1, 0.5, and 2.g/1
TCA exposures). Only the 2.0 g/1 dose of DCA was reported to induce a statistically significant
increase at 21-days of exposure of PCO activity over control (~ 1.8-fold of control) with the 0.1
and 0.5 g/1 exposure PCO activity to be slightly less than control values (~ 20% less). Thus,
although there was no increase in % liver/body weight at 0.1 g/1 TCA, the PCO activity was
reported to be increased by -50% after 21 days. A 13% increase in liver weight at 0.5 g/1 TCA
was reported to be associated with 2.2-fold of control level of PCO activity and a 33% increase
in liver weight after 2.0 g/1 TCA to be associated with 4.1-fold of control level of PCO activity.
Thus, increases in PCO activity were not necessarily correlated with concurrent TCA-induced
increases in liver weight and the magnitudes of increase in liver weight between 0.5g/l and 2.0
g/1 TCA (2.5-fold) was greater than the corresponding increase in PCO activity (1.8-fold of
control). Although there was a 20-fold difference in TCA dose, the magnitude of increase in
PCO activity between 0.1 and 2.0 g/1 TCA was ~ 2.7-fold. As stated above, the 4-fold difference
in TCA dose at the two highest levels resulted in a 2.5 fold increase in liver weight. For DCA,
the increases in liver weight at 0.1 and 0.5 g/1 DCA exposures were not associated with increased
PCO activity after 21 days of exposure. The 2.0 g/1 DCA exposure concentration was reported to
induce 1.8-fold of control PCO activity. After 71 days of treatment, TCA induced a dose-related
increase in PCO activity that was ~ twice the magnitude as that reported at 21 days (i.e. ~ 9-fold
greater at 2.0 g/1 level). After 71 days, for DCA the 0.1 and 0.5 g/1 doses produced a statistically
significant increase in PCO activity (~ 1.5-fold and 2.5-fold of control, respectively). The
administration of 1.25 g/1 clofibric acid in drinking water was used as a positive control and
reported to induce ~6-7-fold of control PCO activity at 21 and 71 days of exposure.
Laurate hydroxylase activity was reported to be elevated significantly only by TCA at 21
days (2.0g/l TCA dose only) and to increased to approximately the same extent (~ 1.4 to 1.6-fold
of control values) at all doses tested. For 0.1 g/1 DCA the laurate hydroxylase activity was
reported to be similar to that of 0.1 g/1 TCA (~ 1.4-fold of control) but to be ~ 1.2-fold of control
at both the 0.5 and 2.0 g/1 DCA exposures. At 71 days, both the 0.5 and 2.0 g/1 TCA exposures
induced a statistically significant increase in laurate hydroxylase activity (i.e., 1.6-fold and 2.5-
fold of control, respectively) with no change after DCA exposure. The actual data rather than %
of control values were reported for laurate hydroxylase activity. The control values for laurate
hydroxylase activity varied 1.7-fold between 21 and 71 days experiments. The results for 8-
OHdG levels are discussed in Section 3.4.2.3 below. Of note is that the increases in PCO
activity noted for DCA and TCA were not associated with 8-OHdG levels (which were
unchanged, see Section 3.4.2.3 below) and also not with changes laurate hydrolase activity or %
liver/body weight ratio increases observed after either DCA or TCA exposure. A strength of this
study is that is examined exposure concentrations that were lower than those examined in many
other short-term studies of DCA and TCA.
02/12/09
E-118
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.3.2.4 Bull et al. 1990
The focus of this study was the determination of "dose-response relationships in the
tumorigenic response to these chemicals [sic DCA and TCA] in B6C3F1 mice, determine the
nature of the non-tumor pathology that results from the administration of these compounds in
drinking water, and test the reversibility of the response." Male and female B6C3F1 mice (age
37 days) were treated from 15 to 52 weeks with neutralized TCA and TCA. A highly variable
number and generally low number of animals were reported to be examined in the study with
n=5 for all time periods except for 52 weeks where in males the n= 35 for controls, n=l 1 for 1
g/L DCA, n = 24 for 2 g/L DCA, n= 11 for lg/L TCA, and n= 24 for 2 g/L TCA exposed mice.
Female mice were only examined after 52 weeks of exposure and the number of animals
examined was n= 10 for control, 2g/L DCA, and 2 g/L TCA exposed mice. "Lesions to be
examined histologically for pathological examination were selected by a random process" with
lesions reported to be selected from 31 of 65 animals with lesions at necropsy. 73 of 165 lesions
identified in 41 animals were reported to be examined histologically. All hyperplastic nodules,
adenomas and carcinomas were lumped together and characterized as hepatoproliferative lesions.
Accordingly there were only exposure concentrations available for dose-response analyses in
males and only "multiplicity of hepatoproliferative lesions" were reported from random samples.
Thus, this data cannot be compared to other studies and is unsuitable for dose-response with
inadequate analysis performed on random samples for pathological examination. The authors
state that some of the lesions taken at necropsy and assumed to be proliferative were actually
histologically normal, necrotic, or an abscess as well. It is also limited by a relatively small
number of animals examined in regard to adequate statistical power to determine quantitative
differences. Similar concerns were raised by Caldwell et al (2007b) with a subsequent study
(e.g., Bull, 2002). For example, the authors report that 5/11 animals had "lesions" at lg/L TCA
at 52 weeks and 19/24 animals had lesions at 2 g/L TCA at 52 weeks. However, while 7 lesions
were examined in 5 mice bearing lesions at 1 g/L TCA, only 16 of 30 lesions from 11 of the 19
animals bearing lesions examined in the 2 g/L TCA group. Therefore, almost half of the mice
with lesions were not examined histologically in that group along with only half of the "lesions."
The authors reported the effects of DCA and TCA exposure on liver weight and %
liver/body changes (m ± SEM) and these results gave a pattern of hepatomegaly generally
consistent with short-term exposure studies. The authors report "no treatment produced
significant changes in the body weight or kidney weight of the animals (data not shown)" In
male mice (n = 5) at 37 weeks of exposure, liver weights were reported to be 1.6 ± 0.1, 2.5 ± 0.1,
and 1.9 ± 0.1 g for control, 2 g/1 DCA, and 2 g/1 TCA exposed mice, respectively. The %
liver/body weights were reported to be 4.1 ± 0.3, 7.3 ± 0.2, and 5.1. ± 0.1 % for control, 2 g/1
DCA, and 2 g/1 TCA exposed mice, respectively. In male mice at 52 weeks of exposure, liver
weights were reported to be 1.7 ± 0.1, 2.5 ± 0.1, 5.1 ± 0.1 , 2.2 ± 0.1, and 2.7 ± 0.1 g for control
(n = 35), 1 g/1 DCA (n = 11), 2 g/1 DCA (n= 24), 1 g/1 TCA (n = 11), and 2 g/1 TCA (n = 24)
exposed mice, respectively. In male mice at 52 weeks of exposure, % liver/body weights were
reported to be 4.6 ± 0.1, 6.5 ± 0.2, 10.5 ± 0.4, 6.0 ± 0.3, and 7.5 ± 0.5 % for control, 1 g/1 DCA,
2 g/1 DCA, 1 g/1 TCA, and 2 g/1 TCA exposed mice, respectively. For female mice (n = 10) at
02/12/09
E-119
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
52 weeks of exposure, liver weights were reported to be 1.3 ± 0.1, 2.6 ± 0.1, and 1.7 ± 0.1 g for
control, 2 g/1 DCA, and 2 g/1 TCA exposed mice, respectively. The % liver/body weights were
reported to be 4.8 ± 0.3, 9.0 ± 0.2, and 6.0. ± 0.3 % for control, 2 g/1 DCA, and 2 g/1 TCA
exposed mice, respectively. Although the number of animals examined varied 3-fold between
treatment groups in male mice, the authors reported that all DCA and TCA treatments were
statistically increased over control values for liver weight and % body/liver weight in both
genders of mice. In terms of % liver/body weight ratio, female mice appeared to be as
responsive as males at the exposure concentration tested. Thus, hepatomegaly reported at these
exposure levels after short-term exposures appeared to be further increased by chronic exposure
with equivalent levels of DCA inducing greater hepatomegaly than TCA.
Interestingly, after 37 weeks of treatment and then a cessation of exposure for 15 weeks
liver weights were assessed in control male mice, 2 g/1 DCA treated mice, and 2 g/1 TCA treated
mice (n = 11 for each group but results for controls were pooled and therefore n = 35). Liver
weights were reported to be 1.7 ± 0.1, 2.2 ± 0.1, and 1.9 ± 0.1 g for control, 2 g/1 DCA, and 2 g/1
TCA exposed mice, respectively. The % liver/body weights were reported to be 4.6 ± 0.1, 5.7 ±
0.3, and 5.4. ± 0.2 % for control, 2 g/1 DCA, and 2 g/1 TCA exposed mice, respectively. After 15
weeks of cessation of exposure, liver weight and % liver/body weight were reported to still be
statistically significantly elevated after DCA or TCA treatment. The authors partially attribute
the remaining increases in liver weight to the continued presence of hyperplastic nodules in the
liver. The authors state that because of the low incidence of lesions in the control group and the
two groups that had treatments suspended, all the lesions from these groups were included for
histological sectioning. However, the authors present a table indicating that, of the 23 lesions
detected in 7 mice exposed to DCA for 37 weeks, 19 were examined histologically. Therefore,
groups that were exposed for 52 weeks had a different procedure for tissue examination as those
at 37 weeks. In terms of liver tumor induction, the authors stated that "statistical analysis of
tumor incidence employed a general linear model ANOVA with contrasts for linearity and
deviations from linearity to determine if results from groups in which treatments were
discontinued after 37 weeks were lower than would have been predicted by the total dose
consumed." The multiplicity of tumors observed in male mice exposed to DCA or TCA at 37
weeks and then sacrificed at 52 weeks were reported by the authors to have a response in animals
that received DCA very close to that which would be predicted from the total dose consumed by
these animals. The response to TCA was reported by the authors to deviate significantly (P=
0.022) from the linear model predicted by the total dose consumed. Multiplicity of lesions per
mouse and not incidence was used as the measure. Most importantly the data used to predict the
dose response for "lesions" used a different methodology at 52 weeks than those at 37 weeks.
Not only were not all animal's lesions examined but foci, adenomas, and carcinomas were
combined into one measure. Therefore foci, of which a certain percentage have been commonly
shown to spontaneously regress with time, were included in the calculation of total "lesions".
Pereira and Phelps (1996) note that in initiated mice treated with DCA, the yield of altered
hepatocytes decreases as the tumor yields increase between 31 and 51 weeks of exposure
suggesting progression of foci to adenomas. Initiated and non-initiated control mice also had
fewer foci/mouse with time. Because of differences in methodology and the lack of discernment
02/12/09
E-120
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
between foci, adenomas, and carcinomas for many of the mice exposed for 52 weeks, it is
difficult to compare differences in composition of the "lesions" after cessation of exposure. For
TCA treatment the number of animals examined for determination of which "lesions" were foci,
adenomas, and carcinomas was 11 out of the 19 mice with "lesions" at 52 weeks while all 4 mice
with lesions after 37 weeks of exposure and 15 weeks of cessation were examined. For DCA
treatment the number of animals examined was only 10 out of 23 mice with "lesions" at 52
weeks while all 7 mice with lesions after 37 weeks of exposure and 15 weeks of cessation were
examined. Most importantly, when lesions were examined microscopically then did not all turn
out to be preneoplastic or neoplastic. Two lesions appeared "to be histologically normal" and
one necrotic. Not only were a smaller number of animals examined for the cessation exposure
than continuous exposure but only the 2 g/1 exposure levels of DCA and TCA were studied for
cessation. The number of animals bearing "lesions" at 37 and then 15 week cessation weeks was
7/11 (64%) while the number of animals bearing lesions at 52 weeks was 23/24 (96%) after 2 g/1
DCA exposure. For TCA the number of animals bearing lesions at 37 weeks and then 15 weeks
cessation was 4/11 (35%) while the number of animals bearing lesions at 52 weeks was 19/24
(80%>). While suggesting that cessation of exposure diminished the number of "lesions,"
conclusions regarding the identity and progression of those lesion with continuous vs. non-
continuous DCA and TCA treatment are tenuous.
Macroscopically, the "livers of many mice receiving DCA in their drinking water
displayed light colored streaks on the surface" at every sacrifice period and "corresponded with
multi-focal areas of necrosis with frequent infiltration of lymphocytes." At the light microscopic
level, the lesions were described to also be present in the interior of the liver as well. For TCA-
treated mice, "similar necrotic lesions were also observed... but at a much lower frequency,
making it difficult to determine if they were treatment-related." Control animals were reported
not to show degenerative changes. "Marked cytomegaly" was reported for mice treated with
either 1 or 2 g/1 DCA "throughout the liver" In regard to cell size the authors did not give any
description in the methods section of the paper as to how sections were selected for
morphometric analysis or what areas of the liver acinus were examined but reported after 52
weeks of treatment the long axis of hepatocytes measured (mean ± S.E.) 24.9 ± 0.3, 38.5 ± 1.0,
and 29.3 ±1.4 [j,m in control, DCA- and TCA- treated mice, respectively.
Mice treated with TCA (2 g/1) for 52 weeks were reported to have livers with
"considerable dose-related accumulations of lipofuscin " However, no quantitative analyses
were presented. A series of figures representative of treatment showed photographs (1000X) of
lipofuscin fluorescence indicating greater fluorescence in TCA treated liver than control or DCA
treated liver.
A series of photographs of H&E sections in the report (Figures 2 a, b and c) are shown as
representative histology of control mice, mice treated with 2 g/1 DCA and 2 g/1 TCA. The area
of the liver from which the photographs were taken did not include either portal tract or central
veins and the authors did not give the zone of the livers from which they were taken. The figure
representing TCA treatment shows only a mild increase in cell volume in comparison to controls,
02/12/09
E-121
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
while for DCA treatment the hepatocyte diameter was greatly enlarged, pale stained so that
cytoplasmic contents appear absent, nuclei often pushed to the cell perimeter, and the sinusoids
appearing to be obscured by the swollen hepatocytes. The apparent reduction of sinusoidal
volume by the enlarged hepatocytes raises the possibility of decreased blood flow through the
liver which may have been linked to focal areas of necrosis reported for this high exposure level.
In a second set of figures Glycogen accumulation was shown with PAS staining at the same level
of power (400X) for the same animals. In control animals PAS positive material was not
uniformly distributed between or within hepatocytes but send to show a zonal pattern of
moderate intensity. PAS positive staining (which the authors reported to be glycogen) appeared
to be slightly less than controls but with a similar pattern in the photograph representing TCA
exposure. However, for DCA the photograph showed a uniform and heavy stain within each
hepatocyte and across all hepatocytes. The authors stated in the results section of the paper that
"the livers of TC A-treated animals displayed less evidence of glycogen accumulation and it was
more prominent in periportal than centrilobular portions of the liver acinus." In their abstract
they state "TCA produced small increases in cell size and a much more modest accumulation of
glycogen." Thus, the statement in the text, which is suggestive that TCA induced an increase in
glycogen over controls that was not as much as that induced by DCA, and the statement in the
abstract which concludes TCA exposure increased glycogen is not consistent with the
photographs. In the photograph shown for TCA there is less not more PAS positive staining
associated with TCA treatment in comparison to controls. In Sanchez and Bull (1990) the
authors report that "TCA exposure induced a much less intense level of PAS staining that was
confined to periportal areas" but do not compare PAS staining to controls but only to DCA
treatment. In the discussion section of the paper the authors state "Except for a small increase in
liver weight and cell size, the effects produced by DCA were not observed with TCA." Thus,
there seems to be a discrepancy with regard to what the effects of TCA are in relation to control
animals from this report.that has caused confusion in the literature. Kato-Weinstein et al (2001)
reported that in male mice exposed to DCA and TCA the DCA increased glycogen and TCA
decreased glycogen content of the liver using chemical measurement of glycogen in liver
homogenates and using ethanol-fixed sections stained with PAS, a procedure designed to
minimize glycogen loss).
2.3.2.5. Nelson et al. 1990
Nelson et al (1990) reported that they used the same exposure paradigm as Herren-
Freund et al. (1987), with little description of methods used in treatment of the animals. Male
B6C3F1 mice were reported to be exposed to DCA (1 or 2 g/1) or TCA (1 or 2 g/1) for 52 weeks.
The number of animals examined for non-tumor tissue was 12 for controls. The number of
animals varied from 2 to 8 for examination of non-tumor tissue, hyperplastic nodules, and
carcinoma tissues for c-Myc expression. There was no description for how hyperplastic nodules
were defined and whether they included adenomas and foci. For the 52-week experiments, the
results were pooled for lesions that had been obtained by exposure to the higher or lower
concentrations of DCA or TCA (i.e., the TCA results are for lesions induced by either 1.0 g/1 or
2.0 g/1 TCA). A second group of mice were reported to be given either DCA or TCA for 37
02/12/09
E-122
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
weeks and then normal drinking water for the remaining time till 52 weeks with no
concentrations given for the exposures to these animals. Therefore it is impossible to discern
what dose was used for tumors analyzed for c-Myc expression in the 37-week treatment groups
and if the same dose was used for 37 and 52 week results. Autoradiography was described for 3
different sections per animal in 5 different randomly chosen high power fields per section. The
number of hyperplastic nodules or the number of carcinomas per animal induced by these
treatments was not reported nor the criteria for selection of lesions for c-myc expression.
Apparently a second experiment was performed to determine the expression of c-H-ras.
Whereas in the first experiment there were no hyperplastic nodules, in the second 1 control
animal was reported to have a hyperplastic nodule. The number of control animals reported to
be examined for non-tumor tissue in the second group was 12. The numbers of animals in the
second group was reported to vary from 1 to 7 for examination of non-tumor tissue, hyperplastic
nodules, and carcinoma tissues for c-H-ras expression. The number of animals per group for the
investigation of H-ras did not match the numbers reported for that of c-Myc. The number of
animals treated to obtain the "lesion" results was not presented (i.e., how many animals were
tested to get a specific number of animals with tumors that were then examined). The number of
lesions assessed per animal was not reported.
At 52 weeks of exposure, hyperplastic nodules (n= 8 animals) and carcinomas (n= 6
animals) were reported to have ~ 2-fold expression of c-Myc relative to non-tumor tissue (n= 6
animals) after DCA treatment. After 37 weeks of DCA treatment and cessation of exposure,
there was a ~ 30% increase in c-Myc in hyperplastic nodules (n=4 animals) that was not
statistically significant. There were no carcinomas reported at this time. After 52 weeks of TCA
exposure, there was ~ 2-fold of non-tumor tissue reported for c-Myc in hyperplastic nodules (n=
6 animals) and ~ 3-fold reported for carcinomas (n=6 animals). After 37 weeks of TCA
exposure there was ~ 2-fold c-Myc in hyperplastic nodules (n= 2 animals) that was not
statistically significant and ~ 2.6 fold increase in carcinomas (n=3 animals) that was reported to
be statistically significant over non-tumor tissue. There was no difference in c-Myc expression
between untreated animals and non-tumor tissue in the treated animals.
The authors reported that c-Myc expression in TCA-induced carcinomas was "almost 6
times that in control tissue (corrected by subtracting nonspecific binding)," and concluded that c-
Myc in TCA-induced carcinomas was significantly greater than in hyperplastic nodules or
carcinomas and hyperplastic nodules induced by DCA. However the c-myc expression reported
as the number of grains per cells was ~ 2.6-fold in TCA-induced carcinomas and ~ 2-fold in
DCA-induced carcinomas than control or non-tumor tissue at 52 weeks. The hyperplastic
nodules from DCA- and TCA-treatments at 52 weeks gave identical ratios of - 2-fold In 3
animals per treatment, c-Myc expression was reported to be similar in "selected areas of high
expression" for either DCA or TCA treatments of 52 weeks.
There did not appear to be a difference in c-H-ras expression between control and non-
tumor tissue from DCA- or TCA- treated mice. The levels of c-H-ras transcripts were reported
to be "slightly elevated" in hyperplastic nodules induced by DCA (~ 67%) or TCA (~ 43%) but
02/12/09
E-123
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
these elevations were not statistically significant in comparison to controls. However,
carcinomas "derived from either DCA- or TCA-treated animals were reported to have
significantly increased c-H-ras levels relative to controls." The fold increase of non-tumor tissue
at 52 weeks for DCA-induced carcinomas was ~2.5-fold and for TCA induced carcinomas -2.0
fold. Again the authors state that "if corrected for non-specific hybridization, carcinomas
expressed approximately 4 times as much c-H-ras than observed in surrounding tissues" Given
that control and non-tumor tissue results were given as the controls for the expression increases
observed in "lesions", it is unclear what this the usefulness of this "correction" is. The authors
reported that "focal areas of increased expression of c-H-ras were not observed within
carcinomas."
The limitations of this experiment include uncertainty as to what doses were used and
how many animals were exposed to produce animals with derive tumors. In addition results of
differing doses were pooled and the term hyperplastic nodule, undefined. The authors state that
c-Myc expression in itself is not sufficient for transformation and that its over expression
commonly occurs in malignancy. They also state that "Unfortunately, the limited amount of
tissue available prevented a more serious pursuit of this question in the present study." In regard
to the effects of cessation of exposure, the authors do not present data on how many animals
were tested with the cessation protocol, what doses were used, and how many lesions comprised
their results and thus comparisons between these results and those from 52 weeks of continuous
exposure are hard to make. Quantitatively, the small number of animals, whose lesions were
tested, was 2-4 for the cessation groups. Bull et al. (1990) is given as the source of data for the
cessation experiment (see Section 2.3.2.1 above).
2.3.2.6. DeAngelo et al. 1999
The focus of this study was to "determine a dose response for the hepatocarcinogenicity
of DCA in male mice over a lifetime exposure and to examined several modes of action that
might underlie the carcinogenic process." As DeAngelo et al pointed out, many studies of DCA
had been conducted at high concentrations and for less than lifetime studies, and therefore of
suspect relevence to environmental concentrations. This study is one of the few that examined
DCA at a range of exposure concentrations to determine a dose-response in mice. The authors
concluded that DCA-induced carcinogenesis was not dependent on peroxisome proliferation or
chemically sustained proliferation. The number of hepatocellular carcinomas/animals was
reported to be significantly increased over controls at all DCA treatments including 0.05 g/1 and
a NOEL not observed. Peroxisome proliferation was reported to be significantly increased at 3.5
g/L DCA only at 26 weeks and did not correlate with tumor response. No significant treatment
effects on labeling of hepatocytes (as a measure of proliferation) outside proliferative lesions
were also reported and thus that DCA-induced liver cancer was not dependent on peroxisome
proliferation or chemically sustained cell proliferation.
Male B6C3F1 mice were 28-30 days of age at the start of study and weighed 18-21 grams
(or ~ 14% range). They were exposed to 0, 0.05, 0.5, 1.0, 2.0, and 3.5 g/1 DCA via drinking
02/12/09
E-124
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
water as a neutralized solution. The time-weighted mean daily water consumption calculated
over the 100-week treatment period was reported to be 147, 153, 158, 151, 147, and 124 (84% of
controls) ml/kg/day for 0, 0.05, 0.5, 1, 2, and 3.5 g/1 DCA, respectively. The number of animals
reported as used for interim sacrifices were 35, 30, 30, 30 and 30 for controls, 0.5 g/1, 1.0 g/1, 2.0
g/1, and 3.5 g/1 DCA treated groups respectively (i.e., 10 mice per treatment group at interim
sacrifices of 26, 52 and 78 weeks). The number of animals at final sacrifice were reported to be
50, 33, 24, 32, 14 and 8 for controls, 0.05g/l, 0.5 g/1, 1.0 g/1, 2.0 g/1, and 3.5 g/1 DCA treated
groups respectively The number of animals with unscheduled deaths before final sacrifice were
reported to be 3, 2, 1,9, 11 and 8 for controls, 0.05g/l, 0.5 g/1, 1.0 g/1, 2.0 g/1, and 3.5 g/1 DCA
treated groups respectively. The Authors reported that early mortality tended to occur from liver
cancer. The number of animals examined for pathology were reported to be 85, 33, 55, 65, 51
and 41 for controls, 0.05g/l, 0.5 g/1, 1.0 g/1, 2.0 g/1, and 3.5 g/1 DCA treated groups respectively
The experiment was conducted in two parts with control, 0.5 g/1, 1.0 g/1, 2.0 g/1 and 3.5 g/1
groups treated and then 1 months later a second group consisting of 30 control group mice and
35 mice in a 0.05 g/1 DCA exposure group studied. The authors reported not difference in
prevalence and multiplicity of hepatocellular neoplasms in the two groups so that data were
summed and reported together. The # of animals reported as examined for tumors were n=10
animals, with controls reported to be 35 animals split among 3 interim sacrifice times - exact
number per sacrifice time is unknown. The number of animals reported "with pathology" and
assumed to be included in the tumor analyses from Table 1, and the sum of the number of
animals "scheduled for sacrifice that survived till 100 weeks" and "interim sacrifices" do not
equal each other. For the 1 g/1 DCA exposure group, 30 animals were sacrificed at interim
periods, 32 animals were sacrificed at 100 weeks, 9 animals were reported to have unscheduled
deaths, but of those 71 animals only 65 animals were reported to have pathology for the group.
Therefore some portion of animals with unscheduled deaths must have been included in the
tumor analyses. The exact number of animals that may have died prematurely but included in
analyses of pathology for the 100 week group is unknown. In Figure 3 of the study, the authors
reported prevalence and multiplicity of hepatocellular carcinomas following 79 to 100 weeks of
DCA exposure in their drinking water. The number of animals in each dose group used in the
tumor analysis for 100 weeks was not given by the authors. Given that the authors included
animals that survived past the 78 interim sacrifice period but died unscheduled deaths in their
100 week results, the number must have been greater than those reported as present at final
sacrifice. A comparison of the data for the 100-week data presented in Table 3a and Figure 3
shows that the data reported for 100 weeks is actually for animals that survived from 79 to 100
weeks. The authors report a dose-response that is statistically significant from 0.5 to 3.5 g/1
DCA for hepatocellular carcinoma incidence and a dose-response in hepatocellular carcinoma
multiplicity that is significantly increased over controls from 0.05 to 0.5 g/1 DCA that survived
79 to 100 weeks of exposure (i.e., 0, 8-, 84-, 168-, 315- and 429 mg/kg/day dose groups with
prevalences of 26%, 33%, 48%, 71%, 95%, and 100%), respectively, and multiplicities of 0.28,
0.58, 0.68, 1.29, 2.47, and 2.90, respectively). Hepatocellular adenoma incidence or multiplicity
was not reported for the 0.05 g/1 DCA exposure group.
02/12/09
E-125
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
In Table 3 of the report, the time course of hepatocellular carcinomas and adenoma
development are given and summarized below.
Prevalence
52 weeks control = 0% carcinomas, 0% adenoma
0.5 g/L DC A =0/10 carcinoma, 1/10 adenomas
1.0 g/L DC A = 0/10 carcinomas, 1/10 adenomas
2.0 g/L DC A = 2/10 carcinomas, 0/10 adenomas
3.5 g/L DC A =5/10 carcinomas, 5/10 adenomas
78 weeks control = 10% carcinomas, 10% adenomas
0.5 g/L DC A = 0/10carcinoma, 1/10 adenomas
1.0 g/L DC A = 2/10 carcinomas, 2/10 adenomas
2.0 g/L DC A =5/10 carcinomas, 5/10 adenomas
3.5 g/L DC A = 7/10 carcinomas, 5/10 adenomas
100 weeks control = 26% carcinoma, 10% adenoma
0.5 g/L DCA = 48%) carcinoma, 20% adenomas
1.0 g/L DCA = 71%) carcinomas, 51.4% adenomas
2.0 g/L DCA = 95% carcinomas, 42.9% adenomas
3.5 g/L DCA = 100%) carcinomas, 45% adenomas
Multiplicity
(lesions/animal m± SEM)
Carcinomas
Adenomas
0
0
0
0.10 ±0.09
0
0.10 ±0.09
0.20 ±0.13
0
0.70 ±0.25
0.80 ±0.31
0.10±0.10
0.10 ±0.09
0
0.10 ±0.09
0.20 ±0.13
0.20 ±0.13
1.0 ±0.47
1.00 ± 0.42
1.20 ±0.37
1.00 ±0.42
0.28 ±0.07
0.12 ±0.05
0. 68 ±0.17
0.32 ±0.14
1.29 ±0.17
0.80 ±0.17
2.47 ± 0.29
0.57 ±0.16
2.90 ±0.40
0.64 ± 0.23
The authors reported hepatocellular carcinomas and number of lesions/animal in mice
that survived 79-100 wk of exposure (they combined exposure groups to be animals after the
week 78 sacrifice time that did and did not make it to 100 weeks). This is the same data reported
above for the 100 week exposure with the inclusion of the 0.05 g/1 DCA data. The difference
between number of animals at interim and final sacrifices and those "with pathology" and used
in the tumor analysis but most likely coming from unscheduled deaths is reported below as
"extra" and varied across treatment groups.
Dose = Prevalence of HC #HC/animal N = atl00weeks extra added in
Control = 26% 0.28 50 0
0.05 g/L = 33% 0.58 33 0
0.5 g/L = 48% 0.68 24 1
1 g/L= 71% 1.29 32 3
2 g/L= 95% 2.47 14 7
3.5 g/L =100% 2.90 8 3
These data show a dose-related increase in tumor formation and decrease in time-to-
tumor associated with DCA exposure at the lowest levels examined. These findings are limited
02/12/09
E-126
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
by the small number of animals examined at 100 weeks but especially those examined at
"interim sacrifice" periods (n= 10). The data illustrate the importance of examining multiple
exposure levels at lower concentrations at longer durations of exposure and with an adequate
number of animals to determine the nature of a carcinogenic response.
Preneoplastic and non-neoplastic hepatic changes were reported to have been described
previously and summarized as large preneoplastic foci observed at 52 weeks with multiplicities
of 0.1, 0.1, 0.2 and 0.16 for 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA exposure respectively. At 100
weeks all values were reported to be significant (0.03, 0.06, 0.14, 0.27 for 0.5 g/1, 1 g/1, 2 g/1, and
3.5 g/1 DCA exposure respectively. Control values were not reported by the authors. The
authors reported that the prevalence and severity of hepatocellular cytomegaly and of
cytoplasmic vacuolization with glycogen deposition to be dose-related and considered significant
in all dose groups examined when compared to control liver. However, no quantitative data was
shown. The authors reported a severity index of 0 = none, 1 = < 25%, 2 = 50-75% and 4 = 75%
of liver section for hepatocellular necrosis and report at 26 weeks scores (n= 10 animals) of 0.10
±0.10, 0.20 ±0.13, 1.20 ±0.38, 1.20 ± 0.39 and 1.10 ± 0.28 for control, 0.5g/l, 1 g/1, 2 g/1 and
3.5 g/1 DCA treatment groups, respectively. Thus there appeared to be a treatment but not dose-
related increase in hepatocellular necrosis that is does not involve most of the liver from 1 g/1 to
3.5 g/1 DCA at this time point. At 52 weeks of exposure the score for hepatocellular necrosis
was reported to be 0, 0, 0.20 ± 0.13, 0.40 ± 0.22 and 1.10 ± 0.43 for control, 0.5g/l, 1 g/1, 2 g/1
and 3.5 g/1 DCA treatment groups, respectively. At 78 weeks of exposure the score for
hepatocellular necrosis was reported to be 0, 0, 0, 0.30 ± 0.21 and 0.20 ± 0.13 for control, 0.5g/l,
1 g/1, 2 g/1 and 3.5 g/1 DCA treatment groups, respectively. Finally, the final sacrifice time when
more animals were examined the extent of hepatocellular necrosis was reported to be 0.20 ±
0.16, 0.20 ± 0.08, 0.42 ±0.15, 0.38 ± 0.20 and 1.38 ± 0.42 for control, 0.5g/l, 1 g/1, 2 g/1 and 3.5
g/1 DCA treatment groups, respectively. Thus, there was not reported increase in hepatocellular
necrosis at any exposure period for 0.5g/l DCA treatment and the mild hepatocellular necrosis
seen at the three highest exposure concentrations at 26 weeks had diminished with further
treatment except for the highest dose at up to 100 weeks of treatment. Clearly the pattern of
hepatocellular necrosis did not correlate with the dose-related increases in hepatocellular
carcinomas reported by the authors and was not increased over control at the 0.5 g/1 DCA level
where there was a DCA-related tumor increase.
The authors cite previously published data and state that cyanamid -insensitive
palmitoyl coenzyme A (CoA) oxidase activity (a marker of peroxisome proliferation) data for the
26 week time point plotted against 100 wk hepatocellular carcinoma prevalence of animals
bearing tumors was significantly enhanced at concentrations of DCA that failed to induce
"hepatic PCO" activity. The authors report that neither 0.05 nor 0.5 g/1 DCA had any marked
effect on PCO activity and that it was "only significantly increased after 26 weeks of exposure to
3.5 g/1 DCA and returned to control level at 52 weeks (data not shown)." In regards to
hepatocyte labeling index after treatment for 5 days with tritiated thymidine, the authors report
that animals examined in the dose-response segment of the experiment at 26 and 52 weeks were
examined but no details of the analysis were reported. The authors comment on the results from
02/12/09
E-127
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
this study and a previous one that included earlier time points of study and stated that there were
"no significant alterations in the labeling indexes for hepatocytes outside of proliferative lesions
at any of the DCA concentrations when compared to the control values with the exception of
0.05 g/1 DCA at 4 weeks (4.8 ± 0.6 vs. 2.7 ± 0.4 control value; data not shown)."
The effects of DCA on body weight, absolute liver weight and % liver/body weight were
given in Table 2 of the paper for 26, 52, 78 and 100 weeks exposure. For 52 and 78 week studies
10 animals per treatment group were examined. Liver weights were not determined for the
lowest exposure concentration (0.05 g/1 DCA) except for the 100 week exposure period. At 26
weeks of exposure there was not a statistically significant change in body weight among the
exposure groups (i.e., 35.4 ± 0.7, 37.0 ± 0.8, 36.8 ± 0.8, 37.9 ± 0.6, and 34.6 ± 0.8 g for control,
0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). Absolute liver weight was reported to have a
dose-related significant increase in comparison to controls at all exposure concentrations
examined with liver weight reaching a plateau at the 2 g/1 concentration (i.e., 1.86 ± 0.07, 2.27 ±
0.10, 2.74 ± 0.08, 3.53 ± 0.07, and 3.55 ± 0.1 g for control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA,
respectively). The % liver/body weight ratio increases due to DCA exposure were reported to
have a similar pattern of increase (i.e., 5.25 ±0.11, 6.12 ± 0.16, 7.44 ± 0.12, 9.29 ± 0.08, and
10.24 ± 0.12 % for control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). This represented
a 1.17-fold, 1.41-fold, 1.77-fold, and 1.95-fold of control % liver/body weight at these exposures
at 26 weeks.
At 52 weeks of exposure there was not a statistically significant change in body weight
among the exposure groups except for the 3.5 g/1 exposed group in which there was a significant
decrease in body weight (i.e., 39.9 ± 0.8, 41.7 ± 0.8, 41.7 ± 0.9, 40.8 ± 1.0, and 35.0 ± 1.1 g for
control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). Absolute liver weight was reported
to have a dose-related significant increase in comparison to controls at all exposure
concentrations examined with liver weight reaching a plateau at the 2 g/1 concentration (i.e., 1.87
± 0.13, 2.39 ± 0.04, 2.92 ± 0.12, 3.47 ± 0.13, and 3.25 ± 0.24 g for control, 0.5 g/1, 1 g/1, 2 g/1,
and 3.5 g/1 DCA, respectively). The % liver/body weight ratio increases due to DCA exposure
were reported to have a similar pattern of increase (i.e., 4.68 ± 0.30, 5.76 ± 0.12, 7.00 ± 0.15,
8.50 ± 0.26, and 9.28 ± 0.64 % for control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively).
For liver weight and % liver/body weight there was much larger variability between animals
within the treatment groups compared to controls and other treatment groups. There were no
differences reported for patterns of change in body weight, absolute liver weight, and %
liver/body weight between animals examined at 26 weeks and those examined at 52 weeks. At
78 weeks of exposure there was not a statistically significant change in body weight among the
exposure groups except for the 3.5 g/1 exposed group in which there was a significant decrease in
body weight (i.e., 46.7 ± 1.2, 43.8 ± 1.5, 43.4 ± 0.9, 42.3 ± 0.8, and 40.2 ± 2.2 g for control, 0.5
g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). Absolute liver weight was reported to have a
dose-related increase in comparison to controls at all exposure concentrations examined but none
were reported to be statistically significant (i.e., 2.55 ± 0.14, 2.16 ± 0.09, 2.54 ± 0.36, 3.31 ±
0.63, and 3.93 ± 0.59 g for control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). The %
liver/body weight ratio increases due to DCA exposure were reported to have a similar pattern of
02/12/09
E-128
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
increase over control values but only the 3.5 g/1 exposure level was reported to be statistically
significant (i.e., 5.50 ± 0.35, 4.93 ± 0.09, 5.93 ± 0.97, 7.90 ± 1.55, and 10.14 ± 1.73 % for
control, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DC A, respectively). Finally, for the animals reported to
be sacrificed between 90 and 100 weeks there was not a statistically significant change in body
weight among the exposure groups except for the 2.0 and 3.5 g/1 exposed groups in which there
was a significant decrease in body weight (i.e., 43.9 ± 0.8, 43.3 ± 0.9, 42.1 ± 0.9, 43.6 ± 0.7, 36.1
± 1.2 g, and 36.0 ± 1.3 g for control, 0.05 g/1, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively).
Absolute liver weight did not show a dose-response pattern at the two lowest exposure levels but
was elevated with the 3 highest doses with the two highest being statistically significant (i.e.,
2.59 ± 0.26, 2.74 ± 0.20, 2.51 ± 0.24, 3.29 ± 0.21, 4.75 ± 0.59 and 5.52 ± 0.68 g for control, 0.05
g/1, 0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively). The % liver/body weight ratio increases
due to DCA exposure were reported to have a similar pattern of increase over control values but
only the 2.0 and 3.5 g/1 exposure levels were reported to be statistically significant (i.e., 6.03 ±
0.73, 6.52 ± 0.55, 6.07 ± 0.66, 7.65 ± 0.55, 13.30 ± 1.62 and 15.70 ± 2.16 % for control, 0.05 g/1,
0.5 g/1, 1 g/1, 2 g/1, and 3.5 g/1 DCA, respectively).
It must be recognized that liver weight increases, especially in older mice, will reflect
increased weight due to tumor burden and thus DCA-induced hepatomegaly will be somewhat
obscured at the longer treatment durations. However, by 100 weeks of exposure there did not
appear to be an increase in liver weight at the 0.05 g/1 and 0.5 g/1 exposures while there was an
increase in tumor burden reported. Examination of the 0.5 g/1 exposure group from 26 to 100
weeks shows that slight hepatomegaly, reported as either absolute liver weight increase over
control or change in % liver/body ratio, was present by 26 weeks (i.e., 22% increase in liver
weight and 17% increase in %liver/body weight), decreased with time, and while similar at 52
weeks, was not significantly different from control values at 78 or 100 weeks durations of
exposure. However, tumor burden was increased at this low concentration of DCA.
The authors present a figure comparing the number of hepatocellular carcinomas per
animal at 100 weeks compared with the % liver/body weight at 26 weeks and show a linear
correlation (r2 = 0.9977). Peroxisome proliferation and DNA synthesis, as measured by tritiated
thymidine, were reported to not correlate with tumor induction profiles and were also not
correlated with early liver weight changes induced by DCA exposure. Most importantly, in a
paradigm that examined tumor formation after up to 100 weeks of exposure, DCA-induced
tumor formation was reported to occur at concentrations that did not also cause cytotoxicity and
at levels 20 to 40 times lower than those used in "less than lifetime" studies reporting concurrent
cytotoxicity.
2.3.2.7. Carter et al. 2003
The focus of this study was to present histopathological analyses that included
classification, quantification and statistical analyses of hepatic lesions in male B6C3F1 mice
receiving DCA at doses as low as 0.05g/l for 100 weeks and at 0.5, 1.0, 2.0, and 3.5 g/1 for
between 26 and 100 weeks. This analysis used tissues from the DeAngelo et al.. (1999) (two
02/12/09
E-129
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
blocks from each lobe and all lesions found at autopsy). This study used the following
diagnostic criteria for hepatocellular changes. Altered hepatic Foci (AHF) were defined as
histologically identifiable clones that were groups of cells smaller than a liver lobule that did not
compress the adjacent liver. Large Foci of Cellular Alteration (LFCA) were defined as lesions
larger than the liver lobule that did not compress the adjacent architecture (previously referred to
as hyperplastic nodules by Bull et al., 1990) but had different staining. These are not non-
neoplastic proliferative lesions termed "hepatocellular hyperplasia" that occur secondary to
hepatic degeneration or necrosis. Adenomas (ADs) showed growth by expansion resulting in
displacement of portal triad and had alterations in both liver architecture and staining
characteristics. Carcinomas (CAs) were composed of cells with a high nuclear-to-cytoplasmic
ration and with nuclear pleomorphism and atypia that showed evidence of invasion into the
adjacent tissue. They frequently showed a trabecular pattern characteristic of mouse
hepatocellular CAs.
The report grouped lesions as eosinophilic, basophilic and/or clear cell, and dysplastic.
"Eosinophilic lesions included lesions that were eosinophilic but could also have clear cell,
spindle cell or hyaline cells. Basophilic lesions were grouped with clear cell and mixed cell (i.e.,
mixed basophilic, eosinophilic, hyaline, and/or clear cell) lesions." The authors reported that
"this grouping was necessary because many lesions had both a basophilic and clear cell
component and a few <10 % had an eosinophilic or hyaline component." "Lesions with foci of
cells displaying nuclear pleomorphism, hyperchromasia, prominent nucleoli, irregular nuclear
borders and/or altered nuclear to cytoplasmic ratios were considered dysplastic irrespective of
their tinctorial characteristics." Therefore, Carter et al. (2003) lumped mixed phenotype lesions
into the basophilic grouping so that comparisons with the results of Bull et al. (2002) or Pereira
(1996), which segregate mixed phenotype from those without mixed phenotype, cannot be done.
This report examined type and phenotype of preneoplastic and neoplastic lesions pooled
across all time points. Therefore, conclusions regarding what lesions were evolving into other
lesions have left out the factor of time. Bannasch (1996) reported that examining the evolution
of foci through time is critical for discerning neoplastic progression and described foci evolution
from eosinophilic or basophilic lesions to more basophilic lesions. Carter et al. (2003) suggest
that size and evolution into a more malignant state are associated with increasing basophilia, a
conclusion consistent with those of Bannasch (1996). The analysis presented by Carter et al
(2003) also suggested that there was more involvement of lesions in the portal triad which may
give an indication where the lesions arose. Consistent with the results of DeAngelo et al (1999),
Carter et al. (2003) reported that "DCA (0.05 - 3.5 g/1) increased the number of lesions per
animal relative to animals receiving distilled water and shortened the time to development of all
classes of hepatic lesions." They also concluded that "although this analysis could not
distinguish between spontaneously arising lesions and additional lesions of the same type
induced by DCA, only lesions of the kind that were found spontaneously in control liver were
found in increased numbers in animals receiving DCA." "Development of eosinophilic,
basophilic and/or clear cell and dysplastic AHF was significantly related to DCA dose at 100
weeks and overall adjusted for time." The authors concluded that the presence of isolated,
02/12/09 E-130
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
highly dysplastic hepatocytes in male B6C3F1 mice chronically exposed to DC A suggested
another direct neoplastic conversion pathway other than through eosinophilic or basophilic foci.
It appears that the lesions being characterized as carcinomas and adenomas in DeAngelo
et al (1999) were not the same as those by Carter et al. (2003) at 100 weeks even though they
were from the same tissues. Carter et al. identified all carcinomas as dysplastic despite tincture
of lesion and subdivided adenomas by tincture. If the differing adenoma multiplicities are
summed for Carter et al. they do not add up to the same total multiplicity of adenoma given by
DeAngelo et al. It is unclear how many animals were included in the differing groups in both
studies for pathology. The control and high-dose groups differ in respect to "animals with
pathology" between DeAngelo et al. and the "number of animals in groups" examined for lesions
in Carter et al. Neither report gave how many animals with unscheduled deaths were treated in
regards to how the pathology data was included in presentation of results. Given that DeAngelo
et al. represents animals at 100 weeks as also animals from 79 -100 weeks exposure, it is
probable that the animals that died after 79 weeks were included in the group of animals
sacrificed at 100 weeks. However, the number of animals affecting that result (which would be a
mix of exposure times) for either DeAngelo et al., or Carter et al., is unknown from published
reports. In general, it appears that Carter et al. (2003) reported more adenomas/animal for their
100 week animals than DeAngelo et al. (1999) did, while DeAngelo et al. reported more
carcinomas/animal. Carter et al. reported more adenomas/animal than controls while DeAngelo
et al. reported more carcinomas/animal than controls at 100 weeks of exposure.
In order to compare this data with others (e.g., Pereira, 1996) for estimates of multiplicity
by phenotype tincture it would be necessary to add foci and LFCA together as foci, and
adenomas and carcinomas together as tumors. It would also be necessary to lump mixed foci
together as "basophilic" from other data sets as was done for Carter et al. in describing
"basophilic lesions". If multiplicity of carcinomas and adenomas are summed from each study
to control for differences in identification between adenoma and carcinoma, there are still
differences in the two studies in multiplicity of combined lesions/animal with DeAngelo giving
consistently higher estimates. However, both studies show a dose response of tumor multiplicity
with DCA and a difference between control values and the 0.05 DCA exposure level. Error is
introduced by having to transform the data presented as a graph in Carter et al. (2003). Also no
SEM is given for the Carter data.
Comparison of data from Carter et al. (2003) and DeAngelo et al. (1999):
Exposure
Total
Total
Total
Total
Sum of
Sum of
level of
adenoma
adenoma
carcinoma
carcinoma
adenomas
adenomas
DCA at
multiplicity
multiplicity
multiplicity
multiplicity
and
and
79-100
(Carter)
(DeAngelo)
(Carter)
(De Angelo)
carcinoma
carcinoma
weeks
fe/L)
multiplicity
(Carter)
multiplicity
(DeAngelo)
0
0.22
0.12
0.05
0.28
0.27
0.40
0.05
0.48
-
<0.025
0.58
-0.50
-
02/12/09
E-131
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
0.5
0.44
0.32
0.20
0.68
0.64
1.0
1.0
0.52
0.80
0.30
1.29
0.82
2.09
2.0
0.60
0.57
1.55
2.47
2.15
3.27
3.5
1.48
0.64
1.30
2.90
2.78
3.54
In regard to other histopathological changes, the authors report that "necrosis was found
in 11.3% of animals in the study and the least prevalent toxic or adaptive response. No focal
necrosis was found at 0.5 g/L. The incidence of focal necrosis did not differ from controls at 52
or 78 weeks and only was greater than controls at the highest dose of 3.5 g/L at 100 weeks.
Overall necrosis was negatively related to the length of exposure and positively related to the
DCA dose. Necrosis was an early and transitory response. There was no difference in necrosis 0
and 0.05 g/L or 0.5 g/L. There was an increase in glycogen at 0.5 g/L at the perioportal area.
There was no increase in steatosis but a dose-related decrease in steatosis. Dysplastic LFCA
were not related to necrosis indicating that these lesions do not represent, regenerative or
reparative hyperplasia. Nuclear atypia and glycogen accumulation were associated with
dysplastic adenomas. Necrosis was not related to occurrence of dysplastic adenomas. Necrosis
was of borderline significance in relation to presence of hepatocellular carcinomas. Necrosis
was not associated with dysplastic LFCAs or Adenomas." They concluded that "the degree to
which hepatocellular necrosis underlies the carcinogenic response is not fully understood but
could be significant at higher DCA concentrations (>_1 g/L),"
2.3.2.8. Stauber and Bull 1997
This study was designed to examine the differences in phenotype between altered hepatic
foci and tumors induced by DCA and TCA. Male B6C3F1 mice (7 weeks old at the start of
treatment) were treated with 2.0 g/1 neutralized DCA or TCA in drinking water for 38 or 50
weeks, respectively. They were then treated with additional exposures (n = 12) of 0, 0.02, 0.1,
0.5, 1.0, 2.0 g/L DCA or TCA for an additional 2 weeks. Three days prior to sacrifice in DCA-
treated mice or 5 days for TCA-treated mice, animals had miniosmotic pumps implanted and
administered BrdU. Immunohistochemical staining of hepatocytes from randomly selected
fields (minimum of 2000 nuclei counter per animal) from 5 animals per group were reported for
14- and 28-day treatments. It was unclear how many animals were examined for 280- and 350-
day treatments from the reports. The percentage of labeled cells in control livers was reported to
vary between 0.1 and 0.4 % (i.e., 4-fold). There was a reported ~ 3.5-fold of control level for
TCA labeling at 14 day time period and a ~ 5.5 fold for DCA. At 28 days there was ~ 2.5 fold of
control for TCA but a ~ 2.3-fold decrease of control for DCA. At 280 days there was no data
reported for TCA but for DCA ~ 2-fold decrease in labeling of control. At 350 days there was
no data for DCA but a reported ~ 2.3 fold decrease in labeling of control with TCA. The
authors reported that the increases at day 14 for TCA and DCA exposure and the decrease at day
28 for DCA exposure were statistically significant although a small number of animals were
examined. Thus, although there may be some uncertainty in the exact magnitude of change,
there was at most ~5-fold of control labeling for DCA within after 14 days of exposure that was
02/12/09
E-132
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
followed by a decrease in DNA synthesis by day 28 of treatment. These data show that
hepatocytes undergoing DNA synthesis represented a small population of hepatocytes with the
highest level with either treatment less than 1% of hepatocytes. Rates of cell division were
reported to be less than control for both DCA and TCA by 40 and 52 weeks of treatment.
In this study the authors reported that there was no necrosis with the 2.0 g/L DCA dose
for 52 weeks and conclude that necrosis is a recurring but inconsistent result with chronic DCA
treatment. Histological examination of the livers involved in the present study found little or no
evidence of such damage or overt cytotoxicity. It was assumed that this effect has little bearing
on data on replication rates. Foci and tumors were combined in reported results and therefore
cannot be compared the results Bull et al (2002) or to DeAngelo et al. (1999). Prevalence rates
were not reported. Data were reported in terms of "lesions" with DCA-induced "lesions"
containing a number of smaller lesions that were heterogeneous and more eosinophilic with
larger "lesions" tending to less numerous and more basophilic. For TCA results using this
paradigm, the "lesions" were reported to be less numerous, more basophilic, and larger than
those induced by DCA. The DCA-induced larger "lesions" were reported to be more
"uniformly reactive to c-Jun and c-Fos but many nuclei within the lesions displaying little
reactivity to c-Jun. The authors stated that while most DCA-induced "lesions" were
homogeneously immunoreactive to c-Jen and C-Fos (28/41 lesions), the rest were stained
heterogeneously. For TCA-induced lesions, the authors reported not difference in staining
between "lesions" and normal hepatocytes in TCA-treated animals. Again, of note is that not
only were "lesions" comprised of foci and tumors at different stages of progression reported in
these results, but that also DCA and TCA results were reported for different durations of
exposure.
2.3.2.9. Pereira 1996
The focus of this study was to report the dose-response relationship for the carcinogenic
activity of DCA and TCA in female B6C3F1 mice and the characteristics of the lesions.
Female B6C3F1 mice (7-8 weeks of age) were given drinking water with either DCA or TCA at
2.0, 6.67, or 20 mmol/L and neutralized with sodium hydroxide to a pH or 6.5-7.5. The control
received 20 mmol/L sodium chloride. Conversion of mmol/L to g/L was as follows: 20.0
mmol/L DCA = 2.58 g/L, 6.67 mmol/L DCA = 0.86 g/L, 2.0 mmol/L = 0.26 g/L, 20.0 mmol/L
TCA = 3.27 g/L, 6.67 mmol/L TCA =1.10 g/L, 2.0 mmol/L TCA = 0.33 g/L. The
concentrations were reported to be chosen so that the high concentration was comparable to
those previously used by us to demonstrate carcinogenic activity. The mice were exposed till
sacrifice at 360 (51 weeks), or 576 days (82 weeks) of exposure. Whole liver was reported to be
cut into ~ 3 mm blocks and along with representative section of the visible lesions fixed and
embedded in paraffin and stained with H&E for histopathological evaluation of foci of altered
hepatocytes, hepatocellular adenomas, and hepatocellular carcinomas. The slides were reported
to be evaluated blind. Foci of altered hepatocytes in this study were defined as containing 6 or
more cells and hepatocellular adenomas were distinguished from foci by the occurrence of
compression at greater than 80% of the border of the lesion.
02/12/09
E-133
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Body weights were reported to be decreased only the highest dose of DC A from 40
weeks of treatment onward. For TCA there were only 2 examination periods (weeks 51 and 82)
that had significantly different body weights from control and only at the highest dose.
Liver/body weight percentage was reported in comparison to concentration graphically and
shows a dose-response for DCA with steeper slope than that of TCA at 360 and 576 days of
exposure. The authors report that all three concentrations of DCA resulted in increased
vacuolation of hepatocytes.(probably due to glycogen removal from tissue processing). Using a
score of 1 - 3, (with 0 indicating the absence of vacuolization, +1 indicating vacuolated
hepatocytes in the periportal zone, + 2 indicating distribution of vacuolated hepatocytes in the
midzone, and +3 indicating maximum vacuolization of hepatocytes throughout the liver), the
authors also reported "the extent of vacuolization of the hepatocytes in the mice administered 0,
2.0, 6.67 or 20.0 mmol/1 DCA was scored as 0.0, 0.80 ± 0.08, 2.32 ± 0.11, or 2.95 ± 0.05,
respectively."
Cell proliferation was reported to be determined in treatment groups containing 10 mice
each and exposed to either DCA or TCA for 5, 12, or 33 days with animals implanted with
miniosmotic pumps 5 days prior to sacrifice and administered BrdU. Tissues were
immunohistochemically stained for BrdU incorporation. At least 2000 hepatocytes/mouse were
reported to be evaluated for BrdU-labeled and unlabeled nuclei and the BrDU-labeling index was
calculated as the percentage of hepatocytes with labeled nuclei. Pereira (1996) reported a dose-
related increase in BrDU labeling in 2000 hepatocytes that was statistically significant at 6.67
and 20.mmol/L DCA at 5 days of treatment but that labeling at all exposure concentrations
decreased to control levels by day 12 and 33 of treatment. The largest increase in BrdU labeling
was reported to be a 2-fold of controls at the highest concentration of DCA after 5 days of
exposure. For TCA all doses (2.0, 6.67 and 20 mmol/L) gave a similar and statistically
significant increase in BrDU labeling by 5 days of treatment (~ 3-fold of controls) but by days 12
and 33 there were no increases above control values at any exposure level. Given the low level
of hepatocyte DNA synthesis in quiescent control liver, these results indicate a small number of
hepatocytes underwent increased DNA synthesis after DCA or TCA treatment and that by 12
days of treatment these levels were similar to control levels in female B6C3F1 mice.
Incidence of foci and tumors in mice administered DCA or TCA (prevalence or number
of animals with tumors of those examined at sacrifice) in this report are given below:
Prevalence:
Treatment
at 82 weeks
N Foci
number
%
adenomas
number
%
carcinomas
number %
20.0 mmol NACL 90 10
20.0 mmol DCA 19 17
6.67 mmol DCA 28 11
2.0 mmol DCA 50 7
11.1 2
89.5* 16
39.3* 7
14.0 3
2.2 2
84.2* 5
25.0* 1
6.0 0
2.2
26.3*
3.6
0
02/12/09
E-134
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
20.0 mmol TCA
18
11
61.1*
7 38.9*
5 27.8'
6.67 mmol TCA
27
9
33.3*
3 11.1
5 18.5
2.0 mmol TCA
53
10
18.9
4 7.6
0 0
*p <0.05
Treatment at
51 weeks
20.0 mmol NACL
40
0
0
1 2.5
0 0
20.0 mmol DCA
20
8
40.0*
7 35*
1 5
6.67 mmol DCA
20
1
5
3 15
0 0
2.0 mmol DCA
40
0
0
0 0
0 0
20.0 mmol TCA
20
0
0
2 15.8
5 25*
6.67 mmol TCA
19
0
0
3 7.5
0 0
2.0 mmol TCA
40
3
7.5
3 2.5
0 0
Multiplicity:
Treatment
N
foci/mouse
adenomas/mouse
carcinomas/mouse
at 82 weeks
20.0 mmol NACL
90
0.11 ±0.03
0.02 ±0.02
0.02 ±0.02
20.0 mmol DCA
19
7.95 ±2.00**
5.58 ± 1.14**
0.37 ± 0.17*
6.67 mmol DCA
28
0.39 ± 0.11*
0.32 ± 0.13*
0.04 ± 0.04
2.0 mmol DCA
50
0.14 ±0.05
0.06 ±0.03
0
20.0 mmol TCA
18
1.33 ±0.31**
0.61 ±0.22*
0.39 ± 0.16*
6.67 mmol TCA
27
0.41 ±0.13*
0.11 ±0.06
0.22 ±0.10*
2.0 mmol TCA
53
0.26 ±0.08
0.08 ± 0.04
0
*p <0.05 ** p<0.01
Treatment at
N
Foci/mouse
Adenoma/mouse
Carcinoma/mouse
51 weeks
20.0 mmol NACL
40
0
0.03 ±0.03
0
20.0 mmol DCA
20
0.60 ±0.22**
0.45 ±0.17**
0.10±0.10
6.67 mmol DCA
20
0.05 ±0.05
0.20 ±0.12
0
2.0 mmol DCA
40
0
0
0
20.0 mmol TCA
20
0
0.15 ± 0.11
0.50 ± 0.18*
6.67 mmol TCA
19
0
0.21 ±0.12
0
2.0 mmol TCA
40
0.08 ± 0.04
0.08 ± 0.04
0
02/12/09
E-135
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
These data show the decreased power of using fewer than 50 mice, especially at shorter
durations of exposure. By 82 weeks of exposure increased adenoma and carcinomas induced by
TCA or DC A treatment are readily apparent.
The foci of altered hepatocytes and the tumors obtained from this study were reported to
be basophilic, eosinophilic or mixed containing both characteristics. DCA was reported to
induce a predominance of eosinophilic foci and tumors, with over 80% of the foci and 90% of
the tumors in the 6.67 and 20.0 mmol/L concentration groups being eosinophilic. Only
approximately half of the lesions were characterized as eosinophilic with the rest being
basophilic in the group administered 2.0 mmol/L DCA. The eosinophilic foci and tumors were
reported to consistently stained immunohistochemically for the presence of GST-71, while
basophilic lesions did not stain for GST-71, except for a few scattered cells or small areas
comprising less than 10% of foci. The foci of altered hepatocytes in the TCA treatment groups
were approximately equally distributed between basophilic and eosinophilic in tincture.
However, the tumors were predominantly basophilic lacking GST-71 (21 of 28 or 75%) including
all 11 hepatocellular carcinomas. The limited numbers of lesions, i.e., 14, in the sodium chloride
(vehicle control) group were characterized as 64.3, 28.6, and 7.1% basophilic, eosinophilic, and
mixed, respectively.
Phenotype:
Treatment
N
%Foci
at 51 and 82 weeks
basophilic
eosinophilic
mixed
20.0 mmol NACL
10
70.0
30.0
0
20.0 mmol DCA
150
3.3
96.7
0
6.67 DCA
11
18.2
81.8
0
2.0 mmol DCA
7
42.8
57.2
0
20.0 mmol TCA
22
36.4
54.6
9.1
6.67 mmol TCA
11
45.5
54.5
0
2.0 mmol TCA
13
38.5
61.5
0
Treatment
N
Tumors
at 51 and 82 weeks
basophilic
eosinophilic
mixed
20.0 mmol NACL
4
50.0
25.0
25.5
20.0 mmol DCA
105
2.9
96.1
1.0
6.67 DCA
10
10
90
0
2.0 mmol DCA
3
0
100
0
20.0 mmol TCA
18
61.1
22.2
16.7
6.67 mmol TCA
6
100
0
0
2.0 mmol TCA
4
100
0
0
02/12/09
E-136
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
These data for female B6C3F1 mice shows that DC A and TCA treatment induced a
mixture of basophilic or eosinophilic foci. The pooling of the data between time and adenoma
vs. carcinoma decreases the ability to ascertain the phenotype of tumor due to treatment or the
progression of phenotype with time as well as the small number of tumor examined at lower
exposure concentrations. Foci that occurred at 51 and 82 weeks were presented as one result.
Adenomas and carcinoma data were pooled as one endpoint (n= number of total foci or tumors
examined). Therefore evolution of phenotype between less to more malignant stages of tumor
were lost.
2.3.2.10. Pereira and Phelps 1996
The focus of this study was to determine tumor response and phenotype in MNU-treated
mice after DCA or TCA exposure. The concentrations of DCA or TCA were the same as Pereira
(1996). For Pereira (1996) the animals were reported to be 7-8 weeks of age when started on
treatment and sacrificed after 360 or 576 days of exposure (51 or 82 weeks). For this study and
Tao et al. (2004), animals were reported o be 6 weeks of age when exposed to DCA or TCA via
drinking water and to be 31 or 52 weeks of age at sacrifice. Thus, exposure time would be ~ 24
or 45 weeks. A control group of non-MNU treated animals was presented for female B6C3F1
mice treated for 31 or 52 weeks and are discussed below. Although this paradigm appears to be
the same paradigm as those reported in Pereira (1996), fewer animals were studied. The number
of animals in each group varied between 8 controls and 14 animals in the 2.0 mmol/1 treatment
groups. In mice that were not treated with MNU, but were treated with either DCA or TCA at 31
weeks, there were no reported statistically significant treatment-related effect upon the yield of
foci or altered hepatocytes and liver tumors but the number of animals examined was small and
therefore of limited power to detect a response. The results below indicate a DCA-related
increase in foci and percentage of mice with foci.
Treatment till 31 weeks of age:
Treatment
No
Foci/mouse
incidence %
Adenomas/mouse
incidence %
20.0 mmol NACL
15
0.13 ± 0.13
6.7
0.13 ±0.13
not reported
20.0 mmol DCA
10
0.40 ±0.16
40%
0
0%
6.67 DCA
10
0.10±0.10
10%
0
0%
2.0 mmol DCA
15
0
0
0
0
20.0 mmol TCA
10
0
0
0
0
6.67 mmol TCA
10
0
0
0
0
2.0 mmol TCA
15
0
0
0
0
See Section 4.2.3 for further discussion of the results of co-exposures to MNU and DCA
or TCA from this study.
02/12/09
E-137
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.3.2.11. Ferreira-Gonzalez et al. 1995
The focus of this study was the investigation of differences in H-ras mutation spectra in
hepatocellular carcinomas induced by TCA or DCA in male B6C3F1 mice. 28-day old mice
were exposed for 104 weeks to 0. 1.0 g or 3.5 g/1 DCA or 4.5 g/1 TCA that was pH adjusted.
Tumors observed from this treatment were diagnosed as either hepatocellular adenomas or
carcinomas. DNA was extracted from either spontaneous, DCA- or TCA-induced hepatocellular
carcinomas. Samples for analysis were chosen randomly in the treatment groups of which 19%
of untreated mice had spontaneous liver hepatocellular carcinomas (0.26 carcinomas/animal),
DCA treatment induced 100% prevalence at 3.5 g/1 (5.06 carcinomas/animal) and 10.6%
carcinomas at 1.0 g/1 (1.29 carcinomas/animal). TCA treatment was reported to induce 73.3%
prevalence at 4.5 g/1 (1.5 carcinomas/animal). The number of samples analyzed was 32 for
spontaneous carcinomas, 33 for mice treated with 3.5 g/1 DCA, 13 from mice treated with 1.0 g/
DCA, and 11 from mice treated with 4.5 g/1 TCA. This study has the advantage of comparison
of tumor phenotype at the same stage of progression (hepatocellular carcinoma), for allowance of
the full expression of a tumor response (i.e. 104 weeks), and an adequate number of spontaneous
control lesions for comparison with DCA or TCA treatments. However, tumor phenotype at an
endstage of tumor progression reflects of tumor progression and not earlier stages of the disease
process.
There were no ras mutations detected except at H-61 in DNA from spontaneously arising
tumors of control mice. Only 4/57 samples from carcinogen-treated mice were reported to
demonstrate mutation other than in the second exon of H-ras. In spontaneous liver carcinomas,
58%) were reported to show mutations in H-61 as compared with 50%> of tumor from 3.5 g/1
DCA-treated mice and 45%> of tumors from 4.5.g/1 TCA-treated mice. Thus, there was a
heterogeneous response for this phenotypic marker for the spontaneous, DCA-, and TCA-
treatment induced hepatocellular carcinomas.
All samples positive for mutation in the exon 2 of H-ras were sequenced for the
identification of the base change responsible for the mutation. The authors noted that H-ras
mutations occurring in spontaneously developing hepatocellular carcinomas from B6C3F1 male
mice are largely confined to codon 61 and involve a change from CAA to either AAA or CGA or
CTA in a ratio of 4:2:1. They noted that in this study, all of the H-ras second codon mutations
involved a single base substitution in H-61 changing the wild-type sequence from CAA to AAA
(80%>), CGA (20%) or CTA for the 18 hepatocellular carcinomas examined. In the 16
hepatocellular carcinomas from 3.5 g/1 DCA treatment with mutations, 21%> were AAA
transversions, 50%> were CGA transversions, and 29%> were CTA transversions. For the 6
hepatocellular carcinomas from 1.0 g/1 DCA with mutations, 16%> were an AAA transversion,
50%o were a CGA transversion, and 34%> were a CTA transversion. For the 5 hepatocellular
carcinomas from 4.5 g/1 TCA with mutations, 80%> were AAA transversions, 20%> CGA
tranversions, and 0%> were CTA transversions. The authors note that the differences in
frequency between DCA and TCA base substitutions did not achieve statistical significance due
to the relatively small number of tumors from TCA-treated mice. They note that the finding of
02/12/09
E-138
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
essentially equal incidence of H-ras mutations in spontaneous tumors and in tumors of
carcinogen-treated mice did not help in determining whether DCA and TCA acted as
"genotoxic" or "non-genotoxic" compounds.
2.3.2.12. Pereira et al. 2004
Pereira et al (2004) exposed 7-8 week old female B6C3F1 mice treated with "AIN-76A
diet" to neutralized 0, or 3.2 g/1 DCA in the drinking water and 4.0 or 8.0 g/kg L-methionine
added to their diet. The final concentration of methionine in the diet was estimated to be 11.3 and
15.3 g/kg. Mice were sacrifice 8 and 44 weeks after exposure to DCA with body and liver
weights evaluated for foci, adenomas, and hepatocellular carcinomas. No histological
descriptions were given by the authors other than tinctoral phenotype of foci and adenomas for a
subset of the data. The number of mice examined was 36 for the DCA + 8.0 g/kg methionine or
4.0 g/kg methionine group sacrificed at 44 weeks. However for the DCA-only treatment group
the number of animals examined was 32 at 44 weeks and for those groups that did not receive
DCA but either methionine at 8.0 or 4.0 g/kg, there were only 16 animals examined. All groups
examined at 8 weeks had 8 animals per group. Liver glycogen was reported to be isolated from
30-50 mg of whole liver. Peroxisomal acyl-CoA oxidase activity was reported to be determined
using lauroyl-CoA as the substrate and was considered a marker of peroxisomal proliferation.
Whole liver DNA methylation status was analyzed using a 5-MeC antibody.
Methionine (8.0 g/kg) and DCA co-exposure was reported to result in the death of 3 mice
while treatment with methionine (4.0 g/kg) and DCA or methionine (8.0g/kg) alone was reported
to kill one mouse in each group. The authors reported that "There was an increased in body
weight during weeks 12 to 36 in the mice that received 8.0 g/kg methionine without DCA. There
was no other treatment-related alteration in body weight." However the authors do not present
the data and initial or final body weights were not presented for the differing treatment groups.
DCA treatment was reported to increase % liver/body weight ratios at 8 and 44 weeks to about
the same extent (i.e., ~ 2.4 -fold of control at 8 weeks and 2.2-fold of control at 44 weeks).
Methionine co-exposure was reported to not affect that increase (~ 2.4-fold, 2.2-fold, and 2.1-
fold of control after DCA treatment alone, DCA/4 g/kg methionine, and DCA/8 mg/kg
methionine treatment for 8 weeks, respectively). There was a slight increase in % liver/body
weight ratio associated with 8.0 g/kg methionine treatment alone in comparison to controls (~
7%) at 8 weeks with no difference between the two groups at 44 weeks.
After 8 weeks of only DCA exposure, the amount of glycogen in the liver was reported to
be ~ 2.09-fold of the value for untreated mice (115 mg/g vs. 52.5 mg/g glycogen in treated vs.
control, respectively at 8 weeks). Both 4 g/kg and 8 g/kg methionine co-exposure reduced the
amount of DCA-induced glycogen increase in the liver (~ 1.64-fold of control for DCA /4.0 g/kg
methionine and ~ 1.54-fold of control for DCA/8.0 mg/kg methionine). Thus, for treatment with
DCA alone or with the two co-exposure levels of methionine, the magnitude of the increase in
liver weight was greater than that of the increase in liver glycogen (i.e., 2.42-fold vs. 2.09-fold of
control % liver/body weight vs. glycogen content for DCA alone, 2.20-fold vs. 1.64-fold of
02/12/09
E-139
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
control % liver/body weight vs. glycogen content for DCA/4.0 g/kg methionine, 2.10-fold vs.
1.54-fold of control % liver/body weight vs. glycogen content for DCA/8.0 g/kg methionine).
Thus, the magnitudes of treatment-related increases were higher for % liver/body weight than for
glycogen content in these groups. In regard to percentage of liver mass that glycogen
represented, the control value for this study is similar to that presented by Kato-Weinstein et al
(2001) in male mice (~ 60 mg glycogen per gram liver) and represents ~ 6% of liver mass.
Therefore a doubling of the amount of glycogen is much less than the 2-fold increases in liver
weight observed for DCA exposure in this paradigm. This data suggests that DCA-related
increases in liver weight gain are not only the result of increased glycogen accumulation, and
that methionine co-exposure is affecting glycogen accumulation to a much greater extent than
the other underlying processes that are contributing to DCA-induced hepatomegaly after 8 weeks
of exposure. The authors reported that 8-weeks of DCA exposure alone did not result in a
significant increase in cell proliferation as measured by PCN index (neither data or methods were
shown). This is consistent with other data showing that DCA effects on DNA synthesis were
transient and had subsided by 8 weeks of exposure.
The levels of lauroyl-CoA oxidase activity were reported to be increased (~ 1.33-fold of
control) by DCA treatment alone at 8 weeks and to be slightly reduced by 8 g/kg methionine
treatment alone (~ 0.83 fold of control). Methionine co-exposure was reported to have little
effect on DCA-induced increases in lauroyl-CoA oxidase activity. The levels of DNA
methylation were reported to be increased by 8.0 g/kg methionine only treatment at 8 weeks ~
1.32-fold of control, and reduced by DCA only treatment to ~ 0.44-fold of control. DCA and 4.0
g/kg methionine co-exposure gave similar results as controls (within 2%). Co-exposures of DCA
and 8.0 g/kg methionine treatments were reported to increase DNA methylation 1.22-fold of
controls after 8 weeks of co-exposure.
In the 44-week study, the authors report that foci and hepatocellular adenomas were
found. However, the authors do not report the incidences of these lesions in their study groups
(how many of the treated animals developed lesions). As noted above, the numbers of animals in
these groups varied widely between treatments (e.g. n= 36 for DCA and co-exposure to 8.0g/kg
methionine but only n= 16 for 8 g/kg methionine treatment alone). Although reporting
unscheduled deaths in the 8.0g/kg methionine and DCA co-exposure groups, the authors did not
indicate whether these mortalities occurred in the 44-week or 8-week study groups.
Multiplicities of foci and adenoma data were presented. DCA was reported to induce 2.42 ±
0.38 foci/mouse and 1.28 ± 0.31 adenomas/mouse (m ± SE) after 44 weeks of treatment. The
DCA-induced foci and adenomas were reported to stain as eosinophilic with "relatively large
hepatocytes and nuclei". The authors did not present data on the % of foci and adenomas that
were eosinophilic using this paradigm. The addition of 4.0 or 8.0 g/kg methionine to the AIN-
76A diet was reported to reduce the number of DCA-induced adenomas/mouse to 0.167 ± 0.093
and 0.028 ± 0.028, respectively. However, the addition of 4.0 g/kg methionine to the DCA
treatment was reported to increase the number of foci/mouse (3.4 ± 0.46 foci/mouse). The
addition of 8.0 g/kg methionine to the DCA treatment was reported to yield 0.94 ± 0.24
foci/mouse. There were no foci or tumors in the 16 mice that received either the control diet or
02/12/09
E-140
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the 8.0 g/kg methionine treatment without DCA. The authors did not report whether methionine
treatment had an effect on the tincture of the foci or adenomas induced by DCA.
Therefore, a very high level of methionine supplementation to an AIN-760A diet, was
shown to affect the number of foci and adenomas, i.e., decrease them, after 44 weeks of co-
exposure to very high exposure concentration of DCA. However, a lower level of methionine
co-exposure increased the incidence of foci at the same concentration of DCA. Methionine
treatment alone at the 8 g/kg level was reported to increase liver weight, decrease lauroyl-CoA
activity and to increase DNA methylation. No histopathology was given by the authors to
describe the effects of methionine alone. Co-exposure of methionine with 3.2 g/1 DCA was
reported to decrease by ~ 25% DCA-induced glycogen accumulation and increase mortality, but
not to have much of an effect on peroxisome enzyme activity (which was not elevated by more
than 33% over control for DCA exposure alone). The authors suggested that their data indicate
that methionine treatment slowed the progression of foci to tumors. Whether, these results
would be similar for lower concentrations of DCA and lower concentrations of methionine that
were administered to mice for longer durations of exposure, cannot be ascertained from this data.
It is possible that in a longer-term study, the number of tumors would be similar. Whether,
methionine treatment co-exposure had an effect on the phenotype of foci and tumors was not
presented by the authors in this study. Such data would have been valuable to discern if
methionine co-exposure at the 4.0 mg/kg level that resulted in an increase in DCA-induce foci,
resulted in foci of a differing phenotype or a more heterogeneous composition than DCA
treatment alone.
2.3.2.13. DeAngelo et al. 2008
In this study, neutralized TCA was administered in drinking water to male B6C3 F1 mice
(28-30 days old) in three studies. In the first study control animals received 2 g/1 sodium
chloride while those in the second study were given 1.5 g/1 neutralized acetic acid (HAC) to
account for any taste aversion to TCA dosing solutions. In a third study deionized water served
as the control. No differences in water uptake were reported. Mean initial weights were reported
to not differ between the treatment groups (19.5 ± 2.5 g - 21.4 ± 1.6 g or ~ 10% difference). The
first study was reported to be conducted at the U.S. EPA laboratory in Cincinnati, OH in which
mice were exposed to 2 g/1 sodium chloride, or 0.05, 0.5, or 5 g/1 TCA in drinking water for 60
weeks. There were 5 animals at each concentration that were sacrificed at 4, 15,31, and 45
weeks with 30 animals sacrificed at 60 weeks of exposure. There were 3 unscheduled deaths in
the 0.05 g/1 TCA group leaving 27 mice at final necropsy. For the other exposure groups there
were 29 or 30 animals at final necropsy. In the second study, also conducted in the same
laboratory, mice were reported to be exposed to 1.5 g/1 neutralized acetic acid or 4.5 g/1 TCA for
104 weeks. Serial necropsies were conducted (5 animals per group) at 15, 30, and 45 weeks of
exposure and on, 10 animals in the control group at 60 weeks. For this study, a total of 25
animals were sacrificed in interim necropsies in the 1.5 g/1 HAC group and 15 in the 4.5 g/1 TCA
group. There were 7 unscheduled deaths in the HAC group and 12 in the 4.5 g/1 TCA group
leaving 25 animals at final necropsy and 30 animals in the final necropsy groups, respectively.
02/12/09
E-141
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Study 3 was conducted at the U.S EPA laboratory in RTP NC. Mice were exposed to deionized
water or 0.05 or 0.5 g/1 TCA in the drinking water for 104 weeks with serial necropsies (n=8 per
group) conducted at 26, 52, and 78 weeks. There were 19-21 animals reported at interim
sacrifices and 17 unscheduled deaths in the deionized water group, 24 unscheduled deaths in the
0.05 g/1 TCA group, and 24 unscheduled deaths in the 0.5 g/1 TCA group. This left 34 mice at
final necropsy in the control group, 29 mice in the 0.05 g/1 TCA group, and 27 mice in the 0.5 g/1
group.
At necropsy, liver, kidneys, spleen and testes weights were reported to be taken and
organs examined for gross lesions. Tissues were prepared for light microscopy and stained with
H& E. At termination of the exposure periods, a complete rodent necropsy was reported to be
performed. Representative blocks of tissue were examined only in 5 mice from the high dose
and control group with the exception of gross lesions, liver, kidney, spleen and testis at interim
and terminal sacrifices. If the number of any histopathologic lesions in a tissue was
"significantly increased above that in control animals" then that tissue was reported to be
examined in all TCA dose groups. For Study # 3 a second contract pathologist reviewed 10% of
the described hepatic lesions. No "major differences" were reported between the two pathologic
diagnoses. The prevalence and multiplicity of hepatic tumors were reported to be derived by
performing a histopathologic examination of surface lesions and four sections cut from each of
four tissue blocks excised from each liver lobe. Tumor prevalence was reported to be calculated
as the percentage of the animals with a neoplastic lesion compared to the number of animals
examined. Tumor multiplicity was reported to be calculated by dividing the number of each
lesion or combined adenomas and carcinomas by the number of animals examined.
Preneoplastic large foci of cellular alteration were also observed over the course of the study.
The prevalence and severity of hepatocellular cytoplasmic alterations, inflammation, and
necrosis were reported to be determined using a scale based on the amount of liver involved of 1
= minimal (occupying 25%), 2 = mild (occupying 25-50%), 3 = moderate (occupying 50 -75%)
and 4 = marked (occupying >75%). The only "significant change outside of the liver" was
reported to be testicular degeneration. Lactate dehydrogenase (LDH) was determined in arterial
blood collected at 30 and 60 weeks (study 1) and 4, 30, and 104 weeks (study 2). Cyanide
insensitive palmitoyl coenzyme A oxidase (PCO) was also reported to be measured. Five days
prior to sacrifice, tritiated thymidine (studies 1 and 2) or BrdU (study 3) was administered via
miniosmotic pumps and the number of hepatocyte nuclei with grain counts >6 were scored in
1000 cells or chromogen pigment over nuclei (BrdU). The labeling index was calculated by
dividing the number of labeled hepatocyte nuclei by total number of hepatocytes scored. Total
neoplastic and preneoplastic lesions (multiplicity) were counted individually or combined (AD
and CA) for each animal. The analysis of tumor prevalence data were reported to include only
those animals examined at the scheduled necropsies or animals surviving to week 60 (Study 1) or
longer than 78 weeks (Studies 2 and 3). The data from all the scheduled necropsies was
combined for an overall test of treatment-related effect.
02/12/09
E-142
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For Study #1 (60-week exposure) all TCA treated groups experienced a decrease in
drinking water consumption with the decreases in drinking water for the 0.5 g/1 and 5 g/1 TCA
exposure groups reported as statistically significant by the authors. The water consumption in
ml/kg-day was reported to be reduced by 11%, 17%, and 30% in the 0.05g/l, 0.5 g/1 and 5 g/1
TCA treated groups compared to 2 g/1 NaCl control animals as measured by time-weighted mean
daily water consumption measured over the study. The control value was reported to be 171
ml/kg/day. Although the 0.05 g/1 exposure concentrations were not measured, the 0.5 g/1 and 5
g/1 solutions were within 4% of target concentrations. The authors estimated that the mean daily
doses were 0, 8 mg/kg, 68 mg/kg and 602 mg/kg per day. For the 102 week studies the mean
water consumption with deionized water was reported to be 112 ml/kg/day and 132 ml/kg/day
for control animals given 1.5 g/L HAC. Therefore, there appeared to be a 35% decrease in water
consumption between the controls in Study #1 given 2 g/1 NaCl and controls in a Study # 3 given
deionized water but conducted at a different laboratory. There appeared to be a 23% reduction in
water consumption between animals given 2 g/1 NaCl and those given 1.5 g/1 HAC at the same
laboratory (Study # 2). As the concentrations of TCA were increased, there would be a
corresponding increase in the amount of sodium hydroxide needed to neutralize the solutions and
a corresponding increase in salts in the solution as well as TCA. The authors did not address nor
discuss the differences in drinking water consumption between the differing control solutions
between the studies. DeAngelo et al. (1999) reported mean drinking water consumption of 147
ml/kg/day in control mice of over 100 weeks and that the highest dose of DCA (3.5 g/1) reduced
drinking water consumption by 26%. Carter et al. (1995) reported that DCA at 5 g/1 to decrease
drinking water consumption by 64% and 46% but 0.5 g/1 DCA to not affect drinking water
consumption. While reporting that Study #1 showed that increasing TCA concentration
decreased drinking water consumption, the drinking water consumption in Studies #2 and #3
were similar between controls and TCA exposure groups with both being less than the control
and low TCA concentration values reported in Study # 1 (i.e. in Study #2 the 1.5 g/1 HAC and
4.5 g/1 TCA drinking water consumption was ~ 130 ml/kg/day and in Study # 3 the drinking
water consumption was -112 ml/kg/day for the deionized water control and 0.05 g/1 and 0.5 g/1
TCA exposure groups). Thus, the drinking water concentrations for Study #3 was ~ 35% less
than for the control values for Study #1 and was also ~ 25% less than for DeAngelo et al (1999).
The reasons for the apparently lower drinking water averages for Study #3 and the lack of effect
of the addition of 0.5 g/1 TCA that was reported in Study #1 and in other studies, was not
discussed by the authors.
In Study #1, there was little difference between exposure groups (n=5) noted for the final
body weights (mean range of 27.6 - 28.1 g) in mice sacrificed after 4 weeks of exposure.
However, absolute liver weight and % liver/body weight ratios increased with TCA dose. The %
liver/body weight ratios were 5.7 ± 0.4%, 6.2 ± 0.3%, 6.6 ± 0.4%, and 7.7. ± 0.6% for the 2 g/1
NaCl control, 0.05 g/1, 0.5 g/1, and 5 g/1 TCA exposure groups, respectively. These represent
1.09-fold, 1.16-fold, and 1.35-fold of control levels that were statistically significant. At 15
weeks of exposure the fold increases in % liver/body weight ratios were 1.14-fold, 1.16-fold,
1.47-fold of controls for 0.05 g/1 0.5 g/1 and 5 g/1 TCA. At 31 weeks of exposure the fold
increases in % liver/body weight ratios were 0.98-fold, 1.09-fold, 1.59-fold of controls for 0.05
02/12/09
E-143
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
g/1 0.5 g/1 and 5 g/1 TCA. At 45 weeks of exposure the fold increases in % liver/body weight
ratios were 1.13-fold, 1.45-fold, 1.98-fold of controls for 0.05 g/1 0.5 g/1 and 5 g/1 TCA. At 60
weeks of exposure the % liver/body weight ratios were 0.94-fold, 1.25-fold, 1.60-fold of controls
for 0.05 g/1 0.5 g/1 and 5 g/1 TCA. Thus, the range of increase at the lowest level of TCA
exposure (i.e., 0.05 g/1) was 0.94-fold to 1.14-fold of controls. These data consistently show
TCA-induced increases in liver weight from 4 to 60 weeks of the study that were dose-related.
For the 0.5 g/1 exposure group, the magnitude of the increase compared to control was reported
to be about the same between weeks 4 and 30 with the highest increase reported to be at week 45
(1.45-fold of control). In regard to the correspondence with magnitude of difference in dose of
TCA and liver weight increase, there was ~ 2-fold increase in liver weight gain corresponding to
10-fold increases in TCA concentration at 4 weeks of exposure. For the 4 and 15-week
exposures there was ~ 3.3- and 3.9-fold difference in liver weight that corresponded to a 100-
fold difference in exposure concentration of TCA (i.e., 0.05 vs. 5.0 g/1 TCA).
The small number of animals examined, n=5, limit the power of the study to determine
the change in % liver/body weight up to 45 weeks, especially at the lowest dose. However the
0.05 g/1 TCA exposure groups at 4 week and 15 weeks were reported to significantly increase %
liver/body weight ratios. The % liver/ body weight ratios for all of the treatment groups and the
ability to detect significant changes were affected by changes in final body weight and changing
numbers of animals. After 4 to 30 weeks of exposure, the final body weights of mice increased
in control animals but were within 11% of each other between weeks 31 and 60. The %
liver/body weight ratios in controls decreased from 4 to 31 weeks and were slightly elevated by
60 weeks compared to the 31-week level. Although control values were changing, there
appeared to be no difference between control values and treated values in final body weight for
any duration of exposure with the exception of the 5 g/1 TCA exposure group after 60 weeks of
exposure, which was decreased by ~ 15%. At the 31-week and 60-week exposure durations, the
0.05 g/1 TCA groups did not have increased % liver/body weight ratios over controls.
In Study #2, conducted in the same laboratory but with a 1.5 g/1 HAC solution used for
control groups, there was less than 5% difference in final body weights between control mice
give HAC and those treated with 4.5 g/1 TCA up to 45 weeks. However final body weight was
reduced by TCA treatment by 104 weeks by ~ 15%. Between the interim sacrifices of 15, 30,
and 45 weeks, the % liver/body weight ratios in control mice were similar at 15 and 45 weeks (~
4.8%>) but greater in the 30-week control group (5.3% or ~ 10 % greater than other interim
control groups). The TCA-induced increases in body weight were 1.60-fold, 1.40-fold, and 1.79-
fold of control for the 15, 30, and 45 week groups exposed to 4.5 g/1 TCA in Study # 2. The
smaller magnitude of TCA-induced liver weight increase at 30-weeks that that for 15 and 45
weeks, was a reflection of the increased % liver/body weight ratio reported for the HAC control
at that time point.
Comparisons can be made between Study #1 and Study #2 for 4.5 g/1 or 5.0 g/1 TCA
exposure levels and controls for 15, 30/31 and 45 weeks of exposure to ascertain the consistency
of response from the same laboratory. Although the two studies had differing control solutions
02/12/09
E-144
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and reported different drinking water consumption overall, they were exposing the TCA groups
to almost the same concentration of TCA in the same buffered solutions for the same periods of
time with the same number of mice per group. Between Study #1 and Study #2, there were
consistent % liver/body weight ratios induced by either 5.0 g/1 TCA and 4.5 g/1 TCA at weeks 15
and 30/31 (i.e. within 3% of each other). The % liver/body ratios for these exposure groups
ranged from 7.3% - 7.7% between weeks 15 and 30/31 for the ~ 5.0 g/1 TCA exposure in both
studies. Final body weights were within 10%. While the % liver/body weight ratios induced by
~ 5.0 g/1 TCA were similar, the magnitude of increase in comparison to the controls was 1.47-
fold and 1.59-fold of control for Study #1, and 1.60-fold and 1.40-fold of control for Study #2
after 15 and 30/31 weeks of exposure, respectively. At 45 weeks, the % liver/body weight ratios
were within 11% of each other (9.4% vs. 8.4%) and final body weights were within 2 % of each
for this exposure concentration between the two studies giving a 1.98-fold and 1.79-fold of
control % liver/body weight, respectively. Thus, the apparent magnitude of TCA-induced
increase in % liver/body weight was affected by control values used as the basis for comparison.
The % liver/body weights reported for either 4.5 g/1 TCA or 5.0 g/1 TCA exposure groups for
weeks 15 and 30/31 was similar between the two studies conducted in the same laboratory.
Study #3 was conducted in a separate laboratory, interim sacrifice times were not the
same as for Study # 1, the number of animals examined differed (n= 5 for Study #1 and n=8 for
Study #3), and control animals studied for comparative purposes were given different drinking
water solutions (deionized water vs. 2 g/1 NaCl). Most importantly the body weights reported at
52 weeks was much grated than that reported at 45 weeks for Studies #1 and #2. However, a
comparison of TCA-induced liver weight gain and the effects of final body weight can be made
between the 0.05 g/1 and 0.5 g/1 TCA exposure groups at 30 weeks (Study #1) and 26 weeks
(Study #3), at 45 weeks and 60 weeks (Study # 1), and 52 weeks (Study # 3). At 31 weeks there
was <2% difference in mean final body weights between control and the two TCA-treatment
groups in Study #1. There was also little difference between the TCA-treated groups at week in
Study #3 at week 26 and the TCA treatment groups in at week 31 in Study # 1 (i.e., range of 42.6
g - 43.5 g for 0.05 and 0.5 g/1 TCA treatments in Studies # 1 and #3). However in Study # 3, the
control value was 12% lower than that of Study #1 for mean final body weight. Based on final
body weights, there would be an expectation of similar results between the two studies at the 26
and 30 week time points. At the 45 week (Study #1), and 52-week (Study #3), and 60-week
(Study #1) durations of exposure, the mean final body weights varied little between their
corresponding control groups at each sacrifice time (less than 4% variation between control and
TCA-treated groups). However, there was variation in mean final body weights between the
differing sacrifice times. Control and TCA-treated groups were reported to have lower mean
final body weights at 45 weeks of exposure in Study #1 than at either 30 weeks or at 60 weeks.
The 45-week mean final body weights in Study #1 were also reported to be lower than those at
52 weeks in Study #3. Control mean body weight values were 28% higher at 52 weeks in Study
#3 than 45 weeks in Study #1 and 15% higher for 60 weeks in Study #1. In essence, for Study #
1 mean final body weights went down between 31 and 45 weeks of exposure and then went back
up at 60 weeks of exposure for control mice (~ 43 g, ~ 40 g, and ~ 44 g for 31, 45 and 60 weeks,
respectively) as well as for both TCA concentrations. However, for Study #3 final mean body
02/12/09
E-145
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
weights went up between 26 and 52 weeks of exposure for control mice (~ 39 g vs. ~51 g) and
for both TCA concentrations. While for Study # 1 the % liver/body weight ratios were 0.98-fold
and 1.09-fold of control at 31 weeks of exposure, at week 45 the ratios were 1.13-fold and 1.45-
fold of control, and at week 60 they were 0.94-fold and 1.25-fold of controls for the 0.05 g/1 and
0.5 g/1 TCA exposure levels, respectively. For Study # 3, the pattern differed than that of Study
# 1. There was a 1.07-fold and 1.18-fold of control % liver/body weight for 26 weeks but a 0.92-
fold and 1.04-fold of control % liver/body weight change at 52 weeks of exposure at 0.05 g/1 and
0.5 g/1 TCA exposure, respectively.
Thus, there appeared to be differences in control and the treatment groups at the 26 week
sacrifice groups in Study # 3 that was not apparent at the 52-week sacrifice time. Overall, the
final body weights appeared to be similar between controls and TCA treatment groups at the 52-
week sacrifice time in Study # 3 and at the 31-, 45-, and 60-week sacrifice times in Study #1.
However, although consistent within sacrifice times, the final body weights differed between the
various sacrifice times in Studies # 1 and # 3. The patterns of % liver/body weight at differing
and similar sacrifice times appeared to differ between the Study # 1 and Study #3 at the same
concentrations of TCA. The largest difference appeared to be between week 45 group in Study #
1 and week 52 group in Study #3 where both concentrations of TCA were reported to induce
increases in % liver/body weight in one study but to have little difference in the other. The
differences in mean final body weights between these two sacrifice times were also the largest
although control and TCA-treatment groups had little difference on this parameter. Similar to
the work of Kjellstrand et al with TCE (Kjellstrand et al., 1983a), the groups with the lower body
weight appeared to have the greatest response in liver weight increase.
These data illustrate the variability in findings of % liver weight induction between
laboratories, studies, choice of controls solutions, and the affects of final body weights on this
parameter. They also illustrate the limitations for determining either the magnitude or pattern of
liver weight increases using a small number of test animals. As animals age the size of their
liver changes but also during the latter parts of the lifespan, foci and spontaneously occurring
liver tumors can affect liver weight. The results of Study #1 show a consistent dose-response in
TCA liver weight increases at 4 and 15 week time periods over a range of concentration from
0.05 g/1 to 5 g/1 TCA.
In regard to non-neoplastic pathological changes the authors reported that" Increased
incidences and severity of centrilobular cytoplasmic alterations, inflammation, and necrosis were
the only nonproliferative changes seen in livers of animals exposed to TCA for 60 weeks (Tables
7-9; Study 1. Incidences were between 21 and 93%; severity ranged from minimal to mild; and
some lesions were transient. Centrilobular cytoplasmic alterations (Table 7) were the most
prominent nonproliferative lesion. The incidence and severity were dose related and
significantly increased at all TCA concentrations. Centrilobular alterations are a low-grade
degeneration of the hepatocytes characterized by an intense eosinophilic cytoplasm with deep
basophilic granularity (microsomes) and slight hepatomegaly. The distribution ranged from
centrilobular to diffuse. The incidence of inflammation was increased significantly in the 5 g/1
02/12/09
E-146
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
TCA treatment group (Table 8), but was significantly lower in the 0.05- and 0.5 g/1 groups
between 31 and 45 weeks, but abated by 60 weeks. There was a significant dose-related trend,
but a significant increase in severity was only found at 5 g/1. No alteration in the severity of this
lesion was observed. The occurrence and severity of nonproliferative lesions in animals exposed
to 0.5 and 4.5 g/1 TCA for 104 weeks were similar to those observed at 60 weeks (data not
shown). No pathology outside the liver was observed except for a significant dose-related trend
and incidence of testicular tubular degeneration at 0.5 and 5 g/1 TCA."
The results shown in Table # 7 by the authors for the 60-week TCA-exposed mice did
not show a dose-response for either incidence or severity of centrilobular cytoplasmic alterations.
They reported a 7%, 48%, 21% and 93% incidence and a 0.10 ± 0.40, 0.70 ± 0.82, 0.34 ± 0.72
and 1.60 ± 0.62 mean severity score for control, 0.05, 0.5, and 5.0 g/1 TCA exposure groups,
respectively. Thus, for control, 0.05 and 0.5 g/1 TCA exposure there was less than minimal (i.e.,
score of 1 or occupying less than 25% of the microscopic field) severity of this finding for the 27
to 30 mice examined in each group. Only slight hepatomegaly is noted by the authors to be
included in their description of the centrilobular cytoplasmic alteration. Interestingly, the
elevation of this parameter for both incidence and severity in the 0.05 g/1 TCA exposed group
compared to 0.5 g/1 exposure group did not correspond to an increase in % liver/body weight for
this same exposure group. While the % liver/body weight ratio was 32% higher, the incidence
and severity of this lesion were reported to be half that in the 0.5 g/1 vs. 0.05 g/1 exposure groups
after 60 days of TCA exposure. Thus, TCA-induced hepatomegaly did not appear to be
associated with this centrilobular cytoplasmic change. Similarly the incidence of hepatic
inflammation was reported to be 10%, 0%, 7%, and 24% and severity, 0.11 ± 0.40, 0.09 ± 0.30,
0.12 ± 0.33, and 0.29 ± 0.48 for control, 0.05, 0.5, and 5.0 g/1 TCA exposure groups,
respectively. Thus, at no TCA exposure concentration was the incidence more than 24% and the
severity was considerably less than minimal. The reported results for hepatic necrosis were
pooled from data from the 5 mice exposed for either 30 or 45 weeks (n= 10 total). No incidences
of necrosis were reported for either control or 0.05 g/1 TCA exposed mice. At 0.5 g/1 TCA 3/10
mice were reported to have necrosis but at a severity level of 0.50 ± 0.97. At 5.0 g/1 TCA 5/10
mice were reported to have necrosis but at a severity level of 1.30 ± 1.49. The limitations of the
small number of animals pooled in this data are obvious. However, there does not appear to be
much more than minimal necrosis at the highest dose of TCA between 30 and 45 weeks and this
response is reported by the authors to be transient.
Serum LDH activity was reported by the authors for 31 and 60 week TCA exposures in
Study #1. They state that "There was a dose-related trend at 31 weeks; serum LDH was
significantly increased at 0.5 and 5 g/1 TCA (161 ± 39 and 190 ± 44, respectively vs. 100 ± 28 IU
for the control). LDH activity returned to control levels at 60 weeks. Similarly, elevated LDH
levels were observed at early time periods for 0.5 and 4.5 g/L TCA during the 104 week
exposure (data not shown: Studies 2 and 3)." The data presented by the author for Study #1 are
from 5 animals/group for the 30-week results and 30 animals/group for the 60-week results. Of
interest is for the 60-week data, there appears to be 50% decreased in LDH activity at 0.05 and ~
25% decrease in LDH activity at 0.5 g/1 TCA treatment with the LDH level reported to be the
02/12/09
E-147
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
same as control for the 5 g/1 TCA exposure group. For the 31-week data, in which only 5
animals were tested in each treatment group, there appeared to be a slight increase at the 0.5 g/1
(60% increase over control) and 5 g/1 (90% increase over control) treatment groups. The data for
necrosis detected by light microscopy and by LDH level is consistent with no changes from
control detected at the 0.05 g/1 TCA treatment group and less than minimal necrosis of on a 60%
increase in LDH level over control reported for 0.5 g/1 TCA treatment. Even at the highest dose
of 5.0 g/1 TCA there is still little necrosis or LDH release reported over control.
Data for testicular tubular degeneration was reported for Study # 1 after 60-weeks of
TCA exposure. The incidence of testicular tubular degeneration was reported to be 7%, 0%,
14%), and 21% for mice exposed to 2.0 g/1 NaCl, 0.05 g/1, 0.5 g/1 and 5.0 g/1 TCA. The severity
of the lesions was reported to be 0.10 ± 0.40, 0, 0.17 ± 0.47, and 0.21 ± 0.41 with a significant
trend with dose reported by the authors for severity and for the 0.5 g/1 and 5 g/1 treatment groups
to be significantly increased over control incidence levels. Of note, similar to the % liver/body
weight ratios and hepatic inflammation values for this data set, the values for testicular tubular
degeneration were slightly higher in control mice than 0.05 g/1 TCA exposed mice. In regard to
mean severity levels for testicular degeneration, although still minimal, there was little difference
between the results for reported for the 0.5 g/1 TCA and 5.0 g/1 TCA exposed mice.
In regard to peroxisome proliferation, liver PCO activity was presented for up to 60
weeks (Study #1) and 104 weeks (Study #2). Similar to the data for LDH activity, ~ 30 animals
were examined at the 60-week time point but only 5 animals per exposure group were examined
for 4-, 15-, 31-, and 45-week results. The data are presented in a figure and in some instances
hard to determine the magnitude of change. Similar to other reports, the baseline level of PCO
activity was variable between control groups and ranged 2.7-fold (~ 1.49 to 4.06 nmol NAD
reduced/min/mg protein given by the authors). There appeared to be little change in PCO
activity between the 0.05 g/1 TCA exposure and control levels for up to 45 weeks of exposure
(i.e., the groups with n= 5) in Study # 1. For the 60-week group the 0.05g/l TCA group PCO
activity was ~ 1.7-fold of control but was not statistically significant. For the 0.5 g/1 TCA
treatment groups, the increase ranged from ~1.3-fold to 2.7-fold of control after 4-, 15-, 31-, and
45-weeks of exposure with the largest differences reported at 4 and 60 weeks (i.e., 2.2-fold and
2.7-fold of control, respectively). For the 5.0 g/1 TCA exposure groups, the increase ranged
from ~ 3.2-fold to ~ 5.7-fold of control after 4, 15, 31, and 45 weeks of exposure. While the
data at 60-weeks had the most animals examined (~ 30 vs. 5) with ~ 1.7-fold, 2.7-fold, and 4.5-
fold of control PCO activity, at this time period the authors report the occurrence of tumors had
already occurred. At the earlier time points of 4 and 15 weeks, there was a difference in the
magnitude TCA-induced increase in PCO activity. As displayed graphically, at 4 weeks the
PCO increase was ~ 1.3-fold, 2.4-fold, and 5.3-fold of control for 0.05, 0.5, and 5.0 g/1 TCA,
respectively, while at 15 weeks, the PCO levels were decreased by 5%, increased to 1.3-fold, and
increased to 3.2-fold of control with only the 5.0 g/1 treatment group difference to be statistically
significant.
02/12/09
E-148
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For Study # 2 the authors present a figure (figure #4) that states that PCO values were
given for mice given HAC or 4.5 g/1 TCA for 4-60 weeks. However the data presented in Figure
#4 appears to be for 15-, 30-, 45- and 104-week exposures. The number of mice is not given in
the figure but the methods section states that serial section were conducted on 5 mice/group for
these interim sacrifice periods. The number of mice examined for PCO activity at 104 weeks
was not given by the authors but the number of mice at final sacrifice was given as 25. The
levels of PCO in the control tissues varied by ~ 33% for weeks 15 to 45 but there was a ~ 5-fold
difference between the level reported at 104 weeks and that for the earlier time periods in control
mice shown in the figures (~ 2.23 vs. 0.41 nmol NAD reduced/min/mg protein as given by the
authors). The increase over control induced by 4.5 g/1 TCA in Study # 2 was shown to be ~ 6.9-
fold, 4.8-fold, 3.6-fold, and 19-fold of controls for 15, 30, 45 and 104 weeks, respectively.
Therefore, at a comparable level of TCA exposure (-5.0 g/1), number of mice examined
(n=5), and durations of exposure (15, 30, and 45 weeks), the increase in PCO activity induced by
~ 5.0 g/1 TCA varied between 3.2- to 5.7-fold of control in Study #1 and between 3.6- to 6.9-fold
of control in Study #2. There was not a consistent pattern between the two studies in regard to
level of PCO induction from ~ 5 g/1 TCA and duration of exposure. The lowest TCA-induced
PCO activity increase was recorded at 15 weeks in Study #1 (i.e., 3.2-fold of control) and highest
PCO activity increase was recorded at 15 weeks in Study #2 (i.e., 6.9-fold of control). No PCO
data was reported for data in Study #3 with the exception of the authors stating that "PCO
activity was significantly elevated for the 0.5 g/1 TCA exposure over the 104 weeks (study 3).
The extent of the increases was similar to those measured for 0.5 g/1 TCA (200-375%: data not
shown) in Study 1." No other details are given for PCO activity in Study #3.
Hepatocyte proliferation was reported by the authors to be assessed by either
incorporation of tritiated thymidine (Studies #1 and #2) or BrdU (Study # 3) into hepatocyte
nuclei. As noted previously, these techniques measure DNA synthesis and not necessarily
hepatocyte proliferation. The authors did not report if specific areas of the liver were analyzed
by autoradiographs or how many autoradiographs were examined in the analyses they conducted.
For later time points of examination (60 -104 weeks) the authors did not indicate whether
hepatocytes in foci or adenomas were excluded from DNA synthesis reports. The authors
present data for what are clearly, 31, 45, and 60 week exposure for Study # 1 as the percent
tritiated thymidine labeled nuclei. An early time point that appears to be 8 weeks is also given.
However for Study # 1 only 4 week and 15 week durations were tested so it cannot be
established what time period the earlier time point represents. What is very apparent from the
data presented for Study # 1 is that the baseline level of tritiated thymidine incorporation was
relatively high and highly variable for the 5 animals examined (~ 8% of hepatocytes were
labeled). There did not appear to be an apparent pattern of TCA treatment groups at this
timepoint with the 0.05 and 5.0 g/1 TCA groups having a similar percentage of labeled
hepatocytes and for 0.5 g/1 TCA reported to have a 60% reduction in labeled hepatocytes. After
31 weeks of exposure the control values were reported to be 2% of hepatocytes labeled. The
authors report that only the 5.0 g/1 TCA group had a statistically significant increase of control
and was elevated to ~ 6% of hepatocytes. The two lower doses of TCA had similar reported
02/12/09
E-149
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
incidences of labeled hepatocytes of 4.5% that were not reported to be statistically significant.
For the 45-week exposure period in Study #1, the control value was reported to be 1.2% with
only the 5.0 g/1 TCA value reported to be statistically significantly increased at 3.2% and the
other two TCA groups to be similar to control. Finally, for the 60 week group from Study #1,
the control value was reported to be 0.6% of hepatocytes labeled and the only the 0.5 g/1 TCA
dose reported to be statistically significantly increased over control at 3.2%. What is clear from
this study is that the control value for the unidentified early time point is much higher than the
other values. There should not be such a large difference in mature mice nor such a high level.
The difference in control values between the earlier time point and the 31-week time point was
4-fold. The difference between the earlier time point and the 45-week time point was ~ 7-fold.
There did not appear to be an increase in hepatocyte tritiated thymidine labeling due to any
concentration of TCA at the early unidentified time point (~ week 10 from the figure) from
Study # 1. There was no dose-response apparent for the other study periods and the % of
hepatocytes labeled were 3% or less. These results indicated DNA synthesis was not increased
by 10-60 week exposures to TCA exposure that induced increased liver tumor response.
For Study # 2 results were reported for tritiated thymidine incorporation into hepatocytes
in a figure that was labeled as 4.5 g/1 TCA and control tissue for 104 weeks but showed data for
15, 30, and 45 weeks of exposure. Of note is that the control values for this study were much
lower than that reported for Study #1. The % of hepatocytes labeled with tritiated thymidine was
reported to be ~ 2% for the 15 week exposure period and less than 1% for the 30- and 45-week
exposure periods. For the 4.5 g/1 TCA exposures the % hepatocytes labeled with tritiated
thymidine were ~ 2-4% at all time points with only the 45 week period identified by the authors
as statistically significant.
For Study # 3, rather than tritiated thymidine, BrdU was used as a measure of DNA
synthesis. The results are presented in Figure # 8 of the report in which the 0.5 g/1 TCA
concentration is mislabeled as 0 g/1 and the Figure is mislabeled as having a duration of 104
weeks but the data is presented for 26, 52, and 78 weeks of exposure. The % of hepatocytes at
26 weeks was reported to be ~ 1-2% for the control, 0.05 and 0.5 g/1 TCA groups. At 52 weeks
the control value was ~ 1% the 0.05 g/1 TCA value was less than 0.1% and the 0.5 g/1 TCA value
was ~ 3.5% but not statistically significant. At 78 weeks of exposure the control value was
reported to be ~ 0.2% with only the 0.05 g/1 TCA group having a statistically significant increase
over control.
From this data, the estimated control values for DNA synthesis at similar time points of
exposure ranged from 0.4% to 2% at 26-31 weeks and ~0.1 to 1.2% at 45-52 weeks. The results
for Study #1 and # 2 were inconsistent in regard to the magnitude of tritiated thymidine
incorporation but consistent in that there was a lot of variability in these measurements, not a
consistent pattern with time that was TCA-dose related, and ,even at the highest dose of TCA,
did not indicate much of an increase in cell proliferation 15-45 weeks of exposure. Similarly the
results for Studies #1 and #3 indicate that the two lower doses of TCA there were not generally
02/12/09
E-150
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
statistically significant increases in DNA synthesis from 15-45 weeks of exposure although there
was an increase in liver tumor response at later time points.
The authors reported that "all gross and microscopic histopathological alterations were
consistent across the three studies." However, the histological descriptions that follow were
focused on the liver for both neoplastic and non-neoplastic parameters. As stated above, only a
few animals (n= 5) from the control and high TCA dose level were examined for lesions other
than liver, kidneys, spleen and testes. Thus, whether other neoplastic lesions were induced by
TCA exposure cannot be determined from this set of studies.
Study # 1 was conducted for 60 weeks. Although of short duration and using 30 or less
animals, the authors reported in the text that "a significant trend with dose was found for liver
cancer. The prevalence and multiplicity of adenomas (38%; 0.55 ± 0.15) or carcinoma (38%;
0.42 ± 0.11) were statistically significant at 602 mg/kg/day TCA compared to control (7%; 0.07
± 0.05) [sic for both adenoma and carcinoma the same value was given, mean ± SD], When
either an adenoma or a carcinoma was present, statistical significant was seen at both 5 g/1 (55%;
1.00 ± 0.19) and 0.5 g/1 (38%: 0.52 ± 0.14 TCA exposure groups compared to control (13%; 0.13
± 0.06). No significant change in liver neoplasia were reported to be observed by the authors at
0.05 g/1 TCA. Preneoplastic large foci of cellular alteration (24%) were seen in the 5 g/1 TCA
control compared to control."
Although not statically significant, there was an incidence of 15% adenoma in the 0.05
g/1 TCA treatment group (n= 27) and a multiplicity of 0.15 ± 0.07 adenomas/mouse reported
with both values being twice that of the values given for the controls (n= 30). The incidence and
multiplicity for carcinomas was approximately the same for the 0.05 g/1 TCA treatment group
and the control group. Given the small number of animals examined, the study was limited in its
ability to determine statistical significance for the lower TCA exposure level. The fold increases
of incidence and multiplicity of adenomas at 60 weeks was 2.1-fold, 3.0-fold, 5.4-fold of control
incidence and 2.1-fold, 3.4-fold, and 7.9-fold of control multiplicity for 0.05 g/1, 0.5 g/1, and 5 g/1
exposure to TCA. For multiplicity of adenomas and carcinomas combined there was a 1.46-fold,
4.0-fold, and 7.68-fold of control values. Analysis of tumor prevalence data for this study
included only animals examined at scheduled necropsy. Since most animals survived till 60
weeks, most were included and a consistent time point for tumor incidence was reported.
There are significant discrepancies for reporting of data for tumor incidences in this
report for the 104 week data. While the methods section and Table describing the dose
calculation and animal survival indicate that Study #3 control animals were administered
deionized water and those from Study# 2 were given HAC, Table 6 of the report gives 2 g/1 NaCl
as the control solution given for Study #2 and 1.5g/l HAC for Study #3. A comparison of the
descriptions of animal survival and tumor incidence and multiplicity between the results given in
DeAngelo et al. (2008) and George et al, (2000) shows not only that the control data presented in
DeAngelo et al (2008) for Study #3 to be the same data as that presented by George et al (2000)
previously, but also indicates that rather than 1.5 g/1 HAC, the tumor data presented in DeAngelo
02/12/09
E-151
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
et al (2008) is for mice exposed to deionized water. DeAngelo et al (2008) did not report that
this data was from a previous publication.
Descriptor
George et al., 2000
DeAngelo et al., 2008
Species
mouse
mouse
Strain
B6C3F1
B6C3F1
Gender
Male
Male
Age
28-30 days
28-30 days
Source
Charles River, Portage
Charles River, Portage
Mean Initial body wt
19.5 ± 2.5 g
19.5 ± 2.5 g
Water consumption
111.7 ml/kg/day
1 12 ml/kg/day
Laboratory
RTP NC
RTP NC
# Animals at start
72
72
# Animals at interim sac.
22
21
# Unscheduled deaths
16
17
# Animals at final sacrifice
34
34
# Animals for pathology
65
63
Adenoma incidence
21.4 0 o
21%
Adenoma multiplicity
0.2 1 ± 0.06
0.21 ± 0.06
Carcinoma incidence
54.8%
55%
Carcinoma multiplicity
0.74 ±0.12
0.74 ±0.12
For Studies # 2 and # 3 tumor prevalence data was reported in the methods section of the
report to include necropsies of animals that survived greater than 78 weeks and thus included
animals that were scheduled for necropsy but also those which were moribund and sacrificed at
differing times. Thus, for the longer times of study, there was a mixture of exposure durations
that included animals that were ill and sacrificed early and those that survived to the end of the
study. Animals that were allowed to live for longer periods or who did not die before scheduled
sacrifice times had a greater opportunity to develop tumors. However, animals that died early
may have died from tumor-related causes. The mislabeling of the tumor data in DeAngelo et al
(2008) has effects on the interpretation of results for if the tumor results table was not mislabeled
it would indicated 17 animals were included in the liver tumor analysis that were not included in
the final necropsy and that the 7 unscheduled deaths could not account for the total number of
"extra" mice included in the tumor analysis so some of the animals had to have come from
interim sacrifice times (78 weeks or less) and that for Study #3 the data from 9 animals at
terminal sacrifice were not used in the tumor analysis. Not only was the control data mislabeled
for Study #3, but the control data was also apparently mislabeled for Study #2 as being 2.0 g/1
NaCl rather than 1.5 g/1 HAC. Of the 42 animals used for the tumor analysis in Study #3, only
34 were reported to have survived to interim sacrifice so that 8 animals were included from
unscheduled deaths. However, the authors report that there were 17 unscheduled deaths in the
study not all were included in the tumor analysis. The basis for the selection of the 8 animals for
tumor analysis was not give by the authors.
02/12/09
E-152
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Not only are the numbers of control animals used in the tumor analysis different between
two studies (25 mice in Study #2 and 42 mice in Study #3), but the liver tumor results reported
for Study #2 and Study #3 were very different. Of the 42 "control" mice examined from Study
#3, the incidence and multiplicity of adenomas was reported to be 21% and 0.21 ± 0.06,
respectively. For carcinomas, the incidence and multiplicity was reported to be 55% and 0.74 ±
0.12, respectively, and for the incidence and multiplicity of adenomas and carcinomas combined
reported to be 64% and 0.93 ± 0.12, respectively. For the 25 mice reported by the authors for
Study # 2 to have been treated with "2.0g/l NaCl" but were probably exposed to 1.5 g/1 HAC, the
incidence and multiplicity of adenomas was 0%. For carcinomas, the incidence and multiplicity
was reported to be 12% and 0.20 ± 0.12, respectively and for the incidence and multiplicity of
adenomas and carcinomas combined to be 12% and 0.20 ± 0.12, respectively. Therefore while ~
64% the 42 control mice in Study #3 were reported to have adenomas and carcinomas, only 12%
of the 25 mice were reported to have adenomas and carcinomas in Study #2 for 104-weeks.
While the effect of using fewer mice in one study vs. the other will be to reduce the
power of the study to detect a response, there are additional factors that raise questions regarding
the tumor results. Not only were the tumor incidences were reported to be higher in control
mice from Study # 3 than Study #2, but the number of unscheduled deaths was reported to also
be 2-fold higher. The age, gender, and strain of mouse were reported to be the same between
Study #2 and #3 with only the vehicles differing and weight of the mice to be reported to be
different. Although the study by George et al (2000) describes the same control data set as for
Study #3 as being for animals given deionized water, there is uncertainty as to the identity of the
vehicle used for the tumor results reported for Study #3 and there are some discrepancies in
reporting between the two studies. As discussed below in Section 2.5, the differences in the
weight of the mice between Studies #1, #2 and #3 is critical to the issue of differences in
background tumor rate and hence interpretability of the study.
As noted by Leakey et al. (2003b), the greatest correlation with liver tumor incidence and
body weight appears between the ages of 20 and 60 weeks in male mice. As reported in Section
2.5., the mean 45-week body weight reported for control male B6C3F1. mice in the George et al
(2000) study, which is the same control data as DeAngelo et al (2008) was ~ 50 g. This is a
much greater body weight than reported for Study # 1 at 45 weeks (i.e., 39.6 g) and for Study #
2 at 45 weeks ( i.e., 39.4 g). Using probability curves presented by Leakey et al (2003b), the
large background rate of 64% of combined adenomas and carcinomas for Study # 3 is in the
range predicted for such a large body weight (i.e. ~ 65%). Such a high background incidence
compromises a 2-year bioassay as it prevents demonstration of a positive dose-response
relationship. Thus, Study # 3 of DeAngelo et al (2008) is not comparable to the results in Study
# 1 and # 2 for the determination of the dose-response for TCA.
The accurate determination of the background liver tumor rate is very important in
determining a treatment related effect. The very large background level of tumor incidence
reported for Study # 3 makes the detection of a TCA-related change in tumor incidence at low
exposure levels very difficult to determine. Issues also arise as to what the source of the tumor
02/12/09
E-153
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
data was in the TCA-treatment and control groups in Study #3. While 29 mice exposed to
0.05g/l TCA were reported to have been examined at terminal sacrifice, 35 mice were used for
liver tumor analysis. Similarly, while 27 mice exposed to 0.5 g/1 TCA were reported to have
been examined at terminal sacrifice, 37 mice were used for tumor analysis. Finally, for the 42
control animals examined for tumor pathology in the control group, 34 were examined at
terminal sacrifice. Clearly more animals were included in the analyses of tumor incidence and
multiplicity than were sacrificed at the end of the experiment. What effect differential addition
of the results from mice not sacrificed at 104 weeks and the selection bias that may have resulted
from their inclusion on these results cannot be determined. Not only were the background levels
of tumors reported to be increased in the control animals in Study #3 compared to Study #2 at
104 weeks, but the rate of unscheduled deaths was doubled. This is also an expected
consequence of using much larger mice (Leakey et al 2003b).
For the 35 mice examined after 0.05 g/1 TCA in Study # 3, the incidence and multiplicity
of adenomas was reported to be 23% and 0.34 ± 0.12, respectively. For carcinomas, the
incidence and multiplicity was reported to be 40% and 0.71 ±0.19, respectively, and for the
incidence and multiplicity of adenomas and carcinomas combined reported to be 57% and 1.11 ±
0.21, respectively. For the 37 mice examined after 0.5 g/1 TCA in Study # 3, the incidence and
multiplicity of adenomas was reported to be 51% and 0.78 ±0.15, respectively. For carcinomas,
the incidence and multiplicity was reported to be 78% and 1.46 ± 0.21, respectively, and for the
incidence and multiplicity of adenomas and carcinomas combined reported to be 87% and 2.14 ±
0.26, respectively. Thus, at 0.5 g/1 TCA the results presented for this study for the "104 week"
liver tumor data were significantly increased over the reported control values However, these
results are identical to those reported in Study #3 for a 10-fold higher concentration of TCA (4.5
g/1 TCA) for the same 104 weeks of exposure but in the much larger mice. Of the 36 animals
exposed to 4.5 g/1 TCA in Study #2 and included in the tumor analysis, 30 animals were reported
to be examined at 104 weeks. The incidence and multiplicity of adenomas was reported to be
59% and 0.61 ± 0.16, respectively. For carcinomas, the incidence and multiplicity was reported
to be 78%) and 1.50 ± 0.22, respectively, and for the incidence and multiplicity of adenomas and
carcinomas combined reported to be 89% and 2.11 ± 0.25, respectively.
The importance of selection and determination of the control values for comparative
purposes of tumor induction are obvious from this data. The very large difference in control
values between Study #2 and Study #3 is the determinant of the magnitude of the dose response
for TCA after 104 weeks of exposure. The tumor response for 0.5 g/1 and 4.5 g/1 TCA exposure
between the two experiments was identical. Therefore only the background tumor rate
determined the magnitude of the response to treatment. If a similar control values (i.e., a
historical control value) were used in these experiments, there would appear to be no difference
in TCA-tumor response between 0.5 and 4.5 g/1 TCA at 104 weeks of exposure. DeAngelo et al
(1999) report for male B6C3F1 mice exposed only water for 79 to 100 weeks the incidence of
carcinomas to be 26% and multiplicity to be 0.28 lesions/mouse. For 100-week data, the
incidence and prevalence of adenomas was reported to be 10% and 0.12 ± 0.05 and for
carcinomas to be 26% and 0.28 ± 0.07. Issues with reporting for that study have already been
02/12/09
E-154
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
discussed in Section 2.3.2.5. However, the data for DeAngelo et al. (1999) are more consistent
with the control data for "1.5 g/1 HAC" for Study #2 in which there were 0% adenomas and 12%
carcinomas with a multiplicity of 0.20 ± 0.12, than for the control data for Study #3 in which
64% of the control mice were reported to have adenomas and carcinomas and the multiplicity
was 0.93 ± 0.12. If either the control data from DeAngelo et al. (1999) or Study #2 were used
for comparative purposes for the TCA-treatment results of Study #2 or #3, there would be a
dose-response between 0.05 and 0.5 g/1 TCA but no difference between 0.5 g/1 TCA and 4.5 g/1
TCA after 100 weeks of exposure. The tumor incidence would have peaked at ~ 90% in the 0.5
and 4.5 g/1 TCA exposure groups. These results would be more consistent with the 60-week
results in Study #1 in which 0.5 and 5 g/1 TCA exposure groups already had incidences of 38%
and 55%) of adenomas and carcinomas combined, respectively, compared to the 13% control
level. With increased time of exposure the differences between the two highest TCA exposure
concentrations may diminish as tumor progression is allowed to proceed further. However, the
use of the larger and more tumor prone mice in Study #3 also increases the tumor incidence at
the longer period of study.
The authors also presented data for multiplicity of combined adenomas or carcinomas for
mice sacrificed at weeks 26, 52, and 78 for Study #3 (n= 8 per group). No indication of
variability of response, incidence data, statistical significance, or data for adenomas vs.
carcinomas, or the incidence of adenomas was reported. The authors reported that "neoplastic
lesions were first found in the control and 0.05 g/1 TCA groups at 52 weeks. At 78 weeks,
adenomas or carcinomas were found in all groups (0.29, 0.20, and 0.57 tumors/animals for
control, 0.05 g/1 TCA, and 0.5 g/1 TCA, respectively)." Because no other data were presented at
the 52 and 78 week time points in this study, these results cannot be compared to those presented
for Study #1 which was conducted for 60 weeks. Of note, the results presented from Study #1
for 60 weeks of exposure to control, 0.05 g/1 or 0.5 g/1 TCA exposure in 27-30 mice show a 13%,
15%) and 38% incidence of hepatocellular adenomas and carcinomas and a multiplicity of 0.13 ±
0.06, 0.19 ± 0.09, and 0.52 ± 0.14, respectively. Both the incidence and multiplicity of adenomas
were 2-fold higher in the 0.05 g/1 TCA treatment group than for the control. However, the
interim data presented by the authors from Study #3 for 52 weeks of exposure in only 8 mice per
group gives a higher multiplicity of adenomas and carcinomas for control animals (-0.25) than
for either 0.05 g/1 TCA or 0.5 g/1 TCA treatments. Again, comparisons between Study #2 and #3
are difficult due to difference in mouse weight.
Of note, there are no descriptions given in this report in regard to the phenotype of the
tumors induced by TCA or for the liver tumors reported to occur spontaneously in control mice.
Such information would have been of value as this study reports results for a range of TCA
concentration and for 60 and 100 weeks of exposure. Insight could have been gained as to the
effects of differing concentrations of TCA exposure, whether TCA-induced liver tumors had a
similar phenotype as those occurring spontaneously, as well as information in regard to effects
on tumor progression and heterogeneity.
02/12/09
E-155
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Although only examining tissues from 5 mice from the control and high-dose groups only
at 104 weeks at organ sites other than the liver, the authors report that "neoplastic lesions at 104
weeks (Studies #2 and #3) at organ sites other than the liver were found in the lung, spleen,
lymph nodes, duodenum (lymphosarcoma), seminal vesicles, skin, and thoracic cavity of control
and treated animals. All were considered spontaneous for the male B6C3F1 mouse and did not
exceed the tumor incidences when compared to a historical control database (Haseman et. al.,
1984; NIEHS, 1998)." No data was shown. The limitations involved in examining only 5
animals in the control and high-dose groups, and the need to examine the concurrent control data
in each experiment, especially given the large variation in liver tumor response between long-
term studies carried out in the two different laboratories used for Study #2 and Study #3 using
the same strain and gender of mouse, make assertions regarding extrahepatic carcinogenicity of
TCA from this study impossible to support.
A key issue raised from this study is whether changes in any of the parameters measured
in interim sacrifice periods before the appearance of liver tumors (i.e., 4-15 weeks)
corresponded to the induction of liver tumors. The first obstacle for determining such a
relationship is the experimental design of these studies in which only a full range of TCA
concentrations is treated for 60 weeks of exposure with a small number of animals available for
determination of a carcinogenic response (i.e., 30 animals or less in Study #1) and a very small
number of animals (n=5 group) examined for other parameters. Also as stated above, PCO
activity was highly variable between controls and between treatment groups (e.g., the PCO
activity for Study # 1 and #2 at ~ 5 g/1 exposure for 15 weeks). On the other hand, most of the
animals that were examined at terminal sacrifice were also utilized for the tumor results without
the differential deletion or addition of "extra" animals for the tumor analysis. For the 60-week
data in Study # 1 there appeared to be a consistent dose-related increase in the incidence and
multiplicity of tumors after TCA exposure. The TCA-induced increases in liver tumor responses
can be compared with both increased liver weight and PCO activity that were also reported to be
increased with TCA dose as earlier events. Although the limitations of determining the exact
magnitude of responses has already been discussed, as shown below, the incidence and
multiplicity of adenomas show a dose-related increase at 60 weeks. However, the magnitude of
differences in TCA concentrations was not similar to the magnitude of increased liver tumor
induction by TCA after 60 weeks of exposure.
First of all, the greater occurrence of TCA-induced increases in adenomas than
carcinomas reported after 60 weeks of exposure would be expected for this abbreviated duration
of exposure as they would be expected to occur earlier than carcinomas. For adenoma induction,
there was a ~ 2-fold increase between the 0.05 g/1 dose of TCA and the control group for
incidence (7% vs. 15%) and multiplicity (0.07 vs. 0.15 tumors/animals). However, an additional
10-fold increase in TCA dose (0.5 g/1) only resulted in a reported 1.8-fold greater incidence
(15% vs. 21%) and 2.2-fold increase in multiplicity (0.15 vs. 0.24 tumors/animal) of control
adenoma levels. An additional 10-fold increase in dose (5.0g/l vs. 0.5 g/1 TCA) resulted in a 2.2-
fold increase in incidence (21% vs. 38%) and 2.9-fold increase in multiplicity (0.24 vs. 0.55
tumors/animal) of control adenoma levels. Thus, a 100-fold difference in TCA exposure
02/12/09
E-156
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
concentration resulted in differences of 4-fold of control incidence and 6-fold of control
multiplicity for adenomas. For adenomas or carcinomas combined (a parameter that included
carcinomas for which only the two highest exposure levels of TCA were reported to increase
incidence and multiplicity) the incidences were reported to be 13%, 15% 38%, and 55%, and the
multiplicity reported to be 0.13, 0.19, 0.52, and 1.00 for control, 0.05, 0.5, and 5.0 g/1 TCA at 60
weeks. For multiplicity of adenomas or carcinomas, the 0.05 g/1 TCA exposure induced a 1.5-
fold increase over control. An additional 10-fold increase in TCA (0.5 g/1) induced a 6-fold
increase in tumors/animal. An additional 10-fold increase in TCA (5.0 g/1 vs. 0.5 g/1) induced an
additional 2.2-fold increase in tumors/animal. Therefore, using combinations of adenomas or
carcinomas, there was a 13-fold increase in multiplicity that corresponded with a 100-fold
increase in dose.
The results for adenoma induction at 60 weeks of TCA exposure (i.e., ~ 2-fold increased
incidences and 2- to 3-fold increases in multiplicity with 10-fold increases in TCA dose) are
similar to the ~ 2-fold increase in liver weight gain resulting from 10-fold differences in dose
reported at 4-weeks of exposure. For PCO activity there was a ~ 30% increase in PCO activity
from control at 0.05 g/1 TCA. A 10-fold increase in TCA exposure concentration (0.5 g/1)
resulted in an additional ~ 5-fold increase in PCO activity. However another 10-fold increase in
TCA concentration (0.5 vs. 5 g/1) resulted in a 3-fold increase in PCO activity. The 100-fold
increase in TCA dose (0.05 vs. 5 g/1 TCA) was correlated with a 14-fold increase in PCO
activity. For 15 weeks of TCA exposure there was no difference in 0.05 and control PCO
activity and only a 30% difference between the 0.05 and 0.5 g/1 TCA exposures. There was a 7-
fold difference in PCO activity between the 0.5 and 5.0 g/1 TCA exposure concentrations. The
increases in PCO activity and liver weight data at 15-weeks did not fit the magnitude of increases
in tumor multiplicity or incidence data at 60 weeks as well as did the 4-week data. However, the
TCA-induced increase in tumors at 60 weeks (especially adenomas) seemed to correlate more
closely with the magnitude of liver weight increase than for PCO activity at both 4 and 15 weeks.
In regard to Studies #1 and #2 there are consistent periods of study for % liver/body
weight with the consistency of the control values being a large factor in the magnitude of TCA-
induced liver weight increases. As discussed above, there were differences in the magnitude of
% liver/body weight increase at the same concentration between the two studies (e.g., a 1.47-fold
of control % liver/body weight in the 5 g/1 TCA exposed group in Study #1 and 1.60-fold of
control in Study #2 at 15 weeks). For the two studies that had extended durations of exposure
(Studies #2 and #3) the earliest time period for comparison of % liver/body weight is 26 weeks
(Study #3) and 30 weeks (Study #2). If those data sets (26 weeks for Study # 3 and 30 weeks for
Study # 2) are combined, 0.05, 05, and 4.5 g/1 TCA gives a % liver body/weight increase of 1.07-
fold, 1.18-fold and 1.40-fold over concurrent control levels. Using this parameter, there appears
to be a generally consistent pattern as that reported for Study #1 at weeks 4 and 15. Generally, a
10-fold increase in TCA exposure concentration resulted in ~ 2.5-fold increased in additional
liver weight observed at ~ 30 weeks of exposure which correlated more closely with adenoma
induction at 60 weeks than did changes in PCO activity. A similar comparison between Studies
02/12/09
E-157
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
of longer duration (Studies #2 and #3) could not be made for PCO activity as data were not
reported for Study #3.
For 104-week studies of TCA-tumor induction (Studies #2 and #3) the lower TCA
exposure levels (0.05 and 0.5 g/1 TCA) were assayed in a separate experiment and by a separate
laboratory than the high dose (5.0 g/1 TCA) and most importantly in larger more tumor prone
mice. The total lack of similarity in background levels of tumors in Study #2 and #3, the
differences in the number of animals included in the tumor analyses, and the low number of
animals examined in the tumor analysis at 104 weeks (less than 30 for the TCA treatment
groups) makes the determination of a dose-response TCA-induced liver tumor formation after
104-weeks of exposure problematic. The correlation of % liver/body weight increases with
incidence and multiplicity of liver tumors in Study #1 and the similarity of dose-response for
early induction of % liver/body weight gain between Study #1 suggest that there should be a
similarity in tumor response. However, as noted above, the 104-week studies had very
difference background rates of spontaneous tumors reported in the control mice between Study#
2 and #3.
The table below shows the incidence and multiplicity data for Studies #2 and #3 along
with the control data for DeAngelo et al. (1999) for the same paradigm. It also provides an
estimate of the magnitude of increase in liver tumor induction by TCA treatments if the control
values from the DeAngelo et al. (1999) data set were used as the background tumor rate. As
shown below, the background rates for Study #2 are more consistent with those of DeAngelo et
al (1999). Whereas there was a 2:1 ratio of multiplicity for adenomas and adenomas and
carcinomas between 0.5 and 5.0 g/1 TCA after 60 weeks of exposure, there was no difference in
any of the data (i.e., adenoma, carcinoma, and combinations of adenoma and carcinoma
incidence and multiplicity) for these exposure levels in Study #2 and #3 for 104 weeks. The
difference in the incidences and multiplicities for all tumors was 2-fold between the 0.05 and 0.5
g/1 TCA exposure groups in Study #2. These results are consistent with the two highest
exposure levels reaching a plateau of response with a long enough duration of exposure (~ 90%
of animals having liver tumors) and with the 2-fold difference in liver tumor induction between
concentrations of TCA that differed by 10-fold, reported in Study #1.
If either the control values for Study #2 or the control values from DeAngelo et al (1999)
were used for as the background rate of spontaneous liver tumor formation, the magnitude of
liver tumor induction by the 0.05 g/1 TCA over control levels differs dramatically from that
reported as control tumor rates in Study #3. To put the 64% incidence data for carcinomas and
adenomas reported in DeAngelo et al. (2008) for the control group of study #3 in context, other
studies cited in this review for B6C3F1 mice show a much lower incidence in liver tumors in
that: (1) the NCI (1976) study of TCE reports a colony control level of 6.5% for vehicle and
7.1% incidence of hepatocellular carcinomas for untreated male B6C3F1 mice (n= 70 - 77) at 78
weeks, (2) Herren-Freund et al (1987) report a 9% incidence of adenomas in control male
B6C3F1 mice with a multiplicity of 0.09 ± 0.06 and no carcinomas (n= 22) at 61 weeks, (3)
NTP (1990) report an incidence of 14.6% adenomas and 16.6% carcinomas in male B6C3F1
02/12/09
E-158
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mice after 103 weeks (n= 48), and (4) Maltoni et al (1986) report that B6C3F1 male mice from
the "NCI source" had a 1.1% incidence of "hepatoma" (carcinomas and adenomas) and those
from "Charles River Co." had a 18.9% incidence of "hepatoma" during the entire lifetime of the
mice (n= 90 per group). The importance of examining an adequate number of control or treated
animals before confidence can be placed in those results in illustrated by Anna et al (1994) in
which at 76 weeks 3/10 control male B6C3F1 mice that were untreated and 2/10 control animals
given corn oil were reported to have adenomas but from 76 to 134 weeks, 4/32 mice were
reported to have adenomas (multiplicity of 0.13 ± 0.06) and 4/32 mice were reported to have
carcinomas (multiplicity of 0.12 ± 0.06).
Using concurrent control values reported in Study #3, there is no increase in incidence of
multiplicity of adenomas and carcinomas for the 0.05g/l exposure group. However, compared to
either the control data from DeAngelo et al. (1999) or the control data from Study #3, there is a ~
2-3-fold or ~ 5-fold increased in incidence or multiplicity of liver tumors, respectively. Thus,
trying to determine a correspondence with either liver weight increases or increases in PCO
activity at earlier time points will be depend on the confidence placed in the concurrent control
data reported in Study #3 in the 104 week studies. As noted previously, the use of larger tumor
prone mice in Study #3 limits its usefulness to determine the dose-response for TCA.
The authors provide a regression analysis for "tumors/animal" or multiplicity as a percent
of control values and PCO activity for the 60-week and 104-week data. Whether adenomas and
carcinomas combined or individual tumor type were used was not stated. Also comparing PCO
activity at the end of the experiments, when there was already a significant tumor response rather
than at earlier time points, may not be useful as an indicator of PCO activity as a key event in
tumorigenesis. A regression analysis of this data is difficult to interpret because of the dose
spacing of these experiments as the control and 5 g/1 exposure levels will basically determine the
shape of the dose response curve. The 0.05 g/1 and 0.5 g/1 exposure groups in the regression
were so close to the control value in comparison to the 5 g/1 exposure, that the dose response will
appear linear between control and the 5.0 g/1 value with the two lowest doses not affectly
changing the slope of the line (i.e., "leveraging" the regression). The value of this analysis is
limited by (1) the use of tumor prone larger mice in Study #3 that had large background rates of
tumors which make inappropriate the apparent combination of results from study #2 and #3 for
the multiplicity as percentages of control values (2) the low and varying number of animals
analyzed for PCO values and the variability in PCO control values (3) the appropriateness of
using PCO values from later time points, and (4) the dose-spacing of the experiment.
Similarly, the authors report a regression analysis that compares "percent of
hepatocellular neoplasia" which again is indicated by tumor multiplicity with TCA dose as
represented by mg/kg/day. This regression analysis also is of limited value for the same reasons
as that for PCO with added uncertainty as the exposure concentrations in drinking water have
been converted to an internal dose and each study gave different levels of drinking water with
one study showing a reduction of drinking water at the 5 g/1 level. The authors attempt to
identify a NOEL for tumorigenicity using tumor multiplicity and TCA dose. However, it is not
02/12/09
E-159
-------
1
2
3
4
5
6
7
8
9
10
11
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
an appropriate descriptor for these data, especially given that "statistical significance" of the
tumor response is the determinant of the conclusions regarding a dose in which there is no TCA-
induced effect. Only the 60-week experiment (i.e., study # 1) is useful for the determination of
tumor dose-response due to the issues related to appropriateness of control in study #3. A power
calculation of the 60-week study shows that the type II error, which should be > 50% and thus
greater than the chances of "flipping a coin", was 41% and 71% for incidence and 7% and 15%
for multiplicity of adenomas for the 0.05 g/1 and 0.5 g/1 TCA exposure groups. For the
combination of adenomas and carcinomas, the power was 8% and 92% for incidence and 6% and
56%) for multiplicity at 0.05 g/1 and 0.5 g/1 TCA exposure. Therefore, the designed experiment
could accept a false null hypothesis, especially in terms of tumor multiplicity, at the lower
exposure doses and erroneously conclude that there is no response due to TCA treatment.
02/12/09
E-160
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
TCA-induced increases in liver tumor occurrence and other parameter over control after 60 weeks
(Study #1)
Dose TCA Adenomas Adenomas or carcinomas % liver/body weight PCO activity
g/1 Incidence Multiplicity Incidence Multiplicity 4-week 15-week 4-week 15-week
NaCl 7% 0.07 13% 0.13
0.05 15% (2.1-fold) 0.15 (2.1-fold) 15% (1.2-fold) 0.19(1.5-fold) 1.09-fold 1.14-fold 1.3-fold 1.0-fold
0.5 21% (3.0-fold) 0.24 (3.4-fold) 38% (2.9-fold) 0.52 (4.0-fold) 1.16-fold 1.16-fold 2.4-fold 1.3-fold
5.0 38% (5.4-fold) 0.55 (7.9-fold) 55% (4.2-fold) 1.00 (7.7-fold) 1.35-fold 1.47-fold 5.3-fold 3.2-fold
TCA-induced increases in liver tumor occurrence after 104 weeks (Studies # 2and #3)
Dose TCA
Adenomas
Carcinomas
Adenomas or Carcinomas
Incidence
Multiplicity
Incidence
Multiplicity
Incidence
Multiplicity
Study #3
1.5 g/1 HAC (H20?)
21%
0.21
55%
0.74
64%
0.93
0.05 g/1 TCA
23%
0.34
40%
0.71
57%
1.11
(1.1-fold)
(1.6-fold)
(0.7-fold)
(1.0-fold)
(0.9-fold)
(1.2-fold)
0.5 g/1 TCA
51%
0.78
78%
1.46
87%
2.14
(2.4-fold)
(3.7-fold)
(1.4-fold)
(2.0-fold)
(1.4-fold)
(2.3-fold)
Study # 2
2.0 g/1 NaCl (HAC?) 0%
4.5 g/1 TCA 59%
(?)
0
0.61
(?)
12%
78%
(6.5-fold)
0.20
1.50
(7.5-fold)
12%
89%
(7.4-fold)
0.20
2.14
(11-fold)
DeAngelo et al. (1999)
H20 10% 0.12 26% 0.28
0.05g/TCA (S #3) (2.3-fold) (2.8-fold) (1.5-fold) (2.5-fold)
02/12/09
E-161
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1 0.5 g/1 TCA (S #3) (5.1-fold) (6.5-fold) (3.0-fold) (5.2-fold)
2 5.0 g/1 TCA (S #2) (5.9-fold) (6.5-fold) (3.0-fold) (5.4-fold
02/12/09
E-162
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.3.2.14. DeAngelo et al. 1997
The design of this study appears to be similar to that of DeAngelo et al. (2008) but to have
been conducted in F344 rats. 28-30 day old rats that were reported to be of similar weights were
exposed to 2.0 g/1 NaCl, 0.05, 0.5, or 5.0 g/1 TCA in drinking water for 104 weeks. There were
groups of animals sacrificed at 15, 30, 45 and 60 weeks (n= 6) for PCO analysis. There were 23,
24, 19, and 22, animals reported to be examined at terminal sacrifice at 104 weeks and 23, 24, 20,
and 22 animals reported to be used in the liver tumor analysis reported by the authors for the
control, 0.05, 0.5, and 5.0 g/1 treatment groups, respectively. Complete pathological exams were
reported to be performed for all tissues from animals in the high dose TCA group at 104 weeks.
No indication is given as to whether a complete necropsy and pathological exam was performed
for controls at terminal sacrifice. Tritiated thymidine was reported to be administered at interim
sacrifices five days prior to sacrifice and to be examined with autoradiography. The 5 g/1 TCA
treatment group was reported to have a reduction in growth to 89.3% of controls.
For water consumption TCA vs. reported to slightly decrease water consumption at all
doses with a 7%, 8%, and 4% decrease in water consumption reported for 0.05, 0.5 and 5.0 g/1
TCA, respectively. Body weight was decreased by 5.0 g/1 TCA dose only through 78 weeks of
exposure to 89.3% of the control value. All of the % liver/body weight ratios were reported to be
slightly decreased (1-4%) by all of the exposure concentrations of TCA but the data shown does
not indicate if the liver weight data was taken at interim sacrifice times and appears to be only for
animals at terminal sacrifice of 104 weeks.
No data was shown for hepatocyte proliferation but the authors reported no TCA treatment
effects. For PCO there was a 2.3-fold difference between control values between the 15-week
and 104-week data. For the 0.05 and 0.5 g/1 TCA treatment groups there was not a statistically
significant difference reported between control and treated group PCO levels. At 15 weeks the
PCO activity was reduced by 55%, increased to 1.02-fold, and increased 2.12-fold of control for
0.05, 0.5 and 5.0 g/1 TCA exposures, respectively. For the 30 week exposure groups, the 0.05 and
0.5 g/1 TCA groups were reported to have PCO levels within 5% of the control level. However,
for the 5.0 g/1 TCA treatment groups there was ~2-fold of control PCO activity at the 15, 30, 45
and 60 weeks and at 104 weeks there was a 4-fold of control PCO activity. Of note is that the
control PCO value was lowest at 104 weeks while the TCA treatment group was similar to
interim values.
For analysis of liver tumors, there were 20-24 animals examined in each group. Unlike
the study of DeAngelo et al.(2008), it appeared that most of the animals that were sacrificed at
104 weeks were used in the tumor analysis without addition of "extra" animals or deletion of
animal data. The incidence of adenomas was reported to be 4.4%, 4.2%, 15%, and 4.6% and the
incidence of hepatocellular carcinomas was reported to be 0, 0, 0, and 4.6% for the control, 0.05,
0.5, and 5.0 g/1 TCA exposure groups. The multiplicity or tumors/animal was reported to be 0.04,
0.08, 0.15, and 0.05 for adenomas and 0, 0, 0, and 0.05 for carcinomas for the control, 0.05, 0.5,
and 5.0 g/1 TCA exposure groups. Although there was an increase in the incidence of adenomas
02/12/09
E-163
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
at 0.5 g/1 and an increase in carcinomas at 5.0 g/1 TCA, they were not reported to be statistically
significant by the authors. Neither were the increase in adenoma multiplicity at the 0.05 and 0.5
g/1 exposures. However, using such a low number of animals per treatment group (n= 20 -24)
limits the ability of this study to determine a statistically significant increase in tumor response
and to be able to determine that there was no treatment-related effect. A power calculation of the
study shows that the type II error, which should be > 50% and thus greater than the chances of
"flipping a coin", was less than 6% for incidence and multiplicity of tumors at all exposure DCA
concentrations with the exception of the incidence of adenomas for 0.5 g/1 treatment group
(58.7%). Therefore, the designed experiment could accept a false null hypothesis, especially in
terms of tumor multiplicity, at the lower exposure doses and erroneously conclude that there is no
response due to TCA treatment. Thus, while suggesting a lower response than for mice for TCA-
induced liver tumors, the study is inconclusive for determination of whether TCA induces a
carcinogenic response in the liver of rats. The experimental design is such that extrahepatic
carcinogenicity of TCA in the male rat cannot be determined.
2.3.2.15. DeAngelo et al., 1996.
In this study, 28-day-old male F344 rats were given drinking water containing DCA at
concentrations of 0, 0.05, 0.5, or 5.0 g/1 with another group was provided water containing 2.0 g/1
NaCl for 100 weeks. This experiment modified its exposure protocol due to toxicity (peripheral
neuropathy) such that the 5.0 g/1 group was lowered to 2.5 g/1 at 9 weeks and then 2.0 g/1 at 23
weeks and finally to 1.0 g/1 at 52 weeks. When the neuropathy did not reverse or diminish, the
animals were sacrificed at 60 weeks and excluded from the results. Based on measured water
intake in the 0, 0.05 and 0.5 g/L groups, the time-weighted average doses were reported to be 0,
3.6, and 40.2 mg/kg-day respectively. This experiment was conducted at an EPA laboratory in
Cincinnati and the controls for this group were given 2.0 g/1 NaCl (study #1). In a second study
rats were given either deionized water or 2.5 g/1 DCA which was also lowered to 1.5 g/1 at 8
weeks and to 1.0 g/1 at 26 weeks of exposure (study #2).
Although 23 animals were reported to be sacrificed at terminal sacrifice that had been
given 2 g/1 NaCl, the number of animals reported to be examined in this group for hepatocellular
lesions was 3. The incidence data for this group for adenomas was 4.4% so this is obviously a
typographical error. The number of rats included in the water controls for tumor analysis was
reported to be 33 which was the same number as those at final sacrifice. The number of animals
at final sacrifice was reported to be 23 for 2 g/1 NaCl, 21 for 0.05 g/1 DCA, 23 for 0.5g/l DCA in
experiment #1 and 33 for deionized water and 28 for the initial dose of 2.5 g/1 DCA in experiment
#2. Although these were of the same strain, the initial body weight was 59.1 g vs. 76 g for the 2.0
g/1 control group vs. deionized water group. The treatment groups in both studies were similar to
the deionized water group. The % liver/body weights were greater (4.4% vs. 3.7% in the NaCl
vs. deionized water control groups (~ 20%). The number of unscheduled deaths was greater in
study # 2 (22%) than in study # 1 (12%). Interim sacrifice periods were conducted.
02/12/09
E-164
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
As with the DeAngelo et al. (2008) study in mice, the number of animals reported at final
sacrifice was not the same as the number examined for liver tumors in study # 1 (5 more animals
examined than sacrificed at the 0.05 g/1 DCA and 6 more animals examined than sacrificed at the
0.5 g/1 DCA exposure groups) with n= 23, n= 26, and n= 29 for the 2g/l NaCl, 0.05 g/1 DCA and
0.5 g/1 DCA groups utilized in the tumor analysis . For study #2 the same number of rats was
reported to be sacrificed as examined. The source of the extra animals for tumor analysis in study
#1, whether from interim sacrifice or unscheduled deaths, was not given by the authors and is
unknown. Carcinomas prevalence data was not reported for the control group or 0.05 g/1 DCA
group in study #1 and multiplicity data was not reported to the control group, or 0.05 g/1 DCA
group. Multiplicity was not reported for adenomas in the 0.05 g/1 DCA group in Study # 1.
There was a lack of hepatocyte DNA synthesis and necrosis reported at any dose group
carried out to final sacrifice at 100 weeks. The authors reported that the incidence of adenomas to
be 4.4% in 2 g/1 NaCl control, 0 in 0.05 g/1 DCA, and 17.2% in the 0.5 g/1 DCA exposure groups.
For carcinomas no data was reported for the control or 0.05 g/1 DCA group but an incidence of
10.3%) was reported for the 0.5 g/1 DCA group. The authors reported increased hepatocellular
adenomas and carcinomas in male F344 rats although not data was reported for carcinomas in the
control and 0.05 g/1 exposure groups. They reported that for 0.5 g/1 DCA 24.1 % vs. 4.4%>
adenomas and carcinomas combined (Study # 1) and 28.6 vs. 3.0% (study #2) at what was
initially 2.5 g/1 DCA but continuously reduced). Tumor multiplicity was significantly was
reported to be increased in the 0.5 g/1 DCA group (0.04 adenomas and carcinomas/animal in
control vs. 0.31 in 0.5 g/1 DCA in study #1 and 0.03 in control vs. 0.36 in what was initially 2.5
g/1 DCA in study # 2). The issues of use of a small number of animals, additional animals for
tumor analysis in study # 1, and most of all the lack of a consistent dose for the 2.5 g/1 animals in
study # 2, are obvious limitations for establishment of a dose-response for DCA in rats.
2.3.2.16. Richmond et al., 1995
This study was conducted by the same authors as DeAngelo et al. (1996) and appears to
report results for the same data set for the 2 g/1 NaCl control, 0.05 g/1 DCA and 0.5 g/1 DCA
exposed groups. Of note is that while DeAngelo et al (1996) refer to the 28-day old rats as
"weanlings" the same aged rats are referred to as "adults" in this study. Male Fischer 344 rats
were administered time-weighted average concentrations of 0, 0.05, 0.5, or 2.4 g/1 DCA in
drinking water. Concentrations were kept constant but due to hind-limb paralysis all 2.4 g/1 DCA
exposed rats had been sacrificed by 60 weeks of exposure. In the 104-week sacrifice time, there
were 23 rats reported to be analyzed for incidence of hepatocellular adenomas and carcinomas in
the control group, 26 rats in the 0.05 g/1 DCA group and 29 rats in the 0.5 g/1 DCA exposed
group. This is the same number of animals included in the tumor analysis reported in DeAngelo et
al. (1996). Tumor multiplicity was not given. Richmond et al (1995) reported that there was a
4% incidence of adenomas reported in the 2.0 g/1 NaCl control animals, 0%> at 0.05 g/1 DCA, and
21%) in the 0.5 DCA group at 104 weeks. These figures are similar to those reported by
DeAngelo et al (1996) for the same data set with the exception of a 17.2%> incidence of adenomas
reported for the 0.5 g/1 DCA group. There were no hepatocellular carcinomas reported in the
02/12/09
E-165
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
control or 0.05 g/1 exposure groups but a 10% incidence reported in the 0.5 g/1 DCA exposure
group at 104 weeks of exposure. While carcinomas were not reported by DeAngelo et al (1996)
for the control and 0.05 g/1 groups they are assumed to be zero in the summary data for
carcinomas and adenomas combined. The 10% incidence at 0.5 g/1 DCA is similar to the 10.4%
incidence reported for this group by DeAngelo et al (1996). At 60 weeks at 2.4 g/1 DCA, the
incidence of hepatocellular adenoma was reported to be 26% and hepatocellular carcinoma to be
4%. This is not similar to the values reported by DeAngelo for 2.5 g/1 DCA that was
continuously decreased so that the estimated final concentration was 1.6 g/1 DCA for 100 weeks,
for those animals, the incidence of adenomas was reported by DeAngelo et al (1996) to be 10.7%
and carcinomas 21.4%, probably more a reflect of longer exposure time allowing for adenoma to
carcinoma progression. The authors did not report any of the results of DCA-induced increases of
adenomas and carcinomas to be statistically significant. As it appears the same data set was used
for the 2.g/1 NaCl control, 0.05 g/1 DCA and 0.5 g/1 DCA exposure groups as was reported in
DeAngelo et al (1996), the same issues arise as regarding the differences in numbers of animals
were included in tumor analysis than were reported to have been present at final sacrifice. As
stated previously for the DeAngelo et al., (1997) study of TCA in rats, the use of small numbers
of rats limits the detection of and ability to determine whether there was no treatment-related
effects, especially at the low concentrations of DCA exposure.
2.4. Summaries and Comparisons Between TCE, DCA, and TCA Studies
There are a number of studies to TCE that have reported effects on the liver. However,
the study of this compound is difficult as its concentration does not remain stable in drinking
water, some studies have been carried out using TCE with small quantities of a carcinogenic
stabilizing agent, some studies have been carried out in whole body inhalation chambers that
resulted in additional oral administration and for which individual animal data were not recorded
throughout the experiment, and the results of gavage studies have been limited by gavage related
deaths and vehicle effects. In addition some studies have been conducted using the i.p route of
administration which results in route-related toxicity and inflammation. For many studies, liver
effects consisted of measured increases in liver weight with little or no description of attendant
histological changes induced by TCE treatment. A number of studies were conducted at a few
relatively high doses with attendant effects on body weight, indicative of systemic toxicity and
affecting TCE-induced liver weight gain. Although, many studies have been performed in male
mice, the inhalation studies of Kjellstrand et al. indicate that male mice, regardless of strain
appear to have a greater variability in response, as measured by TCE-induced liver weight gain,
and susceptibility to TCE-induced decreases in body weight than female mice. However, the
body of the TCE literature is consistent in identifying the liver as a target of TCE-induced affects
and with the most commonly reported change to be a dose-related TCE-induced increase in liver
weight in multiple species, strains, and genders from both inhalation and oral routes of exposure.
The following sections will not only summarize results for studies of TCE reported in
Sections 2.1 - 2.2, but provide comparison of studies of either TCA or DCA that have used
similar paradigms or investigated similar parameters described in Sections 2.3.1 and 2.3.2. A
02/12/09
E-166
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
synopsis of the results from studies of chloral hydrate and it comparison with TCE results is
presented in the next section (Section 2.3.5). While the study of Bull et al. (2002), described in
Section 2.2.21, presents data for combinations of DCA or TCA exposure for comparisons of
tumor phenotype with those induced by TCE, the examination of co-exposure studies of TCE
metabolites in rodents that are also exposed to a number of other carcinogens, and descriptions of
the toxicity data for brominated haloacetates that also occur with TCE in the environment, are
presented in Section 4.
2.4.1. Summary of Results For Short-term Effects of TCE.
In regard to early changes in DNA synthesis, the data for TCE is very limited The study
by Mirsalis et al. (1989) used an in vivo-in vitro hepatocyte DNA repair and S-phase DNA
synthesis in primary hepatocytes from male Fischer-344 rats (180-300g) and male and female
B6C3F1 mice (20-29 g for male mice and 18-25 g female mice) administered TCE by gavage in
corn oil. They reported negative results 2-12 hours after treatment from 50-1000 mg/kg TCE in
rats and mice (male and female) for unscheduled DNA synthesis and repair using 3 animals per
group. After 24 and 48 hours of 200 or 1000 mg/kg TCE in male mice (n= 3) and after 48 hours
of 200 (n= 3) or 1000 (n= 4) mg/kg TCE in female mice, similar values of 0.30 to 0.69% of
hepatocytes were reported as undergoing DNA synthesis in those hepatocytes in primary culture
with only the 1000 mg/kg TCE dose in male mice at 48 hours giving a result considered to be
positive (~ 2.2%). No statistical analyses were performed on these measurements which were
obviously limited by both the number of animals examined and the relevance of the paradigm.
TCE-induced increases in liver weight have been reported to occur quickly. The
inhalation study of Okino et al. (1991) in male rats demonstrates that liver weight and metabolism
were increased with as little as 8 hours of TCE exposure (500 and 2000 ppm) and as early as 22
hours after cessation of such exposures with little concurrent hepatic necrosis. Laughter reported
increase liver weight in SV129 mice in their 3-days study (see below). Tao et al. (2000) reported
a 1.26-fold of control % liver/body weight in female B6C3Flmice fed 1000 mg/kg TCE in corn
oil for 5 days. Elcombe et al. (1985) and Dees and Travis (1993) reported gavage results in mice
and rats after 10 days exposure to TCE which showed TCE-induced increases in liver weight (see
below for more detail on dose-response). Tucker et al (1982) reported that 14 days of exposure to
24 mg/kg and 240 mg/kg TCE via gavage to induce a dose-related increase in liver weight in male
CD-I mice but did not show the data.
TCE-induced increases in % liver/body weight ratios have been studied most extensively
in B6C3F1 and Swiss mice. Both strains have been shown to have a TCE-induced increase in
liver tumors from long-term exposure as well (see Section 2.4.2. below). A number of studies
have provided dose-response information for TCE-induced increases in liver weight from 10 days
to 13 weeks of exposure in mice. Most studies have reported that the magnitude of increase in
TCE exposure concentration is similar to the magnitude increase of % liver/body weight increase.
For example a 2-fold increase in TCE exposure has often resulted in a 2-fold increase in the %
change in liver/body weight over control (i.e., 500 mg/kg TCE induces a 20% increase in liver
02/12/09
E-167
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
weight and 1000 mg/kg TCE induces a 50% increase in liver weight as reported by Elcombe et
al., 1985). The range in which this relationship is valid has been reported to vary from 100 mg/kg
TCE at 10 days (Dees and Travis, 1993) to 1600 mg/kg (Buben and O'Flaherty, 1985) at 6 weeks
and up to 1500 mg/kg TCE for 13 weeks (NTP, 1990). The consistency in the relationship
between magnitude of liver weight increase and TCE exposure concentration has been reported
for both genders of mice, across oral and inhalation routes of exposure, and across differing
strains of mice tested. For rats, there are fewer studies with fewer exposure levels tested, but both
Berman et al. (1995) and Melnick et al. (1987) report that short-term TCE exposures from 150
mg/kg to ~ 2000 mg/kg induced % liver/body weight that increased proportionally with the
magnitude of TCE exposure concentration.
Dependence of PPARa activation for TCE-liver weight gain has been investigated in
PPARa null mice by both Nakajima et al. (2000) and Laughter et al (2004). After 2 weeks of 750
mg/kg TCE exposure to carefully matched SV129 wild type or PPARa-null male and female mice
(n= 6 group), there was a reported 1.50-fold of control in wild-type and 1.26-fold of control %
liver/body weight in PPARa-null male mice by Nakajima et al. (2000). For female mice, there
was ~ 1.25-fold of control % liver/body weight ratios for both wild-type and PPARa-null mice.
Thus, TCE-induced liver weight gain was not dependent on a functional PPARa receptor in
female mice and some portion of it may have been in male mice. Both wild-type male and female
mice were reported to have similar increases in the number of peroxisome in the pericentral area
of the liver and TCE exposure and, although increased 2-fold, were still only ~ 4% of cytoplasmic
volume. Female wild type mice were reported to have less TCE-induced elevation of very long
chain acyl-CoA synthetase, D-type peroxisomal bifunctional protein, mitochondrial trifunctional
protein a subunits a and P, and cytochrome P450 4A1 than males mice, even though peroxisomal
volume was similarly elevated in male and female mice. The induction of PPARa protein by
TCE treatment was also reported to be slightly less in female than male wild-type mice (2.17-fold
vs. 1.44-fold of control, respectively).
Laughter et al (2004) also studied SV129 wild type and PPARa-null male mice treated
with 3 daily doses of TCE in 0.1% methyl cellulose for either 3 days (1500 mg/kg TCE) or 3
weeks (0, 10, 50, 125, 500, 1000, or 1500 mg/kg TCE 5 days a week). However, not only is the
paradigm not comparable to other gavage paradigms, but no initial or final body weights of the
mice were reported and thus the influence of differences in initial body weight on % liver/body
weight determinations could not be ascertained. In the 3-day study, while control wild type and
PPARa-null mice were reported to have similar % liver/body weight ratios (~ 4.5%), at the end of
the 3-week experiment the % liver/body weight ratios were reported to be increased in the
PPARa-null male mice (5.1%). TCE treatment for 3 days was reported to increase the %
liver/body weight ratio 1.4-fold of control in the wild type mice and 1.07-fold of control in the
null mice. In the 3-week study, wild-type mice exposed to various concentrations of TCE had %
liver/body weights that were reported to be within ~ 2% of control values except for the 1000
mg/kg and 1500 mg/kg groups ( ~ 1.18 - fold and 1.30 - fold of control levels, respectively). For
the PPARa-null mice the variability in % liver/body weight was reported to be greater than that of
the wild-type mice in most of the groups and the baseline level of % liver/body weight ratio also
02/12/09
E-168
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1.16-fold greater. TCE exposure was apparently more toxic in the null mice with death at the
1500 mg/kg TCE exposure level resulting in the prevention of recording of % liver/body weights.
At 1000 mg/kg TCE exposure level there was a reported 1.10-fold of control % liver/body weight
in the PPARa-null mice. None of the increases in % liver/body weight in the null mice were
reported to be statistically significant by Laughter et al. (2004). However, the statistical power of
the study was limited due to low numbers of animals and increased variability in the null mice
groups. The % liver/body weight after TCE treatment that was reported in this study was actually
greater in the null mice than the wild-type male mice at the 1000 mg/kg TCE exposure level (5.6
± 0.4 % vs. 5.2 ± 0.5%, for null and wild-type mice, respectively). At 1-weeks and at 3-weeks,
TCE appeared to induce increases in liver weight in PPARa-null mice, although not reaching
statistical significance in this study. At a 1000 mg/kg TCE exposure for 3 weeks % liver/body
weights were reported to be 1.18-fold of control in wild-type and 1.10-fold of control in null
mice. Although the experiments in Laughter et al. for DCA and TCA were not conducted using
the same paradigm, the TCE-induced increase in % liver/body weight more closely resembled the
dose-response pattern for DCA than for DCA wild-type SV129 and PPARa-null mice.
Many studies have used cyanide-insensitive palmitoyl Co A oxidase activity (PCO) as a
surrogate for peroxisome proliferation. Of note is that several studies have shown that this
activity is not correlated with the volume or number of peroxisomes that are increased as a result
of exposure to TCE or it metabolites (Nakajima et al 2000; El combe et al 1985: Nelson et al
1989). This activity appears to be highly variable both as a baseline measure and in response to
chemical exposures. Laughter et al. (2004) presented data showing that WY -14,643 induced
increases in PCO activity varied up to 6-fold between experiments in wild-type mice. They also
showed that PCO activity, in some instances, was up to 6-fold of wild type mice values in
untreated PPARa-null mice. Parrish et al. (1996) noted that control values between experiments
varied as much as a factor of 2-fold for PCO activity and thus their data were presented as percent
of concurrent controls. Goldsworthy and Popp (1987) reported that 1000 mg/kg TCE induced a
6.25-fold of control PCO activity in B6C3F1 mice in two 10-day experiments. However for F344
rats, the increases over control between two experiments conducted at the same dose were
reported to vary by > 30%. Finally, Melnick et al. (1987) have reported that corn oil
administration alone can elevate PCO activity as well as catalase activity.
For TCE there are two key 10-days studies (Elcombe et al., 1985; Dees and Travis, 1993)
that examine the effects of short-term exposure in mice and rats via gavage exposure and attempt
to determine the nature of the dose response in a range of exposure concentrations that include
levels below which there is concurrent decreased body weights. Although they have limitations,
they reported generally consistent results. In regard to liver weight in mice, gavage exposure to
TCE at concentrations ranging from 100 to 1500 mg/kg TCE produced increases in liver/body
weight that was dose-related (Elcombe et al., 1985; Dees and Travis, 1993).
Elcombe et al. (1985) reported a small decrease in DNA content with TCE treatment
(consistent with hepatocellular hypertrophy) that was not dose-related, increased tritiated
thymidine incorporation in whole mouse liver DNA that was that was treatment but not dose-
02/12/09
E-169
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
related (i.e., a 2-, 2-, and 5-fold of control values in mice treated with 500, 1000, and 1500 mg/kg
TCE), and slightly increased numbers of mitotic figures that were treatment but not dose-related
and not correlated with DNA synthesis as measured by thymidine incorporation. Elcombe et al
(1985) reported an increase in peroxisome volume after TCE exposure that was correlated with
the magnitude of increase in peroxisomal-associated enzyme activity at the only dose in which
both were tested. Peroxisome increases after TCE treatment in mice livers were identified as
being pericentral in location. After TCE treatment, increased peroxisomal volumes in B6C3F1
mice were reported to be not dose-related (i.e., there was little difference between 500 to 1500
mg/kg TCE exposures). The TCE-induced increases in peroxisomal volumes were also not
correlated with the reported increases in thymidine incorporation or mitotic activity in mice.
Neither TCE-induction of peroxisomes or hepatocellular proliferation, as measured by either
mitotic index or thymidine incorporation, was correlated with TCE-induced liver weight
increases. Elcombe et al. (1985) only measured PCO activity in a subset of B6C3F1 mice at the
1000 mg/kg TCE exposure level for 10 days of exposure and reported an 8-fold of control PCO
activity and a 1.5-fold of control catalase activity. This result was similar to that of Goldsworthy
and Popp (1987) who reported 6.25-fold of control PCO activity in male B6C3F1 mice exposed
to 1000 mg/kg/day TCE for 10 days in two separate experiments.
Similar to Elcombe et al., who reported no difference in response between 500 and 1000
mg/kg TCE treatments, Dees and Travis (1993) reported that incorporation of tritiated thymidine
in DNA from mouse liver was elevated after TCE treatment and the mean peak level of tritiated
thymidine incorporation occurred at 250 mg/kg TCE treatment level remaining constant for the
500 and 1000 mg/kg treated groups. Dees and Travis (1993) specifically report that mitotic
figures, although very rare, were more frequently observed after TCE treatment, found most often
in the intermediate zone, and found in cells resembling mature hepatocytes. They reported that
there was little tritiated thymidine incorporation in areas near the bile duct epithelia or close to the
portal triad in liver sections from both male and female mice. They also reported no evidence of
increased lipofuscin and that increased apoptoses from TCE exposure "did not appear to be in
proportion to the applied TCE dose given to male or female mice" (i.e., the mean number of
apopotosis 0, 0, 0, 1 and 8 for control, 100, 250, 500, and 1000 mg/kg TCE treated groups,
respectively). Both Elcombe et al. (1985) and Dees and Travis (1993) reported no changes in
apoptosis other than increased apoptosis only at a treatment level of 1000 mg/kg TCE.
Elcombe et al. (1985) reported increased in % liver/body weight after TCE treatment in
both the Osborne-Mendel and Alderly Park rat strain, although to a smaller extent than in mice.
For both strains, Elcombe et al. (1985) reported no TCE-induced changes in body weight at doses
ranging from 500 to 1500 mg/kg. For male Osborne-Mendel rats administration of TCE in corn
oil gavage resulted in a 1.18-fold, 1.26-fold, and 1.30-fold of control % liver/body weight at 500
mg/kd/day, 1000 mg/kg/day, and 1500 mg/kg/day exposures, respectively. For Alderly Park rats
those increases were 1.14-fold, 1.17-fold, and 1.17-fold of control at the same respective exposure
levels for 10 days of exposure. In regard to liver weight increases, Melnick et al (1987) reported a
1.13-fold and 1.23-fold of control % liver/body weight in male Fischer 344 rats fed 600
mg/kg/day and 1300 mg/kg/day TCE in capsules, respectively. There was no difference in the
02/12/09
E-170
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
extent of TCE-induced liver increase between the two lowest dosed group administered TCE in
corn oil gavage (~ 20% increase in % liver/body weight at 600 mg/kd and 1300 mg/kg TCE) for
14 days. However, the magnitude of increases in % liver/body weight in these groups was
affected by difference between control groups in liver weight although initial and final body
weights appeared to be similar. By either type of vehicle, Melnick et al. (1987) reported
decreases in body weights in rats treated with concentrations of TCE 2200 mg/kg/day or greater
for 14 days. Similarly, Nunes et al (2001) reported decreased body weight in Sprague-Dawley
rats administered 2000 mg/kg/day for 7 days in corn oil. Melnick et al (1987) reported that both
exposures to either 600 or 1300 mg/kg/day TCE in capsules did not result in decreased body
weight and caused less than minimal focal necrosis randomly distributed in the liver. At 2200 and
4800 mg/kg TCE fed via capsule, Melnick et al. (1987) reported that although there was
decreased body weight in rats treated at these exposures, there was little TCE-induced necrosis,
and no evidence of inflammation, cellular hypertrophy or edema with TCE exposure. Similarly,
Berman et al. (1995) reported increases in liver weight gain at doses as low as 50 mg/kg TCE, no
necrosis up to doses of 1500 mg/kg, and hepatocellular hyper trophy only at the 1500 mg/kg level
in female Fischer 344 rats.
For rats, Elcombe et al. (1985) reported an increase over untreated rats of 1.13-fold of
control PCO activity in Alderly Park rats after 1000 mg/kg/day TCE exposure for 10 days, while
Goldsworthy and Popp (1987) reported a 1.8-fold and 2.39-fold of control in male Fischer 344
rats at the same exposure in two separate experiments. Melnick et al. (1987) reported PCO
activity of 1.23-fold and 1.75-fold of control in male Fischer 344 rats fed 600 mg/kg/day and
1300 mg/kg/day TCE for 14 days in capsules. For rats treated by gavage with 600 mg/kg/day or
1200 mg/kg day TCE corn oil, they reported 1.16-fold and 1.29-fold of control values. However,
control levels of PCO were 16% higher in corn oil controls than in untreated controls. In addition
Melnick et al. (1987) reported little catalase increases in rats fed TCE via capsules in food (less
than 6% increase) but a 1.18-fold and 1.49-fold of control catalase activity in rats fed 600
mg/kg/day or 1200 mg/kg/TCE via corn oil gavage, indicative of a vehicle effect.
The data from Elcombe et al.(1985) included reports of TCE-induced pericentral
hypertrophy and eosinophilia for both rats and mice but with "fewer animals affected at lower
doses." In terms of glycogen deposition, Elcombe report "somewhat" less glycogen pericentrally
in the livers of rats treated with TCE at 1500 mg/kg than controls with less marked changes at
lower doses restricted to fewer animals. They do not comment on changes in glycogen in mice.
Dees and Travis (1993) reported TCE-induced changes to "include an increase in eosinophilic
cytoplasmic staining of hepatocytes located near central veins, accompanied by loss of
cytoplasmic vacuolization." Since glycogen is removed using conventional tissue processing and
staining techniques, an increase in glycogen deposition would be expected to increase
vacuolization and thus the report from Dees and Travis is consistent with less not more glycogen
deposition. Neither study produced a quantitative analysis of glycogen deposition changes from
TCE exposure. Although not explicitly discussing liver glycogen content or examining it
quantitatively in mice, these studies suggest that TCE-induced liver weight increases did not
02/12/09
E-171
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
appear to be due to glycogen deposition after 10 days of exposure and any decreases in glycogen
were not necessarily correlated with the magnitude of liver weight gain either.
For both rats and mice the data from Elcombe et al (1985) showed that tritiated thymidine
incorporation in total liver DNA observed after TCE exposure did not correlate with mitotic index
activity in hepatocytes with both Elcombe et al. (1985) and Dees and Travis (1993) reporting a
small mitotic indexes and evidence of periportal hepatocellular hypertrophy from TCE exposure.
Neither mitotic index or tritiated thymidine incorporation data support a correlation with TCE-
induced liver weight increase in the mouse. If higher levels of hepatocyte replication had
occurred earlier, such levels were not sustained by 10 days of TCE exposure. Both Elcombe et al.
(1985) and Dees and Travis (1993) present data that represent "a snapshot in time" which does
not show whether increased cell proliferation may have happened at an earlier time point and then
subsided by 10 days. These data suggest that increased tritiated thymidine levels were targeted to
mature hepatocytes and in areas of the liver where greater levels of polyploidization occur. Both
Elcombe et al.(1985) and Dees and Travis (1993) show that tritiated thymidine incorporation in
the liver was ~ 2-fold of controls between 250 -1000 mg/kg TCE, a result consistent with a
doubling of DNA. Thus, given the normally quiescent state of the liver, the magnitude of this
increase over control levels, even if a result of proliferation rather than polyploidization, would be
confined to a very small population of cells in the liver after 10 days of TCE exposure. Laughter
et al. (2004) reported that there was an increase in DNA synthesis after aqueous gavage exposure
to 500 and 1000 mg/kg TCE given as 3 boluses a day for 3 weeks with BrdU given for the last
week of treatment. An examination of DNA synthesis in individual hepatocytes was reported to
show that 1% and 4.5% of hepatocytes had undergone DNA synthesis in the last week of
treatment for the 500 and 1000 mg/kg doses, respectively. Both Elcombe et al (1985) and Dees
and Travis (1993) show TCE-induced changes for several parameters at the lowest level tested
without toxicity and without evidence of regenerative hyperplasia or sustained hepatocellular
proliferation. In regards to susceptibility to liver cancer induction, the more susceptible
(B6C3F1) vs. less susceptible (Alderly Park/Swiss) strains of mice to TCE-induced liver tumors
(Maltoni et al., 1988), the "less susceptible" strain was reported by Elcombe et al (1985) to have,
a greater baseline level of liver weight/body weight ratio, a greater baseline level of thymidine
incorporation as well as greater responses for those endpoints due to TCE exposure. However,
both strains showed a hepatocarcinogenic response after TCE exposure, although there are
limitations regarding determination of the exact magnitude of response for these experiments as
previously discussed.
2.4.2. Summary of Results For Short-term Effects of DCA and TCA: Comparisons With
TCE
Short-term exposures from DCA and TCA have been studied either through gavage or in
drinking water. Palatability became an issue at the highest level of DCA tested in drinking water
experiments (5 g/1) which caused a significant reduction of drinking water intake in mice of 46 to
64% (Carter et al 1995). Decreases in drinking water consumption have also been reported for a
range of concentrations of DCA and TCA from 0.05 g/1 to 5.0 g/1, in both mice and rats, and with
02/12/09
E-172
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
generally the higher concentrations producing the highest decrease in drinking water (Carter et al.,
1995; Mather et al., 1990; DeAngelo et al., 1997; DeAngelo et al., 1999; DeAngelo et al., 2008).
However, results within studies (e.g., DeAngelo et al., 2008) and between studies have been
reported to vary as to the extent of the reduction in drinking water from the presence of TCA or
DCA. Some drinking water studies of DC A or TCA have not reported drinking water
consumption as well. Therefore, although in general DCA and TCA studies have do not include
vehicle effects, such as corn oil, they have been affected by differences in drinking water
consumption not only changing the dose received by the rodents and therefore potentially the
shape of the dose-response curve, but also the effects of dehydration are potentially added to any
chemically-related reported effects.
Studies have attempted to determine short-term effects on DNA by TCE and its
metabolites. Nelson and Bull (1988) administered TCE male Sprague Dawley rats and male
B6C3F1 mice measured the rate of DNA unwinding under alkaline conditions 4 hours later. For
rats there was a significantly increased rate of unwinding at the two highest dose and for mice
there was a significantly increased level of DNA unwinding at a lower dose. In this same study,
DCA was reported to be most potent in this assay with TCA being the lowest, while chloral
hydrate (CH) closely approximated the dose-response curve of TCE in the rat. In the mouse the
most potent metabolite in the assay was reported to be TCA followed by DCA with CH
considerably less potent. Nelson and Bull (1988) and Nelson et al (1989) have reported increases
in single strand breaks after DCA and TCA exposure. However, Styles et al (1991) (for mice)
and Chang et al (1992) (for mice and rats) did not. Austin et al. (1996) note that the alkaline
unwinding assay, a variant of the alkaline elution procedure, is noted for its variability and
inconsistency depending on the techniques used while performing the procedure. In regard to
oxidative damage as measured by TBARS for lipid peroxidation and 8-OHdG levels in DNA,
increases appear to be small (less than 50% greater than control levels) and transient after DCA
and TCA treatment in mice (see Section 3.4.2.3) with TCE results confounded by vehicle or route
of administration effects.
Although there is no comparative data for TCE, the study of Styles et al. (1991) is
particularly useful for determining effects of TCA from 1 to 4 days of exposure in mice. Styles et
al (1991) reported no change in "hepatic" DNA uptake of tritiated thymidine up to 36 hours, a
peak at 72 hours (~ 6-fold of control), and falling levels by 96 hours (~ 4-fold of controls) after
500 mg/kg TCA gavage exposure. Incorporation of tritiated thymidine observed for individual
hepatocytes decreased between 24 and 36 hours, rose slowly back to control levels at 48 hours,
significantly increased by 72 hours, and then decreased by 96 hours. Thus, increases in "hepatic"
DNA tritiated thymidine uptake did not capture the decrease observed in individual hepatocytes at
36 hours. By either measure the population of cells undergoing DNA synthesis was small with
the peak level being less than 1% of the hepatocyte population. Zonal distribution of labeled
hepatocytes were decreased at 36 hours in all zones, appeared to be slightly greater in perioportal
than midzonal cells with centrilobular cells still below control levels by 48 hours, similarly
elevated over controls in all zones by 72 hours, and to have returned to near control levels in the
midzonal and centrilobular regions but with periportal areas still elevated by 96 hours. These
02/12/09
E-173
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
results are consistent with all hepatocytes showing a decrease in DNA synthesis by 36 hours and
then a wave of DNA synthesis to occur, starting at the periportal zone and progressing through the
liver acinus that is decreased by 4 days after exposure.
Along with changes in liver weight, DNA synthesis, and glycogen accumulation, several
studies of DCA and TCA have focused on the extent of peroxisome proliferation as measured by
changes in peroxisome number, cytoplasmic volume and enzyme activity induction as potential
"key events" occurring from shorter-term exposures that may be linked to chronic effects such as
liver tumorigenicity. As noted above in Section 2.4.1, TCE-induced liver weight gain has been
reported to not be dependent on a functional PPARa receptor in female mice while some portion
of increased liver weight may have been in male mice. Also as noted cyanide-insensitive
palmitoyl CoA oxidase activity (PCO) has also been reported to not be correlated with the volume
or number of peroxisomes that are increased as a result of exposure to TCE or it metabolites
(Nakajima et al 2000; Elcombe et al 1985: Nelson et al 1989) and to be highly variable both as a
baseline measure and in response to chemical exposures (e.g., variation of up to 6-fold between
after WY-14,643 exposure in mice). Also as noted, above the vehicle used in many TCE gavage
experiments, corn oil, has been reported to elevate PCO activity as well as catalase activity.
A number of short-term studies have examined the effects of TCA and DCA on liver
weight increases and evidence of peroxisome proliferation and changes in DNA synthesis. In
particular two studies of DCA and TCA used a similar paradigm presented by Elcombe et al
(1985) and Dees and Travis (1993) for TCE effects in mice. Nelson et al. (1989) report findings
from gavage doses of unbuffered TCA (500 mg/kg) and DCA (500 mg/kg) in male B6C3F1 mice
and Styles et al. (1991) also providing data on peroxisome proliferation using the same paradigm.
Nelson et al. (1989) reported levels of PCO activity in mice administered 500 mg/kg DCA or
TCA for 10 days with 250 mg/kg Clofibrate administration serving as a positive control. DCA
and TCA exposure were reported to not affect body weight, but both to significantly increase liver
weight (1.63-fold of control for DCA and 1.30-fold of control for TCA treatments), and %
liver/body weight ratios (1.53-fold of control for DCA and 1.16-fold of control for DCA
treatments). PCO activity was reported to be significantly increased by ~ 1.63-fold, 2.7-fold, and
5-fold of control for DCA, TCA and Clofibrate treatments, respectively and indicated that both
DCA and TCA were weaker inducers of this activity than Clofibrate. Results from randomly
selected electron photomicrographs showed an increase in peroxisomes per unit area but gave a
different pattern than PCO enzyme activity (i.e., 2.5-fold and 2.4-fold of control peroxisome
volume for DCA and TCA, respectively). Evidence of gross hepatotoxicity was reported to not
occur in vehicle or TCA-treated mice. Light microscopic sections were reported to show TCA
and control hepatocytes to have the same intensity of PAS staining, but with slightly larger
hepatocytes occurring in TCA-treated mice throughout the liver section with architecture and
tissue pattern of the liver intact. For DCA, the histopathology was reported to be markedly
different than control mice or TCA treated mice. DCA was reported to induce a marked increase
in the size of hepatocytes throughout the liver with an approximately 1.4-fold of control diameter
that was accompanied by increased PAS staining (indicative of glycogen deposition). All DCA-
treated mice were reported to have multiple white streaks grossly visible on the surface of the
02/12/09
E-174
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
liver corresponding with subcapsular foci of coagulative necrosis that were not encapsulated,
varied in size, and accompanied by a slight inflammatory response characterized by neutrophil
infiltration.
A quantitative comparison of effects from equivalent exposures of TCE, TCA, and DCA
(500 mg/kg for 10 days in mice via corn oil gavage for TCE) can be drawn between the Elcombe
et al. (1985), Dees and Travis (1993), Styles et al. (1991), and Nelson et al. (1989) data for
relationship to control values for % liver/body weight, peroxisome volume and activity (PCO),
and qualitatively for glycogen deposition.
Model
% Liver/
Body Wt.
B6C3F1
Male
Peroxisome
Volume
Nelson et al. (1989)**
TCA 1.16-fold 2.4-fold
DCA 1.53-fold 2.5-fold
Peroxisome Glycogen
Enzyme Activity Deposition
2.7-fold
1.63-fold
no change
increased
Styles et al. (1991)
B6C3F1 TCA NR
Male
1.9-fold
NR
NR
Elcombe et al. (1985)
B6C3F1
Male
TCE 1.20-fold
8-fold
NR
NR
Alderly Park TCE 1.43-fold
Male (Swiss)
Dees and Travis (1993)
B6C3F1
Male
TCE 1.05-fold*
4-fold
NR
NR
NR
NR
NR
B6C3F1 TCE 1.18-fold NR
Female
NR
NR
02/12/09
E-175
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
* statistically significant although small increase. ** Unbuffered. NR = not reported as no
analysis was performed for this dose or the authors did not report this finding (i.e., did not note a
change in glycogen in description of exposure-related changes).
Although using a similar species, route of exposure, and dose, the comparison of
responses for TCE and it's metabolites shown above are in male mice and also are reflective of
variability in strain, and variability and uncertainty of initial body weights. As described in more
detail in Section 2.2, initial age and body weight have an impact on TCE-related increases in liver
weight. Male mice have been reported to have greater variability in response than female mice
within and between studies and most of the comparative data for the 10-day 500 mg/kg doses of
TCE or its metabolites were from studies in male mice. Corn oil, used as the vehicle for TCE
gavage studies but not those of its metabolites, has been noted to specifically affect peroxisomal
enzyme induction, body weight gain, and hepatic necrosis, specifically, in male mice (Merrick et
al 1989). Corn oil alone has also been reported to increase PCO activity in F344 rats and to
potentiate the induction of PCO activity of TCA (DeAngelo et al., 1989). Thus, quantitative
inferences regarding the magnitude of response in these studies are limited by a number of
factors.
The variability in the magnitude of TCE-induced increases in % liver/body weight across
studies in readily apparent but for TCE, TCA and DCA there is an increase in liver weight in mice
at this dose after 10 days of exposure. The volume of the peroxisomal compartment in
hepatocytes was reported to be more greatly increased from TCE-treatment by Elcombe et al.
(1985) than for either TCA or DCA by Nelson et al. (1989) or Styles et al. (1991). However, the
control values for the B6C3F1 mice were half that of the other strain reported by Elcombe et al.
(1985) and this parameter in general did not match the pattern of PCO activity values reported for
TCA and DCA (Nelson et al., 1989). There is no PCO activity data at this dose for TCE but
Elcombe et al. (1985) reported that the magnitude of TCE-induced increase in peroxisome
volume was similar to that of PCO activity at the only dose where both were tested (1000 mg/kg
TCE). However, Elcombe et al. (1985) reported increased peroxisomal volumes in B6C3F1 mice
after 10 days of TCE treatment were not dose-related (i.e., there was little difference between 500,
1000 and 1500 mg/kg TCE exposures in the magnitude of TCE-induced increases in peroxisomal
volume). The lack of dose-response for TCE-induced peroxisomal volume increases was not
consistent with increases in % liver/body weight that increased with increasing TCE exposure
concentration. Also as noted above, PCO activity appears to be highly variable in untreated and
treated rodents and to vary between experiments and between studies.
From the above comparison it is clears that TCE, DCA and TCA exposures were
associated with increased liver weight in mice but a question arises as to what changes account
for the liver weight increases. For TCE and TCA 500 mg/kg treatments, changes in glycogen
were not reported in the general descriptions of histopathological changes (Elcombe et al., 1985;
Styles et al., 1991; Dees and Travis, 1993) or were specifically described by the authors as being
similar to controls (Nelson et al., 1989). However for DCA, glycogen deposition was specifically
noted to be increased with treatment, although no quantitative analyses was presented that could
02/12/09
E-176
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
give information as to the nature of the dose-response (Nelson et al., 1989). Issues in regard to
not only whether TCE and its metabolites each gives a similar response for a number of
parameters, but what potential changes may be associated with carcinogenicity from long-term
exposures can be examined by a comparison of the dose-response curves for these parameters
from a range of exposure concentrations and durations of exposure. In addition, if glycogen
accumulation results from DC A exposure, what proportion of DCA-induced liver weight
increases result from such accumulation or other events that may be similar to those occurring
with TCE exposure (see Section 4.2.4 below)?
As noted above in Section 2.4.1., TCE-induced changes in liver weight appear to be
proportional to the exposure concentration across route of administration, gender and rodent
species. As an indication of the potential contribution of TCE metabolites to this effect, a
comparison of the shape of the dose-response curves for liver weight induction for TCE and its
metabolites is informative. A number of studies of TCA and DCA in drinking water, conducted
from 10-days to 4 weeks, have attempted to measure changes in liver weight induction,
peroxisomal enzyme activity, and changes in DNA synthesis predominantly in mice to provide
insight into the MOA(s) for liver cancer induction (Parrish et al., 1996; Sanchez and Bull 1990;
Carter et al 1995; DeAngelo et al.,1989, 2008).
Direct comparisons are harder to make between the gavage studies of TCE for 10 days of
exposure and those of DCA and TCA via drinking water. Similar to 10-day gavage exposures to
TCE, 14-day exposures to TCA or DCA via drinking water were reported to induce dose-related
increases in liver weight in male B6C3F1 mice (0.3, 1.0 , and 2.0 g/1 TCA or DCA) with a greater
increase in liver weight from DCA than TCA at 2 g/1 and a difference in the shape of the dose-
response curve (Sanchez and Bull, 1990). They reported a 1.08-fold, 1.31-fold, and 1.62-fold of
control liver weight for DCA and a 1.15-fold, 1.22-fold, and 1.38-fold of control values for TCA
at 0.3 g/1, 1.0 g/1 and 2.0 g/1 concentrations, respectively (n= 12-14 mice). While the magnitude
of difference between the exposures was ~ 6.7-fold between the lowest and highest dose, the
differences between TCA exposure groups for change in % of liver weight was ~ 2.5, but for
DCA the slope of the dose-response curve for liver weight increases appeared to be closer to the
magnitude of difference in exposure concentrations between the groups (i.e. a difference of 7.7-
fold between the highest and lowest dose for liver weight induction).
DeAngelo et al. (1989) reported that after 14 days of exposure to 5 g/1 or 2 g/1 TCA in
male mice, the magnitudes of the difference in the increase in dose (2.5-fold) was generally
higher than the increase % liver/body weight ratios at these doses (i.e., ~ 40% for the Swiss-
Webster, C3H, and for one of the B6C3F1 mouse experiments, and for the C57BL/6 mouse there
was no difference in liver weight induction between the 2 g/1 and 5g/1 TCA exposure groups).
There was a range in the magnitude of % liver/ body weight ratio increases between the strains of
mice with liver weight induction reported to range between 1.26-fold to 1.66-fold of control
values for the 4 strains of mice at 5 g/1 TCA and to range between 1.16-fold to 1.63-fold of
control values at 2 g/1 TCA. One strain, B6C3F1, was chosen to compare responses between
DCA and TCA. At 1 g/1, 2 g/1 and 5 g/1 TCA or DCA, DCA was reported to induce a greater
02/12/09
E-177
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
increase in liver weight that TCA (i.e. 1.55-fold vs. 1.39-fold of control % liver/body weight ratio
for 5.0 g/1 DC A vs. TCA, respectively). At the 5 g/1 exposures DCA induced ~ 40% greater %
liver/body weight than TCA. Although as noted above, the majority of the data from this study in
mice did not indicate that the magnitude of difference in exposure concentration was the same as
that of liver weight induction for TCA, in the particular experiment that examined both DCA and
TCA, the increase in % liver/body weight ratios were similar to the magnitude of difference in
dose between the 2 g/1 and 5 g/1 exposure concentrations for both DCA and TCA (i.e. 2- to 2.5-
fold increase in liver weight change corresponding to a 2.5-fold difference in exposure
concentration).
Carter et al. (1995) examined 0.5 g/1 and 5.0 g/1 exposures to DCA in B6C3F1 male mice
and reported that % liver/body weights were increased consistently from 0.5 g/1 DCA treatment
from 5 days to 30 days of treatment (i.e., a range of 1.05-fold to 1.16-fold of control). For 5.0 g/1
DCA exposure the range of increase in % liver/body weight was reported to be 1.37-fold to 2.04-
fold of control for the same time period. At the 15 days of exposures the % liver/body weight
ratios were 1.67-fold and 1.12-fold of control for 5.0 g/1 and 0.5 g/1 DCA and at 30 days were
1.99-fold and 1.16-fold, respectively. The difference in magnitude of dose and % liver/body
weight increase is difficult to determine given that the 5 g/1 dose of DCA reduced body weight
and significantly reduced water consumption by ~ 50%. The differences in DCA-induced %
liver/body weights were ~ 6-fold for the 15, 25, and 30-day data between the 0.5 and 5 g/1 DCA
exposures rather than the 10-fold difference in exposure concentration in the drinking water.
Parrish et al (1996) reported that for male B6C3F1 mice exposed to TCA or DCA (0, 0.01,
0.5 and 2.0 g/1) for 3 or 10 weeks, the 4- to 5-fold magnitude of difference in doses resulted in
increases in % liver/body weight for the 21-day and 71-day exposures that were greater for DCA
than TCA. The % liver/body weight ratio were 0.98-fold, 1.13-fold, 1.33-fold of control levels at
0.1 g/1, 0.5 g/1 and 2.0 g/1 TCA and for DCA were 1.02-fold, 1.24-fold, and 1.46-fold of control
levels, respectively, after 21 days of exposure. Both TCA and DCA exposures at 0.1 g/1 resulted
in difference in % liver/body weight change of 2% or less. For TCA, although there was a 4-fold
increase in magnitude between the 0.5 and 2.0 g/1 TCA exposure concentrations, the magnitude of
increase for % liver/body weight increase was 2.5-fold between them at both 21 and 71 days of
exposure. For DCA, the 4-fold difference in dose between the 0.5 and 2.0 g/1 DCA exposure
concentrations were reported to result in a ~ 2-fold increase in % liver/body weight increase at 21
days and ~ 4.5-fold increase at 71 days.
DeAngelo et al. (2008) studied 3 exposure concentrations of TCA in male B6C3F1 mice,
which were an order of magnitude apart, for 4 weeks of exposure. The % liver/body weight
ratios were 1.09-fold, 1.16-fold, and 1.35-fold of control levels, for 0.05, 0.5, and 5.0 g/1 TCA
exposures, respectively. The 10-fold differences in exposure concentration of TCA resulted in ~
2-fold differences in % liver/body weight increases. No dose-response inferences can be drawn
from the 4-week study of DCA and TCA in B6C3F1 male mice by Kato-Weinstein et al. (2001)
but 2 g/1 DCA and 3 g/1 TCA in drinking water were reported to induce % liver/body weights of
1.42-fold and 1.33-fold of control, respectively (n=5).
02/12/09
E-178
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The majority of short-term studies of DC A and TCA in mice have been conducted in the
B6C3F1 strain and in males. Studies conducted from 14 to 30 days show a consistent increase in
% liver/body weight induction by TCA or DC A. Accordingly an examination of all of the data
from Parrish et al. (1996), Sanchez and Bull (1990), Carter et al. (1995), Kato-Weinstein et al.
(2001), and DeAngelo et al. (1989; 2008) from 14 to 30 days of exposure in male B6C3F1 mice
can give an approximation of the dose-response differences between DCA and TCA for liver
weight induction as shown in the table below and figure. Although the data for B6C3F1 mice
from Sanchez and Bull (1990) is reported as the fold of liver weight rather that % liver/body
weight increase, it is included in the comparison as both reflect increase in liver weight. Similar
data can be assessed for TCE for comparative purposes. Short duration studies (10-42 days) were
selected because (i) in chronic studies, liver weight increases are confounded by tumor burden,
(ii) multiple studies are available, and (iii) in this duration range, Kjellstrand et al. (1981) reported
that TCE-induced increases in liver weight plateau, and (iv) TCA studies do not show significant
duration-dependent differences in this duration range. These comparisons are presented in the
table below.
DeAngelo et al (1989) and Carter et al (1995) used up to 5 g/1 DCA and TCA in their
experiments with Carter et al (1995) noting a dramatic decrease in water consumption in the 5 g/1
DCA treatment groups (46%- 64% reduction) which can affect body weight as well as dose
received. DeAngelo et al (1989) did not report drinking water consumption. The drinking water
consumption was reported by DeAngelo et al. (2008) to be reduced by 11%, 17%, and 30% in the
0.05g/l, 0.5 g/1 and 5 g/1 TCA treated groups compared to 2 g/1 NaCl control animals over 60
weeks. DeAngelo et al. (1999) reported mean drinking water consumption to be reduced by 26%
in mice exposed to 3.5 g/1 DCA over 100 weeks. Carter et al. (1995) reported that DCA at 5 g/1 to
decrease drinking water consumption by 64% and 46% but 0.5 g/1 DCA to not affect drinking
water consumption. Thus it appears that the 5 g/1 concentrations of either DCA or TCA can
significantly affect drinking water consumption as well as inducing reductions in body weight.
Accordingly, an estimation of the shape of the dose-response curve for comparative purposes
between DCA or TCA drinking water studies is best examined at concentrations at 2g/l or less,
especially for DCA.
The dose-response curves for similar concentrations of DCA and TCA are presented in
Figure 4.4.1 for durations of exposure from 14-28 days in the male B6C3F1 mouse, which was
the most common sex and strain used. For this comparative analysis an average is provided
between two values for a given concentration and duration of exposure for comparison with other
doses and time points. As noted in the discussion of individual experiments, there appears to be a
linear correlation between dose in drinking water and liver weight induction up to 2 g/1 of DCA.
However, the shape of the dose-response curve for TCA appears to be quite different (i.e., lower
concentrations of TCA inducing larger increase that does DCA but then the response reaching an
apparent plateau for TCA at higher doses while that of DCA continues to increase). As shown by
DeAngelo et al (2008), 10-fold differences in the magnitude of exposure concentration to TCA
corresponded to ~ 2-fold differences in liver weight induction increases. In addition, TCA studies
02/12/09
E-179
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
did not show significant duration-dependent difference in liver weight induction in this duration
range as shown in the table below.
Of interest is the issue of how the dose-response curves for TCA and DCA compare to
that of TCE in a similar model and dose range. Since TCA and DCA have strikingly different
dose-response curves, which one if either best fits that of TCE and thus can give insight as to
which is causative agent for TCE's effects in the liver? In the case of the TCE database in the
mouse two strains have been predominantly studied, Swiss and B6C3F1, and both have been
reported to get liver tumors in response to chronic TCE exposure. Rather than administered in
drinking water, oral TCE studies have been conducted via oral gavage and generally in corn oil
for 5 days of exposure per week. The study by Goel et al. (1992) was conducted in ground-nut
oil. Vehicle effects, the difference between daily and weekly exposures, the dependence of TCE
effects in the liver on its metabolism to a variety of agents capable inducing effects in the liver,
differences in response between strains, and the inherent increased variability in use of the male
mouse model all add to increased difficulty in establishing the dose-response relationship for TCE
across studies and for comparisons to the DCA and TCA database. Despite difference in
exposure route, etc, a consistent pattern of dose-response emerges from combining the available
TCE data. The effects of oral exposure to TCE from 10-42 days on liver weight induction is
shown in Figure 4.4.2 using the data of Elcombe et al (1985), Dees and Travis (1993), Goel et al.
(1992), Merrick et al. (1987), Goldsworthy and Popp (1987), and Buben and O'Flaherty (1985).
More detailed discussion of the 4- to 6-week studies is presented in Section 2.4.3 below (e.g., for
Merrick et al., 1989; Goel et al., 1992; Buben and O'Flaherty, 1985). For this comparative
analysis an average is provided between two values per concentration and duration of exposure
for comparison with other doses and time points. As shown by the 10-day data in B6C3 F1 mice,
there are significant differences in response between studies of male B6C3F1 mice at the same
dose of TCE. This variability is similar to findings from inhalation studies of TCE in male mice
(Kjellstrand etal., 1983a).
As shown in Figure 4.4.2, oral TCE administration in male B6C3F1 and Swiss mice
appeared to induce a dose-related increase in % liver/body weight that was generally proportional
to the increase in magnitude of dose, though as expected, with more variability than observed for
a similar exercise for DCA or TCA in drinking water. Common exposure concentrations
between B6C3F1 and Swiss mice were 100, 500, 1000, 1500 and 2400 mg/kg/day TCE which
corresponded to a 5-fold, 2-fold, 1.5-fold and 1.6-fold difference in the magnitude of dose. For
the data from studies in B6C3 F1 mice, there was no increase reported at 100 mg/k/day TCE but
between 500 and 1000, 1000 and 1500, and 1500 and 2400 mg/kg/day TCE the magnitude of
difference in doses matched that of the magnitude of increase in % liver/body weight (i.e., a 2.6-
fold, 1.4-fold, and 1.7-fold increase in liver weight was matched by a 2-fold, 1.5-fold, and 1.6-
fold increase in TCE exposure concentration at these exposure intervals). However, only 10-day
was available for doses between 100 and 500 mg/kg in B6C3F1 mice and at the lower doses, a
10-day interval may have been too short for the increase in liver weight to have been fully
expressed. The database for the Swiss mice, which has more data from 28 and 42 days of
exposure, support this conclusion. At 28-42 days of exposure there was a much greater increase
02/12/09
E-180
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
in liver weight from TCE exposure in Swiss mice than the 10-day data in B6C3F1 mice. In
Figure 4.4.2, the 10-day data are included for comparative purpose for the B6C3F1 data set and
the Swiss and B6C3F1 datasets combined. Both the combined TCE data and that for only
B6C3F1 mice shows a correlation with the magnitude of dose and magnitude of % liver/body
weight increase. The slope of the dose response curves are both closer to that of DCA than TCA.
The correlation coefficients for the linear regressions presented for the B6C3F1 data is R2 =0.861
and for the combined data sets is R2 = 0.712. Comparisons of the slopes of the dose-response
curves indicate that TCA is not responsible for TCE-induced liver effects. In this regression all
data points were treated equally although some came from several sets of data and others did not.
Of note is that the 2000 mg/kg TCE data point in the combined data set, which is much lower in
liver weight response than the other data, is from one experiment (Goel et al. 1992), from 6 mice,
at one time point (28 days), and one strain (Swiss). Deletion of this data point from the rest of the
23 used in the study results in a better fit to the data of the regression analysis.
A more direct comparison would be on the basis of dose rather than drinking water
concentration. The estimations of internal dose of DCA or TCA from drinking water studies have
been reported to vary with DeAngelo et al. reporting DCA drinking water concentrations of 1.0,
2.0, and 5.0 g/1 to result in 90, 166, and 346 mg/kg/day, respectively. For TCA, 0.05, 0.5, 1.0,
2.0, and 5 g/1 drinking water exposures were reported to result in 5.8 (range 3.6-8.0), 50 (range of
32.5 to 68), 131, 261, and 469 ( range 364 to 602) mg/kg/day doses. The estimations of internal
dose of DCA or TCA from drinking water studies, while varying considerably (DeAngelo et al.
1989, 2008), nonetheless suggest that the doses of TCE used in the gavage experiments were
much higher than those of DCA or TCA. However, only a fraction of ingested TCE is
metabolized to DCA or TCA, as, in addition to oxidative metabolism, TCE is also cleared by
GSH conjugation and by exhalation.
While DCA dosimetry is highly uncertain (Sections 3.3 and 3.5), the mouse PBPK model,
described in Section 3.5 was calibrated using extensive in vivo data on TCA blood, plasma, liver,
and urinary excretion data from inhalation and gavage TCE exposures, and makes robust
predictions of the rate of TCA production. If TCA were predominantly responsible for TCE-
induced liver weight increases, then replacing administered TCE dose (e.g., mg TCE/kg/day) by
the rate of TCA produced from TCE (mg TCA/kg/day) should lead to dose-response curves for
increased liver weight consistent with those from directly administered TCA. Figure 4.4.3 shows
this comparison using the PBPK model-based estimates of TCA production for 4 TCE studies
from 28-42 days in the male NMRI, Swiss, and B6C3F1 mice (Kjellstrand et al. 1983b, Buben
and O'Flaherty 1985, Merrick et al. 1989, Goel et al. 1992) and 4 oral TCA studies in B6C3F1
male mice at 2 g/1 or lower drinking water exposure (DeAngelo et al. 1989, 2008, Parrish et al.
1996, Kato-Weinstein et al. 2001) from 14-28 days of exposure. The selection of the 28-42 day
data for TCE was intended to address the decreased opportunity for full expression of response at
10 days. PBPK modeling predictions of daily internal doses of TCA in terms of mg/kg/day via
produced via TCE metabolism would be are indeed lower than the TCE concentrations in terms of
mg/kg/day given orally by gavage. The predicted internal dose of TCA from TCE exposure
studies are of a comparable range to those predicted from TCA drinking water studies at exposure
02/12/09
E-181
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
concentrations in which palability has not been an issue for estimation of internal dose. Thus
although the TCE data are for higher exposure concentrations, they are predicted to produce
comparable levels of TCA internal dose estimated from direct TCA administration in drinking
water.
Figure 4.4.3 clearly shows that for a given amount of TCA produced from TCE, but going
through intermediate metabolic pathways, the liver weight increases are substantially greater than,
and highly inconsistent with, that expected based on direct TCA administration. In particular, the
response from direct TCA administration appears to "saturate" with increasing TCA dose at a
level of about 1.4-fold, while the response from TCE administration continues to increase with
dose to 1.75-fold at the highest dose administered orally in Buben and O'Flaherty 1985) and over
2-fold in the inhalation study of Kjellstrand et al. (1983b). For this analysis is unlikely that strain
differences can account for this inconsistency in the dose-response curves. TCE-induced
increases in liver weight appear to be generally similar between B6C3F1 and Swiss male mice
(see table below) via oral exposure and between NMRI male and female mice after inhalation,
although the NMRI strain appeared to be more prone to TCE-induced toxicity in male mice and
for females to have a smaller TCE-induced liver weight increase than other strains (Kjellstrand et
al. 1983b). As noted previously, the difference in response between strains and between studies
in the same strain for TCE liver weight increases can be highly variable. Little data exists to
examine this issue for TCA studies although DeAngelo et al (1989) report a range of 1.16-fold to
1.63-fold of control % liver/body weight increase after 14 days exposure at 2 g/1 TCA in the
Swiss-Webster, C3H, C57BL/6, and B6C3F1 strains, with differences also noted between 2
studies of the B6C3F1 mouse.
Furthermore, while as noted previously, oral studies appear to report a linear relationship
between TCE exposure concentration and liver weight induction, the inclusion of inhalation
studies on the basis of internal dose led to a highly consistent dose-response curve for among
TCE study. Therefore, it is unlikely that differing routes of exposure can explain the
inconsistencies in dose-response. The PBPK model predicted that matching average TCA
production by TCE with the equivalent average dose from drinking water-administered TCA also
led to an equivalent AUC of TCA in the liver. Moreover, Dees and Travis (1993) administered
100 to 1000 mg/kg-d TCA by gavage to male and female B6C3F1 mice for 11 days, and did not
observe increases in liver/body weight ratios more than 1.28-fold, no higher than those observed
with drinking water exposures. Finally, the dose-response consistency between TCE inhalation
and gavage studies argues against route of exposure significantly impacting liver weight
increases. Thus, no level of TCA administration appears able account for the continuing increase
in liver weights observed with TCE, quantitatively inconsistent with TCA being the predominant
metabolite responsible for TCE-induced liver weight changes. Thus, involvement of other
metabolites, besides TCA, is implicated as the causes of TCE-induced liver effects.
Additional analyses do, however, support a role for oxidative metabolism in TCE-induced
liver weight increases, and that the parent compound TCE is not the likely active moeity
(suggested previously by Buben and O'Flaherty 1985). In particular, the same studies are shown
02/12/09
E-182
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
in Figure 4.4.4 using PBPK-model based predictions of the area-under-the-curve (AUC) of TCE
in blood and total oxidative metabolism, which produces chloral, TCOH, DCA, and other
metabolites in addition to TCA. The dose-response relationship between TCE blood levels and
liver weight increase, while still having a significant trend, shows substantial scatter and a low R2
of 0.43. On the other hand, using total oxidative metabolism as the dose metric leads to
substantially more consistency dose-response across studies, and a much tighter linear trend with
an R2 of 0.90 (Figure 4.4.4). A similar consistency is observed using liver-only oxidative
metabolism as the dose metric, with R2 of 0.86 (not shown). Thus while the slope is similar
between liver weight increase and TCE concentration in the blood and liver weight increase and
rate of total oxidative metabolism, the data are a much better fit for total oxidative metabolism.
As stated in many of the discussions of individual studies, there is a limited ability to
detect a statistically significant change in liver weight change in experiments that use a relatively
small number of animals. Many experiments have been conducted with 4-6 mice per dose group.
The experiments of Buben and O'Flaherty used 12-14 mice per group giving it a greater ability to
detect a TCE-induced dose response. In some experiments greater care was taken to document
and age and weight match the control and treatment groups before the start of treatment. The
approach taken above for the analyses of TCE, TCA and DCA uses data across several data sets
and gives a more robust description of these dose-response curves, especially at lower exposure
levels. For example, the data from DeAngelo et al. (2008) for TCA-induced % liver/body weight
ratio increases in male B6C3F1 mice were only derived from 5 animals per treatment group after
4 weeks of exposure. The 0.05 g/1 and 0.5 g/1 exposure concentrations were reported to give a
1.09-fold and 1.16-fold of control % liver/body weight ratios which were consistent with the
increases noted in the cross-study database above. However, a power calculation shows that the
type II error, which should be > 50% and thus greater than the chances of "flipping a coin", was
only a 6% and 7% and therefore the designed experiment could accept a false null hypothesis.
Although the qualitative similarity to the linear dose-response relationship between DCA
and liver weight increases is suggestive of DCA being the predominant metabolite responsible for
TCE liver weight increases, due to the highly uncertain dosimetry of DCA derived from TCE, this
hypothesis cannot be tested on the basis of internal dose. Similarly, another TCE metabolite,
chloral hydrate, has also been reported to induce liver tumors in mice, however there are no
adequate comparative data to assess the nature of liver weight increases induced by this TCE
metabolite (see Section 2.5. below). Whether its formation in the liver after TCE exposure
correlates with TCE-induced liver weight changes cannot be determined. Of note is the high
variability in total oxidative metabolism reported in mice and humans of Chapter 3 of the TCE
risk assessment which suggests that the correlation of total TCE oxidative metabolism with TCE-
induced liver effects should lead not only to a high degree of variability in response in rodent
bioassays which is the case (see Section 2.4.4 below) but also make detection of liver effects
more difficult in human epidemiological studies (see Section 4.3.2 of Chapter 3 of the TCE risk
assessment). What mechanisms or events are leading to liver weight increases for DCA, TCA
and TCE can be examined by correlations between changes in glycogen content, hepatocyte
volume, and evidence of polyploidization noted in short-term assays.
02/12/09
E-183
-------
1
2
3
4
5
6
7
8
9
10
11
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Data have been reported regarding the nature of changes the TCE and its metabolites
induce in the liver and are responsible for the reported increases in liver weight. Increased liver
weight may result from increased size or hypertrophy of hepatocytes through changes in glycogen
deposition, but also through increased polyploidization. Increased cell number may also
contribute to increased liver weight. As noted above in Section 2.4.1., hepatocellular hypertrophy
appeared to be related to TCE-induced liver weight changes after short-term exposures.
However, neither glycogen deposition, DNA synthesis, or increases in mitosis appear to be
correlated with liver weight increases. In particular DNA synthesis increases were similar from
250-1000 mg/kg and peroxisomal volume was similar between 500 mg/kg and 1500 mg/kg TCE
exposures after 10 days. Autoradiographs identified hepatocytes undergoing DNA synthesis in
"mature" hepatocytes that were in areas where polyploidization typically takes place in the liver.
02/12/09
E-184
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Liver Weight induction as % Liver/body Weight Fold-of-control in Male B6C3F1 Mice from DCA or TCA
Drinking Water Studies:
Concentration
(g/1)
DCA 0.1
0.3
0.5
1.0
2.0
5.0
TCA 0.05
0.1
0.3
0.5
1.0
2.0
3.0
5.0
14 or 15 days
1.08-fold
1.12-fold
1.31-fold
1.62-fold
1.67-fold
1.15-fold
Duration of Exposure
20 or 21 days 25 days
1.02-fold
1.24-fold, 1.05-fold 1.16-fold
1.46-fold, 2.01-fold 2.04-fold
0.98-fold
1.23-fold, 1.08-fold
1.38-fold, 1.16-fold, 1.26-fold
1.39-fold, 1.35-fold
1.13-fold
1.33-fold
Mean for average of days 14-30
28 or 30 days
1.02-fold
1.08-fold
1.16-fold 1.15-fold
1.31-fold
1.99-fold, 1.42-fold 1.83-fold
1.67-fold
1.09-fold
1.16-fold
1.3 3-fold
1.09-fold
0.98-fold
1.15-fold
1.15-fold
1.16-fold
1.30-fold
1.3 3-fold
1.37-fold
02/12/09
E-185
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Liver Weight induction as % Liver/body Weight Fold-of-control in Male B6C3F1 or Swiss Mice from TCE
Gavage Studies:
10 days
28 days
Concentration
(mg/kg/day)
B6C3F1
42 days
Mean for
average of days 10-42
TCE 100 1.00-fold
250 1.00-fold
500 1.20-fold, 1.06-fold
600
1000 1.50-fold, 1.17-fold, 1.50-fold
1200
1500 1.47-fold
2400
1.36-fold
1.64-fold
1.81-fold
1.00-fold
1.00-fold
1.13-fold
1.36-fold
1.39-fold
1.64-fold
1.47-fold
1.81-fold
Swiss
TCE
100
200
400
500
800
1000
1500
1600
2000
1.43-fold
1.56-fold
1.75-fold
1.32-fold
1.41-fold
1.38-fold
1.12-fold
1.15-fold
1.25-fold
1. 36-fold
1.63-fold
1.12-fold
1.15-fold
1.25-fold
1.38-fold
1.36-fold
1.49-fold
1.75-fold
1.63-fold
1.38-fold
02/12/09
E-186
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2400 1.69-fold 1.69-fold
B6C3F1 and Swiss (data sets combined)
Mean for average of days 10-42
TCE 100 1.06-fold
200 1.15-fold
250 1.00-fold
400 1.25-fold
500 1.26-fold
600 1.36-fold
800 1.36-fold
1000 1.49-fold
1200 1.64-fold
1500 1.61-fold
1600 1.63-fold
2000 1.38-fold
2400 1.75-fold
02/12/09
E-187
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Male B6C3F1 mice liver weight for TCA and DCA in drinking water - days 14-30
Concentration of DCA or TCA (g/l)
1 Figure 4.4.1 (reproduced from Section 4.4). Comparison of average fold-changes in relative
2 liver weight to control and exposure concentrations of 2 g/l or less in drinking water for TCA
3 and DCA in male B6C3F1 mice for 14-30 days (Parrish et al.,1996; Sanchez and Bull, 1990;
4 Carter et al., 1995; Kato-Weinstein et al., 2001; DeAngelo et al., 1989; 2008).
02/12/09
E-188
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Male mice liver weight for TCE oral gavage - days 10-42
Concentration of TCE (mg/kg/day)
Male mice liver weight for TCE oral gavage - days 10-42
Concentration of TCE (mg/kg/day)
1 Figure 4.4.2 (Reproduced from Section 4.4). Comparisons of fold-changes in average relative
2 liver weight and gavage dose of (top panel) male B6C3F1 mice for 10-28 days of exposure
3 (Merrick et al., 1989; Elcombe et al., 1985; Goldsworthy and Popp 1987, Dees and Travis, 1993)
4 and (bottom panel) in male B6C3F1 and Swiss mice
02/12/09
E-189
-------
1
2
3
4
5
6
7
8
9
10
11
12
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
2.5
O)
CO
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Daily AUC TCE in Blood (mg-hr/l) Daily TCE Oxidized (mg/kg-d)
1 Figure 4.4.4 (reproduced from Section 4.4). Fold-changes in relative liver weight for data sets in
2 male B6C3F1, Swiss, and NRMI mice reported by TCE studies of duration 28-42 days
3 (Kjellstrand et al. 1983b, Buben and O'Flaherty 1985, Merrick et al. 1989, Goel et al. 1992)
4 using internal dose metrics predicted by the PBPK model described in section 3.5: (A) dose
5 metric is the median estimate of the daily AUC of TCE in blood, (B) dose metric is the median
6 estimate of the total daily rate of TCE oxidation. Lines show linear regression. Use of liver
7 oxidative metabolism as a dose metric gives results qualitatively similar to (B), with R2=0.86.
02/12/09
E-191
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
By 14 days of exposure, Sanchez and Bull (1990) reported that both dose-related TCA-
and DCA-induced increases in liver weight were generally consistent with changing cell size
increases, but were not correlated with patterns of change in hepatic DNA content, incorporation
of tritiated thymidine in DNA extracts from whole liver, or incorporation of tritiated thymidine in
hepatocytes. There are conflicting reports of DNA synthesis induction in individual hepatocytes
for up to 14 days of DCA or TCA exposure and a lack of correlation with patterns observed for
this endpoint and those of whole liver thymidine incorporation. The inconsistency of whole liver
DNA tritiated thymidine incorporation with that reported for hepatocytes was noted by the
Sanchez and Bull (1990) to be unexplained. Carter et al. (1995) also report a lack of correlation
between hepatic DNA tritiated thymidine incorporation and labeling in individual hepatocytes in
male mice. Carter et al (1995) reported no increase in labeling of hepatocytes in comparison to
controls for any DCA treatment group from 5 to 30 days of DCA exposure. Rather than increase
hepatocyte labeling, DCA induced a decrease with no change reported from days 5 though 15 but
significantly decreased levels between days 20 and 30 for 0.5 g/1 that were similar to those
observed for the 5 g/1 exposures.
The most comparable time period between TCE, TCA and DCA results for whole liver
thymidine incorporation is the 10- and 14-day durations of exposure when peak tritiated
thymidine incorporation into individual hepatocytes and whole liver for TCA and DCA have
been reported to have already passed (Styles et al., 1991; Sanchez and Bull, 1990; Pereira, 1996;
Carter et al., 1995). Whole liver DNA synthesis was elevated over control levels by ~ 2 fold
after from 250 to 1000 mg/kg TCE exposure after 10 days of exposure but did not correlate with
mitosis (Elcombe et al., 1985; Dees and Travis,1993). After 3 weeks of exposure to TCE,
Laughter et al. (2004) reported in individual hepatocytes that 1% and 4.5% of hepatocytes had
undergone DNA synthesis in the last week of treatment for the 500 and 1000 mg/kg TCE levels,
respectively. More importantly, these data show that hepatocyte proliferation in TCE-exposed
mice at 10 days of exposure or for DCA- or TCA-exposed mice for up to 14 days of exposure is
confined to a very small population of cells in the liver.
In regard to cell size, although increased glycogen deposition with DCA exposure was
noted by Sanchez and Bull (1990), lack of quantitative analyses of that accumulation in this study
precludes comparison with DCA-induced liver weight gain. Although not presenting a
quantitative analysis, Sanchez and Bull (1990) reported DCA-treated B6C3F1 mice to have large
amounts of PAS staining material and Swiss-Webster mice to have similar increase despite
reporting differences of DCA-induced liver weight gain between the two strains. The lack of
concordance of the DCA-induced magnitude of increase in liver weight with that of glycogen
deposition is consistent with the findings for longer-term exposures to DCA reported by Kato-
Weinstein et al. (2001) and Pereira et al (2004) in mice (see Section 2.4.4. below). Carter et al.
(1995) reported that in control mice there was a large variation in apparent glycogen content and
also did not perform a quantitative analysis of glycogen deposition. The variability of this
parameter in untreated animals and the extraction of glycogen during normal tissue processing
for light microscopy makes quantitative analyses for dose-response difficult unless specific
02/12/09
E-192
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
methodologies are employed to quantitatively assess liver glycogen levels as was done by Kato-
Weinstein et al (2001) and Pereira et al. (2004).
Although suggested by their data, polyploidization was not examined for DCA or TCA
exposure in the study of Sanchez and Bull (1990). Carter et al (1995) reported that hepatocytes
from both 0.5 and 5 g/1 DCA treatment groups were reported to have enlarged, presumably
polyploidy nuclei with some hepatocyte nuclei labeled in the mid-zonal area. There were
statistically significant changes in cellularity, nuclear size, and multinucleated cells during 30
days exposure to DCA. The percentage of mononucleated cells hepatocytes was reported to be
similar between control and DCA treatment groups at 5- and 10-day exposure. However, at 15
days and beyond, DCA treatments were reported to induce increases in mononucleated
hepatocytes. At later time periods there were also reports of DCA-induced increases nuclear
area, consistent with increased polyploidization without mitosis. The consistent reporting of an
increasing number of mononucleated cells between 15 and 30 days could be associated with
clearance of mature hepatocytes as suggested by the report of DCA-induced loss of cell nuclei.
The reported decrease in the numbers of binucleate cells in favor of mononucleate cells is not
typical of any stage of normal liver growth (Brodsky and Uryvaeva, 1977). The linear dose-
response in DCA-induced liver weight increase was not consistent with the increased numbers of
mononucleate cells and increase nuclear area reported from day 20 onward by Carter et al (1995).
Specifically, the large differences in liver weight induction between the 0.5 g/1 treatment group
and the 5 g/1 treatment groups at all times studied also did not correlate with changes in nuclear
size and % of mononucleate cells. Thus, DCA-induced increases in liver weight were not a
function of cellular proliferation, but probably included hypertrophy associated with
polyploidization, increased glycogen deposition and other factors.
In regard to necrosis, Elcombe et al. (1995) reported only small incidence of focal
necrosis in 1500 mg/kg TCE-exposed mice and no necrosis at exposures up to 1000 mg/kg for
10 days as did Dees and Travis (1993). Sanchez and Bull (1990) report DCA-induced localized
areas of coagulative necrosis both for B6C3F1 and Swiss-Webster mice at higher exposure
levels (1 or 2 g/1) by 14 days but not at the 0.3 g/1 level or earlier time points. For TCA
treatment, necrosis was reported to not be associated with TCA treatment for up to 2 g/1 and up
to 14 days of exposure. Carter et al (1995) reported that mice given 0.5 g/1 DCA for 15, 20, and
25 days had midzonal focal cells with less detectable or no cell membranes, loss of the coarse
granularity of the cytoplasm, with some cells having apparent karyolysis, but for liver
architecture to be normal.
As for apoptosis, Both Elcombe et al. (1985) and Dees and Travis (1993) reported no
changes in apoptosis other than increased apoptosis only at a treatment level of 1000 mg/kg
TCE. Rather than increases in apoptosis, peroxisome proliferators have been suggested to
inhibit apoptosis as part of their carcinogenic MOA (see Section 3.4.1). However, the age and
species studied appear to greatly affect background rates of apoptosis. Snyder et al. (1995)
report that control mice were reported to exhibit apoptotic frequencies ranging from ~ 0.04 to
0.085%, that over the 30-day period of their study the frequency rate of apoptosis declined, and
02/12/09
E-193
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
suggest that this pattern is consistent with reports of the livers of young animals undergoing
rapid changes in cell death and proliferation. They reported rat liver to have a greater the
estimated frequency of spontaneous apoptosis (~ 0.1%) and therefore greater than that of the
mouse. Carter et al (1995) reported that after 25 days of 0.5 g/1 DCA treatment apoptotic bodies
were reported as well as fewer nuclei in the pericentral zone and larger nuclei in central and
midzonal areas. This would indicate an increase in the apoptosis associated potential increases
in polyploidization and cell maturation. However, Snyder et al (1995) report that mice treated
with 0.5 g/1 DCA over a 30-day period had a similar trend as control mice of decreasing
apoptosis with age. The percentage of apoptotic hepatocytes decreased in DCA-treated mice at
the earliest time point studied and remained statistically significantly decreased from controls
from 5 to 30 days of exposure. Although the rate of apoptosis was very low in controls,
treatment with 0.5g/l DCA reduced it further (-30-40% reduction) during the 30-day study
period. The results of this study not only provide a baseline of apoptosis in the mouse liver,
which is very low, but also to show the importance of taking into account the effects of age on
such determinations. The significance of the DCA-induced reduction in apoptosis reported in
this study, from a level that is already inherently low in the mouse, to account for the MOA for
induction of DCA-induced liver cancer is difficult to discern.
2.4.3. Summary TCE Subchronic and Chronic Studies
The results of longer-term (Channel et al., 1998; Toraason et al., 1999; Parrish et al.,
1996) studies of "oxidative stress" for TCE and its metabolites are discussed in Section 3.4.2.3.
Of note are the findings that the extent of increased enzyme activities associated with
peroxisome proliferation do not appear to correlate with measures of oxidative stress after
longer term exposures (Parrish et al. 1996) and single strand breaks (Chang et al., 1992).
Similar to the reports of Melnick et al (1987) in rats, Merrick et al. (1989) report that
vehicle (aqueous or gavage) affects TCE- induced toxicity in mice. Vehicle type made a large
difference in mortality, extent of liver necrosis, and liver weight gain in male and female
B6C3F1 mice after 4 weeks of exposure. The lowest dose used in this experiment was 600
mg/kg/day in males and 450 mg/kg/day in females. Administration of TCE via gavage using
Emulphor resulted in mortality of all of the male mice and most of the female mice at a dose in
corn oil that resulted in few deaths. However, use of Emulphor vehicle induced little if any
focal necrosis in males at concentrations of TCE in corn oil gavage that caused significant focal
necrosis, indicating vehicle effects.
As discussed above in Section 2.4.2, the extent of TCE-induced liver weight increases
was consistent between 4 and 6 weeks of exposure and between 10-day and 4 week exposure at
higher dose levels. In general, the reported elevations of enzymatic markers of liver toxicity and
results for focal hepatocellular necrosis were not consistent and did not reflect TCE dose-
responses observed for induction of liver weight increases (Merrick et al., 1989). Female mice
given corn oil and male and female mice given TCE in Emulphor were reported to have "no to
negligible necrosis" although they had increased liver weight from TCE exposure. Using a
02/12/09
E-194
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
different type of oil vehicle, Goel et al. (1992) exposed male Swiss mice to TCE in groundnut
oil at concentrations ranging from 500 to 2000 mg/kg for 4 weeks and reported no changes in
body weight up to 2000 mg/kg, although there was a 15% decrease at the highest dose, but
increases TCE-induced increase in % liver/body weight ratio. At a dose of 1000 and 2000
mg/kg, liver swelling, vacuolization, and widespread degenerative necrosis of hepatocytes was
reported along with marked proliferation of "endothelial cells" but no quantitation regarding the
extent or location of hepatocellular necrosis was reported, nor whether there was a dose-
response relationship in these events. They reported a TCE-related dose-response in catalase,
liver protein but decreased induction at the 2000 mg/kg level where body weight had decreased.
Three studies were published by Kjellstrand et al. that examined effects of TCE
inhalation primarily in mice using whole body inhalation chambers (Kjellstrand et al. 1981,
1983a, 1983b). Liver weight changes were used as the indication of TCE-induced effects. The
quantitative results from these experiments had many limitations due to their experimental
design including failure to determine body weight changes for individual animals and inability
to determine the exact magnitude of TCE due to concurrent oral TCE ingestion from food and
grooming behavior. An advantage of this route of exposure is that there were not confounding
vehicle effects. The results from Kjellstrand et al. (1981) are particularly limited by
experimental design errors but reported similar increases in liver weight gain in gerbils and rats
exposed at 150 ppm TCE. For rats, Kjellstrand et al (1981) do report increases in liver/body
weight ratios of 1.26-fold and 1.21-fold of control in male and female rat 30 days of continuous
TCE inhalation exposure. The unpublished report of Woolhiser et al (2006) reports 1.05-fold,
1.07-fold, and 1.13-fold of control % liver/body weight changes in 100 ppm, 300 ppm and 1000
ppm-exposure groups that are exposed for 6 hours/day, 5 days/week for 4 weeks in groups of 8
female CD rats. At the two highest exposure levels, body weight was reduced by TCE
exposure. If the 150 ppm continuous exposure concentrations of Kjellstrand are analogous to
750 ppm exposures using the paradigm of Woolhiser et al (2006). Therefore, the very limited
inhalation database for rats does indicate TCE-related increases in liver weight.
The study of Kjellstrand et al. (1983 a) employed a more successful experimental design
that recorded liver weight changes in carefully matched control and treatment groups to
determine TCE-treatment related effects on liver weight in 7 strains of mice after 30 days of
continuous inhalation exposure at 150 ppm TCE. Individual animal body weight changes were
not recorded so that such an approach cannot take into account the effects of body weight
changes and determine a relative % liver/body weight ratio. The data presented in this report
was for absolute liver weight changes between treated and non-treated groups with carefully
matched average body weights at the initiation of exposure. A strength of the experimental
design is its presentation of results between duplicate experiments and thus to show the
differences in results between similar exposed groups that were conducted at different times.
This information gives a measure of variability in response with time. Mouse strain groups, that
did not experience TCE-induced decreased body weight gain in comparison to untreated groups
(i.e., DBA and "wild" type mice), represented the most accurate determination of TCE-induced
liver weight changes given that systemic toxicity that affects body weight can also affect liver
02/12/09 E-195
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
weight. The C57BL, B6CBA, and NZB groups all had at least one group out of two of male
mice with changes in final body weight due to TCE exposure. Only one group of NMRI mice
were reported in this study and that group had TCE-induced decreases in final body weight.
The A/sn group not only had both male groups with decreased final body weight after TCE
exposure (along with differences between exposed and control groups at the initiation of
exposure) but also a decrease in body weight in one of the female groups and thus appears to be
the strain with the greatest susceptibility to TCE-induced systemic toxicity. In strains of male
mice in which there was no TCE-induced affects on final body weight ("wild type" and DBA),
the influence of gender on liver weight induction and variability of the response could be more
readily assessed. In "wild type" mice there was a 1.76-fold and 1.80-fold of control liver
weight in groups 1 and 2 for female mice, and for males a 1.84-fold and 1.62-fold of control
liver weight for groups 1 and 2, respectively. For DBA mice there was a 1.87-fold and 1.88-
fold of control liver weight in groups 1 and 2 for female mice, and for males a 1.45-fold and
2.00-fold of control liver weight for groups 1 and 2, respectively. Of note, as described
previously, the size of the liver is under strict control in relation to body size. An essential
doubling of the size of the liver is a profound effect with the magnitude of liver weight size
increase physiologically limited..
Overall, the consistency between groups of female mice of the same strain for TCE-
induced liver weight gain, regardless of strain examined, was striking as was the lack of body
weight changes at TCE exposure levels that induced body weight changes in male mice. In the
absence of body weight changes, the difference in TCE-response in female mice appeared to be
reflective of strain and initial weight differences. Groups of female mice with higher body
weights, regardless of strain, generally had higher increases in TCE-induced liver weight
increases. For the C57BL and As/n strains, female mice starting weights were averaged 17.5 g
and 15.5 g, while the average liver weights were 1.63-fold and 1.64-fold of control after TCE
exposure, respectively. For the B6CBA, "wild" type, DBA, and NZB female groups the
starting body weights averaged 22.5 g, 21.0 g, 23.0 g, and 21.0 g, while the average liver
weights were 1.70-fold, 1.78-fold, 1.88-fold, and 2.09-fold of control after TCE exposure,
respectively. The NMRI group of female mice, did not follow this general pattern and had the
highest initial body weight for the single group of 10 mice reported (i.e., 27 g) associated with
1.66-fold of control liver weight.
The results of Kjellstrand et al. (1983a) suggested that there was more variability
between male mice than female mice in relation to TCE-induced liver weight gain. More strains
exhibited TCE-induced body weight changes in male mice than female mice suggesting
increased susceptibility of male mice to TCE toxicity as well as more variability in response.
Initial body weight also appeared to be a factor in the magnitude of TCE-induced liver weight
induction rather than just strain. In general, the strains and groups within strain that had TCE-
induced body weight decreases had smaller TCE-induced increase in liver weight. Therefore,
only examining liver weight in males as an indication of TCE treatment effects would not be an
accurate predictor of strain sensitivity nor the magnitude or response at doses that also affect
body weight. The results from this study show that comparison of the magnitude of TCE
02/12/09
E-196
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
response, as measured by liver weight increases, should take into account, strain, gender, initial
body weight and systemic toxicity. It shows a consistent pattern of increased liver weight in
both male and female mice after TCE exposure of 150 ppm for 30 days.
Kjellstrand et al. (1983b) presented data in the NMRI strain of mice (a strain that
appeared to be more prone to TCE-induced toxicity in male mice and a smaller TCE-induced
increase in liver weight in female mice) after inhalation exposure of 37 to 300 ppm TCE. They
used the same experimental paradigm as that reported in Kjellstrand et al (1983a) except for
exposure concentration. For female mice exposed to concentrations of TCE ranging from 37
ppm to 300 ppm TCE continuously for 30 days, only the 300 pm group experienced a 16%
decrease in body weight between control and exposed animals and therefore changes in TCE-
induced liver weight increases were affected by changes in body weight only for that group.
Initial body weights in the TCE-exposed female mice were similar in each of these groups (i.e.,
range of 29.2 - 31.6 g, or 8%), with the exception of the females exposed to 150 ppm TCE for
30 days (i.e., initial body weight of 27.3 g), reducing the effects of differences in initial body
weight on TCE-induced liver weight induction. Exposure to TCE continuously for 30 days was
reported to result in a linear dose-dependent increase in liver weight in female mice with 1.06-
fold, 1.27-fold, 1.66-fold, and 2.14-fold of control liver weights reported at 37 ppm, 75 ppm,
150 ppm, and 300 ppm TCE, respectively. In male mice there were more factors affecting
reported liver weight increases from TCE exposure. For male mice both the 150 ppm and 300
ppm exposed groups experienced a 10% and 18% decrease in final body weight after TCE
exposure, respectively. The 37 ppm and 75 ppm groups did not have decreased final body
weight due to TCE exposure but varied by 12% in initial body weight. TCE-induced increases
in liver weight were reported to be 1.15-fold, 1.50-fold, 1.69-fold, and 1.90-fold of control for
37 ppm, 75 ppm, 150 ppm, and 300 ppm TCE exposure in male mice, respectively. The
flattening of the dose-response curve at the two highest doses is consistent with the effects of
toxicity on final body weight.
Kjellstrand et al. (1983b) noted that liver mass increase and the changes in liver cell
morphology were similar in TCE-exposed male and female mice and report that after 150 ppm
exposure for 30 days, liver cells were generally larger and often displayed a fine vacuolization
of the cytoplasm, changes in nucleoli appearance, Kupffer cells of the sinusoid to be increased
in cellular and nuclear size, the intralobular connective tissue was infiltrated by inflammatory
cells and for exposure to TCE in higher or lower concentrations during the 30 days to produce a
similar morphologic picture. For mice that were exposed to 150 ppm TCE for 30 days and then
examined 120 days after the cessation of exposure, liver weights were 1.09-fold of control for
TCE-exposed female mice and the same as controls for TCE-exposed male mice. However, the
livers were not the same as untreated liver in terms of histopathology. The authors reported that
"after exposure to 150 ppm for 30 days, followed by 120 days of rehabilitation, the
morphological picture was similar to that of the air-exposure controls except for changes in
cellular and nuclear sizes." The authors did not present any quantitative data on the lesions
they describe, especially in terms of dose-response, and most of the qualitative description is for
the 150 ppm exposure level in which there are consistent reports of TCE induced body weight
02/12/09
E-197
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
decreases in male mice. Although stating that Kupffer cells were increased in cellular and
nuclear size, no differential staining was applied to light microscopy sections and used to
distinguish Kupffer from endothelial cells lining the hepatic sinusoid in this study. Without
differential staining such a determination is difficult at the light microscopic level and a question
remains as to whether theses are the same cells as described by Goel et al (1992) as a
proliferation of sinusoidal endothelial cells after exposures of 1000 mg/kg/day and 2000
mg/kg/day TCE exposure for 28 days in male Swiss mice. As noted in Section 2.4.2., the
discrepancy in DNA synthesis measures between hepatocyte examinations of individual
hepatocytes and whole liver measures in several reports of TCE metabolite exposure, is
suggestive of increased DNA synthesis in the nonparenchymal cell compartment of the liver.
Thus, non-parenchymal cell proliferation is suggested as an effect of subchronic TCE exposures
in mice without concurrent focal necrosis via inhalation studies (Kjellstrand et al., 1983b) and
with focal necrosis in the presence of TCE in a groundnut oil vehicle (Goel et al., 1992).
Although Kjellstrand et al. (1983b) did not discuss polyploidization, the changes in cell
size and especially the continued change in cell size and nuclear staining characteristics after
120 days of cessation of exposure are consistent with changes in polyploidization induced by
TCE that were suggested in studies from shorter durations of exposure (Elcombe et al., 1985;
Dees and Travis, 1993) and of longer durations (e.g., Buben and O'Flaherty). Of note is that in
the histological description provided by Kjellstrand et al (1983b), there is no mention of focal
necrosis or apoptosis resulting from these exposures to TCE to mice. Vacuolization is reported
and consistent with hepatotoxicity or lipid accumulation, which is lost during routine
histological slide preparation. The lack of reported focal necrosis in mice exposed through
inhalation is consistent with reports of gavage experiments of TCE in mice that do not use corn
oil as the vehicle (Merrick et al., 1989).
Buben and O'Flaherty (1985) reported the effects of TCE via corn oil gavage after six
weeks of exposure at concentrations ranging from 100 to 3200 mg/kg day. This study was
conducted with older mice than those generally used in chronic exposure assays (Male Swiss-
Cox outbred mice between 3 and 5 months of age). Liver weight increases, decreases in liver
glucose-6-phosphate (G6P) activity, increases in liver triglycerides, and increases in serum
glutamate-pyruvate transaminase (SGPT) activity were examined as parameters of liver toxicity.
Few deaths were reported during the 6-week exposure period except at the highest dose and
related to central nervous system depression. TCE exposure caused dose-related increases in %
liver/body weight with a dose as low as 100 mg/kg/day were reported to cause a statistically
significant increase (i.e., 112% of control). The increases in liver size were attributed to
hepatocyte hypertrophy, as revealed by histological examination and by a decrease in the liver
DNA concentration, and although enlarged, were reported to appear normal. A dose-related
trend toward triglyceride concentration was also noted. A dose-related decrease in glucose-6-
phophatase activity was reported with similar small decreases (~ 10%) observed in the TCE
exposed groups that did not reach statistical significance until the dose reached 800 mg/kg TCE
exposure. SGPT activity was not observed to be increased in TCE-treated mice except at the
two highest doses and even at the 2400 mg/kg dose half of the mice had normal values. The
02/12/09
E-198
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
large variability in SGPT activity was indicative of heterogeneity of this response between mice
at the higher exposure levels for this indicator of liver toxicity. Such variability of response in
male mice is consistent with the work of Kjellstrand et al. Thus, the results from Buben and
O'Flaherty (1985) suggest that hepatomegaly is a robust response that was reported to be
observed at the lowest dose tested, dose-related, and not accompanied by overt toxicity.
In terms of histopathology, Buben and O'Flaherty (1985) reported swollen hepatocytes
with indistinct borders; their cytoplasm was clumped and a vesicular pattern was apparent and
not simply due to edema in TCE-treated male mice. Karyorhexis (the disintegration of the
nucleus) was reported to be present in nearly all specimens from TCE-treated animals and
suggestive of impending cell death, not present in controls, and to appear at a low level at 400
mg/kg TCE exposure level and slightly higher at 1600 mg/kg TCE exposure level. Central
lobular necrosis was present only at the 1600 mg/kg TCE exposure level and at a very low level.
Buben and O'Flaherty report increased polyploidy in the central lobular region for both 400
mg/kg and 1600 mg/kg TCE and described as hepatic cells having two or more nuclei or
enlarged nuclei containing increased amounts of chromatin, but at the lowest level of severity or
occurrence. Thus, the results of this study are consistent with those of shorter-term studies via
gavage which report hepatocellular hypertrophy in the centralobular region, increased liver
weight induced at the lowest exposure level tested and at a level much lower than those inducing
overt toxicity, and that TCE exposure is associated with changes in ploidy.
The National Toxicology Program 13- week study of TCE gavage exposure in 10
F344/N rats [125 to 2000 mg/kg (males) and 62.5 to 1000 mg/kg (females)] and in B6C3Flmice
[375 to 6000 mg/kg] reported all rats survived the 13-week study, but males receiving 2000
mg/kg exhibited a 24% difference in final body weight. The study descriptions of pathology in
rats and mice were not very detailed and included only mean liver weights. The rats had
increased pulmonary vasculitis at the highest concentration of TCE and that viral titers were
positive for Sendai virus and no liver effects were noted for them in the study. For mice, liver
weights (both absolute and % liver/body weight) were reported to increase in a dose-related
fashion with TCE -exposure and to be increased by more than 10% in 750 mg/kg TCE- exposed
males and 1500 mg/kg or more TCE-exposed females. Hepatotoxicity was reported as
centrilobular necrosis in 6/10 males and 1/10 females exposed to 6000 mg/kg TCE and
multifocal areas of calcifications scattered throughout 3000 mg/kg TCE exposed male mice and
only a single female 6000 mg/kg dose, considered to be evidence of earlier hepatocellular
necrosis. One female mouse exposed to 3000 mg/kg TCE also had a hepatocellular adenoma, an
extremely rare lesion in female mice of this age (20 weeks). However, at the lowest dose of
exposure, was a consistent decrease in liver weigh in female and male mice after 13 weeks of
TCE exposure.
Kawamoto et al (1988) exposed rats to 2g/kg TCE subcutaneously for 15 weeks and
reported TCE-induced increases in liver weight. They also reported increase in cytochrome
P450, cytochrome b-5, and NADPH cytochrome c reductase. The difficulties in relating this
route of exposure to more environmentally relevant ones is discussed in Section 2.2.11.
02/12/09
E-199
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For 2-year or lifetime studies of TCE exposure a consistent hepatocarcinogenic response
has been observed in mice of differing strains and genders and from differing routes of
exposure. However, for rat studies some studies have been confounded by mortality from
gavage error or the toxicity of the dose of TCE administered. In some studies, a relative
insensitive strain of rat has been used. However, in general it appears that the mouse is more
sensitive than the rat to TCE-induced liver cancer. Three studies give results the authors
consider to be negative for TCE-induced liver cancer in mice, but have either design and/or
reporting limitations, or are in strains and paradigms with apparent low ability for liver cancer
induction or detection.
Fukuda et al (1983) reported a 104-week inhalation bioassay in female Crj:CD-l (ICR)
mice and female Crj :CD (SD) rats exposed to 0, 50, 150 and 450 ppm TCE (n=50). There were
no reported incidences of mice or rats with liver tumors for controls indicative of relatively
insensitive strains used in the study for liver effects. While TCE was reported to induce a
number of other tumors in mice and rats in this study, the incidence of liver tumors was less than
2% after TCE exposure. Of note is the report of cystic cholangioma reported in 1 group of rats.
Henschler et al. (1980) exposed NMRI mice and WIST random bred rats to 0. 100 and
500 ppm TCE for 18 months (n=30). This study is limited by short duration of exposure, low
number of animals, and low survival in rats. Control male mice were reported to have one
hepatocellular carcinoma and 1 hepatocellular adenoma with the incidence rate unknown. In the
100 ppm TCE exposed group, 2 hepatocellular adenomas and 1 mesenchymal liver tumor were
reported. No liver tumors were reported at any dose of TCE in female mice or controls. For
male rats, only 1 hepatocellular adenomas at 100 ppm was reported. For female rats no liver
tumors were reported in controls, but 1 adenoma and 1 cholangiocarcinoma was reported at 100
ppm TCE and at 500 ppm TCE, 2 cholangioadenomas, a relatively rare biliary tumor, was
reported. The difference in survival in mice, did not affect the power to detect a response, as
was the case for rats. However, the low number of animals studied, abbreviated exposure
duration, and apparently low sensitivity of this paradigm (i.e., no background response in
controls) suggests a study of limited ability to detect a TCE carcinogenic liver response. Of note
is that both Fukuda et al (1983) and Henschler et al. (1980) report rare biliary cell derived
tumors in rats in relatively insensitive assays.
Van Duuren et al. (1979), exposed mice to 0.5 mg/mouse to TCE via gavage once a
week in 0.1 ml trioctanion (n= 30). Inadequate design and reporting of this study limit that
ability to use the results as an indicator of TCE carcinogenicity.
The NCI (1976) study of TCE was initiated in 1972 and involved the exposure of
Osborn-Mendel rats and B6C3F1 mice to varying concentrations of TCE. The animals were co-
exposed to a number of other carcinogens as exhalation as multiples studies and control animals
all shared the same laboratory space. Treatment duration was 78 weeks and animals received
TCE via gavage in corn oil at 2 doses (n= 20 for controls, but n= 50 for treatment groups). For
02/12/09
E-200
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
rats, the high dose was reported to result in significant mortality (i.e., 47/50 high-dose rats died
before scheduled termination of the study). A low incidence of liver tumors was reported for
controls and carbon tetrachloride positive controls in rats from this study. In B6C3F1 mice,
TCE was reported to increase incidence of hepatocellular carcinomas in both doses and both
genders of mice (~ 1170 and 2340 mg/kg for males and 870 and 1740 mg/kg for female mice).
Hepatocellular carcinoma diagnosis was based on histologic appearance and metastasis to the
lung. The tumors were described in detail and to be heterogeneous "as described in the
literature" and similar in appearance to tumors generated by carbon tetrachloride. The
description of liver tumors in this study and tendency to metastasize to the lung are similar to
descriptions provided by Maltoni et al (1986) for TCE-induced liver tumors in mice via
inhalation exposure.
For male rats, noncancer pathology in the NCI (1976) study was reported to include
increased fatty metamorphosis after TCE exposure and angiectasis or abnormally enlarged blood
vessels. Angiectasis can be manifested by hyperproliferation of endothelial cells and dilatation
of sinusoidal spaces. The authors conclude that due to mortality, "the test is inconclusive in
rats." They note the insensitivity of the rat strain used to the positive control of carbon
tetrachloride exposure.
The NTP (1990) study of TCE exposure in male and female F344/N rats, and B6C3F1
mice (500 and 1000 mg/kg for rats, and 1000 mg/kg for mice) is limited in the ability to
demonstrate a dose-response for hepatocarcinogenicity. There was also little reporting of non-
neoplastic pathology or toxicity and no report of liver weight at termination of the study.
However by the end of a 2-year cancer bioassay, liver tumor induction can be a significant
factor in any changes in liver weight. No treatment-related increase in necrosis in the liver was
observed in mice. A slight increase in the incidence of focal necrosis was noted for TCE-
exposed male mice (8% vs. 2% in control) with a slight reduction in fatty metamorphosis in
treated male mice (0 treated vs. 2 control animals) and in female mice a slight increase in focal
inflammation (29% vs. 19% of animals) and no other changes. Therefore this study did not
show concurrent evidence of liver toxicity but did show TCE-induced neoplasia after 2 years of
TCE exposure in mice. The administration of TCE was reported to cause earlier expression of
tumors as the first animals with carcinomas were 57 weeks for TCE-exposed animals and 75
weeks for control male mice.
The NTP (1990) study reported that TCE exposure was associated with increased
incidence of hepatocellular carcinoma (tumors with markedly abnormal cytology and
architecture) in male and female mice. Hepatocellular adenomas were described as
circumscribed areas of distinctive hepatic parenchymal cells with a perimeter of normal
appearing parenchyma in which there were areas that appeared to be undergoing compression
from expansion of the tumor. Mitotic figures were sparse or absent but the tumors lacked
typical lobular organization. Hepatocellular carcinomas had markedly abnormal cytology and
architecture with abnormalities in cytology cited as including increased cell size, decreased cell
size, cytoplasmic eosinophilia, cytoplasmic basophilia, cytoplasmic vacuolization, cytoplasmic
02/12/09
E-201
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
hyaline bodies and variations in nuclear appearance. Furthermore, in many instances several or
all of the abnormalities were present in different areas of the tumor and variations in architecture
with some of the hepatocellular carcinomas having areas of trabecular organization. Mitosis
was variable in amount and location. Therefore the phenotype of tumors reported from TCE
exposure was heterogenous in appearance between and within tumors.
For rats, the NTP (1990) study reported no treatment-related non-neoplastic liver lesions
in males and a decrease in basophilic cytological change reported from TCE-exposure in female
rats. The results for detecting a carcinogenic response in rats were considered to be equivocal
because both groups receiving TCE showed significantly reduced survival compared to vehicle
controls and because of a high rate (e.g., 20% of the animals in the high-dose group) of death by
gavage error.
The NTP (1988) study of TCE exposure in four strains of rats to "diisopropylamine-
stabilized TCE" was also considered inadequate for either comparing or assessing TCE-induced
carcinogenesis in these strains of rats because of chemically induced toxicity, reduced survival,
and incomplete documentation of experimental data. TCE gavage exposures of 0, 500 or 1000
mg/kg per day (5 days per week, for 103 weeks) male and female rats was also marked by a
large number of accidental deaths (e.g., for high-dose male Marshal rats 25 animals were
accidentally killed. Results from a 13-week study were briefly mentioned in the report and
indicated exposure levels of 62.5- 2000 mg/kg TCE were not associated with decreased survival
(with the exception of 3 male August rats receiving 2000 mg/kg TCE) and that the
administration of the chemical for 13 weeks was not associated with histopathological changes.
In regard to evidence of liver toxicity, the 2-year study of TCE exposure reported no evidence of
TCE-induced liver toxicity described as non-neoplastic changes ACI, August, Marshal, and
Osborne-Mendel rats. Interestingly, for the control animals of these four strains there was, in
general, a low background level of focal necrosis in the liver of both genders. In summary, the
negative results in this bioassay are confounded by the killing of a large portion of the animals
accidently by experimental error but TCE-induced overt liver toxicity was not reported.
Maltoni et al. (1986) reported the results of several studies of TCE via inhalation and
gavage in mice and rats. A large number of animals were used in the treatment groups but the
focus of the study was detection of a neoplastic response with only a generalized description of
tumor pathology phenotype given and limited reporting of non-neoplastic changes in the liver.
Accidental death by gavage error was reported not to occur in this study. In regards to effects of
TCE exposure on survival, "a nonsignificant excess in mortality correlated to TCE treatment
was observed only in female rats (treated by ingestion with the compound) and in male B6C3F1
mice. TCE-induced effects on body weight were reported to be absent in mice except for one
experiment (BT 306 bis) in which a slight non-dose correlated decrease was found in exposed
animals. "Hepatoma" was the term used to describe all malignant tumors of hepatic cells, of
different sub-histotypes, and of various degrees of malignancy and were reported to be unique or
multiple, and have different sizes (usually detected grossly at necropsy) from TCE exposure. In
regard to phenotype tumors were described as usual type observed in Swiss and B6C3F1 mice,
02/12/09 E-202
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
as well as in other mouse strains, either untreated or treated with hepatocarcinogens and to
frequently have medullary (solid), trabecular, and pleomorphic (usually anaplastic) patterns.
Swiss mice from this laboratory were reported to have a low incidence of hepatomas without
treatment (1%). The relatively larger number of animals used in this bioassay (n= 90 to 100), in
comparison to NTP standard assays, allows for a greater power to detect a response.
TCE exposure for 8 weeks via inhalation at 100 ppm or 600 ppm may have been
associated with a small increase in liver tumors in male mice in comparison to concurrent
controls during the life span of the animals. In Swiss mice exposed to TCE via inhalation for 78
weeks there a reported increase in hepatomas associated with TCE treatment that was dose-
related in male but not female Swiss mice. In B6C3F1 mice exposed via inhalation to TCE for
78 weeks, the results from one experiment indicated a greater increase in liver cancer in females
than male mice but in a second experiment in males there was a TCE-exposure associated
increase in hepatomas. Although the mice were supposed to be of the same strain, the
background level of liver cancer was significantly different in male mice. The finding of
differences in response in animals of the same strain but from differing sources has also been
reported in other studies for other endpoints (see Section 3.1.2.). However, for both groups of
male B6C3F1 mice the background rate of liver tumors over the lifetime of the mice was less
than 20%.
For rats, there were 4 liver angiosarcomas reported (1 in a control male rat, 1 both in a
TCE-exposed male and female at 600 ppm TCE for 8 weeks, and 1 in a female rat exposed to
600 ppm TCE for 104 weeks) but the specific results for incidences of hepatocellular
"hepatomas" in treated and control rats were not given. Although the Maltoni et al. (1986)
concluded that the small number was not treatment-related, the findings were brought forward
because of the extreme rarity of this tumor in control Sprague-Dawley rats, untreated or treated
with vehicle materials. In rats treated for 104 weeks, there was no report of a TCE treatment-
related increase in liver cancer in rats. This study only presented data for positive findings so it
did not give the background or treatment-related findings in rats for liver tumors in this study.
Thus, the extent of background tumors and sensitivity for this endpoint cannot be determined.
Of note is that the Sprague-Dawley strain used in this study was also noted in the Fukuda et al
(1983) study to be relatively insensitive for spontaneous liver cancer and to also be negative for
TCE-induced hepatocellular liver cancer induction in rats. However, like Fukuda et al (1983)
and Henschler et al (1980), that reported rare biliary tumors in insensitive strains of rat for
hepatocellular tumors, Maltoni et al, (1986) reported a relatively rare tumor type, angiosarcoma,
after TCE exposure in a relatively insensitive strain for "hepatomas." As noted above, many of
the rat studies were limited by premature mortality due to gavage error or premature mortality
(Henschler et al 1980; NCI 1976; NTP 1990; NTP 1988), which was reported not occur in
Maltoni et al (1986).
There were other reports of TCE carcinogenicity in mice from chronic exposures that
were focused primarily on detection of liver tumors with limited reporting of tumor phenotype
or non-neoplastic pathology. Herren-Freund et al. (1997) reported that male B6C3 F1 mice
02/12/09
E-203
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
given 40 mg/1 TCE in drinking water had increased tumor response after 61 weeks of exposure.
However, concentrations of TCE fell by about V2 at this dose of TCE during the twice a week
change in drinking water solution so the actual dose of TCE the animals received was less than
40 mg/1. The % liver /body weight was reported to be similar for control and TCE -exposed
mice at the end of treatment. However, despite difficulties in establishing accurately the dose
received, an increase in adenomas per animal and an increase in the number of animals with
hepatocellular carcinomas were reported to be associated with TCE exposure after 61 weeks of
exposure and without apparent hepatomegaly. Anna et al., (1994) reported tumor incidences for
male B6C3F1 mice receiving 800 mg/kg/day TCE via gavage (5 days/week for 76 weeks). All
TCE-treated mice were reported to be alive after 76 weeks of treatment. Although the control
group contained a mixture of exposure durations (76 -134 weeks) and concurrent controls had a
very small number of animals, TCE-treatment appeared to increase the number of animals with
adenomas, the mean number of adenomas and carcinomas, but with no concurrent TCE-induced
cytotoxicity.
2.4.4. Summary of Results For Sub-chronic and Chronic Effects of DCA and TCA:
Comparisons With TCE
There are no similar studies for TCA and DCA conduced at 6 weeks and with the range
of concentrations examined in Buben and O'Flaherty (1985) for TCE. In general, many studies
of DCA and TCA have been conducted at few and high concentrations, with shortened durations
of exposure, and varying and low numbers of animals to examine primarily a liver tumor
response in mice. However, the analyses presented in Section 2.4.2 gives comparisons of
administered TCA and DCA dose-responses for liver weight increases for a number of studies in
combination as well as comparing such dose-responses to that of TCE and its oxidative
metabolism. As stated above, many subchronic studies of DCA and TCA have focused on
elucidating a relationship between dose and hypothesized events that may be indicators of
carcinogenic potential that have been described in chronic studies with a focus on indicators of
peroxisome proliferation and DNA synthesis. Many chronic studies have focused on the nature
of the DCA and TCA carcinogenic response in mouse liver through examination of the tumors
induced.
Most all of the chronic studies for DCA and TCA have been carried out in mice. As the
database for examination of the ability of TCE to induce liver tumors in rats includes several
studies that have been limited in ability determine a carcinogenic response in the liver, the
database for DCA and TCA in rats is even more limited. For TCA, the only available study in
rats (DeAngelo et al 1997) has been frequently cited in the literature to indicate a lack of
response in this species for TCA-induced liver tumors. Although reporting an apparent dose-
related increase in multiplicity of adenomas and an increase in carcinomas over control at the
highest dose, DeAngelo et al. (1997) use such a low number of animals per treatment group (n=
20-24) that the ability of this study to determine a statistically significant increase in tumor
response and to be able to determine that there was no treatment-related effect are limited. A
power calculation of the study shows that the type II error, which should be > 50%, was less
02/12/09
E-204
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
than 8% probability for incidence and multiplicity of all tumors at all exposure DCA
concentrations with the exception of the incidence of adenomas and adenomas and carcinomas
for 0.5 g/1 treatment group (58%) in which there was an increased in adenomas reported over
control (15% vs. 4%) that was the same for adenomas and carcinomas combined. Therefore, the
designed experiment could accept a false null hypothesis and erroneously conclude that there is
no response due to TCA treatment. Thus, while suggesting a lower response than for mice for
liver tumor induction, it is inconclusive for determination of whether TCA induces a
carcinogenic response in the liver of rats.
For DCA, there are two reported long-term studies in rats (DeAngelo et al, 1996;
Richmond et al., 1995) that appear to have reported the majority of their results from the same
data set and which consequently were subject to similar design limitations and DCA-induced
neurotoxicity in this species. DeAngelo et al (1996) reported increased hepatocellular adenomas
and carcinomas in male F344 rats exposed for 2 years. However, the data from exposure
concentrations at a 5 g/1 dose had to be discarded and the 2.5 g/1 DCA dose had to be
continuously lowered during the study due to neurotoxicity. There was a DCA-induced
increased in adenomas and carcinomas combined reported for the 0.5 g/1 DCA (24.1 % vs. 4.4%
adenomas and carcinomas combined in treated vs. controls) and an increase at a variable dose
started at 2.5 g/1 DCA and continuously lowered (28.6% vs. 3.0% adenomas and carcinomas
combined in treated vs. controls). Only combined incidences of adenomas and carcinomas for
the 0.5 g/1 DCA exposure group was reported to be statistically significant by the authors
although the incidence of adenomas was 17.2% vs. 4% in treated vs. control rats.
Hepatocellular tumor multiplicity was reported to be increased in the 0.5 g/1 DCA group (0.31
adenomas and carcinomas/animal in treated vs. 0.04 in control rats) but was reported by the
authors to not be statistically significant. At the starting dose of 2.5 g/1 that was continuously
lowered due to neurotoxicity, the increased multiplicity of hepatocellular carcinomas was
reported by the authors to be to be statistically significant (0.25 carcinomas/animals vs. 0.03 in
control) as well as the multiplicity of combined adenomas and carcinomas (0.36 adenomas and
carcinomas/animals vs. 0.03 in control rats). Issues that affect the ability to determine the nature
of the dose-response for this study include (1) the use of a small number of animals (n= 23,
n=21 and n= 23 at final sacrifice for the 2.0 g/1 NaCl control, 0.05 g/1 and 0.5 g/1 treatment
groups) that limit the power of the study to both determine statistically significant responses and
to determine that there are not treatment-related effects (i.e. power) (2) apparent addition of
animals for tumor analysis not present at final sacrifice (i.e., 0.05 and 0.5 g/1 treatment groups),
and (3) most of all, the lack of a consistent dose for the 2.5 g/1 DCA exposed animals. Similar
issues are present for the study of Richmond et al (1995) which was conducted by the same
authors as DeAngelo et al (1996) and appeared to be the same data set. The Richmond et al.
(1995) data for the 2 g/1 NaCl. 0.05 g/1 DCA and 0.5 g/1 DCA exposure groups were the same
data set reported by DeAngelo et al (1996) for these groups. Additional data was reported for
F344 rats administered and 2.5 g/1 DCA that, due to hind-limb paralysis, were sacrificed 60
weeks (DeAngelo et al., 1996). Tumor multiplicity was not reported by the authors. There was
a small difference in reports of the results between the two studies for the same data for the 0.5
g/1 DCA group in which Richmond et al. (1995) reported a 21% incidence of adenomas and
02/12/09
E-205
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
DeAngelo et al. (1996) reported a 17.2% incidence. The authors did not report any of the results
of DCA-induced increases of adenomas and carcinomas to be statistically significant. The same
issues discussed above for DeAngelo et al (1996) apply to this study. Similar to the DeAngelo
study of TCA in rats (DeAngelo et al. 1997) the study of DCA exposure in rats reported by
DeAngelo et al. (1996) and Richmond et al (1995), the use of small numbers of rats limits the
detection of treatment-related effects and the ability to determine whether there was no
treatment related effects (type II error), especially at the low concentrations of DCA exposure.
For mice the data for both DCA and TCA is much more extensive and has shown that
both DCA and TCA induced liver tumors in mice. Many of the studies are for relatively high
concentrations of DCA or TCA, have been conducted for a year or less, and have focused on the
nature of tumors induced to ascertain potential MO As and to make inferences as to whether
TCE-induced tumors in mice are similar. As shown previously in Section 2.4.2, the dose-
response curves for increased liver weight for TCE administration in male mice are more similar
to those for DCA administration and TCE oxidative metabolism than for direct TCA
administration. There are two studies in male B6C3F1 mice that attempt to examine multiple
concentrations of DCA and TCA for 2-year studies (DeAngelo et al., 1999; DeAngelo et al.,
2008) at doses that do not induce cytotoxicity and attempt to relate them to subchronic changes
and peroxisomal enzyme induction. However, the DeAngelo et al (2008) study was carried out
in B6C3F1 mice that were of large size and prone to liver cancer and premature mortality
limiting its use for the determination of TCA-dose response in a 2-year bioassay. One study in
female B6C3F1 mice describes the dose-response for liver tumor induction at a range of DCA
and TCA concentrations after 51 or 82 weeks (Pereira, 1996) with a focus on the type of tumor
each compound produced.
DeAngelo et al. (1999) conducted a study of DCA exposure to determine a dose
response for the hepatocarcinogenicity of DCA in male B6C3F1 mice over a lifetime exposure
and especially at concentrations that did not illicit cytotoxicity or were for abbreviated exposure
durations. DeAngelo et al. (1999) used 0.05 g/1, 0.5 g/1, 1.0 g/1, 2.0 g/1 and 3.5 g/1 exposure
concentrations of DCA in their 100-week drinking water study. The number of animals at final
sacrifice was generally low in the DCA treatment groups and variable (i.e., n= 50, n= 33, n= 24,
n= 32, n= 14, and n= 8 for control, 0.05 g/1, 0.5 g/1, 1 g/1, 2.0 g/1 and 3.5 g/1 DCA exposure
groups). It was apparent that animals that died unscheduled deaths between weeks 79 and 100
were included in data reported for 100 weeks. Although the authors did not report how many
animals were included in the 100-week results, it appeared that the number was no greater than
1 for the control, 0.05, and 0.5 exposure groups and varied between 3 and 7 for the higher DCA
exposure groups. The multiplicity or number of hepatocellular carcinomas/animals was
reported to be significantly increased over controls in a dose-related manner at all DCA
treatments including 0.05 g/1 DCA, and a NOEL reported not to be observed by the authors (i.e.,
0.28, 0.58, 0.68, 1.29, 2.47, and 2.90 hepatocellular carcinomas/ animal for control, 0.05, 0.5,
1.0, 2.0, and 3.5 g/1 DCA). Between the 0.5 g/1 and 3.5 g/1 exposure concentrations of DCA the
magnitude of increase in multiplicity was similar to the increases in magnitude in dose. The
incidence of hepatocellular carcinomas were reported to be increased at all doses as well but not
02/12/09
E-206
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
reported to be statistically significant at the 0.05 g/1 exposure concentration. However given
that the number of mice examined for this response (n= 33), the power of the experiment at this
dose was only 16.9% to be able to determine that there was not a treatment related effect. The
authors did not report the incidence or multiplicity of adenomas for the 0.05 g/1 exposure group
in the study and neither did they report the incidence or multiplicity of adenomas and
carcinomas in combination. For the animals surviving from 79 to 100 weeks of exposure, the
incidence and multiplicity of adenomas peaked at 1 g/1 while hepatocellular carcinomas
continued to increase at the higher doses. This would be expected where some portion of the
adenomas would either regress or progress to carcinomas at the higher doses.
DeAngelo et al (1999) reported that peroxisome proliferation was significantly increased
at 3.5 g/1 DCA only at 26 weeks, not correlated with tumor response, and to not be increased at
either 0.05 g/1 or 0.5 g/1 treatments. The authors concluded that DCA-induced carcinogenesis
was not dependent on peroxisome proliferation or chemically sustained proliferation, as
measured by DNA synthesis. DeAngelo et al (1999) reported not only a dose-related increase in
DCA-induced liver tumors but also a decrease in time-to-tumor associated with DCA exposure
at the lowest levels examined. In regards to cytotoxicity there appeared to be a treatment but not
dose-related increase in hepatocellular necrosis that did not involve most of the liver from 1 g/1
to 3.5 g/1 DCA exposures for 26 weeks of exposure that decreased by 52 weeks with no necrosis
observed at the 0.5 g/1 DCA treatment for any exposure period.
Hepatomegaly was reported to be absent by 100 weeks of exposure at the 0.05g/l and 0.5
g/1 exposures while there was an increase in tumor burden reported. However, slight
hepatomegaly was present by 26 weeks in the 0.5 g/1 group and decreased with time. Not only
did the increase in multiplicity of hepatocellular carcinomas increase proportionally with DCA
exposure concentration after 79-100 weeks of exposure, but so did the increases in % liver/body
weight. DeAngelo et al. (1999) presented a figure comparing the number of hepatocellular
carcinomas/animal at 100 weeks compared with the % liver/body weight at 26 weeks that
showed a linear correlation (r2 = 0.9977) while peroxisome proliferation and DNA synthesis did
not correlate with tumor induction profiles. The proportional increase in liver weight with DCA
exposure was also reported for shorter durations of exposure as noted in Section 2.4.2. The
findings of the study illustrates the importance of examining multiple exposure levels at lower
concentrations, at longer durations of exposure and with an adequate number of animals to
determine the nature of a carcinogenic response. Although Carter et al. (1995) suggested that
there is evidence of DCA-induced cytotoxicity (e.g., loss of cell membranes and apparent
apoptosis) at higher levels, the 0.5 g/1 exposure concentration has been shown by DeAngelo et al
(1999) to increase hepatocellular tumors after 100 weeks of treatment without concurrent
peroxisome proliferation or cytotoxicity in mice.
As noted in detail in Section 2.3.2.13, DeAngelo et al. (2008) exposed male B6C3F1
mice to neutralized TCA in drinking water to male B6C3 F1 mice in three studies. Rather than
using 5 exposure levels that were generally 2-fold apart, as was done in DeAngelo et al. (1999)
for DCA, DeAngelo et al (2008) studied only 3 doses of TCA that were an order of magnitude
02/12/09
E-207
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
apart which limits the elucidation of the shape of the dose-response curve. In addition
DeAngelo et al. (2008) contained 2 studies, each conducted in a separate laboratories, for the
104-week data so that the two lower doses were studied in one study and the highest dose in
another. The first study was conducted using 2 g/1 NaCl, or 0.05, 0.5, or 5 g/1 TCA in drinking
water for 60 weeks (Study #1) while the other two were conducted for a period of 104 weeks
(Study #2 with 2.5 g/1 neutralized acetic acid or 4.5 g/1 TCA exposure groups and Study #3 with
deionized water, 0.05 g/1 TCA and 0.5 g/1 TCA exposure groups). In the studies reported in
DeAngelo et al (2008) a small number of animals has been used for the determination of a tumor
response (~ n= 30 at final necropsy), but for the data for liver weight or PCO activity at interim
sacrifices the number was even smaller (n= 5). The % liver/body weight changes at 4 weeks in
Study #1 have been included in the analysis for all TCA data in Section 2.4.2, and are consistent
with that data. Although there was a 10-fold difference in TCA exposure concentration, there
was a 9%, 16% and 35% increase in liver weight over control for the 0.05g/l 0.5 g/1 and 5 g/1
TCA exposures. PCO activity varied 2.7-fold as baseline controls but the increase in PCO
activity at 4 weeks was 1.3-fold, 2.4-fold, and 5.3-fold of control for the 0.05 g/1, 0.5 g/1 and 5
g/1 TCA exposure groups in Study #1. The incidence data for adenomas observed at 60 weeks
was 2.1-fold, 3.0-fold and 5.4-fold of control values and the fold increases in multiplicity were
similar after 0.05 g/1, 0.5 g/1 and 5.0 g/1 TCA. Thus, in general the dose-response for TCA-
induced liver weight increases at 4 weeks was similar to the magnitude of induction of
adenomas at 60 weeks. Such a result is more consistent with the ability of TCA to induce tumors
and increases in liver weight at low doses with little change with increasing dose as shown by
this study and the combined data for TCA liver weight induction by administered TCA
presented in Section 2.4.2.
While the 104-week data from Study's #2 and #3 could have been more valuable for
determination of the dose-response as it would have allowed enough time for full tumor
expression, serious issues are apparent for Study #3 which was reported to have a 64%
incidence rate of adenomas and carcinomas for controls while that of Study #2 was 12%. As
stated in Section 2.3.2.13, the mice in Study #3 were of larger size than those of either Study #1
or #2 and the large background rate of tumors reported is consistent with mice of these size
(Leakey et al., 2003b). However, the large background rate and increased mortality for these
mice limit their use for determining the nature of the dose-response for TCA liver
carcinogenicity. Examination of the data for treatment groups shows that there was no
difference in any of the results between the 0.5 g/1 (Study #3) and 5 g/1 (Study #2) TCA
exposure groups (i.e., adenoma, carcinoma, and combinations of adenoma and carcinoma
incidence and multiplicity) for 104 weeks of exposure. For these same exposure groups, but at
60 weeks of exposure (Study #1), there was a 2-fold increase in multiplicity for adenomas, and
for adenomas and carcinomas combined between the 0.5 and 5.0 g/1 TCA exposure groups. At
the two lowest doses of 0.05 g/1 and 0.5 g/1 TCA from Study #3 in the large tumor prone mice,
the differences in the incidences and multiplicities for all tumors were 2-fold at 104 weeks.
These results are consistent with (1) the two highest exposure levels reaching a plateau of
response after a long enough duration of exposure for full expression of the tumors (i.e., ~ 90%
of animals having liver tumors at the 0.5 g/1 and 5 g/1 exposures) with the additional tumors
02/12/09
E-208
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
observed in a tumor-prone paradigm. Thus, without use of the 0.05 g/1 and 0.5 g/1 TCA data
from Study #3, only the 4.5 g/1 TCA data from study #2 can be used for determination of the
TCA cancer response in a 2-year bioassay.
To put the 64% incidence data for carcinomas and adenomas reported in DeAngelo et al.
(2008) for the control group of Study #3 in context, other studies cited in this review for male
B6C3F1 mice show a much lower incidence in liver tumors with: (1) NCI (1976) study of TCE
reporting a colony control level of 6.5% for vehicle and 7.1%> incidence of hepatocellular
carcinomas for untreated male B6C3F1 mice (n= 70 - 77) at 78 weeks, (2) Herren-Freund et al
(1987) reporting a 9%> incidence of adenomas in control male B6C3F1 mice with a multiplicity
of 0.09 ± 0.06 and no carcinomas (n= 22) at 61 weeks, (3) NTP (1990) reporting an incidence
of 14.6% adenomas and 16.6%> carcinomas in male B6C3F1 mice after 103 weeks (n= 48), and
(4) Maltoni et al (1986) reporting that B6C3F1 male mice from the "NCI source" had a 1.1%)
incidence of "hepatoma" (carcinomas and adenomas) and those from "Charles River Co." had a
18.9%o incidence of "hepatoma" during the entire lifetime of the mice (n= 90 per group). The
importance of examining an adequate number of control or treated animals before confidence
can be placed in those results in illustrated by Anna et al (1994) in which at 76 weeks 3/10
control male B6C3F1 mice that were untreated and 2/10 control animals given corn oil were
reported to have adenomas but from 76 to 134 weeks, 4/32 mice were reported to have
adenomas (multiplicity of 0.13 ± 0.06) and 4/32 mice were reported to have carcinomas
(multiplicity of 0.12 ± 0.06). Thus, the reported combined incidence of carcinomas and
adenomas of 64%> reported by DeAngelo et al (2008) for the control mice of Study # 3, not only
is inconsistent and much higher than those reported in Studies #1 and #2, but also much higher
than reported in a number of other studies of TCE .
Trying to determine a correspondence with either liver weight increases or increases in
PCO activity after shorter periods of exposure will be depend whether data reported in Study #3
in the 104 week studies can be used. DeAngelo et al. (2008) report a regression analyses that
compare "percent of hepatocellular neoplasia," indicated by tumor multiplicity, with TCA
dose, represented by estimations of the TCA dose in mg/kg/day, and with PCO activity for the
60-week and 104-week data. Whether adenomas and carcinomas combined or individual tumor
type were used in these analysis was not reported by the authors. Concerns arise also from
comparing PCO activity at the end of the experiments, when there was already a significant
tumor response, rather than at earlier time points. Such PCO data may not be useful as an
indicator key event in tumorigenesis when tumors are already present. In addition regression
analyses of this data are difficult to interpret because of the dose spacing of these experiments as
the control and 5 g/1 exposure levels will basically determine the shape of the dose response
curve. The 0.05 g/1 and 0.5 g/1 exposure levels are close to the control value in comparison to
the 5 g/1 exposure level, the dose response appears to be linear between control and the 5.0 g/1
value with the two lowest doses not affectly changing the slope of the line (i.e., "leveraging" the
regression). Thus, the value of these analyses is limited by (1) use of data from Study # 3 in a
tumor prone mouse that is not comparable to those used in Studies #1 and #2, (2) the
appropriateness of using PCO values from later time points and the variability in PCO control
02/12/09
E-209
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
values (3) the uncertainty of the effects of palatability on the 5 g/1 TCA results which were
reported in one study to reduce drinking water consumption, and (4) the dose-spacing of the
experiment.
DeAngelo et al (2008) attempt to identify a NOEL for tumorigenicity using tumor
multiplicity data and estimated TCA dose. However, it is not an appropriate descriptor for these
data, especially given that "statistical significance" of the tumor response is the determinant
used by the authors to support the conclusions regarding a dose in which there is no TCA-
induced effect. Due to issues related to the appropriateness of use of the concurrent control in
Study #3, only the 60-week experiment (i.e., study # 1) is useful for the determination of tumor
dose-response. Not only is there not allowance for full expression of a tumor response at the 60-
week time point but a power calculation of the 60-week study shows that the type II error, which
should be > 50% and thus greater than the chances of "flipping a coin", was 41% and 71% for
incidence and 7% and 15% for multiplicity of adenomas for the 0.05 g/1 and 0.5 g/1 TCA
exposure groups. For the combination of adenomas and carcinomas, the power calculation was
8%> and 92% for incidence and 6% and 56% for multiplicity at 0.05 g/1 and 0.5 g/1 TCA
exposure. Therefore, the designed experiment could accept a false null hypothesis, especially in
terms of tumor multiplicity, at the lower exposure doses and erroneously conclude that there is
no response due to TCA treatment.
Pereira (1996) examined the tumor induction in female B6C3 F1 mice and demonstrate
that foci, adenoma, and carcinoma development in mice are dependent on duration of exposure,
or period of observation in the case of controls, for full expression of a carcinogenic response.
In control female mice a 360- vs. 576-day observation period showed that at 360 days no foci or
carcinomas and only 2.5% of animals had adenomas whereas by 576 days of observation, 11%
had foci, 2% adenomas, and 2% had carcinomas. For DCA and TCA treatments, foci,
adenomas, and carcinoma incidence and multiplicity did not reach full expression until 82
weeks at the 3 doses employed (2.58 g/1 DCA, 0.86 g/1 DCA, 0.26 g/1 DCA, 3.27 g/1 TCA, 1.1.0
g/1 TCA, and 0.33 g/1 TCA). Although the numbers of animals were relatively low and variable
at the two highest doses (18-28 mice) there were 50-53 mice studied at the lowest dose level and
90 animals studied in the control group. The results of Pereira (1996) show that not only were
the incidence of mice with foci, adenoma, and carcinomas greatly increased with duration of
exposure, but that concentration also affected the nature and magnitude of the response in
female mice. At 2.86 g/1, 0.86 g/1, 0.26 g/1 DCA exposures and controls, after 82 weeks the
incidence of adenomas in female B6C3 F1 mice was reported to be 84.2%, 25.0%, 6.0% and
2.2%, respectively, and carcinomas to be 26.3 %, 3.6%, 0%, and 2.2%, respectively. For the
multiplicity or number of tumors/animal at these same exposure levels of DCA, the multiplicity
was reported to be 5.58, 0.32, 0.06 and 0.02 adenomas/animal, and 0.37, 0.04, 0 and 0.02
carcinomas/animal. Thus, for DCA exposure in female mice, for ~ 3-fold increases in DCA
exposure concentration, after 82 weeks of exposure there was a similar magnitude of increase in
adenomas incidence with much greater increases in multiplicity. For hepatocellular carcinoma
induction, there was no increase in the incidence or multiplicity or carcinomas between the
control and 0.33 g/1 DCA dose. At 3.27 g/1, 1.10 g/1. 0.33 g/1 TCA and controls, after 82 weeks
02/12/09
E-210
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the incidence of adenomas in female B6C3F1 mice was reported to be 38.9%, 11.1%, 7.6%, and
2.2%), respectively, and carcinomas to be 27.8%, 18.5%, 0 % and 2.2%, respectively. At these
same exposure levels of TCA, the multiplicity was reported to be 0.61, 0.11, 0.08, and 0.02
adenomas/animal, and 0.39, 0.22, 0, and 0.02 carcinomas/animal, respectively. Thus for TCA,
the incidences of adenomas were lower at the two highest doses than DC A and the ~3-fold
differences in dose between the two lowest doses only resulted in ~ 50% increase in incidences
of adenomas. For incidence of carcinomas the ~ 3-fold difference in dose between the two
highest doses only resulted in ~ 50% increase in carcinoma incidence. A similar pattern was
reported for multiplicity after TCA exposure. Foci were also examined and, in general., were
similar to adenomas regarding incidence and multiplicity. Thus, the dose-response curve for
tumor induction in female mice differed between DCA and TCA after 82 weeks of exposure
with TCA having a much less steep dose-response curve than DCA. This is consistent with the
pattern of liver weight increases reported for male B6C3F1 mice in Section 2.4.2.
DeAngelo et al. (1999) report a linear increase in incidence and multiplicity of
hepatocellular carcinomas that is proportional to dose and as well as proportional to the
magnitude of liver weight increase from subchronic exposure to DCA. However the studies of
DeAngelo et al. (2008) and Pereira (1996) are suggestive that TCA induced increase in tumor
incidence are less proportional to increases in dose as are liver weight increases from subchronic
exposure. Given that TCE subchronic exposure also induced an increase in liver weight that
was proportional to dose (i.e., similar to DCA but not TCA), it is of interest as to whether the
dose-response for TCE induced liver cancer in mice was similar. The database for TCE, while
consistently showing a induction of liver tumors in mice, is very limited for making inferences
regarding the shape of the dose-response curve. For many of these experiments multiplicity was
not given only liver tumor incidence. NTP (1990), Bull et al (2002), Anna et al. (1994)
conducted gavage experiments in which they only tested one dose of - 1000 mg/kg/day TCE.
NCI (1976) tested 2 doses that were adjusted during exposure to an average of 1169 mg/kg/day
and 2339 mg/kg/day in male mice with only 2-fold dose spacing in only 2 doses tested. Maltoni
et al (1988) conducted inhalation experiments in 2 sets of B6C3F1 mice and one set of Swiss
mice at 3 exposure concentrations that were 3-fold apart in magnitude between the low and mid-
dose and 2-fold apart in magnitude between the mid- and high-dose. However for one
experiment in male B6C3F1 mice, the mice fought and suffered premature mortality and for two
the experiments in B6C3F1 mice, although using the same strain, the mice were obtained from
differing sources with very different background liver tumor levels. For the Maltoni et al (1988)
study a general descriptor of "hepatoma" was used for liver neoplasia rather than describing
hepatocellular adenomas and carcinomas so that comparison of that data with those from other
experiments is difficult. More importantly, while the number of adenomas and carcinomas may
be the same between treatments or durations of exposure, the number of adenomas may decrease
as the number of carcinomas increase during the course of tumor progression. Such information
is lost by using only a hepatoma descriptor. Maltoni et al (1988) did not report an increase over
control for 100 ppm TCE for the Swiss group and one of the B6C3F1 groups and only a slight
increase (1.12-fold) in the second B6C3F1 group. At 300 ppm TCE exposure, the incidences of
hepatoma were 2-fold of control values for the Swiss, 4-fold of control for group of B6C3F1
02/12/09
E-211
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mice, and 1.6-fold of control for the other group of B6C3F1 mice. At 600 ppm TCE the
incidences of hepatoma were 3.3-fold of control for the Swiss group, 6.1-fold of control for one
group of B6C3F1 mice, and 1.2-fold for the other group of B6C3F1 mice. Thus, for each group
of TCE exposed mice in the Maltoni et al. (1988) inhalation study, the background levels of
hepatomas and the shape of the dose-response curve for TCE-hepatoma induction were variable.
However, an average of the increases, in terms of fold of control, between the 3 experiments
gives a ~ 2.9- fold increase between the low- and mid-dose (100 ppm and 300 ppm) and ~ 1.4-
fold increase between the mid- and high-dose (300ppm and 600 pm) groups. Although such a
comparison obviously has a high degree of uncertainty associated with it, it suggests that the
magnitude of TCE-induced hepatoma increases over control is similar to the 3-fold and 2-fold
difference in the magnitude of exposure concentrations between these doses. Therefore the
increase in TCE-induced liver tumors would roughly proportional to the magnitude of exposure
dose. This result would be similar to the result for the concordance of the increases in liver
weight and exposure concentration observed 28-42 day exposures to TCE (see Section 2.4.2)
using oral data from B6C3F1 and Swiss mice, and inhalation data from NMRI mice. The
available inhalation data for TCE induced liver weight dose-response is from one study in a
strain derived from Swiss mice (Kjellstrand et al., 1983b) and was conducted in male and female
mice with comparable doses of 75 ppm and 300 ppm TCE. However, male mice of this strain
exhibited decreased body weight at the 300 ppm level which can affect %liver/body weight
increases. The magnitude of TCE-induced increases in liver weight between the 75 ppm and
300 ppm exposures were ~ 1.80-fold for males (1.50 vs. 1.90-fold of control liver weights) and
4.2-fold for females (1.27-fold vs. 2.14-fold of control liver weight) in this strain. Female mice
were examined in one study each of Swiss and B6C3F1 mice by Maltoni et al. (1988). Both the
Swiss and B6C3F1 studies reported increases in incidences of hepatomas over controls only at
the 600 ppm TCE level in female mice indicating less of a response than males. Similarly, the
Kjellstrand et al (1983b) data also showed less of a response in females compared to males in
terms TCE induction of liver weight at the 37 to 150 ppm range of exposure in NMRI strain.
While the data for TCE dose-response of liver tumor induction is very limited, it is suggestive of
a correlation of TCE-induced increases in liver weight correlating liver tumor induction with a
pattern that is dissimilar to that of TCA.
Of those experiments conducted at ~ 1000 mg/kg/day gavage dose of TCE in male
B6C3F1 mice for at least 79 weeks (Bull et al., 2002; NCI, 1976; Anna et al., 1994, NTP, 1990)
the control values were conducted in varying numbers of animals (some as low as n=15, i.e.,
Bull et al., 2002) and with varying results. The incidence of hepatocellular carcinomas ranged
from 1.2% to 16.7% (NCI, 1976; Anna et al., 1994, NTP, 1990) and the incidence of adenomas
ranged from 1.2% to 14.6% (Anna et al., 1994, NTP, 1990) in control B6C3F1 mice. After ~
1000 mg/kg/day TCE treatment, the incidence of carcinomas ranged from 19.4% to 62% (Bull
et al., 2002; NCI, 1976; Anna et al., 1994, NTP, 1990) with 3 of the studies (NCI, 1976; Anna et
al., 1994, NTP, 1990) reporting a range of incidences between 42.8% to 62.0%). The incidence
of adenomas ranged from 28% to 66.7% (Bull et al., 2002; Anna et al., 1994; NTP, 1990).
These data are illustrative of the variability between experiments to determine the magnitude
and nature of the TCE response in the same gender (male), strain (B6C3F1), time of exposure
02/12/09
E-212
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(3/4 studies were for 76-79 weeks and 1 for 2 years duration), and roughly the same dose (800 -
1163 mg/kg/day TCE). Given, that the TCE-induced liver response, as measured by liver
weight increase, is highly correlated with total oxidative metabolism to a number of agents that
are hepatoactive agents and hepatocarcinogens, the variability in response from TCE exposure
would be expected to be greater than studies of exposure to a single metabolite such as TCA or
DCA.
Caldwell et al (2007b) have commented on the limitations of experimental paradigms
used to study liver tumor induction by TCE metabolites and show that 51-week exposure
duration has consistently produced a tumor response for these chemicals, but with greater lesion
incidence and multiplicity at 82 weeks. As reported by DeAngelo et al. (1999) and Pereira
(1996), full expression of tumor induction in the mouse does not occur until 78 to 100 weeks of
DCA or TCA exposure, especially at lower concentrations. Thus, use of abbreviated exposure
durations and concurrently high exposure concentrations limits the ability of such experiments
to detect a treatment-related effect with the occurrence of additional toxicity not necessarily
associated with tumor-induction. Caldwell et al (2007b) present a table that shows that the
differences in the ability of the studies to detect treatment-related effects could also be attributed
to a varying and low number of animals in some exposure groups and that because of the low
numbers of animals tested at higher exposures, the power to detect a statistically significant
change is very low and in fact for many of the endpoints is considerably less than "50%
chance." The table below from Caldwell et al. (2008b) illustrates the importance of
experimental design and the limitations in many of the studies in the TCE metabolite database.
02/12/09
E-213
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Table II. Power calculations* for experimental design described in text, using Pereira et al. as an example.
Exposure concentrationf in
female B6C3F1 mice
(Pereira, 1996 and Pereira
and Phelps, 1996)
Number
of
animals
Power calculation for
Foci
Power calculation for
Adenomas
Power calculation for
Carcinomas
20.0 mmol/L NaCl (control)
(82 wk)
90
Null hypothesis
Null hypothesis
Null hypothesis
2.58 g/L DCA
(82 wk)
19
0.03
0.03
0.13
0.86 g/L DCA
(82 wk)
28
0.74
0.20
0.91
0.26 g/L DCA
(82 wk)
50
0.99
0.98
-
3.27 g/L TCA
(82 wk)
18
0.15
0.09
0.14
1.10 g/L TCA
(82 wk)
27
0.60
0.64
0.3
0.33 g/L TCA
(82 wk)
53
0.93
0.91
-
02/12/09
E-214
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1 "[Conversion of mmol/L to g/L from the original reports of Pereira (1996) and Pereira and Phelps (1996) is as follows:
2 20.0 mmol/L DCA = 2.58 g/L, 6.67 mmol/L DCA = 0.86 g/L, 2.0 mmol/L = 0.26 g/L, 20.0 mmol/L TCA = 3.27 g/L,
3 6.67 mmol/L TCA = 1.10 g/L, 2.0 mmol/L TCA = 0.33 g/L.
4
5 *The power calculations represent the probability of rejecting the null hypothesis when in fact the alternate hypothesis
6 is true for tumor multiplicity (i.e., the total # of lesions/# animals). The higher the power number calculated, the more
7 confidence we have in the null hypothesis. Assumptions made included: normal distribution for the fraction of tumors
8 reported, null hypothesis represents what we expected the control tumor fraction to be, the probability of a Type I error
9 was set to 0.05, and the alternate hypothesis was set to four times the null hypothesis value.
02/12/09
E-215
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Bull et al. (1990) examined male and female B6C3F1 mice (age 37 days) exposed from
15 to 52 weeks to neutralized DCA and TCA (1 g/1 or 2 g/1) but tumor data was not suitable for
dose response. They reported effects of DCA and TCA exposure on liver weight and %
liver/body changes that gave a pattern of hepatomegaly generally consistent with short-term
exposure studies. Only 10 female mice were examined at 52 weeks but the female mice were
reported to be as responsive as males at the exposure concentration tested. After 37 weeks of
treatment and then a cessation of exposure for 15 weeks, liver weights % liver/body weight were
reported to be elevated over controls which Bull et al. (1990) partially attribute the remaining
increases in liver weight to the continued presence of hyperplastic nodules in the liver.
Microscopically, livers treated with DCA were reported to have multi-focal areas of necrosis and
frequent infiltration of lymphocytes on the surface and an interior of the liver. For TCA-treated
mice, similar necrotic lesions were reported but at such a low frequency that they were similar to
controls. Marked cytomegaly was reported from exposure to either 1 or 2 g/1 DCA throughout
the liver. Cell size was reported to be increased from TCA and DCA treatment with DCA
producing the greatest change. The 2 g/1 TCA exposures were observed to have increased
accumulations of lipofuscin but no quantitative analysis was done. Photographs of light
microscopic sections, that were supposed to be representative of DCA and TCA treated livers at
2g/l, showed such great hepatocellular hypertrophy from DCA treatment that sinusoids were
obscured. Such a degree of cytomegaly could have resulted in reduction of blood flow and
contributed to focal necrosis observed at this level of exposure.
As discussed in Sections 3.2 and 3.4.2.1, glycogen accumulation has been described to be
present in foci in both humans and animals as a result from exposure to a wide variety of
carcinogenic agents and predisposing conditions in animals and humans. Bull et al (1990)
reported that glycogen deposition was uniformly increased from 2 g/1 DCA exposure with
photographs of TCA exposure showing slightly less glycogen staining than controls. However,
the abstract and statements in the paper suggest that there was increased PAS positive material
from TCA treatment that has caused confusion in the literature in this regard. Kato-Weinstein et
al. (2001) reported that in male B6C3F1 mice exposed to DCA and TCA, the DCA treatment
increased glycogen and TCA decreased glycogen content of the liver by using both chemical
measurement of glycogen in liver homogenates and by using ethanol-fixed sections stained with
PAS, a procedure designed to minimize glycogen loss. Kato-Weinstein et al. (2001) reported
that glycogen rich and poor cells were scattered without zonal distribution in male B6C3F1 mice
exposed to 2 g/1 DCA for 8 weeks. For TCA treatments they reported centrilobular decreases in
glycogen and ~ 25% decreases in whole liver by 3 g/1 TCA. Kato-Weinstein et al. (2001)
reported whole liver glycogen to be increased ~1.50-fold of control (90 vs. 60 mg glycogen/g
liver) by 2 g/1 DCA after 8 weeks exposure male B6C3F1 mice with a maximal level of glycogen
accumulation occurring after 4 weeks of DCA exposure. Pereira et al. (2004) reported that after
8 weeks of exposure to 3.2 g/1 DCA liver glycogen content was 2.20-fold of control levels (155.7
vs. 52.4. mg glycogen/g liver) in female B6C3F1 mice. Thus, the baseline level of glycogen
content reported by (~ 60 mg/g) and the increase in glycogen after DCA exposure was consistent
between Kato-Weinstein et al. (2001) and Pereira et al. (2004). However, the increase in liver
weight reported by Kato-Weinstein et al (2001) of 1.60-fold of control % liver/body weight
02/12/09
E-216
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
cannot be accounted for by the 1.50-fold of control glycogen content. Glycogen content only
accounts for 5% of liver mass so that 50% increase in glycogen cannot account for the 60%
increase liver mass induced by 2 g/1 DC A exposure for 8 weeks reported by Kato-Weinstein
(2001). Thus, DCA-induced increases in liver weight are occurring from other processes as well.
Carter et al. (2003) and DeAngelo et al. (1999) reported increased glycogen after DCA treatment
at much lower doses after longer periods of exposure (lOOweeks). Carter reported increased
glycogen at 0.5 g/1 DCA and DeAngelo et al (1999) at 0.03 g/1 DCA in mice. However there is
no quantitation of that increase.
The issues involving identification of MO A through tumor phenotype analysis are
discussed in detail below for the more general case of liver cancer as well as for specific
hypothesized MOAs (see Sections 3.1.4., 3.1.8., 3.2.1., and 3.4.1.5). For TCE and its
metabolites, c-Jun staining, H-rats mutation, tincture, heterogeneity in dysplacity have been used
to describe and differentiate liver tumors in the mouse.
Bull et al. (2002) reported 1000 mg/kg TCE administered via gavage daily for 79 weeks
in male B6C3F1 mice to produce liver tumors and also reported deaths by gavage error (6 out of
40 animals). The limitations of the experiment are discussed in Caldwell et al (2007b)
Specifically, for the DCA and TCA exposed animals, the experiment was limited by low
statistical power, a relatively short duration of exposure, and uncertainty in reports of lesion
prevalence and multiplicity due to inappropriate lesions grouping (i.e., grouping of hyperplastic
nodules, adenomas, and carcinomas together as "tumors"), and incomplete histopatholology
determinations (i.e., random selection of gross lesions for histopathology examination). For the
TCE results, a high prevalence (23/36 B6C3F1 male mice) of adenomas and hepatocellular
carcinoma (7/36) was reported. For determinations of immunoreactivity to c-Jun, as a marker of
differences in "tumor" phenotype, Bull et al. (2002) included all lesions in most of their
treatment groups, decreasing the uncertainty of his findings. However, for immunoreactivity
results hyperplastic nodules, adenomas, and carcinomas were grouped and thus, changes in c-Jun
expression between the differing types of lesions were not determined. Bull et al. (2002)
reported lesion reactivity to c-Jun antibody to be dependent on the proportion of the DCA and
TCA administered after 52 weeks of exposure. Given alone, DCA was reported to produce
lesions in mouse liver for which approximately half displayed a diffuse immunoreactivity to a c-
Jun antibody, half did not, and none exhibited a mixture of the two. After TCA exposure alone,
no lesions were reported to be stained with this antibody. When given in various combinations,
DCA and TCA co-exposure induced a few lesions that were only c-Jun+, many that were only c-
Jun-, and a number with a mixed phenotype whose frequency increased with the dose of DCA.
For TCE exposure of 79 weeks, TCE-induced lesions were reported to also have a mixture of
phenotypes (42% c-Jun+, 34% c-Jun-, and 24% mixed) and to be most consistent with those
resulting from DCA and TCA co-exposure but not either metabolite alone.
Stauber and Bull (1997) exposed male B6C3F1 mice (7 weeks old at the start of
treatment) to 2.0 g/1 neutralized DCA or TCA in drinking water for 38 or 50 weeks, respectively
and then exposed (n = 12) to 0, 0.02, 0.1, 0.5, 1.0, 2.0 g/L DCA or TCA for an additional 2
02/12/09
E-217
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
weeks. Foci and tumors were combined in reported results as "lesions" and prevalence rates
were not reported. The DCA-induced larger "lesions" were reported to be more "uniformly
reactive to c-Jun and c-Fos but many nuclei within the lesions displaying little reactivity to c-Jun.
Stauber and Bull (1997) stated that while most DCA-induced "lesions" were homogeneously
immunoreactive to c-Jun and C-Fos (28/41 lesions), the rest were stained heterogeneously. For
TCA-induced lesions, the authors reported no difference in staining between "lesions" and
normal hepatocytes in TCA-treated animals. These results are slightly different that those
reported by Bull et al. (2002) for DC A, who report c-Jun positive and negative foci in DCA-
induced liver tumors but no mixed lesions. Because "lesions" comprised of foci and tumors,
different stages of progression reported in these results. The duration of exposures also differed
between DCA and TCA treatment groups that can affect phenotype. The shorter duration of
exposure can also prevent full expression of the tumor response.
Stauber et al (1998) presented a comparison of in vitro results with "tumors" from
Stauber and Bull (1997) and note that 97.5% of DCA-induced "tumors" were c-Jun + while none
of the TCA-induced "tumors" were c-Jun +. However, the concentrations used to give tumors in
vivo for comparison with in vitro results were not reported. This appears to differ from the
heterogeneity of result for c-Jun staining reported by Bull et al (2002) and Stauber and Bull
(1997). There was no comparison of c-Jun phenotype for spontaneous tumors with the authors
stating that because of such short time, no control tumors results were given. However, the
results of Bull et al (2002) and Stauber and Bull (1997), do show TCA-induced lesions to be
uniformly c-Jun negative and thus the phenotypic marker was able to show that TCE-induced
tumors were more like those induced by DCA than TCA.
The premise that DCA induced c-Jun positive lesions and TCA-induced c-Jun negative
lesions in mouse liver was used as the rationale to study induction of "transformed" hepatocytes
by DCA and TCE treatment in vitro. Stauber et al. (1998) isolated primary hepatocytes from 5-8
week old male B6C3F1 mice (n=3) and subsequently cultured them in the presence of DCA or
TCA. In a separate experiment 0.5 g/1 DCA was given to mice as pretreatment for 2 weeks prior
to isolation. The authors assumed that the anchorage-independent growth of these hepatocytes
was an indication of an "initiated cell." DCA and TCA solutions were neutralized before use.
After 10 days in culture with DCA or TCA (0, 0.2, 0.5 and 2.0 mM), concentrations of 0.5 mM
or more DCA and TCA both induced an increase in the number of colonies that was statistically
significant, increased with dose with DCA, and slightly greater for DCA. In a time course
experiment the number of colonies from DCA treatment in vitro peaked by 10 days and did not
change through days 15- 25 at the highest dose and, at lower concentrations of DCA, increased
time in culture induced similar peak levels of colony formation by days 20-25 as that reached by
10 days at the higher dose. Therefore, the number of colonies formed was independent of dose if
the cells were treated long enough in vitro. However, not only did treatment with DCA or TCA
induce anchorage independent growth but untreated hepatocytes also formed larger numbers of
colonies with time, although at a lower rate than those treated with DCA. The level reached by
untreated cells in tissue culture at 20 days was similar to the level induced by 10 days of
exposure to 0.5 mM DCA. The time course of TCA exposure was not tested to see if it had a
02/12/09
E-218
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
similar effect with time as did DC A. The colonies observed at 10 days were tested for c-Jun
expression with the authors noting that "colonies promoted by DCA were primarily c-Jun
positive in contrast to TCA promoted colonies that were predominantly c-Jun negative." Of the
colonies that arose spontaneously from tissue culture conditions, 10/13 (76.9%) were reported to
be c-Jun +, those treated with DCA 28/34 (82.3%) were c-Jun +, and those treated with TCA
5/22 (22.7%>) were c-Jun +. Thus, these data show heterogeneity in cell in colonies but with
more were c-Jun + colonies occurring by tissue culture conditions alone and in the presence of
DCA, rather than in the presence of TCA. The authors reported that with time (24, 48, 72, and
96 hours) of culture conditioning the number of c-Jun+ colonies was increased in untreated
controls. The authors reported that DCA treatment delayed the increase in c-Jun+ expression
induced by tissue culture conditions alone in untreated controls while TCA treatment was
reported to not affect the increasing c-Jun+ expression that increased with time in tissue culture.
This results seems paradoxical given that DCA induced a higher number of colonies at 10 days
of tissue culture than TCA and that most of the colonies were c-Jun positive. The number of
colonies was greater for pretreatment with DCA, but the magnitude of difference over the control
level was the same after DCA treatment in vitro without and without pretreatment. As to the
relationship of c-Jun staining and peroxisome proliferators as a class, as pointed out by Caldwell
and Keshava (2006), although Bull et al. (2004) have suggested that the negative expression of c-
jun in TCA-induced tumors may be consistent with a characteristic phenotype shown in general
by peroxisome proliferators as a class, there is no supporting evidence of this.
An approach to determine the potential MO As of DCA and TCA through examination of
the types of tumors each "induced" or "selected" was to examine H-ras activation (Ferreira-
Gonzalez et al., 1995; Anna et al., 1994; Bull et al., 2002; Nelson et al., 1990. This approach has
also been used to try to establish an H-ras activation pattern for "genotoxic" and "non-genotoxic"
liver carcinogens compounds and to make inferences concerning peroxisome proliferator-
induced liver tumors. However, as noted by Stanley et al. (1994), the genetic background of the
mice used and the dose of carcinogen may affect the number of activated H-ras containing
tumors which develop. In addition, the stage of progression of "lesions" (i.e., foci vs. adenomas
vs. carcinomas) also has been linked the observance of H-ras mutations. Fox et al. (1990) note
that tumors induced by phenobarbital (0.05%> drinking H20, 1 yr), chloroform (200 mg/kg corn
oil gavage, 2 times weekly for 1 year) or Ciprofibrate (0.0125%) diet, 2 years) had a much lower
frequency of H-ras gene activation than those that arose spontaneously (2-year bioassays of
control animals) or induced with the "genotoxic" carcinogen benzidine-2 HC1 (120 ppm,
drinking H20, 1 yr) in mice. In that study, the term "tumor" was not specifically defined but a
correlation between the incidence of H-ras gene activation and development of either a
hepatocellular adenoma or hepatocellular carcinoma was reported to be made with no
statistically significant difference between the frequency of H-ras gene activation in the
hepatocellular adenomas and carcinomas. Histopathological examination of the spontaneous
tumors, tumors induced with benzidine-2HCL, Phenobarbital, and chloroform was not reported
to reveal any significant changes in morphology or staining characteristics. Spontaneous tumors
were reported to have 64%> point mutation in codon 61 (n= 50 tumors examined) with a similar
response for Benzidine of 59%> (n= 22 tumors examined), whereas for Phenobarbital the
02/12/09
E-219
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mutation rate was 7% (n= 15 tumors examined), chloroform 21% (n= 24 tumors examined) and
Ciprofibrate 21% (n= 39 tumors examined). The Ciprofibrate-induced tumors were reported to
be more eosinophilic as were the surrounding normal hepatocytes. Hegi et al. (1993) tested
Ciprofibrate-induced tumors in the NIH3T3 co-transfection-nude mouse turnorigenicity assay,
which the authors state is capable of detecting a variety of activated proto-oncogenes. The
tumors examined (Ciprofibrate-induced or spontaneously arising) were taken from the Fox et al.
study (1990), screened previously, and found to be negative for H-ras activation. With the
limited number of samples examined, Hegi et al concluded that ras proto-oncogene activation or
activation of other proto-oncogenes using the nude mouse assay were not frequent events in
Ciprofibrate-induced tumors and that spontaneous tumors were not promoted with it. Using the
more sensitive methods, the H-ras activation rate was reported to be raised from 21 to 31% for
Ciprofibrate-induced tumors and from 64 to 66% for spontaneous tumors. Stanley et al. (1994)
studied the effect of Methylclofenapate (MCP) ( 25 mg/kg for up to 2 years), a peroxisome
proliferator, in B6C3F1 (relatively sensitive) and C57BL/10J (relatively resistant) mice for H-ras
codon 61 point mutations in MCP-induced liver tumors (hepatocellular adenomas and
carcinomas). In the B6C3F1 mice the number of tumors with codon 61 mutations was 11/46 and
for C57BL/10J mice 4/31. Unlike the findings of Fox et al (1990), Stanley et al (1994) reported
an increase in the frequency of mutation in carcinomas, which was reported to be twice that of
adenomas in both strains of mice, indicating that stage of progression was related to the number
of mutations in those tumors, although most tumors induced by MCP did not have this mutation.
In terms of liver tumor phenotype, Anna et al. (1994) reported that the H-ras codon 61
mutation frequency was not statistically different in liver tumors from DCA and TCE-treated
mice from a highly variable number of tumors examined. In regard to mutation spectra in H-ras
oncogenes in control or spontaneous tumors, the patterns were slightly different but mostly
similar to that of DCA-induced tumors (0.5% in drinking water). From their concurrent controls
they reported that H-ras codon 61 mutations in 17% (n=6) of adenomas and 100% (n=5) of
carcinomas. For historical controls (published and unpublished) they reported mutations in 73%
(n=33) of adenomas and mutations in 70% (n=30) of carcinomas. For tumors from TCE treated
animals they reported mutations in 35% (n=40) of adenomas and 69% (n=36) of carcinomas,
while for DCA treated animals they reported mutations in 54% (n=24) of adenomas and in 68%
(n=40) of carcinomas. Anna et al 1994) reported more mutations in TCE-induced carcinomas
than adenomas.
The study of Ferreira-Gonzalez (1995) in male B6C3 F1 mice has the advantage of
comparison of tumor phenotype at the same stage of progression (hepatocellular carcinoma), for
allowance of the full expression of a tumor response (i.e. 104 weeks), and an adequate number of
spontaneous control lesions for comparison with DCA or TCA treatments. However, tumor
phenotype at an endstage of tumor progression reflects of tumor progression and not earlier
stages of the disease process. In spontaneous liver carcinomas, 58% were reported to show
mutations in H-61 as compared with 50% of tumor from 3.5 g/1 DCA-treated mice and 45% of
tumors from 4.5.g/1 TCA-treated mice. Thus, there was a heterogeneous response for this
phenotypic marker for the spontaneous, DCA-, and TCA- treatment induced hepatocellular
02/12/09
E-220
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
carcinomas and not a pattern of reduced H-ras mutation reported for a number of peroxisome
proliferators. A number of peroxisome proliferators have been reported to have a much smaller
mutation frequency that spontaneous tumors [e.g,. 13-24% H-ras codon 61 mutations after
Methylclofenopate depending on mouser strain, Stanley et al. (1994): 21 to 31% for
Ciprofibrate-induced tumors and from 64 to 66% for spontaneous tumors, Fox et al. (1990 ) and
Hegi et al (1993)].
Bull (2000) suggested that "the report by Anna et al (1994) indicated that TCE-induced
tumors possessed a different mutation spectra in codon 61 of the H-ras oncogene than those
observed in spontaneous tumors of control mice." Bull (2000) stated that "results of this type
have been interpreted as suggesting that a chemical is acting by a mutagenic mechanism" but
went on to suggest that it is not possible to a priori rule out a role for selection in this process and
that differences in mutation frequency and spectra in this gene provide some insight into the
relative contribution of different metabolites to TCE-induced liver tumors. Bull (2000) noted
that data from Anna et al. (1994), Ferreira-Gonzalez et al (1995), and Maronpot et al (1995)
indicated that mutation frequency in DCA-induced tumors did not differ significantly from that
observed in spontaneous tumors. Bull (2000) also noted that the mutation spectra found in DCA-
induced tumors has a striking similarity to that observed in TCE-induced tumors, and DCA-
induced tumors were significantly different than that of TCA-induced liver tumors.
Bull et al. (2002) reported that mutation frequency spectra for the H-ras codon 61 in
mouse liver "tumors" induced by TCE (n= 37 tumors examined) were reported to be
significantly different than that for TCA (n= 41 tumors examined), with DCA-treated mice
tumors giving an intermediate result (n= 64 tumors examined). In this experiment, TCA-induced
"tumors" were reported to have more mutations in codon 61 (44%) than those from TCE (21%)
and DCA (33%). This frequency of mutation in the H-ras codon 61 for TCA is the opposite
pattern as that observed for a number of peroxisome proliferators in which the number of
mutations at H-ras 61 in tumors has been reported to be much lower than spontaneously arising
tumors (see Section 3.4.1.5.). Bull et al. (2002) noted that the mutation frequency for all TCE
,TCA or DCA tumors was lower in this experiment than for spontaneous tumors reported in
other studies (they had too few spontaneous tumors to analyze in this study), but that this study
utilized lower doses and was of shorter duration than that of Ferreira-Gonzalez (1995). These
are additional concerns in addition to the effects of lesion grouping in which a lower stage of
progression is group with more advanced stages. In a limited subset of tumors that were both
sequenced and characterized histologically, only 8 of 34 (24%) TCE-induced adenomas but 9/15
(60%) of TCE-induced carcinomas were reported to have mutated H-ras at codon 61, which the
authors suggest is evidence that this mutation is a late event.
Thus, in terms of H-ras mutation, the phenotype of TCE-induced tumors appears to be
more like DCA-induced tumors (which are consistent with spontaneous tumors), or those
resulting from a co-exposure to both DCA and TCA (Bull et al., 2002), than from those induced
by TCA. As noted above, Bull et al. (2002) reported the mutation frequency spectra for the H-
ras codon 61 in mouse liver tumors induced by TCE to be significantly different than that for
02/12/09
E-221
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
TCA, with DCA-treated mice tumors giving an intermediate result and for TCA-induced tumors
to have a H-ras profile that is the opposite than those of a number of other peroxisome
proliferators. More importantly, these data suggest that using measures, other than dysplasticity
and tincture, mouse liver tumors induced by TCE are heterogeneous in phenotype.
With regard to tincture, Stauber and Bull (1997) reported the for male B6C3F1 mice,
DCA-induced "lesions" contained a number of smaller lesions that were heterogeneous and more
eosinophilic with larger "lesions" tending to less numerous and more basophilic. For TCA
results using this paradigm, the "lesions" were reported to be less numerous, more basophilic,
and larger than those induced by DCA. Carter et al. (2003) used tissues from the DeAngelo et al.
(1999) and examined the heterogeneity of the DCA-induced lesions and the type and phenotype
of preneoplastic and neoplastic lesions pooled across all time points. Carter et al. (2003)
examined the phenotype of liver tumors induced by DCA in male B6C3 F1 mice and the shape
of the dose-response curve for insight into its MOA. They reported a dose-response of
histopathologic changes (all classes of premalignant lesions and carcinomas) occurring in the
livers of mice from 0.05-3.5 g/L DCA for 26-100 weeks and suggest foci and adenomas
demonstrated neoplastic progression with time at lower doses than observed DCA genotoxicity.
Preneoplastic lesions were identified as eosinophilic, basophilic and/or clear cell (grouped with
clear cell and mixed cell) and dysplastic. Altered foci were 50% eosinophilic with about 30%
basophilic. As foci became larger and evolved into carcinomas they became increasingly
basophilic. The pattern held true through out the exposure range. There was also a dose and
length of exposure related increase in atypical nuclei in "non-involved" liver. Glycogen
deposition was also reported to be dose-dependent with periportal accumulation at the 0.5 g/1
exposure level. Carter et al. (2003) suggested that size and evolution into a more malignant state
are associated with increasing basophilia, a conclusion consistent with those of Bannasch (1996)
and that there a greater periportal location of lesions suggestive as the location from which they
arose. Consistent with the results of DeAngelo et al (1999), Carter et al. (2003) reported that
DCA (0.05 - 3.5 g/1) increased the number of lesions per animal relative to animals receiving
distilled water, shortened the time to development of all classes of hepatic lesions, and that the
phenotype of the lesions were similar to those spontaneously arising in controls. Along with
basophilic and eosinophilic lesions or foci, Carter et al (2003) concluded that DCA-induced
tumors also arose from isolated, highly dysplastic hepatocytes in male B6C3F1 mice chronically
exposed to DCA suggesting another direct neoplastic conversion pathway other than through
eosinophilic or basophilic foci.
Rather than male B6C3F1 mice, Pereira (1996) studied the dose-response relationship for
the carcinogenic activity of DCA and TCA and characterized their lesions (foci, adenomas and
carcinomas) by tincture in females (the generally less sensitive gender). Like the studies of TCE
by Maltoni et al (1986), female mice were also reported to have increased liver tumors after TCA
and DCA exposures. Pereira (1996) pool lesions were pooled for phenotype analysis so the
affect of duration of exposure could not be determined nor adenomas separated from carcinomas
for "tumors." However, as the concentration of DCA was decreased the number of foci was
reported by Pereira (1996) to be decreased but the phenotype of the foci to go from primarily
02/12/09 E-222
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
eosinophilic foci (i.e. ~ 95% eosinophilic at 2.58 g/1 DCA) to basophilic foci (~ 57%
eosinophilic at 0.26 g/1). For TCA the number of foci was reported to ~ 40 basophilic and ~ 60
eosinophilic regardless of dose. Spontaneously occurring foci were more basophilic by a ratio of
7/3. Pereira (1996) described the foci of altered hepatocytes and tumors induced by DCA in
female B6C3F1 mice to be eosinophilic at higher exposure levels but at lower or intermittent
exposures to be half eosinophilic and half basophilic. Regardless of exposure level, half of the
TCA-induced foci were reported to be half eosinophilic and half basophilic with tumors 75%
basophilic. In control female mice, the limited numbers of lesions were mostly basophilic, with
most of the rest being eosinophilic with the exception of a few mixed tumors. The limitations of
descriptions tincture and especially for inferences regarding peroxisome proliferator from the
description of "basophilia" is discussed in Section 3.4.1.5.
The results appear to differ between male and female B6C3F1 mice in regard to tincture
for DCA and TCA at differing doses. What is apparent is that the tincture of the lesions is
dependent on the stage of tumor progression, agent (DCA or TCA), gender, and dose. Also what
is apparent from these studies is the both DCA and TCA are heterogeneous in their tinctoral
characteristics as well as phenotypic markers such as mutation spectra or expression of c-Jun.
The descriptions of tumors in mice reported by the NCI, NTP, and Maltoni et al studies
are also consistent with phenotipic heterogeneity as well as consistency with spontaneous tumor
morphology (see Section 3.4.1.5). As noted in section 3.1, hepatocellular carcinomas observed
in humans are also heterogeneous. For mice, Maltoni et al (1986) described malignant tumors of
hepatic cells to be of different sub-histotypes, and of various degrees of malignancy and were
reported to be unique or multiple, and have different sizes (usually detected grossly at necropsy)
from TCE exposure. In regard to phenotype tumors were described as usual type observed in
Swiss and B6C3F1 mice, as well as in other mouse strains, either untreated or treated with
hepatocarcinogens and to frequently have medullary (solid), trabecular, and pleomorphic
(usually anaplastic) patterns. For the NCI (1976) study, the mouse liver tumors were described
in detail and to be heterogeneous "as described in the literature" and similar in appearance to
tumors generated by carbon tetrachloride. The description of liver tumors in this study and
tendency to metastasize to the lung are similar to descriptions provided by Maltoni et al (1986)
for TCE-induced liver tumors in mice via inhalation exposure. The NTP (1990) study reported
TCE exposure to be associated with increased incidence of hepatocellular carcinoma (tumors
with markedly abnormal cytology and architecture) in male and female mice. Hepatocellular
adenomas were described as circumscribed areas of distinctive hepatic parenchymal cells with a
perimeter of normal appearing parenchyma in which there were areas that appeared to be
undergoing compression from expansion of the tumor. Mitotic figures were sparse or absent but
the tumors lacked typical lobular organization. Hepatocellular carcinomas were reported to have
markedly abnormal cytology and architecture with abnormalities in cytology cited as including
increased cell size, decreased cell size, cytoplasmic eosinophilia, cytoplasmic basophilia,
cytoplasmic vacuolization, cytoplasmic hyaline bodies and variations in nuclear appearance.
Furthermore, in many instance several or all of the abnormalities were reported to be present in
different areas of the tumor and variations in architecture with some of the hepatocellular
02/12/09 E-223
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
carcinomas having areas of trabecular organization. Mitosis was variable in amount and
location. Therefore the phenotype of tumors reported from TCE exposure was heterogenous in
appearance between and within tumors from all 3 of these studies.
Caldwell and Keshava (2006) report "that Bannasch (2001) and Bannasch et al. (2001)
describe the early phenotypes of preneoplastic foci induced by many oncogenic agents (DNA-
reactive chemicals, radiation, viruses, transgenic oncogenes and local hyperinsulinism) as
insulinomimetic. These foci and tumors have been described by tincture as eosinophilic and
basophilic and to be heterogeneous. The tumors derived from them after TCE exposure are
consistent with the description for the main tumor lines of development described by Bannasch
et al (2001) (see Section 3.4.1.5). Thus, the response of liver to DCA (glycogenosis with
emergence of glycogen poor tumors) is similar to the progression of preneoplastic foci to tumors
induced from a variety of agents and conditions associated with increased cancer risk."
Furthermore Caldwell and Keshava (2006) note that Bull et al. (2002) report expression of IR to
be elevated in tumors of control mice or mice treated with TCE, TCA and DCA but not in non-
tumor areas suggesting that this effect is not specific to DCA.
There is a body of literature that has focused on the effects of TCE and its metabolites
after rats or mice have been exposed to "mutagenic" agents to "initiate" hepatocarcinogenesis
and this is discussed in Section 4.2 below. TCE and its metabolites were reported to affect tumor
incidence, multiplicity, and phenotype when given to mice as a co-exposure with a variety of
"initiating" agents and with other carcinogens. Pereira and Phelps (1996) reported that MNU
alone induced basophilic foci and adenomas. MNU and low concentrations of DCA or TCA in
female mice were reported to induce heterogeneous for foci and tumor with a higher
concentration of DCA inducing more eosinophilic and a higher concentration of TCA inducing
more tumors that were basophilic. Pereira et al (2001) reported that not only dose, but gender
also affected phenotype in mice that had already been exposed to MNU and were then exposed
to DCA. As for other phenotypic markers, Lantendresse and Pereira (1997) reported that
exposure to MNU and TCA or DCA induced tumors that had some commonalities, were
heterogeneous, but for female mice were overall different between DCA and TCA as co-
exposures with MNU.
Stop experiments which attempt to ascertain the whether progression differences exist
between TCA and DCA have used higher concentrations at much lower durations of exposure.
A question arises as to whether the differences in results between those animals in which
treatment was suspended in comparison to those in which had not had been conducted so that full
expression of response had not been allowed rather than "progression" as well as the effects of
using large doses. After 37 weeks of treatment and then a cessation of exposure for 15 weeks
Bull et al. (1990) reported that after 15 weeks of cessation of exposure, liver weight and %
liver/body weight were reported to still be statistically significantly elevated after DCA or TCA
treatment. The authors partially attribute the remaining increases in liver weight to the continued
presence of hyperplastic nodules in the liver. In terms of liver tumor induction, the authors
02/12/09
E-224
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
stated that "statistical analysis of tumor incidence employed a general linear model ANOVA
with contrasts for linearity and deviations from linearity to determine if results from groups in
which treatments were discontinued after 37 weeks were lower than would have been predicted
by the total dose consumed." The multiplicity of tumors observed in male mice exposed to DCA
or TCA at 37 weeks and then sacrificed at 52 weeks were reported by the authors to have a
response in animals that received DCA very close to that which would be predicted from the
total dose consumed by these animals. The response to TCA was reported by the authors to
deviate significantly (P= 0.022) from the linear model predicted by the total dose consumed.
Multiplicity of lesions per mouse and not incidence was used as the measure. Most importantly
the data used to predict the dose response for "lesions" used a different methodology at 52 weeks
than those at 37 weeks. Not only were not all animal's lesions examined, but foci, adenomas,
and carcinomas were combined into one measure. Therefore foci, of which a certain percentage
have been commonly shown to spontaneously regress with time, were included in the calculation
of total "lesions". Pereira and Phelps (1996) note that in MNU-treated mice that were then
treated with DCA, the yield of altered hepatocytes decreases as the tumor yields increase
between 31 and 51 weeks of exposure suggesting progression of foci to adenomas. Initiated and
non-initiated control mice were reported to also have fewer foci/mouse with time. Because of
differences in methodology and the lack of discernment between foci, adenomas, and carcinomas
for many of the mice exposed for 52 weeks, it is difficult to compare differences in composition
of the "lesions" after cessation of exposure in the Bull et al (1990) study. For TCA treatment the
number of animals examined for determination of which "lesions" were foci, adenomas, and
carcinomas was 11 out of the 19 mice with "lesions" at 52 weeks while all 4 mice with lesions
after 37 weeks of exposure and 15 weeks of cessation were examined. For DCA treatment the
number of animals examined was only 10 out of 23 mice with "lesions" at 52 weeks while all 7
mice with lesions after 37 weeks of exposure and 15 weeks of cessation were examined. Most
importantly, when lesions were examined microscopically then did not all turn out to be
preneoplastic or neoplastic. Two lesions appeared "to be histologically normal" and one
necrotic. Not only were a smaller number of animals examined for the cessation exposure than
continuous exposure but only the 2 g/1 exposure levels of DCA and TCA were studied for
cessation. The number of animals bearing "lesions" at 37 and then 15 week cessation weeks was
7/11 (64%) while the number of animals bearing lesions at 52 weeks was 23/24 (96%) after 2 g/1
DCA exposure. For TCA the number of animals bearing lesions at 37 weeks and then 15 weeks
cessation was 4/11 (35%) while the number of animals bearing lesions at 52 weeks was 19/24
(80%>). While suggesting that cessation of exposure diminished the number of "lesions,"
conclusions regarding the identity and progression of those lesion with continuous vs. non-
continuous DCA and TCA treatment are tenuous.
2.5. Studies of Chloral Hydrate
Given that total oxidative metabolism appears to be highly correlated with TCE-induced
increases in liver weight in the mouse rather than merely the presence of TCA, other metabolites
are of interest as potential agents mediating the effects observed for TCE. Recently Caldwell
and Keshava provided a synopsis of the results of more recent studies involving chloral hydrate
02/12/09
E-225
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(CH) (Caldwell and Keshava, 2006). A large fraction of TCE oxidative metabolism appears to
go through chloral hydrate, with subsequent metabolism to TCA and trichloroethanol (Chiu et
al., 2006b). Merdink et al. (2008) demonstrated that CH administered to humans can be
extremely variable and complex in its pharmacokinetic behaviour with a peak plasma
concentration of CH in plasma 40-50 times higher than observed at the same time interval for
other subjects. Studies of CH toxicity in rodents are consistent with the general presumption that
oxidative metabolites are important for TCE-induced liver tumors, but whether CH and its
metabolites are sufficient to explain all of TCE liver tumorigenesis remains unclear, particularly
because of uncertainties regarding how DC A may be formed (Chiu et al., 2006b). Studies of CH
may enable a comparison between toxicity of TCE and CH and may help elucidate its role in
TCE effects. As with other TCE metabolites, the majority of the studies have focused on the
mouse liver tumor response. For rats, while the limited data suggests that there is less of a
response than mice to CH, those studies are limited in power or reporting.
Daniel et al. (1992) exposed adult male B6C3F1 (C57B1/6JC male mice bred to
C3 Heb/Fej female mice) 28-day old mice to CH, 2-chloroacetaldehyde, or DC A in 2 different
phases (1 and II) with initial weights ranging from 9.4 to 13.6 g. The test compounds were
buffered and administered in drinking water for 30 and 60 weeks (n=5 for interim sacrifice), and
for 104 weeks (n=40). The concentration of CH was 1 g/1 and for DC A 0.5 g/1 and the estimated
doses of DC A were 85 mg/kg/day, 93 mg/kg/day and 166 mg/kg/day for the DC A group I, DC A
group II, and CH exposed group, respectively. Microscopic examination of tissues was
conducted for all tissues for five animals of the CH groups with liver, kidneys, testes, and spleen,
in addition to all gross lesions, reported to be examined microscopically in all of the 104-week
survivors. The initial body weight for drinking water controls was reported to be 12.99 ± 3.04 g
for group I (n= 23) and 10.48 ± 1.70 for group II (n =10). For DC A treated animals, initial
body weights were 13.44 ± 2.57 g for group I (n=23) and 9.65 ± 2.72 g for group II (n= 10). For
the CH treated group the initial body weights were reported to be 10.42 ± 2.49 g (n= 40). It is
not clear from the report what control group best matched, if any, the CH group. Thus, the mean
initial body weights of the groups as well as the number of animals varied considerably in each
group (i.e., ~ 40% difference in mean body weights at the beginning of the study). The number
of animals surviving till the termination of the experiment was 10, 10, 16, 8, and 24 for the
control group I, control group II, DC A group I, DC A group II, and CH groups, respectively. An
increase in absolute and relative liver weight vs. reported to be observed at 30 weeks for DC A
and CH groups and at 60 weeks for CH but data was not shown in the study. At 104 weeks, the
data for the surviving control groups were combined as was that for the 2 DC A treatment groups.
Of note was that for CH treated survivors (n=24) water consumption was significantly reduced in
comparison to controls. Absolute liver weight was reported to be 2.09 ± 0.6 g, 3.17 ± 1.3 g and
2.87 ± 1.1. g for control, DC A and CH treatment groups, respectively. The % liver to body
weight was reported to be similarly elevated (1 57-fold of control for DC A and 1.41-fold of
control for CH) at 104 weeks. At 104 weeks the treatment-related liver lesions in histological
sections were reported to be most prominently hepatocytomegaly and vacuolization in DCA-
treated animals. Cytomegaly was also reported to be in 5%, 92%, and 79 % of control, DC A and
CH treatment groups, respectively. Cytomegaly in CH treated mice was described as minimal
02/12/09 E-226
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and associated with an increased number of basophilic granules (rough endoplasmic reticulum).
Hepatocellular necrosis and chronic active inflammation were reported to be mildly increased in
both prevalence and severity in all treated groups. The histological findings, from interim
sacrifices (n=5), were considered by the authors to be unremarkable and were not reported.
Liver tumors were increased by DC A and CH treatment. The % incidence of liver carcinomas
and adenomas combined in the surviving animals was 15%, 75%, and 71% in control, DC A and
CH treated mice, respectively. In the CH treated group, the incidence of hepatocellular
carcinoma was 46%. The number of tumors/animals was also significantly increased with CH
treatment. Most importantly, morphologically the authors noted that there did not appear to be
any discernable differences in the visual appearance of the DC A- and CH-induced tumors.
George et al. (2000) exposed male B6C3F1 mice and male F344/N rats to CH in drinking
water for 2 years (up to 162.6 mg/kg/day). Target drinking water concentrations were 0, 0.05,
0.5, and 2 g/1 CH in rats and 0, 0.05, 0.5 and 1.0 g/1 CH in mice. Groups of animals (n=6/group)
were sacrificed at 13 (rats only), 26, 52 and 78 weeks following the initiation of dosing with
terminal sacrifices at week 104. A complete pathological examination was performed on 5 rats
and mice from the high-dose group, with examination primarily of gross lesions except for liver,
kidney, spleen and testes. BrdlJ incorporation was measured in the interim sacrifice groups in
rats and mice with PCO examined at 26 weeks in mice. In rats, the number of animals surviving
> 78 weeks and examined for hepatocellular proliferative lesions was 42, 44, 44, and 42 for the
control, 7.4, 37.4 and 163.6 mg/kg/day CH treatment groups, respectively. Only 32, 36, 35, and
32 animals were examined at the final sacrifice time. Only the lowest treatment group had
increased liver tumors which were marginally significantly increased by treatment. The % of
animals with hepatocellular adenomas and carcinomas was reported to be 2.4, 14.3, 2.3 and 6.8
% in male rats. In mice, preneoplastic foci and adenomas were reported to be increased in the
livers of all CH treatment groups (13.5-146.6 mg/kg/day) at 104 weeks. The incidences of
adenomas were reported to be statistically increased at all dose levels, the incidences of
carcinomas significantly increased at the highest dose, and time-to-tumor decreased in all CH-
treatment groups. The % incidence of hepatocellular adenomas was reported to be 21.4%,
43.5%, 5 1.3%, and 50% in control, 13.5, 65.0 and 146.6 mg/kg day treatment groups,
respectively. The % incidence of hepatocellular carcinomas was reported to be 54.8%, 54.3%,
59.0% and 84.4% in these same groups. The resulting % incidence of hepatocellular adenomas
and carcinomas was reported to be 64.3%, 78.3%, 79.5% and 90.6%. The number of mice
surviving > 78 weeks was reported to be 42, 46, 39, and 32 and the number surviving to final
sacrifice to be 34, 42, 3 1, and 25 for control, 13.5, 65.0 and 146.56 mg/kg/day, respectively. CH
exposure was reported to not alter serum chemistry, hepatocyte proliferation (i.e., DNA
synthesis), or hepatic PCO activity (an enzyme associated with PPARa agonism) in rats and
mice at any of the time periods monitored (all interim sacrifice periods for BrdU incorporation,
52 or 78 weeks for serum enzymes, and 26 weeks for PCO) with the exception of 0.58 g/L
chloral hydrate at 26 weeks slightly increasing hepatocyte labeling (~ 2-3 fold increase over
controls) in rats and mice but the % labeling still represented 3% or less of hepatocytes. With
regard to other carcinogenic endpoints only five animals were examined at the high dose, thereby
limiting the study's power to determine an effect. Control mice were reported to have a high
02/12/09
E-227
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
spontaneous carcinoma rate (54%), thereby limiting the ability to detect a treatment-related
response. No descriptions of the foci or tumor phenotype were given. However, of note is the
lack of induction of PCO response with CH at 26 weeks of administration in either rats or mice.
Leakey et al. (2003a) studied the effects of CH exposure (0, 25, 50, and 100 mg/kg, 5
days/week, 104-105 weeks via gavage) in male B6C3F1 mice with dietary control used to
manipulate body growth (n= 48 for 2 year study and n= 12 for the 15-month interim study).
Dietary control was reported to decrease background liver tumor rates (incidence of 15-20%)
and was reported to be associated with decreased variation in liver-to-body weight ratios, thereby
potentially increasing assay sensitivity. In dietary-controlled groups and groups fed ad libitum,
liver adenomas and carcinomas (combined) were reported to be increased with CH treatment.
With dietary restriction there was a more discernable CH tumor-response with overall tumor
incidence reduced, and time-to-tumor increased by dietary control in comparison to ad libitum
fed mice. Incidences of hepatocellular adenoma and carcinoma overall rates were reported to be
33%), 52%, 49%o, and 46%> for control, 25mg/kg, 50 mg/kg, and 100 mg/kg ad libitum-fed mice,
respectively. For dietary controlled mice the incidence rates were reported to be 22.9%, 22.9%,
29.2% and 37.5% for controls, 25mg/kg, 50 mg/kg, and 100 mg/kg CH, respectively. Body
weights were matched and carefully controlled in this study.
After 2 years of CH treatment the heart weights of ad libitum-fed male mice administered
100 mg/kg CH were reported to be significantly less and kidney weights of the 50 and 100
mg/kg less than vehicle controls. No other significant organ weight changes due to CH
treatment were reported to be observed in either diet group except for liver. The liver weights of
CH treated groups for by dietary groups were reported to be increased at 2 years and the absolute
liver weights of dosed groups to be generally increased at 15 months with %> liver/body weight
ratios increased in CH treated dietary-controlled mice at 15 months. There was 1 0-fold, 0.87-
fold, and 1 08-fold of control % liver/body weight for ad libitum fed mice exposed to 25, 50 and
100 mg/kg CH, respectively. For dietary controlled mice, there was 1 05-fold, 1 08-fold, and
1.11 -fold of control % liver/body weight for the same dose groups at 15 months. Thus, there was
no corresponding dose-response for % liver/body weight in the ad libitum-fed mice, which were
reported to show a much larger variation in liver-to-body- weight ratios (i.e., the standard
deviation and standard errors were 2- to 17-fold lower in dietary controlled groups than for ad
libitum-fed groups). Liver weight increases at 15-months did not correlate with 2-year tumor
incidences with this group. However for dietary controlled groups the increase in % liver/body
weights at 15 months were generally correlated with increases in liver tumors at 2 years. The
incidences of peripheral or focal fatty change were reported to be increased in all CH-treated
groups of ad libitum-fed mice at 15 months (~ half the animals showed these changes for all dose
groups, with no apparent dose-response). Of the enzymes associated with PPARa agonism (total
CYP, CYP2B isoform, CYP4A, or 1 auric acid P-hydroxylase activity), only CYP4A and lauric
acid P-hydroxylase activity were significantly increased at 15 months of exposure in the dietary -
restricted group administered 100 mg/kg CH with no other groups reported showing a
statistically significant increased response (n= 12/group). Although not statistically significant,
the 100 mg/kg CH exposure group of ad libitum-fed mice also had an increase in CYP4A and
02/12/09 E-228
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1 auric acid |3-hydroxylase activity. The authors reported that the increase in magnitude of
CYP4A and 1 auric acid P-hydroxylase activity at 100 mg/kg CH at 15 months in dietary
controlled mice correlated with the increase incidence of mice with tumors. However, there was
no correlation of tumor incidence and the increased enzyme activity associated with peroxisome
proliferation in the ad libitum-fed mice. No descriptions of liver pathology were given other
than incidence of mice with fatty liver changes. Hepatic trial ondi aldehyde concentration in ad
libitum fed and dietary controlled mice did not change with CH exposure at 15 months but the
dietary controlled groups were all ~ half that of the ad libitum-fed mice. Thus, while overall
increased tumors observed in the ad libitum diet correlated with increased malondialdehyde
concentration, there was no association between CH dose and malondialdehyde induction for
either diet.
Induction of peroxisome-associated enzyme activities was also reported for shorter times
of CH exposure. Seng et al. (2003) described CH toxicokinetics in mice at doses up to 1,000
mg/kg/day for 2 weeks with dietary control and caloric restriction slightly reducing acute
toxicity. Laurie acid P-hydroxylase and PCO activities were reported to be induced only at doses
> 100 mg/kg in all groups, with dietary-restricted mice showing the greatest induction.
Differences in serum levels of TCA, the major metabolite remaining 24 hr after dosing, were
reported not to correlate with hepatic 1 auric acid P-hydroxylase activities across groups.
Leuschner and Beuscher (1998) examined the carcinogenic effects of CH in male and
female Sprague-Dawley rats (69-79 g , 25-29 days old at initiation of the experiment)
administered 0, 15, 45, and 135 mg/kg CH in unbuffered drinking water 7 days/week (n=
50/group) for 124 weeks in males and 128 weeks in females. Two control groups were noted in
the methods section without explanation as to why they were conducted as two groups. The
mean survival for males was similar in treated and control groups with 20%, 24%, 20%, 24%,
and 20 % of control I, control II, 15, 45, and 135 mg/kg CH-treated groups, respectively,
surviving till the end of the study. For female rats, the % survival was 12%, 30%, 24%, 28%,
and 16% for of control I, control II, 15, 45, and 135 mg/kg CH-treated groups, respectively. The
authors report no substance-related influence on organ weights and no macroscopic evidence of
tumors or lesions in male or female rats treated with CH for 124 or 128 weeks. However, no
data is presented on the incidence of tumors using this paradigm, especially background rates.
The authors report a statistically significant increase in the incidence of hepatocellular
hypertrophy in male rats at the 135 mg/kg dose (14/50 animals vs. 4/50 and 7/50 in controls I
and II). For female rats, the incidence of hepatocellular hypertrophy was reported to be 10/50
rats (control I) and 16/50 (control II) rats with 18/50, 13/50 and 12/50 female rats having
hepatocellular hypertrophy after 15, 45, and 135 mg/kg CH, respectively. The lack or reporting
in regard to final body weights, histology, and especially background and treatment group data
for tumor incidences, limit the interpretation of this study. Whether this paradigm was sensitive
for induction of liver cancer cannot be determined.
From the CH studies in mice, there is an apparent increase in liver adenomas and
carcinomas induced by CH treatment by either drinking water or gavage with all available
02/12/09
E-229
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
studies performed in male B6C3F1 mice. However, the background levels of hepatocellular
adenomas and carcinomas in these mice in George et al (2000) and body weight data from this
study show it is from a tumor prone mouse. Comparisons with concurrent studies of mice
exposed to DC A revealed that while both CH and DC A induced hepatomegaly and cytomegaly,
DCA-induced cytomegaly was accompanied by vacuolization while that of CH to be associated
with increased number of basophilic granules (rough endoplasmic reticulum) which would
suggest separate effects. However, the morphology of the CH-induced tumors was reported to
be similar between DC A and CH-induced tumors (Daniels et al., 1992). Using a similar
paradigm (2-year study of B6C3F1 male mice), De Angelo et al. (1999) and Carter et al. (2003)
described DCA-induced tumors to be heterogeneous. This is the same description given for
TCE-induced tumors in the studies by NTP, NCI, and Maltoni et al. and to be a common
description for tumors caused by a variety of carcinogenic agents. Similar to the studies cited
above for CH, DeAngelo et al (1999) reported that PCO levels were only elevated at 26 weeks at
3.5 g/1 DC A and had returned to control levels by 52 weeks. Similar to CH, no increased
tritiated thymidine was reported for DC A at 26 and 52 weeks with only 2-fold of control values
reported at 0.05 g/1 at 4 weeks. Leakey et al. (2003a) reported that ad libitum fed male mice
exhibited a similar degree of increased incidence of peripheral or focal fatty change at 15 months
for all CH doses but not enzymes associated with peroxisome proliferation. While dietary
restriction seemed to have decreased background levels of tumors and increased time-to-tumor,
CH-gave a clear dose-response in dietary restricted animals. However, while the overall level of
tumor induction was reduced there was a greater induction of PPARa enzymes by CH.
Induction of liver tumors by CH observed in ad libitum fed mice were not correlated with
PPARa induction, with dietary restriction alone appearing to have greater levels of 1 auric acid co-
hydrolase activity in control mice at 15 months. Seng et al. (2003) report that 1 auric acid |3-
hydroxylase and PCO were induced only at exposure levels > 100 mg/kg CH, again with dietary
restricted groups showing the greatest induction. Such data argues against the role of
peroxisome proliferation in CH-liver tumor induction in mice.
2.6. Serum Bile Acid Assays
Serum bile acids (SBA) have been suggested as a sensitive indicator of hepatotoxicity to
a variety of halogenated solvents with an advantage of increased sensitivity and specificity over
conventional liver enzyme tests that primarily reflect the acute perturbation of hepatocyte
membrane integrity and "cell leakage" rather than liver functional capacity (i.e., uptake,
metabolism, storage, and excretion functions of the liver) (Bai et al., 1992b; Neghab et al., 1997).
While some studies have reported negative results, a number of studies have reported elevated
SB A in organic solvent-exposed workers in the absence of any alterations in normal liver
function tests. These variations in results have been suggested to arise from failure of some
methods to detect some of the more significantly elevated SB A and the short-lived and reversible
nature of the effect (Neghab et al., 1997). Neghab et al. (1997) have reported that occupational
exposure to 1,1,2-trichloro-1,2,2-trifluoroethane and trichloroethylene has resulted in elevated
SB A and that several studies have reported elevated SB A in experimental animals to chlorinated
solvents such as carbon tetrachloride, chloroform, hexachlorobutadiene, tetrachloroethylene, 1,1,
02/12/09
E-230
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1 -trichloroethane, and trichloroethylene at levels that do not induce hepatotoxicity (Bai et al.
1992a, b; Ham dan and Stacey 1993,; Wang and Stacey 1990). Toluene, a non-halogenated
solvent, has also been reported to increase SB A in the absence of changes in other hepatobiliary
functions (Neghab and Stacey, 1997). Thus, disturbance in SAB appears to be a generalized
effect of exposure to chlorinated solvents and non-chlorinated solvents and not specific to TCE
exposure.
Neghab et al. (1997) reported that 8 hour time-weighted averages exposure to TCE of 8.9
ppm, measured in the breathing zone using a charcoal tube personal sampler for the whole mean
duration of exposure of 3 .4 years, to have not significant changes in albumin, bilirubin, alkaline
phosphatase, alanine aminotransferase, 5"-nucleosidase, y-glutamyltransferase, but to have
significantly increased total serum bile acids. Not only were total bile acids significantly
increased in these TCE-exposed workers compared to controls (~ 2-fold of control), but,
specifically, deoxycholic acid and subtotal of free bile acids were increased. Neghab et al.
(1997) do not show the data, but also report that "despite the apparent overall low level of
exposure, there was a very good correlations (r = 0.94) between the degree of increase in serum
concentration of total bile acids and level of TCE". Neghab et al (1997) note that while a
sensitive indicator or exposure to such solvents in asymptomatic workers, there is no indication
that actual liver injury occurs in conjunction with SAB increases.
Wang and Stacey (1990) administered TCE in corn oil via i.p. injection to male Sprague-
Davvley rats (300-500 g) at concentrations of 0.01, 0.1, 1, 5, and 10 mmol/kg on 3 consecutive
days (n = 4, 5, or 6) with liver enzymes and SB A examined 4 hours after the last TCE treatment.
At these dose, there were not differences between treated and control animals in regard to
alkaline phosphatase and sorbitol dehydrogenase concentrations and an elevation of alanine
aminotransferase only at the highest dose. However, there was generally a reported dose-related
increase in cholic acid, chenodeoxycholic acid, deoxycholic acid, taurocholic acid,
tauroursodeoxycholic acid with cholic acid and taurochlolic acid increased at the lowest dose.
The authors report that "examination of liver sections under light microscopy yielded no
consistent effects that could be ascribed to trichloroethylene." In the same study a rats were also
exposed to TCE via inhalation (n = 4) at 200 ppm for 28 days, and 1000 ppm for 6 hours/day.
Using this paradigm, cholic acid and taurocholic acid were significantly elevated at the 200 ppm
level, (-10 -fold and ~ 5-fold of control, respectively) with very large standard errors of the
mean. At the 1000 ppm level (6 hours, day) cholic acid and taurocholic acid were elevated to ~
2-fold of control but neither was statistically significant. The large variability in responses
between rats and the low number of rats tested in this paradigm limit its ability to determine
quantitative differences between groups. Nevertheless, without the complications associated
with i.p. exposure (see Section 2.2.1 above), both inhalation exposure of TCE at a relative low
exposure level was also associated with increased SB A levels. The authors stated that "no
increases in alanine amino transferase levels were observed in the rats exposed to
trichloroethylene via inhalation." No histopathology results were reported for rats exposed via
inhalation. As stated by Wang and Stacey (1990), "intraperitoneal injection is not particularly
relevant to humans" which was the rationale given for the inhalation exposure experiments in the
02/12/09
E-231
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
study. They point out that intestinal interactions require consideration because a major
determinant of SB A is their absorption from the gut and intestinal flora may play a role in bile
acid metabolism. They also note that grooming done by the experimental rats would probably
give small exposure via ingestion of TCE as well. However, Wang and Stacey (1990) reported
consistent results in terms of TCE-induced changes in SB A at relatively low concentrations by
either inhalation or i.p. routes of exposure that were not associated with other measures of
toxicity.
Ham dan and Stacey (1993) administered TCE in corn oil (1 mmol/kg) in male Sprague
Davvley rats (300 - 400 g) and followed the time-course of SB A elevation, TCE concentration
and trichloroethanol in the blood at 2, 4, 8, and 16 hours after dosing ( n= 4,5, or 6 per group).
Liver and blood concentration of TCE were reported to peak at 4 hours while those of
trichloroethanol peaked at 8 hours after dosing. TCE levels were not detectable by 16 hours in
either blood or liver while those of trichloroethanol were still elevated. Elevations of SB A were
reported to parallel those of TCE with cholic acid and taurochloate acid reported to show the
highest levels of bile acids. The dose given was based on that reported by Wang and Stacey
(1990) to give no hepatotoxicity but an increase in SBA. The authors state that liver injury
parameters were checked and found unaffected by TCE exposure but do not show the data.
Thus, it was TCE concentration and not that of its metabolite that was most closely related to
changes in SBA and after a single exposure, the effect was reversible. In an in vitro study by Bai
and Stacey (1993), TCE was studied in isolated rat hepatocytes with TCE reported to cause a
dose-related suppression of initial rates of cholic acid and taurocholic acid but with no significant
effects on enzyme leakage and intracellular calcium contents, further supporting a role for the
parent compound in this effect. The authors noted that the changes in SBA result from
interference with a physiological process rather "than an event associated with significant
pathological consequences "
State of Science of Liver Cancer MOAs
The experimental evidence in mice shows that TCE and its metabolites induce foci,
hepatocellular adenomas, and carcinomas that are heterogeneous in nature as indicated by
phenotypic differences in tincture, mutational markers, or gene expression markers. The tumors
induced by TCE are reflective of phenotypes that are either similar to those induced by mixtures
of DC A and TCA exposure, or more like those induced by DCA. These tumors have been
described to be similar also to those arising spontaneously in mice or from chemically induced
hepatocarcinogenesis and to arise from preneoplastic foci, and in the case of DCA, single
dysplastic hepatocytes as well as foci. Hepatocellular carcinoma (HCC) observed in humans
also has been described to be heterogeneous and to be associated with formation of preneoplastic
nodules. Although several conditions have been associated with increased risk of liver cancer in
humans, the mechanism of HCC is unknown at this time. A great deal of attention has been
focused on predicting which cellular targets (e.g., "stem-cell" or mature hepatocyte) are
associated with HCC as well as on phenotypic markers in HCC that can provide insight not only
into MOA and origin of tumor, but also for prediction of clinical course. Examination of
02/12/09
E-232
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
pathways and epigenetic changes associated with cancer, and the relationship of these changes to
liver cancer are also discussed below. The field of cancer research has been transformed by the
recent discoveries of epigenetic changes and their role in cancer and chronic disease states. The
following discussion describes these advances but also the issues involved with the technologies
that have emerged to describe them (see Section 3.1.2., below). Exposure to TCE and its
metabolites, like many others, induces a heterogeneous response, even in a relatively
homogeneous genetic paradigm as the experimental laboratory rodent model. The importance of
phenotypic anchoring is a major issue in the study of any MO As using these new technologies of
gene expression pattern. Although a large amount of information is now available using
microarray technologies and transgenic mouse models, specifically for TCE and in study of
suggested MO As for TCE and its metabolites, use of these approaches has limitations that need
to be considered in the interpretation of data and conclusions derived from such data, especially
quantitative conclusions.
For TCE and its metabolites, the extent of acute to subchronic induction of hepatomegaly
correlated with hepatocellular carcinogenicity, although each had differing factors contributing
to that hepatomegaly from periportal glycogen deposition to hepatocellular hypertrophy and
increased polyploidy. The extent of transient DNA synthesis, peroxisome proliferation, or
cytotoxicity was not correlated with carcinogenicity. Hepatomegaly is also a predictor of
carcinogenicity for a number of other compounds in mice and rats. Allen et al. (2004) examined
the NTP database (87 compounds for rat and 83 for mice) and tried to correlate specific
hepatocellular pathology in prechronic studies with carcinogenic endpoints in the chronic 2-year
assays. The best single predictor of liver cancer in mice was hepatocellular hypertrophy.
Hepatocellular cytomegaly and hepatocyte necrosis also contributed, although the numbers of
positive findings were less than hypertrophy. With regard to genotoxicity studies, there was no
evidence of a correlation between mouse liver tumor chemicals and Salmonella or micronucleus
assay outcome. None of the prechronic liver lesions examined were correlated with either
Salmonella or Micronucleus assays. In rats no single prechronic liver lesions (when considered
individually) was a strong predictor of liver cancer in rats. The most predictive lesions was
hepatocellular hypertrophy. There was not significant correlation between liver tumors/toxicity
and the 2 mutagenicity measures. Although the lack of correlation with the mutagenicity assays
could be interpreted as rodent assays predominantly identifying non-genotoxic liver carcinogens,
this conclusion could be questioned because it is solely dependent on Salmonella mutagenicity
and additional genotoxic endpoints could conceivably shift the association between liver cancer
and genotoxicity towards a more positive correlation. As to questions of the usefulness of the
mouse bioassay, the two mutagenicity assays did not correlated with rat results either and an
important indicator for carcinogenicity would be lost.
Examination of tumor phenotype from TCE, DCA and TCA exposures in mice shows a
large heterogeneity which is also consistent with the heterogeneity observed in human HCC (see
Section 3.1.8., below). The heterogeneity of tumor phenotype has been correlated with survival
outcome and tumor aggressiveness in humans and in transgenic mouse models that share some of
the same perturbations in gene pathway expression (see Sections 3.1.8. and 3.2.1, below). An
02/12/09 E-233
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
examination of common pathway disturbances that may be common to all cancers and those of
liver tumors shows that there are pathways in common but that there is greater heterogeneity in
disturbance of hepatic pathways in cancer that may make is useful as a marker of disturbance of
cancer pathway that may lead to different targets of carcinogenicity depending on the cellular
context and target. Thus, although primate and human liver may not be as susceptible to HCC as
the rodent liver, the pathways leading to HCC in rodents and humans appear to be similar and
heterogeneous, with some indicative of other susceptible cellular targets for neoplasia in a
differing context.
3.1. State of Science for Cancer and Specifically Human Liver Cancer
3.1.1. Epigenetics and Disease States (Transgenerational Effects, Effects of Aging and
Background changes)
Recently, Wood et al. (2007) published their work on "genomic landscapes" of human
breast and colorectal cancers that significantly forwards the understanding of "key events"
involved with induction of cancer. They state that there are ~ 80 DNA mutations that alter
amino acid in a typical cancer but that examination of the overall distribution these mutations in
different cancers of the same type leads to a new view of cancer genome landscapes: they are
composed of a handful of commonly mutated genes "mountains" but are dominated by a much
larger number of infrequently mutated gene "hills." "Statistical analyses suggested that most of
the ~ 80 mutation in an individual tumor were harmless and that <15 were likely to be
responsible for driving the initiation, progression, or maintenance of the tumor." "Historically
the focus of cancer research has been on the gene mountains, in part because they were the only
alterations that could be identified with available technologies. However, our data show that vast
majority of mutations in cancers do not occur in such mountains. This new view of cancer is
consistent with the idea that a large number of mutations, each associated with a small fitness
advantage, drive tumor progression. It is the "hills" and not the "mountains" that dominate the
cancer genomic landscape." The large number of "hills" actually reflects alterations in a much
smaller number of cell signaling pathways. Indeed, pathways rather than individual genes
appear to govern the course of tumorigenesis." "It is becoming increasingly clear that pathways
rather than individual genes govern the course of tumorigenesis. Mutations in any of several
genes of a single pathway can thereby cause equivalent increases in net cell proliferation." "This
new view of cancer is consistent with the idea that a large number of mutations, each associated
with a small fitness advantage, drive tumor progression." Thus, when pathways are altered the
same phenotype can arise from alterations in any of several genes.
Consistent with the arguments put forth by Wood et al. (2007) for mutations in cancer is
the additional insight into pathway alterations by epigenomic mechanisms which can act
similarly as mutation. Weidman et al. (2007) report that "cell phenotype is not only dependent
on its genotype but also on its unique epigenotype, which is shaped by developmental history
and environmental exposures. The human and mouse genome projects identified approximately
15,500 and 29,000 CpG islands, respectively. Hypermethylation of CpG-rich regions of gene
promoters inhibit expression by blocking the initiation of transcription. DNA methylation is also
02/12/09 E-234
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
involved in the allelic inactivation of imprinted genes, the silencing of genes on the inactive X
chromosome, and the reduction of expression of transposable elements. Because epigenomic
modifications are copied after DNA synthesis by DNMT1, they are inherited during somatic cell
replication." "Inherited and spontaneous or environmentally induced epigenetic alterations are
increasingly being recognized as early molecular events in cancer formation. Furthermore, such
epigenetic alterations are potentially more adverse than nucleotide mutations because their
effects on regional chromatin structure can spread, thereby affecting multiple genetic loci.
Although tumor suppressor gene silencing by DNA methylation occurs frequently in cancer,
genome-wide hypomethylation is one of the earliest events to occur in the genesis of cancer.
Demethylation of the genome can lead to the reactivation of transposable elements, thereby
altering the transcription of adjacent genes, the activation of oncogenes such as H-Ras, and
biallelic expression of imprinted loci (e.g., loss of IGF2 imprinting)." Thus, epigenetic
modification may be worse than mutation in terms of cancer induction.
Dolinoy et al. (2007) report on the role of environmental exposures on the epigenome,
especially during critical periods of development and their role in adult disease susceptibility.
They report that "aberrant epigenetic gene regulation has been proposed as a mechanism of
action for non-genotoxic carcinogenesis, imprinting disorders, and complex disorders including
Alzheimer's disease, schizophrenia, asthma, and autism. Epigenetic modifications are inherited
not only during mitosis but also can be transmitted transgenerationally (Rakyan et al., 2002;
Rakyan et al., 2003; Anway et al., 2005). The influence on environmental factors on epigenetic
gene regulation may also persist transgenerationally despite lack of continued exposure in
second, third, and fourth generations (Anway et al., 2005). Therefore if the genome is compared
to the hardware in a computer, the epigenome is the software that directs the computer's
operation." "The epigenome is particularly susceptible to deregulation during gestation, neonatal
development, puberty and old age. Nevertheless, it is most vulnerable to environmental factors
during embryogenesis because DNA synthetic rate is high, and the elaborate DNA methylation
pattern and chromatin structure required for normal tissue development is established during
early development." "83 imprinted genes have been identified in mice and humans with 29 or
about one third being imprinted in both species. Since imprinted genes are functionally haploid,
they are denied the protection from recessive mutations that diploidy would normally afford.
Imprinted genes that have been linked to carcinogenesis include IGF2 (Bladder, lung, ovarian
and others), IGF2R (breast, colon, lung, and others), and Neuronatin (pediatric leukemia)."
Bjornsson et al. (2008) recently reported that not only were time-dependent changes in
global DNA methylation within the same individuals in 2 separate populations in widely
separated geographic locations, these changes showed familial clustering in both increased and
decreased methylation. These results were not only suggested to support the relationship of age-
related loss of normal epigenetic patterns as a mechanism for late onset of common human
diseases but also that losses and gains of DNA methylation observed over time in different
individuals could contribute to disease with the example provided of cancer which is associated
with both hypomethylation and hypermethylation through activation of oncogenes and silencing
of tumor suppressor genes. The study also showed considerable interindividual age variation,
02/12/09
E-235
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
with differences accruing over time within individuals that would be missed by studies that
employed group averaging.
The review by Reamone-Buettner and Borlak (2007) provide insight into the role of non-
coding RNAs in diseases such as cancer. They report that "a large number of non-coding RNAs
(ncRNAs) play important role in regulating gene expressions, and advances in the identification
and function of eukaryotic ncRNAs, e.g., microRNAs and their function in chromatin
organization, gene expression, disease etiology have been recently reviewed. The regulatory
pathways mediated by small RNAs are usually collectively referred to as RNA interference
(RNAi) or RNA-mediated silencing. RNAi can be triggered by small double-stranded RNA
(dsRNA) either introduced exogenously into cells as small interfering siRNAs or that have been
produced endogenously from small non-coding RNAs known as microRNAs (miRNAs). The
dsRNAs are characteristically cleaved by the ribonucleas Ill-enzyme Dicer into 21- to 23 nt
duplexes and the resulting fragments base-pair with complementary mRNA to target cleavage or
to repress translation." "Two mechanisms exist of miRNA-mediated gene regulation,
degradation of the target mRNA, and translational repression. Whether one or the other of these
mechanisms is used depends on the degree of the complementary between the miRNA and target
mRNA. For a near perfect match, the Argonaute protein in the RNA-induced silencing complex
(RISC) cleaves the mRNA target, which is destined for subsequent degradation by ribonucleases.
In the situation of a less degree of complimentarity, commonly occurring in humans, the
translational repression mechanism is used to control gene expression. However, the exact
mechanism for translational inhibition is unclear." The varying degrees in complimentarity
would help explain the large number of genes that could be affected by miRNA and pleiotropic
response.
The review by Feinberg et al (2006) specifically addresses the epigenetic progenitor
origin of human cancer. They conclude that epigenetic alterations are ubiquitous and serve as
surrogate alterations for genetic change (oncogene activation, tumour-suppressor-gene
silencing), by mimicking the effect of genetic change. They report that:
Advances in characterizing epigenetic alterations in cancer include global alterations,
such as hypomethylation of DNA and hypoacetylation of chromatin, as well as gene-
specific hypomethylation and hypermethylation. Global DNA hypomethylation leads to
chromosomal instability and increased tumour frequency, which has been shown in vitro
and in vivo in mouse models, as well as gene-specific oncogene activation, such as R-ras
in gastric cancer, and cyclin D2 and maspin in pancreatic cancer. In addition, the
silencing of tumour-suppressor genes is associated with promoter DNA hypermethylation
and chromatin hypoacetylation, which affect divergent genes such as retinoblastoma 1
(RBI), pl6 (also known as cyclin-dependent kinase inhibitor 2A (CDKN2A), von
Hippel-Lindau tumor suppressor (VHL), and MutL protein homologue (MLH1).
Genetic mechanisms are not the only path to gene disruption in cancer. Pathological
epigenetic changes - non-sequence-based alteration that are inherited through cell
02/12/09
E-236
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
division - are increasingly being considered as alternatives to mutations and chromosomal
alterations in disrupting gene function. These include global DNA hypomethylation,
hypermethylation and hypomethylation of specific genes, chromatin alterations and loss
of imprinting. All of these can lead to aberrant activation of growth-promoting genes and
aberrant silencing of tumour-suppressor genes.
Most CG dinucleotides are methylated on cytosine residues in vertebrate genomes. CG
methylation is heritable, because after DNA replication the DNA methyltransferase 1,
DNMT1, methylates unmethylated CG on the base-paired strand. CG dinucleotides
within promoters within promoters tend to be protected from methylation. Although
individual genes vary in hypomethylation, all tumours have shown global reduction of
DNA methylation. This is a striking feature of neoplasia.
In addition to global hypomethylation, promoters of individual genes show increased
DNA methylation levels. Hypermethylation of tumour-suppressor genes can be tumour-
type specific. An increasing number of genes are found to be normally methylated at
promoters but hypomethylated and activated in the corresponding tumours. These
include R-RAs in gastric cancer, melanoma antigen family A, l(MAGEl) in melanoma,
maspin in gastric cancer, S100A4 in colon cancer, and various genes in pancreatic cancer.
Our genetic material is complexed with proteins in the form of histones in a one-to-one
weight ratio. Core histones H2A, H2B, H3 and H4 form nucleosome particles that
package 147 bp of DNA, and the linker histone HI packages more DNA between core
particles, forming chromatin. It is chromatin and not just DNA, that is the substrate for
all processes that affect genes and chromosomes. In recent years, it has become
increasingly evident that chromatin, like DNA methylation, can impart memory to genetic
activity. There are dozens of post-translational histone modifications. Studies in many
model systems have shown that particular histone modifications are enriched at sites of
active chromatin (histone H3 and H4 hyperacetylation, lysing at 4 and H3 (H3-K4)
dimethylation and trimethylation, and H3-K79 methylation) and others are enriched at
sites of silent chromatin (H3-K9 and H3-K27 methylation). These and other histone
modifications survive mitosis and have been implicated in chromatin memory.
Overproduction of key histone methyltransferases that catalyze the methylation of either
H3-K4 or H3-K27 residues are frequent events in neoplasia. Global reductions in
monoacetylated H4-K16 and trimethylated H4-K20 are general features of cancer cells.
Genomic imprinting is parent-of -origin-specific gene silencing. It results from a germ-
line mark that causes reduced or absent expression of a specific allele of a gene in
somatic cells of the offspring. Imprinting is a feature of all mammals affecting genes that
regulate cell growth, behaviour, signaling, cell cycle and transport; moreover, imprinting
is necessary for normal development. Imprinting is important in neoplasia because both
gynogenotes (embryos derived only from the maternal genetic complement) and
02/12/09 E-237
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
androgenotes (embryos derived only from the paternal genetic complement) form
tumours - ovarian teratomas, and hydtidiform moles/ choriocarcinomas, respectively.
Loss of imprinting (LOI) refers to activation of the normally silenced allele, or silencing
of the normally active allele, of an imprinted gene. LOI of the insulin-like growth factor
2 gene (IGF2) accounts for half of Wilms tumours in children. LOI of IGF2 is also a
common epigenetic variant in adults and is associated with a fivefold increased frequency
of colorectal neoplasia. LOI of IGF2 might cause cancer by increasing the progenitor cell
population in the kidney in Wilm's tumor and in the gastrointestinal tract in colorectal
cancer.
Feinberg et al. (2006) propose that epigenetic changes can provide mechanistic unity to
understanding cancer, they can occur earlier and set the stage for genetic alterations, and have
been linked to the pluripotent precursor cells form which cancers arise. "To integrate the idea of
these early epigenetic events, we propose that cancer arises in three steps; an epigenetic
disruption of progenitor cells, an initiating mutation and genetic and epigenetic pasticity."
The first step involves an epigenetic disruption of progenitor cells in a given organ or
system, which leads to a polyclonal precursor population of neoplasia-ready cells. These
cells represent a main target of environmental, genetic and age-dependent exposure that
largely accounts for the long latency period of cancer. Epigenetic disruption might
perturb the normal balance between undifferentiated progenitor cells and differentiated
committed cells within a given anatomical compartment, either in number or in their
capacity for aberrant differentiation, which provides a common mechanism of neoplasia.
All tumours show global changes in DNA methylation, and DNA methylation is clonally
inherited through cell division. Because the conventional genetic changes in cancer are
also clonal, global hypomethylation would have to occur universally, at the same moment
as the mutational changes, which seems unlikely. This suggests that global DNA
hypomethylation (and global reductions of specific histone modifications) precedes
genetic change in cancer. Similarly, hypermethylation of tumour-suppressor genes has
been observed in the normal tissue of patients in which the same gene is hypermethylated
in the tumour tissue. Recent data demonstrate LOI of IGF2 throughout the normal
colonic epithelium of patients who have LOI-associated colorectal cancer. LOI is
associated with increased risk of intestinal cancers in both humans and mice. A specific
change in the epithelium is seen in mice that are engineered to have biallelic expression
of IGF2 - a shift in the proportion of progenitor to differentiated cells throughout the
epithelium; a similar abnormality was observed in humans with LOI of IGF2.
The proposed existence of the epigenetically disrupted progenitors of cancer implies that
the earliest stages in neoplastic progression occur even before what a pathologist would
recognize as a benign pre-neoplastic lesion. Such alterations are inherently polyclonal.
This is in contrast with the widely accepted model of cancer as a monoclonal disorder
that arises from an initiating mutation- a model that was proposed and accepted when
02/12/09
E-238
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
little was known about epigenetic phenomena in cancer.
Thus, Feinberg et al. (2006) provide a hypothesis for the latency period of cancer and
suggest that epigenetic changes predate mutational ones in cancer. Tissues which look
phenotypically "normal" may harbor epigenetic changes and predispositions toward neoplasia.
In regard to what cells may be targets or epigenetic changes that can be "progenitor cells" in the
case of cancer, Feinberg et al. (2006) define such cell having "capacity for self-renewal and
pluripotency - over their tendency toward limited replicative potential and differentiation."
Within the liver, there are multiple cell types that would fit such a definition including those who
are considered "mature" (see Section 3.1.4, below). Feinberg et al. (2006) also note that
epigenetic states can be continuously modified to become heterogeneous at all states of the
neoplastic process.
Telomere erosion results in chromosome shortening and uncapped ends that begin to fuse
and the resulting dicentric chromosomes break at anaphase. DNA palindromes have
recently been found to form at high levels in cancer cells. Like telomere erosion, DNA
palindrome formation can lead to genetic instability by initiating bridge-breakage-fusion
cycles. However, it is not known how or exactly when palindromes form, although they
appear early in cancer progression. Epigenetic instability can also promote cancer
through pleiotropic alterations in the expression of genes that modify chromatin.
Epigenetic changes are reversible but the changes can initiate irreversible genetic
changes. Permanent epigenetic changes can have an epigenetic basis. On a background of
cancer-associated epigenetic instability, the effects of mutations in oncogenes and tumour
-suppressor genes might be exacerbated. Therefore the risk of developing malignancy
would be much higher for a given mutations event if it occurred on the background of
epigenetic disruption.
The environmental dependence of cancer fits an epigenetic model generally for human
disease - the environment might influence disease onset not simply through mutational
mechanisms but in epigenetically modifying genes that are targets for either germline or
acquired mutation; that is, by allowing genetic variates to be expressed. Little is known
about epigenetic predispositions to cancer, but a recent twin study indicates that, similar
to cancer risk, global epigenetic changes show striking increase with age.
Environmental insults might affect the expression of tumour-progenitor genes, leading to
both genetic and epigenetic alterations. Liver regeneration after tissue injury leads to
widespread hypomethylation and hypermethylation of individual genes; both of these
epigenetic changes occur in cancer.
In regard to the implications of epigenomic changes and human susceptibility to toxic
insult, the review by Szyf (2007) provides additional insights.
02/12/09
E-239
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The basic supposition in the field has been that the interindividual variations in response
to xenobiotic are defined by genetic differences and that the main hazard anticipated at
the genomic level from xenobiotic is mutagenesis or physical damage to DNA. In
accordance with this basic hypothesis, the main focus of attention in pharmacogenetics
has been on identifying polymorphisms in genes encoding drug metabolizing enzymes
and receptors. New xenobiotics were traditionally tested for their genotoxic effects.
However, it is becoming clear that epigenetic programming plays an equally important
role in generating interindividual phenotypic differences, which could affect drug
response. Moreover, the emerging notion of the dynamic nature of the epigenome and its
responsibility to multiple cellular signaling pathways suggest that it is potentially
vulnerable to the effects of xenobiotics not only during critical period in development but
also later in life as well. Thus, non-genotoxic agents might affect gene function through
epigenetic mechanisms in a stable and long-term fashion with consequences, which might
be indistinguishable form the effects of physical damage to the DNA. Epigenetic
programming has the potential to persist and even being transgenerationally transmitted
(Anway et al., 2005) and this possibility creates a special challenge for toxicological
assessment of safety of xenobiotics. Any analysis of interindividual phenotype diversity
should therefore take into account epigenetic variations in addition to genetic sequence
polymorphisms. Whereas, a germ-line polymorphism is a static property of an individual
and might be mapped in any tissue at any point in life, epigenetic differences must be
examined at different time points and at diverse cell types.
Karpinets and Foy (2005) propose that epigenetic alterations preceed mutations and that
succeeding mutations are not random but in response to specific types of epigenetic changes the
environment has encouraged. This mechanism was also suggested as to both explain the delayed
effects of toxicant exposure and the bystander effect of radiation on tumor development, which
are inconsistent with the accepted mechanism of direct DNA damage. "In a study of ionizing
radiation, non-irradiated cells acquired mutagenesis through direct contact with tells whose
nuclei had previously been irradiated with alpha-particles (Zhou et al., 2003). Molecular
mechansms underlying these experimental findings are not known but it is believed that it may
be a consequence of bystander interactions involving intercellular signaling and production of
cytokines (Lorimore et al., 2003)."
Caldwell and Keshava (2006) report that "aberrant DNA methylation has emerged in
recent years as a common hallmark of all types of cancers with hypermethylation of the promoter
region of specific tumor suppressor genes and DNA repair genes leading to their silencing (an
effect similar to their mutation), and genomic hypomethylation (Ballestar and Esteller, 2002;
Berger and Daxenbichler, 2002; Herman et al.,1998; Pereira et al. 2004; Rhee et al., 2002).
Whether DNA methylation is a consequence or cause of cancer is a long-standing issue
(Ballestar and Esteller, 2002). Fraga et al. (2004, 2005) report global loss of monoacetylation
and trimethylation of histone H4 as common a hallmark of human tumor cells but suggest
genomone-wide loss of 5-methylcytosine (associated with the acquisition of a transformed
phenotype) does not exist as a static predefined value throughout the process of carcinogenesis
02/12/09
E-240
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
but as a dynamic parameter (i.e., decreases are seen early and become more marked in later
stages)."
3.1.2. Emerging Technologies, DNA and siRNA, miRNA Microarrays- promise and
limitations for MOA
Currently new approaches are emerging for the study of changes in gene expression and
protein production induced by chemical exposure that could be related to their toxicity and serve
as an anchor for determining similar patterns between rodent models and human diseases or risks
of chemically-induced health impacts. Such approaches have the promise to extend the
definitions of "genotoxic" and "non-genotoxic" effects which with the advent of epigenomic
study have become obsolete as they assume only alteration of the DNA sequence is important in
cancer induction and progression. However, not only is phenotypic anchoring an issue in
regard to the differing cell types, regions, and lobes of the liver (see Section 1.2, above), it is also
an issue for overall variability of response between animals and is critical for interpretation of
microarray and other genomic database approaches. As shown in the discussions of TCE effects
in animal models, TCE treatment resulted in a large variability in response between what are
supposed to be relatively homogeneous genetically similar animals and there was an apparent
difference in response between studies using the same paradigm. It is important that as varying
microarray approaches and analyses of TCE toxicity or of potential MO As are published, the
issue of phenotypic anchoring at the cellular to animal level is addressed. Several studies of
TCE microarray results and those of PPARa agonists have been reported in the literature in an
attempt to discern MO As. Issues related to conduct of these experiments and interpretation of
their results are listed below.
Perhaps one of the most important studies of this issue has been reported by Baker et al.
(2004). The ILSIHESI formed a hepatotoxicity working group to evaluate and compare
biological and gene expression responses in rats exposed to well-studied hepatotoxins (Clofibrate
and methapyrilene), using standard experimental protocol and to address the following issues: a)
how comparable are the biological and gene expression data from different laboratories running
identical in vivo studies b) how reproducible are the data generated across laboratories using the
same microarray platform c) how do data compare using different microarray platforms; d) how
do data compare using RNA from pooled and individual animals; e) do the gene expression
changes demonstrate time- and dose-dependent responses that correlate with known biological
markers of toxicity? (Baker et al., 2004). The rat model studied was the male Sprague-Dawley
rat (57 or 60-66 days of age) exposed to 250 or 25 mg/kg/day Clofibrate for 1, 3 or 7 days. Two
separate in vivo studies were conducted: one at Abbott Laboratories and on at GSK (UK). There
was a difference in biological response between the two laboratories. The high dose (250
mg/kg/day) group at day 3 had a 15% increase in liver weight relative to body weight in the GSK
study, compared with a 3% liver weight increase in the Abbott study. At 7-days, there was a
31% liver weight increase in the GSK study and 15% in the Abbott study. Observed changes in
clinical chemistry parameters also indicated difference in the biological response of the in vivo
study concordant with difference in liver weight. A significant reduction in total cholesterol
02/12/09
E-241
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
levels was seen in the GSK study at the high dose for all time points. However, the Abbott study
demonstrated a significant reduction only at one dose and time point. The incidence of mitotic
figures also differed between the labs. In both studies there was a 2-3 times greater Acyl-CoA
enzyme (ACOX) activity at the high dose but no difference from control in the low dose. Again
the GSK lab gave greater response. For microarrays, GSK and ULR pooled samples from each
treatment group of 4 animals. U.S. EPA did some of the microarray analyses as well as GSK
and ULR (GSK in UK). It is apparent that although the changes in genes were demonstrated by
both laboratories, there were quantitative differences in the fold change values observed between
the two sites.
The U.S. EPA analyzed gene expression in individual RNA samples obtained from day 7
high and low-dose animals that had been treated at Abbot. GSK (US) and ULR analyzed gene
expression in pooled RNA from day 7 high and low dose animals treated at GSK (UK). Gene
expression data from individual animal samples indicated that 7 genes were significantly
upregulated (maximum of 7.2-fold) and 12 were down regulated (maximum of 4.3-fold decrease)
in the high dose group. The low dose group generated only one statistically significant gene
expression change, namely heat shock protein 70 (HSP70), In comparison, expression changes
in the 7-day pooled high-dose samples analyzed by GSK (U.S.) ranged from 43.3-fold to a 3.5-
fold decrease. Changes in these same samples analyzed by ULR ranged from a 4.9 fold increase
to a 4.3-fold decrease. As an example, the microarray fold change at 7-day 250 mg/kg/day
Clofibrate showed a 3.8-fold increase for US EPA individual animals sampled, and 2.2-fold
increase for pooled samples by ULR, and a 20.3-fold increase in pooled samples by GSK (US)
for CYP4A1 (Baker et al., 2004). Thus, these results show a very large difference not only
between treatment groups but between pooled an non-pooled data and between labs analyzing
the same RNA.
Not only was there a difference in DNA microarray results but a comparison of gene
expression data from day 7 high-dose samples obtained using quantitative realtime PCR versus
data generated using cDNA microarrays is shown a quantitative difference but qualitative similar
patterns. Although both methods of quantitative real time PCR on the pooled sample showed the
PPARa gene to be down regulated, the GSK (IS) pooled sample microarray analysis indicated
upregulation; the URL pooled and U.S. EPA individual microarray analyses showed no change.
The microarray for PPARa at 7-day 250 mg/kg/day Clofibrate showed no change for individual
animals (U.S EPA), no change for pooled samples (ULR) and upregulation of 1.8-fold value for
pooled samples for GSK(US). The quantitative real time PCR on the pooled sample using
Taqman gave a 4.5-fold down regulation and using SYBR Green gave a 1.2-fold down
regulation of PPARa.
Baker et al. (2004) reported that the pooling of samples for microarray analysis has been
used in the past to defray the cost of microarray experiments, reduce the effect of biological
variation, and in some cases overcome availability of limiting amounts of tissues. Unfortunately
this approach essentially produced a sample size (n) of one animal. Repeated microarray
experiments with such pooled RNA produces technical replicates as opposed to true biological
replicates and thus does not allow calculation of biologically significant changes in gene
02/12/09
E-242
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
expression between different dose groups or time points. Another possible consequence of
pooling is to mask individual gene changes and leave open the possibility of introducing error
due to individual outlier responses.
Woods et al. (2007a) note that "because toxicogenomics is a relatively novel technology,
there are a number of limitations that must be resolved before array data is widely accepted.
Microarray studies have been touted as being highly sensitive for detecting toxic responses at
much earlier time points and/or lower doses than histopathology, clinical chemistry or other
traditional toxicological assays can detect. However, based on the nature of the assay,
measurements of extreme levels of gene expression - low or high -are thought to be unreliable.
Also the reproducibility of microarray experiments has raised concerns. "Batch effects" based
on the day, user, and laboratory environment have been observed in array datasets. To address
these concerns, confirmation of microarray-derived gene expression profiles is typically
performed using quantitative real time polymerase chain reaction (RT-PCR) or Northern blot
analysis."
In addition to the issues raised above, Waxman et al. (2007) raise issues regarding how
quantitative realtime PCR experiments are conducted. They state that cancer development
affects almost all pathways and genes including the "housekeeping" genes, which are involved in
the cell's common basic functions [e.g., glyceraldehyde-3-phosphate dehydrogenase (GADPH),
beta actin (ACTB), TATA-binding protein (TBP), ribosomal proteins (RP), and many more].
However, "many of these genes are often used to normalize quantitative real-time RT-
PCR(qPCR) data to account for experimental differences, such as differences in RNA quantity
and quality, the overall transcriptional activity and differences in cDNA synthesis. GADPH and
ACTB are most commonly used for normalization, including studies of cancer. Waxman et al
(2007) suggest that despite the fact that it has been shown that these genes are differentially
expressed in cancers, including colorectal-, prostate-, and bladder-cancer, some qPCR studies on
hepatocellular carcinoma used GAPDH or ACTB for normalization. Since many investigations
on cancer include multiple comparisons, and analyze different stages of the disease, such as
normal tissue, pre-neoplasm, and consecutive stages of cancer, "it crucial to find an appropriate
gene for normalization" whose expression is constant throughout all disease stage and not
response to treatment. For liver cancers associated with exposure to hepatitis C virus (HCV),
Waxman et al (2007) reported that differing states, including preneoplastic lesions (cirrhosis and
dysplasia) and consecutive stages of hepatocellular carcinoma, had differential expression of
"housekeeping" genes and that using them for normalization had an effect on the fold change of
qPCR data and on the general direction (up or down) of differentially expressed genes. For
example, GAPDH was strongly upregulated in advanced and very advanced stages of
hepatocellular carcinoma (in some samples up to 7-fold) and ACTB was up-regulated 2- to 3-
fold in many advanced and very advanced tumor samples. Waxman et al. (2007) conclude that
"microarray data are known to be highly variable. Due to its higher dynamic range qPCR is
thought to be more accurate and therefore is often used to corroborate microarray results.
Mostly, general direction (up and down-regulation) and rank order of the fold-changes are
similar, but the levels of the fold changes of microarray experiments differ compared to qPCR
02/12/09
E-243
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
data and show a marked tendency of being smaller. This effect is more pronounced as the fold
change is very high."
In relation to use of gene expression and indicators of cancer causation, Volgelstein and
Kinzler (2004) make important points regarding their use:
Levels of gene expression are unreliable indicators of causation because disturbance of
any network invariably leads to a multitude of such changes only peripherally related to
the phenotype. Without better ways to determine whether an unmutated but interesting
candidate gene has a causal role in neoplasia, cancer researchers will likely be spending
precious time working on genes only peripherally related to the disease they wish to
study.
This is important caveat for gene expression studies for MOA that are "snapshots in time"
without phenotypic anchoring and even more applicable to experimental paradigms where there
is ongoing necrosis or toxicity in addition to gene changes that may or may not be associated
with neoplasia.
For an endpoint that is not as complex as neoplasia, there are issues regarding uses of
microarray data. In regard to the determination of acute liver toxicity caused by one of the most
studied hepatotoxins, acetaminophen, and its correlation with microarray data, Beyer et al.
(2007) also have reported the results of landmark study examining issues regarding use of this
approach. "The biology of liver and other tissues in normal and disease states increasingly is
being probed using global approaches such as microarray transcriptional profiling. Acceptance
of this technology is based principally on a satisfactory level of reproducibility of data among
laboratories and across platforms. The issue of reproducibility and reliability of genomics data
obtained from similar (standardized) biological experiments performed in different laboratories
is crucial to the generation and utility of large databases of microarray results. While several
recent studies uncovered important limitation of expression profiling of chemical injury to cells
and tissues (Baker et al 2004; Beekman et al 2006; Ulrich et al 2004), determining the effects of
intralaboratory variables on the reproducibility, validity, and general applicability of the results
that are generated by different laboratories and deposited into publicly available databases
remains a gap." "The National Institutes of Environmental Health Sciences (NIEHS) established
the Toxicogenomics Research Consortium to apply the collective and specialized expertise from
academic institutions to address issues in integrating gene expression profiling, bioinformatics,
and general toxicology. Key elements include developing standardized practices for gene
expression studies and conducting systematic assessments of the reproducibility of traditional
toxicity endpoints and microarray data within and among laboratories. To this end the
consortium selected the classical hepatotoxicant acetaminophen (APAP) for its proof of concept
experiments. Despite more than 30 years of research on APAP, we are far from a complete
understanding of the mechanisms of liver injury, risk factors, and molecular markers that predict
clinical outcome after poisoning. APAP-induced hepatotoxicity was performed at seven
geographically dispersed Centers. Parallel studies with N-acetyl-m-aminophenol (AMAP), the
02/12/09
E-244
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
non-hepatotoxic isomer of APAP, provided a method to isolate transcripts associated with
hepatotoxicity" (Beyer et al., 2007).
Beyer et al identified potential sources of interlaboratory variability when microarray
analyses were conducted by one laboratory on RNA samples generated in different laboratories
but using the same experimental paradigm and source of animals. Toxic injury by APAP
showed variability across Centers and between animals [e.g., % liver affected by necrosis (<20%
to 80% at one time period and 0% to 60% at another), control animal serum ALT (3-fold
difference), and in glutathione depletion (<5% to > 60%) between centers]. There was
concordance between APAP toxicity as measured in individual animals (rather than expressed as
just a mean with SE) and transcriptional response. Of course the variability between gene
platforms and processing of the microarrays had been reduced by using the same facility to do all
of the microarray analyses. However, the results show that phenotypic anchoring of gene
expression data is required for biologically meaningful meta-analysis of genomic experiments.
Woods et al. (2007a) note that "improvements should continue to be made on statistical
analysis and presentation of microarray data such that it is easy to interpret. Prior to the current
advances in bioinformatics, the most common way of reporting results of microarray studies
involved listing differentially expressed genes, with little information about the statistical
significance or biological pathways with which the genes are associated." However, there are
issues with the use of "Classifiers" or predictive genomic computer programs based on genes
showing altered expression in association with the observed toxicities. "Although these metrics
built on different machine learning algorithms could be useful in estimating the severity of
potential toxicities induced by compounds, the applications of these classifiers in understanding
the mechanisms of drug-induced toxicity are not straightforward. In particular this approach is
unlikely to distinguish the upstream causal genes from the downstream responsive genes among
all the genes associated with an induced toxicity. Without knowledge of the causal sufficiency
order, designing experiments to test predicted toxicity in animal models remains difficult" (Dai
et al., 2007). Ulrich (2003) states limitation of microarray analysis to study nuclear receptors
(e.g., PPARa). "Nuclear receptors comprise a large group of ligand-activated transcription
factors that control much of cellular metabolism. Toxicogenomics is the study of the structure
and output of the entire genome as it related and responds to adverse xenobiotic exposure.
Traditionally, the genes regulated by nuclear receptors in cells exposed to toxins have been
explored at the mRNA and protein levels using northern and western blotting techniques.
Though effective when studying the expression of individual genes, these approaches do not
enable the understanding of the myriad of genes regulated by individual receptors or of the
crosstalk between receptors." "Discovery of the multiple genes regulated by each receptor type
has thus been driven by technological advances in gene expressional analysis, most commonly
including differential display, RT-PCR and DNA microarrays., and in the development or
receptor transgenic and knockout animal models. There is much cross talk between receptors
and many agonists interact with multiple receptors. Off target effects cannot be predicted by
target specificity. Though RCR can affect transcription directly, much of its effects are exerted
through heterodimeric binging with other nuclear receptors (PXR, CAR, PPARa, PPARy, FXR,
02/12/09
E-245
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
LXR, TR)" (Ulrich, 2003).
Another tool recent developed is gene silencing by introduction of siRNA. Dai et al.
(2007) note issues involved in the siRNA to change gene expression for exploration of MO A etc.
to include the potential of off-target effects, incomplete knockdown, and non-targeting of splice
variants by the selected siRNA sequence. Using knockdown of PPARa in mice, Dai et al. (2007)
report "PPARa knockdown was variable between mice ranging from ~ 80 % knockdown to little
or no knockdown and that differing siRNAs gave different patterns of gene expression with some
grouped with PPARa -/- null mice but others grouped with expression patterns of mice injected
with control siRNA or Ringers buffer alone and showing no PPARa knockdown. Dai et al
concluded that it is possible that it is the change in PPARa levels that is important for perturbing
expression of genes modulated by PPARa rather than the absolute levels of PPARa. Not only is
the finding of variability in knockdowns by siRNA technologies important but The finding that
level of PPAR is not necessarily correlated with function and that it could be the change and not
absolute level that matters in modulation in gene expression by PPARa is of importance as well.
How an animal responds to decreased PPARa function may also depend on its gender. Dai et al.
(2007) observed more dramatic phenotypes in female vs. male mice treated with siRNA and
noted that in aged PPARa -/- mice, Costet et al. (1998) have reported sexually dimorphic
phenotypes including obesity and increased serum triglyceride levels in females, and steatosis
and increased hepatic triglyceride levels in males.
In regard to the emerging science and preliminary reports of the effects of microRNA as
oncogenes and tumor suppressors and of possible importance to hypothesized MO As for liver
cancer, the same caveats as described for DNA microarray analyses all apply along with
additional uncertainties. miRNAs repress their targeted mRNAs by complementary base pairing
and induction of the RNA interference pathway. Zhang et al. (2007) report Northern blot
detection of gene expression at the mRNA level and its correlation with miRNA expression in
cancer cells as well as realtime PCR. These PCR-based analyses quantify miRNA precursors
and not the active mature miRNAs. However, they report that the relationship between pri-
miRNA and mature miRNA expression has not been thoroughly addressed and is critical in order
to use real time PCR analysis to study the function of miRNAs in cancers. They go on to state
that "although Northern Blotting is a widely used method for miRNA analysis, it has some
limitations, such as unequal hybridization efficiency of individual probes and difficulty in
detecting multiple miRNAs simultaneously. For cancer studies, it is important to be able to
compare the expression pattern of all known miRNAs between cancer cells and normal cells.
Thus, it is better to have methods which detect all miRNA expression at a single time."
"Although Northern blot analysis, real-time PCR, and miRNA microarray can detect the
expression of certain miRNAs and determine which miRNAs may be associated with cancer
formation, it is difficult to determine whether or not miRNAs play a unique role in cancers. Also
these techniques cannot directly determine the correlation between mRNA expression levels and
whether the up-regulation or down-regulation of certain miRNAs is the cause of cancer or a
downstream effect of the disease." "Many miRNA genes have been found that are significantly
overexpressed in different cancers. All of them appear to function as oncogenes; however, only
02/12/09
E-246
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
a few of them have been well characterized." Zhang et al. (2007) suggest that bioinformatic
studies indicate that numerous genes are the targets of miR-17-92: more than 600 for miR-19a
and miR-20, two members of the miR-17-92 cluster." Cho et al. (2007) state that "though more
than 530 miRNAs have been identified in human, much remains to be understood about their
precise cellular function and role in the development of diseases." "Although each miRNA can
control hundreds of target genes, it remains a great challenge to identify the accurate miRNA
targets for cancer research." Thus, miRNAs have multiple target so like other transcription
factors may have pleotropic effects that are cell, timing, and context specificity.
Vogelstein and Kinzler (2004) go on to state "in the last decade many important gene
responsible for the genesis of various cancers have been discovered." Most importantly they and
other suggest that pathways rather than individual gene expression should be the focus of study.
As a specific example, Volgelstein and Kinzler note "another example of the reason for focusing
on pathways rather than individual genes has been provided by studies of TP53 tumor-suppressor
gene. The p53 protein is a transcription factor that normally inhibits cell growth and stimulates
cell death when induced by cellular stress. The most common way to disrupt the p53 pathway is
through a point mutation that inactivates its capacity to bind specifically to its cognate
recognition sequence. However, there are several other ways to achieve the same effects,
including amplification of the MDM2 gene and infection with DNA tumor viruses whose
products bind to p53 and functionally inactivate it."
In regard to cellular anchoring for gene expression or pathway alterations associated with
cancer and the importance of "context" of gene expression changes, Vogelstein and Kinzler
(2004) give several examples. "In solid tumors the important of the interactions between stroma
and epithelium is becoming increasingly recognized (e.g., the importance of the endothelial
cell)" "One might expect that a specific mutation of a widely expressed gene would have
identical or at least similar effects in different mammalian cell types. But this is not in general
what is observed. Different effects of the same mutation are not only found in distinct cell types;
difference can even be observed in the same cell types, depending on when the mutation
occurred during the tumorigenic process. The RAS gene mutations provide informative
examples of these complexities. KRAS2 gene mutation in normal pancreatic duct cells seem to
initiate the neoplastic process, eventually leading to the development of pancreatic cancer. The
same mutations occurring in normal colonic or ovarian epithelial cells lead to self-limiting
hyperplastic or borderline lesions that do not progress to malignancy. In many human and
experimental cancers, RAS genes seem to function as oncogenes. But RAS genes can function as
suppressor genes under other circumstances, inhibiting tumorigenesis after administration of
carcinogens to mice. These and similar observation on other cancer genes are consistent with the
emerging notion that signaling molecules play multiple roles at multiple time, even in the same
cell type. However, the biochemical bases for such variations among cancer cells are almost
unknown."
In regard to the major pathways and mediators involved in cancer several investigators
have reported a coherent set that are involved in many types of cancers. Vogelstein and Kinzler
02/12/09
E-247
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
(2004) note that major pathways and mediators include p53, RB, WNT, E-cadherin, GL1, APC,
ERK, RAS:GTP, P13K,SMAD, RTK BAD, BAX, and H1F1. In regard to coherence and site
concordance between animal and human data, the disturbance of a pathway in one species may
result in the different expression of tumor pattern in another but both linked to a common
endpoint of cancer. Thus, pathways rather than a single mutation should be the focus of MO A
and cancer as several actions can be manifested by one pathway or change at one time that lead
to cancer.
Vogelstein and Kinzler (2004) also note that pathways that are common to "cancer" are
also operative in liver cancer where, as a heterogeneous disease, multiple pathways have been
implicated in differing manifestations of this disease. Thus, liver cancer may be an example its
multiple forms are analogous to differing sites being affected by common pathways that are
affected in "cancer." Pathway concordance may not always show up as site concordance as
expression of cancer between species. Liver cancer may be the example where many pathways
can lead a cancer that is characterized by its heterogeneity.
3.1.3. Etiology, Incidence and Risk Factors for HCC
The review article of Farazi and DePinho (2006) provides and excellent summary of the
current state of human liver cancer in terms of etiology and incidence. The 5-year survival rate
of individuals with liver cancer in the United States is only 8.9% despite aggressive conventional
therapy with lethality of liver cancer due in part from its resistance to existing anticancer agents,
a lack of biomarkers that can detect surgically respectable incipient disease, and underlying liver
disease that limits the use of chemotherapeutic drugs. Chen et al. (2002) report that surgical
resection is considered the only "curative treatment" but > 80 of patients have widespread
hepatocellular carcinoma (HCC) at the time of diagnosis and are not candidates for surgical
treatment. Among patients with localized HCC who undergo surgery, 50% suffer a recurrence.
Primary liver cancer is the fifth most common cancer worldwide and the third most common
cause of cancer mortality. Hepatocellular carcinoma (HCC) accounts for between 85% and 90%
of primary liver cancers (El-Serag and Rudolph, 2007). Seitz and Stickel (2006) report that
epidemiological data from the year 2000 indicate that more than 560,000 new cases of HCC
occurred worldwide, accounting for 5.6% of all human cancers and that HCC is the fifth most
common malignancy in men and the eighth in women. Overall, incidence rates of HCC are
higher in males compared to females. In almost all populations, males have higher liver cancer
rates than females, with male:female ratios usually averaging between 2:1 and 4:1 and the largest
discrepancies in rates (>4:1) found in medium-risk European populations (El-Serag and Rudolph,
2007). Experiments show a 2- to 8-fold of control HCC development in male mice as well
supporting the hypothesis that androgens influence HCC progression rather than sex-specific
exposure to risk factors (El-Serag and Rudolph, 2007). El-Serag and Rudolph (2007) also report
that "in almost all areas, female rates peak in the age group 5 years older than the peak age
group for males. In low risk population (e.g., U.S.) the highest age-specific rates occur among
persons aged 75 and older. A similar pattern is seen among most high-risk Asian populations.
In contrast male rats in high-risk African populations (e.g., Gambia) ten to peak between ages 60
02/12/09
E-248
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and 65 before declining, whereas female rates peak between 65 and 70 before declining."
Age adjusted incidence rates for HCC are extremely high in East and Southeast Asia and
in Africa but in Europe, there is a gradually decreasing prevalence from South to North. HCC
incidence rates also vary greatly among different populations living in the same region and vary
by race (e.g., for all ages and sexes in the US, HCC rates are 2 times higher in Asian than in
African Americans, whose rates are 2 times higher than those in whites) ethnic variability likely
to include differences in the prevalence and acquisition time of major risk factors for liver
disease and HCC (El-Serag and Rudolph, 2007). Worldwide HCC incidence rate doubled during
the last two decades and younger age groups are increasingly affected (El-Serag, 2004). The
high prevalence of HCC in Asia and Africa may be associated with widespread infection with
hepatitis B (HBV) and hepatitis C (HCV) but other risk factors include chronic alcohol misuse,
non alcoholic fatty liver disease (NAFLD), tobacco, oral contraceptives, and food contamination
with aflatoxins (Seitz and Stickel, 2006). El-Serag and Rudolph (2007) report HCC to be the
fastest growing cause of cancer-related death in men in the United States with age-adjusted HCC
incidence rates increasing more than 2-fold between 1985 and 2002 and that, overall, 15%-50%
of HCC patients in the United States have not established risk factors.
Although liver cirrhosis is present in a large portion of patients with HCC, it is not always
present. Fattrovich et al (2004) report that "differences of geographic area, method of
recruitment of the HCC cases (medical or surgical) and the type of material studied (liver biopsy
specimens, autopsy, or partial hepatectomies) may account for the variable prevalence of HCC
without underlying cirrhosis (7% to 54%) quoted in a series of studies. Percutaneous liver
biopsy specimens are subject to sampling error. However, only a small proportion of patients
with HCC without cirrhosis have absolutely normal liver histology, the majority of them
showing a range of fibrosis intensity from no fibrosis are all to septal and bridging fibrosis,
necroinflammation, steatosis, and liver cell dysplasia."
Farazi and DePinho (2006) note that for diabetes, a higher indices of HCC has been
described in diabetic patients with no previous history of liver disease associated with other
factors. El-Serag and Rudolph (2007) report that in their study of VA patients (173,643 patients
with and 650,620 patients without diabetes), that HCC incidence doubled among patients with
diabetes and was higher among those with a longer follow-up of evaluation. "Although most
studies have been conducted in low HCC rate areas, diabetes also has been found to be a
significant risk factor in areas of high HCC incidence such as Japan. Taken together, available
data suggest that diabetes is a moderately strong risk factor for HCC."
Non-alcoholic fatty liver disorders (NAFLD) and non-alcoholic steatohepatitis contribute
to the development of fibrosis and cirrhosis and therefore might also contribute to HCC
development. The pathogenesis of NAFLD includes the accumulation of fat in the liver which
can lead to reactive oxygen species in the liver with necrosis factor a (TNFa) elevated in
NAFDL and alcoholic liver disease (Seitz and Stickel 2006). Abnormal liver enzymes not due to
alcohol, viral hepatitis, or iron overload are present in 2.8% to 5.5% of the US general population
02/12/09
E-249
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and may be due to NAFLD in 66% to 90% of cases (Adams and Lindor, 2007). Primary NAFLD
occurs most commonly and is associated with insulin-resistant states, such as diabetes and
obesity with other conditions associated with insulin resistance, such as polycystic ovarian
syndrome and hypopituitarism also associated with NAFLD (Adams and Lindor, 2007). The
steatotic liver appears to be susceptible to further hepatotoxic insults, which may lead to
hepatocyte injury, inflammation, and fibrosis, but the mechanisms promoting progressive liver
injury are not well defined (Adams and Lindor, 2007). Substrates derived from adipose tissue
such as FFA, TNF-a, leptin, and adiponectin have been implicated with oxidative stress
appearing to be important leading to subsequent lipid peroxidation, cytokine induction, and
mitochondrial dysfunction. Liver disease was the third leading cause of death among NAFLD
patients compared to the 13th leading cause among the general population, suggesting that liver-
related mortality is responsible for a proportion of increased mortality risk among NAFLD
patients (Adams and Lindor, 2007).
The relative risk for HCC in type 2 diabetics has been reported to be approximately 4 and
increases to almost 10 for consumption of more than 80 g of alcohol per day (Hassan et al.,
2002). El-Serag and Rudolph (2007) report that "it has been suggested that many cryptogenic
cirrhosis and HCC cases represent more severe forms of nonalcoholic fatty liver disease
(NAFLD), namely nonalcoholic steato hepatitis (NASH). Studies in the United States evaluating
risk factors for chronic liver disease or HCC have failed to identify HCV, HBV, or heavy alcohol
intake in a large proportion of patients (30-40%). Once cirrhosis and HCC are established, it is
difficult to identify pathologic features of NASH. Several clinic-based controlled studies have
indicated that HCC patients with cryptogenic cirrhosis tend to have clinical and demographic
features suggestive of NASH (predominance of women, diabetes, and obesity) as compared with
age- and sex-matched HCC patients of well defined vial or alcoholic etiology. The most
compelling evidence for an association between NASH and HCC is indirect and come from
studies examining HCC risk with 2 conditions strongly associated with NASH: obesity and
diabetes. In a large prospective cohort in the US, followed up for 16 years, liver cancer mortality
rates were 5 times greater among men with the greatest baseline body mass index (range 35-40)
compared with those with a normal body mass index. In the same study, the risk of liver cancer
was not as increase in women, with a relative risk of 1.68. Two other population-based cohort
studies from Sweden and Denmark found excess HCC risk (increased 2- to 3-fold) in obese men
and women compared with those with a normal body mass index." "Finally, liver disease occurs
more frequently in those with more severe metabolic disturbances, with insulin resistance itself
shown to increase as the disease progresses. Several developed countries most notably the
United States, are in the midst of a burgeoning obesity epidemic. Although the evidence linking
obesity to HCC is relatively scant, even small increase in risk related to obesity could translate
into a large number of HCC cases." Thus, even a small increase in risk related to obesity could
result in a large number of HCC cases and the latency of HCC may make detection of increased
HCC risk not detectable for several years.
Other factors are involved as not every cirrhotic liver progresses to HCC. Seitz and
Stickel (2006) suggest that 90 to 100% of those who drink heavily suffer from alcoholic fatty
02/12/09
E-250
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
liver, 10-35% of those evolve to alcoholic steatohepatitis, 8-20% of those evolve to alcoholic
cirrhosis, and 1-2% of those develop HCC. Hepatitis C (HCV) infects approximately 170
million individuals worldwide with approximately 20% of chronic HCV cases developing liver
cirrhosis and 2.5% developing hepatocellular carcinoma (HCC). Infection with hepatitis B
(HVB), a non-cytopathic, partially double stranded hepatotropic DNA virus classified as a
member of the hepadnaviridae family, is also associated with liver cancer risk with several lines
of evidence supporting the direct involvement of HBV in the transformation process (Farazi and
DePinho, 2006). El-Serag and Rudolph (2007) suggest that "Epidemiologic research has shown
that the great majority of adult-onset HCC cases are sporadic and that many have at lease 1
established non-genetic risk factor such as alcohol abuse or chronic HCV or HBV infection.
However, most people with these known environmental risk factors never develop cirrhosis or
HCC, whereas a sizable minority of HCC case develop among individuals without any known
risk factors." "Genetic epidemiology studies in HCC, similar to several other conditions, have
fallen short of early expectations that they rapidly and unequivocally would result in
identification of genetic variants conveying substantial excess risk of disease and thereby
establish the groundwork for effective genetic screening for primary prevention."
3.1.4. Issues Associated with Target Cell Identification
Another outstanding and important question in HCC pathogenesis involves the cellular
origin of this cancer. The liver is made up of a number of cell types showing different
phenotypes and levels of differentiation. Which cell types are targets of hepatocarcinogens and
are those responsible for human HCC is a matter of intense debate. Studies over the last decade
provide evidence of several types of cells in the liver that can repopulate the hepatocyte
compartment after a toxic insult. "Indeed, although the existence of a liver stem cell is often
debated, most experts agree that progenitor liver cells are activated, in response to significant
exposure to hepatotoxins. Also, progenitor cells derived from nonhepatic sources, such as bone
marrow and pancreas, have been demonstrated recently to be capable of differentiating into
mature hepatocytes under correct microenvironmental conditions (Gandillet et al., 2003). At
present, analyses of human HCCs for oval cell markers, comparison of their gene-expression
patterns with rat fetal hepatoblasts and the cellular characteristics of HCC from various animal
models have provided contrasting results about the cellular origin of HCC and imply dual origins
from either oval cells or mature hepatocytes. The failure to identify a clear cell of origin for
HCC might stem from the fact that there are multiple cells of origin, perhaps reflecting the
developmental plasticity of the hepatocyte lineage. The resolution of the HCC cell of origin
issue could affect the development of useful preventative strategies to target nascent neoplasms,
foster an understanding of how HCC-relevant genetic lesions function in that specific cell-
development context and increase our ability to develop more accurate mouse models in which
key genetic events are targeted to the appropriate cellular compartment (Farazi and DePinho,
2006). Two reviews by Librecht (2006) and Wu and Chen (2006) provide excellent summaries
of the issues involved in identifying the target cell for HCC and the review by Roskams et al
(2004) provides a current view of the "oval cell" its location and human equivalent. Recent
reports by Best and Coleman (2007) suggest another type of liver cell is also capable of
02/12/09
E-251
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
proliferation and differentiating into small hepatocytes (i.e., small hepatocyte-like progenitor
cell).
The review by Librecht (2006) provides and excellent description of the controversy and
data supporting different views of the cells of origin for HCC. "In recent years, the results of
several studies suggest that human liver tumors can be derived from hepatic progenitor cells
rather than from mature cell types. The available data indeed strongly suggest that most
combined hepatocellular-cholangiocarcinomas arise from hepatic progenitor cells (HPCs) that
retained their potential to differentiate into the hepatocyte and biliary lineages. Hepatic
progenitor cells could also be the basis for some hepatocellular carcinomas and hepatocellular
adenomas, although it is very difficult to determine the origin of an individual hepatocellular
carcinoma. There is currently not enough data to make statements regarding a hepatic progenitor
cell origin of cholangiocarcinoma. The presence of hepatic progenitor cell markers and the
presence and extent of the cholangiocellular component are factors that are related the prognosis
of hepatocellular carcinomas and combined hepatocellular-cholangiocarcinomas, respectively."
"The traditional view that adult human liver tumors arise from mature cell types has been
challenged in recent decades." "HPCs are small epithelial cells with an oval nucleus, scant
cytoplasm and location in the bile ductules and canals of Hering. HPCs can differentiate towards
the biliary and hepatocytic lineages. Differentiation towards the biliary lineage occurs via
formation of reactive bile ductules, which are anastamosing ductules lined by immature biliary
cells with a relatively large and oval nucleus surrounded by a small rim of cytoplasm.
Hepatocyte differentiation leads to the formation of intermediate hepatocyte-like cells which are
defined as polygonal cells with a size intermediate between than of HPCs and hepatocytes. In
most liver diseases, hepatic progenitor cells are "activated" which means that they proliferate and
differentiate towards the hepatocytic and/or biliary lineages. The extent of activation is
correlated with disease severity." "HPCs and their immediate biliary and hepatocytic progeny
not only have a distinct morphology, but they also express several markers, with many also
present in bile duct epithelial cells. Immunohistochemistry using antibodies against these
markers facilitates the detection of HPCs. The most commonly used markers are cytokeratin
(CK) 19 and CK7." "The proposal that a human hepatocellular carcinoma does not necessarily
arise from mature hepatocyte, but could have HPC origin, has classically been based on three
different observations. Each of them, however, gives only indirect evidence that can be
disputed." "Firstly, it has been shown that HPCs are the cells of origin of HCC in some animal
models of hepatocarcinogenesis, which has led to the suggestion that this might also be the case
in humans. However, in other animal models, the HCCs arise from mature hepatocytes and not
from HPCs or reactive bile ductular cells (Bralet et al 2002; Lin et al 1995- DEN treated rats).
Since it is currently insufficiently clear which of these animal models accurately mimics human
hepatocarcinogenesis, one should be careful about extrapolating data regarding HPC origin of
HCC in animal models to the human situation." "Secondly, liver diseases that are characterized
by the presence of carcinogens and development of dysplastic lesions also show HPC activation.
Therefore, the suggestion has been made that HPCs form a "target population" for carcinogens,
but this is only a theoretical possibility not supported by experimental data." "Thirdly, several
studies have shown that a considerable proportion of HCCs express one or more HPC markers
02/12/09 E-252
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
that are not present in normal mature hepatocytes. Due to the fact that most HPC markers are
also expressed in the biliary lineage, the term "biliary marker" has been used in some of these
studies. The "maturation arrest" hypothesis states that genetic alterations occurring in a HPC, or
its immediate progeny, cause aberrant proliferation and prevent its normal differentiation.
Further accumulation of genetic alterations eventually leads to malignant transformation of these
incompletely differentiated cells. The resulting HCC expresses HPC markers as evidence of its
origin. However, expression of HPC markers can also be interpreted in the setting of the
"dedifferentiation" hypothesis, which suggests that the expression of HPC markers is acquired
during tumor progression as a consequence of accumulating mutations. For example,
experiments in which human HCC cells lines were transplanted into nude mice have nicely
shown that the expression of HPC marker, CK19, steadily increased when the tumors became
increasingly aggressive and metastasized to the lung, Thus, the expression of CK19 in a HCC
does not necessarily mean that the tumor has a HPC origin, but it can also be mutation-induced,
acquired expression associated with tumor progression. Both possibilities are not mutually
exclusive." "For an individual HCC that expresses a HPC marker, it remains impossible to
determine whether this marker reflects the cellular origin and/or is caused by tumor progression.
This can only be elucidated by determining whether HCC contains cells that are ultrastructurally
identical to HPCs in non-tumor liver.
Similarly, the review by Wu and Chen (2006) also presents a valuable analysis of these
issues and state the "The question of whether hepatocellular carcinomas arises from the
differentiation block of stem cells or dedifferentiation of mature cells remains controversial.
Cellular events during hepatocarcinogenesis illustrate that HCC may arise for cells at various
stages of differentiation in the hepatic stem cell lineage." "The role of cancer stem cells has been
demonstrated for some cancers, such as cancer of the hematopoietic system, breast and brain.
The clear similarities between normal stem cell and cancer stem cell genetic programs are the
basis of the a proposal that some cancer stem cells could derived form human adult stem cells.
Adult mesenchymal stem cells (MSC) may be targets for malignant transformation and undergo
spontaneous transformation following long-term in vitro culture, supporting the hypothesis of
cancer stem cell origin. Stem cells are not only units of biological organization, responsible for
the development and the regeneration of tissue and organ systems, but are also targets of
carcinogenesis. However, the origin of the cancer stem cell remains elusive." "Three levels of
cells that can respond to liver tissue renewal or damage have been proved (1) mature liver cells,
as "unipotential stem cells", which proliferate under normal liver tissue renewal and respond
rapidly to liver injury, (2) oval cells, as bipotential stem cells, which are activated to proliferate
when the liver damage is extensive and chronic or if proliferation of hepatocytes is inhibited; and
(3) bone marrow stem cells, as multipotent liver stem cells, which have a very long proliferation
potential. There are two major nonexclusive hypotheses of the cellular origin of cancer; from
stem cells due to maturation arrest or from dedifferentiation of mature cells. Research on hepatic
stem cells in hepatocarcinogenesis has entered a new era of controversy, excitement and great
expectations." "The two major hypotheses about the cellular origination of HCC have been
discussed for almost 20 years. Debate has centered on whether or not HCC originates from the
differentiation block of stem cells or dedifferentiation of mature cells. Recent research suggests
02/12/09
E-253
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
that HCC may originate from the transdifferentiation of bone marrow cells. In fact, there might
be more than one type of carcinogen target cell. The argument about the origination of HCC
becomes much clearer when viewed from this viewpoint: poorly differentiated HCC orginate
from bone marrow stem cells and oval cells, while well-differentiated HCC originates form
mature hepatocytes." "The cellular events during hepatocarcingenesis illustrate that HCC may
arise from cells at various stages of differentiation in the hepatocyte lineage. There are four
levels of cells in the hepatic stem cell lineage: bone marrow cell, hepato-pancreas stem cell, oval
cell and hepatocyte. HSC and the liver are known to have a close relationship in early
development. Bone marrow stem cells could differentiate into oval cells, which could
differentiate into heptaocytes and duct cells. The development of pancreatic and liver buds in
embryogenesis suggests the existence of a common progenitor cells to both the pancreas and
liver. All of the four levels of cells in the stem cell lineage may be targets of
hepatocarcinogenesi s."
Along with the cell types described as possible targets and participants in HCC, Best and
Coleman (2007) describe yet another type of cell in the liver that can respond to hepatocellular
injury which they term small hepatocyte-like progenitor cells and conclude that they are not the
progeny of oval cells, but represent a distinct liver progenitor cell population. Another potential
regenerative cell is the small hepatocyte-like progenitor cell (SHPC). SHPCs share some
phenotypes with hepatocytes, fetal hepatoblasts, and oval cells, but are phenotypically distinct.
They express markers such as albumin, transferring, and AFP and possess bile canaliculi and
store glycogen.
A recent review by Roskams et al. (2004) provides a current view of the "oval cell" its
location and human equivalent. They conclude that "while similarities exist between the
progenitor cell compartment of human and rodent livers, the different rodent models are not
entirely comparable with the human situation, and use of the same term has created confusion as
to what characteristics may be expected in the human ductular reaction. For example, a defining
feature of oval cells in many rodent models of injury is production of alpha-fetoprotein, whereas
ductular reactions in humans rarely display such expression. Therefore we suggest that the "oval
cell" and "oval -like cell" no longer be used in description of human liver."
In the chronic hepatitis and cancer model of Vig et al. (2006) it is not the oval cells or
SHPCs that are proliferating but the mature hepatocytes, thus supporting theories that it is not
only oval cells that are causing proliferations leading to cancer. Vig et al., (2006) also report that
studies in mice an humans indicate that oval cells also may give rise to liver tumors and that oval
cells commonly surround and penetrate human liver tumors, including those caused by hepatitis
B. Tarsetti et al. (1993) suggest that although some studies have suggested that oval cells are
directly involved in the formation of HCC others assert that HCC originates from preneoplastic
foci and nodules derived from hepatocytes and report that HCC evolved in their model of liver
damage from hepatocytes, presumably hepatocellular nodules, and not from oval cells. They
also suggest that proliferation alone may not lead to cancer. Recent studies that follow the
progression of hepatocellular nodules to HCC in humans (see Section 3.2.4 below) suggest an
02/12/09
E-254
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
evolution from nodule to tumor.
3.1.5. Status of Mechanism of Action for Human HCC
The underlying molecular mechanisms leading to hepatocarcinogeneis remain largely
unclear (Yeh et al., 2007). Although HCC is multistep, and its appearance in children suggest a
genetic predisposition exists, the inability to identify most of the predisposing genes and how
their altered expression relates to histological lesions that are the direct precursors to HCC, has
made it difficult to identify the rate limiting steps in hepatocarcingenesis (Feitelson et al., 2002).
Calvisi et al. (2007) report that although the major etiological agents have been identified, the
molecular pathogenesis of HCC remains unclear and that while deregulation of a number of
oncogenes (e.g., c-Myc, cyclinDl and [i-catenin and tumor suppressor genes including P16INK4A,
P53, E-cadherin, DLC-1, andpRb) have been observed at different frequencies in HCC, the
specific genes and the molecular pathways that play pivital roles in liver tumor development
have not been identified. Indeed rather than simple patterns of mutations, pathways that are
common to cancer have been identified through study of tumors and through transgenic mouse
models. Branda and Wands (2006) state that the molecular factors and interactions involved in
hepatocarcinogenesis are still poorly understood but are particularly true with respect to genomic
mutations, "as it has been difficult to identify common genetic changes in more than 20% to 30%
of tumors." As well as phenotypically heterogeneous, "it is becoming clear that HCCs are
genetically heterogeneous tumors." The descriptions of heterogeneity of tumors and of pathway
disruptions common to cancer are also shown for liver tumors (see Sections 3.1.6, and 3.1.8,
below). However, many of these studies focus on the end process and of examination of the
genomic phenotype of the tumor for inferences regarding clinical course, aggressiveness of
tumor, and consistency with other forms of cancer. As stated above, the events that produce
these tumors from patients with conditions that put them at risk,are not known.
El-Serag and Rudolph (2007) suggest that risk of HCC increases with at the cirrhosis
stage when liver cell proliferation is decreased and that acceleration of carcinogenesis at this
stage may result from telomere shortening (resulting in limitations of regenerative reserve and
induction of chromosomal instability), impaired hepatocyte proliferation (resulting in cancer
induction by loss of replicative competition), and altered milieu conditions that promote tumor
cell proliferation. "When telomeres reach a critically short length, chromosome uncapping
induces DNA damage signals, cell-cycle arrest, senescence, or apoptosis. Telomeres are
critically short in human HCC and on the single cell level telomere shortening correlated with
increasing aneuploidy in human HCC." "Chemicals inhibiting hepatocyte proliferation
accelerate carcinogen-induced liver tumor formation in rats as well as the expansion and
transformation of transplanted hepatocytes. It is conceivable that abnormally proliferating
hepatocytes would not expand in healthy regenerating liver but would expand quickly and
eventually transform in the growth restrained cirrhotic liver". "Liver mass is controlled by
growth factors - mass loss through could provide a growth stimulatory macroenviroment. For
the microenvironment, cirrhosis activates stellate cells resulting in increased production of
extracellular matrix proteins, cytokines, growth factors, and products of oxidative stress."
02/12/09
E-255
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Like other cancers, genomic instability is a common feature of human HCC with various
mechanisms thought to contribute, including telomere erosion, chromosome segregation defects,
and alteration in DNA damage-response pathways. In addition to genetic events associated with
the development of HCC (p53 inactivation, mutation in P-catenin, overexpression of ErbB
receptor family members, and overexpression of the MET receptor whose ligand is HGF) various
cancer-relevant genes seem to be targeted on the epigenetic level (methylation) in human HCC
(Farazi and DePinho, 2006). Changes in methylation have been detected in the earliest stages of
hepatocarcinogenesis and to a greater extent in tumour progression (Lee et al., 2003). Seitz and
Stickel (2006) report that aberrant DNA hypermethylation (a silencing effect on genes) may be
associated with genetic instability as determined by the loss of heterozygosity and microsatellite
instability in human HCC due to chronic viral hepatitis and that modifications of the degree of
hepatic DNA methylation have also been observed in experimental models of chronic
alcoholism. Farazi and DePinho (2006) report that two of the key molecules that involved in
DNA damage response, p53 and BRCA2, seem to have roles in destabilizing the HCC genome
(Collin 2005). The inactivation of p53 through mutation or viral oncoprotein sequestration is a
common event in HCC and p53 knock in mouse models containing dominant point mutations
have been shown to cause genomic instability. However, Farazi and DePinho (2006) note that
despite documentation of deletions or mutations in these and other DNA damage network genes,
their direct roles in the genomic instability of HCC have yet to be established in many genetic
model systems.
Telomere shortening has been described as a key feature of chronic hyperproliferative
liver disease (Urabe et al 1996; Miura et al 1997; Rudolf et al 2001: Kitada et al 1995),
specifically occurring in the hepatocyte compartment. These observations have fueled
speculation that telomere shortening associated with chronic liver disease and hepatocyte
turnover contribute to the induction of genomic instability that drives human HCC (Farazi and
DePinho, 2006). Defects in chromosome segregation during mitosis result in aneuploidy, a
common cytogenetic feature of cancer cell including HCC (Farazi and DePinho, 2006).
Several studies have attempted to categorize genomic changes in relation to tumour state.
In general, high levels of chromosomal instability seem to correlate with the de-differentiation
and progression of HCC (Wilkens et al., 2004). Several studies have suggested certain
chromosomal changes to be specific to dysplastic lesions, early -stage and late-stage HCCs, and
metastases. It is important to note that the studies that have attempted to compare genomic
profiles and tumour state are few in number, often did not classify HCCs on the basis of etiology,
and used relatively low-resolution genome-scanning platforms (Farazi and DePinho, 2006).
Farazi and DePinho (2006) note that it should be emphasized that although genome -etiology
correlates reported in some studies, are intriguing, several studies have failed to uncover
significant differences in genomic changes between different etiological groups, although the
outcome might related to small sample sizes and the low-resolution genome -scanning platform
used.
02/12/09
E-256
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
3.1.6. Pathway and Genetic Disruption Associated with HCC and Relationship to Other
Forms of Neoplasia
In their landmark paper, Hanahan and Weinberg (2000) suggested that the vast catalog of
cancer cell genotypes were a manifestation of six essential alterations in cell physiology that
collectively dictate malignant growth; self-sufficiency in growth signals, insensitivity to growth
-inhibitory (antigrowth signals), elevation of programmed cell death (apoptosis), limitless
replication potential, sustained angiogenesis, and tissue invasion and metastasis. They proposed
that these six capabilities are shared in common by most and perhaps all types of human tumors
and, while virtually all cancers must acquire the same six hallmark capabilities, their means of
doing so would vary significantly, both mechanistically and chronologically. It was predicted
that in some tumors, a particular genetic lesions may confer several capabilities simultaneously,
decreasing the number of distinct mutational steps required to complete tumorigenesis. Loss of
the p53 tumor suppressor was cited as an example that could facilitate both angiogenesis and
resistance to apoptosis and to enable the characteristic of genomic instability. The paths that cells
could take on their way to becoming malignant were predicted to be highly variable, and within
a give cancer type, mutation of a particular target genes such as ras or p53 could be found only
in a subset of otherwise histologically identical tumors. Furthermore, mutations in certain
oncogenes and tumor suppressor genes could occur early in some tumor progression pathways
and late in others. Genes known to be functionally altered in "cancer" were identified as
including Fas,Bcl2, Decoy R, Bax, Smads, TFGpR, pl5, pl6, Cycl D, Rb, HPV E7, ARF,
PTEN, Myc, Fos, Jun, Ras, Abl, NF1, RTK, TGFa, Integrins, E-cadherin, Src, P-catenin, APC,
and WNT.
Branda and Wands (2006) report that two signal transduction cascades that appear to be
very important are insulin/IFG-l/IRS-l/MAPK and Wnt/Frizzled/ P-catenin pathways which are
activated in over 90% of HCC tumors (Branda and Wands, 2006). Feitelson et al. (2002)
reported that "In addition to NF-kB, up-regulated expression of rhoB has been reported in some
HCCs. RhoB is in the ras gene family, is associated with cell transformation, and may be a
common denominator to both viral and non-viral hepatocarcingenesis. Activation of ras and NF-
kB, combined with down regulation of multiple negative growth regulatory pathways, then, may
contribute importantly to early steps in hepatocarcinogenesis. Thus viral proteins may alter the
patterns of hepatocellular gene expression by transcriptional trans-regulation." "Another early
event appears to involve the mutation of P-catenin, which is a component of the Wnt signal
transduction pathway whose target genes include c-myc, c-jun, cyclin Dl, fibronectin, the
connective tissue growth factor WISP, and matrix metaolloproteinases." Boyault et al (2007)
report that "altogether, the principle carcinogenic pathways known to be deregulated in HCC are
inactivation of TP53, Wnt/wingless activation mainly through CTNNB1 mutations activating P-
catenin- and AXIN1-inactivating mutations, retinoblastoma inactivation through RBI and
CDKN2A promoter methylation and rare gene mutations, insulin growth factor activation
through IGF2 overexpression, and IGF2R-inactiving mutations."
02/12/09
E-257
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
El-Serag and Rudolph suggest that "in general, the activation of oncogenic pathways in
human HCC appears to be more heterogeneous compared with other cancer types." El-Serag
and Rudolph (2007) report that the p53 pathway is a major tumor-suppressor pathway that (1)
limits cell survival and proliferation (replicative senescence) in response to telomere shortening
(2) induces cell-cycle arrest in response to oncogene activation (oncogene-induced senescence),
(3) protects genome integrity, and (4) is affected at multiple levels in human HCC. "p53
mutations occur in aflatoxin induced HCC (>50%) and with lower frequency (20-40%) in HCC
not associated with aflatoxin." In addition, "the vast majority of human HCC overexpresses
gankyrin, which inhibits both Rb checkpoint and p53 checkpoint function." "The pl6/Rb
checkpoint is another major pathway limiting cell proliferation in response to telomere
shortening, DNA damage, and oncogene activation. In human HCC the Rb pathway is disrupted
in more than 80% of cases, with repression of pl6 by promoter methylation being the most
frequent alteration. Moreover, expression of gankyrin (an inhibitor of p53 and Rb checkpoint
function) is increased in the vast majority of human HCCs, indicating that the Rb checkpoint is
dysfunctional in the vast majority of human HCCs." "The frequent inactivation of p53 in human
HCC indicates that abrogation of p53-dependent apoptosis could promote hepatocarcinogenesis.
The role of impairment of p53-independent apoptosis for hepatocarcinogenesis remains to be
defined." "Activation of the P-catenin pathway frequently occurs in mouse and human HCC
involving somatic mutations, as well as transcriptional repression of negative regulators. An
activation of the Akt signaling and impaired expression of phosphatase and tensin homolog
(PTEN) (a negative regulator of Akt) have been reported in 40-60% of Human HCC." They
suggest that although Myc is a potent oncongene inducing hepatocarcinogenesis in mouse models
the data on human HCC are heterogeneous and further studies are required.
3.1.7. Epigenetic Alterations in HCC
The molecular pathogenesis of HCC remains largely unknown but it is presumed that the
development and progression of HCC are the consequence of cumulative genetic and epigenetic
events similar to those described in other solid tumors (Calvisi et al., 2006). Calvisi et al (2007)
provide a good summary of DNA methylation status and cancer as well as it's status in regard to
HCC:
Aberrant DNA methylation occurs commonly in human cancers in the forms of genome-
wide hypomethylation and regional hypermethylation. Global DNA hypomethylation
(also known as demethylation) is associated with activation of protooncogenes, such as c-
Jun, c-Myc, and c-HA-Ras, and generation of genomic instability. Hypermethylation on
CpG islands located in the promoter regions of tumor suppressor genes results in
transcriptional silencing and genomic instability. CpG hypermethylation (also known as
de novo methylation) acts as an alternative and/or complementary mechanisms to gene
mutations causing gene inactivation, and it is now recognized as an important mechanism
in carcinogenesis. Although the mechanism(s) responsible for de novo methylation in
cancer are poorly understood, it has been hypothesized that epigenetic silencing depends
on activation of a number of proteins known as DNA methyltransferases (DNMTs) that
posses de novo methylation activity. The importance of DNMTs in CpG methylation
02/12/09
E-258
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
was substantiated by the observation that genetic disruption of both DNMT1 and
DNMT3b genes in HCT116 cell lines nearly eliminated methyltransferase activity.
However, more recent findings indicate that the HCT116 cells retain a truncated,
biologically active form of DNMT1 and maintain 80% of their genomic methylation.
Further reduction of DNMT1 levels by an siRNA approach resulted in decreased cell
viability, increased apoptosis, enhanced genomic instability, checkpoint defects, and
abrogation of replicative capacity. These data show that DNTM1 is required for cell
survival and suggest that DNTM1 has additional functions that are independent of its
methyltransferase activity. Concomitant overexpression of DNMT1, -3A, and -3b has
been found in various tumors including HCC. However, no changes in the expression of
DNMTs were found in other neoplasms, such as colorectal cancer, suggesting the
existence of alternative mechanisms. In HCC, a novel DNMT3b splice variant, known as
DNMT3b4 is overexpressed. DNMT3b4 lacks DNMT activity and competes with
DNMT2b3 for targeting of pericentromeric satellite regions in HCC, resulting in DNA
hypomethylation of these regions and induction of chromosomal instability, further
linking aberrant methylation and generation of genomic alterations.
It is now well accepted that methylation changes occur early and ubiquitously in cancer
development. The case has been made that tumor cell heterogeneity is due, in part, to
epigenetic variation in progenitor cells and that epigenetic plasticity together with genetic
lesions drive tumor progression (Feinberg et al., 2006).
A growing number of genes undergoing aberrant CpG island hypermethylation in HCC
have been discovered, suggesting that de novo methylation is an important mechanism
underlying malignant transformation in the liver. However, most of the previous studies
have focused on a single or a limited number of genes, and few have attempted to analyze
the methylation status of multiple genes in HCC and associated chronic liver diseases. In
addition, the functional consequence(s) of global DNA hypomethylation and CpG island
hypermethylation in human liver cancer has not been investigated to date. Furthermore,
to our knowledge no comprehensive analysis of CpG island hypermethylation involving
activation of signaling pathways has been performed.
Calvisi et al. (2007) report that global gene expression profiles show human HCC to
harbor common molecular features that differ greatly from those of non-tumorous surrounding
tissues, and that human HCC can be subdivided into 2 broad but distinct subclasses that are
associated with length of patient survival. They further suggest that aberrant methylation is a
major event in both early and late stages of liver malignant transformation and might constitute a
critical target for cancer risk assessment, treatment, and chemoprevention of HCC. Calvisi et al
(2007) conducted analysis of methylation status of genes selected based on their capacity to
modulate signaling pathways (Ras, Jak/Stat, Wingless/Wnt, and RELN) and/or biologic features
of the tumors (proliferation, apoptosis, angiogenesis, invasion, DNA repair, immune response,
and detoxification). Normal livers were reported to show the absence of promoter methylation
for all genes examined. At least 1 of the genes involved in inhibition of Ras (ARH1, CLU,
DAB2, HDAB21P, HIN-1, HRASL, LOX, NORE1A, PAR4, RASSF1A, RASSF2, RASSF3,
RASSF4, RIG, RRP22, and SPRY2 and-4), Jak/Stat (ARH1,CIS, SHP1, PIAS-1, PIAS-y, SOCS1,
02/12/09
E-259
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
-2, and -3, SYK, and GRIM-19), and Wnt/p-catenin (ARC, E-cadherin, y-catenin, SFRP1, -2, -4,
and -5, DKK-1 and -3, WIF-1 andHDPR1) pathways was affected by de novo methylation in all
HCC. A number of these genes were also reported to be highly methylated in the surrounding
non-tumorous liver. In contrast, inactivation of at least 1 of these genes implicated in the RELN
pathway (DAB1, reelin) was detected differentially in HCC of subclasses of tumor that had
difference in tumor aggressiveness and progression. Epigenetic silencing of multiple tumor
suppressor genes maintains activation of the Ras pathway with a major finding in the Calvisi et
al (2007) study to be the concurrent hypermethylation of multiple inhibitors of the Ras pathway
with Ras was significantly more active in HCC than in surrounding or normal livers. Also
important, was the finding of no significant association between methylation patterns and
specific etiologic agents (i.e., HVB, HVC, ethanol, etc) was detected further substantiating the
conclusion that aberrant methylation is a ubiquitous phenomenon in hepatocarcingenesis.
"Current evidence suggests that hypomethylation might promote malignant transformation via
multiple mechanisms, including chromosome instability, activation of protooncogenes,
reactivation of transposable elements, and loss of imprinting." "The degree of DNA
hypomethylation progressively increased from non-neoplastic livers to fully malignant HCC,
indicating that genomic hypomethylation is an important prognostic factors in HCC, as reported
for brain, breast, and ovarian cancer." Calvisi et al. (2007) also report that regional CpG
hypermethylation was also enhanced during the course of HCC disease and that the study of
tumor suppressor gene promoters showed that CpG methylation was frequently detected both in
surrounding non-tumorous livers and HCC.
3.1.8. Heterogeneity of Preneoplastic and HCC Phenotypes
A very important question for the treatment of HCC in humans is early detection.
Research has focused on identification of lesions that will progress to HCC and to also determine
from the phenotype of the nodule and genetic expression its cell source, likely survival, and
associations with etiologies and MO As. As with rodent models where preneoplastic foci have
been observed to be associated with progression to adenoma and carcinoma, nodules observed in
humans with high risk for HCC have been observed to progress to HCC. In humans
histomorphology of HCC is notoriously heterogeneous (Yeh et al., 2007). Although much
progress has been made, there is currently not universally accepted staging system for HCC
partly because of the natural course of early HCC is unknown and the natural progression of
intermediated and advanced HCC are quite heterogeneous (Thorgeirsson, 2006). Nodules are
heterogeneous as well with differences in potential to progress to HCC. Chen et al (2002) report
that standard clinical pathological classification of HCC has limited valued in predicting the
outcome of treatment as the phenotypic diversity of cancer is accompanied by a corresponding
diversity in gene expression patterns. There is also histopathological variability in the
presentation of HCC in geographically diverse regions of the world with some slow growing,
differentiated HCC nodules surrounded by a fibrous capsule are common among Japanese but, in
contrast, a"febrile" form of HCC, characterized by leukocytosis, fever, and necrosis within a
poorly differentiated tumor to be common in South African blacks (Feitelson et al., 2002).
02/12/09
E-260
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
A multistep process is suggested histologically, where HCC appears within the context of
chronic hepatitis and/or cirrhosis within regions of the liver cell dsyplasia or adenomatous
hyperplasia (Feitelson et al., 2002). Kobayashi et al. (2006) report that the higher the grade of
the nodule the higher the percentage that will progress to HCC with 18.8% of all nodules and
regenerative lesions going on to become HCC, 53.3% remaining unchanged, and 27.9%
disappearing in the observation period of 0.1 to 8.9 years. Borzio et al (2003) report that the rate
of liver malignant transformation was 40% in larger regenerative nodules, low-grade dysplastic,
and high-grade dysplastic nodules with higher grade of dysplasia extranodular detection of large
cell change and hyperchroic pattern associated with progression to HCC. Yeh et al (2007) report
that nuclear staining for Ki-67 and Topo II-a (a nuclear protein targeted by several
chemotherapeutic agents) significantly increased in the progression from cirrhosis, through high
grade dysplastic nodules to HCC whereas the scores for TGF-a in these lesions showed an
inverse relationship. "In comparison with 18 HCC arising in noncirrhotic livers, the expression
of TGF-a is significantly stronger in cirrhotic liver than in noncirrhotic parenchyma and its
expression is also stronger in HCC arising in cirrhosis than in HCC arising in noncirrhotic
patients." They concluded that initiation in cirrhotic and noncirrhotic liver may have different
pathways with Transforming growth factor-a (a mitogen activated the EFGR) playing a relative
more important role in HCC from cirrhotic liver. Over expression of TGF-a in the liver of
transgenic mice induced increased proliferation, dysplasia, adenoma and carcinoma. Yeh et al.
(2007) concluded that such high-grade dysplastic nodules are precursor lesions in
hepatocarcingenesis and that TGF-a may play an important role in the early events of liver
carcinogenesis.
Moinzadeh et al. (2005) reported in a meta-analysis of all available (n = 785 HCCs) that
gains and losses of chromosomal material were most prevalent in a number of chromosomes and
that amplifications and deletions occurred on chromosomal arms in which oncogenes (e.g. MYC
and 8q24) and tumor suppressor genes (e.g., RBI on 13ql4) are located as well a modulators of
the WNT-signaling pathway. However, in multifocal HCC, nodules arising de novo within a
single liver have a different spectrum of genetic lesions. "Hence, there are likely to be many
paths to hepatocellular carcinoma, and this is why it has been difficult to assign specific
molecular alterations to changes in hepatocellular phenotype, clinical, or histopathological
changes that accompany tumor development" (Feitelson et al., 2002).
Serum alpha-fetoprotein (AFP) is commonly used as tumor marker for HCC. Several
reports have linked HCC to cytokines in an attempt to find more specific markers of HCC. Jia et
al (2007) report that AFP marker allows for identification of a small set of HCC patients with
smaller tumors, and these patients have a relatively long-term survival rate following curative
treatment. "Presently the only approach to screen for the presence of HCC in high-risk
populations is the combination of serum AFP and ultrasonagraphy. However, elevated AFP is
only observed in about 60 to 7-% of HCC patients and to a lesser extent (33-65%) in patients
with smaller HCCs. Moreover, nonspecific elevation of serum AFP has been found in 15% to
58%) of patients with chronic hepatitis and 11%> to 47% of patients with liver cirrhosis." Soresi et
al (2006) report that Serum IL-6 levels are low in physiological conditions, but increase
02/12/09 E-261
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
considerably pathological conditions such as trauma, inflammation and neoplasia. In tumors IL-
6 may be involved in promoting the differentiation and growth of target cells. "Many works have
reported high serum IL-6 levels in various lifer diseases such as acute hepatitis, primary biliary
cirrhosis, chronic hepatitis (hepatitis C) and HCV-correlated liver cirrhosis and in hepatocellular
carcinoma." Soresi et al (2006) report that patients with HCC group had higher IL-6 values than
those with cirrhosis and that "higher-staged" patients had the highest IL-6 levels. Hsia et al
(2007) also examined IL-6, IL-10 and hepatocyte growth factor (HGF) as potential markers for
HCC. "The expression of IL-6 or IL-10 or higher level of HGF or AFP was observed only 0-3%
of normal subjects. Patients with HCC more frequently had higher IL-6 and IL-10 levels, where
as HGF levels in HCC patients were not significantly elevated compared to patients with chronic
hepatitis or non-HCC tumors (but greater than controls). Among patients with low AFP level,
IL-6 or IL-10 expression was significantly associated with the existence of HCC. Patients with
large HCC (>5 cm) more often had increased IL-6, IL-10 or AFP levels. Serum levels of IL-6
and IL-10 are frequently elevated in patients with HCC but not in benign liver disease or non-
HCC tumors."
Nuclear DNA content and ploidy have also been the subjects of several studies through
the years for identification of pathways for prediction of survival or origin of tumors. Nakajima
et al. (2004) report that p53 loss can contribute to the propagation of damaged DNA in daughter
cells through the inability to prevent the transmission of inaccurate genetic material, considered
to be one of the major mechanisms for the emergence of aneuploidy in tumors with inactivated
p53 protein and the increasing ploidy in HCC was associated with disturbance in p53. McEntee
et al. (1991) reported that specimens from 74 patients who underwent curative resection for
primary HCC and analyzed for DNA content, [i.e., tumors were classified as DNA aneulploid if
a separate peak was present from its standard large diploid peak (2C) and tetaploid peak (4C)]
were 33% were DNA diploid, 30% were DNA tetraploid/ polyploidy, and 37% were aneuploid
of the primary tumors examined. Non-tumor controls were diploid and survival was not
different between patients with diploid vs. nondiploid tumors. Zeppa et al (1998) reported ploidy
in 84 hepatocellular carcinomas diagnosed by fine-needle aspiration biopsy to have 68 cases
were aneuploid and 16 euploid (9 diploid and 7 polyploid) with median survival of 38 months for
patients with diploid HCC and 13 months for aneuploid HCC. Lin et al (2003) report in their
study of fine needle aspiration of HCC that "the ratio of S and G2/M periods of DNA, which
reflect cell hyperproliferation, in the group with HCC tumors> 3cm in diameter were markedly
higher than those of the group with nodules< 3 cm in diameter and the group with hyperplastic
nodules." "DNA analysis of aspiration biopsy tissues acquired from intrahepatic benign
hyperplastic nodules showed steady diploid (2c) peak that stayed in G1 period. DNA analysis of
aspiration biopsy tissues acquired from HCC nodules showed S period of hyperproliferation and
G2/M period. The DNA analysis of HCC nodules showed aneuploid peak." They concluded that
the biological behavior of the cell itself that the normal tissue, reactive tissue and benign tumor
all have normal diploid DNA but like most other malignant tumors, HCC appears to have
polyploid DNA, especially aneuploid DNA." Attallah et al. (1999) report small needle liver
biopsy data to show HHC to be 21.4% diploid, 50% aneuploid and 28.6% tetraploid and that
higher ploidies (aneuploid and tetraploid) were observed in human liver cancer than residual
02/12/09
E-262
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
tissues, although in some cases there was increased aneuploidy (cirrhosis, 37%, hepatitis ~ 50%).
Of note for the study is the lack of appropriate control tissue and uncertainty as to how some of
their diploid cells could have been binucleate tetraploid cells. Anti et al, (1994) reported
reduction in binuclearity in the chronic hepatitis and cirrhosis groups that was significantly
correlated with a rise in the diploid/polyploidy ratio and that precancerous and cancerous nodules
within cirrhotic liver show an increased tendency toward diploidy or the emergence of aneuploid
populations. They note that a number of investigators have noted significantly increased
hepatocyte diploidization during the early stages of chemically induced carcinogenesis in rat
liver, but other experimental findings indicate that malignant transformation can occur after any
type of alteration in ploidy distribution. On the other hand, Melchiorri et al (1994) note that
several studies using flow cytometric or image cytometric methods reported high DNA ploidy
values in 50-11% of the examined HCCs and that the presence of aneuploidy was significantly
related to a poor patient prognosis. They report that the DNA content of mononucleated and
binucleated hepatocytes obtained by ultrasound-guided biopsies of 10 macroregenerative nodules
without histologic signs of atypia the lesions with the greater fraction of fraction of
mononucleated hepatocytes were diagnosed as HCCs during the clinical follow-up with results
also suggesting that diploid and tetraploid stem cell lines are the main lines of the HCCs as well
as a reduction in the percentage of binucleated hepatocytes in HCC. Gramantieri et al (1996)
report that the percentage of binucleated cells was reduced in most of HCC they studied (i.e., the
mean percentage of binucleated cells 9% in comparison to 24% found in normal liver) and that
most HCC, as many other solid neoplasms, showed altered nuclear parameters.
Along with reporting pathways that are perturbed in HCC, emerging evidence also shows
that signatures of pathway are predictive of clinical characteristics of HCC. A number of studies
have examined gene expression in tumors to try to determine which pathways may have been
disturbed in an attempt to predict survival and treatment options for the patients and to
investigate possible MO As for the tumor induction and progression. Chen et al. (2002)
described a systematic characterization of gene expression patterns in human liver cancers using
cDNA microarrays to study tumor and non-tumor liver tissues in HCC patients and of note did
quality assurance on their microarray chips (many studies do not report that they have done so)
and examined the effects of hepatitis virus on its subject and identified people with it. Most
importantly, Chen et al (2002) provided phenotypic anchoring of each tumor with its genetic
profile rather than pooling data. The hierarchical analysis demonstrated that clinical samples
could be divided into two major clusters, one representing HCC samples and the other with a few
exceptions, representing non-tumor liver tissues. Most importantly, expression patterns varied
significantly among the HCC and non-tumor liver samples and that samples from HBV-infected,
hepatitis C virus infected, and noninfected individuals were interspersed in the HCC branch.
Thus, tumors from people infected with HVB, HVC and noninfected people with HCC were
interspersed in the HCC pattern and could be discerned based on etiology. One cluster of genes
was highly expressed in HCC samples compared with non-tumor liver tissues included a
"proliferation cluster" comprised of genes whose functions are required for cell cycle
progression and whose expression levels correlate with cellular proliferation rates with most of
the genes in this cluster are specifically expressed in the G2/M phase. Gene profiles for HCC
02/12/09
E-263
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
were consistent with fewer molecular features of differentiated normal hepatocytes. Chen et al
(2002) noted that both normal and liver tumors are complex tissue compose of diverse cells and
that distinct patterns of gene expression seemed to provide molecular signatures of several
specific cell types including expression of two clusters of genes associated with T and B
lymphocytes, presumably reflecting lymphocytic infiltration into liver tissues, and genes
associated with stellate cell activation. This important finding acknowledges that HCC are not
only heterogeneous in hepatocyte phenotype but are made up of many other nonparenchymal cell
types and that gene expression patterns reflect that heterogeneity. A gene cluster was also
identified at a higher level in HCC that included several genes typically expressed in endothelial
cells, including CD34, which is expressed in endothelial cells in veins and arteries but not in the
endothelial cells of the sinusoids in non-tumor liver and which mar reflect disruption of the
molecular program that normally regulate blood vessel morphogenesis in the liver.
Of great importance was the investigation by Chen et al. (2002) of whether samples from
multiple sites in a single HCC tumor, or multiple separate tumor nodules in one patient, would
share a recognizable gene expression signature. With a few instructive exceptions, all the tumor
samples from each patient clustered were reported to cluster together. To further examine the
relationship among multiple tumor samples from individual patients, we calculated the pairwise
comparison for all pairs of samples and samples some primary tumors multiple times. Tumor
patterns of gene expression were more highly correlated those seen in samples from the same
patient than other patient but every tumor had a distinctive and characteristic gene expression
pattern, recognizable in all samples taken from different areas of the same tumor. For multiple
discrete tumor masses obtained from six patients, three of these patients had multiple tumors
with a shared distinctive gene expression pattern but in three other patients, expression patterns
varied between tumor nodules and the difference provided new insights into the sources of
variation in molecular and biological characteristics of cancers. Thus, in some patients multiple
tumors were from the same clone, as demonstrated by a similar gene expression profile, but for
some patients multiple tumors were arising from differing clones within the same liver. In
regard to whether the distinctive expression patterns characteristic of each tumor reflect the
individuality of the tumor, or are they determined by the patient in whom the tumor arose?,
analysis of the expression patterns observed in the two tumor nodules from one patient showed
that the two tumors were not more similar than those of an arbitrary pair of tumors from different
patients. These results show the heterogeneity of HCC and that "one gene pattern" will not be
characteristic of the disease.
However, HCC did have a pattern that differed from other cancers. Chen et al (2002)
analyzed the expression patterns of 10 randomly selected HCC samples and 10 liver metastases
of other cancers and reported that the HCC samples and the metastatic cancers clustered into two
distinct groups, based on difference in their patterns of gene expression. Although some of the
HCC samples were poorly differentiated and expressed the genes of the liver-specific cluster at
very low levels compared to with either normal liver or well-differentiated HCC, the genes of the
liver-specific cluster were reported to be consistently expressed at higher levels in HCC than in
tumors of nonliver origin with metastatic cancers originating from the same tissue typically
02/12/09
E-264
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
clustered together, expressing genes characteristic of the cell types of origin. Thus, liver cancer
was distinguishable from other cancer even though very variable in expression and
differentiation state.
In attempt to create molecular prognostic indices that can be used for identification of
distinct subclasses of HCC that could predict outcome, Lee et al (2004a) report two subclasses of
HCC patients characterized by significant differences in the length of survival and also identified
expression profiles of a limited number of genes that accurately predicted the length of survival.
Total RNAs from the 19 normal livers, including "normal liver in HCC patients," were pooled
and used as a reference for all microarray experiments and thus variations between patients and
especially differences due to conditions predisposing HCC, were not determined. DNA
microarray data using hierarchical clustering was reported to yield two major clusters, one
representing HCC tumors, and the other representing non-tumor tissues with a few exceptions
that were not characterized by the authors. Lee et al (2004a) report that along with 2 distinctive
subtypes of gene expression patterns in HCC there was heterogeneity among HCC gene
expression profiles and that one group had an overall survival time of 30.8 months and the other
83.7 months. Only about half the patients in each group were reported to have cirrhosis.
Expression of typical cell proliferation markers such as PCNA and cell cycle regulators such as
CDK4, CCNB1, CCNA2 and CKS2 was greater in one class than the other of HCC.
The report by Boyault et al. (2007) attempted to compare etiology and genetic
characterization of the tumors they produce and confirms the heterogeneity of HCC, some
without attendant genomic instability. Boyault et al. (2007) reported that genetic alterations are
indeed closely associated with clinical characteristics of HCC that define 2 mechanisms of
hepatocarcinogenesis. "The first type of HCC was associated with not only a high level of
chromosome instability and frequent TP53 and AXIN1 mutations but also was closely linked to
HBV infections and a poor prognosis. Conversely, the second subgroup of HCC tumors was
chromosome-stable, having a high incidence of activating P-catenin alteration and was not
associated with viral infection." Boyault et al (2007) reported that in a series of 123 tumors,
mutations in the CTNNB1 (encoding P-catenin), TP53, ACINI, TCF1, PIK3CA and KRAS
genes in 34, 31, 13, 5, 2 and 1 tumors were identified, respectively. No mutations were found in
NRAS, HRAS, and EGFR. Hypermethylation of the CDKN2A and CDH1 promoter was
identified in 35% and 16% of the tumors, respectively. Boyault et al (2007) grouped tumors by
genomic expression as well as other factors. HCC groups associated with high rate of
chromosomal instability were reported to be enriched with over expression of cell-cycle/
proliferation/DNA metabolism genes. They concluded that "the primary clinical determinant of
class membership is HBV infection and the other main determinants are genetic and epigenetic
alterations, including chromosome instability, CTNNB1 and TP53 mutations, and parental
imprinting. Tumors related to HCV and alcohol abuse were interspersed across subgroups G3-
G6. Boyault et al (2007) suggested that there results indicate that HBV infection early in life
leads to a specific type of HCC that has immature features with abnormal parental gene
imprinting selections, possibly through the persistence of fetal hepatocytes or alternatively
through partial dedifferentiation of adult hepatocytes. "These G1 tumors are related to high-risk
02/12/09
E-265
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
populations found in epidemiological studies."
3.2. Animal Models of Liver Cancer
There are obvious differences between rodents and primate and human liver, and there is
a difference in background rates of susceptibility to hepatocarcinogenesis. With strains of mice
there are large differences in responses to hepatotoxins (e.g., acetaminophen) and to
hepatocarcinogens as well as background rates of hepatocarcinogenecity. Maronpot (2007)
reports that modulators of murine hepatocarcinogenesis, such as diet, hormones, oncogenes,
methylation, imprinting, and cell proliferation/apoptosis are among multiple mechanistically
associated factors that impact this target organ response in control as well as in treated mice, and
suggests that there is no one simple paradigm to explain the differential strain sensitivity to
hepatocarcinogenesis. Because of the variety of studies with differing protocols used to generate
susceptibility data, direct comparisons among strains and stocks is problematic but in regard to
susceptibility to carcinogenicity the C3H/HeJ and C57BL/6J mouse have been reported to have
up to a 40-fold difference in liver tumor multiplicity (Maronpot, 2007). However, as noted
above, TCE causes liver tumors in C6C3F1 and Swiss mice with studies of trichloroethylene
metabolites dichloroacetic acid, trichloroacetic acid, and chloral hydrate suggesting that both
dichloroacetic acid and trichloroacetic acid are involved in trichloroethylene-induced liver
tumorigenesis. Many effects reported in mice after dichloroacetic acid exposure are consistent
with conditions that increase the risk of liver cancer in humans and can involve GST Xi, histone
methylation, and overexpression of IGF2 (Caldwell and Keshava, 2006). The heterogeneity of
liver phenotype observed in mouse models is also consistent with human HCC. These data lend
support to the qualitative relevance of the mouse model for TCE-induced cancer risk.
Bannasch et al. (2003) made important observations that have implications regarding the
differences in susceptibility between rodent and human liver cancer. They stated that "Although
the classification of such nodular liver lesions in rodents as hyperplastic or neoplastic has
remained controversial, persistent nodules of this type are considered neoplasms, designated as
adenomas. In human pathology, the situation appears to be paradoxical because adenomas are
only diagnosed in the noncirrhotic liver, yet a confusing variety terms avoiding the clearcut
classification as an adenoma has been created for nodular lesions in liver cirrhoses, not
withstanding that the vast majority hepatocellular carcinomas develop in cirrhotic livers. Even
if a portion of these nodular lesions would be regarded as adenomas, being integrated into an
adenoma-carcinoma sequence as observed in many animal experiments, clinical and
epidemiological records of liver neoplasms, including both benign and malignant forms, would
increase considerably. This would not only bring hepatic neoplasia further into focus of human
neoplasia in general, but also shed new light on the classification of some chemicals producing
high incidence of liver neoplasms in rodents, but appearing harmless to humans according to
epidemiological evaluations solely based on the incidence of hepatocellular carcinoma in
exposed populations." Thus, that in humans only HCCs are recorded but in animals adenomas
are counted as neoplasms, may indicate that the scope of the problem of liver cancer in humans
may be underestimated.
02/12/09
E-266
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Tumor phenotype differences have been reported for several decades through the work of
Bannasch et al. The predominant cell line of foci of altered hepatocytes (FAH) have excess
glycogen storage early in development that appears to be similar to that shown by DCA
treatment. Bannasch et al. (2003) report that "the predominant glycogenotic-basophilic cell line
FAH reveals that there is an overexpression of the insulin receptor, the IGF-1 receptor, the
insulin receptor substrates-1/2 and other components of the insulin-stimulated signal transduction
pathway." Bannasch states that foci of this type have increased expression of GSTpi and insulin
has also been shown to induce the expression of GSTpi but that hyperinsulin-induced foci do not
show increased GSTpi. Cellular dedifferentiation during progression from glycogenotic to
basophilic cell populations is associated with downregulation in insulin signaling. The
amphophilic-basophilic cell lineage of peroxisome proliferators and hepadnaviridae were
reported to have foci that mimic effects of thyroid hormone with mitochondrial proliferation and
activation of mitochondrial enzymes. Bannasch et al. (2003) state that "the unequivocal
separation of 2 types of compounds, usually classified as initiators and promoters, remains a
problem at the level of the foci because at least the majority of chemical hepatocarcinogens seem
to have both initiating and promoting activity, which may differ in quantitative rather than
qualitative terms from one compound to another." "Whereas genetic mutations have been
predominantly postulated to initiate hepatocarcinogenesis for many years, more recently
epigenetic changes have been increasingly discussed as a plausible cause of the evolution of
preneoplastic foci characterized by metabolic changes including the expression of GSTpi."
Su and Bannasch (2003) report that glycogen-storing foci represents early lesion with the
potential to progress to more advance glycogen-poor basophilic lesions through mixed cell foci
and resulting hyperproliferative lesions and are associated with HCC in man. Small-cell change
(SCC) of liver parenchyma (originally called liver cell dysplasia of small cell size) is reported to
share cytological and histological similarities to early well defined HCC. Close association
between SCC and more advanced (basophilic) foci indicates that foci often progress to HCC
through SCC in humans. SCC were reported to be present in all basophilic foci. Previous
studies were cited that showed that the biochemical phenotype of human FAH, mainly including
glycogen storing clear cell foci and clear cell-predominated mixed cell foci, were observed in
more than 50% of cirrhotic livers with or without HCC. FAH of clear and mixed cell types were
observed in almost all livers bearing HCC, and in chronic liver diseases without HCC but at a
lower frequency. Su and Bannasch (2003) report that "the finding of mixed cell foci (MCF)
mainly in livers with high-risk or cryptogenenic cirrhosis indicates that these are more advanced
precursor lesions in man, in line with earlier observations in experimental animals. Considering
their preferential emergence in cirrhotic livers of the high-risk group, their unequivocally
elevated proliferative activity, and the resulting large size with frequent nodular transformation,
we suggest that mixed cell populations are endowed with a high potential to progress to HCC in
humans, as previously shown in rats." In human HCC, irregular areas of liver parenchyma with
marked cytoplasmic amphophilia, phenotypically similar to the amphophilic preneoplastic foci in
rodent liver exposed to different hepatocarcinogenic chemicals (e.g. DHEA a peroxisome
proliferator) or the hepadnaviruses were reported to present in 45% of the specimens from
cirrhotic livers examined. "However, more data are needed to elucidate the nature of the
02/12/09
E-267
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
oncocytic and amphophilic lesions regarding their role in HCC development."
With respect to the ability respond to a mitogenic stimulus, differences between primate
and rodent liver response to a powerful stimulus, such as partial hepatectomy, have been noted
that indicate that primate and human liver respond differently (and much more slowly) to such a
stimulus. Gaglio et al. (2002) report after 60% partial hepatectomy in Rhesus macaques (Macaca
mulatto), the surface area of the liver remnant was restored to its original preoperative value over
a 30 day period. The maximal liver regeneration occurred between days 14 and 21, with
thickening of liver cell plates, binucleation of hepatocytes, Ki-67 and PCNA expression
(occurring in hepatocytes throughout the lobule at a maximum labeling index of 30%), and
mitoses parallel increased most prominently between post-hepatectomy days 14 and 30.
However, cytokines associated with inducing proliferation were elevated much earlier. TGF-a,
IL-6, HGF, IL-6 and TNF-a mRNA persisted until day 14, with peak elevations of IL-6, TNF-a,
occurring 24 hours later surgery, and IL-6 reduced to control levels by day 14. Gaglio et al.
(2002) suggest that their results clearly indicate that the pattern and timing of liver regeneration
observed in this non-human primate model are significantly different when comparing different
species (e.g., peak expression of Ki-67 in a 60% partial hepatectomy model in rats occurs within
hours following partial hepatectomy) and that the difference in timing and pattern of maximal
hepatocellular regeneration cannot be explained simply by differences in size of animals (e.g.,
60% partial hepatectomy in dogs produced liver regeneration peaks at 72 hours with weights
approximating the weights of the Rhesus macaques). They note that previous studies in humans,
who underwent 40-80%) partial hepatectomy, reveal a similar delay in peak liver regeneration
based on changes in serum levels of ornithine decarboxylase and thymidine kinase, further
highlighting significant interspecies differences in liver regeneration. For C57BL/6 X 129 mice
Fujita et al (2001) report that after partial hepatectomy, the liver had recovered more than 90%> of
its weight within 1 week. This difference in response to a mitogenic stimulus has impacts on the
interpretations of comparisons between rodent and primate liver responses to chemical exposures
which give a transient increases in DNA synthesis or cell proliferation such as PPARa agonists.
Also, as stated above, the primate and human liver, while having a significant polyploidy
compartment, do not have the extent of polyploidization and the early onset of that has been
observed in the rodent. However, as noted by Lapis et al (1995), exposure to DEN has proven to
be a highly potent hepatocarcinogen in non-human primates, inducing malignant tumors in 100%
of animals with an average latent period of 16 months when administered at 40 mg/kg
intraperitoneally every 2 weeks.
In regard to species extrapolation of epigenomic changes between humans and rodents,
Weidman et al. (2007) caution that "Although we do predict some overlap between mouse and
human candidate imprinted genes identified through our machine-learning approach, it is likely
that the most significant criterion in species-specific identification will differ. This difference
underscored the importance for increased caution when assessing human risk from
environmental agents that alter the epigenome using rodent models; the molecular pathways
targeted may be independent."
02/12/09
E-268
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Despite species differences, the genome of the mouse has been sequenced and many
transgenic mouse models are being used to study the consequences of gene expression
modulation and pathway perturbation to study human diseases and treatments. However the use
of transgenic models must be used with caution in trying to determine to determine MO As and
the background effects of the transgene (including background levels of toxicity) and specificity
of effects must be taken into account for interpretation of MO A data, especially in cases where
the knockout in the mouse causes significant liver necrosis or steatosis (Keshava and Caldwell,
2006; Keshava et al 2006; Caldwell and Keshava, 2006; Caldwell et al 2007b). For the
determination of effects of pathway perturbation and similarity to human HCC phenotype,
mouse transgenic models have been particularly useful with tumors produced in such models
shown to correlate with tumor aggressiveness and survival to human counterparts.
3.2.1. Similarities with Human and Animal Transgenic Models
Mice transgenic for transforming growth factor -a (a member of the EGF family and a
ligand for the ErfB receptors) develop HCCs (Farazi and DePinho, 2006). Compound TGFa and
MYC transgenic mice show increase hepatocarcinogenesis that is associated with the disruption
of TGFpi signaling and chromosomal losses, some of which are syntenic to those in human
HCCs that include the retinoblastoma (RB) tumour suppressor locus (Sargent et al 1999). Lee et
al. (2004b) investigated whether comparison of global expression patterns of orthologous genes
in human and mouse HCCs would identify similar and dissimilar tumor phenotypes, and thus
allow the identification of the best-fit mouse models for human HCC. The molecular
classification of HCC on the basis of prognosis in Lee et al. (2004a) was further compared with
gene-expression profiles of HCCs from seven different mouse models (Lee et al., 2004b). Lee et
al (2004b) characterized the gene expression patters of 68 HCC from 7 different mouse models;
two chemically induced (Ciprofibrate and diethylnitrosamine), four transgenic (targeted
overexpression of Myc, E2F1, Myc andE2Fl, and Myc and Tgfa in the liver). HCCs from some
of these mice (MYC, E2F1 and MYC-E2F1 transgenics) showed similar gene-expression
patterns to the ones of HCCs from patients with better survival. Murine HCCs derived for
MYC-TGFa transgenic model or diethylnitrosamine-treated mice showed similar gene-
expression patterns to HCCs from patients with poor survival. The authors report that Myc Tgfa
transgenic mice typically have a poor prognosis, including earlier and higher incident rates of
HCC development, higher mortality, higher genomic instability and higher expression of poor
prognostic markers (e.g., AFP) and that Myc and Myc/E2fl transgenic mice have relatively
higher frequency of mutation in P-catenin (Catnb) and nuclear accumulation of P-catenin that are
indicative of lower genomic instability and better prognosis in human HCC.
Lee et al (2004b) indentified three distinctive HCC clusters, indicating that gene
expression pattern of mouse HCC are clearly heterogeneous and reported that Ciprorfibrate-
induced HCCs and HCCs from Acox-/- mice were closely clustered and well separated from
other mouse models. However, are several issues regarding this study that give limitations to
some of its conclusions regarding the Acox -/- mouse and Ciprofibrate treatment. The Acox-/-
mouse is characterized by profound hepatonecrosis which confounds conclusions regarding gene
02/12/09
E-269
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
expression related to PPARa agonism made by the authors. There was very limited reporting of
the animal models (DEN and Clofibrate) protocols used. Only 3 tumors were examined for
Clofibrate treatment and it is unknown if the tumors were from the same animals. Similarly only
3 tumors were examined from DEN treatment which has been shown to produce heterogeneous
tumors and to produce necrosis in some paradigms of exposure. Myc/E2F1 and E2F1 mice were
split in both clusters that were compared with human HCCs. The authors used previously
published data from Meyer et al. (2003) for tumors from Acoxl"1" null mice, DENA-treated mice
and Ciprofibrate-treated mice.
Meyer et al (2003) examined three tumors from 2 C57BL/6j mice fed Ciprofibrate for 19
months and three tumors from 2 C57BL/6j mice injected with DEN at 2-3 months but the age at
which tumors appear was not given by the authors. Pooled mRNA from animals of varying age
(5-15 months old was used for controls. mRNAs that differed by 2-fold in tumors were
reported to be: 60 genes up-regulated and 105 genes down-regulated in Acoxl"1" null mice
tumors; 136 genes up-regulated and 156 genes down-regulated in Ciprofibrate-induced tumors;
and 61 genes up-regulated and 105 genes down-regulated in DEN-induced tumors. The authors
state that "Each tumor class revealed a somewhat different unique expression pattern." There
were "genes that were general liver tumor markers in all three types of tumors" with 38 genes
commonly deregulated in all three tumor types. On note, the cell cycle genes (CDK4,
CDC25Am CDC7 and MAPK3) cited by Lee et al (2004b) as being more highly expressed in
DEN-induced tumors were not reported to be changed in DEN tumors in Meyer et al (2003) or to
be altered in the Acoxl"1" null mice or mice treated with Ciprofibrate. Finally, the distinction
between groups may be dominated by gene expression changes in a large number of genes that
are related to PPAR activation but not related to hepatocarcinogenesis.
Calvisi et al. (2004a) used transgenic mice to study pathway alterations and tumor
phenotype and to further examine the premise that genomic alterations (genetic and epigenetic)
characteristic of HCC can describe tumors into 2 broad categories, the first category
characterized by activation of the Wnt/Wingless pathway via disruption of P-catenin function
and chromosomal stability and the second by chromosomal instability. Increased coexpression
of c-myc with TGF-a or E2F-1 transgenic mice was reported to result in a dramatic synergistic
effect on liver tumor development when compared with respective monotransgenic lines,
including shorter latency period, and more aggressive phenotype whereas P-catenin activation is
relatively common in HCCs developed in c-myc and c-myc/TGF-pi transgenic mice, rare in the
c-myc/TGF-a transgenic line which also has genomic instability. Calvisi et al. (2004a) also
report that P-catenin staining correlated with histopathologic type of liver tumors. Eosinophilic
tumors with abnormal nuclear staining of P-catenin were predominant in neoplastic lesions
characteristic of c-myc and c-myc/E2Fl lesions. Poorly differentiated HCCs with basophilic or
clear-cell phenotypes developed more frequently in c-myc/TGF-a and TGF-a mice and often
showed a reduction or loss of P-catenin immunoreactivity. P-catenin mutation was associated
with a more benign phenotype. Calvisi et al. (2004a) note that the relationship between P-
catenin activation, tumor grade, and clinical outcome in human HCC remains controversial.
02/12/09
E-270
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
"There are studies that show a significant correlation between P-catenin nuclear accumulation, a
high grade of HCC tumor differentiation, and a better prognosis, whereas others find that nuclear
accumulation of P-catenin may be associated with poor survival or that it does not affect clinical
outcome."
Calvisi et al. (2004b) report for E-cadherin a variety of morphologenetic events, including
cell migration, separation and formation of boundaries between cell layers and differentiation of
each cell layer into functionally distinct structures. Loss of expression of E-cadherin was
reported to result in dedifferentiation, invasiveness, lymph node or distant metastasis in a variety
of human neoplasms including HCC and that the role of E-cadherin might be more complex that
previously believed. In order to elucidate the role of E-cadherin in the sequential steps of liver
carcinogenesis, we have analyzed the expression patterns of E-cadherin in a collection of
preneoplastic and neoplastic liver lesions from c-Myc, E2F1, c-Myc/TGF-a and c-Myc/E2Fl
transgenic mice. In particular, we have investigated the relevance of genetic, epigenetic and
transcriptional mechanisms on E-cadherin protein expression levels. Our data indicate that loss
of E-cadherin contributes to HCC progression in c-Myc transgenic mice by promoting cell
proliferation and angiogenesis, presumably through the upregulation of HIF-la and VEGF
proteins. The c-Myc line, was most like wild type and lost E-cadherin in the tumors, c-
Myc/TGF-a dysplastic lesion were reported to show overexpression of E-cadherin mainly in
pericentral areas with E2F1 clear cell carcinoma showed intense staining of E-cadherin. Since
reduction or loss of E-cadherin expression is primarily determined by loss of heterozygosity at
the E-cadherin locus or by its promoter hypermethylation in human HCC Calvisi et al (2004b)
determined the status of the E-cadherin locus and promoter methylation in wild type livers and
tumors from transgenic mice by microsatellite analysis and methylation specific PCR,
respectively. "Wild-type livers and HCCs, regardless of their origins, showed the absence of
LOH at the E-cadherin locus. E-cadherin promoter was not hypermethylated in wild-type, c-
Myc/TGF-a and E2F1 livers. No E-cadherin promoter hypermethylation was detected in c-Myc
and c-Myc/E2Fl HCCs with normal levels of E-cadherin protein. In striking contrast, seven of
20 (35%) of c-Myc and two of four (50%) c-Myc/E2Fl HCCs with downregulation of E-
cadherin displayed E-cadherin promoter hypermethylation. These results suggest that promoter
hypermethylation might be responsible for E-cadherin downregulation in a subset of c-Myc and
c-Myc/E2Fl HCCs." "The molecular mechanisms underlying down-regulation of E-cadherin in
c-Myc tumors remain poorly understood at present. No LOH at the E-cadherin locus was
detected in the c-Myc HCCs whereas only a subset of c-Myc tumors displayed hypermethylation
of the E-cadherin promoter. Furthermore, no association was detected between E-cadherin
downregulation and protein levels of transcriptional repressors, Snail, Slug or the tumor
suppressor WT1, in disagreement with the finding that overexpression of Snail suppresses E-
cadherin in human HCC." "E-cadherin might play different and apparently opposite roles, which
depend on specific tumor requirements in both human and murine liver carcinogenesis."
Importantly, the results of Calvisi et al. (2004b) that hypermethylation of promoters can be
associated with down regulation of a gene in mouse liver tumors similar to human HCC and that
tumors can have the same behaviour with methylation change as with loss of hetererozygosity.
02/12/09
E-271
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
This report also gives evidence of the usefulness of the mouse model to study human
liver cancer as it shows the similarity of dysfunctional regulation in mouse and human cancer
and the heterogeneity within and between mouse lines tumors with differing dysfunctions in
gene expression. This parallels human cancer where there is heterogeneity in tumors from one
person and every tumor has its own signature. Finally, this report correlates differing pathway
perturbations with mouse liver phenotypes similar to those reported in experimental
carcinogenesis models and for TCE and its metabolites.
Farazi and DePinho (2006) suggest that "as comparative array CGH analysis of various
murine cancers has shown that such aberrations often target syntenic loci in the analogous human
cancer type, we further suggest that comparative genomic analysis of available mouse model of
mouse HCC might be particularly helpful in filtering through the complex human cancer
genome. Ultimately mouse models that share features with human HCCs could serve as valuable
tools for gene identification and drug development. However, one needs to keep in mind key
differences between mice and humans. For example, as noted in certain human HCC cases,
telomere shortening might drive the genomic instability that enables the accumulation of cancer-
relevant changes for hepatocarcinogenesis. As mice have long telomeres, this aspect of
hepatocarcingenesis might be fundamentally different between the species and provide additional
opportunities for model refinement and testing of this mechanism through use of a telomere
deficient mouse model. These and other cross-species difference, and limitations in the use of
human cell-culture systems, must be considered in any interpretation of data from various model
systems" (Farazi and DePinho, 2006). Thus, these mouse models of liver cancer inductions are
qualitatively able to mimic human liver cancer and support the usefulness of mouse models of
cancer.
3.3. Hypothesized Key Events in HCC Using Animal Models
3.3.1. Changes in ploidy
As stated above in Section 1.1, increased polyploidization has been associated with
numerous types of liver injury and appears to result from exposure to TCE and its metabolites as
well as changes in the number of binucleate cells. Hortelano et al (1995) reported that cytokines
and NO can affect ploidy and further suggests a role of these changes for carcinogenesis in
general. Vickers et al. (1996) noted that while both DEN and 17 a-ethinylestradiol have been
reported to enhance the proportion of diploid hepatocytes, initiators like N-nitrosomorpholine are
reported to increase the proportion of hypertrophied and polyploidy hepatocytes. The
relationship of such changes to cancer induction has been studied in transgenic mouse models
and in models involved with mitogens of differing natures.
Melchiorri et al. (1993) report the response pattern of the liver to acute treatment with
primary mitogens in regard to ploidy changes occurring in rat liver following two different types
of cell proliferation: compensatory regeneration induced by surgical partial hepatectomy and
direct hyperplasia induced by the mitogens lead nitrate and Nafenopin (a PPARa agonist) in 8
02/12/09
E-272
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
week old male Wistar rats. Feulgen stain was used and DNA content quantified by image
cytometry in mononucleate and binucleate cells. Mitotic index was determined in the same
samples. The term "diploid" was used to identify cells with a single, diploid nucleus and
tetraploid for cells containing 2 diploid nuclei or one tetraploid nucleus referred (bi- and
mononucleate, respectively). Octoploid cells were identified as either binucleate or
mononucleate. "During liver regeneration following surgical PH an increase in the mitotic index
with a peak at 24 hours was observed. The most striking effect associated with the regenerative
response was the almost complete disappearance of binucleate cells, tetraploid (2 X 2c) as well
as octoploid (4 X 2c) with only < 10% of the control values being present 3 days after PH."
"Concomitantly, an increase in mononucleate tetraploid (4c) as well as mononucleate octoploid
(8c) cells was observed, resulting at 3 days after PH in a population made up of almost entirely
(98%) by mononucleated cells." However, lead nitrate treatment was reported to induce rapid
increase in the formation of binucleate cells occurred with, 3 days after treatment, their number
accounting for 40% of the total cell population versus 22% binucleate cells in control rats and
2% in PH animals killed at the same time point. The increased binuclearity was reported to be
observed only in the 4 X 2c cells (25% vs. 6% of the controls) and in 8 X 2c cells (3.7% versus
0.1% of controls). The increase in 4 X 2c and 8 X 2c cells was reported to be accompanied by a
concomitant reduction in 2 X 2c cells with the change induced in cellular ploidy by lead nitrate
resulting in 37% of cells being either 8c or 16c. However, at the same time point, cells having a
ploidy higher than 4c were reported to account for only 11% in PH rats and 9% in control
animals. Changes in the ploidy pattern were reported to be preceded by an increased mitotic
activity which was maximal 48 hours after treatment with lead nitrate. The increase in mitotic
index in lead nitrate-treated rats was associated with a striking increase in the labeling index of
hepatocytes (60.1% versus 3% of control rats) and to an almost doubling of hepatic DNA content
in 3 days after lead nitrate. Melchiorri et al. (1993) concluded that the entire cell cycle appeared
to be induced by lead nitrate but that the finding of a high increase of binucleate cells suggested
that lead nitrate-induced liver growth, unlike liver regeneration induced by partial hepatectomy
was characterized by an uncoupling between cell cycle and cytokinesis which raised questions
whether lead nitrate-induced liver growth resulted in a true increase in cell number or is only the
expression of an increased hepatocyte ploidy. They reported that part of the increase in DNA
content observed 3 days after lead nitrate was indeed expression of polyploidizing process due to
acytokinetic mitoses but that a consistent increase in cells number (+26%) was also induced by
lead nitrate treatment.
After Nafenopin treatment, Melchiorri et al. (1993) reported that the increase in DNA
content was increased 22% over controls and was much lower than induced by lead nitrate and
that Nafenopin did not induce significant changes in binucleate cell number. However, a shift
towards a higher ploidy class (8c) was reported to be observed following Nafenopin and the 21%
increase in DNA content seen after Nafenopin treatment was almost entirely due to increase in
the ploidy state with only 7% increase in cell number. Melchiorri et al (1993) examined whether
hepatocytes characterized by high ploidy content (highly differentiated cells) would be
preferentially eliminated by apoptosis. An increase in apoptotic bodies was reported to be
associated with the regression phase after lead nitrate treatment (when liver mass is reduced) but
02/12/09
E-273
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
despite the elimination of excess DNA, the changes in ploidy distribution induced by lead nitrate
were found to persist suggested that polyploidy cells were not preferentially eliminated by
apoptosis during the regression phase of the liver. Melchiorri et al (1993) note that other studies
in rat exposed to the mitogens cyproterone acetate (CPA) and the peroxisome proliferator
methylclofenapate (MCP) also reported a very strong decline in binucleate cells with a
concomitant increase in mononucleate tetraploid cells in the liver similar to the pattern described
after partial hepatectomy.
Lalwani et al. (1997) reported the results of 1000 ppm WY-14,643 exposure in male
Wistar rats after 1, 2, and 4 weeks and suggested that an early wave of nuclear division occurred
at the early stages of exposure without cumulative effects on cell proliferation. Consistent with
hepatomegaly, WY-14,643-treated were reported to exhibit multifocal hepatocellular
hypertrophy and karyomegaly by routine microscopic analysis. For binucleate hepatocytes, there
were no reported differences between WY-14,643 and controls for days 4 and 11 but an increase
in the number at day 25 in WY-14,643-treated animals compared to controls. Increases in the
diameter of nuclei were shown by WY-14,643 treatment from Day 11 and 25 with increasing
numbers of cells displaying larger nuclear diameters. The mitotic index was reported to not be
significantly changed in WY-14,643 treated rats compared to controls. Mitotic figures did not
appear to survive the treatment necessary for flow cytometric analyses. Proliferating nuclear cell
antigen (PCNA) was increased on day 4 in WY-14,643- treated animals compared to controls
whereas no differences were found on days 11 and 25. However, immunohistochemistry was
reported to show remarkable increases in BrdU-labeled nuclei in liver sections after 4 days of
labeling with the populations of BrdU-labeled cell declining over the course of treatment. The
labeling index was high and approximately 80% of the BrdU-labeled cells were in periportal
areas. PCNA-expressing cells were increased in the periportal area of the liver. Intense nuclear
staining of PCNA was evident as an indicator of DNA replication in S phase. Microscopic
examination showed BrdU labeling only in periportal hepatocytes, whereas no significant
labeling was observed in non-parenchymal cells, indicating that the replicative activity was
confined to the liver cells. Lalwani et al (1997) suggested that their results showed that events
related to cell proliferation occur in the initial phase of WY-14,643 treatment in rats but not
followed by changes in the rate of DNA synthesis as the treatment progressed. They note that
Marsman et al (1988) observed constant increases in DNA synthesis by [3H]-thymidine
authoradiography with up to 1 year of continuous administration of WY-14,643, whereas the rate
of DNA synthesis or the BrdU labeling index in their study declined after the first 4 weeks of
treatment. They suggest that the increased percentage of cells appearing in G2-M phase and the
analysis of liver nuclear profiles suggest that the progression of these additional cells (i.e., cells
that are stimulated to enter the cell cycle by the test agent) through the cell cycle is arrested in
the late stages of the cell cycle. "Unlike BrdU labeling, which demonstrated DNA synthesis
activity over the 4-day labeling period, the PCNA labeling index represents levels of the protein
product at an interval post treatment. PCNA expression in cells exposed to chemicals or to WY
may not provide true representation of S phase or proliferative activity because PCNA-
expressing nuclei were also found in G0=G1 and G2-M phases." Lalwani et al. (1997)
02/12/09
E-274
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
concluded that cell proliferation alone does not appear to constitute a determining process
leading to tumors in most tissues and sustained cell replication may not be a primary feature of
peroxisome proliferator-induced hepatocarcinogenesis. Miller et al. (1996) note that studies
with methylclofenapate (MCP) in Alpk:AP rats indicate that DNA synthesis occurs primarily in
one hepatocyte subpopulation as defined by ploidy status, the binucleated tetraploid (2 X 2N)
hepatocytes and that this preferential hepatocyte DNA synthesis is manifested by dramatic
alterations in hepatocyte ploidy subclasses, i.e., significant increases in mononucleate tetraploid
(4N) hepatocytes concomitant with decreases in 2 X 2N hepatocytes. They reported results in
male Fischer 344 rats were 13 weeks old (an agent in which polyploidization had reached a
plateau) exposed to 1000 ppm WY-14,643 and MCP (gavage via corn oil at 8 mg/ml or 25
mg/kg MCP once daily) for 2, 5, and 10 days (n = 4). WY-14,643 and MCP were reported to
induce significant increases in the octoploid hepatocyte class that coincided with decreases in the
tetraploid hepatocyte class. However, MCP did not induce this shift until day 5 of exposure.
These results show an approximate doubling of 8N hepatoctyes but still a very low number of the
total hepatocyte population that does not reach greater than 7% and is still only ~ twice that of
control values and thus does not present itself with a very large target population. There was no
real effect on 4N hepatocytes due to these treatments and the % of hepatocytes that were 4 N
stayed ~ 70% and were thus the majority cell type in the liver. Miller et al (1996) note the
importance of maturation and/or strain for these analyses there are maturation-dependent
differences in the distribution and mitogenic sensitivity of hepatoctyes in the various subclasses.
Hasmall and Roberts (2000) note that despite their differing hepatocarcinogenecity, both
DCB (a nonhepatocarcinogen in Fischer 344 rats) and DEHP at the doses and routes used in the
NTP bioassays induced similar profiles of S-phase labeling index (LI). A large and rapid peak
during the first 7 days (1115% and 1151%) of control for DEHP and DCB, respectively) was
followed by a return to control levels. They suggest that the size of the S-phase response does
not necessarily determine hepatocarcinogenic risk and that the subpopulation in which S-phase is
induced may be a better correlate with subsequent hepatocarcinogenecity. They compared the
effects on polyploidy/nuclearity and on the distribution of S-phase labeled cells with ETU, the
peroxisome proliferator MCP, and phenobarbitone. Male F334 rats 7-9 weeks old were exposed
to MCP (0.1%) in diet), ETU 83 ppm diet, phenobarbitone (500 mg/ml drinking water) for 7 days.
The number of rats for 7 day study was not given by the authors. Hasmall and Roberts (2000)
reported that treatment of rats with MCP, ETU or phenobarbitone for 7 days had no significant
effect on the ploidy profile as compared with corn oil controls (data not shown) but that MCP
and phenobarbitone did induce significant changes in nuclearity. MCP reduced the binuclear
tetraploid (2 X 2N) population and increased the mononuclear octoploid (8N) population.
Phenobarbitone similarly increased the proportion of cells in the 4N population. ETU had no
effect on the nuclearity profile as compared with control. However, what the authors describe for
their results in polidy and nuclearity are different than those presented in their figures. There
were significant differences between controls that the authors did not characterize and there
appeared to be a greater difference between controls than some of the treatments.
Gupta et al. (2000) report that in transgenic mice with overexpression of transforming
growth factor a (TGF a), liver-cell turnover increases, along with the onset of hepatic
02/12/09 E-275
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
polyploidy, whereas hepatocellular carcinoma originating in these animals contain more diploid
cells and notes that Co-expression of c-Myc and TGF a transgenes in mouse hepatocytes was
associated with greater degrees of polyploidy as well as increased development of hepatocellular
carcinoma. Gupta notes that in the presence of ongoing liver injury and continuous depletion of
parenchymal cells, hepatic progenitor cells (including oval cells) are eventually activated but
what roles polyploid cells play in this process requires further study. In the working model by
Gupta, sustained disease by chronic hepatitis, metabolic disease, toxins, ect., may lead to
hepatocyte polyploidy and loss, and the emergence of rapidly cycling progenitor or escape cell
clones with the onset of liver cancer.
Conner et al. (2003) describe the development of transgenic mouse models in which
E2F1 and/or c-Myc was overexpressed in mouse liver. The E2F1 and c-Myc transcription
factors are both involved in regulating key cellular activities including growth and death and,
when overexpressed, are capable of driving quiescent cells into S-phase in the absence of other
mitogenic stimuli and are potent inducers of apoptosis operating at least through one common
pathway involving p53. Deregulation of their expression is also frequently found in cancer cells
(Conner et al., 2003). Conner et al. (2003) reported that although both c-Myc and E2F1 mono-
transgenic mice were prone to liver cancer, E2F1 mice developed HCC more rapidly and with a
higher frequency and that the combined expression of these two transcription factors
dramatically accelerated HCC growth compared to either E2F1 or c-Myc mono-transgenic mice.
All three transgenic lines were reported to show a low but persistent elevation of hepatocyte
proliferation before an onset of tumor growth. Ploidy was shown to be affected differently by c-
Myc and E2F1, and suggested distinct differences by which these two transcription factors
control liver proliferation/maturation. Both transgenic alterations induced liver cancer but had
differing effects on polyploidization suggestive that liver cancer can arise from either type of
mature hepatocyte.
c-Myc single-transgenic mouse showed a continuous high cell proliferation preceded the
appearance of pre-neoplastic lesions which was also true, although to a lesser extent, in the E2F1
mouse. At 15 weeks of age, all of the transgenic mouse lines were reported to have a high
incidence (>60%) of hepatic dysplasia with mitotic indices were equivalent in c-Myc/E2Fl, and
c-Myc livers, but 2-fold higher than the mitotic index in E2F1 and very low in wild type mice.
Thus, the combination of the two transgenes did not have and additive effect on proliferation.
An analysis of the DNA content in hepatocyte nuclei isolated from 4- to 15- week old mice was
reported to show that in young wild type livers, the majority of nuclei had a diploid DNA content
with a smaller proportion of tetraploid nuclei and as the mice aged, the number of tetraploid and
octoploid nuclei increased consistent with the previous findings of others. However, c-Myc mice
were reported to demonstrate a premature polyploidization with the number of 2N nuclei in c-
Myc livers was almost 2-fold less, while the proportion of 4N nuclei increased more than 2.5
fold at 4 weeks of age with the most prominent ploidy alteration being an increase in the fraction
of hepatocytes with octaploid nuclei (~ 200-fold higher). The percentage of polyploidy cells was
reported to continue to rise in 15 week old c-Myc livers with the majority of hepatocytes having
nuclei with 4N and 8N DNA content, with an attendant increase in binucleated hepatocytes and
02/12/09
E-276
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
increase in average cell size. In striking contrast, E2F1 hepatocytes were reported to not undergo
normal polyploidization with aging. The majority of E2F1 nuclei were reported to remain in the
diploid state and to be almost identical in E2F1 mice at 4 and 15 weeks of age with the
percentage of binucleated hepatocytes also reduced. In c-Myc/E2Fl mice, the age-related
changes in ploidy distribution were reported to resemble those found in both c-Myc and in E2F1
single transgenic mice. At a young age, c-Myc/E2Fl mice, similar to E2F1 mice, were reported
to retain significantly more diploid nuclei than c-Myc mice. However, as mice aged, the majority
of c-Myc/E2Fl hepatocytes, similar to c-Myc cells but in contrast to findings in E2F1 cells,
became polyploid. Consistent with a more progressive polyploidization, the DNA content was
significantly higher in both c-Myc/E2Fl and c-Myc livers. Conner et al. (2003) report that other
known modulators of ploidy in the liver are the tumor suppressor p53, pRb, and the cell cycle
inhibitor p21 as well as, genes involved in the control of the cell cycle progression such as cyclin
A, cyclin B, cyclin D3, and cyclin E.
Along with increased liver cancer, Conner et al. (2003) note that the C-Myc mice also
experienced a persistent liver injury as evidenced by significant elevation of circulating levels of
aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase along with the
appearance of a frequent oval/ductular proliferation. However, oval cell proliferation may be a
marker of hepatocyte damage but not be the cells responsible for tumor induction (Tarsetti et al.,
1993). Conner et al. (2000) report that in E2F1 is overexpressed in the liver there is both
oncogenic and tumor-suppressive properties. In regard to liver morphological changes, E2F1
transgenic mice were reported to uniformly develop pericentral dysplasia and foci adjacent to
portal tracts followed by the abrupt appearance of adenomas and subsequent malignant
conversion with all of the animals having foci by 2-4 months and by 8-10 months most having
adenomas with dysplastic changes remaining confined to the pericentral regions of the liver
lobule. In regard to phenotype, the majority of the foci were composed of small round cells, with
clear-cell phenotype but eosinophilic, mixed and basophilic foci were also seen. In adenomas
with malignant transformation to HCC, there appeared to be high mitotic indices, blood vessel
invasion and central collection of deeply basophilic cells with large nuclei giving a "nodule- in-
nodule: appearance. Macrovesicular hepatic steatosis was first noted in some E2F1 transgenic
livers at 6-8 months and by 10-12 months 60% of animals had developed prominent fatty
change. Hepatic steatosis has been noted in several transgenic mouse models of liver
carcinogenesis (Conner et al., 2000). These results raise interesting points of regional difference
in tumor formation which can be lost in analyses using whole liver and that the phenotype of foci
and tumors are similar to those seen from chemical carcinogenesis. The occurrence of
hepatotoxicity in these transgenic mice is also of note.
3.3.2 Hepatocellular Proliferation and Increased DNA Synthesis
Caldwell et al (2007b) have presented a discussion of the role of proliferation in cancer
induction. They state that "in the case of CC14 exposure, hepatocyte proliferation may be related
to its ability to induce liver cancer at necrogenic exposure levels, but the nature of this
proliferation is fundamentally different from peroxisome proliferators or other primary mitogens
02/12/09 E-277
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
that cause hepatocyte proliferation without causing cell death (Coni et al., 1993; Ledda-
Columbano et al., 1993, 1998, 2003; Menegazzi et al., 1997; Columbano and Ledda-Columbano,
2003). After initiation with a mutagenic agent, the transient proliferation induced by primary
mitogens has not been shown to lead to cancer-induction, while partial hepatectomy or
necrogenic treatments of CC14 result in the development of tumors [Ledda-Columbano et al.,
1993; Gelderblom et al., 2001]." Roskams et al (2003) notes that partial hepatectomy does not
cause hepatocellular carcinoma in normal mice without initiation. Melchiorri et al (1993) report
that a series of studies has shown tha acute proliferative stimuli provided by primary mitogens,
unlike those of the regenertative type such as those elicited by surgical or chemical partial
hepatectomy, do not support the initiation phase and do not effectively promote the growth of
initiated cells (Columbano et al 1990; Columbano et al 1987; Ledda-Columbano et al 1989).
They note that, the finding that most of these chemicals, with the exception of WY, induce only a
very transient increase in cell proliferation raises the question whether such a transient induction
of liver cell proliferation might be related to liver cancer appearing 1-2 years later. They note
that mitogen-induced liver growth differs from compensatory regeneration in several aspects (1)
it does not require an increased expression of hepatocyte growth factor mRNA in the liver (2) it
is not necessarily associated with an immediate early genes such as c-fos and c-jun; (3) it results
in an excess of tissue and hepatic DNA content that is rapidely eliminatated by apoptotic cell
death following withdrawls of the stimulus.
Other studies have questioned the importance of a brief wave of DNA synthesis in
induction of cancer. Chen et al (1995) note that Jirtle et al. (1991) and Schulte-Hermann et al.
(1986) reported that during a 2-week period of treatment with lead, DNA synthesis was
increased most in centrolobular hepatocytes and that the predominantly centrilobular distribution
of the labeled nuclei may have been due largely to the brief wave of mitogenic response, because
from the fifth day onward DNA synthesis activity returned to control level even though lead
nitrate treatment continued. They concluded that sustained cell proliferation may ben more
important than a brief wave of increased DNA synthesis. Chen et al. (1995) reported that a
number of different agents acting via differing MO As will induce periportal proliferation.
Vickers et al. (1996) reported that mitogenic response induced by acute 17 a-ethinylestradiol
adminstration is randomly distributed throughout the hepatic lobule, while continuous
adminstration increases the proportion of diploid cells. Richardson et al. (1986) reported that the
lobular distribution in the correlation of hepatocyte inititiation and akylation reported in their
model of carcinogenicity did "not support that early proliferation is associated with cancer as at 7
days there is a transient increase in the lobes least likely to get a tumor and no difference
between the lobes at 14 and 28 days DEN although there is a difference in tumor formation
between the lobes." Cells undergoing DNA synthesis may not be in the same zone of the liver
where other hypothesized "key events" take place. Tanaka et al. (1992) note that the distribution
of hepatocyte proliferation in the periportal area was in contrast to the distribution of peroxisome
proliferation in the centrilobular area of Clofibrate treated rats. Melnick et al. (1996) note that
replicative DNA synthesis commonly has been evaluated by measurement of the fraction of cells
incorportating bromodeoxyuridine or tritiated thymidine into DNA during S-phase of the cell
cycle (S-phase labeling index) but that the S-phase labeling index would not be identical to the
cell division rate when replication of DNA does not progress to formation of two viable daughter
02/12/09
E-278
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
cells. "The general view at an international symposium on cell proliferations and chemical
carcinogenesis was that although cell replication is involved inextricably in the development of
cancers, chemically enhanced cell division does not reliably predict carcinogenicity (Melnick et
al ,1993)." They note tha the finding that enzyme-altered hepatic foci were not induced in rats
fed WY-14,643 for 3 weeks followed by partial hepatectomy indicates that early high levels of
replicative DNA synthesis and peroxisome proliferation are not sufficient activities for initiation
of hepatocarcinogenesis.Baker et al., (2004) reported that, similar to the pattern of transient
increases in DNA synthesis reported for TCE metabolites, Clofibrate exposure induced the
upregulation of a variety of cell prolifertation-associated genes (e.g., G2/M specific cyclin Bl,
cyclin-dependent kinase 1, DNA topoisomerase II alpha, c-myc protooncogene, pololike serien-
threonine protein kinase, and cell divisions control protein 20) began on or before day 1 and
peaked at some point between days 3 and 7. By day 7, cell proliferation genes were down
regulated. The chronology of this gene expression agrees with the histologic dianosis of mitotic
figures in the tissue, where an increase in mitotic figures was detected in the day 1 and most
notably day 3 high and low-dose groups. However, by day 7, the incidence of mitotic figures
had decreased. The clustering of genes associated with the G2/M transition point suggests that in
the rats, the polyploidy cells arrested at G2/M are those that are proceeding through the cell
cycle.
A dose-response for increased DNA-synthesis also seems to be lacking for the model
PPARa agonist, WY-14,643 suggesting that the transient increase reported by Eacho et al (1991)
for this compound that then increases later at necrogenic exposure levels, are not related to its
carcinogenic potential. Wada et al. (1992) reported that in male Fischer 344 rats exposed to a
range of WY-14,643 concentrations (5 - 1000 ppm) that liver weight gain at the lowest dose
gave a sustained response for many weeks and gave increased cell labeling in the first week.
Peroxisomes proliferation as measure by electron microscopy increases started at 50 ppm
exposures but by enzymatic means, peroxisomal activities were elevated at the 5 ppm dose. Of
note is the difference in distribution in hepatocellular proliferation which was not where the
hypertrophy or where the lipofuscin increases were observed. The authors note that these data
suggest that 50 and 1000 ppm WY-14,643 should give the same carcinogencity if peroxisome
proliferation or sustained proliferation are the "key events." The study of Marsman et al (1992)
is very important in that it not only shows that clofibric acid dose not have sustatined
proliferation, but it also shows that it and WY-14,643 at 50 ppm did not induced apoptosis in
rats. It is probably that use of WY-14,643 are high concentrations may induce apoptosis in a
manner not applicable to other peroxisome proliferators or to treatment with WY-14,643 at 50
ppm. This study also confirms that exposure to WY-14,643 at 50 ppm and WY-14,643 at 1000
ppm induces similar effects in regards to hepatocyte proliferation and peroxisomal proliferation.
The study by Eacho et al. (1991) also gives a reference point for the degree of
hepatocytes undergoing transient DNA synthesis from WY-14,643 and Clofibrate and how much
smaller it is for TCE and its metabolites which generally involve less than 1% of hepatocytes.
"The labeling index of BrdU was 7.2% on day 3 and 15.5% on dayt 6 after clofibric acid but by
day 10 and 30 labeling index was the same as controls at -1-2%." "For WY the labeling index
was 34.1% at day 3 and 18.6% at day 6. At day 10 the labeling index was 3.3% and at day 30
02/12/09
E-279
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
was 6%, representing 6.6- and 15-fold of respective controls. Control levels were ~ 0.5 to 1%."
"The labeling index was increased to 32% by 0.3% LY171883 and to 52% by 0.05% Nafenopin.
The 0.005%) and 0.1% dietary doses of WY increased the 7 day labeling index to a comparable
level (55% - 58%)."
Yeldani et al. (1989) report results showing that until foci appear, cell proliferation has
ceased to increase over controls after the first week for ciprofibrate-induced
hepatocarcinogenesis and also shows the importance of using age matached controls and not
pooled controls for comparative purposes of proliferation as well as how low proliferative rates
are in control animals. The results of Barass et al. (1993) are important in suggesting when age
of animals is important when doing quantitation of labeling indexes. Studies such as that
conducted by Pogribny et al. (2007) that only give the replication rate as a ratio to control will
make the proliferation levels look progressive when in fact they are more stable with time, it is
just the controls that change with age as a comparison point.
3.3.3. Nonparenchymal Cell Involvement in Disease States Including Cancer
The recognition that not only parenchymal cells but also nonparenchymal cells play a
role in HCC has resulted in studies of their role in initiation as well as progression of neoplasia.
The role of the endothelial cell in controlling angiogenesis, a prerequisite for neoplastic
progression, and the role of the Kupffer cell and its regulation of the cytokine milieu that
controls many hepatocyte functions and responses have been reported. However, as pointed out
by Pikarsky et al. (2004) and by the review by Nickoloff et al. (2005) the roles of inflammatory
cytokines in cancer are context and timing specific and not simple. For TCE, non-parenchymal
cell proliferation has been observed after inhalation (Kjellstrand et al., 1983b) and gavage (Goel
et al., 1992) exposures of ~ 4 weeks duration.
3.3.3.1. Epithelial cell control of liver size and cancer - angiogenesis
The epithelium is key in controlling restoration after partial hepatectomy and not
surprisingly HCC growth. Greene et al (2003) hypothesized that the control of physiologic
organ mass was similar to the control of tumor mass in the liver and that specifically, the
proliferation of hepatocytes after partial hepatectomy, like the proliferations of neoplastic cells in
tumors, requires the synthesis of new blood vessels to support the rapidly increasing mass. They
report that a peak in hepatocyte production of VEGF, an endothelial mitogen, corresponds to an
increase of VEGF receptor expression on endothelial cells after partial hepatectomy and the rate
of endothelial proliferation.. Fibroblast growth factor and transforming gowth factor-alpha,
which stimulate endothelial cells, are secreted by hepatoctyes 24 hours after partial hepatectomy
but endothelial cells were reported to secrete hepatocyte growth factor, a potent heaptocyte
mitogen that is also proangiogenic. The secretion of transforming growth factor -beta by
endothelial cells 72 hours after partial hepatectomy was reported to inhibit hepatocyte
proliferation. Thus, Greene et al. (2003) suggested that endothelial cells and hepatocytes of the
regenerating liver influence each other, and both populations are required for the regulation of
02/12/09
E-280
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the regenerative process.
3.3.3.2. Kupffer cell control of proliferation and cell signals, role in early and late effects
Vickers et al. (1996) have rported that Kupffer cells are increased in number in
prenoplastic foci but are decreased in hepatocellular carcinom and that other studies have
demonstrated that both sinusoidal endothelial cells and Kupffer cells within hepatocellular
carcinoma cells in humans stain positive for mitotic activity although the number of non-
parenchymal cells compared to parenchymal cells may be reduced. Lapis et al. (1995) reported
that Kupffer cells contain lysozyme in their cytoplasmic granules, vacuoles and phagosomes and
some cells show a positive reaction in the rough endoplasmic reticulum, perinuclear cisternae
and the Golgi zone and that in human monocytes the lysozyme is co-localized with the CD68
antigen and myeloperoxidase. They also report that in rodent hepatocarcinogenesis increased
numbers of Kupffer cells were observed in preneoplastic foci, whereas abnormally low numbers
were present following progression to hepatocellular carcinoma.and that "the Kupffer cell count
in human HCC has also been shown to be very low and varies with different histological form."
They reported that for monkey HCCs, that the proportion of endothelial elements remained
constant (the parenchymal/endothelial cell ratio), however there was a striking reduction in the
areas occupied by Kupffer cells. While healthy control livers contained the highest number of
Kupffer cells, in the tumor-bearing cases the non-neoplastic, non-cirrhotic liver adjacent to the
HCC nodules had a significantly lower number of Kupffer cells and the number decreased
further in the non-neoplastic portions of cirrhotic livers. Within HCC nodules the Kupffer cell
count was greatly reduced with no significant changes were observed between the cirrhotic areas
and the carcinomas, however, the tumors contrained fewer lysozyme and CD68 positive cells.
Lapis et al. (1995) note that "since other cell types withing the liver sinusoids (monocytes and
polypmorphs) and portal macrophage were also positive, it was important to identify the star-like
morphology of the Kupffer cells. The results of the two indepent observers assessment of the
morphology and enumberation of Kupffer cells were quite consistent and differed by only 3%."
"The loss of Kupffer cells in the HCC may possibly result from capillarization of the sinusoids,
which has been observed during the process of liver cirrhosis and carcinogenesis. Capillarization
entails the sinusoidal lining endothelial cells losing their fenestrations."
3.3.3.3. Nf-kB and TNF-a - context, timing and source of cell signaling molecules
A large body of literature has been devoted to the study of nuclear factor k B for its role
not only in inflammation and a large number of other processes, but also for its role in
carcinogenesis. However, the effects of these cytokines are very much dependent on their
cellular context and the timing of their modulation. As described by Adli and Baldwin (2006),
"The classic form of NF-kB is composed of a heterodimer of the p50 and p65 subunits, which is
preferentially localized in the cytoplasm as an inactive complex with inhibitor proteins of the IkB
family. Following exposure to a variety of stimuli, including inflammatory cytokines and LPS,
IkBs are phosphorylated by the IKKa/p complexes then accumulate in the nucleus, where they
transcriptionally regulate the expression of genes involved in immune and inflammatory
02/12/09
E-281
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
responses." The five members of the mammalian NF-kB family, p65 (RelA), RelB, c-Rel,
P50/pl05 (NF-KB1) and p52/pl00 (NF-kB2), exist in unstimulated cells as homo- or
heterodimers bound to IkB family proteins. Transcriptional specificity is partially regulated by
the ability of specific NF-kB dimmers to preferentially associate with certain members of the IkB
family. Individual NF-kB responses can be characterized as consisting of waves of activation
and inactivation of the various NF-kB members (Hayden and Ghosh, 2004). While the function
of NF-kB in many contexts have been established, it is also clear that there is great diversity in
the effects and consequences of NF-kB activation with NF-kB subunits not necessarily
regulating the same genes in an identical manner and in all of the different circumstances in
which they are induced. The context within which NF-kB is activated, be it the cell type or the
other stimuli to which the cell is exposed, is therefore a critical determinant of the NF-kB
behavior (Perkins and Gilmore, 2006).
Balkwill et al. (2005) report that "the NF-kB pathway has dual actions in tumor
promotion: first by preventing cell death of cells with malignant potential, and second by
stimulating production of proinflammatory cytokines in cells of infiltrating myeloid and
lymphoid cells. The proinflammatory cytokines signal to initiated and/or otherwise damaged
epithelial cells to promote neoplastic cell proliferation and enhance cell survival. However, the
tumor promoting role of NF-kB may not always predominate. In some cases, especially early
cancers, activation of this pathway may be tuor suppressive (Perkins, 2004). Inhibiting NF-kB in
keratinocytes promotes squamous cell carcinogenesis by reducing growth arrest and terminal
differentiation of initiated keratinocytes (Seitz et al., 1998)." Other inflammatory mediators have
also been associated with oncogenesis. Balkwill et al. (2005) reported that TNFa is frequently
detected in human cancers (produced by epithelial tumor cells, as in for instance, ovarian and
renal cancer) or stromal cells (as in breast cancer). They also report that the loss of hormonal
regulation of IL-6 is implicated in the pathogenesis of several chronic diseases, including B cell
malignancies, renal cell carcinoma, and prostate, breast, lung, colon, and ovarian cancers. Over
100 agents, such as antioxidants, proteosome inhibitors, NSAIDs, and immunosuppressive
agents are NF-kB inhibitors with none being entirely specific (Balkwill et al., 2005). Thus,
alterations in these cytokines, and the cells that produce them, are implicated as features of
"cancer" rather than specific to HCC.
Balkwill et al. (2004) report that "Two mouse models of inflammation-associated cancer
now implicate the gene transcription factor NF-kB and the inflammatory mediator known as
tumour-necrosis factor a (TNF- a) in cancer progression. Using a mouse model of inflammatory
hepatitis that predisposes mice to liver cancers, Pikarsky et al. present evidence that the survival
of hepatocytes - liver cells - and their progression to malignancy, are regulated by NF-kB. NF-
kB is an important transcription factor that controls cell survival by regulating programmed cell
death, proliferation and growth arrest. Pikarsky et al. find that the activation state of NF-kB, and
its localization in the cell, can be controlled by TNF-a produced by neighboring inflammatory
cells (collectively known as stromal cells)" Pikarsky et al (2004) reported that that the
inflammatory process triggers hepatocyte NF-kB through upregulation of TNF-a in adjacent
endothelial and inflammatory cells. Switching off NF-kB in mice from birth to seven months of
02/12/09
E-282
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
age, using hepatocyte-specific inducible iKB-super repressor transgene, had no effect on the
course of hepatitis, nor did it affect early phases of hepatocyte transformation. By contrast,
suppressing NF-kB inhibition through anti- TNF-a treatment or induction of the IKB-super
repressor in later stages of tumor development resulted in apoptosis of transformed hepatocytes
and failure to progress to hepatocellular carcinoma. The Mdr2 knockout hepatocytes in
Pikarsky's model of hepatocarcinogenicity were distinguishable from wild-type cells by several
abnormal features; high proliferation rate, accelerated hyperploidy and dysplasia. Pikarsky et al
(2004) reported that NF-kB knockout and double mutant mice displayed comparable degrees of
proliferation, hyperploidy and dysplasia implied that NF-kB is not required for early neoplastic
events. Thus, activation of NF-kB was not important in the early stages of tumour development,
but was crucial for malignant conversion. "Greten et al reporting in Cell, come to a similar
conclusion by studying a mouse colitis-asociated cancer model. Their work does not directly
implicate TNF-a, but instead found enhanced production of several pro-inflammatory mediators
(cytokines) including TNF-a„ in the tumour microenvironment during the development of
cancer. An important feature of both studies is that NF-kB activation was selectively ablated in
different cell compartments in developing tumour masses, and at different stages of cancer
development." However, Balkwill et al (2004) also note that TNF-a and NF-kB have many
different effects, depending on the context in which they are called into play and the cell type
and environment.
In contrast, inhibition of Nf-KB impaired HCC progression in a mouse model of
cholestatic hepatitis (Pikarsky et al., 2004). El-Serag and Rudolph (2007) note that "the
influence of inflammatory signaling on hepatocarcingenesis can be context dependent; deletion
of Nf-KB-dependent inflammatory responses enhanced HCC formation in carcinogen treated
mice (Sakurai et al 2006). Similarly, deletion of Nf-KB essential modulator/ I kappa P kinase
(NEMO/IKK), an activator of Nf-KB, induced steatohepatitis and HCC in mice (Luedde et al.,
2007). Maeda et al. (2005) reported that hepatocyte specific deletion of IKKp (which prevents
NF-kB activation) increased DEN-induced hepatocarcingenesis and that a deletion of IKKp in
both hepatocytes and hematopoietic-derived cells, however, had the opposite effect, decreasing
compensatory proliferation and carcinogenesis. They suggest that thes results, which differ
from previous suggestion that the tumor-promoting function of NF-kB is excreted in hepatocytes
(Pikarsky et al., 2004) and suggest that chemicals or viruses that interfere with NF-kB activation
in hepatocytes may promote HCC development.
Alterations in NF-kB levels have been suggested as a key event for the
hepatocarcinogenicity PPARa agonists. The event associated with PPAR effects has been the
extent of NF-kB activation as determined through DNA binding. As reported by Tharappel et al
(2001), NF-kB activity is assayed with eletrophoretic modbility shift assay (EMSAs) with
nuclear extracts prepared from frozen live tissue as a measure of DNA binding of NF-kB.
Increase transcription of downstream targets of NF-kB activity have also been measured. It has
been suggested that PPARa may act as a protective mechanism against liver toxicity. Ito et al
(2007) cite repression of NF-kB by PPARa to be the rationale for their hypothesis that PPARa-
null mice may be more vulnerable to tumorigenesis induced by exposure to environmental
02/12/09
E-283
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
carcinogens. However, as shown in Section 3.4.1.2., although DEHP was reported to also induce
glomerularnephritis more often in PPARa-null mice as suggested Kamijo et al (2007) due of the
absence of PPARa- dependent anti-inflammatory effect of antagonizing the oxidative stresss and
NF-kB pathway, there was no greater or lesser susceptibility to DEHP-induced liver
carcinogencicity in the PPARa null mice.
Because PPARa is known to exert anti-inflammatory effects by inducing expression of
iKBa, which antagonizes NFkB signaling, the expression of IicBa has been measured in some
studies (Kamijo et al (2007) as well as expression of TNR1 mRNA to evaluate the sensitivity to
the inflammatory response. Ito et al (2007) report that in wild type mice there did not appear to
be a difference between controls and DEHP treatment for p65 immunoblot results. DEHP
treatment was also reported to not induce p65 or p52 mRNA either or influence the expression
levels of TNFa, IkBa, IkBp and IL-6 mRNA in wild type mice. Tharappel et al. (2001) treated
with WY-14,643, gemfibrozil or Dibutyl phthalate in rats and reported elevated NF-kB DNA
binding in rats with WY-14,642 to have sustained response but not others. WY increased DNA
binding activity of NF-kB at 6, 34 or 90 days. Gemfibrozil and DEHP increased NF-kB activity
to a lesser extent and not at all times in rats. For gemfibrozil there was only a 2-fold increase in
binding at 6 days with no increase at 34 days and increase only in low dose at 90 days. In rats
treated with Dibutyl phthalate, there no change at 6 days, at 34 days there was an increase at high
and low dose, at 90 days only low dose animals showed a change. In pooled tissue from WY-
14,643- treated animals, the complex that bound the radiolabled NF-kB fragment did contain
both p50 and p65. Both WY-14,643 and gemfibrozil were reported to produce tumors in rats
with Dibutyl pthalate untested in rats for carcinogenicity. Thus, early changes in NF-kB were
not supported as a key event and WY-14,643 to have a pattern that differed from the other
PPARa agonists examined.
In regard to the links between inflammation and cancer, Nickoloff et al. (2005) in their
review of the issue, caution that such a link is not simple. They note that "dissecting the
mediators of inflammation in cutaneous carcinogenic pathways has revealed key roles for
prostaglandins, cyclooxygenase-2, tumor necrosis factor-a, AP-1, NF-kB, signal transducer and
activator of transcription (STAT)3, and others Several clinical conditions associated with
inflammation appear to predispose patients to increased susceptibility for skin cancer including
discoid lupus erythematosus, dystrophic epidermolysis bullosa, and chronic wound sites.
Despite this vast collection of data and clinical observations, however, there are several
dermatological setting associated with inflammation that do not predispose to conversion to
lesions into malaignancies such as psoriasis, atopic dermatitis, and Darier's disease." Nickoloff
et al. (2005) suggest that such a "link may not be as simple as currently portrayed because certain
types of inflammatory processes in skin (and possibly other tissues as well) may also serve a
tumor suppressor function. Over the past few months, several publications in leading biomedical
journals grappled with an important issue in oncology, namely defining potential links between
chronic tissue damage, inflammation, and the development of cancer. Balkwill and Coussens
(2004) reviewed the role of the NF-kB signal transduction pathway that can regulate
inflammation and also promote malignancy. Their review summarixed the latest findings
02/12/09
E-284
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
revealed in a letter to Nature by Pikarsky et al. (2004). Using Mdr2 knockout mice in which
hepatitis is followed by hepatocellular carcinoma, Pikarsky et al. implicated TNFa upregulatoin
in tumor promotion of HCC, and suggest that TNFa and NF-kB are potential targets for cancer
prevention in the context of chronic inflammation. A similar conclusion was reached with
respect to NF-kB by an independent group of investigators using a model of experimental
dextran sulfate-induced colitis, in which inactivation of the IkB kinase resulted in reduced
colorectal tumors (Greten et al, 2004). Although there are many other clinical condition
supporting the concept of inflammation is a critical component of tumor progression (e.g., reflux
esophagitis/ esophageal cancer; inflammatory bowel disease/ colorectal cancer), there is at least
one notable example that does not fit this paradigm. As described below, psoriasis is a chronic
cutaneous inflammatory disease, which is seldom if ever accompanied by cancer suggesting the
relationship between tissue repair, inflammation, and development may not be as simple as
portrayed by the aforementioned reviews and experimental results. Besides psoriasis, other
noteworthy observations pointing to more complexity include the observation that in the Mdr2
knockout mice, we rarely detect bile duct tumors despite extensive inflammation, NF-kB
activation, and abundant proliferation of bile ducts in portal spaces (Pikarsky et al., 2004).
Moreover, in a skin cancer mouse model, NF-kB was shown to inhibit tumor formation (Dajee et
al., 2003). Thus, the composition of inflammatory mediators, or the properties of the responding
epithelial cells (e.g., signaling machinery, metabolic status), may dictate either tumor promotion
or tumor suppression. Chronic inflammation and tissue repair can trigger pro-oncogenic events,
but also that tumor suppressor pathways may be upregulated at various sites of injurty and
chronic cytokine networking."
"One cannot easiliy dismiss the many dilemmas raised by the psoriatic plaque that
confound a simple link between the tissue repair, inflammation, and carcinogenesis. Since it is
easily visible to the naked eye, and patients may suffer from such lesions for decades, it is
difficult to argue that various skin cancers such as squamous cell carcinoma, basal cell
carcinoma, or melanoma actually do develop within plaques by are being overlooked by patients
and dermatologists. Remarkably, psortiatic plaques are intentionally exposed to mutagenic
agents including excessive sunlight, topical administration of crude coal tar, or parenteral DNA
cross-linking agent -psoralen followed by ultraviolet light. Moreover these treatments are
known to induce skin cancer in nonlesional skin. Thus since psoriatic skin is characterized by
altered differentiation, angiogenesis, increased telomerase activity, proliferative changes, and
apoptosis resistance, one would expect that each and every psoriatic plaque would be converted
to cancer, or at least serve as fertile soil for the presence of non-epithelial skin cancers over
time." "In conclusion, it would seem prudent to remember the paradigm proposed by Weiss
(1971) in which he suggested that premalignant cells do not comprise an isolated island, but are a
focus of intense tissue interactions. The myriad inflammatory effects of the tumor
microenvironment are important for understanding tumor development, as well as tumor
suppression and senescence, and for the design for efficacious prevention strategies against
inflammation-associate cancer" (Nickoloff et al., 2005).
02/12/09
E-285
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
3.3.4. Gender Influences on Susceptibility
As discussed previously, male humans and rodents are generally more likely to get HCC.
The increased risk of liver tumors from estrogen supplements in women has been documented.
In mice male TCE exposure has been shown to have greater variability in response and greater
effects on body weight in males (Kjellstrand et al., 1983a, 1983b) but to also induce dose-related
increases in liver weight and carcinogenic response in female mice as well as males (see Section
2.3.3.2). Recent studies have attempted to link differences in inflammatory cytokines and gender
differences in susceptibility.
Lawrence et al. (2007) suggest that "studies of Naugler et al. (2007) and Rakoff-Nahoum
and Medzhitov (2007), advance our understanding of the mechanisms of cancer-related
inflammation. They describe an important role for an intracellular signaling protein called
MyD88 in the development of experimental liver and colon cancers in mice. MyD88 function
has been well characterized in the innate immune response (Akira and Takeda, 2004), relaying
signals elicited by pathogen-associated molecules and by the inflammatory cytokine interleukin-
1 (IL-1)." "The conclusion from Naugler et al. (2007) and Rakoff-Nahoun and Medzhitov is that
MyD88 may function upstream of NF-kB in cells involved in inflammation-associated cancer.
Immune cells infiltrate the microenvironment of a tumor. Naugler et al. (2007) and Rakoff-
Nahoun and Medzhitov (2007) suggest that the development of liver and intestinal cancers in
mice may depend on a signaling pathway in infiltrating immune cells that involved the protein
MyD88, the transcription factor NF-kB, and the pro-inflammatory cytokine 11-6. TLR binds a
ligand which acts on MyD88 which acts on NF-kB which leads to secretion of inflammatory
cytokine IL-6 which leads to promotion of tumor cell survival and proliferation."
Naugler et al. (2007) suggested gender disparity in MyD88-dependent IL-6 production
was linked to differences in cancer susceptibility using the DEN model (a mutagen with
concurrent regenerative proliferation at a single high dose) with a single injection of DEN.
Partial hepatectomy was reported to induce no gender-related difference in IL-6 increase. After
DEN treatment the male mouse had 275 ng/ml as the peak IL-6 levels 12 hours after DEN and
for female mice the peak was reported to be 100 ng/ml 12 hours after DEN administration. This
is only about a 2.5 fold difference between genders. 11-6 mRNA induction was reported for mice
4 hours after DEN while at 4 hours, at a time when there was no difference in serum IL-6
between male and female mice. It was not established that the 4 hour results in mRNA translated
to the differences in serum at 12 hour between the sexes. The magnitude of mRNA differences
does not necessarily hold the same relationship as the magnitude in serum protein. In fact, there
was not a linear correlation between mRNA induction and IL-6 serum levels.
A number of issues complicate the interpretation of the results of the study. The study
examined an acute response for the chronic endpoint of cancer and may not explain the
differences in gender susceptibility for agents that do not cause necrosis. The DEN was
administered in 15-day old mice (which had not reached sexual maturity) for tumor information
at a much lower dose than used in short-term studies of inflammation and liver injury in which
02/12/09
E-286
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
mature mice were used. If large elevations of IL-6 are the reason for liver cancer, why doesn't a
partial hepatectomy induce liver cancer in itself? The percentage of proliferation at 36 and 48
hours after partial hepatectomy was the same between the sexes. If a 2.5-fold difference in IL-6
confers gender susceptibility, it should do so after partial hepatectomy and lead to cancer. For
female mice, partial hepatectomy showed alterations in a number of parameters. However,
partial hepatectomy does not cause cancer alone. The 5-fold increase 4 hours after DEN
induction of IL-6 mRNA in male mice is in sharp contrast to the 27 fold induction of IL-6 1 hour
after partial hepatectomy (in which at 4 hours the IL-6 had diminished to 6-fold). There
appeared to be variability between experiments (e.g., the difference in males between
experiments appears to be the same magnitude as the difference between male and female in one
experiment and the baseline of IL-6 mRNA induction appeared to be highly variable between
experiments as well as absolute units of ALT in serum 24 and 48 hours after DEN treatment that
tended to be greater that the effects of treatments) with experiments using very few animals
(n=3) for most treatment groups. Of note is that the MyD88 -/- male mice still had a background
level of necrosis similar to that of WT mice at 48 hours after DEN treatment a time, long after
the peak of IL-6 mRNA induction and IL-6 serum levels were reported to have peaked .One of
the key issues regarding this study is whether difference in IL-6 reported here lead to a in
proliferation and does that difference within 48 hours of a necrotizing dose of a carcinogen
change the susceptibility to cancer? This report shows that male and female mice have a
difference in necrosis after CCL4 and a difference in proliferation. But are early differences in
IL-6 at 4 hours related to the same kind of stimulus that leads to necrosis and concurrent
proliferation? The amount of proliferation between male and female mice 48 hours after DEN
was very small and the study was conducted in a very few mice (n = 3). At 36 hours the
proliferation was almost the same between the genders and about 0.6% of cells. The baseline of
proliferation also differed between genders but the variation and small number of animals made
it insignificant statistically. At 48 hours the differences in proliferation between male and female
mouse were more pronounced but still quite low (2% for males and ~ 1% for females). Is the
change in proliferation just a change in damage by the agent? Given the large variation in serum
ALT and by inference necrosis, is there an equal amount of variability in proliferation? This
study gives only limited information for DEN treatment.
The difference in incidence of HCC was reported to be greater than that of proliferation
between genders and of other parameters although differences in tumor multiplicity or size
between the genders are never given in the paper. Most importantly, comparisons between the
short-term changes in cytokines and indices of acute damage are for adult animals that are
sexually mature and at doses that are 4 times (100 vs. 25 mg/kg) that of the sexually immature
animals who are going through a period of rapid hepatocyte proliferation (15 day old animals).
It is therefore difficult to extrapolate between the two paradigms to distinguish the effects of
hormones and gender on the response. Finally, the work of Rakoff-Hahoun and Medzhitov
(2007) showed that it is the effect of tumor progression and not initiation that is affected by
MyD88 (a signaling adaptor to Toll-like receptors). Thus looking at parameters at the initiation
phase at necrotic doses for liver tumors may not be relevant.
02/12/09
E-287
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
3.3.5. Epigenomic Modification
There are several examples of chemical exposure to differing carcinogens that have lead
to progressive loss of DNA methylation (i.e., DNA hypomethylation) including TCE and its
metabolites. The evidence for TCE and its metabolites is specifically discussed in Section
3.4.2.2. below. Other examples of carcinogens that have been noted to change DNA methylation
are early stages of tumor development include ethionine feeding, phenobarbitol, arsenic,
dibromoacetic acid, and stress. However, it has not yet been established whether epigenetic
changes induced by carcinogens and found in tumors play a causative role in carcinogenesis or
are merely a consequence of the transformed state (Tryndyak et al., 2006). Tamoxifen has been
reported to be an example of a chemical that induced genetic and epigenetic changes (Tryndyak
et al., 2006).
Pogribny et al. (2007) report the effect of WY-14,643 on global mouse DNA
hypomethylation exposed at 1000 ppm for 1 week, 5 weeks or 5 months. What is of particular
note in this study is that at this exposure level, one commonly used for MOA studies using WY-
14,643 to characterize the effects of PPARa agonists as a class, there was significant
hepatonecrosis and mortality reported by Woods et al. (2007b). Both wild type and PPARa-/-
null mice were examined. In wild type mice DNA syntheses was elevated 3-, 13-, and 22-fold of
time-matched controls after 1 week, 5 weeks, and 5 months of WY 14,543 treatment. Changes in
ploidy were not examined. After 5 weeks of exposure, the ratio of unmethylated CpG cites in
whole liver DNA was the same for WY-14,643 treatment and control but by 5 months there was
an increase in hypomethylation in WY-14,643 treated wild type mice. The authors did not
report whether foci were present or not which could have affected this result. The similarity in
hypomethylation at 5 days and 5 weeks, that also had a small probability of foci development, is
suggestive of foci affecting the result at 5 months. For PPAR -/- mice there was increased
hypomethylation reported at 1 week and 5 weeks after WY-14,643 treatment that was not
statistically significant with so few animals studied and at 5 months the null mice had decreased
hypomethylation compared to 1 and 5 weeks. The authors note that, methylation of c-Myc genes
was reported to not be affected by long-term dietary treatment with WY-14,643 even though
WY-14,643-related hypomethylation of c-Myc gene early after a single dose of WY-14,643 has
been observed (Ge et al., 2001a).The authors concluded "thus, alterations in the genome
methylation patterns with continuous exposure to non-genotoxic liver carcinogens, such as WY,
may not be confined to specific cell proliferation-related genes."
Pogribny et al. (2007) reported Histone H3 and H4 trimethylation status in wild type and
PPAR null mice to show a rapid and sustained loss of histone H3K9 and histone H4K20
trimethylation in wild type mice fed WY-14,643 from 1 week to 5 months. There was no
progressive loss in histone hypomethylation with the same amount of demethylation occurring at
5 days, 5 weeks, and 5 months in wild type mice fed WY. The change from control was ~ 60%
reduction. The control values with time were not reported and all controls were pooled to give
one value (n = 15). For PPAR-/-1 mice there was a slight decrease with WY-14,643 treatment (
~ 15%) reported. In wild type mice, WY-14,643 treatment was reported to have no effect on the
02/12/09
E-288
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
major histone methyltransferase, Suv39hl, while expression of another (PRDM/Rizl) increased
significantly as early as on week of treatment and remained elevated for up to five months. The
effect on expression of Suv420h2 (responsible for histone H4K20 trimethylation) was more
gradual and the amounts of this protein in livers of mice fed Wy-14m643 were reported to be
lower than in control. The authors did not examine these parameters in the null mice so the
relationship of these effects to receptor activation cannot be determined. Pogribny et al. (2007)
report hypomethylation of retroelements (LTRIAP, LINE1 and LINE2 retrotransposons)
following long-term exposure to WY-14,643 which the authors concluded can have effects on
the stability of the genome. Again these results are for whole liver that may contain foci.
Nevertheless, these findings raise questions about other target organs and a more general
mechanism for WY-14,643 effects than a receptor mediated one. The lack of effects on c-Myc
and the irrelevance of the transient proliferation through it reported here gives more evidence of
the irrelevance of a MOA dependent on transient proliferation. The authors stated that studies
show that a sustained loss of DNA methylation in liver is an early and indispensable event in
hepatocarcinogenesis induced by long-term exposure of both genotoxic and non-genotoxic
carcinogens in rodents. Thus, this statement argues against making such a distinction in MOA
for "genotoxic" and "non-genotoxic" carcinogens. Finally, the use of a dose which Woods et al.
(2007b) demonstrate to have significant hepatonecrosis and mortality, limits the interpretation of
these results and their relevance to models of carcinogenesis without concurrent necrosis.
Strain sensitivity to hepatocarcinogenicity has been investigated in terms of short-term
changes in methylation. Bombail et al. (2004) reported that a tumour-inducing dose of
phenobarbital reduced the overall level of liver DNA methylation in a tumour-sensitive
(B6C3F1) mouse strain but that the same dose of phenobarbital did not alter global methylation
level in a more tumour-resistant strain (C57BL/6), although the compound increased hepatocyte
proliferation in both strains (Counts et al., 1996). Bombail et al reported that "In a similar study,
Watson and Goodman (2002) used a PCR-based technique to measure DNA methylation
changes specifically in GC-rich regions of the mouse genome." Watson and Goodman (2002)
found that, that in these areas of the genome, exposure to phenobarbital caused an increase in
methylation in dosed animals compared with control animals. Again, the change was more
pronounced in tumour-prone C3H/He and B6C3F1 strains than in the less sensitive C57BL/6
strain. They also reported increased DNA synthesis in C57BL/6 mice but decreased global
methylation in the B6C3F1 strain after PB administration 1-2 weeks. The lifetime spontaneous
tumor rates were reported to be less than 5% in C57BL/6 mice but up to 80% in C3H/He mice.
Counts et al. (1996) reported that cell proliferation and global hepatic methylation status in
relatively liver tumor susceptible B6C3F1 with relatively resistant C57BL6 mice following
exposure to PB and/or chlorine/methionine deficient (CMD) diet. Cell proliferation was reported
to be higher in C57BL6 mice while transient hypomethylation occurred to a greater extent in
B6C3F1 mice after phenobarbital treatment while dual administration of CMD and PB led to
enhance cell proliferation and greater global hypomethylation with similar trends in terms of
strain sensitivities as compared with either treatment alone (i.e., greater increase in cell
proliferation in C57B1/6 and greater levels of hypomethylation in B6C3F1). Thus, the authors
concluded that B6C3F1 mice have relatively low capacity to maintain the nascent methylation
02/12/09 E-289
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
status of their hepatic DNA. However, on the whole, the control values for methylation for the
C57BL/6 mice appear to be slightly higher than the B6C3F1 mice. Thus, claims that the liver
tumor sensitive B6C3F1 had more global hypomethylation after a promoting stimulus which
could be related to tumor sensitivity, is tempered by the fact that resistant strain had a higher
control baseline of hypomethylation. The baseline level of LI or hepatocyte proliferation appears
to be slightly higher in the C57BL/6 mouse. In addition, the largest strain difference in
hypomethylation after a CMD diet was at week 12 (135% of control for the B6C3F1 strain and
151% of control for the C57BL/6 strain) and this pattern was opposite that for the 1 week time
point. Thus, the suggestion by Counts et al. (1996), that the inability to maintain methylation
status by the B6C3F1 strain, is also not supported by the data for CMD diet.
3.4. Specific Hypothesis for MOA of TCE Hepatocarcinogenicity in Rodents
3.4.1. PPARa Agonism as the MOA for Liver Tumor Induction - The State of the
Hypothesis
PPARa receptor activation has been suggested to be the MOA for TCA liver tumor
induction and for TCE liver tumor induction to occur primarily as a result of the presence of its
metabolite TCA (NAS, 2006). However, as discussed previously (Section 2.1.10), TCE-induced
increases in liver weight have been reported in male and female mice that do not have a
functional PPARa receptor (Nakajima et al., 2000). The dose-response for TCE-induced liver
weight increases differs from that of TCA (see Section 2.4.2). The phenotype of the tumors
induced by TCE have been described to differ from those by TCA and to be more like those
occurring spontaneously in mice, those induced by DCA, or those resulting from a combination
of exposures to both DCA and TCA (see Section 2.4.4). As to whether TCA induces tumors are
induced through activation of the PPARa receptor, the tumor phenotype of TC A-induced mouse
liver tumors has been reported to have a pattern of H-ras mutation frequency that is opposite that
reported for other peroxisome proliferators (see Section 2.4.4.; Bull et al 2002; Stanley et al.,
1994; Fox et al., 1990; Hegi et al., 1993). While TCE, DCA, and TCA are weak peroxisome
proliferators, liver weight induction from exposure to these agents has not correlated with
increases in peroxisomal enzyme activity (e.g. PCO activity) or changes in peroxisomal number
or volume. However, liver weight induction from subchronic exposures appears to be a more
accurate predictor of carcinogenic response for DCA, TCA and TCE in mice (see Section 2.4.4).
The database for cancer induction in rats is much more limited than that of mice for
determination of a carcinogenic response to these chemicals in the liver and the nature of such a
response.
The MOA for peroxisome proliferators has been the subject of research and debate for
several decades. It has evolved from an "oxidative damage" due to increased peroxisomal
activity to a MOA framework example developed by Klaunig et al. (2003) that reported that
described causal inferences to be made for hepatocarcinogenesis after a chemical exposure was
shown to activate of the PPAR-a receptor with concurrent perturbation of cell proliferation and
apoptosis, and selective clonal expansion. Of note although inhibition of apoptosis was proposed
02/12/09
E-290
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
as part of the sequellae of PPARa activation, as noted in Section 2.4.1., no changes in apoptosis
in mice exposed to TCE have been reported with the exception of mild enhanced apoptosis at
1000 mg/kg/day dose but more importantly that for mice the rate of apoptosis decreases as mice
age and appear to be lower than that of rats. While DCA exposure has been noted to reduce
apoptosis, the significance of DCA-induced reduction in apoptosis, from a level that is already
inherently low in the mouse, for the MOA for induction of DCA-induce liver cancer is difficult
to discern.
Klaunig et al. based causal inferences on the attenuation of these events in PPAR-a-null
mice in response to the prototypical agonist WY-14,643 with a number of intermediary events
considered to be associative (e.g., expression of peroxisomal and non-peroxisome genes,
peroxisome proliferation, inhibition of gap junction intracellular communication, hepatocyte
oxidative stress as well as Kupffer cell-mediated events). The data set for di(2-ethylhexyl)
phthalate (DEHP) was prominently featured as an example of "PPAR-a induced
hepatocarcinogenesis." For DEHP PPAR-a activation was described as the initial key event with
evidence lacking for a direct effect but supported primarily supported by evidence from PPAR-a-
knockout mice treated with WY-14,643. Klaunig et al. concluded that".. .all the effects
observed are due only to the activation of this receptor and the downstream events resulting from
this activation and that no other modes of action are operant"
Although that PPARa receptor activation is the sole MOA for DEHP has been cited by
several reports (including IARC 2000), several articles have questioned the adequacy of this
proposed MOA (Melnick 2001; Melnick 2002; Melnick 2003; Melnick et al., 2007; FIFRA
2004; Caldwell and Keshava, 2006; Caldwell et al., 2008b; Keshava and Caldwell, 2006; and
Keshava et al., 2006). New information is now available that also questions several of the
assumptions inherent in the proposed MOA by Klaunig et al. and the dismissal of PPARa
agonists as posing a health risk to humans. Specific questions have been raised about the use of
WY-14,643 as a prototype for PPARa (especially at necrogenic doses) and use of the PPARa -/-
null mouse in abbreviated bioassays to determine carcinogenic hazard.
3.4.1.1. Heterogeneity of PPARa agonist effects and inadequacy of WY-14,643 paradigm as
prototype for class
Inferences regarding the carcinogenic risk posed to humans by PPARa agonists have
been based on limited epidemiology studies in humans that were not designed to detect such
effects. However, as noted by Nissen et al. (2007) the PPARa receptor is pleiotropic, highly
conserved, has "cross talk" with a number of other nuclear receptors, and plays a role in several
disease states. "The fibrate class of drugs, which are PPARa agonists intended to treat
dyslipidemia and hypercholesterolemia, have recently been associated with a number of serious
side effects." While these reports of clinical side effects are for acute or subchronic conditions
do not (and would not be expected to) be able to detect liver cancer from fibrate treatment, they
clearly demonstrate that compounds activating the PPAR receptors may produce a spectrum of
effects in humans and the difficulty in studying and predicting the effects from PPAR agonism.
02/12/09
E-291
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Graham et al. (2004) recently reported significantly increased incidence of hospitalized
rhabdomyolysis in patients treated with fibrates both alone and in combination with statins.
Even though pharmaceutical companies have spent a great deal of effort to develop agonists
which are selective for desired effects, the pleiotropic nature of the receptor continues to be an
obstacle.
Also, fibrates, WY-14,643 and other PPARa agonists pan agonists for other PPARs.
Shearer and Hoekstra (2003) note that fibrates, including Fenofibrate, Clofibrate, Bezafibrate,
Ciprofibrate, Gemfibrozil, and Beclofibrate are all drugs that were discovered prior to the
cloning of PPARa and without knowledge of their mechanism of action but with optimization of
lipid lowering activity was carried out by administration of candidates to rodents. They report
that many PPARa ligands, including most of the common fibrate ligands, show only modest
selectivity over the other subtypes with, for example, fenofibric acid and WY-14,643 showing <
10 fold selectivity for activation of human PPARa compared to PPARy and/or PPARS. In human
receptor transactivation assays they report:
Human receptor transactivation assays of EC50:
WY-14,643 = 5.0 |im for PPARa, 60 |im for PPAR y, 35 |im for PPARS.
Clofibrate = 55 [j,m for PPARa, -500 [j,m for PPAR y, inactive at 100 [j,m for PPARS
Fenofibrate = 30 [j,m for PPARa, 300 [j,m for PPAR y, inactive at 100 [j,m for PPARS
Bezafibrate = 50 [j,m for PPARa, 60 [j,m for PPAR y, 20 [j,m for PPARS.
Murine receptor transactivation assay of EC50:
WY = 0.63 [j,m for PPARa, 32 [j,m for PPAR y, inactive at 100 [j,m for PPARS
Clofibrate = 50 [j,m for PPARa, -500 [j,m for PPAR y, inactive at 100 [j,m for PPARS
Fenofibrate =18 [j,m for PPARa, 250 [j,m for PPAR y, inactive at 100 [j,m for PPARS
Bezafibrate = 90 [j,m for PPARa, 55 [j,m for PPAR y, 110 [j,m for PPARS.
Thus these data show the relative effective concentrations and "potency for PPAR
activity" of various agonists in humans and rodents, rodent and human responses may vary
depending on agonist, agonists vary in what they activate between the differing receptors and
that there is a great deal of transactivation of these drugs.
For fibrates specifically, a study by Nissen et al. (2007) reports that in current practice, 2
fibrates, Gemfibrozil and Fenobibrate, are still widely used to treat a constellation of lipid
abnormalities known as atherogenic dyslipidemia and note that currently available fibrates are
weak ligands for the PPARa receptor and may interact with other PPAR systems. They note that
the pharmaceutical industry has sought to develop new, more potent and selective agents within
this class but, most importantly, that none of the novel PPARa agonists has achieved regulatory
approval and that according to a former safety officer in the U.S. Food and Drug Administration
(FDA) (El-Hage, 2007) that more than 50 PPAR modulating agents have been discontinued due
02/12/09
E-292
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
to various types of toxicity (e.g., elevations in serum creatinine, rhabdomylosis, "multi-species,
multi-site increases in tumor with no safety margin for clinical exposures, and adverse
cardiovascular outcomes) but without scientific publications describing the reasons for
termination of the development programs. Nissen et al. report differences in effect between a
more highly selective and potent PPARa agonist and the less potent and specific one in humans.
They note "a recent large study of Fenofibrate in patients with diabetes showed no significant
reduction in morbidity but a trend toward increased all-cause mortality (Keech et al. 2005, 2006).
Whether this potential increase in mortality is derived from compound specific toxicity of
Fenofibrate or is an adverse effect of PPARa activation remains uncertain." In addition to the
lack of publication of effects from PPAR agonists in human trials in which toxicity can be
examined as noted by Nissen et al., the Keech study is illustrative of the problem in trying to
ascertain liver effects from fibrate treatment in humans as the focus of the outcomes was
coronary events in a study of 5 years duration in a older diabetic population. As stated above,
the challenges the pharmaceutical industry and the risk assessor face in determining the effects of
PPAR agonists is "that these compounds and drugs modulate the activity of a large number of
genes, some of which produce unknown effects." Nissen et al. further note that "Accordingly,
the beneficial effects of PPAR activation appear to be associated with a variety of untoward
effects which may include, oncogenesis, renal dysfunction, rhabdomylosis, and cardiovascular
toxicity. Recently, the FDA began requiring 2-year preclinical oncogenicity studies for all
PPAR-modulating agents prior to exposure of patients for durations of longer than 6 months (El-
Hage, 2007)." Thus, while existing evidence for liver cancer in humans is null rather than
negative, there remains a concern for oncogenicity and many obstacles for determining such
effects through human study. The heterogeneity in response to PPARa agonists and the
heterogeneity of effects they cause (Keshava and Caldwell, 2006) are evident from these reports.
Many studies have used the effects of WY-14,643 at a very high dose and extrapolated
those findings to PPARa agonists as a class. However, this diverse group of chemicals have
varying potencies and effects for the "key events" described by Klaunig et al. (2003) (Keshava
and Caldwell, 2006). The standard paradigm used with WY-14,643 to induced liver tumors in
all mice exposed by 1 year (an abbreviated bioassay), uses a large dose that has also has been
reported to produced liver necrosis, which can have an effect of cell proliferation and gene
expression patterns, and to also induce premature mortality (Woods et al., 2007b). As stated
above, WY- 14,643 also has a short peak of DNA synthesis that peaks after a few days of
exposure, recedes, and then unlike most PPARa agonists studied (e.g., Clofibrate, clofibric acid,
Nafenopin, Ciprofibrate, DEHP, DCA, TCA and LY-171883) has a sustained proliferation at the
doses studied (Tanaka et al., 1992; Barrass et al 1993; Marsman et al., 1992; Eacho et al., 1991;
Lake et al 1993; Yeldani et al., 1989; David et al 1999; Marsman et al., 1988; Carter et al., 1995;
Sanchez and Bull 1990). Clofibrate has been shown to have a decrease in proliferation gene
expression shortly after its peak (see Section 3.2.2). As shown in above, WY-14,643
hepatocellular increases in DNA synthesis did not appear to have a dose-response (see Section
3.4.2), only WY-14,643 had a sustained elevation of Nf-KB (gem and dibutyl phthalate did not)
(see Section 3.4.3.3), and the effects on DNA methylation occurred at 5 months and not earlier
time points (when Foci were probably present) and effects of histone trimethylation were
02/12/09 E-293
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
observed to be the same from 1 weeks to 5 months (see Section 3.4.5). Such effects on the
epigenome suggest other effects of WY, -14,643 other than receptor activation and are not
specific to just WY-14,643 and are found in a number of conditions leading to cancer and in
tumor progression (see Sections 3.2.1 and 3.2.7.).
In their study of PPARa-independent short-term production of reactive oxygen species
from induced by large concentrations of WY-14,643 and DEHP in the diet, Woods et al. (2007c)
examined short-term exposures to (0.6% w/w DEHP or 0.05% or 500 pm WY-14,643 for 3 days,
1 weeks or 3 weeks) and reported that WY-14,643 induced a dramatic increase in bile flow that
was not observed from DEHP exposure. By 1 week of exposure there was a 5% increase in bile
flow for DEHP treatment but a 240% increase in bile flow for WY-14,643 treatment and by 3
weeks the difference in bile volume between treated and control was 12% for DEHP and 1100%
for WY-14,643 treated animals. The level of radical production was reported to be comparable
at 2 hours (after injection of 1.2 g/kg) and 3 weeks (0.6% w/w in diet) of DEHP exposure (~ 3
fold) (number of animals unknown). In this oxygen radical formation as measured by spin
trapping in the bile was reported to be decreased after 3 days of treatment after DEHP and WY-
14,643 treatment. However the large changes in bile flow by WY-14,643 treatment limit the
interpretation of this data along with a small number of animals examined in this study (e.g., 6
control and DEHP animals and 3 animals exposed to WY-14,643 at 3 days), a 30% variation in
% liver/body weight ratios between control groups, and the insensitivity of the technique. In an
earlier study oxidative stress appears to be correlated with neither cell proliferation nor
carcinogenic potency (Woods et al., 2006). Woods et al. (2006) reported WY-14,643Y or DEHP
to induce an increase in free radicals at 2 hrs, a decrease at 3 days then an increase at 3 weeks for
both. However, radical formation did not correlate with the proliferative response, as DEHP
fails to produce a sustained induction of proliferative response in rodent liver but WY-14,643
does, and both WY-14,643 and DEHP gave a similar pattern of radical formation that did not
vary much from controls which is in contrast to their carcinogenic potency.
Although assumed to be a reflection of cell proliferation in many studies of WY-14,643
and by Klaunig et al. (2003), DNA synthesis recorded using the standard exposure paradigm for
WY-14,643, can also be a reflection of hepatocyte, nonparenchymal cell or inflammatory cell
mitogenesis (in the case of necrosis induced inflammation), from changes in hepatocyte ploidy,
or a combination of all. Other peroxisome proliferators have been shown to have a decrease in
proliferation gene expression shortly after their peaks (e.g., Clofibrate, see Section 3.2.2) and
both Methylclofenapate and Nafenopin have been shown to increase cell ploidy with Nafenopin
having the majority of its DNA synthesis a reflection of increased ploidy with only a small
percentage as increases in cell number (Section 3.4.1.). Several authors have also noted
increases in ploidy for WY-14,643 (see Section 3.4.1).
The Tg.AC genetically modified mouse was used to study 14 chemicals administered by
the topical and oral (gavage and/or diet) routes by Eastin et al. (2001). Clofibrate was considered
clearly positive in the topical studies but not WY-14,643 regardless of route of administration.
Based on the observed responses it was concluded by the workgroup (Assay Working Groups)
02/12/09
E-294
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
that the Tg.AC model was not overly sensitive and possesses utility as an adjunct to the battery
of toxicity studies used to establish human carcinogenic risk. The difference in result between
Clofibrate and WY-14,643 is indicative of a different MOA for the two compounds.
Similarly, at large exposure concentrations Boerrigter (2004) investigated the response of
male and female lacZ-plasmid transgenic mice treated at 4 months of age with 6 doses of 2,333
mg/kg DEHP, 200 mg/kg WY-14,643 or 90 mg/kg Clofibrate over a two week period mutation
frequencies assayed at 21 days following the last exposure. DEHP and WY-14,643 were shown
to significantly elevate the mutant frequency in both male and female liver DNA while
Clofibrate, as the dose level studied, was apparently non-mutagenic in male and female liver
(i.e., six-dose exposure to DEHP or WY-14,643 over a two week period significantly increased
the mutant frequency in liver of both female and male mice by approximately 40%). The author
notes that "the laxZ plasmid-based transgenic mouse mutation assay is somewhat unique among
other commercially available models (e.g. mutamouse and big blue), by virtue of its ability to
accurately quantify both point mutations and large deletions including those which originate in
the lacZ plasmid catamer and extend into the 3' flanking genomic region. It should be noted that
to date there is no single, agreed upon protocol for conducting mutagenicity assays with
transgenic rodents although several aspects have been upon by the Transgenic Mutation Assays
workgroup of the International Workshop on Genotoxicity Procedures."
For several chemicals both rats and mice demonstrate evidence of receptor activation
through peroxisome proliferation and peroxisome-related gene expression but only one develops
cancer. The herbicide, 2,4- dichlorophenoxyacetic acid (2,4-D), is a striking example of the
problems that would be associated with only using evidence of PPARa receptor activation to
make conclusions about MOA of liver tumors. 2,4-D is structurally similar to the PPARa
agonist Clofibrate and has been shown at similar concentrations to increase peroxisome number
and size, increase hepatic carnitine acetyltransferase activity and catalase, and decrease serum
triglycerides and cholesterol in rats (Vainio et al., 1983). Peroxisome number was also increased
in Chinese hamsters to a similar level as with Clofibrate at the same exposure concentration after
9 days of exposure to 2,4-D (Vainio et al., 1982). In mice, Lundgren et al. (1987) report that
2,4-D exposure statistically increased the liver-somatic index over controls after a few days
exposure and increased mitochondrial protein, microsomal protein, carnitine acetyltransferase,
PCO activity, cytochrome oxidase, cytosolic epoxide hydrolase, microsomal epoxide hydrolase,
microsomal P450 content, and hepatic cytosolic epoxide hydrolase in mouse liver. Thus, 2,4-D
activates the PPARa receptor, with associated changes in peroxisome-related gene expression, in
multiple species and at similar doses to Clofibrate.. However, Charles et al. (1996) and Charles
and Leeming (1998) report that in several 2-year studies that there were no 2,4-D-induced
increases in liver tumors in F344 rats, CD1 rats, B6C3F1 mice and CD-I mice. Another
example, is provided by Gemfibrozil, known as (5-2[2,5-dimethylphenoxy]2-2-
dimethylpentanoic acid) and [2,2-dimethyl-5-(2,5-xylyoxy) valeric acid], a therapeutic agent
that activates the PPARa receptor and is a peroxisome proliferator, but is carcinogenic only in
male rats but not female rats, nor in either gender of mouse (Contrera et al., 1997). Gemfibrozil
causes tumors in pancreas, liver, adrenal, and testes of male rats and causes increases in absolute
02/12/09
E-295
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and relative liver weights in both rats and mice (Fitzgerald et al., 1981). Gemfibrozil, is a highly
effective lipid and cholesterol lowering drugs in humans and in mice (Olivier et al 1988).
However, although Gemfibrozil activates the PPARa receptor and induces peroxisome
proliferation in mice, it does not induce liver tumors in that species. In the longterm study of
Bezafibrate, Hays et al (2005) note that the role of this receptor in hepatocarcinogenesis has only
been examined using one relatively specific PPARa agonist (WY-14,643) and report that
Bezafibrate can induce the expression of a number of PPARa target genes (acyl CoA oxidase
and CYP4a) and increased liver weight in PPARa knockout mice that is not dependent on
activation of PPARP or PPARy. As noted by Boerrigter (2004), "In contrast to DEHP and WY-
14,643, Clofibrate produced hepatocellular carcinomas in rats only while no increase in the
incidence of tumors was reported in mice (Gold and Zeiger 1997). However, Clofibrate induces
peroxisome proliferation in both rats and mice (Lundgren and DePierre 1989) but only produced
hepatocellular carcinomas in rats (Gold and Zeiger, 1997)." Melnick et al. (1996) note that
similar levels of peroxisomal induction were observed in rats exposed to DEHP and DEHA
[di(2-ethylhexyl) adipate] at doses comparable to those used in the bioassays of these chemicals.
However, DEHP but not DEHA gave a positive liver tumor response in 2-year studies in rats. In
an evaluation of the carcinogenicity of tetrachloroethylene, an expert panel of the IARC
concluded that the weak induction of peroxisome proliferation by this chemical in mice was not
sufficient to explain the high incidence of liver tumors observed in an inhalation bioassay.
In adult animals, apoptosis acts as a safeguard to prevent cells with damaged DNA from
progressing to tumor, but like cell proliferation, alterations in apoptosis are common to many
MO As. In addition, only short term data is available on changes in apoptosis due to PPARa
agonists, and long-term changes have not been investigated (Rusyn et al., 2006). For example,
although a decrease in apoptosis has also suggested to be an important additional molecular
event that may affect the number of cells in rodent liver following exposure to the peroxisome
proliferator DEHP, apoptosis rates have not investigated past 4 days of exposure and thus the
time-course of this event is uncertain. The antiapoptotic effects of PPAR agonists appear to also
be dependent on non-parenchymal cells (i.e., Kupffer cells) which do not express PPARa and
could be a transient event (Rusyn et al., 2006). Morimura et al. (2006) report evidence for
exposure to WY-14,643 that does not support a role for PPARa -mediated apoptosis in tumor
formation (see Section 3.5.1.3., below) as well as appearing to be specific to WY-14,643 (see
Section 3.4.3.3.).
The lack of a causal relationship of transient DNA synthesis increases and
hepatocarcinogenesis has been raised by many (Caldwell et al 2008b) and is discussed in Section
3.4.2. as well as the changes in ploidy (see Section 3.4.1). In regard to gene expression profiles,
many studies have focused on gene profiles during the early transient proliferative phase or have
identified genes primarily associated with peroxisome proliferation as "characteristic" or relevant
to those associated with tumor induction. Several have focused on the number of genes whose
expression "goes up" or "goes down" from a small number of animals. Caldwell and Keshava
(2006) presented information on WY-14,643, dibutyl phthalate, Gemfibrozil and DEHP, and
noted inconsistent results between PPARa agonists, paradoxes between mRNA and protein
02/12/09 E-296
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
expression, strain, gender, and species differences in response to the same chemical, and time-
dependent differences in response for several enzymes and glutathione.
3.4.1.2. New information on causality and sufficiency for PPARa receptor activation
In its review of the EPA's draft risk assessment of perflurooctanoic acid (PFOA), the
Science Advisory Panel (FIFRA 2004) expressed concerns about whether PPARa agonism
constitutes the sole MO A for PFOA effects in the liver and the relevance to exposed fetuses,
infants, and children. In part based on uncertainties regarding the Klaunig et al. (2003) proposed
MO A, they concluded that the tumors induced by PFOA were relevant to human risk assessment.
The hypothesis that activation of the PPARa receptor is the sole mode of action
hepatocarcinogenesis induced by DEHP and many other chemicals is further called into question
by recent studies. In the case of DEHP, Klaunig et al (2003) assumed that WY-14,643 and
DEHP would operate through the same key events and that long-term bioassays of DEHP in
PPARa -/- knockout mice would be negative and hence demonstrate the need for receptor
activation for hepatocarcinogenesis from DEHP.
The fallacy of these assumption are illustrated by the recent report of the first 2-year
bioassay of DEHP in PPARa -/- knockout mice (Sv/129 background strain) that reported DEHP-
induced hepatocarcinogenesis (Ito et al., 2007). Similar to other studies, the PPAR-/- mice had
slightly increased liver weights in comparison to controls and treated wild type mice (~ 12%
increase over controls). In fact statistical analysis of the incidence data shows that adenomas
were significantly increased in PPARa -/- mice compared with wild type mice exposed to 500
ppm DEHP and that a significant dose-response trend for adenomas and adenomas plus
carcinomas was observed in PPARa -/- mice. Overall, the cancer incidences were consistent
with a previous study of DEHP (David et al., 1999) in B6C3F1 mice at the same doses for nearly
the same exposure duration. A strength of this study is that it was conducted at much lower
more environmentally relevant doses that did not significantly increase liver enzymes as
indications of toxicity. As noted by the authors, DEHP was reported to also induce
glomerularnephritis more often in PPARa-null mice because of the absence of PPARa-
dependent anti-inflammatory effect of antagonizing the oxidative stress and NF-kB pathway
(Kamijo et al., 2007). Thus, these data support that hypothesis that there is no difference in liver
tumor incidences between PPARa -/- mice and wild type mice in a standard non-abbreviated
exposure bioassay that does not exceed the maximal tolerated doses and that DEHP can induce
hepatotoxicity as well as other effects independent of action of the PPARa receptor.
Adenomas
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
ftftHl
m
O 100
ppm DEHP
~ fto-Wild ~ Ito-knockout ~ David-79wk ~ David-total
+ - p<=0.05 by 1 -tail Fisher exact test as compared to
control; * - p <=0 .05 by 1 - and 2-tail Fisher exact test as
compared to control in same study
E-2
Carcinomas
25.0% —
20.0%
§2 15.0%
5
"g 10.0%
5.0%
0.0%
ppm DEHP
~ ft o-Wild ~ fto-knockout
~ David-79wk ~ David-total
No statistically-significant differences across all
studies and doses.
-100 500
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The study of Yang et al. (2007a) informs as to the sufficiency of PPARa receptor
activation and subsequent molecular event for hepatocarcinogenesis in mice. The study used a
VP16PPARa transgene under control of the liver-enriched activator protein (LAP) promoter to
constituitively activate the PPARa receptor in mouse hepatocytes. LAP- VP16PPARa
transgenic mice showed a number of effects associated with PPARa receptor activation
including decreased serum triglycerides and free fatty acids, peroxisome proliferation, enhanced
hepatocyte DNA synthesis and induction of cell-cycle genes and those described as "PPARa
targets" to comparable levels reported for WY-14,643 exposure. Hepatocyte proliferation as
determined the labeling index into hepatocyte nuclei was increased after 2 weeks of WY-14,643
treatment over controls (20.5% vs. 1.6% in control livers) with the LAPVP16 PPARa mice
giving a similar results (20.8% vs. 1.0% in control livers). The authors note that transgenic mice
did not appear to have positive labeling of nonparenchymal cell nuclei that were present in the
WY-14,643 treated animals. The transferase-mediated dUTP nick end-labeling assay results
were reported to show that there was no difference in apoptosis in wild type mice treated with
WY-14,643, the transgenic mice or and controls. In a small number of animals, microsomal
genes (CYP4A), peroxisomal (ACOX, BIEN - the bifunctional enzyme) and mitochondrial fatty
oxidation genes (LCAD - long chain acyl CoA dehydrogenase and VLCAD - very long chain
acyl CoA dehydrogenase) were expressed in the transgenic mice with WY-14,643 also
increasing expression of these genes in wild type mice but with less LPL than the transgenic
mice. Hepatic coA oxidation, were increased to a similar level in wild type mice treated with
WY-14,643 and the transgenic mice (n = 3-4) and were statistically different than controls. LAP-
VP 16PPARa transgenic mice (8 weeks of age) exhibited hepatomegaly (~ 50 increase %
body/liver weight over controls), and an accumulation of lipid due to triglycerides but not
cholesterol. However, compared to wild-type mice exposed to WY-14,643 for two weeks, the
extent of hepatomegaly was reduced (i.e., % liver/body weight increase of ~ 2.5 fold with WY-
14,643 treatment), no hepatocellular hypertrophy or eosinophilic cytoplasms and no evidence of
non-parenchymal cell proliferation were observed in the LAP- VP16PPARa transgenic mice.
At ~1 year of age, Yang et al (2007a) reported there to be no evidence of preneoplastic
lesions or hepatocellular neoplasia in LAP- VP16PPARa transgenic mice, in contrast to 11
months of exposure to WY-14,643 in wild type mice. Microscopic examination of liver sections
were consistent with the gross findings, as hepatocellular carcinomas and hepatic lesions were
observed in the long-term WY-14,643 treated wild type mice, but not in > 20 LAP-VP16 PPARa
mice at the age of over 1 year in the absence of dox. There was no quantitative information on
tumors given nor of foci development in the WY-14,643 mice. As noted by Yang et al. (2007a)
PPARa activation only in mouse hepatocytes is sufficient to induce peroxisome proliferation and
increased DNA synthesis but not to induce liver tumors. Thus, "hepatocyte proliferation"
indentified by Klaunig et al. (2003) as a "causal event" in their PPARa MOA is not sufficient to
induce hepatocarcinogenesis. These data not only call into question the adequacy of the MOA
hypothesis proposed by Klaunig et al. (2003), but suggest only multiple mechanisms and also
multiple cell types may be involved in hepatocarcinogenicity caused by chemicals that are also
PPARa agonists.
02/12/09
E-298
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
3.4.1.3 Use of the PPAR-/- knockout and humanized mouse
Great importance has been attached to the results reported for PPARa -/- mice and their
humanized counterpart with respect to inferences regarding the MOA or peroxisome
proliferators and whether short-term chemical exposures or abbreviated bioassays conducted
with these mice can show that a PPARa MOA is involved. Consequently, the use of these
models warrants scrutiny. Compared to untreated wild type mice, liver weights in knockout
mice or humanized mice have been reported to be elevated (Voss et al., 2006; Laughter et al.,
2004; Morimura et al., 2006) and within 10% of each other (Peters et al.,. 1997). In order to be
able to assign affects to a test chemical tested in knockout mice, a better characterization is
needed of the baseline differences between PPARa -/- knockout and wild-type mice. This is
particularly important for examining weak agonists because the changed they induce may be
small and need to be confidently distinguished from differences due to the loss of the receptor
alone. As shown by the Ito et al. (2007) study and as noted by Maronpot et al. (2004), there is a
need for lifetime studies to characterize background or spontaneous tumor patterns and life spans
(including those of the background strain). While the original work by Lee et al. (1995)
describes "the mice homozygous for the mutation were viable, healthy, and fertile and appeared
normal," the authors did not describe the survival curves for this model nor their background
tumor rate. In fact, further work has shown that they carry a background of chronic conditions,
including: (1) chronic diseases such as obesity and steatosis (Akiyama et al., 2001; Costet et al.,
1998); (2) altered hepatic of hepatocellular structure and function, such as vacuolated
hepatocytes (Voss et al., 2006, Anderson et al., 2004), also seen in "humanized" mice (Cheung et
al., 2004); and (3) altered lipid metabolism, including reduced glycogen stores, blunted hepatic
and cardiac fatty acid oxidation enzyme system response to fasting, elevated plasma free fatty
acids, fatty liver (steatosis), impaired gluconeogenesis, and significant hepatic insulin resistance
(Lewitt el al., 2001). Howroyd et al. (2004) reported decreased longevity and enhancement of
age-dependent lesions in PPARa -/- mice.
These baseline differences from wild-type mice may render them more susceptible to
toxic responses or shorten their lifespans with chemical exposure. For example, after
administration of 250 microliters CCL4/kg, all male and 40% of female PPARa knockout mice
were dead or moribund after 2 days of treatment, whereas 25 % of male wild-type mice and none
of the female wild-type mice exhibited outward signs of toxicity (Anderson et al., 2004). Hays
et al. (2005) reported that 100% of PPARa knockout have cholestasis after 1 year of Bezafibrate
treatment with higher bile acid concentration than wild-type mice. Lewitt et al (2001) noted that
male knockout mice have more marked accumulation of hepatic fat, hypercholesterolemia and to
be particularly sensitive to fasting with some dying if fasted for more than 24 hours. Sexual
dimorphism but especially increased susceptibility of the male mouse has been reported for
knockout mice with pure Sv/129 backgrounds (Lewitt et al., 2001; Anderson et al 2004) as well
as those with a suggested C57BL/6N background (Djouadi et al., 1998, Costet et al., 1998).
Akiyama et al. (2001) showed an apparent greater sexual dimorphism in mice with a pure Sv/129
background than C57BL/6N in regard to weight gain from 2 to 9 months but not in changes in
body weight or liver weight between wild-type and knockout animals. Adipose tissue, serum
02/12/09 E-299
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
triglycerides and cholesterol were altered in the knockout animals. Given that the experiment
was only carried out for 9 months, changes in body fat, liver weight and lipid levels may be
greater as the animals get older and steatosis is more prevalent. The dramatic effect on survival
as well as gender difference by the increased expression of lipoprotein lipase in the PPARa
knockout mouse with further genetic modification is demonstrated by Nohammer et al (2003)
who reported 50% mortality in 6 months and 100% mortality within 11 months of age while
females survived. These differences could affect the results of tumor induction for PPARa
agonists with less potency than WY-14,643 that do not produce tumors so rapidly. In addition
these studies suggest the need for careful consideration of the effects of use of different
background strains for the knockout and the need for careful characterization of the background
responses of the mouse model and the effects of the use of different background strains for the
knockout. Morimura et al. (2006) reported that, using the B6 background strain, there were only
foci at time periods but knockouts with the SV129 background had multiple tumors after WY-
14,643 treatment.
PPARa knockout mice have also been used to examine the dependence on PPARa on
changes in cell signaling, protein production, or liver weight. However, to be useful, the changes
incurred just by loss of the PPARa should also be well described. Reported differenced between
PPARa-knockout and wild-type mice can impact the sensitivity and specificity of these markers
of for the hypothesized MOA. In regards to altered cell signaling, Wheeler et al. (2003) note that
in normal cells p21waf and p27kipl inhibit the Cdk/cyclin complexes responsible for cell cycle
progression through Gl/S transition. While these cellular signaling molecules are down-
regulated in response to partial hepatectomy in normal mice, they remain elevated in PPARa
knockout mice along with decreased DNA synthesis. Fumonisins are hepatocarcinogens that
have been associated changes in apoptosis and tissue generation, and increased acyl-CoA
oxidase and CYP4A (markers of PPARa activation) (Martinez-Larranaga et al., 1996). Voss et
al. (2006) report that the average number of hepatic apoptotic foci per mouse induced by
Fumonisins were 3-fold higher and liver mitotic figures counts were 2-fold lower in PPARa
knockout in comparison to wild-type mice, thus illustrating a difference in proliferative response
in the mice. PPARa-null mice have increased apoptosis and decreased mitosis with fumonisin
treatment. Voss et al. (2006) also report several differences in gene expression in wild-type and
PPARa knockout mice that ranged from 0.3% to 483% of the activity of wild-type mice. The
complex expression patterns of gene expression and determining their mechanistic implications
in regard to hepatotoxicity and carcinogenicity are difficult. Certainly the vast number of genes
whose expression is affected by WY-14,643 (1012 genes as cited by Voss et al., 2006) illustrates
such complexity. Voss et al. (2006) conclude that studies should consider dose- and time course-
related effect as well as species and strain-related differences in the expression of gene products.
The "humanized" PPARa mouse has a human copy of PPARa inserted into a PPARa
knockout mouse. It is inserted in a tetracycline response system so that in the absence of DOX
only human PPAR alpha is transcribed in humanized mouse liver and not in other tissues. A
rigorous examination of newly emerging studies regarding the "humanized" mouse is warranted.
There are two papers that have been published using the humanized PPARa mouse (Cheung et
02/12/09
E-300
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
al., 2004; Morimura et al., 2006). Many of the issues described above for PPARa -/- mice are of
concerned for the humanized knockout mouse. In addition, the placement of the humanized
PPAR gene is a potential confounding factor, as discussed by Morimura et al. (2006): "It also
cannot be ruled out that the hPPARa mice are resistant to the hepatotoxic effects of peroxisome
proliferators due to the site of expression of the human receptor. The cDNA was placed under
control of the tetracycline regulatory system and the liver-specific Cebp/B promoter that is
preferentially expressed in hepatocytes."
In the Cheung et al. (2004) report, the humanized mouse was fed WY-14,643 for 2 or 8
weeks (age not given for the mice). WY-14,643 and Fenobrate were reported to decrease serum
total triglyceride levels in wild and humanized mice to about the level seen in PPARa -/- mice
(which were already suppressed without treatment). Hepatomegaly and increase in hepatocyte
size were observed in the PPARa -humanized mice fed WY-14,643 for 2 weeks but less than
wild mice. By contrast, Morimura et al., (2006) state that the humanized mice did not exhibit
hepatomegaly after treatment with WY-14,643. Cheung et al (2004) present figures that show
increased vacuolization of hepatocytes in a control humanized mouse in comparison to wild type
mice. Vacuolization increased with WY-14,643 treatment in the humanized mouse. Therefore
there was a background level of liver dysfunction in these mice even with humanized PPARa.
Vacuolization is consistent with fatty liver observed in the non-humanized PPARa-/- mouse.
The authors reported that the humanized mouse did not have increased #s of peroxisomes after
WY treatment. However, they present a figure for genes encoding peroxisomal, mitochondrial,
and microsomal fatty acid oxidation enzymes that shows they were still markedly increased in
PPARa -humanized mice following 8 weeks of exposure to WY-14,643. Therefore there is a
paradox in these reported results.
Morimura et al. (2006) provided a useful example to illustrate the many issues associated
with interpreting studies with genetically-altered animals. While this study is suggestive of a
difference in susceptibility to tumor induction between wild-type and PPARa humanized mice, a
conclusion that human PPARa is refractory to liver tumor induction is not sufficiently supported
by this study. This study had uneven durations of exposure and follow-up and reported
substantial toxicity or mortality that limit the interpretation of the observed tumor rates. For
example, the 6 week-old male "humanized" mice had a 44-week experimental period but for
"wild-type" mice that period was 38 weeks. In addition, for humanized mice, 10 mice were
treated with 0.1% WY-14,643 with 20 controls, but for wild-type mice, 9 mice were given 0.1%
WY with 10 controls. Furthermore, wild-type, WY-14,643-treated animals has suppressed
growth and only a 50% survival to 38 weeks, so an effective LD50 has been used for this length
of exposure. Specifically, of the 10 wild-type WY-14,643 treated mice, 3 died of toxicity and 2
were killed due to morbidity and their tissues examined. Humanized mice had a similar growth
for animals treated with WY-14,643 and controls with only one mouse killed because of
morbidity. Therefore the reported results, including tumor numbers, are for a mixture of
different exposure durations and ages of animals. In addition the results of the study were
reported for only on exposure level.
02/12/09
E-301
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Furthermore, it is interesting that while control humanized mice had no adenomas, WY-
14,643 treated humanized mice had one. Morimura et al. (2006) noted that this adenoma had a
morphology "similar to spontaneous mouse liver tumor with basophilic and clear hepatocytes,"
whereas the tumors in wild-type mice treated with WY-14,643 were more diffusely basophilic.
If the humanized animals were allowed to liver their natural lifespan, this raises the possibility
that WY-14,643 may induce tumors that are similar to other carcinogens rather than those which
have been characterized as "characteristic" of peroxisome proliferators (see Section 3.5.1.5)
when human PPARa is present. Therefore the humanized PPARa rather than mouse PPARa
may have an association with a tumor phenotype characteristic of other MO As but this study
need to be carried out for a longer period of exposure and with more animals to make that
determination. The baseline tumor response of PPARa humanized mice needs to be
characterized as well as tumors exposure to WY-14,643 or other carcinogens acting through
differing MO As. The numbers of foci were not reported, but "altered foci" were detected in one
humanized mouse with WY-14,643 treatment and one without treatment. The phenotypes of the
foci were not given by the authors
As discussed above, changes in liver weights have been used as a marker for
susceptibility to liver tumor induction and the issues regarding baseline differences in PPARa -/-
mice are equally relevant for PPARa humanized mice. Morimura et al. (2006) reported that
absolute liver weight for control humanized mice at 44 weeks was 1.57 g (n=10). The absolute
liver weight for wild control mice was 1.1 g (n=9) at 38 weeks. The final body weights differed
by 14% but liver weights differed by 30%. Therefore, even though comparing different aged
mice, the control humanized mice had greater liver size than the wild type control mice on an
absolute and relative basis. This is consistent with humanized knockout mice having greater
sized livers and a baseline of hepatomegaly. With treatment, Morimura et al. (2006) report that
PPARa humanized mice treated with WY-14,643 had greater absolute and relative liver weights
than controls but less elevations than wild-type treated animals. However, because half of the
wild-type animals died, it is difficult to discern if liver weights were reported for moribund
animals sacrificed as well as animals that survived to 38 weeks for wildtype mice treated with
WY-14,643. However, it appears that moribund animals were included that were sacrificed early
for treated groups and that values from the animal killed at 27 weeks were added in with those
surviving till 45 weeks in the PPARa humanized mice treated with WY-14,643 group.
With respect to the gene expression results reported by Morimura et al. (2006), it is
important to note that they are for liver homogenates with a significant portion of the nuclei from
nonparenchymal cell of the liver (e.g., Kupffer and stellate cells). Thus, the results represent
changes resulting from a mixture of cell types and from differing zones of the liver lobule, with
potentially different gene changes merged together. Livers without macroscopic nodules were
used for western blot and but could have contained small foci in the homogenate as well. The
gene expression results were also reported for an exposure level of WY-14,643 that is an LD50
in wild type mice and could reflect toxicity responses rather than carcinogenic ones. The
samples were also obtained at the end of the experiment (with a mix of durations of exposure)
and may not reflect key events in the causation of the cancer but events that are downstream.
02/12/09
E-302
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
These limitations notwithstanding, it is interesting that expression of p53 gene was
reported by Morimura et al. (2006) to be increased in PPARa humanized mice treated with WY-
14,643 compared to all other groups. Furthermore, of the cell cycle genes that were tested, (i.e.,
CD1, Cyclin-dependent Kinases 1 and 4, and c-myc ) there was a slightly greater level of c-myc
and CD1 in control PPARa humanized mice than control wild type mice as a baseline. This
could indicate that there was already increased cell cycling going on in the control PPARa
humanized mouse and could be related to the increased liver size. Treatment with WY-14,643
induced an increase in cycling genes in wild-type mice in relation to its control, but whether that
induction was greater than control levels for PPARa humanized mice for c-myc and CDk4 was
not reported by the authors. Apoptosis genes were reported to have little difference between
control PPARa humanized and wild type mice but to have a greater response induced by WY-
14,643 in humanized mice forp53 andp21. There was no consistent or large change in
apoptosis genes in response to exposure to WY-14,643 in wild type mice. The increased
response of apoptosis genes in PPARa humanized mice without corresponding tumor formation
does not support that response as a key event in the MOA (neither does the lack of response from
WY-14,643 in wild type mice). For genes associated with PPARa peroxisomal (Acox),
microsomal (CYP4a) mitochondrial fatty oxidation (Mead) and especially malic enzyme, there
was a greater response in wild type than PPARa humanized mouse after treatment with WY-
14,643. However, this is somewhat in contrast to Cheung et al. (2004), who reported increased
in some genes encoding peroxisomal, mitochondrial, and microsomal fatty oxidation enzymes in
the PPARa humanized mouse after treatment with WY-14,643.
The results reported by Yang et al. (2007b) use another type of "humanized" mouse to
study PPARa effects. Yang et al. (2007b) used a PPARa humanized transgenic mouse on a
PPAR-/- background that has the complete human PPARa (hPPARa) gene on a PAC genomic
clone, introduced onto the mouse ppara-null background and express human PPARa not only in
the liver but also in other tissues. Mice were administered WY-14,643 or Fenofibrate [0.1% or
0.2% (w/w)]. The authors show a figure representing expression of the hPPARa for two mice
with the tissue used for the genotyping with great variation in expression between the two cloned
mice indicated by intensity of staining. The authors state that in agreement with RNA
expression, hPPARa protein was highly expressed in the liver of hPPARaPAC mice to an extent
similar to the mPPARa in wild-type mice. They report that following two weeks of Fenofibrate
treatment, a robust induction of mRNA expression of genes encoding enzymes responsible for
peroxisomal (ACOX), mitochondrial (MCAD,LCAD), microsomal (CYP4A) and cytosolic
(ACOT) fatty acid metabolism were found in liver, kidney and heart of both wild-type and
hPPARaPAC mice indicating that hPPARa functions in the same manner as mPPARa to regulate
fatty acid metabolism and associated genes. However, the authors did no measure in Fenofibrate
treated animals, only WY-14,643 raising the issue of whether there a difference in the relative
mRNA expression of genes for ACOX etc and lipids between the two peroxisomal proliferator
treatments. The expression of enzymes associated with PPAR induction was presented only for
mice treated with Fenofibrate. However, the lipids results were presented only for mice treated
with WY-14,643. Therefore it cannot be established that these two agonists give the same
response for both parameters. Also for the enzymes the relative expressions to wild type controls,
02/12/09
E-303
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the absolute expression and its variation between animals is not reported. It appears that the
peroxisomal enzyme induction by Fenofibrate is the same in the wild type and transgenic mice.
However, in Figure 4 of the paper the mice treated with WY-14,643 instead of Fenofibrate were
presented for the peroxisome proliferator marker enzyme PMP70 in total liver protein gel. There
appear to be more PMP70 in the transgenic mice than wild type mice as a baseline. The PMP70
appeared to be similar after WY-14,643 treatment. However, only one gel was given and no
other quantitation was given by the authors.
The authors state that "in addition WY-14,643 and Fenofibrate treatment produced
similar effect to the liver specific humanized PPARa mouse line (Cheung et al 2004)."
However, the results were not the same between Fenofibrate and WY-14,643 and the mouse line
used by Cheung et al. had background differences in response and pathology. In one figure in
the paper there appears to be a difference in background level of serum total triglyceride between
the wild type and hPPARaPAC mice that the authors do not note. The power of using such few
mice does not help to sort out any significant differences in background level of triglycerides.
The authors note that WY-14,643 treatment also resulted in decreased serum triglycerides levels
in hPPARaPAC mice consistent with the induction of expression of genes encoding fatty acid
metabolism and that the hypolipidemic effects of fibrates are generally explained by increased
expression of lipoprotein lipase (LPL) and decreased expression of apolipoprotein C- III (Apo
C-III) (Auwerx et al., 1996). However, the alteration of these genes by WY-14,643 treatment
was only observed in wild type mice and not in hPPARaPAC mice suggesting that the
hypolipidemic effect observed in hPPARaPAC mice are not through LPL and APO C-III. The
authors do not note that there could be a difference in the regulation of these pathways by the
transgene rather than how the normal gene is regulated and the pathways it affects. The reason
for examining this question with WY-14,643 treatment rather than with Fenofibrate treatment is
not addressed by the authors, especially since the other "markers" of peroxisomal gene induction
appear to be affected by Fenofibrate in the wild type and hPPARaPAC mice.
Hepatomegaly was reported to be observed in the hPPARaPAC mice following two weeks
of WY-14,643 treatment as revealed by the increase liver to body weight ratio compared to
untreated hPPARaPAC mice but to be markedly lower when compared to wild type mice under
the same treatment. Histologically, the livers of the wild type mice treated with WY-14,643
were hypertrophic with clear eosinophilic regions with these phenotypic effects were observed in
both wild-type and hPPARaPAC mice. The % liver/body weight was reported to increase from ~
4% in wild type mice to ~ 9% after WY-14,643 treatment and from ~ 4% in hPPARaPAC to little
less that 6% after treatment with WY-14,643. In wild type mice treated with WY-14,643 the
labeling index was 21.8% compared with 1.1% in untreated wild type controls. In hPPARaPAC
mice, WY-14,643 treatment was reported to give an average labeling index of 1.0% compared
with 0.8% in the untreated control hPPARaPAC mice. Treatment with WY-14,643 treatment was
reported to result in a marked induction in the expression of CDK4 and cyclin D1 in the livers of
wild-type mice but to be unaffected hPPARaPAC mice treated with WY-14,643. These data
were reported to be in agreement with the liver-specific PPARa-humanized mice that showed not
increase in incorporation of BrdU into hepatocytes upon treatment with WY-14,643 (Cheung et
02/12/09
E-304
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
al. 2004) and further confirmed that activation of hPPARa dose not induce hepatocyte
proliferation. However, the authors present a figure an example with one liver each with no
quantitation given by the authors for BrdU incorporation. It is not clear whether the pictures
were taken from the same area of the liver or how representative they are. The numbers of mice
were never reported for the labeling index. The data presented do suggest that there was
hypertrophy and hepatomegaly in the humanized mice and but not proliferation in this particular
WY,-14643 model. Of interest would be investigation of proliferation by other peroxisome
proliferators besides WY-14,643 at this necrogenic dose as it is WY that is the anomaly to
continue with proliferation at 2 weeks. The photomicrographs presented by the authors are so
small and at such low magnification that little detail can be discerned from them. There are no
portal triads or central veins to orient the reader as to what region of the liver has been affected
and where if any there would be hepatocellular vacuolization
To determine whether peroxisome proliferation occurred in the hPPARaPAC mice upon
administration of PPs, Yang et al. (2007b) examined by Western Blot analysis the protein levels
of the major peroxisomal membrane protein 70 (PMP70) a marker of peroxisome proliferation).
After two weeks treatment of 1000 ppm WY-14,643, induction of PMP70 was reported to be
observed in the wild type mice as well as in hPPARaPAC mice. The authors suggested that this
result indicates that peroxisomal proliferator treatment induced peroxisomal proliferation in
hPPARaPAC mice. The results of this study indicate that hepatomegaly and peroxisome
proliferation occur in this humanized mouse model when treated with large concentrations of
WY-14,643. Thus these results are inconsistent with claims that peroxisome proliferators cannot
cause hepatomegaly or peroxisome proliferation in humans or that humans are refractory to these
effects. Like the lipid effects, they show a broader spectrum of effects may occur in humans and
decreases the specificity of these effects as species specific. However due to the model
compound being WY-14,643 at a necrogenic dose of 1000 ppm, the effect may not be seen in
humans using the lower potency peroxisome proliferators. It would have been useful for this
study to include and examination of these effects with Fenofibrate rather than WY-14,643 and
then attempting to extrapolate such effects to other peroxisome proliferators. The authors often
attribute the effects of peroxisome proliferators to those reactions induced by WY-14,643 and do
not acknowledge that the changes induced by WY-14,643 may be different -especially in regards
to hepatocellular DNA synthesis in which other peroxisome proliferators can cause liver tumors
without sustained proliferation that WY-14,643 induces, especially at a necrogenic dose
Yang et al. (2007b) report the results of induction of various genes by WY-14.643 in wild
type and hPPARaPAC mice by microarray analysis followed by confirmation and quantitation by
qPCR and report that more genes were induced by WY-14,643 in wild type mice than in
hPPARaPAC mice. They report that "importantly, the oncogene c-myc was not induced in
hPPARaPAC mice. Moreover, genes encoding cell surface proteins such as Anxa2, CD39, CD63,
Ly6D, and CD24a, and several other genes such as Cidea, Cidec, Dhrs8 andHsdllb were also
not induced in hPPARaPAC mice. Interestingly, Sult2al was only induced in hPPARaPAC mice
and not in WT mice; this gene is also induced in human hepatocytes by PP (Fang et al., 2005).
The regulation of several of these genes has previously been demonstrated through a PPARa-
02/12/09
E-305
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
dependent mechanism. Additional studies will be necessary to fully explore the molecular
regulatory mechanism and the functional implication associated with these differently regulated
genes." As stated in Section 3.2.2 above, there are several limitations for interpretations of the
results such as those presented by Yang et al. (2007b) which include the lack of phenotypic
anchoring for the results. The authors have shown changes from whole liver and have listed
changes in genes between wild type and humanized mice on a PPAR -/- background that in itself
with bring about changes in gene expression. The authors acknowledge difficulties in
determining what their reported gene changes mean. The authors do not indicated the context of
how the mice were treated, whether these are pooled results, and when the samples were taken.
It is assumed to be whole liver.
Yang et al. (2007b) report that "activation of PPARa alters hepatic miRNA expression
(Shah et al., 2007). They report that let-7C, a miRNA critical in cell growth and shown to target
c-myc, was inhibited by WY-14,643 treatment in wild type mice and that the expression levels of
both pri-let-7C and mature let-7C were significantly higher in hPPARaPAC mice compared to
wild type mice. Treatment with WY-14,643 was reported to decrease the expression of Pi-let-7C
and mature let-7C in wild type mice but in hPPARaPAC mice. The authors note that "in addition,
the induction of c-myc by WY-14,643 treatment in wild type mice did not occur in WY-14,643
treated hPPARaPAC mice. This is in agreement with the previous observation in liver-specific
humanized PPARa (Shah et al 2007) and further indicates the activation of human PPARa does
not cause a change in hepatic miRNA and c-myc gene expression." A qPCR analysis of pri-let-
7C following 2 weeks WY-14,632 treatment was reported for wild type and hPPARaPAC mice
(n=3-4). There appeared to be ~ 20 times more Let-7C expression in hPPARaPAC mice than
control wild mice as a baseline. The gel given by the authors showed a very small difference in
wild type mice in Let-7C northern blot analysis between a control wild type and WY-14,643-
treated wild type mouse. There appeared to be no difference in the hPPARaPAC mice between
control and WY-14,643 treatment and a larger stained area than the control wild type mice. The
relative c-my expression between the hPPARaPAC mice and wild type control mice did not
correlate with changes in Let-7C expression. Thus, the amount of decrease by treatment with
WY-14,632 in wild type mice appeared to be extremely small compared to the much greater
baseline expression in the hPPARaPAC mice. The change brought by WY-14,632 treatment in
wild type mice was a small change compared to the 20-fold greater baseline expression in the
hPPARaPAC mice. The authors stated that the expression of the c-Myc regulator was higher in the
hPPARaPAC mice indicating over regulation of cell division and an inability for hepatocytes to
proliferate. However, their results showed that there was a greater difference in regulatory
baseline function of the PPAR using this paradigm and this construct. Are these differences due
to human PPAR or to the way PPAR was put back into PPAR-/- mouse and expected to
function? If the experiment included mouse PPAR put back in this way on a null background,
what would such an experiment show? Are these results representative of the PPAR or how it is
now controlled and expressed? In addition, what would the study of other peroxisome
proliferators besides WY-14,643 show in regard to changes in miRNA. Are these results
reflective of a just the transient effect that is prolonged in a special case? As discussed in
Section 3.2.2. there are issues with microarray data in addition to the newly emerging field of
02/12/09
E-306
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
miRNA arrays which include phenotypic anchoring and whether they are from whole liver or
pooled samples. The results given in this report are for relative Let-7C expression given and not
absolute values. The changes in baseline Let-7C expression between the wild type and the
hPPARaPAC mice did not correlate with the magnitude of difference in northern blot analysis and
did not correlate at all with c-myc expression reported in this study. Thus, a direct correlation
between the effect of Let-7C expression and function and effects from WY-14,643 was not
supported. The relative expression was reported but the variation of baseline expression of the
"PPAR controlled genes" was not. Given that one of the first figures reported a large difference
between animals in expression of the human PPAR gene in the transgenic animals, how did this
difference affect the results given here as relative changes downstream?
Yang et al. (2007b) conclude that the hPPARaPAC mice represent the most relevant model
for humans since, the tissue distribution of PPARa is similar to that observed in wild-type mice
and the hPPARa in hPPARaPAC mice is under regulation of its native promoter. Indeed up-
regulation of hepatic mPPARa in wild type mice by fasting was mirrored by the hPPARa in
hPPARaPAC mice. However, there was no demonstration that the artificial chromosome that is
replicating along with other DNA is controlled sterically by the same control since it is not on
the mouse genome in the same place as the native PPAR. There is also not a demonstration of
how stable the baseline of PPAR DNA expression is in this mouse model - does it vary as much
or more than native PPAR between mice? The authors state that "induction of PPARa target
genes for fatty acid metabolism and a decrease in serum triglycerides by PP in hPPARaPAC mice
indicates that hPPARa is functional in the mouse environment with respects to regulation of fatty
acid metabolism. This is in agreement with the liver-specific PPARa humanized mice that also
exhibit these responses (Cheung et al., 2004). Indeed the DNA binding domain of hPPARa is
100% homologous with that of the mouse suggesting that both bind to the same PPRE binding
site in the promoter region of target genes. Transfection of hPPAR into murine hepatocytes
increased PPs induced peroxisome proliferation related effects (Macdonald et al., 1999). These
results suggest that hPPARa and mPPARa do not differ in induction of target genes with known
PPRE." However, replacement with human PPAR in the Cheung et al. model is not sufficient to
prevent the same types of toxicity as seen with PPAR knockouts on the hepatocytes such as
steatosis.
Yang et al. (2007b) note that "the increased LPL an decreased expression of apo C-III are
proposed to explain the hypolipidemic effects of PPS (Auwerx et al., 1996). However,
hPPARaPAC mice treated with PP exhibit lowered serum triglycerides without alteration of the
expression of LPL and apoC-III. This indicates the hypolipidemic effects in rodents are
mediated via other molecular regulatory mechanisms. It is also suggested that the activation of
PPARa by PPs stimulates hepatic fatty acid oxidation and thereby diminishing their
incorporation into triglycerides and secretion of VLDL (Froyland et al., 1997). Consistent with
this idea, a robust induction of the genes encoding enzymes for fatty acid oxidation by PP in
hPPARaPAC mice were observed. Thus, the exact mechanism by which PPs exert their
hypolipidemic effects needs reexamination." However, the use of two different peroxisome
proliferators (i.e., WY-14,643 and Fenofibrate) for two types of effects (peroxisomal and lipid)
02/12/09
E-307
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
may be the cause of some paradoxes here in terms of MO A for lipid effects. The baseline
differences in the hPPARaPAC mice for serum total triglicerides was not explored by these
authors and the small number of animals used make conclusions difficult about the magnitude of
difference. The differences in baseline expression for LPL are not discernable in the graphic
representation of the results.
Yang et al. (2007b) note that "on the other hand, the difference in the affinity of ligands
for the human and mouse PPARa receptor was proposed to account for the species difference.
The ligand binding domain of hPPARa is 94% homologous with that of the mouse. In vitro
transactivation assays have previously shown that WY has a higher affinity for rodent PPARa
than human PPARa, while Fenofibrate has similar affinity for rodent and human PPARa
(Shearer and Hoekstra 2003; Sher et al., 1993). In the present study WY and Fenofibrate exhibit
the same capacity to induce known PPARa target genes in the liver, kidney and heart in both
wild-type and hPPARaPAC mice." The statement by the authors that Fenofibrate and WY-
14,643 had the same affinity "as shown by this study" is not true. The two treatments were not
studied for the same enzymes or genes in the data reported in the study. Both WY-14,643 and
Fenofibrate can induce PPARa targets but it was not shown to the same extent. Yang et al.
(2007b) state that "This is in agreement with the liver-specific PPARa humanized mice that also
exhibit a similar capacity to induce PPARa target genes in liver by WY and Fenofibrate (Cheung
et al., 2004). Thus, the ligand affinity difference between mouse and human PPARa may not be
critical under the conditions of these studies." Alternatively, these results could reflect that these
studies were conducted with two different agonists with different affinities and responses due to
receptor activation.
Finally, a useful comparison to make are the differences between wild type mice, PPARa
-/- mice that serve as the background for the transgenic human mouse models, and both
transgenic models. The small and variable number of animals examined in these studies is
readily apparent. The results of the Cheung et al. (2004) humanized mouse model and those
reported for Yang et al. (2007b) show differences in the study designs including PPARa
agonists studied for particular effects and results reported for similar treatments.
02/12/09
E-308
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Comparison between results for Yang et al. 2007b and Cheung et al. 2004:**
Effect
Wild type
mice
PPAR -/- knockout mice
Humanized mice
(liver only)
Humanized PAC
mice
Triglycerides
Cheung
(n=6-9)
Control 145 mg/ml
0.1% WY-14,643 60mg/ml
(2 wks)
0.2% Fenofibrate 85 mg/ml
(2 wks)
Yang
(n=4-6)
Control 95 mg/ml
0.1 % WY-14,643 55mg/ml
(2wks)
Cheung
(n=6-9)
Control 100 mg/ml
0.1% WY-14,643 115 mg/ml
(2 wks)
0.2% Fenofibrate 85 mg/ml
(2 wks)
Cheung
(n=6-9)
Control 175 mg/ml
0.1%WY-14,643 60 mg/ml
(2 wks)
0.2% Fenofibrate 85 mg/ml
(2 wks)
Yang
(n=4-6)
Control 120 mg/ml
0.1%WY-14,643 75 mg/ml
(2 wks)
BrdU
incorporation
Cheung
(n=5)
Control 1.6%
0.1% WY-14,643 57.9 %
(8 wks)
Yang
(n=4-6)
Control 1.1%
0.1% WY-14,643 21.8%
(2 wks)
Not done
Cheung
(n=5)
Control 1.6 %
0.1% WY-14,643 2.8%
(8wks)
Yang
(n=4-6)
Control 0.8 %
0.1% WY-14,643 1.0%
(2 wks)
Hepatomegaly*
(% liver body
weight ratio)
Cheung
(n=5-9)
Control 4 %
0.1% WY-14,643 11 %
(2 wks)
0.2% Fenofibrate 8.5%
(2 wks)
Cheung
(n=5-9)
Control 5 %
0.1% WY-14,643 5 %
(2 wks)
0.2% Fenofibrate 5.5%
(2 wks)
Cheung
(n=5-9)
Control 4.5 %
0.1% WY-14,643 7 %
(2 wks)
0.2% Fenofibrate 7 %
(2 wks)
02/12/09
E-309
-------
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Yang
Yang
(n=4-6)
(n=4-6)
Control 4 %
Control 4 %
0.1% WY-14,643 9 %
0.1% WY 6%
(2 wks)
(2 wks)
*Percentages are approximate values extrapolated from figures for hepatomegaly.** The ages of the humanized knockout mice are not given
for Cheung et al. 2004 but are 8-10 weeks for Yang et al. 2007b.
02/12/09
E-310
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
As shown above, the effect on the PPARa-/- by the knockout included decreased
triglyceride levels and slightly increased liver weight. Although treatment with WY-14,643 and
Fenofibrate were reported to decrease triglyceride levels in wild-type mice, paradoxically so did
knocking out the receptor. Exposures to WY-14,643 appeared to induce a slight increase and
Fenofibrate a slight decrease in triglyceride levels in PPARa -/- mice but the variability of
response and small number of animals in the experiments limited that ability to discern a
quantitative difference in the treatments. In the study by Cheung et al. (2004) it appears that the
insertion of humanized PPARa restored the baseline and treatment responses for triglyceride
levels. Overall, the results reported by Yang et al. (2007b) appeared to have a lower level of
triglycerides in control wild type mice that was similar in magnitude to the treatment effect
reported by Fenofibrate by Cheung et al. (2004). However, there also appeared to be restoration
of this effect in the humanized mouse model of Yang et al. (2007b). In regard to DNA synthesis,
both Cheung et al. (2004) and Yang et al. (2007b) only gave results for WY-14,643 and for
different durations of exposure so they were not comparable. It appeared that ~ 60 % of
hepatocytes were labeled by 8 weeks of WY-14,643 treatment (Cheung et al., 2004) compared to
~ 20% after 2 weeks of exposure. Again this highlights the difference between using WY-14,643
as a model for the PPARa as a class at times when almost all other PPARa agonists have ceased
to increase DNA synthesis or have reductions in this parameter. The background changes due to
the PPARa -/- knockout were not reported so that the effects of the knockout could not be
ascertained. It appeared that insertion of humanized PPARa did not results in restoration of WY-
14,643 -induced DNA synthesis. The correlation with this parameter and any focal areas of
necrosis were not discussed by the authors of the study. In regard to hepatomegaly, Fenofibrate
and WY-14,643 appeared to both give an increase in liver weight in the humanized mouse model
of Cheung et al., (2004) with little effect in the knockout mouse. For Fenofibrate there was little
difference in liver weight gain in the wild type mouse and that of the humanized mouse model of
Cheung et al., (2004). However, Fenofibrate was not tested in the humanized mouse model of
Yang et al. (2007b). In that model only WY-14,643 was used but there was still an increase in
liver weight. Thus, in terms of effects on liver weight gain and triglyceride levels both models
gave comparable results and appeared to indicate that insertion humanized PPARa would restore
some of the effects of the knockout. However, the results from both experiments highlight the
need for adequate numbers of animals and other PPARa agonists to be tested besides WY-14,463
at such a high dose and certainly for longer periods of time to ascertain whether such
manipulations will affects carcinogenicity.
3.4.1.4 NF-kB activation
NF-kB activation has also been proposed as a key event in the induction of liver cancer
through PPARa activation. As discussed in Sections 3.2.6. and 3.4.3.3., activation of the NF-kB
pathway is implicated in carcinogenesis, non-specific for a particular MOA for liver cancer, and is
context dependent on its effects. Its specific actions depend on the cell type and type of agent or
signal that activates translocation of the complex. NF-kB is not only involved in biological
processes other than tumor induction, but also exhibits some apparently contradictory behaviors
(Perkins and Gilmore, 2006). Although many studies point to a tumor-promoting function of NF-
02/12/09
E-311
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
kB subunits, evidence also exists for tumor suppressor functions. NF-kB actions are associated
with TNF and INK among many other cell signaling systems and molecules and it has functions
that alter proliferation and apoptosis. NF-kB activation reported in some studies may be
associated with early Kupffer cell responses and be associative but not key events in the
carcinogenic process. However, most assays look at total NF-kB expression in the whole liver
and at the early periods of proliferation and apoptosis. The origin of the NF-kB is crucial as to its
effect in the liver. For instance hepatocyte specific deletion of IKKp increased DEN-induced
hepatocarcinogenesis but a deletion of IKKp in both hepatocytes and Kupffer cells however, had
the opposite effect (Maeda et al., 2005).
3.4.1.5. Phenotype as an indicator of a PPARa MO A
As discussed previously (see Sections 3.1.5, and 3.1.8.) foci of altered hepatocytes (FAH)
precede both hepatocellular adenomas and carcinomas in rodents and, in humans with chronic
liver diseases that predispose them to hepatocellular carcinomas. Striking similarities in specific
changes of the cellular phenotype of preneoplastic FAH are emerging in experimental and human
hepatocarcinogenesis, irrespective of whether this was elicited by chemicals, hormones,
radiation, viruses, or, in animal models, by transgenic oncogenes or Helicobacter hepaticus.
Several authors have noted that the detection of phenotypically similar FAH in various animal
models and in humans prone to developing or bearing hepatocellular carcinomas favors the
extrapolation from data obtained in animals to humans (Bannasch et al., 2003; Su and Bannasch,
2003; Bannasch et al., 2001). In regard to phenotype by tincture Caldwell and Keshava (2006)
state:
In addition, the term "basophilic" in describing preneoplastic foci or tumors
can be misleading. The different types of FAH have been related to three
main preneoplastic hepatocellular lineages: 1) the glycogenotic-basophilic
cell lineage, 2) its xenomorphic-tigroid cell variant, and 3) the amphophilic-
basophilic cell lineage. Specific changes of the cellular phenotype of the
first two lineages of FAHs are similar in experimental and human
hepatocarcinogenesis, irrespective of whether they were elicited by DNA-
reactive chemicals, hormones, radiation, viruses, transgenic oncogenes and
local hyperinsulinism as described by the first two FAH's and this
similarity favors extrapolation from data obtained in animals to humans
(Bannasch et al., 2003; Su and Bannasch, 2003; Bannasch et al., 2001). In
contrast, the amphophilic cell lineage of hepatocarcinogenesis has been
observed mainly after exposure of rodents to peroxisome proliferators or to
hepadnaviridae (Bannasch et al., 2001).
Bannasch (1996) describes "amphophilic" FAH and tumors induced by
peroxisome proliferators to maintain the phenotype as the foci progress to
tumors. They are glycogen poor from the start with increased numbers of
mitochondria, peroxisomes and ribosomes. The author further states that
02/12/09
E-312
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
the "homogenous basophilic" descriptions by others of foci induced by WY
are really amphophilic. Agents other than peroxisome proliferators can
induce "acidophilic" or "eosinophilic" (due to increased smooth
endoplasmic reticulum) or glycognotic foci which tend to progress to
basophilic stages (due to increased ribosomes).
Tumors and foci induced by peroxisome proliferators have been suggested
to have a phenotype of increased mitochondrial proliferation and
mitochondrial enzymes (thyromimetic rather than insulinomimetic)
(Keshava and Caldwell, 2006).
Tumors from peroxisome proliferators in Kraupp-Grasl et al. (1990) and Grasl-Kraupp et
al. (1993) for rat liver tumors were characterized as weakly basophilic with some eosinophilia
and as similar to the description given by Bannasch et al as amphophilic. However, a number of
recent studies indicate that other "classic" peroxisome proliferators may have a different
phenotype than has been attributed to the class through studies of WY-14,643. A recent study of
di (2-ethylhexyl) phthalate (DEHP), another peroxisome proliferator assumed to induce liver
tumors through activation of the PPARa receptor, reported the majority of liver FAH to be of the
first two types after a lifetime of exposure to DEHP with a dose-related tendency for increased
numbers of amphophilic FAHs in rats (Voss et al., 2005). As stated previously, the MOA of
DEHP-induced liver tumors in mice also appears to not be dependent on PPARa activation.
Michel et al. (2007) report the phenotype of tumors and foci in rats treated with clofibric
acid at a very large dose (5000 ppm for 20 months) and note that in controls the first type of foci
to appear was tigroid on day 264 and their incidence increased with time representing the most
abundant type in this group. They report no adenomas or carcinomas after up to 607 days after
given saline injection in the control animals. DEN treatment was examined up to 377 days only
with tigroid, eosinophilic and clear cell foci at that time observed at that time. Clofibric acid was
examined up to 607 days and tigroid and clear cell foci were reported to be the first to appear on
day 264 with no other foci class. By day 377 there were tigroid, eosinophilic and clear cell foci
but no basophilic foci reported with clofibric acid treatment and, although only a few animals
were examined, 2/5 had adenomas but not carcinomas. By day 524 all types of foci were seen
(including basophilic for the first time) and there were adenomas and carcinomas in 2/5 animals.
By 607 days a similar pattern was observed without adenomas but 3/6 animals showing
carcinomas. Although the number of animals examined is very small, these results indicate that
clofibric acid was not inducing primarily "basophilic foci" as reported for peroxisome
proliferators but the first foci are tigroid and clear cell foci. Basophilic foci did not appear till
day 524 similar to control values for foci development and distribution. However unlike
controls, clofibric acid induced eosinophilic and clear cell foci earlier. This is inconsistent with
the phenotype ascribed to peroxisome proliferators as exemplified by WY-14,643.
In regard to y-glutathione -S-transferase (GST-71) and y-transpeptidase (GGT), Rao et al
(1986) fed 2 male F344 rats a diet of 0.1% WY-14,643 for 19 months or 3 F344 rats 0.025%
02/12/09 E-313
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Ciprofibrate for 15-19 months and reported "altered areas,"(AA) "neoplastic nodules" (NN), and
hepatocellular carcinomas (HCC). For WY-14,643 treatment 107 AA, 75 NN, and 5 HCC, and
for Ciprofibrate treatment 107 AA, 27 NN, and 16 HCC were identified. In the WY-14.643-
treated rats, HCC, and NN were both GGT and GST-71 negative (96-100%) with 87% of AA was
negative for both. In Ciprofibrate-treated rats NN and HCC were negative for both markers
(95%) but only 46% of AA were negative for both markers. Thus, a different pattern for tumor
phenotype was reported for WY-14,643 and another peroxisome proliferator, Ciprofibrate, in this
study as well.
In addition, GGT phenotype is reported to not be specific to weakly basophilic foci. GGT
staining was reported to be negative in eosinophilic tumors after initiation and promotion.
Kraupp-Grasl et al. (1990) note differences among PPARa agonists in their ability to promote
tumors and suggest they not necessarily be considered a uniform group. Caldwell and Keshava
(2006) suggest that the reports of a simple designation of "basophilic" is not enough to associate
a foci as caused by peroxisome proliferators [Bannasch, 1996; Grasl-Kraupp et al.,1993; Kraupp-
Grasl et al., 1990], Increased basophilia of tumors and increased numbers of carcinomas is
consistent with the progressive basophilia described by Bannasch (1996), as many adenomas
progress to carcinomas."
It should be noted that the amphophilic foci and tumors described by Bannasch et al. were
primarily studied in rats. Morimura et al. (2006) noted that WY-14,643 induced diffusely
basophilic tumors in mice and therefore identified the WY-14,643 tumors in a way consistent
with the descriptions of amphophilic tumors by Bannasch et al. The tumor induced by WY-
14,643 in their humanized mouse was reported to be similar to those arising spontaneously in the
mouse. However, the mouse response could differ from the rat, especially for PPARa agonists
other than WY-14,643.
H-ras activation and mutation studies have attempted to assign a pattern to peroxisome
proliferator-induced tumors as noted in Section 2.3.3.2., above. However also as noted in
Section 2.3.3.2, the genetic background of the mice used, the dose of carcinogen and the stage of
progression of "lesions" (i.e., foci vs. adenomas vs. carcinomas) may affect the number of
activated H-ras containing tumors which develop. Fox et al. (1990) note that tumors induced by
Ciprofibrate (0.0125% diet, 2 years) had a much lower frequency of H-ras gene activation than
those that arose spontaneously (2-year bioassays of control animals) or induced with the
"genotoxic" carcinogen benzidine-2 HC1 (120 ppm, drinking H20, 1 yr) and that the
Ciprofibrate-induced tumors were reported to be more eosinophilic as were the surrounding
normal hepatocytes than spontaneously occurring tumors. Anna et al. (1994) also stated that
mice treated with Ciprofibrate had a markedly lower frequency of tumors with activated H-ras
but that the spectrum of mutations in tumors was similar those in "spontaneous tumors." Hegi et
al. (1993) tested Ciprofibrate-induced tumors from Fox et al. (1990) in the NIH3T3
cotransfection-nude mouse tumorigenicity assay and concluded that ras protooncogene activation
were not frequent events in Ciprofibrate-induced tumors and that spontaneous tumors were not
promoted with it. Stanley et al. (1994) studied the effect of Methylclofenapate (MCP), a
02/12/09
E-314
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
peroxisome proliferator, in B6C3F1 (relatively sensitive) and C57BL/10J (relatively resistant)
mice for H-ras codon 61 point mutations in MCP-induced liver tumors (hepatocellular adenomas
and carcinomas). In the B6C3F1 mice, ~ 24% of MCP-induced tumors had codon 61 mutations
and for C57BL/10J mice ~ 13%. The findings of an increased frequency of H-ras mutation in
carcinomas compared to adenomas in both strains of mice is suggestive that these mutations were
related to stage of progression. Thus in mice, the phenotype of tumors did not appear to be
readily distinguishable from spontaneous tumors based on tincture for peroxisome proliferators
other than WY-14,643, but did have more of a signature in terms of H-ras mutation and
activation.
The expression of c-Jun has been used to discern TCE tumors from those of its
metabolites. However, as pointed out by Caldwell and Keshava (2006), although Bull et al.
(2004) have suggested that the negative expression of c-jun in TCA-induced tumors may be
consistent with a characteristic phenotype shown in general by peroxisome proliferators as a
class, there is no supporting evidence of this. While increased mitochondrial proliferation and
mitochondrial enzymes (thyromimetic rather than insulinomimetic) properties have been ascribed
to peroxisome proliferator-induced tumors, and the studies cited in Bull et al. (2004) have not
examined TCA-induced tumors for these properties.
3.4.1.6. Human relevance
In its framework for making conclusions about human relevance, the U.S. EPA Cancer
Guidelines (U.S. EPA 2005) asks that critical similarities and differences between test animals
and humans be identified. Humans possess PPARa at sufficient levels to mediate the human
hypolipidemic response to peroxisome-proliferating fibrate drugs. Fenofibrate and Ciprofibrate
induce treatment related increases in liver weight, hypertrophy, numbers of peroxisomes,
numbers of mitochondria, and smooth endoplasmic reticulum in cynomologous monkeys at 15
days of exposure (Hoivik et al., 2004). Given the species difference in the ability to respond to a
mitogenic stimulus such as partial hepatectomy (see Section 3.3.) lack of hepatocellular DNA
synthesis at this time point is not unexpected and as Rusyn et al. (2006) note examination at
differing time point may produce differing results. It is therefore generally acknowledged that "a
point in the rat and mouse key events cascade where the pathway is biologically precluded in
humans in principle cannot be identified."(Klaunig et al. 2003; NAS 2006). Thus, from a
qualitative standpoint, the effects described above are plausible in humans.
As for quantitative differences, there are two key issues. First, as stated in the Cancer
Guidelines, when considering human relevance, "Any information suggesting quantitative
differences between animals and humans is flagged for consideration in the dose-response
assessment." Therefore, while Klaunig et al. (2003) and NAS (2006) go on to suggest that
"this mode of action is not likely to occur in humans based on differences in several key steps
when taking into consideration kinetic and dynamic factors," under the Cancer Guidelines,
such "kinetic and dynamic factors" need to be made explicit in the dose-response assessment,
and should not be part of the qualitative characterization of hazard. Second, the discussion
02/12/09
E-315
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
above points to the lack of evidence supporting associations between the postulated events and
carcinogenic potency. Thus, because inter-species differences in carcinogenicity do not appear
to be associated with inter-species differences in postulated events, they do not provide reliable
metrics with which to make inferences about relative human sensitivity.
3.4.2. Other TCE Metabolite Effects That May Contribute to its Hepatocarcinogenicity
While the focus of most studies of TCA has been its effects on peroxisomal proliferation,
DCA has been investigated for a variety of effects that are also observed either in early stages of
oncogenesis (glycogen deposition) or conditions that predispose patients to liver cancer. Some
studies have examined microarray profiles in attempt to study the MOA or TCE (see section 3.2.2
for caveats regarding such approaches). Caldwell and Keshava have provided a review of these
studies which is provided below.
3.4.2.1. DCA effects and glycogen accumulation correlations with cancer
As noted previously, DCA administration has been reported to increase the observable
amount of glycogen in mouse liver via light microscopy and, although to not be primarily
responsible for DCA-induced liver mass increases, to be increase whole liver glycogen as much
by 50% (Kato-Weinstein et al., 2001). Given that TCE and DCA tumor phenotypes indicate a role
for DCA in TCE hepatocarcinogenicity (see Section 2.3.3.2., above), Caldwell and Keshava
(2006) described the correlations with effects induced by DCA that have been associated with
hepatocarcinogenicity.
A number of studies suggest DCA-induced liver cancer may be linked to its
effects on the cytosolic enzyme glutathione (GST)-S-transferase-zeta. GST-
zeta is also known as maleylacetoacetate isomerase and is part of the
tyrosine catabolism pathway whose disruption in type 1 hereditary
tyrosinemia has been linked to increased liver cancer risk in humans. GST-
zeta metabolizes maleylacetoacetate (MAA) to fumarylacetoacetate (FAA)
which displays apoptogenic, mutagenic, aneugenic, and mitogenic activities
(Bergeron et al., 2003; Jorquera and Tanguary, 2001; Kim et al 2000).
Increased cancer risk has been suggested to result from FAA and MAA
accumulation (Tanquary et al 1996). Cornett et al. (1999) reported DCA
exposure in rats increased accumulation of maleylacetone (a spontaneous
decarboxylation product of MAA), suggesting MAA accumulation.
Ammini et al. (2003) report depletion of the GST-zeta to be exclusively a
post-transcriptional event with genetic ablation of GST-zeta causing FAA
and MAA accumulation in mice. Schultz et al. (2002) report that
elimination of DCA is controlled by liver metabolism via GST-zeta in mice,
and that DCA also inhibits the enzyme (and thus its own elimination) with
young mice being the most sensitive to this inhibition. On the other hand,
older mice (60 weeks) had a decreased capacity to excrete and metabolize
02/12/09
E-316
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
DCA in comparison with younger ones. The authors suggest that
exogenous factors that deplete or reduce GST-zeta will decrease DCA
elimination and may increase its carcinogenic potency. They also suggest
that, due to suicide inactivation of GST-zeta, an assumption of linear
kinetics can lead to an underestimation of the internal dose of DCA at high
exposure rates. In humans, GST-zeta has been reported to be inhibited by
DCA and to be polymorphic (Tzeng et al 2000; Blackburn et al., 2001,
2000). Board et al. (2001) report one variant to have significantly higher
activity with DCA as a substrate than other GST zeta isoforms, which could
affect DCA susceptibility.
Individuals with glycogen storage disease or with poorly controlled diabetes
have excessive storage of glycogen in their livers (glycogenosis) and
increased risk of liver cancer (LaVecchia et., 1994; Adami et al., 1996;
Wideroff et al., 1997; Rake et al., 2002). In an animal model where
hepatocytes are exposed to a local hyperinsulinemia from transplanted islets
of Langerhans and the remaining tissue is hypoinsulinemic, insulin induces
alterations that resemble preneoplastic foci of altered hepatocytes (FAH)
and develop into hepatocellular tumors in later stages of carcinogenesis
(Evert et al., 2003). A number of studies have reported suppression of
apoptosis, decreases in insulin, and glycogenosis in mice liver by DCA at
levels that also induce liver tumors (Bull, 2004; Bull et al., 2004; Lingohr et
al., 2001). In isolated murine hepatocytes, Lingohr et al. (2002) reported
DCA-induced glycogenosis was dose related, occurred at very low doses
(10 |iM), occurred without the presence of insulin, was not affected by
insulin addition, was dependent on phosphatidylinositol 3-kinase (P13K)
activity, and was not a result of decreased glycogen breakdown. The
authors noted that PI3K is also known to regulate cell proliferation and
apoptosis in hepatocytes, and that understanding these mechanisms may be
important to understanding DCA-induced carcinogenesis. They also report
insulin receptor (IR) protein levels decreased to 30% of controls in mice
liver after up to 52 weeks of DCA treatment. Activation of the IR is also the
principal pathway by which insulin stimulates glycogen synthetase (the rate
limiting enzyme of glycogen biosynthesis). However, in DCA-induced
liver tumors IR protein was elevated as well as mitogen-activated protein
kinase (a downstream target protein of the IR) phosphorylation. DCA-
induced tumors were glycogen poor (Lingohr et al., 2001). The authors
suggest that normal hepatocytes down-regulate insulin-signaling proteins in
response to the accumulation of liver glycogen caused by DCA and that the
initiated cell population, which does not accumulate glycogen and is
promoted by DCA treatment, responds differently from normal hepatocytes
to the insulin-like effects of DCA.
02/12/09
E-317
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Gene expression studies of DCA show a number of genes identified with
cell growth, tissue remodeling, apoptosis, cancer progression, and
xenobiotic metabolism to be altered in mice liver at high doses (2 g/1 DCA)
in drinking water (Thai et al., 2001, 2003). After 4 weeks, RNA expression
was altered in 4 known genes (alpha-1 protease inhibitor, cytochrome B5,
stearoyl-CoA desaturase and caboxylesterase) in two mice (Thai et al.,
2001). Except for Co-A desaturase, a similar pattern of gene change was
reported in DCA-induced tumors (10 tumors from 10 different mice) after
93 weeks. Using cDNA microarray in the same mice, Thai et al. (2003)
identified 24 genes with altered expression, of which 15 were confirmed by
Northern blot analysis after 4 weeks of exposure. Of the 15 genes, 14
revealed expression suppressed two- to fivefold and included: MHR 23 A,
cytochrome P450 (CYP), 2C29, CYP 3A11, serum
paraoxonase/arylesterase 1, liver carboxylesterase, alpha-1 antitrypsin, ER
p72, GST-pi 1, angiogenin, vitronectin precursor, cathepsin D, plasminogen
precursor (contains angiostatin), prothrombin precursor and integrin alpha 3
precursor. An additional gene, CYP 2A4/5, had a twofold elevation in
expression. After 93 weeks of treatment with 3.5 g/1 DCA, Northern blot
analyses of total RNA isolated from DCA-induced hepatocellular
carcinomas showed similar alteration of expression (11 of 15). It was noted
that peroxisome proliferator-activated receptor (PPAR)a and IR gene
expression were not changed by DCA treatment. Genes involved in
glycogen or lipid metabolism were not tested.
Although it has not been possible to directly determine whether DCA is
produced from TCE at carcinogenic levels, there is indirect evidence that
DCA is formed from TCE in vivo and contributes to liver tumor
development. Pretreatment with either DCA or TCE inhibits GST-zeta
while TCA pretreatment does not (Schultz et al., 2002; Bull et al., 2004).
TCE treatment decreased Vmax for DCA metabolism to 49% of control
levels with a 1 g/kg TCE dose resembling effects those of 0.05 g/1 DCA
(Schultz et al., 2002).
3.4.2.2. Genetic profiling data for TCE: gene expression and methylation status studies
Caldwell and Keshava (2006) and Keshava and Caldwell (2006) report on both genetic
expression studies and studies of changes in methylation status induced by TCE and its
metabolites (See sections 2.3.2. and 2.3.3, above) as well as differences and difficulties in the
patterns of gene expression between differing PPARa agonists. In Section 4.2.2. (below), the
effects of co-exposures of DCA, TCA and Chloroform on methylation status are discussed. In
particular are concerns for the interpretation of studies which employ pooling of data as well as
interpretation of "snapshots in time of multiple gene changes." For the Laughter et al. (2004)
study in particular, it is not clear whether transcription arrays were performed on pooled data (no
02/12/09
E-318
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
data on variability between individual animals was provided and the methodology section of the
report is not transparently written in this regard). The issue of phenotypic anchoring also arises
as data on % liver/body weight indicates significant variability within TCE treatment groups,
especially in PPARa-null mice. For studies of gene expression using microarrays Bartosiewicz
et al (2001) used a screening analysis of 148 genes for xenobiotic-metabolizing enzymes, DNA
repair enzymes, heat shock proteins, cytokines, and housekeeping gene expression patterns in the
liver in response TCE. The TCE-induced gene induction was reported to be highly selective;
only Hsp 25 and 86 and Cyp2a were up-regulated at the highest dose tested. Collier et al. (2003)
reported differentially expressed mRNA transcripts in embryonic hearts from Sprague-Dawley
rats exposed to TCE with sequences down-regulated with TCE exposure appearing to be those
associated with cellular housekeeping, cell adhesion, and developmental processes. TCE was
reported to induce up-regulated expression of numerous stress-response and homeostatic genes.
For the Laughter et al. (2004) study, transcription profiles using macroarrays containing
approximately 1,200 genes were reported in response to TCE exposure. Forty three genes were
reported to be significantly altered in the TCE-treated wild-type mice and 67 genes significantly
altered in the TCE-treated PPARa knockout mice. Out of the 43 genes expressed in wild-type
mice upon TCE exposure, 40 genes were reported by the authors to be dependent on PPARa and
included genes for CYP4al2, epidermal growth factor receptor, and additional genes involved in
cell growth. However, the interpretation of this information is difficult because in general,
PPARa knockout mice have been reported to be more sensitive to a number of hepatotoxins
partly because of defects in the ability to effectively repair tissue damage in the liver (Shankar et
al., 2003; Mehendale, 2000) and because a comparison of gene expression profiles between
controls (wild-type and PPARa knockout) were not reported.
As stated previously, knockout mice in this study also responded to TCE exposure with
increased liver weight, had increased background liver weights, also had higher baseline levels of
hepatocyte proliferation than wild-type mice. Nakajima et al. (2000) reported that the number of
peroxisomes in hepatocytes increased by 2 fold in wild-type mice but not in PPARa knockout
mice. However, TCE induced increased liver weight in both male and female wild type and
knockout mice, suggesting hepatic effects independent of PPARa activation. In regards to
toxicity, after three weeks of TCE treatment (0 tol,500 mg/kg via gavage), Laughter et al. (2004)
reported toxicity at thel500 mg/kg level in the knockout mice that was not observed in the wild-
type mice — all knockout mice were moribund and had to be removed from the study.
Differences in experimental protocol made comparisons between TCE effects and those of its
metabolites difficult in this study (see Section 2.1.13 above).
As reported by Voss et al. (2006), dose-, time course-, species-, and strain-related
differences should be considered in interpreting gene array data. The comparison of differing
PPARa agonists presented in Keshava and Caldwell (2006) illustrate the pleiotropic and varying
liver responses of the PPARa receptor to various agonists, but did imply that these responses were
responsible for carcinogenesis.
02/12/09
E-319
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
As discussed above in Section 3.3.5 and in Caldwell and Keshava (2006), "Aberrant DNA
methylation has emerged in recent years as a common hallmark of all types of cancers, with
hypermethylation of the promoter region of specific tumor suppressor genes and DNA repair
genes leading to their silencing (an effect similar to their mutation) and genomic hypomethylation
(Ballestar and Esteller 2002; Berger and Daxenbichler 2002; Herman et al. 1998; Pereira et al.
2004; Rhee et al. 2002). Whether DNA methylation is a consequence or cause of cancer is a long-
standing issue (Ballestar and Esteller 2002). Fraga et al. (2004, 2005) reported global loss of
monoacetylation and trimethylation of hi stone H4 as a common hallmark of human tumor cells;
they suggested, however, that genomewide loss of 5-methylcytosine (associated with the
acquisition of a transformed phenotype) exists not as a static predefined value throughout the
process of carcinogenesis but rather as a dynamic parameter (i.e., decreases are seen early and
become more marked in later stages) "
Although little is known about how it occurs, a hypothesis has also been proposed that that
the toxicity of TCE and its metabolites may arise from its effects on DNA methylation status. In
regard to methylation studies, many are co-exposure studies as they have been conducted in
initiated animals, and as stated above, some are very limited in regard to the reporting and conduct
of the study. Caldwell and Keshava (2006) review the body of work regarding TCE, DCA and
TCA. Methionine status has been noted to affect the emergence of liver tumors. As noted by
Counts et al. (1996) a choline/methionine deficient diet for 12 months did not increase liver tumor
formation in C3H/HeN mice but is tumorigenic to B6C3F1 mice. Tao et al (2000) and Pereira et
al (2004) have studied the effects of excess methionine in the diet to see if it has the opposite
effects as a deficiency (i.e. and reduction in a carcinogenic response rather than enhancement). As
noted above for Tao et al. (2000), the administration of excess methionine in the diet is not
without effect. The data of Tao et al. (2000) suggest that % liver/body weight ratios are affected
by short-term methionine exposure (300 mg/kg) in female B6C3F1 mice. Pereira et al (2004)
reported that very high level of methionine supplementation to an AIN-760A diet, affected the
number of foci and adenomas after 44 weeks of co-exposure to 3.2.g/1 DCA. While the highest
concentration of methionine (8.0 g/kg) was reported to decrease both the number of DCA-induce
foci and adenomas, the lower level of methionine co-exposure (4.0 g./kg) increased the incidence
of foci. Co-exposure of methionine (4.0 or 8.0 g/kg) with 3.2 g/1 DCA was reported to decrease
by ~ 25% DCA-induced glycogen accumulation, increase mortality, but not to have much of an
effect on peroxisome enzyme activity (which was not elevated by more than 33% over control for
DCA exposure alone). Methionine treatment alone at the 8 g/kg level was reported to increase
liver weight, decrease lauroyl-CoA activity and to increase DNA methylation. The authors
suggested that their data indicate that methionine treatment slowed the progression of foci to
tumors. Given that increasing hypomethylation is associated with tumor progression, decreased
hypomethylation from large doses of methionine are consistent with a slowing of progression.
Whether, these results would be similar for lower concentrations of DCA and lower
concentrations of methionine that were administered to mice for longer durations of exposure,
cannot be ascertained from this data. It is possible that in a longer-term study, the number of
tumors would be similar. Whether, methionine treatment co-exposure had an effect on the
phenotype of foci and tumors was not presented by the authors in this study. Such data would
02/12/09
E-320
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
have been valuable to discern if methionine co-exposure at the 4.0 mg/kg level that resulted in an
increase in DCA-induce foci, resulted in foci of a differing phenotype or a more heterogeneous
composition than DCA treatment alone. Finally, a decrease in tumor progression by methionine
supplementation is not shown to be a specific event for the MOA for DCA-induced liver
carcinogenicity.
Tao et al. (2000) reported that 7 days of gavage dosing of TCE (1000 mg/kg in corn oil),
TCA (500 mg/kg, neutralized aqueous solution), and DCA (500 mg/kg, neutralized aqueous
solution) in 8-week old female B6C3F1 mice resulted in not only increased liver weight but also
increased hypomethylation of the promoter regions of c-Jun and c-Myc genes in whole liver DNA
(data shown for 1-2 mice per treatment). Treatment with methionine was reported to abrogate this
response only at a 300 mg/kg i.p. dose with 0-100 mg/kg doses of methionine having no effect.
Ge et al. (2001b) reported DCA- and TCA-induced DNA hypomethylation and cell proliferation
in the liver of female mice at 500 mg/kg and decreased methylation of the c-myc promoter region
in liver, kidney and urinary bladder. However, increased cell proliferation preceded
hypomethylation. Ge et al (2002) also reported hypomethylation of the c-myc gene in the liver
after exposure to the peroxisome proliferators 2,4-dichlorophenoxyacetic acid (2,4-D)(l,680
ppm), dibutyl phthalate (DBP) (20,000 ppm), Gemfibrozil (8,000 ppm), and Wy-14,643 (50-500
ppm, with no effect at 5 or 10 ppm) after six days in the diet. Caldwell and Keshava (2006)
concluded that hypomethylation did not appear to be a chemical-specific effect at these
concentrations. As noted above in section 3.3.5, chemical exposure to a number of differing
carcinogens have been reported to lead to progressive loss of DNA methylation..
Caldwell and Keshava (2006) also note similar changes in methylation after initiation and
treatment with DCA or TCA. "After initiation by N-methyl-N-nitrosourea (25 mg/kg) and
exposure to 20 mmL/L DCA or TCA (46 weeks), Tao et al. (2004) report similar hypomethylation
of total mouse liver DNA by DCA and TCA with tumor DNA showing greater hypomethylation.
A similar effect was noted for region-2 (DMR-2) of the insulin-like growth factor-II (IGF-II)
gene. The authors suggest that hypomethylation of total liver DNA and the IGF-II gene found in
non-tumorous liver tissue would appear to be the result of a more prolonged activity and not cell
proliferation, while hypomethylation of tumors could be an intrinsic property of the tumors. Over
expression of IGF-II gene in liver tumors and preneoplastic foci has been shown in both animal
models of hepatocarcinogenesis and humans, and may enhance tumor growth, acting via the over-
expressed IGF-I receptor (Scharf et al., 2001; Werner and Le Roith, 2000). IGF-I is the major
mediator of the effects of the growth hormone; it thus has a strong influence on cell proliferation
and differentiation and is a potent inhibitor of apoptosis (Furstenberger et al., 2002). Normally,
expression of IGF-II in liver is greater during the fetal period than the adult, but is over-expressed
in human hepatocarcinomas due to activation of fetal promoters (Scharf et al., 2001) and loss of
imprinting (Khandawala et al., 2000). Takeda et al. (1996) report IGF-II expression in the liver is
monoallelic (maternally imprinted) in the fetal period is relaxed during the postnatal period,
(resulting in biallelic expression), and is imbalanced in human hepatocarcinomas (leading to
restoration of monoallelic IG-II expression)."
02/12/09
E-321
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
However, Bull (2004) and Bull et al. (2004) have recently suggested that hypomethylation
and peroxisome proliferation occur at higher exposure levels than those that induce liver tumors
for TCE and its metabolites. They report that a direct comparison in the no-effect level or low-
effect level for induction of liver tumors in the mouse and several other endpoints shows that, for
TCA, liver tumors occur at lower concentrations than peroxisome proliferation in vivo but that
PPARa activation occurs at a lower dose than either tumor formation or peroxisome proliferation.
A similar comparison for DCA shows that liver tumor formation occurs at a much lower exposure
level than peroxisome proliferation, PPARa activation, or hypomethylation. In addition, they
report that these chemicals are effective as carcinogens at doses that do not produce cytotoxicity.
3.4.2.3. Oxidative Stress
Several studies have attempted to study the possible effects of "oxidative stress" and DNA
damage resulting from TCE exposures. The effects of induction of metabolism by TCE, as well
as through co-exposure to ethanol, have been hypothesized to in itself increase levels of "oxidative
stress" as a common effect for both exposures (see Section 4.2.4. below). Oxidative stress has
been hypothesized to be the MOA for peroxisome proliferators as well, but has been found to
neither be correlated with cell proliferation nor carcinogenic potency of peroxisome proliferators
(see Section 3.4.1.1). As a MOA, it is not defined or specific as the term "oxidative stress" is
implicated as part of the pathophysiologic events in a multitude of disease processes and is part of
the normal physiologic function of the cell and cell signaling.
In regard to measures of oxidative stress, Rusyn et al. (2006) noted that although an
overwhelming number of studies draw a conclusion between chemical exposure, DNA damage,
and cancer based on detection of 8-hydroxy-2' deoxyguanosine (8-OHdG), a highly mutagenic
lesion, in DNA isolated from organs of in vivo treated animals, a concern exists as to whether
increases in 8-OHdG represent damage to genomic DNA, a confounding contamination with
mitochondrial DNA, or an experimental artifact. As described in Section 2.2.8., the study by
Channel et al (1998) demonstrated that corn oil as vehicle had significant effects on measures of
"oxidative stress" such as thiorbarbiturate acid-reactive substances (TBARS). Also as noted
previously (Sections 2.1.1. and 2.2.11.), studies of TCE which employ the i.p. route of
administration can be affected by inflammatory reactions resulting from that routes of
administration and subsequent toxicity that can involve oxygen radical formation from
inflammatory cells.
The issues with interpretation of the Channel et al (1998) study of TCE administered via
corn oil gavage to mice have already been discussed in Section 2.1.7 above. The TBARS results
indicated suppression of TBARS with increasing time of exposure to corn oil alone with data
presented in such a way for 8-OHdG and total free radical changes that the pattern of corn oil
administration was obscured. It was not apparent from that study that TCE exposure induced
oxidative damage in the liver.
02/12/09
E-322
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Toraason et al. (1999) measured 8-OHdG and a "free radical-catalized isomer of
arachidonic acid and marker of oxidative damage to cell membranes, 8-Epi-prostaglandin F2a
(8epiPGF)", excretion in the urine and TBARS (as an assessment of malondialdehyde and marker
of lipid peroxidation) in the liver and kidney of male Fischer rats (150-200 g) exposed to single 0,
100, 500, or 1000 mg/kg TCE i.p. injections in Alkamuls vehicle (n= 6/group). Two sequential
urine samples were collected 12 hours after injection and animals were sacrificed at 24 hours with
DNA collected from liver tissues and TBARS measured in liver homogenates. The mean body
weights of the rats were reported to vary by 13% but the liver weights varied by 44% after the
single treatments of TCE (i.e., in contrast to the large volume of the literature that reports TCE-
induced increases in liver weight, the 500 and 1000 mg/kg exposed rats were reported to have
reduced liver weight by 44% in comparison to the control values). Using this paradigm, 500
mg/kg TCE was reported to induce stage II anesthesia and a 1000 mg/kg TCE to induce level III
or IV (absence of reflex response) anesthesia and burgundy colored urine with 2/6 rats at 24 hours
comatose and hypothermic. The animals were sacrificed before they could die and the authors
suggested that they would not have survived another 24 hours. Thus, using this paradigm there
was significant toxicity and additional issues related to route of exposure. Urine volume declined
significantly during the first 12 hours of treatment and while water consumption was not
measured, it was suggested by the authors to be decreased due to the moribundity of the rats.
Given that this study examined urinary markers of "oxidative stress" the effects on urine volume
and water consumption, as well as the profound toxicity induced by this exposure paradigm, limit
the interpretation of the study. The authors noted that because both using volume and creatinine
excretion were affected by experimental treatment, urinary excretion of 8-OHdG changed
significantly based on the mode of data expression. Excretion of 8epiPGF was reported to be no
different from controls 12-24 hours and decreased 24 hours after TCE exposure at the two highest
levels. Excretion of 8-OHdG was reported to not be affected by any exposure level of TCE and if
expressed on the basis of 24- hours decreased. TBARS concentration per gram of liver was
reported to be increased at the 500 and 1000 mg/kg TCE exposure levels (~ 2-3) fold. The effects
of decreased liver size in the treated animals for this measure in comparison to control animals,
was not discussed by the authors. For 8-OHdG measures in the liver and lymphocytes, the
authors reported that "cost prohibited analysis of all of the tissues samples" so that a subset of
animals was examined exhibiting the highest TBARS levels. The number of animals used for this
determination was not given nor the data except for 500 mg/kg TCE exposure level. TCE was
reported to increase 8-OHdG/dG in liver DNA relative to controls to about the same extent in
lymphocytes from blood and liver (~ 2-fold) with the results for liver reported to be significant.
The issues of bias in selection of the data for this analysis, as well as the issues already stated for
this paradigm limit interpretation of this data while the authors suggest that evidence of oxidative
damage was equivocal.
DCA and TCA have also been investigated using similar measures. Larson and Bull
(1992) exposed male B6C3F1 mice [26 ± 3 g (SD)]to a single dose of 0, 100, 300, 1000, or 2000
mg/kg/day TCA or 0, 100, 300, or 1000 mg/kg/day DCA in distilled water by oral gavage (n= 4).
Fischer 344 rats (237 ± 4 g) received a single oral dose of 0, 100 or 1000 mg/kg DCA or TCA (n=
4 or 5) TBARS was measured from liver homogenates and assumed to be malondialdehyde. The
02/12/09
E-323
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
authors stated that a preliminary experiment had shown that maximal TBARS was increased 6
hours after a dose of DCA and 9 hours after a dose of TCA in mice (data shown) and that by 24
hours TBARS concentrations had declined to control values (data not shown). However, time-
course information in rats was not presented and the same times used for both species, (i.e., 6- and
9- hours time periods after administration of DCA and TCA) for examination of TBARS activity.
A dose of 100 mg/kg DCA (rats or mice) or TCA (mice) did not elevate TBARS concentrations
over that of control liver with this concentration of TCA not examined in rats. For TCA, there
was a slight dose-related increase in TBARS over control values starting at 300 mg/kg in mice
(i.e., 1.68-fold, 2.02- -fold, and 2.70-fold of control for 300, 1000, and 2000 mg/kg TCA). For
DCA there were similar increases over control for both the 300 and 1000 mg/kg dose levels in
mice (i.e., 3.22-fold and 3.45-fold of control, respectively). For rats the 1000 mg/kg and 2000
mg/kg levels of TCA were reported to show a statistically significant increase in TBARS over
control (i.e., 1.67-fold and 2.50-fold, respectively) with the 300 and 1000 mg/kg level of DCA
showing similar increases but with only the 300 mg/kg-induced change statistically significant
different than control values (i.e., 3.0-fold and 2.0-fold of control, respectively). Of note, is the
report that the induction of TBARS in mice is transient and has subsided within 24 hours of a
single dose of DCA or TCA, that the response in mice appeared to be slightly greater with DCA
than TCA at similar doses, and that for DCA, there was similar TBARS induction between rats
and mice at similar dose levels.
A study by Austin et al. (1996) appears to a follow-up publication of the preliminary
experiment cited in Larson and Bull (1992). Male B6C3F1 mice (8 weeks old) were treated with
single doses of DCA or TCA in buffered solution (300 mg/kg) with liver examined for 8-OHdG.
The authors stated that in order to conserve animals, controls were not employed at each time
point. For DCA the time course of 8-OHdG was studied at 0, 4, 6, and 8 hours after
administration and for TCA at 0, 6, 8, and 10 hours after of a 300 mg/kg dose (n= 6). There was a
statistically significant increase over controls in 8-OHdG for the 4- and 6-hour time points for
DCA (~ 1.4-fold and 1.5-fold of control, respectively) but not at 8 hours in mice. For TCA, there
was a statistically significant increase in 8-OHdG at 8 and 10 hours for TCA (~ 1.4 and 1.3-fold of
control, respectively).
The results for PCO and liver weight for Parrish et al. (1996) are discussed in Section
2.3.2.2 above for male B6C3F1 mice exposed to TCA or DCA (0, 0.01, 0.5 and 2.0 g/1) for 3 or 10
weeks (n= 6). The study focused on an examination of the relationship with measures of
peroxisome proliferation and oxidative stress. The dose-related increase in PCO activity at 21
days (~ 1.5-fold, 2.2-fold, and ~ 4.1-fold of control, for 0.1, 0.5, and 2.g/1 TCA) was reported to
not be increased similarly for DCA. Only the 2.0 g/1 dose of DCA was reported to induce a
statistically significant increase at 21-days of exposure of PCO activity over control (~ 1.8-fold of
control). After 71 days of treatment, TCA induced dose-related increases in PCO activities that
were ~ twice the magnitude as that reported at 21 days (i.e. ~ 9-fold greater at 2.0 g/1 level).
Treatments with DCA at the 0.1 and 0.5 g/1 exposure levels produced statistically significant
increase in PCO activity of - 1.5-fold and 2.5-fold of control, respectively. The administration of
02/12/09
E-324
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
1.25 g/1 clofibric acid in drinking water, used as a positive control, gave ~ 6-7-fold of control PCO
activity at 21 and 71 days exposure.
Parrish et al (1996) reported that laurate hydroxylase activity was reported to be elevated
significantly only by TCA at 21 days and to approximately the same extent (~ 1.4 to 1.6-fold of
control) increased at all doses tested. At 71 days both the 0.5 and 2.0 g/1 TCA exposures induced
a statistically significant increase in laurate hydroxylase activity (i.e., 1.6-fold and 2.5-fold of
control, respectively) with no change reported after DCA exposure. The actual data rather than %
of control values were reported for laurate hydroxylase activity with the control values varying
1.7-fold between 21 and 71 days experiments. Levels of 8-OHdG in isolated liver nuclei were
reported to not be altered from 0.1, 0.5, or 2.0 g/1 TCA or DCA after 21 days of exposure and this
negative result was reported to remain even when treatments were extended to 71 days of
treatment. The authors noted that the level of 8-OHdG increased in control mice with age (i.e., ~
2 fold increase between 71-day and 21-day control mice). Clofibric acid was also reported not to
induce a statistically significant increase of 8-OHdG at 21 days, but to produce an increase (~ 1.4-
fold of control) at 71 days. Thus, the increases in PCO activity noted for DCA and TCA were not
associated with 8-OHdG levels (which were unchanged) and also not with changes laurate
hydrolase activity observed after either DCA or TCA exposure. Of note is the variability in both
baseline levels of PCO and laurate hydrolase activity. Also of note, is that the authors report
taking steps to minimize artifactual responses for their 8-OHdG determinations. The authors
concluded that their data does not support an increase in steady state oxidative damage to be
associated with TCA initiation of cancer and that extension of treatment to time periods sufficient
to insure peroxisome proliferation failed to elevate 8-OHdG in hepatic DNA. The increased 8-
OHdG at 10 weeks after Clofibrate administration but lack of 8-OHdG elevation at similar levels
of PCO induction by were also noted by the authors to suggest that peroxisome proliferative
properties of TCA were not linked to oxidative stress or carcinogenic response.
As noted above for the study of Leakey et al (2003a) (Section 2.3.4.), hepatic
malondialdehyde concentration in ad libitum fed and dietary controlled mice did not change with
chloral hydrate (CH) exposure at 15 months but the dietary controlled groups were all ~ half that
of the ad libitum-fed mice. Thus, while overall increased tumors observed in the ad libitum diet
correlated with increased malondialdehyde concentration, there was no association between CH
dose and malondialdehyde induction for either diet.
4. Effects of co-exposures on MOA - internal and external
exposures to mixtures including alcohol
Caldwell et al. (2008b) recently published a review of the issues and studies involved
with the effects of co-exposures to TCE metabolites that could be considered internal (i.e., an
internal co-exposure for the liver) and co-exposures to metabolites and other commonly
occurring chemicals that are present in the environment. As they stated:
02/12/09
E-325
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Human exposure to a pollutant rarely occurs in isolation. EPA's Cumulative Exposure
project and subsequent National Air Toxics Assessment have demonstrated that
environmental exposure to a number of pollutants, classified as potential human
carcinogens, is widespread [U.S. EPA, 2006;Woodruff et al., 1998], Interactions between
carcinogens in chemical mixtures found in the environment have been a concern for
several decades. Furthermore, how these interactions affect the mode of action (MOA)
by which these chemicals operate and how such effects may modulate carcinogenic risk is
of concern as well. Thus, an understanding of the MOA(s) of a pollutant can help
elucidate its potential carcinogenic risk to humans, and can also help identify susceptible
subpopulations through their intrinsic factors (e.g., age, gender, and genetic
polymorphisms of key metabolic and clearance pathways) and extrinsic factors (e.g. co-
exposures to environmental contaminants, ethanol consumption, and pharmaceutical use).
Trichloroethylene (TCE) can be a useful example for detailing the difficulties and
opportunities for investigating such issues because, for TCE, there is both internal
exposure to a "chemical mixture" of multiple carcinogenic metabolites [Chiu et al.,
2006a, b] and co-exposures from environmental contamination of TCE metabolites, and
from pollutants that share common metabolites, metabolic pathways, MO As, and targets
of toxicity with TCE.
Typically, ground water or contaminated waste sites can have a large number of
pollutants that vary in regard to information available to support the characterization of
their potential hazard, and that have differing MO As and targets. For example,
Veeramachaneni et al. (2001) reported reproductive effects in male rabbits, resulting from
exposure to drinking water containing concentrations of chemicals typical of ground
water near hazardous waste sites. The drinking water exposure mixture contained
arsenic, chromium, lead, benzene, chloroform, phenol, and TCE. Even at 45 weeks after
the last exposure, mating desire/ability, sperm quality, and Leydig cell function were
subnormal. However, while the exposure levels are relevant to human environmental
exposures, design of this study precludes a conclusion as to which individual toxicant, or
combination of the seven toxicants, caused the effects. Thus, this study exemplifies he
problems associated with studying a multi-mixture milieu. Studies of the interactions of
TCE metabolites or common co-exposures that report the interactions of 2 or 3 chemicals
at one time are easier to interpret.
Since EPA published its 2001 draft assessment, several approaches have been reported
that include examination of tumor phenotype, gene expression, and development of
physiologically-based pharmacokinetic (PBPK) models to assess possible effects of co-
exposure. They attempt to predict whether such co-exposures would produce additivity
of response or if co-exposure would change the nature of responses induced by TCE or its
metabolites. In addition, new studies on co-exposure to DBA may help identify a co-
exposure of concern. These studies may give potential insights into possible MO As and
modulators of TCE toxicity. More recent information on the toxicity of individual
metabolites of TCE [Caldwell and Keshava, 2006] may be helpful in trying to identify
02/12/09 E-326
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
which are responsible for TCE toxicity, but may also identify the effects of environmental
co-exposures.
Recently, EPA sought advice from the National Academy of Sciences (NAS) [Chiu et al.,
2006a] with the NAS charge questions including the following. (1) What TCE
metabolites, or combinations of metabolites, may be plausibly involved in the toxicity of
TCE? (2) What chemical co-exposures may plausibly modulate TCE toxicity? (3) What
can be concluded about the potential for common drinking water contaminants such as
other solvents and/or haloacetates to modulate TCE toxicity? (4) What can be concluded
about the potential for ethanol consumption to modulate TCE toxicity? Thus, the
understanding of the effects of co-exposure, in the context of MO A, is an important
element in understanding the risk of a potential human carcinogen.
U.S. EPA's draft TCE risk assessment [U.S. EPA, 2001] identified several factors
involving co-exposure to TCE metabolites, environmental contaminants, and ethanol that
could lead to differential sensitivity to TCE toxicity. Research needs identified there, as
well as in previous reviews [Bull, 2000; Pastino et al., 2000], included further elucidation
of the interaction of TCA and DCA in TCE-induced liver tumors and a better
understanding of the functional relationships among risk factors. The complexity of
TCE's potential interactions with chemical co-exposures from either common
environmental co-contaminants or common behaviors such as alcohol consumption
mirrors the complexity of the metabolism and the actions of TCE metabolites. Thus, TCE
presents a good case study for further exploration of the effects of co-exposure on MOA.
The following Sections first reiterates the findings of Bull et al (2002) in regard to simple
co-exposures of DCA and TCA which can be experienced as an internal co-exposure after TCE
exposure. A number of studies have examined the effects of TCE or its metabolites after
previous exposure to presumably genotoxic carcinogen to not only determine the effect of the co-
exposure on liver carcinogenicity but also to use such paradigms to distinguish between the
effects of TCA and DCA. Finally, not only is TCE a common co-exposure with its own
metabolites, but is also a common co-exposure with other solvents, and the brominated analogues
of TCA and DCA. The available literature is examined for potential similarities in target and
effects that may cause additional concern. The effects of ethanol on TCE toxicity is examined as
well as the potential pharmacokinetic modulation of risk using recently published reports of
physiologically based pharmacokinetic (PBPK) models that may be useful in predicting co-
exposure effects.
4.1. Internal Co-exposures to TCE Metabolites: Modulation of Toxicity and
Implications for TCE MOA
Exposure to TCE will produce oxidative metabolites in the liver as an internal co-
exposure. As stated above, the phenotypic analysis of TCE-induced tumors have similarities to
combinations of DCA and TCA and in some reports to more closely resemble DCA-induce
02/12/09
E-327
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
tumors in the mouse. Results from Bull et al. (2002) are presented in Section 2.2.22. for the
treatment of mice to differing concentrations of DC A and TCA in combination and the
resemblance of tumor phenotype to that of TCE. In regard to cancer dose-response, the most
consistent treatment-related increase in response occurred with combinations of exposure to DCA
and TCA that appeared to increase lesion multiplicity when compared to effects from individual
chemicals separately. Bull et al. (2002) presented results for "selected" lesions examined for
pathology characterization that suggest co-exposure of 0.5 g/1 DCA with either 0.5 g/1 or 2 g/1
TCA had a greater than additive effect on the total number of hyperplastic nodules. In addition
co-exposure to 0.1 g/1 DCA and 2 g/1 TCA was reported to have a greater than additive effect on
the total number of adenomas, but not carcinomas, induced. The random selection of lesions for
the determination of potential treatment-related effects on incidence and multiplicity, rather than
characterization of all lesions, increases the uncertainty in this finding.
4.2. Initiation Studies as Co-exposures
There is a body of literature that has focused on the effects of TCE and its metabolites
after rats or mice have been exposed to "mutagenic" agents to "initiate" hepatocarcinogenesis.
Given that most of these "initiating agents" have many effects that are not only mutagenic but
also on epigenetic, that the dose and exposure paradigm modify these effects, that "initiators"
they can increased tumor responses alone, and the tumors that arise from these protocols are
reflective of simultaneous actions of both "initiator" and "promoter," paradigms that first expose
rats or mice to a "mutagen" and then to other carcinogenic agents can be described as a co-
exposure protocols. As stated previously, DEN and N-nitrosomorpholine have been reported to
increase differing populations of mature hepatocytes with DEN not only being a mutagen but
also to induce concurrent hepatocyte regeneration at a high dose. Thus, the effects of the TCE or
its metabolites are hard to discern from the effects of the "initiating" agent in terms of MOA. As
demonstrated in the studies of Pereira et al. below, the gender also determines the nature of the
tumor response using these protocols. In addition, when the endpoint for examination is tumor
phenotype the consequences of tumor progression are hard to discern from the MOA of the
agents when differing concentrations, different durations of exposure, lesions counted as
"tumors" included different stages of tumor progression (foci to carcinoma), and highly variable
and low numbers of animals are examined. However, differences in phenotype of tumors
resulting from such co-exposures, like the co-exposure studies cited above for co-exposures of
TCE metabolites, can help discern that exposure to TCE metabolites results in differing actions
as demonstrated by differing effects in the presence of co-carcinogens. As stated above, Kraupp-
Grasl et al. (1990) note differences among PPARa agonists in their ability to promote tumors
suggest they should not necessarily be considered a uniform group.
4.2.1. Herren-Freund et al. 1987
The results of TCE exposure alone were reported previously (2.2.17) for this study. This
study's focus was on the effect of TCE, TCA, DCA and Phenobarbital on hepatocarcinogenicity
in male B6C3F1 mice after "initiation" at 15 days with 2.5 or 10 j_ig/g body weight of
02/12/09
E-328
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
ethylnitrosourea (ENU) and then subsequent exposure to TCE and other chemicals in drinking
water begging at 4 weeks of age (an age when the liver is already undergoing rapid growth).
DCA and TCA were given in buffered solutions and sodium chloride given in the water of
control animals. The experiment was reported to be terminated at 61 weeks because the "mice
started to exhibit evidence of tumors." Concentrations of TCE were 0, 3 and 40 mg/1, of DCA
and TCA 0, 2 and 5 g/1, and of Phenobarbital 0 and 500 mg/1. The number of animals examined
in each group ranged from 16 to 32. ENU alone in this paradigm was reported to induce
statistically significant increases in adenomas and hepatocellular carcinomas (39% incidence of
adenomas and 39% incidence of carcinomas vs. 9% and 0% for controls) at the 10 jag/g dose
(n=23), but not at 2.5 j_ig/g dose (n= 22). The effects of high doses of DCA and TCA alone have
already been discussed for other studies, as well as the lack of statistical power using a paradigm
with so few and variable numbers of animals, the limitations of an abbreviated duration of
exposure which does not allow for full expression of a carcinogenic response, and problems of
volatilization of TCE in drinking water. DCA and TCA treatments at these levels (5 g/1) were
reported to increase adenomas and carcinomas irrespective of ENU pretreatment and to ~ the
same extent with and without ENU. TCE at the highest dose was reported to increase the number
of animals with adenomas (37% vs. 9% in control) and carcinomas (37% vs. 0% in controls) but
only the # of adenomas/animal was statistically significant as the number of animals examined
was only 19 in the TCE group. Phenobarbital was reported to have no effect on ENU tumor
induction using this paradigm.
4.2.2. Parnell et al. 1986
This study used a rat liver foci bioassay (y-glutamyltranspeptidase, i.e., GGT) for hepatic
foci after at 3 and 6 month using protocols that included partial hepatectomy, DEN (10 mg/kg) or
TCA (1500 ppm in drinking water) treatment, and then promotion with 5000 ppm TCA (i.e., 5
g/1) for 10, 20, or 30 days and phenobarbital (500 ppm) in male Sprague-Dawley rats (5-6 weeks
old at partial hepatectomy). The number of animals per group ranged from 4-6. PCO activities
were given for various protocols involving partial hepatectomy, DEN, TCA and Phenobarbital
treatments but there was no controls values given that did not have a least one of these
treatments. Overall it appeared there was a slight decrease of PCO activity in rats treated with
partial hepatectomy/DEN/Phenobarbital treatments and a slight increase over other treatments for
rats treated with partial hepatectomy/DEN/ 5000 ppm TCA or just TCA from 2 weeks to 6
months of sampling. In regard to GGT-positive foci, the partial hepatectomy/DEN/Phenobarbital
group (n= 6) was reported to have more positive foci at 3 or 6 months than rats "initiated" with
TCA and PB after partial hepatectomy or partial hepatectomy/Phenobarbital treatment alone
(2.05 foci/cm2 vs. ~ .05 - 0.10 foci/cm2 for all other groups). The number of GGT positive foci
in rats without any treatment were not studied or presented by the authors. For "promotion"
protocols the number of GGT positive foci induced by the partial
hepatectomy/DEN/Phenobarbital protocol at 3 and 6 months, appeared to be reduced when
Phenobarbital exposure was replaced by TCA co-exposure but there was no dose-response
between the 50, 500 and 5000 ppm. However, TCA treatment along with partial hepatectomy
and DEN treatment did increase the levels of foci (means of 0.71 - 0.39 foci/ cm2 at 3 months
02/12/09
E-329
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
and 1.83-2.45 foci/cm2 at 6 months) over treatment of just partial hepatectomy and DEN (0.05 ±
0.20 foci/ cm2 at 3 months and 0.30 ± 0.39 foci/ cm2 at 6 months). For the TCA animals treated
only with 5000 ppm TCA, the number of GGT positive foci at 3 months was 0.23 ±0.16 foci/
cm2 and at 6 months 0.03 ± 0.32 foci/ cm2 with no values for untreated animals presented. For
the positive control (partial hepatectomy/DEN/Phenobarbital) the number of GGT positive foci
increased from 3 to 6 months (1.65 ± 0.23 foci/ cm2 and at 6 months 7.61 ± 0.72 foci/ cm2). The
authors concluded that "although TCA is reported to cause hepatic peroxisomal stimulation in
rats and mice, the results of this study indicate that it is unlikely TCA's effects are related to the
promoting ability seen here. The minimal stimulation of, 10 to 20% over controls of
peroxisomal associated, PCO activity in TCA exposed rats was seen only at the 5000 ppm level
and only within the promotion protocol. This finding is in contrast to the promoting activity seen
at all three concentrations of TCA."
4.2.3. Pereira and Phelps 1996
The results for mice that were not "initiated" by exposure to MNU, but exposed to DCA
or TCA, are discussed in Section 2.3.2.6. However, differences in responses after initiation are
useful for showing differences between single and co-exposures as well as differences between
DCA and TCA effects. On day 15 of age, female B6C3F1 mice received an i.p. injection of
MNU (25 mg/kg) and at 7 weeks of age received DCA (2.0, 6.67, or 20 mmol/1), TCA (2.0, 6.67
mmol, or 20 mmol/1), or NaCl continuously for 31 or 51 weeks of exposure. The number of
animals studied ranged from 6 to 10 in 31-week groups and 6 to 39 in the 52-week groups.
There was a "recovery group" in which mice received either 20 mmol/1 DCA (2.58 g/1 DCA)
(n= 12) or TCA (3.27 g/1 TCA) (n= 11) for 31 weeks and then switched to saline for 21 weeks
until sacrifice at 52 weeks. Strengths of the study included the reporting of hepatocellular
lesions as either foci, adenomas, or carcinomas and the presentation of incidence and
multiplicity of each separately reported for the treatment paradigms. Limitations included the
low and variable number of animals in the treatment groups.
MNU was reported to not "significantly" induce foci or altered hepatocytes, adenomas,
or carcinomas at 31 (n= 10) or 51 weeks (n= 39). However, MNU did increase the incidence
and number/mouse of foci, adenomas and carcinomas at the 52 week sacrifice time in
comparison to saline controls, albeit at lower levels than observed in DCA or TCA co-
treatments groups (e.g., 10% vs. 0% foci, 17.5% vs. 2.5% adenomas, and 10% vs. 0% incidence
of carcinomas at 52 weeks for MNU-treated mice vs. saline control). Co-exposure of DCA
(20.0 mmol/1) for 52 weeks in MNU-treated mice increased the number of foci and
hepatocellular adenomas with the authors reporting "the yield of total lesions/mouse increased
as a second order function of the concentration of DCA (correlation coefficients > 0.998)."
TCA co-exposure in MNU-treated mice was reported to not result in a significant difference in
yield of foci or altered hepatocytes with either continuous 52 week or 31-week exposure, but
exposures to 20.0 or 6.67 mmol/1 TCA did result in increased yield of liver tumors with both
exposure protocols (see below).
02/12/09
E-330
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For TCA treatment in MNU treated mice, the incidences of foci were similar (12.5% vs.
18.2%) but the number of foci/ mouse was ~ 3-fold greater in the cessation protocol than with
continuous exposure. The incidence of adenomas was reported to be the same (~ 66%) as well
as the number of adenomas/animal between continuous and cessation exposures. For
carcinomas, there was a greater incidence for mice with continuous TCA exposure (83% vs.
36%) as well as a greater number of carcinomas/mouse (~ 4-fold) than for those initiated mice
with cessation of TCA exposure. As noted above, the number of animals treated with TCA was
low and variable (e.g., 23 mice studied at 52 weeks 20.0 mmol/1 TCA, and 6 mice at 52 weeks
6.6.7 mmol/1 TCA), limiting the ability to discern a statistically significant effect in regard to
dose-response. The concentration- response relationship for tumors/mouse after 31 and 51
weeks was reported to be best represented by linear progression.
A comparison of results for animals treated with MNU and 20.0 mmol/1 DCA or TCA
for 31 weeks and sacrificed at 31 weeks and those which were treated with MNU and DCA or
TCA for 31 weeks and then sacrificed at 52 weeks is limited by the number of animals exposed
(n= 10 for 31 week sacrifice DCA or TCA, n= 12 for DCA recovery group, and n= 11 for TCA
recovery group). No carcinoma data was reported for animals exposed at 31 weeks and
sacrificed at 31 weeks making comparisons with recovery groups impossible for this parameter
and thus determinations about progression from adenomas to carcinomas. For the MNU and
DCA-treated animals, the incidence or number of animals reported to have foci at 31 weeks was
reported to be 80% but 38.5% for in the recovery group. For adenomas the incidence was
reported to be 50% for DCA-treated animals at 31 weeks and 46.2% for the recovery group. For
MNU and TCA-treated animals, the incidence of foci at 31 weeks was reported to 20% and
18.2%) for the recovery group. For adenomas the incidence was reported to be 60% for the
TCA-treated animals at 31 weeks and 63.6% for the recovery group. Thus, this limited data set
shows a decrease in incidence of foci for the MNU and DCA-treated recovery group but no
change in incidence of foci for TCA or for adenomas for DCA- or TCA-treatment between
those sacrificed at 31 weeks and those sacrificed 21 weeks later. In regard to multiplicity, the
number of foci/mouse was reported to be 2.80 ± 0.20 for the 31-week DCA group and 0.46 ±
0.18 for the recovery group (mean ± SEM). The number of adenomas/mouse was reported to be
1.80 ± 0.83 for the 31-week group and 0.69 ± 0.26 for the recovery group. Thus, both the
number of foci and adenomas per mouse was reported to be decreased after the recovery period
for MNU and DCA treated mice. Given that the number of animals with foci was decreased by
half, the concurrent decrease in foci/mouse is not surprising. For TCA treatments, the numbers
of foci/mouse were reported to be 0.20 ±0.13 for the 31-week group and 0.45 ± 0.31 for the
recovery group. The number of adenomas/mouse for TCA-treatment groups was reported to be
1.30 ± 0.45 for the 31-week group and 0.91 ± 0.28 for the recovery group. For the MNU and
TCA-treated mice the numbers of foci/mouse were reported to be increased and the number of
adenomas/mouse reported to be slightly lower. Because carcinoma data is not presented for the
31 week group, it is impossible to determine whether the TCA adenomas regressed to foci or the
TCA adenomas progressed to carcinomas and more foci apparent with increased time.
02/12/09
E-331
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For the comparison of the numbers of foci, adenomas, or carcinomas per mouse that
were reported for the mice exposed at 31 weeks and sacrificed and those exposed for 52 weeks,
issues arise as to the impact of such few animals studied at 31 weeks, and the differing
incidences of lesions reported for these mice on tumor multiplicity estimates. The number of
animals studied who treated with MNU and 20.0 mmol/1 DCA or TCA for 31 weeks and then
sacrificed was n= 10, while the number of animals exposed to 20.0 mmol/1 DCA or TCA for 52
weeks was 24 for the DCA group and 23 for the TCA group. The number of animals treated at
lower concentrations of DCA or TCA were even lower at the 31-week sacrifice (e.g., n= 6 for
MNU and 6.67 mmol/1 DCA at 31 weeks) and also for the 52-week durations of exposure (e.g.,
n= 6 for MNU and 6.6.7 mmol/1 TCA).
At 31 weeks, 80% of the animals were reported to have foci and 50% to have foci after
52 weeks of exposure to 20.0 mmol/1 DCA and MNU treatment. Thus, similar to the "recovery"
experiment, the number of animals with foci decreased even with continuous exposure between
31 and 52 weeks. For adenomas, 20.0 mmol DCA exposure for 31 weeks was reported to
induce adenomas in 50% of mice and after 52 weeks of exposure to induce adenomas in 73% of
mice. For TCA, the number of animals with foci was reported to be 20% at 31 weeks and 12%
at 52 weeks after exposure to 20.0 mmol/1 TCA after MNU treatment and similar to the
incidence of foci reported for the TCA-recovery group. For 20.0 mmol TCA, adenomas
reported in 60% of mice after 31 weeks and in 67% of mice after 52 weeks of exposure and also
similar to the incidence of adenomas reported for the TCA-recovery group. In regard to
multiplicity, the number of foci/mouse was decreased from 2.80 ± 0.20 to 1.46 ± 0.48 between
31 weeks and 52 weeks of 20.0 mmol DCA in MNU exposed mice. The number of
adenomas/mouse was reported to be increased from 1.80 ± 0.83 to 3.62 ± 0.70 between 31
weeks and 52 weeks of 20.0 mmol DCA and MNU exposed mice. For 20.0 mmol/1 TCA, the
number of foci/mouse was 0.20 ±0.13 and 0.13 ± 0.7 for 31- and 52-week exposures. The
number of adenomas/mouse was reported to be 1.30 ± 0.45 and 1.29 ± 0.24 for 31- and 52-week
exposures. Thus, by only looking at foci and adenoma multiplicity data, there would not appear
to be a change between 31 and 52-weeks. However, during progression a shift may occur such
that foci become adenomas with time and adenomas become carcinomas with time. For
carcinomas there was no data reported for 31 week exposure in MNU and DCA- or TCA-treated
mice. However, at 52 weeks 20.0 mmol DCA was reported to induce carcinomas in 19.2% of
mice and 20.0 mmol TCA to induce carcinomas in 83% of mice. The corresponding numbers of
carcinomas/mouse was 0.23 ± 0.10 for 20.0 mmol/1 DCA treatment and 2.79 ± 0.48 for 20.0
mmol/1 TCA treatment at 52 weeks in MNU treated mice. Thus, although fewer than 20% of
MNU-treated mice were reported to have foci at 20.0 mmol TCA, by 52 weeks almost all had
carcinomas with ~ 67% also having adenomas. For DCA, many more mice had foci at 31
weeks (80%) than for TCA and by 52 weeks ~ 70% had adenoma with only ~ 20% reported to
have carcinomas. The incidence data are suggestive that as these high doses of DCA and TCA,
TCA was more efficient inducing progression of a carcinogenic response than DCA in MNU-
treated mice.
02/12/09
E-332
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The authors interpret the decrease in foci and adenomas between animals treated with
MNU and 20.0 mmol/1 DCA for 31 weeks and sacrificed and those sacrificed 21 weeks later to
indicate that these lesions were dependent on continued exposure. However, the total number of
lesions cannot be ascertained because carcinoma data was not reported for 31-week exposures.
Carcinomas were reported in the recovery group at 52 weeks (0.15 ± 0.10 carcinomas/mouse in
15.4% of animals). Of note is that not only did the number of foci/mouse and incidence
decrease between the 31-week group and the recovery group, but also between 31- and 52-
weeks of continuous exposure for the MNU and 20.0 mmol/1 DCA treated groups. Although
derived from very few animals, the 6.67 mmol/1 DCA group reported no change for foci/mouse
but a decrease in the incidence of foci between 31- and 52-weeks of exposure in MNU treated
mice (i.e., 0.67 ± 0.18 foci/mouse in 50% of the animals at 31 weeks and 0.50 ± 0.34 foci/mouse
in 20% of mice treated for 52 weeks). The numbers of foci/mouse for both MNU-treated and
untreated control mice were reported to be decreased between 31 and 51 weeks as well. As
noted in Section 3.1.8., the number of "nodules" in humans, which may be analogous to foci and
adenomas, can spontaneously regress with time rather than becoming hepatocellular
carcinomas. Also as tumors get larger with progression, the number of tumors/mouse can
decrease due to coalescence of tumors and difficulty distinguishing between them. While data
is suggestive of a decrease in the number of adenomas/mouse after cessation of DCA exposure,
the incidence data is similar between the 31-week exposure and recovery groups. Of note is that
the number of carcinomas/mouse and the incidence of carcinomas was reported to be similar
between the MNU-treated mice exposed continuously to 20.0 mmol/1 DCA for 52 weeks and
those which were treated for 31 weeks and then sacrificed at 52 weeks. Also of note is that,
although incidences and multiplicities of foci and adenomas was reported to be relatively low in
the 2.0 mmol/1 DCA exposure groups, at 52-weeks 40% of the mice tested had carcinomas with
0.70 ± 0.40 carcinomas/mouse. This was a greater percentage of animals with carcinomas and
multiplicity than that reported for the highest dose of DCA. This result suggests that the effects
in regard to tumor progression, and specifically for carcinoma induction, differ between the
lowest and highest doses used in this experiment. However, the low numbers of animals
examined for the lower doses, 31-weeks exposures, and in the recovery group decrease the
confidence in the results of this study in regard to the effects of cessation of exposure on tumor
progression.
In regard to tumor phenotype, in MNU-treated female mice that were not also exposed
to either DCA or TCA, all four foci and 86.7% of 15 adenomas were reported to be basophilic
and 13.3 % eosinophilic at the end of the 52 week- study. However, when MNU-treated female
mice were also exposed to DCA the number eosinophilic foci and tumors increased with
increasing dose after 52 weeks of continuous exposure.. At the 20.0 mmol/1 level all 38 foci
examined were eosinophilic and 99% of the tumors (almost all adenomas) were eosinophilic.
At the 2.0 mmol/1 DCA exposure there were no foci examined but about 5 of 9 tumors
examined (~ 2:1 carcinoma:adenoma ratio) were basophilic and the other 4 were eosinophilic.
For TCA co-exposure in MNU-treated mice, the 20 mmol/1 TCA treatment was reported to give
results of 1 of the 3 foci examined to be basophilic and 2 that were eosinophilic. For the 98
tumors examined (~ 2:1 carcinoma/adenoma ratio) 71.4% were reported to be basophilic and
02/12/09
E-333
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
28.6% were eosinophilic. At the 2.0 mmol/1 TCA exposure level, the 2 foci examined were
reported to be basophilic while the 6 tumors (all adenomas) were reported to be 50%
eosinophilic and 50% basophilic. Thus, after 52 weeks female mice treated with MNU and a
high dose of DC A had eosinophilic foci and adenomas and those treated with the high dose of
TCA had a mixture of basophilic and eosinophilic foci and tumors with a 3:1 ratio of tumors
(mostly carcinomas) being basophilic. At the lower doses of either DCA or TCA the tumors
tended to be mostly carcinomas for DCA and adenomas for TCA but both were ~ 50%
basophilic and 50% eosinophilic. The tumors observed from MNU treatment alone were all
adenomas and mostly 87% basophilic. Thus, not only did treatment concentrations of DCA and
TCA give a different result for tumor multiplicity and incidence, but also for tumor phenotype
in MNU treated female mice. Eosinophilic foci and tumors were reported to consistently be
GST-71 positive while basophilic lesions "did not contain GST-rc, except for a few scattered cells
or very small area comprising less than 5% of the tumor."
Thus, exposure to either DCA or TCA increased incidence and number of animals with
lesions (foci, adenomas, or carcinomas) in MNU- vs. non-treated mice (see Section 2.3.2.6
above). These results suggest that the pattern of foci, adenoma and carcinoma incidence,
multiplicity, and progression appeared to differ between TCA and DCA in MNU-treated female
mice. However, the low and variable number of animals used in this study, make quantitative
inferences between DCA and TCA exposures in "initiated" animals, problematic.
4.2.4. Tao et al. 2000
The source of liver tumors for this analysis was reported to be the study of Pereira and
Phelps (1996). Samples of liver "tumors" and "non-involved" liver was homogenized for
protein expression for c-Jun and c-Myc and therefore contained homogeneous cell types for
study. The term "liver tumors" was not defined so it cannot be ascertained as to whether the
lesions studied were altered foci, hepatocellular adenomas, or carcinomas. Liver tissues were
reported to be frozen prior to study which raises issues of m-RNA quality. Although this study
reports that there were no MNU-induced "tumors" the original paper of Pereira and Phelps
(1996) reports that there were four foci and 15 adenomas in MNU-only treated mice. The
authors reported no difference in c-Jun and c-Myc m-RNA from DCA or TCA-induced tumors
from mice "initiated" with MNU. DNA methyltransferase was reported to be decreased in non-
involved liver in MNU-only treated mice in comparison to that from TCA- and DCA-treated
mice. For a comparison between non-involved liver and tumors, tumors were reported to have
a greater level than did non-involved liver.
4.2.5. Lantendresse and Pereira 1997
This study used the tumors from Pereira and Phelps (1996), except for the MNU-treated
only groups and those groups treated with either DCA or TCA but not MNU initiation, to further
study various biomarkers. The omissions were cited as to be due to insufficient tissue. For
immunohistochemical evaluation of the molecular biomarkers other than GST-71, liver specimens
02/12/09
E-334
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
from 7 MNU/ 20.0 mmol DCA- (i.e. 2.58 g/1 DCA) treated and 6 MNU/20.0 mmol TCA - (i.e.,
3.27 g/1 TCA) treated female mice randomly selected. For GST-71, the number of animals from
which lesion specimens were derived, was 24 MNU/DCA-treated and 23 MNU/TCA-treated
mice. The DCA treated mice were reported to have 1-9 lesions/mouse and TCA treated mice 1-3
lesions/mouse. The number of lesions examined for each biomarker varied greatly. For TCA-
induced foci, no foci were examined for any biomarker except 3 lesions for GST-71, while for
DCA 12-15 foci were examined for each biomarker and 38 lesions examined for GST-71.
Similarly for TCA-induced adenomas, there were 8-10 lesions examined for all biomarkers with
32 lesions examined GST-71, while for DCA 12 lesions for all biomarkers with 94 lesions
examined for GST-71. Finally, for TCA-induced carcinomas there were 3-4 lesions examined per
group with 64 lesions examined for GST-71, while for DCA-induced carcinomas there were no
lesions examined for any biomarker except 3 examined for GST-71. The biomarkers used were:
GST-71, TGF-a, TGF-P, c-Jun, c-Fos, c-Myc, Cytochrome oxidase CYP2E1, and Cytochrome
oxidase CYP4A1.
MNU/DCA treatment was reported to produce "predominantly eosinophilic lesions" with
"in general, the hepatocytes of DCA-promoted foci and tumors were less pleomorphic and
uniformly larger and had more distinctive cell borders than the hepatocytes in lesions caused by
TCA. Parenchymal hepatocytes of DCA-promoted mice were uniformly hypertrophied, with
prominent cell borders, and the cytoplasm was markedly vacuolated, which was morphologically
consistent with the previous description of glycogen deposition in these lesions. In contrast,
TCA-promoted proliferative lesions tended to be basophilic, as previously reported, and were
composed of hepatocytes with less distinct cell borders, slight cytoplasmic vacuolization, and
greater variability in nuclear size and cellular size."
The hepatocytes of altered foci and hepatocellular adenomas from MNU-treated female
mice also treated with DCA were reported to stain positively for TGF-a, c-Jun, c-Myc, CYP 2E1,
CYP 4A1, and GST-71. The authors do not present the data for foci and adenomas separately but
as an aggregate and as the number of lesions with < 50% cells stained or the number of lesions
with > 50% cells stained either "minimally to mildly" or "moderately to densely" stained.
Because no carcinomas for DCA were examined and especially because no foci for TCA
analyses were included in the aggregates, it is difficult to compare the profile between TCA and
DCA exposure in initiated animals and to separate these results from the effects of differences in
tumor progression. Thus, any differences seen in these biomarkers due to progression from foci
to adenoma in DCA-induced lesions or from progression of adenoma to carcinoma in TCA-
induce lesions, was lost. If the results for adenomas had been reported separately, there would
have been a common stage of progression from which to compare the DCA and TCA effects on
initiated female mice liver tumors. For DCA-induced "lesions" (~ 50% foci and ~ 50%
adenomas), most lesions had > 50 % cells staining with moderate to dense levels for TGF-a, and
CYP2E1, CYP4A1, and GST-71 and most lesions had <50% cells staining for even minimally to
mild staining for TGF-P and c-Fos. For c-Jun and c-Myc the aggregate DCA-induced "lesions"
were heterogeneous in the amount of cells and the intensity of cell staining for these biomarkers
in MNU-treated female mice.
02/12/09
E-335
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
For the TCA "lesions" (~ 60% adenomas and ~ 30% carcinomas) the authors note that "in
general, the hepatocytes of tumors promoted by TCA demonstrated variable immunostaining.
With the exception of c-Jun, greater than 50% of the hepatocytes in TCA lesions were essentially
negative or stained only minimally to mildly for the protein biomarkers studies. In some
instances, particularly in TCA-promoted tumors, there was regional staining variability within
the lesions, including immunoreactivity for c-Jun and c-Myc proteins, consistent with clonal
expansion or tumor progression." As stated above, the term "lesion" refers to foci and adenomas
for DCA but for adenomas and carcinomas for TCA making inferences as to differences in the
actions of the two compounds through the comparisons of biomarkers confounded by the effects
of tumor progression. The largest differences in patterns between TCA induced "lesions" and
those by DCA appeared to be TGF-a (with no lesions having >50% cells stained mildly or
moderately/densely for TCA-induced lesions), CYP2E1 (with few lesions having >50%) stained
moderately/densely for TCA-induced lesions), CYP4A1 (with no lesions having >50%) stained
mildly or moderately/densely for TCA-induced lesions), and GST-71 (with all lesions having <
50%o cells stained even mildly for TCA-induced lesions). However, as shown by this data, while
the "lesions" induced by TCA and DCA had some commonalities within each treatment, there
was heterogeneity of lesions produced by both treatments in female mice already exposed to
MNU. Overall, the tumor biomarker pattern suggests differences in the effects of DCA and TCA
through differences in tumor phenotype they induce as co-exposures with MNU treated female
mice.
The authors note that non-lesion parenchymal hepatocytes in DCA-treated initiated mice
stained mostly negative for CYP2E1 and CYP4A1, while in TCA-treated mice staining patterns
in parenchymal non-lesions hepatocytes were centrilobular for CYP2E1 and panlobular for CYP
4A1 (a pattern for CYP4A1 that is opposite of that found in the TCA-induced lesions).
4.2.6. Pereira et al. 1997
This study used a similar paradigm as that of Pereira and Phelps (1996) to study co-
exposures of TCA and DCA to female B6C3F1 mice already exposed to MNU. At 15 days the
mice received 25 mg/kg MNU and starting at 6 weeks of age neutralized solutions of either 0,
7.8, 15.6, 25.0 mmol/1 DCA (n= 30 for control and 25 mmol/1 DCA and n=20 for 7.8 and 15.6
mmol/1 DCA), 6.0 or 25.0 mmol/1 TCA (n= 30 for 25.0 mmol/1 TCA and n=20 for 6.0 TCA), or
combinations of DCA and TCA that included 25.0 mmol/1 TCA + 15.6 mmol/1 DCA (n=20), 7.8
mmol/1 DCA + 6.0 mmol/1 TCA (n=25), 15.6 mmol/1 DCA + 6.0 mmol/1 TCA (45), 25.0 mmol/1
DCA + 6.mmol/1 TCA (n= 25). The corresponding concentrations of DCA and TCA in g/1 is 25
mmol = 3.23 g/1, 15.6 mmol = 2.01 g/1 and 7.8 mmol = 1.01 g/1 DCA and 25 mmol= 4.09 g/1 and
6.0 mmol = 0.98 g/1 TCA. Accordingly, the number of animals at the beginning of the study
varied between 20 and 45. At terminal sacrifice (after 44 weeks of exposure) the numbers of
animals examined were less with the lowest number examined to be 17 mice in the 7.8 mmol/1
DCA group and the largest to be 42 in the 15.6 mmol/1 DCA + 6.0 mmol/1 TCA exposed group.
02/12/09
E-336
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
The authors reported that only a total of eight hepatocellular carcinomas were found in the
study (i.e., 25.0 mmol/1 DCA induced 3 carcinomas, 7.8 mmol DCA + 6.0 mmol TCA induced
one carcinoma, and 25.0 mmol/1 TCA induced 4 carcinomas). Thus, they presented data for
foci/mouse, and adenomas/mouse and their sum of both as "total lesions." The incidences of
lesions (i.e., how many mice in the groups had lesions) were not reported. The shortened
duration of exposure (i.e., 44 weeks), the omission of carcinomas from total "lesion" counts
(precluding consideration of progression of adenomas to carcinomas), the lack of reporting of
tumor incidences between groups, and the variable and low numbers of animals examined in
each group make quantitative inferences regarding additivity of these treatments difficult. MNU
treated mice did have a neoplastic response, albeit low using this paradigm. For mice that were
only exposed to MNU (n=30 at terminal sacrifice) the mean number of foci, adenomas and
"lesions" per mouse were 0.21, 0.07 and 0.28, respectively. No data were given for mice without
MNU treatment but few lesions would be expected in controls. Pereira and Phelps (1996)
reported that saline-only treatment in 40 female mice for 51 weeks resulted in 0% foci, 0.03
adenomas/mouse in 2.5% of mice, and 0% carcinomas. In general, it appeared that the numbers
of foci, adenomas and the combination of both reported as "lesions" per mouse that would have
been predicted by the addition of multiplicities given for DCA, TCA and MNU treatments alone,
were similar to those observed as co-exposure treatments. The largest numbers of foci and
adenomas/mouse were reported for the 25.0 mmol/1 DCA and 6.0 mmol/1 TCA treatments in
MNU treated mice (mean of 6.57 "lesions'Vmouse) with the lowest number reported for 7.8
mmol/1 DCA and 6 mmol/1 TCA (mean of 1.16 "lesions'Vmouse).
The authors reported that the foci of altered hepatocytes were predominantly eosinophilic
in DCA-treated female mice initiated with MNU, while those observed after MNU and TCA
treatment were basophilic. MNU treatment alone induced 4 basophilic and 2 eosinophilic foci,
and 2 basophilic adenomas. MNU and DCA treatment was reported to produce only eosinophilic
foci and adenomas at the 25.0 mmol/1 DCA exposure level. At the 7.8 mmol/1 DCA level of
treatment in MNU-treated mice, 2 foci were basophilic, 4 were eosinophilic and the 1 adenoma
observed was reported to be eosinophilic. Thus, the concentration of exposure appeared to alter
the tincture of the foci observed after MNU and DCA exposure using this paradigm. In this
study, MNU and TCA treatment was reported to induce foci and adenomas that were all
basophilic at both 25.0 mmol/1 TCA and 6.0 mmol/1 TCA exposures. After 7.8 mmol/1 DCA +
6.0 mmol/1 TCA exposure, 2/23 foci were basophilic and 21/23 foci were reported to be
eosinophilic while all 4 adenomas reported for this group were eosinophilic.
Irrespective of treatment, eosinophilic foci for were reported to be GST-71 positive and
basophilic foci to be GST-71 negative. An exception was the 4 carcinomas in the group treated
with 25 mmol/1 TCA which were reported to be predominantly basophilic but contained small
areas of GST-71 positive hepatocytes.
It should be noted that the increased dose (up to3.23 g/1 DCA and 4/09 g/1 TCA) raises
issues of toxicity and effects on water consumption as other studies have noted toxicity at highly
doses of DCA and TCA. The use of an abbreviated duration of exposure in the study raises
02/12/09
E-337
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
issues of sensitivity of the bioassay at the lower doses used in the experiment. In particular, was
enough time provided to observe the full development of a tumor response? Finally, a question
arises as what can be concluded from the low numbers of foci examined in the study and the
affect of such low numbers on the ability to discern differences in these foci by treatment. As
with Pereira and Phelps, there appeared to be a difference the nature of the response induced by
co-exposure of MNU to relatively high vs. low DCA concentrations. Of note is that while this
experiment reported no hepatocellular carcinomas at the lowest dose of DCA at 44 weeks (7.8
mmol DCA), Pereira and Phelps (1996) reported that in 9 mice treated with MNU and 2.0 mmol
DCA for 52 weeks, there were no foci but 20% of mice had adenomas (0.20 adenomas/mouse)
and 40% of mice had carcinomas (0.70 carcinomas/mouse).
These results suggest that DCA co-exposure affects TCA-induced lesions. The authors
concluded that mixtures of DCA and TCA appear to be at least additive and likely synergistic
and similar to the pathogenesis of DCA.
4.2.7. Tao et al. 1998
The focus of this study was an examination of tumors resulting from MNU and DCA or
TCA exposure in mice with the source of tumors was reported to be the study of Pereira et al.
(1997). Thus, similar concerns discussed above for that study paradigm are applicable to the
results of this study. The authors stated that there were also two recovery groups in which
exposure was terminated 1 week prior to euthanization at week 44. The Pereira et al (1997)
study does not report a cessation group in the study. In this study the number of animals treated
in the cessation group, the incidences of tumors in the mice, and the number of tumors examined
were not reported. Another group of female B6C3F1 mice (7-8 weeks old) were reported to not
be administered MNU but given 25 mmol/1 DCA (3.23 g/1 DCA), 25 mmol TCA (4.09 g/1 TCA),
or control drinking water for 11 days (n= 7).
Hepatocellular adenomas in DCA-treated mice, adenomas and carcinomas in TCA-treated
mice were reported to be analyzed for %-5-methylcytosine in the DNA of tumor tissues. The
levels of 5-methylcytosine in liver DNA of mice administered DCA or TCA for 11 days were
reported to be reduced in comparison to control tissues (reduced to ~ 36% of control for DCA
and ~ 41% of control for TCA with the control value reported to be ~ 3.5% of DNA methylated).
The number of animals examined was reported to be 7-10 animals per group.
For control liver from mice that had received MNU but not DCA or TCA, and non-
involved liver after 44 week of exposure to either, the levels of 5-methylcytosine were similar
and not different from the ~ 3.5% of DNA methylated in untreated mice in the 11-days
experiment. Thus, initial decreases in methylated DNA shown by exposure to DCA or TCA
alone for 11 days, were not observed in "non-involved" liver of animals exposed to either DCA
or TCA and MNU.
02/12/09
E-338
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
In regard to tumor tissues, the level of 5-methylcytosine in DNA of hepatocellular
adenomas receiving DC A and MNU was reported to be decreased by 36% in comparison to
noninvolved liver from the same animals. When exposure to DCA was terminated for 1 week
prior to sacrifice the level of 5-methylcytosine in the adenomas was reported to be higher and no
longer differed statistically from the non-involved liver from the same animal or liver from
control animals only administered MNU. The number of samples was reported to be 9-16
samples without identification as to how many samples were used for each tumor analysis or how
many animals provided the samples (i.e., were most of the adenomas from on animal?)
For TCA the 5-methylcytosine level was reported to be reduced by 40% in hepatocellular
adenomas and 51% reduction in hepatocellular carcinomas in comparison to noninvolved liver
from the same animals. These levels were also reported to be less than that the control animals
administered only MNU. Termination of exposure to TCA 1 week prior to sacrifice was reported
to not produce a statistically significant change in the level of 5-methylcytosine in either
adenomas or carcinomas. The levels of 5-methylcytosine were reported to be lower in
carcinomas than adenomas (~ 20% reduction) and to be lower in the "recovery" carcinomas than
continuous carcinomas (~ 25%) but were not reported as statistically significant. The results are
reported to have been derived from 8-16 "samples each". Again information on the number of
animals with tumors, whether the tumors were from primarily from one animal, and which DNA
results are from 8 vs. 16 samples, was not provided by the authors. Given that Pereira et al.
(1997), the source for material of this study, reported that treatment of MNU and 25.0 mmol/1
TCA treatment for 44 weeks induced only 4 carcinomas, a question arises as to how many
carcinomas were used for the 44-week 5-methylcytosine results in this study for carcinomas (i.e.,
how can 8-16 samples arise from 4 carcinomas?). In addition, a question arises as to whether
there was a difference in tumor-response in those animals with and without one week of
cessation of exposure which cannot be discerned from this report. The use of highly variable
number of samples between analysis groups and lack of information as to how many tumors were
analyzed adds uncertainty to the validity of these findings. There did not appear to be a
difference in methylation activity from short-term exposure to either DCA or TCA alone in
whole liver DNA extracts. However, the authors conclude that the difference in methylation
status between tumors resulting from MNU and DCA or TCA exposures supports differences in
the action between DCA and TCA.
4.2.8. Stauber et al. 1998
In this study, 5-8 week old male B6C3F1 mice were used for isolation of primary
hepatocytes which were subsequently isolated and cultured in DCA or TCA. In a separate
experiment 0.5 g/1 DCA was given to mice as pretreatment for 2 weeks prior to isolation. The
authors note that and indication of an "initiated cell" is anchorage-independent growth. DCA
and TCA solutions were neutralized before use. The primary hepatocytes from 3 mice per
concentration were cultured for 10 days with DCA or TCA colonies (8 cells or more) determined
in quadruplicate. The levels of DCA used were 0, 0.2, 0.5 and 2.0 mM DCA or TCA. At
concentrations of 0.5 mM or more DCA and TCA both induced an increase in the number of
02/12/09
E-339
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
colonies that was statistically significant and increased with dose with DCA giving a slightly
greater effect. The authors noted that concentrations greater than 2.0 mM were cytotoxic but did
not show data on toxicity for this study.
Of great interest is the time-course experiment from this study in which the number of
colonies from DCA treatment in vitro peaked by 10 days and did not change through days 15- 25
at the highest dose. For the lower concentrations of DCA increased time in culture induced
similar peak levels of colony formation by days 20-25 as that reached by 10 days at the higher
dose. Therefore, the number of colonies formed was independent of dose if the cells were treated
long enough in vitro. The number of colonies that formed in control hepatocyte cultures also
increased with time but at a lower rate than those treated with DCA (2.0 mM DCA gave ~ 2-fold
of control by 25 days of exposure to hepatocytes in culture). However the level reached by cells
untreated in tissue culture alone by 20 days was similar to the level induced by 0.5 mM DCA by
10 days of exposure. This finding raises the issue of what these "colonies" represent as tissue
culture conditions alone transform these cells to what the authors suggest is an "initiated" state.
TCA exposure was not tested with time to see if it had a similar effect with time as did DCA.
At 10 days, colonies were tested for c-Jun expression with the authors noting that
"colonies promoted by DCA were primarily c-Jun positive in contrast to TCA promoted colonies
that were predominantly c-Jun negative." For colonies that arose spontaneously from tissue
culture conditions, 10/13 (76.9%) were reported to be c-Jun +, those treated with DCA 28/34
(82.3%) were c-Jun +, and those treated with TCA 5/22 (22.1%) were c-Jun +. These data show
heterogeneity in cell in colonies although more were c-Jun + with DCA than TCA. The number
of colonies reported in the c-Jun labeling results represent sums between experiments and thus
present total numbers of the control and the of colonies derived from doses of DCA and TCA at
0.2 to 2.0 mM at 10 days. Thus, changes in colony c-Jun+ labeling due to increasing dose cannot
be determined. The authors reported that with time (24, 48, 72, and 96 hours) of culture
conditioning the number of c-Jun+ colonies was increased in untreated controls. DCA treatment
was reported to delay the increase in c-Jun+ expression induced by tissue culture conditions
alone in untreated controls. TCA treatment was reported to not affect the increasing c-Jun+
expression that increased with time in tissue culture. In this instance, tissue culture environment
alone was shown to transform cells and can be viewed as a "co-exposure." DCA pretreatment in
vivo was reported to increase the number of colonies after plating which reached a plateau at 0.10
mM and gave changes as at low a concentration of 0.02mM DCA administered in vitro. The
background level of colony formation varied between controls (i.e., 2-fold different in
pretreatment experiments and non-pretreatment experiments). Therefore although the number of
colonies was greater for pretreatment with DCA, the magnitude of difference over the control
level was the same after DCA treatment in vitro with and without pretreatment.
The authors presented a comparison of "tumors" from Stauber and Bull (1997) and state
that DCA tumors were analyzed after 38 weeks of treatment but that TCA tumors were analyzed
after 52 weeks. They note that 97.5% of DCA-induced "tumors" were c-Jun + while none of the
TCA-induced "tumors" were c-Jun +. The concentrations used to give tumors in vivo for
02/12/09
E-340
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
comparison with in vitro results were not reported. What was considered to be "tumors" from
the earlier report for this analysis was also not noted. Stauber and Bull (1997) reported results
for combination of foci and tumors raising issues as to what was examined in this report. The
authors stated that because of such short time, no control tumors results were given. The short
and variable time of duration of exposure increases the possibility of differences between the in
vivo data resulting from differences in tumor progression as well as a decreased ability by the
shortened time of observation for full expression of the tumor response.
4.3. Co-exposures of Haloacetates and Other Solvents
As noted by Caldwell et al. (2008b), drinking water exposure data suggest co-exposure of
TCE and its haloacetic acid metabolites, TCA and DCA, is not an uncommon event as DCA and
TCA are the two most abundant haloacetates in most water supplies (Weisel et al., 1999,
Boorman et al., 1999). Dibromoacetic acid (DBA) concentrations have also been reported to
range up to approximately 20 |ig/L in finished water and distribution systems (Weinberg et al.,
2002). Caldwell et al. (2008b) have also noted that co-exposure in different media also occurs
with solvents like perchloroethylene (PERC) that may share some MO As, targets of toxicity, and
common metabolites that can therefore potentially affect TCE health risk (Wu and Schaum,
2000). Some of the information contain in the following sections have been excerpted from the
discussions by Caldwell et al (2007b) regarding the implications for the risk of TCE exposure as
modulated by co-exposures to haloacetates and other solvents that have been studied and reported
in the literature.
4.3.1. Carbon tetrachloride, DCA, TCA: Implications for MOA from Co-exposures
Studies of specific combinations of TCE and chemicals co-located in contaminated areas
have been reported by Caldwell et al. (2008b). For carbon tetrachloride "Pretreatment with TCE
in drinking water at levels as low as 15 mM for three days has been reported to increase
susceptibility to liver damage to subsequent exposure to a single IP injection of 1 mM/kg carbon
tetrachloride (CCU) in Fischer 344 rats [Steup et al., 1991], Potential mechanistic explanations
for this observation included altered metabolism, decreased hepatic repair capability, decreased
detoxification ability, or combination of one or more of the above activities. Simultaneous
administration of an oral dose of TCE (0.5ml/kg) has also been reported to increase the liver
injury induced by an oral dose of 0.05 ml/kg CCI4 [Steup et al., 1993], The authors suggested
that TCE appeared to impair the regenerative activity in the liver, thus leading to increased
damage when CCI4 is given in combination with TCE."
As discussed above in Section 4.2, initiation studies are in themselves a co-exposure. The
study of Bull et al. (2004) is included here as it not only used a co-exposure of vinyl carbamate
with TCE metabolites, but also used carbon tetrachloride as a co-exposure as well. The rationale
for this approach was that co-exposure of TCE (and therefore to its metabolites) and CC14 are
likely to occur as they are commonly found together at contaminated sites. Bull et al. (2004)
hypothesized that modification of tumor growth rates is an indication of promotion rather than
02/12/09
E-341
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
effects on tumor number, and that by studying tumor growth rates they could classify carcinogens
by their MOAs. B6C3F1 male mice were initiated with vinyl carbamate (3 mg/kg) at 2 weeks of
age and then treated with DCA, TCA, CCU, (0.1, 0.5, or 2.0 g/1 for DCA and TCA; 50, 100 or
500 mg/kg CCL4 in 5% Alkamuls via gavage) in pair-wise combinations of the three for 18 to 36
weeks. The exposure level of CCL4 to 5, 20 and 50 mg/kg was reported to be reduced at week 24
due to toxicity for CC14. The number of mice in each group was reported to be 10 with the study
divided into 5 segments. There were evidently differences between treatment segments as the
authors state that" because of some significant quantitative differences in results that were
obtained with replicate experiments treated in different time frames, the simultaneous controls
have been used in the analysis and presentation of these data." As with Bull et al. (2002), the
interpretation of the results of the study is limited by a low number of animals per group, short
duration time of exposure and limited examination and reporting of results. For example, a
sample of 100 out of the 8000 lesions identified in the study was examined to verify that the
general descriptor of neoplastic and non-neoplastic lesion was correctly labeled with "tumors"
describing a combination of hyperplastic nodules, adenomas, and carcinomas. No incidence data
was reported by the authors, but general lesion growth information included mean lesion volume
and multiplicity of lesions (numbers of lesions/mouse). Using these reported indices, there
appeared to be differences in treatment-related effects.
Each treatment was examined alone and then in differing combinations with each other.
Mice initiated with vinyl-carbamate, but without further exposure to the other toxicants, were
reported to have a few lesions that were of small size during the examination period (20-36
weeks). At 30 weeks of CCI4 exposure, there was a dose-related response reported for
multiplicity but mean lesion size was smaller at the highest dose in initiated animals. At 36
weeks, DCA exposure was reported to increase multiplicity at the two highest exposure levels
and increased lesion size at all levels compared to initiated-only animals. However, at a similar
level of induction, multiplicity and mean size of those lesions resulting from DCA treatment were
reported to be much smaller in comparison with CCI4 treatment (i.e., a 20-fold difference for
lesion volume). At 36 weeks, treatments with the same concentration of TCA or DCA induced
similar multiplicity, but the mean lesion volume was reported to be approximately 4-fold greater
in tumors induced by DCA as compared to TCA, and in animals treated with DCA multiplicity
had reached a plateau by 24 weeks rather than 36 for those treated with TCA. Thus, using
multiplicity of lesions and lesion volume as indicators of differences in MO A, exposure to CCI4,
DCA, and TCA appeared to produce distinct differences in results in animals previously treated
with vinyl carbamate.
Simultaneous co-exposure of differing combinations of CCI4, DCA, and TCA were
reported to give more complex results between 24 and 36 weeks of observation but to show that
co-exposure had effects on lesion multiplicity and volume in initiated animals. At 36 weeks,
TCA co-exposure appeared to reduce the lesion volume of either DCA- or CCU-induced lesions
after vinyl carbamate treatment. Similarly, DCA co-exposure was reported to reduce the lesion
volume of either TCA- or CCU-induced lesions when each was given alone after vinyl carbamate
treatment. With regard to multiplicity, TCA co-exposure was reported to reduce DCA-induced
02/12/09
E-342
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
multiplicity only at the lowest dose of TC A while co-exposure with DCA increased multiplicity
of CCl4-induced lesions at all exposure levels. At 24 weeks, there appeared to be little effect on
mean lesion volume by any of the co-exposures but DCA co-exposure decreased multiplicity of
TCA-induced lesions (up to 3-fold) while TCA treatment slightly increased the number of CC14-
induced multiplicity (1.6-fold). This study confirms that short duration of exposure to all three of
these chemicals can cause lesions in already exposed to vinyl carbamate, and suggests that
combinations of these agents differentially influence lesion number and growth rates. The
authors have interpreted their results to indicate differences in MOA between such treatments.
However, the limitations of the study limit conclusions regarding how such co-exposure may be
able to affect toxicity and tumor induction and what the MOA is for each of these agents. This is
especially true at lower and more environmentally relevant concentrations given for longer
durations to uninitiated animals (Caldwell et al 2007b).
4.3.2. Chloroform, DCA, and TCA Co-exposures: Changes in Methylation Status
In Section 3.4.2.2, information on the effects of TCE and its metabolites was presented in
regard to effects on methylation status. After 7 days of gavage dosing, TCE, TCA and DCA
were reported to increased hypomethylation of the promoter regions of c-Jun and c-Myc genes in
mouse whole liver DNA, however, Caldwell and Keshava (2006) concluded that
hypomethylation did not appear to be a chemical-specific effect at the concentration used. Bull
et al. (2004) suggested that hypomethylation occurs at higher exposure levels than those that
induce liver tumors for TCE and its metabolites. Along with studies of methylation changes
induced by a exposure to a single agent a Pereira et al (2001) have attempted to examine the
effects on methylation changes from co-exposures. This study was also reviewed by Caldwell et
al (2007b).
Pereira et al. (2001) hypothesized that changes in the methylation status of DNA can be a
key event for MOA for DCA- and TCA- induced liver carcinogenicity through changes in gene
regulation, and that chloroform (CHCI3) co-exposure may result in modification of DNA
methylation. After 17 days of exposure of exposure to CHCI3 (0, 400, 800, 1600 mg/1, n= 6 mice
per treatment group) female B6C3F1 mice were co-exposed to DCA or TCA (500 mg/kg) during
the last 5 days of exposure to chloroform. As noted by Caldwell et al (2007b), Pereira et al.
(2001) reported the effects of hypomethylation of the promoter region of c-Myc gene and on
expression of its mRNA in the whole livers of female B6C3F1 mice and thus, these results
represent composite changes in DNA methylation status from a variety of cell types and for
hepatocytes lumped from differing parts of the liver lobule. When given alone, DCA, TCA, and
to a lesser extent, the highest concentration of CHCI3 (1600 mg/L), was reported to decrease
methylation of the c-myc promoter region. Co-administration of CHCI3 (at 800 and 1600 mg/1)
was reported to decrease DCA-induced hypomethylation while exposure to CHCI3 had no effect
on TCA-induced hypomethylation. Treatment with DCA, TCA, and, to a lesser extent CHCI3,
was reported to increase levels of c-myc mRNA. While expression of c-myc mRNA was
increased by DCA or TCA treatment, increasing co-exposures to CHCI3 were reported to
attenuate the actions of DCA but not TCA. Thus, differences in methylation status and
02/12/09
E-343
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
expression of the c-myc gene induced by DCA or TCA exposure was reported to be differentially
modulated by co-exposure to CHC13. The authors suggest these differences support differing
actions by DCA and TCA. However, whether these changes represent key events in the
induction of liver cancer is a matter of debate, especially as a "snapshot in time" approach for
such a non-specific endpoint (Caldwell et al 2007b).
In a co-exposure study in which an "initiating agent" was used as a co-exposure along
with other co-exposure, Pereira et al. (2001) treated male and female 15-day old B6C3F1 mice
with N-methyl-N-nitrosourea (MNU) (a cause of liver and kidney tumors) and then, starting at 5
weeks of age, treated them with DCA (3.2 g/L) or TCA (4.0 g/L) along with co-exposure to
CHCI3 (0, 800 or 1600 mg/1) for 36 weeks. Mice were reported to be examined for evidence of
promotion of liver and kidney tumors. The numbers of animals in the exposure groups were
highly variable, ranging from 25 (female initiated mice exposed to DCA) to 6 (female initiated
mice exposed to DCA and 1600 mg/L CHCI3), thus limiting the power of the study to ascertain
treatment-related changes. However, unlike Bull et al (2004), all liver tissues were examined
with incidences of foci, adenomas, carcinomas, and both adenoma and carcinoma reported
separately for treatment groups. Multiplicity for a combination of adenomas and carcinomas
were reported as well as the tincture of foci and tumors.
Although as noted by Caldwell et al (2007b), the statistical power of the study to detect
change was very low, an examination of the pattern of tumors induced by co-exposure to MNU
and TCE metabolites in female mice suggested that: (1) DCA exposure increased the incidence
of adenomas but not carcinomas; (2) TCA increased incidence of carcinomas with little change in
adenoma incidence; (3) co-exposure to 800 and 1600 mg/1 of CHCI3 decreased adenoma
incidence by DCA treatment but not TCA; and (4) CHCI3 co-exposure decreased multiplicity of
TCA-induced tumors and foci, but not for DCA. Caldwell et al (2007b) also note that this study
suggests a gender-related effect on tumor induction from this study with; (1) adenoma incidences
similar in male and female mice treated with DCA, but carcinoma incidence greater in males; (2)
adenoma and carcinoma incidences greater in males than females treated with TCA; (3) tumor
multiplicity similar in both genders for DCA treatments, but lower in females mice for TCA; and
(4) less of an inhibitory effect by CHCI3 on adenoma incidence from DCA exposure in male
mice.
Pereira et al. (2001) also described the tinctural characteristics of the specific lesions
induced by their co-exposure treatments. Both foci and tumors induced by DCA exposure in
"initiated" mice were reported to be over 95% eosinophilic in females, while in males, 89% of
the foci were eosinophilic and 91% of tumors were basophilic. Thus, not only was there a
gender-related difference in the incidences of tumors and foci but also foci and tumor phenotype.
CHCI3 co-exposure was reported to change the DCA-induced foci from primarily eosinophilic to
basophilic (i.e., 11% vs. 75% basophilic) in male mice co-exposed to MNU. In male and female
mice, TCA-induced tumors and foci were basophilic with no effect of CHCI3 on phenotype in
MNU treated mice.
02/12/09
E-344
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
4.3.3. Co-exposures to Brominated Haloacetates: Implications for Common MOAs and
Background Additivity to Toxicity.
As noted by Caldwell et al (2007b), along with chlorinated haloacetates and other
solvents, co-exposures with TCE and brominated haloacetates may occur through drinking water.
These compounds may affect TCE toxicity in a similar fashion to their chlorinated counterparts.
As bromide concentrations increase, brominated haloacetates increase in the water supply.
Kato-Weinstein et al. (2001) administered dibromoacetate (DBA), bromochloroacetate
(BCA), bromodichloroacetate (BDCA), TCA, and DCA in drinking water at concentrations of
0.2-3 g/L for 12 weeks to B6C3F1 male mice. The focus of the study was to determine the
similarity in action between the brominated and chlorinated haloacetates. Each of the
haloacetates, given individually, were reported to increase liver/body weight ratios in a dose-
dependent manner. The dihaloactates, DCA, BCA and DBA, caused liver glycogen
accumulation both by chemical measurements in liver homogenates and in ethanol-fixed liver
sections (to preserved glycogen) stained with PAS. For DCA, a maximal level of glycogen
increase was observed at 4 weeks of exposure at a 2 g/1 exposure concentration. They report a
1.60-fold of control % liver/body weight and 1.50-fold of control glycogen content after 8 weeks
of exposure to 2 g/1 DCA in male B6C3F1 mice. The baseline level of glycogen content (~ 60
mg/g) and the increase in glycogen after DCA exposure was consistent with the results reported
by Pereira et al. (2004). The % liver/body weight data for control mice was for animals sacrifice
at 20 weeks of age. The 4-12 week exposure to DCA were staggered so that all animals would
be 20 weeks of age at sacrifice. Thus, the animals were at differing ages at the beginning of DCA
treatments between the groups. However, as with Pereira et al. (2004) the ~ 10% increase in
liver mass that the glycogen increases represent are lower than the total increase in liver mass
reported for DCA exposure. The authors noted possible contamination of BCA with small
percentages of DCA and DBA in their studies (i.e., 84% BCA, 6% DCA and 8% DBA). The
trihaloacetates (TCA and low concentrations of BDCA) were reported to produce slight
decreases in liver glycogen content, especially in the central lobular region in cells that tended to
accumulate glycogen in control animals. These effects on liver glycogen were reported at the
lowest dose examined (i.e., 0.3 g/1). At the highest concentration, BDCA was reported to induce
a pattern of glycogen distribution similar to that of DCA in mice.
All dihaloacetates were reported to reduce serum insulin levels at high concentrations.
Conversely, trihaloacetates were reported to have no significant effects on serum insulin levels.
For the study of peroxisome proliferation and DNA synthesis, mice were treated to BCA, DBA
and BDCA for 2, 4, or 26 weeks. The effects on DNA synthesis were small for all brominated
haloacetates with only DBA reported to show a significant increase in DNA synthesis at 2 and 4
weeks but not at 26 weeks (increase in DNA synthesis was 3-fold of the highest control level). Of
note is the highly variable level of DNA synthesis reported for control values that varied to a
much higher degree (~3- 6-fold variation within control groups at the same time points) than did
treatment-related changes. DBA was the only brominated haloacetate that was reported to
consistently increased PCO activity as a percentage of control values (actual values and
02/12/09
E-345
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
variability between controls were not reported) with a 2-3 fold increase in PCO activity at 0.3 to
3.0 g/1 DBA. DBA-induced PCO activity increases were reported to be limited to 2-4 weeks of
treatment in contrast to TCA, which the authors reported to increase PCO activity consistently
over time.
Tao et al. (2004) reported DNA methylation, glycogen accumulation and peroxisome
proliferation after exposure of female B6C3F1 mice and male Fischer 344 rats exposed to 1 or 2
g/1 DBA in drinking water for 2 to 28 days. DBA was reported to induce dose-dependent DNA
hypomethylation in whole mouse and rat liver after 7 days of exposure with suppression
sustained for the 28-day exposure period. The expression of mRNA for c-Myc in mice and rats
and mRNA expression of the 1GF-11 gene in female mice were reported to be increased during
the same period. Both rats and mice were reported to exhibit increased glycogen with mice
having increased levels at 2 day and rats at 4 days. DBA was reported to cause an increase in
lauroyl-CoA oxidase activity (a marker of peroxisome proliferation) in both mice (after 7 days)
and rats (after 4 days) that was sustained for 28 days. Methylation changes reported here for
DBA exposure in rats and mice are consistent with those reported for TCA and DCA by Pereira
et al. (2001) in mice. The pattern of glycogen accumulation was also similar to that reported for
DCA by Kato-Weinstein et al. (2001) and suggests that the brominated analogues of TCE
metabolites exhibited similar actions as their chlorinated counterparts. In regard to peroxisomal
enzyme activities Kato-Weinstein et al. (2001) reported PCO activity to be limited to 2-4 weeks
with Tao et al (2004) reporting lauroyl-CoA oxidase activity to be sustained for the lengths of the
study (28-days) for DBA.
As noted by Caldwell et al (2007b), given the similarity of DCA and DBA effects, it is
plausible that DBA exposure also induces liver cancer. Melnick et al. (2007) reported the results
of DBA exposure to F344/N rats and B6C3F1 mice exposed to DBA for 3 months or 2 years in
drinking water (0, 0.05, 0.5, or 1.0 g/1 DBA for 2 years). Neoplasms at multiple sites were
reported in both species exposed to DBA for 2 years with no effects on survival and little effect
on mean body weight in either species. Similar to TCE, DCA and TCA, the liver was reported
to be a target of DBA exposure. After 3-months of exposure, there were dose-related increases in
hepatocellular vacuolization and liver weight reported in rats and mice described as "glycogen-
like." The authors report that the major neoplastic effect of DBA in rats was induction of
malignant mesotheliomas in males and increased incidence of mononuclear cell leukemia in
males and females. For mice, the major neoplastic effect of DBA exposure was reported to be
the increased incidence of hepatocellular adenomas and carcinomas at all exposure levels. In
addition to these liver tumors, hepatoblastomas were also reported to be increased in all exposure
groups of male mice and exceeded historical control rates. The incidence of alveolar/bronchiolar
adenoma and carcinoma was reported to be increased in the 0.5 g/L group of male mice along
with marginal increases in alveolar hyperplasia in DBA-treated groups. The authors reported that
the increases in hepatocellular neoplasms were not associated with hepatocellular necrosis or
regenerative hyperplasia and concluded that an early increase in hepatocyte proliferation were
not likely involved in the MOA for DBA because no increases in hepatocyte DNA labeling index
02/12/09
E-346
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
were observed in mice exposed for 26 days and the slight increase that occurred in male F344
rats was not accompanied by an increase in liver tumor response.
As noted by Caldwell et al (2007), the results of Kato-Weinstein et al. (2001), Tao et al.
(2004), and Melnick et al. (2007) are generally consistent for DBA and show a number of
activities that may be common to TCE metabolites (i.e., brominated and chlorinated haloacetate
analogues generally have similar effects on liver glycogen accumulation, serum insulin levels,
peroxisome proliferation, hepatocyte DNA synthesis, DNA methylation status, and
hepatocarcinogenicity). It is therefore plausible that these effects may be additive in situations
of co-exposure. However, as noted by Melnick et al (2007), methylation status, events associated
with PPARa agonism, hepatocellular necrosis, and regenerative hyperplasia are not established
as key events in the MOA of these agents, and the MO As for DC A- and DBA- induced liver
tumors are unknown.
4.3.4. Co-exposures to Ethanol: Common Targets and MOAs
As noted in the U.S. Environmental Protection Agency's draft TCE assessment [U.S.
EPA, 2001], alcohol consumption is a common co-exposure that has been noted to affect TCE
toxicity with TCE exposure cited as potentially increasing the toxicity of methanol and ethanol,
not only by altering their metabolism to aldehydes, but also by altering their detoxification (e.g.,
similar to the "alcohol flush" reported for those who have an inactive aldehyde dehydrogenase
allele). As noted by Caldwell et al (2007b) "chemical co-exposures from both the environment
and behaviors such as alcohol consumption may have effects that overlap with TCE in terms of
active agents, pharmacokinetics, pharmacodynamics, and/or target tissue toxicity." Caldwell et
al (2007b) also note:
In their review of solvent risk (including TCE), Brautbar and Williams (2002) suggest
that laboratory testing that is commonly used by clinicians to detect liver toxicity may not
be sensitive enough to detect early liver hepatotoxicity from industrial solvents and that
the final clinical assessment of hepatotoxicity and industrial solvents must take into
account synergism with medications, drugs of use and abuse, alcohol, age-dependent
toxicity, and nutrition. Although many of these factors may be important, the focus in
this section is on the effects of ethanol. Contemporary literature reports effects similar to
those of TCE's and previous reports indicate ethanol consumption impacts TCE toxicity
in humans, affects the pharmacokinetics and toxicity of TCE in rats, and is also a risk
factor for cancer.
The association between malignant tumors of the upper gastrointestinal tract and liver and
ethanol consumption is based largely on epidemiological evidence, and thought to be
causally related [Bradford et al., 2005; Badger et al., 2003], Studies of the mechanisms of
ethanol carcinogenicity have suggested the importance of its metabolism, including
induction of CYP2E1 associated increases in production of reactive oxygen species and
enhanced activation of a variety of pro-carcinogens, alteration of retinol and retinoic acid
02/12/09
E-347
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
metabolism, and the actions of the metabolite acetaldehyde [Badger et al., 2003], While
ethanol is primarily metabolized by alcohol dehydrogenase, it undergoes simultaneous
oxidation to acetate by hepatic P450s, primarily CYP2E1. Both chronic ethanol
consumption as well as TCE treatment induces CYP2E1 [U.S. EPA, 2001], Oneta et al.
(2002) report that even at moderate chronic ethanol consumption, hepatic CYP2E1 is
induced in humans, which they suggest, may be of importance in the pathogenesis of
alcoholic liver disease; of ethanol, drug, and vitamin A interactions; and in alcohol-
associated carcinogenesis. Induction of CYP2E1 can cause oxidative stress to the liver
from nicotinamide dinucleotide phosphate (NADPH)-dependent reduction of dioxygen to
reactive products even in the absence of substrate, and subsequent apoptosis [Badger et
al., 2003], Bradford et al. (2005) suggest that CYP2E1, and not NADPH oxidase, is
required for ethanol-induced oxidative DNA damage to rodent liver but that NADPH
oxidase-derived oxidants are critical for the development of ethanol-induced liver injury.
There is increasing evidence that acetaldehyde, which is toxic, mutagenic, and
carcinogenic, rather than alcohol is responsible for its carcinogenicity [Badger et al.,
2003], Mitochondrial aldehyde dehydrogenase (ALDH2) disposes of acetaldehyde
generated by the oxidation of ethanol, and ALDH2 inactivity through mutation or
polymorphism has been linked to esophageal cancer in humans (everyday drinkers and
alcoholics) [Badger et al., 2003], For instance, increased esophageal cancer risk was
reported for patients with the ALDH3*1 polymorphism as well as increased acetaldehyde
in their saliva. TCE exposure has also been reported to induce a similar alcohol flush in
humans which may be linked to its ability to decrease ALDH activities at relatively low
concentrations and thus conferring a similar status to individuals with inactive ALDH2
allele [Wang et al., 1999], Whether the MO A for the buildup of acetaldehyde after
ethanol and TCE co-exposure is: (1) the induction of CYP2E1 by TCE resulting in
increased metabolism to acetaldehyde; (2) inhibition of ALDH and thus reduced
clearance of acetaldehyde, or (3) a number of other actions are unknown. Crabb et al.
(2001) reported 20-30% reductions in ALDH2 protein level by PPARa agonists
(Clofibrate treatment in rats and WY treatment in both wild and PPARa-null mice). This
could be another pathway for TCE-induced effects on ethanol metabolism. It is an
intriguing possibility that the reported association between the increased risk of human
esophageal cancer and TCE exposure [Scott and Chiu, 2006] could be related to TCE
effects on mitochondrial ALDH, given a similar association of this endpoint with ethanol
consumption or ALDH polymorphism.
Finally, ethanol ingestion may have significant effects on TCE pharmacokinetics.
Baraona et al. (2002 a,b) reported that chronic, but not acute, ethanol administration
increased the hepatotoxicity of peroxynitrite, a potent oxidant and nitrating agent, by
enhancing concomitant production of nitric oxide and superoxide. They also reported that
nitric oxide mediated the stimulatory effects of ethanol on blood flow. Ethanol markedly
enhanced portal blood flow (2-fold increase), with no changes in the hepatic, splenic, or
pancreatic arterial blood flows in rats.
02/12/09 E-348
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
4.3.5. Co-exposure Effects on Pharmacokinetics: Predictions Using Physiologically Based
Pharmacokinetic (PBPK) Models
Along with experimental evidence that has focused on chronic and acute experiments
using rodents, the potential pharmacokinetic modulation of risk has also been recently published
reports using physiologically based pharmacokinetic (PBPK) models that may be useful in
predicting co-exposure effects. Caldwell et al. (2008b) also examined and discussed these
approaches and note:.
An important issue for prediction of the effects and relationship on MO As by co-exposure
is the degree to which modulation of TCE toxicity by other agents can be quantified.
Pharmacokinetics or the absorption, distribution, metabolism, and elimination of an agent,
can be affected by internal and external co-exposure. Such information can help to
identify the chemical species that may be causally associated with observed toxic
responses, the MO A, and ultimately identify potentially sensitive subpopulations for an
effect such as carcinogenicity.
Physiologically based pharmacokinetic (PBPK) models are often used to estimate and
predict the toxicologically relevant dose of foreign compounds in the body and have been
developed to predict effects on pharmacokinetics that are additive or less or greater than
additive. One of the first such models was developed for TCE [Andersen et al., 1987],
Given that TCE, PERC, and methyl chloroform (MC) are often found together in
contaminated groundwater, Dobrev et al. (2001) attempted to investigate the
pharmacokinetic interactions among the three solvents to calculate defined "interaction
thresholds" for effects on metabolism and expected toxicity. Their null hypothesis was
defined as competitive metabolic inhibition being the predominant result for TCE given
in combination with other solvents. Gas uptake inhalation studies were used to test
different inhibition mechanisms. A PBPK model was developed using the gas uptake
data to test multiple mechanisms of inhibitory interactions (i.e., competitive,
noncompetitive, or uncompetitive) with the authors reporting competitive inhibition of
TCE metabolism by MC and PERC in simulations of pharmacokinetics in the rat.
Occupational exposures to chemical mixtures of the three solvents within their Threshold
Limit Value (TLV)/TWA limits were predicted to result in a significant increase (22%) in
TCE blood levels compared with single exposures.
Dobrev et al. (2002) extended this work to humans by developing an interactive human
PBPK model to explore the general pharmacokinetic profile of two common biomarkers
of exposure, peak TCE blood levels, and total amount of TCE metabolites generated in
rats and humans. Increases in the TCE blood levels were predicted to lead to higher
availability of the parent compound for GSH conjugation, a metabolic pathway that may
be associated with kidney toxicity/carcinogenicity. A fractional change in TCE blood
concentration of 15% for combined TLV exposures of the three chemicals (25/50/350
ppm of PERC/TCE/MC) resulted in a predicted 27% increase of the S-(l, 2-
02/12/09 E-349
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
dichlorovinyl)-L-cysteine (DCVC) metabolites, indicating a nonlinear risk increase due to
combined exposures to TCE. Binary combinations of the solvents produced GST-
mediated metabolite levels almost twice as high as the expected rates of increase in peak
blood levels of the parent compound. The authors suggested that using parent compound
peak blood levels (a less sensitive biomarker) would result in two to three times higher
(i.e., less conservative) estimates of potentially safe exposure levels. In regard to the
detection of metabolic inhibition by PERC and MC, the simulations showed TCE blood
concentrations to be the more sensitive dose metric in rats, but the total of TCE
metabolites to be the more sensitive dose measure in humans. Finally, interaction
thresholds were predicted to be occurring at lower levels in humans than rats.
Thrall and Poet (2000) investigated the pharmacokinetic impact of low-dose co-exposures
to toluene and TCE in male F344 rats in vivo using a real-time breath analysis system
coupled with PBPK modeling. The authors report that, using the binary mixture to
compare the measured exhaled breath levels from high- and low-dose exposures with the
predicted levels under various metabolic interaction simulations (competitive,
noncompetitive, or uncompetitive inhibition), the optimized competitive metabolic
interaction description yielded an interaction parameter Ki value closest to the Michaelis-
Menten affinity parameter (KM) of the inhibitor solvent. This result suggested that
competitive inhibition is the most plausible type of metabolic interaction between these
two solvents.
Isaacs et al. (2004) have reported gas uptake co-exposure data for CHCI3 and TCE. The
question as to whether it is possible to use inhalation data in combination with PBPK
modeling to distinguish between different metabolic interactions was addressed using
sensitivity analysis theory. Recommendations were made for design of optimal
experiments aimed at determining the type of inhibition mechanisms resulting from a
binary co-exposure protocol. This paper also examined the dual nature of inhibition of
each chemical in the pair to each other, which is that TCE and CHCI3 were predicted to
interact in a competitive manner. Even though as stated by Dobrev et al. (2001), other
solvents inhibit TCE metabolism, it is also possible to quantify the synergistic interaction
that TCE has on other solvents, using techniques such as gas uptake inhalation exposures.
Haddad et al. (2000) has developed a theoretical approach to predict the maximum impact
that a mixture consisting of co-exposure to dichloromethane, benzene, TCE, toluene,
PERC, ethylbenzene, m-, p-, and o-xylene, and styrene would have on venous blood
concentration due to metabolic interactions in Sprague-Dawley rats. Two sets of
experimental co-exposures were conducted. The first study evaluated the change in
venous blood concentration after a 4 hour constant inhalation exposure to the 10 chemical
mixtures. This experiment was designed to examine metabolic inhibition for this
complex mixture. The second study was designed to study the impact of possible enzyme
induction by using the same inhalation co-exposure after a 3 day pretreatment with the
same 10 chemical mixture. The resulting venous concentration measurements for TCE
02/12/09
E-350
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
from the first study were consistent with metabolic inhibition theory. The 10-chemical
mixture was the most complex co-exposure used in this study. The authors stated that as
mixture complexity increased, the resulting parent compound concentration time courses
changed less, an observation which is consistent with metabolic inhibition. For the
pretreatment study, the authors found a systematic decrease in venous concentration (due
to higher metabolic clearance) for all chemicals except PERC. Overall, these studies
suggest a complex metabolic interaction between TCE and other solvents.
A PBPK model for TCE including all its metabolites and their interactions can be
considered a mixtures model where all metabolites have a common starting point in the
liver. An integrated approach taking into account TCE metabolites and their metabolic
inhibition and interactions among each other is suggested in Chiu et al. (2006b).
5. Potentially Susceptible Life Stages and Conditions That May
Alter Risk of Liver Toxicity and Cancer
As described in Sections 1.2., 3.2.2., 3.2.6, 4.2.1, 4.2.2, 4.2.3., and 4.2.4, there are a
number of conditions that are associated with increased risk of liver cancer and toxicity that
include age, use of a number of prescription medications including fibrates and statins, disease
state (e.g., diabetes, NALD, viral infections) and exposure to external environmental
contaminants that have an affect on TCE toxicity and targets. Obviously epigenetic and genetic
factors play a role in determining the risk to the individual. In terms of liver cancer, there is
general consensus that despite the associations that have been made with etiological factors and
the risk of liver cancer, the mechanism is still unknown. The MOA of TCE toxicity is also
unknown but exposure to TCE and its metabolites have shown in rodent models to induce liver
cancer and in a fashion that is not consistent with only a hypothesized MOA of PPARa receptor
activation that is in need of revision. However, multiple TCE metabolites have been shown to
also induce liver cancer with varying effects on the liver that have also been associated with early
stages of neoplasia (glycogen storage) or other actions associated with risk of
hepatocarcinogenicity. The growing epidemic of obesity has been suggested to increase the risk
of liver cancer and may reasonably increase the target population for TCE effects on the liver.
Lifestyle factors such as ethanol ingestion have not only been shown to increase liver
cancer risk in those who already have fatty liver, but also to increase the toxicity of TCE.
However, as noted by Caldwell et al (2007b), while there is evidence to suggest that TCE
exposure may increase the risk of liver toxicity and cancer, there are not data to support a
quantitative estimate of how co-exposures may modulate that risk.
These findings can also serve to alert the risk manager to the possibility that multiple
internal and external exposures to TCE that may act via differing MO As for the
production of liver effects. This information suggests a possible lack of "zero"
background exposures and can help identify potential susceptible populations.
02/12/09
E-351
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Background levels of haloacetates in drinking water may add to the cumulative exposure
an individual receives via the metabolism of TCE. The brominated haloacetates
apparently share some common effects and pathways with their chlorinated counterparts.
Thus, concurrent exposure of TCE, its metabolites, and other haloacetates may pose an
additive response as well as an additive dose. However, personal exposures are difficult
to ascertain and the effects of such co-exposures on toxicity are hard to quantify. EPA's
guidance on cumulative risk assessments directs "each office to take into account
cumulative risk issues in scoping and planning major risk assessments and to consider a
broader scope that integrates multiple sources, effects, pathways, stressors, and
populations for cumulative risk analyses in all cases for which relevant data are available"
[U.S. EPA, 1997], Widespread exposure to possible background levels of TCE
metabolites or co-contaminants and other extrinsic factors have the potential to affect
TCE toxicity. However, the available data for co-exposures on TCE toxicity appears
inadequate for quantifying these effects, particularly at environmental levels of
contamination and exposure. Thus, the risk manager and assessor are going to be limited
by not having information regarding either (1) the type of exposure data necessary to
assess the magnitude of co-exposures that may affect toxicity, or (2) the potential
quantitative impacts of these co-exposures that would enable specific adjustments to risk.
Nonetheless, the risk manager should be aware that qualitatively a case can be made that
extrinsic factors may affect TCE toxicity.
6. Uncertainty and Variability
Along with general conclusions about the coherence of data that enable conclusions about
effects on the liver shown through experimental studies of TCE, there have also been extensive
discussions throughout this report regarding the specific limitations of experimental studies
whose design was limited by small and varying groups of animals and variability in control
responses as well as reporting deficiencies. Section 3.2.5. has brought forward the uncertainty in
the MO A for liver cancer in general. The consistency of different animal models with human
HCC is described in Section 3.3., with Section 3.2.2. providing a discussion of the promise and
limitations of emerging technologies to study the MO As of liver can in general and for TCE
specifically. Issues regarding the target cell for HCC and the complexities of studying the MOA
for a heterogeneous disease are described in Sections 3.2.4. and 3.2.8., respectively. Finally, the
uncertainty regarding key events in how activation of the PPARa receptor my lead to
hepatocarcinogenesis and the problems with extrapolation of results using the common paradigm
to study them (exposure to high levels of WY-14,643 in abbreviated bioassays in knockout mice)
are outlined in Section 3.5.1. As such uncertainties are identified future research can focus on
resolving them.
02/12/09
E-352
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
7. References:
Adami HO, Chow WH, Nyrene O, Berne C, Linet MS, Ekbom A, et al. 1996. Excess risk of
primary liver cancer in patients with diabetes mellitus. J Natl Cancer Inst 88:1472-1477.
Adams LA, Lindor KD. 2007. Nonalcoholic fatty liver disease. Ann Epidemiol 17: 863-869.
Adli M, Baldwin AS. 2006. IKK-i/IKKe control constitutive, cancer cell-associated NF-kB
activity via regulation of Ser-536 p65/RelA phosphorylation. J of Biological Chemistry 281(37):
26076-26984.
Aikata H, Takaishi H, Kawakami Y, Takahashi S, Kitamoto M, Nakanish T, Nakamura Y,
Shjimanmoto F, Kajiyama G, Ide T. 2000. Telomere reduction in human liver tissues with age
and chronic inflammation. Exp Cell Res 256: 578-582.
Akira S, Takeda K, 2004. Toll-like receptor signaling. Nat Rev Immunol. 4(7): 449-511
Akiyama TE, Nicol CJ, Fievet C, Staels B, Ward JM, Auwerx J, Lee SST, Gonzalez FJ, Peters
JM. 2001. Peroxisome proliferator-activated receptor-a regulates lipid homeostasis, but is not
associated with obesity. J Biol Chem. 276(42):39088-39093.
Allen DG, Pearse G, Haseman JK, Maronpot RR. 2004. Prediction of rodent carcinogenesis: an
evalution of prechronic liver lesions as forecasters of liver tumors in NTP carcinogenicity studies.
Toxicol Pathol 32: 393-401.
Ammini CV, Fernandez-Canon J, Shroads AL, Cornett R, Cheung J, James MO, et al. 2003.
Pharmacologic or genetic ablation of maleylacetoacetate isomerase increases levels of toxic
tyrosine catabolites in rodents. Biochem Pharmacol 66:2029-2038.
Andersen ME, Gargas ML, Clewell HJ, Severyn KM. 1987. Quantitative evaluation of the
metabolic interactions between trichloroethylene and 1,1-dichloroethylene in vivo using gas
uptake methods
Anderson SP, Howroyd P, Liu J, Qian X, Bahnemann R, Swanson C, Kwak MK, Kensler TW,
Corton JC. 2004. The transcriptional response to a peroxisome proliferator-activated receptor
alpha agonist includes increased expression of proteome maintenance genes. J Biol Chem.
279:52390-8.
Anna CH, Maronpot RR, Pereira MA, Foley JF, Malarkey DE, Anderson MW. 1994. ras proto-
oncogene activation in dichloroacetic-, trichloroethylene- and tetrachloroethylene-induced liver
tumors in B6C3F1 mice. Carcinogenesis 15:2255-2261.
02/12/09
E-353
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Anti M, Marra G, Ludovico Rapaccini G, Rumi C, Bussa S, Fadda G, Vecchio FM, Valenti A,
Percesepe A, Pompili M, Armelad F, Gentiloni N. 1994. DNA ploidy pattern in human chronic
liver diseases and hepatic nodular lesions. Cancer 73(2): 281-288.
Anway MD, Cupp AS, Uzumcu M, Skinner MK. 2005. Epigenetic transgenerational actions of
endocrine disruptors and male fertility. Science. 308: 1466-1469.
Attallah AB, Tabll AA, Salem SF, El-Sadany M, Ibrahim TA, Osman S, El-Dosoky IM. 1999.
DNA ploidy of liver biopsies from patients with liver cirrhosis and hepatocellular carcinoma: a
flow cytometric analysis. Cancer letters 142: 65-69.
Austin EW, Parrish JM, Kinder DH, Bull RJ. 1996. Lipid peroxidation and formation of 8-
hydroxydeoxygaunosine from acute doses of halogenated acetic acids. Fundamental and Applied
Toxicology 31: 77-82.
Auwerx J, SchoonjansK, Fruchart JC, Staels B. 1996. Transcriptional control of triglyceride
metabolism: fibrates and fatty acids change the expression of the LPL and apo C-III genes by
activating the nuclear receptor PPAR. Atherosclerosis 124 (suppl). S29-S37.
Badger TM, Ronis MJ, Seitz HK, Albano E, Ingelman-Sundberg M, Lieber CS. 2003. Alcohol
metabolism: Role in toxicity and carcinogenesis. Alcohol Clin Exp Res 27:336-347
Bai CL, CanfieldPJ, StaceyNH. 1992a. Effects of hexachloro-1,3-butadiene and 1,1,2,2-
tetrachloroethylene on individual serum bile acids. Toxicol Ind Health 8: 191-203.
Bai CL, Canfield PJ, Stacey NH. 1992b. Individual serum bile acids as early indicators of carbon
tetrachloride- and chloroform-induced liver injury. Toxicology 75:221-234.
Bai CL, Stacey NH. 1993. Mechanism of trichloroethylene-induced elevation of individual bile
acids. II. In vitro and in vivo interference by trichloroethylene with bile acid transport in isolated
rat hepatocytes. Toxicol and Appl Pharmacol 121: 296-302.
Balkwill F, Coussens LM. 2004. An inflammatory Link. The NF-kB protein is a key player in
inflammation, It now seens that it might also activate signaling pathways, in both cancer cells
and tumour-associated inflammatory cells, that promote malignancy. Nature 431:405-406.
Balkwill F, Charles KA, Mantovani A. 2005. Smoldering and polarized inflammation in the
initiation and promotion of malignant disease. Cancer Cell 7: 211-217.
Baker VA, Harries HM, Waring JF, Duggan CM, Ni HA, Jolly RA, Yoon LW, De Souza AT,
Schmid JE, Brown RH, Ulrich RG, Rockett JC. 2004. Clofibrate-induced gene expression
changes in rat liver: a cross-laboratory analysis using membrane cDNA arrays. Environmental
Health Perspectives 112(4): 428-438.
02/12/09
E-354
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Balkwill F, Coussens LM. 2004. An inflammatory Link. The NF-kB protein is a key player in
inflammation, It now seens that it might also activate signaling pathways, in both cancer cells
and tumour-associated inflammatory cells, that promote malignancy. Nature 431:405-406.
Ballestar E, Esteller M. 2002. The impact of chromatin in human cancer: linking DNA
methylation to gene silencing. Carcinogenesis 23:1103-1109
Bannasch P. 1996. Pathogenesis of hepatocellular carcinoma: sequential cellular, molecular, and
metabolic changes. Prog Liver Dis. 16:161-97.
Bannasch P. 2001. Hormonal and hormone-like effects eliciting hepatocarcinogenesis. Folia
Histochem Cytobio 139(Suppl 2):28-29.
Bannasch P, Nehrbass D, Kopp-Schneider A. 2001 Significance of hepatic preneoplasia for
cancer chemoprevention. IARC Sci Publ 154:223-240.
Bannasch P, Haertel T, Su Q. 2003. Significance of hepatic preneoplasia in risk identification and
early detection of neoplasia. Toxicol Pathol. 31(1): 134-139.
Baraona E, Shoichet L, Navder K, Lieber CS. 2002a. Mediation by nitric oxide of the stimulatory
effects of ethanol on blood flow. Life Sci 70:2987-2995.
Baraona E, Zeballos GA, Shoichet L, Mak KM, Lieber CS. 2002b. Ethanol consumption
increases nitric oxide production in rats, and its peroxynitrite-mediated toxicity is attenuated by
polyenylphosphatidylcholine. Alcohol Clin Exp Res 26:883-889.
Barass NC, Price RJ, Lake BG, Orton TC. 1993. Comparison of the acute and chronic mitogenic
effects of the peroxisome proliferators methylclofenapate and clofibric acid in rat liver.
Carcinogenesis 14(7): 1451-1456.
Bartosiewicz M, Penn S, Buckpitt A. 2001b. Applications of gene arrays in environmental
toxicology: fingerprints of gene regulation associated with cadmium chloride, benzo(a)pyrene,
and trichloroethylene. Environ Health Perspect 109:71-74.
Beekman JM, Boess F, Hildebrand H, Kalkuhl A Suter L. 2006. Gene expression analysis of the
hepatoxicant methylpyrilene in primary rat hepatocytes. An interlaboratory study. Environ Health
Perspect 114: 92-99.
Berger J, Daxenbichler G. 2002. DNA methylation of nuclear receptor genes—possible role in
malignancy. J Steroid Biochem Mol Bio 180:1-11.
Bergeron A, Jorquera R, Tanguay RM. 2003. La tyrosinemie hereditare: unemaladie du stress du
reticulum endoplasmique?][in French] Med Sci (Paris) 19(10):976-980.
02/12/09
E-355
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Berman E, Schlicht M, Moser VC, MadPhail RC. 1995. A multidisciplinary approach to
toxicological screening: I, Sytem toxicity. J Toxicol Environ Health 45(2): 127-43.
Best DH, Coleman WB. 2007. Treatment with 2-AAF blocks the small hepatocyte-like
progenitor cell response in retrorsine-exposed rats. J Hepatol 46: 1055-1063.
Beyer RP, Fry RC, Lasarev MR, McConnachi LA, Meira LB, Palmer VS, Powel CL, Ross PK,
Bammler TK, Bradford BU, Cranson AB, Cunningham ML, Fannin RD, Higgins GM, Hurban P,
Kayton RJ, Kerr KF, Kosyk O, Lobenhofer ED, Sieber SO, Vliet PA. Weis BK, Wolfinger R,
Woods CG, Freedman JH, Linney E, Kaufman WF, Kavanagh TJ, Paules RA, Rusyn I, Samson
LD, Spencer PS, Suk W, Tennant RJ, Zarbl H, and members of the Toxicogenomics Research
Consortium. 2007. Multicenter study of acetaminophen hepatoxocity reveals the importance of
biological endpoind in genomic analysis. Toxicological Sciences 99(1): 326-337.
Bjornsson HT, Sigurdsson MI, Fallin MD, Irizarry RA, Aspelund T, Cui H, Yu W, Rongione
MA, Ekstrom TJ, Harris TB, Launer LJ, Eiriksdottir G, Leppert MF, Sapienza C, Gudnason V,
Feinberg AP. 2008. Intra-individual change over time in DNA methylation with familial
clustering. JAMA 299(24): 2877-2883
Blackburn AC, Tzeng HF, Anders MW, Board PG. 2000. Discovery of a functional
polymorphism in human glutathione transferase zeta by expressed sequence tag database analysis.
Pharmacogenetics 10:49-57.
Blackburn AC, Coggan M, Tzeng HF, Lantum H, Polekhina G, Parker MW, et al. 2001.
GSTZld: a new allele of glutathione transferase zeta and maleylacetoacetate isomerase.
Pharmacogenetics 11:671-678.
Board PG, Chelvanayagam G, Jermiin LS, TetlowN, Tzeng HF, Anders MW, et al. 2001.
Identification of novel glutathione transferases and polymorphic variants by expressed sequence
tag database analysis. Drug Metab Dispos 29(4 Pt 2):544-547.
Boerrigter ME. 2004. Mutagenecity fo the peroxisome proliferators clofibrate, Wyeth 14,643
and di-2-ethylhexl phthalate in the lacZ plasmid-based transgenic mouse mutation assay. Journal
of Carcinogenesis 3(1): 1-7
Bombail V, Moggs JG, Orphanides G. 2004. Perturbation of epigenetic status by toxicants.
Toxicology letters. 149: 51-58.
Boorman GA, Dellarco V, Dunnick JK, Chapin RE, Hunter S, Hauchman F, Gardner H, Cox M,
Sills RC. 1999. Drinking water disinfection byproducts: Review and approach to toxicity
evaluation. Environ Health Perspect 107:207-217.
02/12/09
E-356
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Borzio M, Fargion S, Borzio F, Fancanzani AL, Croce AM, Stroffolini T, Oldani S, Cotichini R,
Roncalli M. 2003. Impact of large regenerative, low grade and high grade dysplastic nodules in
hepatocellular carcinonoma development. Journal of Hepatology 39: 208-214.
Boyault S, Rickman DS, de Reynies A, Balabaud C, Rebouissou S, Jeannot E, Herault A, Saric J,
Belghiti J, Franco D, Bioulac-Sage P, Laurent-Puig P, Zucman-Rossi J. 2007. Trancriptome
classification of HCC is related to gene alterations and to new therapeutic targets. Hepatology
45(1): 42-52.
Bradford BU, Kono H, Isayama F, Kosyk O, Wheeler MD, Akiyama TE, Bleye L, Krausz KW,
Gonzalez FJ, Koop DR, Rusyn I. 2005. Cytochrome P450 CYP2E1, but not nicotinamide adenine
dinucleotide phosphate oxidase, is required for ethanol-induced oxidative DNA damage in rodent
liver. Hepatology 41:336-344.
Bralet MP, Pichard V, Ferry N. 2002. Demonstration of direct lineage between hepatocytes and
hepatocellular carcinoma in diethylnitrosamine-treated rats. Hepatology 36: 623-630.
Branda M, Wands JR. 2006. Signal transduction cascades and hepatitis B and C related
hepatocellular carcinoma. Hepatology 43(5): 892-902.
Brautbar N, Williams J II. 2002. Industrial solvents and liver toxicity: Risk assessment, risk
factors and mechanisms. Int J Hyg Environ Health 205:479-491.
Brodsky WY, Uryvaeva IV. 1977. Cell polyploidy: its relation to tissue growth and function. Int
Rev Cytol 50: 275-332.
Buben JA, O'Flaherty EJ. 1985. Delineation of the role of metabolism in the hepatotoxicity of
trichloroethylene and perchloroethylene: a dose-effect study. Toxicology and Applied
Pharmacology. 78: 105-122.
Bull RJ. 2000. Mode of action of liver tumor induction by trichloroethylene and its metabolites,
trichloroacetate and dichloroacetate. Environ Health Perspect 108 (S2): 241-259.
Bull, RJ. 2004. Trichloroethylene and liver tumors in mice. Presentation at the U.S. EPA
Symposium on New Scientific Research Related to the Health Effects of Trichlorothylene.
27
Bull RJ, Sanchez IM, Nelson MA, Larson JL, Lansing AJ. 1990. Liver tumor induction in
B6C3F1 mice by dichloroacetate and trichloroacetate. Toxicology 63:341-359.
Bull RJ, Orner GA, Cheng RS, Stillwell L, Stauber AJ, Sasser LB, Lingohr MK, Thrall BD. 2002.
Contribution of dichloroacetate and trichloroacetate to liver tumor induction in mice by
trichloroethylene. Toxicol Appl Pharmacol 182:55-65.
02/12/09
E-357
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Bull RJ, Sasser LB, Lei XC. 2004. Interactions in the tumor-promoting activity of carbon
tetrachloride, trichloroacetate, and dichloroacetate in the liver of male B6C3F1 mice. Toxicology
199:169-183.
Caldwell JC, KeshavaN. 2006. Key issues in the modes of action and effects of trichloroethylene
metabolites for liver and kidney tumorigenesis. Environ Health Perspect 114:1457-1463.
Caldwell JC, Jinot J, DeVoney D, Gift J. 2008(a). Evaluation of evidence for infection as the
mode of action for induction of rat lymphoma. Environ Mol Mutagen 49(2): 155-64.
Caldwell JC, Keshava N, Evans MV. 2008(b). Difficulty of mode of action determination for
trichloroethylene: an example of complex interactions of metabolites and other chemical
exposures. Environ Mol Mutagen 49(2): 142-54
Calvisi DF, Factor VM, Ladu S, Conner EA, Thorgeirsson SS. 2004a. Disruption of P-catenin
pathway or genomic instability define two distinct categories of liver cancer in transgenic mice.
Gastroenterology 126: 1374-1386.
Calvisi DF, Ladu S, Conner EA, Factor VM, Thorgeirson SS. 2004b. Disregulation of E-
cadherin in transgenic models of liver cancer. Laboratory Investigation 84: 1137-1147.
Calvisi DF, Ladu S, Gorden A, Farina M, Conner EA, Lee JS, Factor VM, Thorgeirsson SS.
2006. Ubiquitous activation of Ras and Jak/Stat pathways in human HCC. Gastroenterology
130: 1117-1128.
Calvisi DF, Ladu S, Gorden A, Farina M, Lee JS, Conner EA, Schroeder I, Factor VM,
Thorgeirsson SS. 2007. Mechanistic and prognostic significance of aberrant methylation in the
molecular pathogenesis of human hepatocellular carcinoma. J Clin Invest 117(9): 2713-2711.
Carter JH, Carter HW, DeAngelo AB. 1995. Biochemical pathologic and morphometric
alterations induced in male B6C3F1 mouse liver by short-term exposure to dichloroacetic acid.
Toxicology Letters 81: 55-71.
Carter JH, Carter HW, Deddens JA, Hurst BM, George MH, DeAngelo AB. 2003. A 2-year
dose-response study of lesion sequences during hepatocellular carcinogenesis in the male B6C3F1
mouse given the drinking water chemical dichloroacetic acid. Environ Health Perspect 111(1):
53-64.
Charbonneau Mm Couture J, Plaa GL. 1991. Inhalation versus oral administration of acetone:
effect of the vehicle on the potentiation of CCL4- induced liver injury. Toxicology Letters 57:
47-54.
02/12/09
E-358
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Chang LW, Daniel FB, DeAngelo AB. 1992. Analysis of DNA strand breaks induced in rodent
liver in vivo, hepatocytes in primary culture, and a human cell line by chlorinated acetic acids and
chlorinated acetaldehydes. Environ Mol Mutagen 20L 277-288.
Channel SR, Lantendresse JR, Kidney JK, Grabau JH, Lane JW, Steel-Goodwin L, Gothaus MC.
1998. A subchronic exposure to trichloroethylene causes lipid peroxidation and hepatocellular
proliferation in male B6C3F1 mouse liver. Tox Sci 43:145-154.
Charles JM and Leeming NM. 1998. Chronic dietary toxicity/ oncogenicity studies on 2,4-
dichlorophenoxybutyric acid in rodents. Toxicological Sciences 46:21-30.
Charles JM, Bond DM, Jeffries TK, Yano BL, Stott WT, Johnson KA, Cunny HC, Wilson RD,
Bust JS. 1996. Chronic dietary toxicity/oncogenicity studies on 2,4-dichlorophenoxyacetic acid
in rodents. Fundamental and applied toxicology 11:166-172.
Chen ZY, White CC, He CY, Liu YF, Eaton DL. 1995. Zonal differences in DNA synthesis
activity and cytochrome p450 gene expression in livers of Male F344 rats treated with five
nongenotoxic carcinogens. J Environ Pathol, Toxicol and Oncol 14(2) 83-99.
Chen C, Cheung ST, So S, Fan ST, Barry S, Higgins J, Lai KM, Ji J, Dudoit S, Ng IOI, van de
Rijn M, Botstein D, Brown PO. 2002. Gene expression patters in human liver cancers.
Molecular Biology of the Cell. 13: 1929-1939.
Cheung C, Akiyama TE, Ward JM, Nicol CJ, Feigenbaum L, Vinson C, Gonzalez FJ. 2004.
Diminished hepatocellular proliferation in mice humanized for the nuclear receptor peroxisome
proliferator-activated receptor a. Cancer Res 64: 3849-3854.
Chiu WA, Caldwell JC, Keshava N, Scott CS. 2006a. Key scientific issues in the health risk
assessment of trichloroethylene. Environ Health Perspect 114:1445-1449.
Chiu WA, Okino M, Lipscomb JC, Evans MV. 2006b. Issues in the pharmacokinetics of
trichloroethylene and its metabolites. Environ Health Perspect 114:1450-1456.
Cho WCS. 2007. OncomiRs: the discovery and progress of microRNAs in cancers. Molecular
Cancer 6:50 doi: 10.1186/1476-4598-6-60.
Cogger VC, Warren A, Fraser R, NgU M, McLean AJ, Le Couteur DG. 2003. Hepatic sinusoidal
pseudocapillarization with aging in the non-human primate. Exp Gerontol. 38: 1101-1107.
Collier JM, Selmin O, Johnson PD, Runyan RB. 2003. Trichloroethylene effects on gene
expression during cardiac development. Birth Defects Res A Clin Mol Teratol 67:488-495.
02/12/09
E-359
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Collin SM. 2005. Mechanisms leading to chromosomal instability. Semin. Cancer Biol. 15: 33-
42.
Columbano A and Ledda-Columbano GM. 2003. Mitogenesis by ligands of nuclear receptors: an
attractive model for the study of the molecular mechanisms implicated in liver growth. Cell
Death and Differention. 10 :S19-S21
Columbano A, Ledda-Columbano GM, Lee G, Rajalakshmi S, SarmaDSR. 1987. Instability of
mitogen-induced liver hyperplasia to support the induction of enzyme-altered islands induced by
liver carcinogens. Cancer Res. 47: 5557-5559.
Columbano A, Ledda-Columbano GM, Curto M, Ennas GM, Coni P, Sarma DSR, Pani P. 1990.
Cell proliferation and promotion of rat liver carcinogenesis: different effect of hepatic
regeneration and mitogen-induced hyperplasia on the development of enzyme altered foci.
Carcinogenesis 11: 771-776.
Columbano A, Ledda-Columbano GM, Coni, Faa G, Liguori C, Santacruz G, Pani P. 1985.
Occurence of cell death (apoptosis) during the involution of liver hyperplasia. Lab Inv. 52. 670-
675.
Coni p, Simbula G, Carcereri De Prati A, Menegazzi M, Suzyki H, Sarma D, Ledda-Columbano
GM, Columbano A. 1993. Differences in the steady-state levels of c-fos, c-jun, and c-myc
messenger RNA during mitogen-induced liver growth and compensatory regeneration.
Hepatology 17:1109-1117.
Conner EA, Lemmer ER, Omori M, Wirth PJ, Factor VM, Thorgeirsson SS. 2000. Dual
functions of E2F-1 in a transgenic mouse model of liver carcingenesis. Oncogene 19:5054-5062.
Conner EA, Lemmer ER, Sanchez A, Factor VM, Thorgeirsson SS. 2003. E2F1 blocks and c-
Myc accelerates hepatic ploidy in transgenic mouse models. Bioch Biophys Res Comm 302: 114-
120.
Contrera JF, Jacobs AC, DeGeorge JJ. 1997. Carcinogenicity testing and the evaluation of
regulatory requirements for pharmaceuticals. Reg Toxicol and Pharmacol. 25:130-145.
Cornett R, James MO, Henderson GN, Cheung J, Shroads AL, Stacpoole PW. 1999. Inhibition of
glutathione S-transferase zeta and tyrosine metabolism by dichloroacetate: a potential unifying
mechanism for its altered biotransformation and toxicity. Biochem Biophys Res Commun
262(3):752-756.
Costet P, Legendre C, More J, Edgar A, Galtier P, Pineau T. 1998. Peroxisome proliferator-
activated receptor alphaisoform deficiency leads to progressive dyslipidemia with
sexually dimorphic obesity and steatosis. J Biol Chem 273:29577-29585.
02/12/09
E-360
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Counts JL, Sarmiento JI, Harbions ML, Downing JC, McClain RM, Goodman JI. 1996.
Carcinogenesis. 17(6): 1251-1257.
Crabb DW, Pinaire J, Chou WY, Sissom S, Peters JM, Harris RA, Stewart M. 2001. Peroxisome
proliferator-activated receptors (PPAR) and the mitochondrial aldehyde dehydrogenase (ALDH2)
promoter in vitro and in vivo. Alcohol Clin Exp Res 25:945-952.
Dai X, De Souza AT, Dai H, Lewis DL, Lee CK, Spencer AG, Herweijer H, Hagstrom JE,
Linsley PS, Bassett DE, Ulrich RG, He YD. 2007. PPARa siRNA-treated expression profiles
uncover the causal sufficiency network for compound-induced liver hypertrophy. PLoS
Computational Biology 3(3): 389 -401.
Dajee M, Lazarov M, Zhang JY, Cai T, Green CL, Russell AJ, Marinkovich MP, Tao S, Lin Q,
Kubo Y, Khavari PA. 2003. NF-kappaB blockade and oncogenic Ras trigger invasive human
epidermal neoplasia. Nature. 421: 639-643.
Daniel FB, DeAngelo AB, Stober JA, Olson GR, Page NP. 1992. Hepatocarcinogenicty of
chloral hydrate, 2-chloroacetaldehyde, and dichloroacetic acid in the male B6C3F1 mouse.
Fundamental and Applied Toxicology 19: 159-168.
David RM, Moore MR, Cifone MA, Finney DC, Guest D. 1999. Chronic peroxisome
proliferation and hepatomegaly associated with the hepatocellular tumorigenesis of di(2-
ethylhexyl)phthalate and the effects of recovery. Toxicol Sci 50(2): 195-205.
DeAngelo AB, Daniel FB, McMillan I, P Wernsing, Savage RE. 1989. Species and strain
sensitivity to the induction of peroxisome proliferation by chloroacetic acids. Toxicology and
Applied Pharmacology 101: 285-298.
DeAngelo, AB; Daniel, FB; Most, BM; et al. (1996) The carcinogenicity of dichloroacetic acid in
the male Fischer 344 rat. Toxicology 114:207-221.
DeAngelo AB, Daniel FB, Most BM, Olson GR. 1997. Failure of monochloroacetic acid and
trichloroacetic acid administered in the drinking water to produce liver cancer in male F344/N
rats. Journal of Toxicology and Environmental Health 52: 425-445.
DeAngelo AB, George MH, House DE. 1999. Hepatocarcinogenicity in the male B6C3F1
mouse following a lifetime exposure to dichloroacetic acid in the drinking water: dose-response
determination and modes of action. J Toxicol Environ Health 58:485-507.
DeAngelo AB, Daniel FB, Wong DM, George MH. 2008. The induction of hepatocellular
neoplasia by trichloroacetic acid administered in the drinking water of the male B6C3F1 mouse.
J Toxicol and Environmental Health Part A 71: 1056-1068.
02/12/09
E-361
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Dees C, Travis C. 1993. The mitogenic potentiao of trichloroethylene in B6C3F1 mice.
Toxicology Letters 69:129-137.
Djouadi F, Weinheimer CJ, Saffitz JE, Pitchford C, Bastin J, Gonzalex F, Kelly DP. 1998. A
gender-related defect in lipid metabolism and glucose homeostasis in peroxisome proliferator-
activator Receptor a-deficient mice. J of Clin Invest. 102 (6): 1083-1091.
Dobrev ID, Andersen ME, Yang RS. 2001. Assessing interaction thresholds for trichloroethylene
in combination with tetrachloroethylene and 1,1,1-trichloroethane using gas uptake studies and
PBPK modeling. Arch Toxicol 75:134-144.
Dobrev ID, Andersen ME, Yang RS. 2002. In silico toxicology: Simulating interaction thresholds
for human exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-
trichloroethane. Environ Health Perspect 110:1031-1039.
Dolinoy DC, Weidman JR, Jirtle RL. 2007. Epigenetic gene regulation: linking early
developmental environment to adult disease. Reproductive Toxicology 23: 297-307.
Dorfmueller MA, Henne SP, York RG, Bornschein RL, Manson JM. 1979. Evaluation of
teratology and behavioral toxicity with inhalation exposure of maternal rats to trichloroethylene.
Toxciology 14:153-166.
Eacho PI, Lanier TL, Brodhecker CA. 1991. Hepatcellular DNA synthesis in rats given
peroxisome proliferating agents: comparison of Wy-14,643 to clofibric acid, nafenopin and
LY171883. Carcinogenesis 12(9): 1557-1561.
Eastin WC, Mennear JH, Tennant RW, Stoll RE, Branstetter DG, Bucher JR, McCullough B,
Binder RL, Spalding JW, Mahler JF. 2001. Tg.AC genetically altered mouse: assay working
group overview of available data. Toxicol Pathol 29(suppl):60-80.
Elcombe CR, Rose MS, Pratt IS. 1985. Biochemical, histological and ultrstructural changes in rat
and mouse liver following the administration of trichloroethylene: possible relevance to species
differences in hepatocarcinogenicity. Toxicology and Applied Pharmacology, 79: 365-376.
El-Hage J (2007. Center for Drug Evaluation and Research. U.S. Food and Drug Administration.
Peroxisome proliferator-activated receptor (PPAR) antagonists: preclinical and clinical cardiac
safety considerations. http://www.fda.gov/cder/present/DIA2006/El-Hage CardiacSafetv.ppt
Accessiblitv verified February 27. 2007.
El-Serag HB. 2004. Hepatocellular carcinoma: recent trends in the United States.
Gastroenterology 127 (5 Supp 1): S27-S34.
02/12/09
E-362
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
El-Serag HB, Rudolph KL. 2007. Hepatocellular carcinoma: epidemiology and molecular
carcinogenesis. Gastroenterology 132: 2557-2576.
European Food Safety Authority (EFSA). 2006. Opinion of the scientific panel AFC related to a
new long-term carcinogenicity study on aspartame. Available online May 11, 2006:
http://www.efsa.eu.int/science/afc/afc_opinions/1471_en.html
Evert M, Schneider-Stock R, Dombrowski F. 2003. Apolipoprotein A-IV mRNA overexpression
in early preneoplastic hepatic foci induced by low-number pancreatic islet transplants in
streptozotocin-diabetic rats. Pathol Res Pract 199(6):373-379.
Fang HL, Strom SC, Cai H, Falany CN, Kocarek TA, Runge-Morris M. 2005. Regulation of
human heptic hydroxysteroid sulgotranferase gene expression by the peroxisome proliferator-
activated receptor alpha transcription factor. Mol Pharmacol 67: 1257-1267.
Farazi PA, DePinho RA. 2006. Hepatocellular carcinoma pathogenesis: from genes to
environment. Nature Reviews Cancer 6: 674-687.
Fattovich G, Stroffolini T, Zagni I, Donato F. 2004. Hepatocellular carcinoma in cirrhosis:
incidence and risk factors. Gastroenterology, 127: S35-S50.
Feinberg AP, Ohlsson R, Henikoff S. 2006. The epigenetic progenitor origin of human cancer.
Nature Reviews Genetics 7: 21-33.
Feitelson MA, Sun B, Satiroglu- Tufan NL, Liu J, Pan J, Lian Z. 2002. Genetic mechanisms of
hepatocarcinogenesis. Oncogene 21: 2593-2604.
Ferreira-Gonzalez DeAngelo AB, Nasim S, Garret CT. 1995. Ras oncogene activation during
hepatocarcinogenesis in B6 C3F1 male mice by dichloroacetic and trichloroacetic acids.
Carcinogenesis 16(3): 495-500.
FIFRA SAP. 2004. A Set of Scientific Issues Being Considered by the Environmental Protection
Agency Regarding Proposed Science Policy: PPAR-a Agonist-Mediated Hepatocarcinogenesis
in Rodents and Relevance to Human Health Risk Assessment. Minutes No. 2003-05 (5 March
2004), Federal Insecticide, Fungicide, and Rodenticide Act Science Advisory Panel. Available:
http://www.epa.gov/oscpmont/sap/2003/december9/meetingminutes.pdf/ [accessed 20 September
2005],
Fitzgerald JE, Sanyer JL, Schardein JL, Lake RS, McGuire EJ, de la Inglesia FA. 1981.
Carcinogen bioassay and mutagenicity studies with the hypolipidemic agent gemfibrozil. J Natl
Cancer Inst. 67(5): 1105-16.
02/12/09
E-363
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Fox TR, Schumann AM, Watanabe PG, Yano BL, Maher BM. 1990. Mutational analysis of H-
ras oncogene in spontanesous C57B176 X C3H/He mouse liver tumors and tumors induced with
genotoxic and nongenotoxic hepatocarcinogens. Cancer Research 50: 4014-4019.
Fraga MF, Herranz M, Espada J, Ballestar E, Paz MF, Ropero S, et al. 2004. A mouse skin
multistage carcinogenesis model reflects the aberrant DNA methylation patterns of human
tumors. Cancer Res 64:5527-5534.
Fraga MF, Ballestar E, Villar-Garea A, Boix-Chornet M, Espada J, Schotta G, et al. 2005. Loss
of acetylation at Lysl6 and trimethylation at Lys20 of histone H4 is a common hallmark of
human cancer. Nat Genet 37:391-400
Froyland L, Madsen L, Vaagenes H, Totland GK, Auwerx J, Kryvi H, Staels B, Berge RK. 1997.
Mitochondrion is the principal target for nutritional and pharmacological control of triglyceride
metabolism. J Lipid Res. 38: 1851-1858.
Fujita J, Marino MW, Wada H, Jungbluth AA, Mackrell PJ, Rivadeneira DE, Stapleton PP, Daly
JM. 2001. Effect of TNF gene depletion on liver regeneration after partial hepatectomy in mice.
Surgery 129(1): 48-54.
Fukuda K, Takemoto K, Tsuruta H. 1983. Inhalation carcinogenicity of trichloroethylene in
mice and rats. Industrial Health 21: 243-254.
Gaglio PJ, Liu H, Dash S, Cheng S, Dunne B, Ratterree M, Baskin G, Blanchard J, Bohm R,
Theise ND, LaBrecque D. 2002. Liver regeneration investigated in non-human primate model
(Macaca mulatto). J Hepatol 37: 625-632.
Gandillet A, Alexandre E, Royer C, Cinqualbre J, Jaeck D, Richert L. 2003. Hepatocyte ploidy
in regenerating livers after partial hepatectomy, drug-induced necrosis, and cirrhosis. Eur Surg
Res 35:148-160.
Ge R, Wang W, Kramer PM, Yan S, tao L, Pereira MA. 2001a. Wy-14,643-induced
hypomethylation of the c-myc gene in mouse liver. Toxicol Sci 62: 28-35.
Ge R, Yang S, Kramer PM, Tao L, Pereira MA. 2001b. The effect of dichloroacetic acid and
trichloroacetic acid on DNA methylation and cell proliferation in B6C3F1 mice. J Biochem Mol
Toxicol 15:100-106.
Ge R, Tao L, Kramer PM, Cunningham ML, Pereira MA. 2002. Effect of peroxisome
proliferators on the methylation and protein level of c-myc protooncogene in B6C3F1 mice liver.
J Biochem Mol Toxicol 16(1): 41-47.
02/12/09
E-364
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Gelderblom WCA, Galendo D, Abel S, Swanevelder S, Marasas WFO, Wild CP. 2001. Cancer
initiation by fumonisin B1 in rat liver- Role of cell proliferation. Cancer Lett 169:127-137.
George MH, Moore T, Kilburn S, Olson GR, DeAngelo AB. 2000. Carcinogenicity of Chloral
hydrate administered in drinking water to the Male F344/N rat and Male B6C3F1 mouse.
Toxicologic Pathology 28(4): 610-618.
Goel SK, Rao GS, Pandya KP, Shanker R. 1992. Trichloroethylene toxicity in mice: a
biochemical, hematological and pathological assessment. Indian Journal of Expeimental Biology
30: 402-406.
Gold LS, Zeiger E. 1997. Handbook of carcinogenic potency and genotoxicity databases. Boca
Raton: CRC Press.
Goldsworthy TL, Popp JA. 1987. Chlorinated hydrocarbon-induced peroxisomal enzyme activity
in relation to species and organ carcinogenicity. Toxicol and Appl Pharmacol 88:225-233.
Gorla GR, Malhi H, Gupta S. 2001. Polyploidy associated with oxidative injury attenuates
proliferation potential of cells. J Cell Science. 114(16): 2943 -2951.
Graham DJ, Staffa JA, Shatin D, Andrade SE, Schech SD, La GL, et al. 2004. Incidence of
hospitalized rhabdomyolysis in patients treated with lipid-lowering drugs. JAMA 292:2585-2590.
Gramantieri L, Melchiorri C, Chieco P, Gaiani S, Stecca B, Casali A, Bolondi L. 1996.
Alteration of DNA ploidy and cell nuclearity in human hepatocellular carcinoma associated with
HBV infection. J Hepatol 25: 848-853.
Grasl-Kraupp B, Waldho' r T, Huber W, Schulte-Hermann R. 1993. Glutathione S-transferase
isoenzyme patters in different subtypes of enzyme-altered rat liver foci treated with the
peroxisome proliferator nafenopin or with Phenobarbital. Carcinogenesis 14:2407-2412.
Greene AK, Wiener S, Puder M, Yoshida A, Shi B, Perez-Atayde AR, Efstathiou JA, Holmgren
L, Adamis AP, Rupnick M, Folkman J, 0;Reily MS. 2003. Endothelial-directed hepatic
regeneration after partial hepatectomy. Ann Surgery 237(4): 530-535.
Greten FR, Eckmann L, Greten TF, Park JM, Li ZW, Egan LJ, Kagnoff MF, Karin M. 2004.
IKKbeta links inflammation and tumorigenesis in a mouse model of colitis-associated cancer.
Cell 118(3):285-96
Gupta S. 2000. Hepatic polyploidy and liver growth control. Seminars in Cancer Biology. 10:
161-171.
02/12/09
E-365
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Haddad S, Charest-Tardif G, Krishnan K. 2000. Physiologically based modeling of the maximal
effect of metabolic interactions on the kinetics of components of complex chemical mixtures. J
Toxicol Environ Health 61:209-223
Hamdan H, Stacey NH. 1993. Mechanisms of trichloroethylene-induced elevation of individual
serum bile acids: I. Correlation of trichloroethylene concentrations to bile acids in rat serum.
Toxicol and Appl Pharmacol 121: 291-295.
Hanahan D, Weinberg RA. 2000. The hallmarks of cancer. Cell 100: 57-70.
Hasmall SC, Roberts RA. 2000. The nongenotoxic hepatocarcinogen deithylhexylphthalate and
methylclofenebrate induce DNA synthesis preferentially in octoploid rat hepatocytes.
Toxicologic pathology 28(4): 503-509.
Haseman JK. 1984. Stastitical issues in the design, analysis and interpretation of animal
carcinogenicity studies. Environ Health Persp 58: 385-392.
Hassan MM, Frame A, Patt YZ, El-Serag HB. 2002. Rising prevalence of hepatitis C virus
infection among patients recently diagnosed with hepatocellular carcinoma in the United States. J
Clin Gastroenterol 35: 266-269
Hayden MS, Ghosh S. 2004. Signaling to NF-kappaB. Genes Dev 18: 2195-2224.
Hays T, Rusyn I, Burns AM, Kennett MJ, Ward JM, Gonzalez FJ, et al. 2005. Role of peroxisome
proliferator-activated receptor a (PPAR a) in benzafibrate-induced hepatocarcinogenesis and
cholestasis. Carcinogenesis 26(l):219-227.
Hegi ME, Fox TR, Belinsky SA, Devereux TR, Anderson MW. 1993. Analysius of activated
protooncogenes in B6C3F1 mouse liver tumors induced by ciprofibrate, a potent peroxisome
proliferator. Carcinogenesis 14(1): 145-149.
Henschler D, Romen W, Elsasser HM, Reichaert D, Eder E, Radwan Z. 1980. Carcinogenicity
study of trichloroethyleneby longterm inhalation in three animal species. Arch Toxicol 43: 237-
248 (1980).
Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, et al. 1998. Incidence and functional
consequences of hMLHl promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci
USA 95(12):6870-6785.
Herren-Freund SL, Pereiira MA, Khoury MD, Olson G. 1987. The carcinogenicity of
trichloroethylene and its metabolites trichloroacetic acid and dichloroacetic adid, in mouse liver.
Toxicol Appl Pharmacol. 90: 183-189.
02/12/09
E-366
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Herrera E, Samper E, Martin-Caballero M, Flores JM, Lee HW, Blasco MA. 1999. Disease states
associated with telomerase deficiency appear ealier in mice with short-telomeres. The EMBO
Journal 18(11): 2050-2960.
Hoivik DJ, Quails CW, Mirabile RC, Cariello NF, Kimbrough CL, Colton HM, Anderson SP,
Santostefano MJ, Morgan RJO, Dahl RR, Brown AR, Zhao Z, Mudd PN, Oliver WB, Brown HR,
and Miller RT. 2004. Fibrates induce hepatic peroxisome and mitochondrial proliferation
without evidence of cellular proliferation and oxidative stress in cynomolgus monkeys.
Carcinogenesis 25(9): 1757-1769.
Hortelano S, Dewez B, Genaro AM, Diaz-Guerra MJ, Bosca' L. 1995. Nitric oxide is released in
regenerating liver after parial hepatectomy. Hepatology 21(3): 776-786.
Hsia CY, Huo TI, Chiang SY, Lu MF, Sun CL, Wu JC, Lee PC, Chi CW, Lui WY, Lee SD.
2007. Evaluationo of interleukin-6 , interleukin-10 and human hepatocyte growth factor as tumor
markers for hepatocellular carcinoma. Eur J Surg Oncol 33(2): 208-12.
Howroyd P, Swanson C, Dunn C, Cattley RC, Corton JC. 2004. Decreased longevity and
enhancement of age-dependent lesions in mice lacking the nuclear receptor peroxisome
proliferator-activated receptor a (PPARa). Toxicologic Pathology 32:591-599.
Hsia CY, Huo TL, Chiang SY, Lu MF, Sun CL, Wu JC, Lee PC, Chi CW, Lui WY, Lee SD.
2007. Evaluation of interleukin-6, interleukin-10, and human hepatocyte growth factor as tumor
markers for hepatocellular carcinoma. EJSO 33: 208-212.
IARC 2000. Di(2-ethylhexyl) phthalate. In: IARC monographs on the evaluation of carcinogenic
risk of chemicals to humans Vol 77 Some industrial chemicals. Lyon, France: International
Agency for Research on Cancer, 148. Available online at
http://monographs.iarc.fr/ENG/Monographs/vol 177.pdf-141.
Isaacs KK, Evans MV, Harris TR. 2004. Visualization-based analysis for
a mixed-inhibition binary PBPK model: Determination of inhibition
mechanism. J Pharmacokinet Pharmacodyn 31:215-242.
Ito Y, Yamanoshita O, Asaeda N, Tagawa Y, Lee CH, Aoyama T, Ishihara G, Furuhashi K,
Kamijima M, Gonzalez FJ, Nakajima T. 2007. Di(2-ethylhexyl)phthalate induced hepatic
tumorigenesis through a peroxisome proliferator-activated receptor a-independent pathway. J
Occup Health 49:172-182.
Jia HL, Ye QH, Qin LX, Budhu A, Forgues M, Chen Y, Liu YK, Sun HC, Wang L, Lu HZ, Shen
F, Tang ZH, Wang XW. 2007. Gene expression profiling reveals potential biomarkers of human
hepatocellular carcinoma. Clin Cancer Res 13(4): 1133-1139.
02/12/09
E-367
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Jirtle RL, Meyer SA, brokenbrough JS. 1991. Liver tumor promoter Phenobarbital: a biphasic
modulator of hepatocyte proliferation. Prog Clin Biolo Res 369: 209-216.
Jorquera R, Tanguay RM. 2001. Fumarylacetoacetate, the metabolite accumulating in hereditary
tyrosinemia, activates the ERK pathway and induces mitotic abnormalities and genomic
instability. Hum Mol Genet 10:1741-1752.
Kamijo Y, Hora K, Nakajima T, Kono K, Takahashi K, Ito Y, Higuchi M, Kiyosawa K,
Shigematsu H, Gonzalez Fj, Aoyama T. 2007. Peroxisome proliferator-activated receptor a by
long-term exposure to the plasticizer Di-(2-ethylhexyl)phthalate. Journal of American Society of
Nephrology. 18(1): 176-88.
Karpinets TV, Foy BD. 2005. Tumorigenesis: the adaptation of mammalian cells to sustained
stress environment by epigenetic alterations and succeeding mateched mutations. Carcinogenesis
26(8): 1323-1334.
Kato-Weinstein J, Stauber AJ, Orner GA., Thrall BD, Bull RJ. 2001. Differential effects of
dihalogenated and trihalogenated acetates in the liver of B6C3 F1 mice. Journal of Applied
Toxicology 21: 81-89.
Kawamoto T, Hobara T, Nakamura K, Imamura A, Ogino K, Kobayashi H, Iwamoto S, Sakai T.
1988. Induction of cytochrome P450, cytochrome b-5, NADPH cytochromo c reductase and
change in cytochrome P-450 isozymes with long-term trichloroethylene treatment. Toxicology
53: 239-249.
Keech A. Simes RJ, Barter P, et al 2005. FIELD study investigators. Effects of long-term
fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the
FIELD study): randomized controlled trial [published correction appears in Lancet,
2006:368:1420] Lancet 366:1849-1861.
Keech A, Simes J, Barter P, Best J, Scott R, Taskinen MR. 2006. Correction to the FIELD study
report. Lancet 368(9545): 1415.
Keshava N and Caldwell JC. 2006. Key Issues in the Role of Peroxisome Proliferator-Activated
Receptor Agonism and Cell Signaling in Trichloroethylene Toxicity. Environmental Health
Perspectives. 114 (9): 1464-1470.
Keshava N, Chiu W, Caldwell JC. 2007. Keshava and Caldwell respond. Environmental Health
Perspectives. 115(1):A4-A5.
Kim HJ, Odend'hal S, Bruckner JV. 1990. Effect of oral dosing vehicles on the acute
hepatoxicity of carbon tetrachloride in rats. Toxicol Appl Pharmacol 102: 34-49.
02/12/09
E-368
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Kim SZ, Kupke KG, Ierardi-Curto L, Holme E, Greter J, Tanguay RM, et al. 2000.
Hepatocellular carcinoma despite long-term survival in chronic tyrosinaemia I. J Inherit Metab
Dis 23:791-804.
Kitada T, Seki S, Kawakita N, Kuroki T, Monna T. 1995. Telomere shortening in chronic liver
diseases. Biochem Biophys Res Commun 211: 33-39.
Kjellstrand P, Lanke J, Bjerkemo M, Zetterqvist L, Mansoon L. 1980. Irreversible effects of
trichloroethylene exposure on the central nervous system. Scand J Work Environ Health. 6: 40-
47.
Kjellstrand P, Kanje M, Mansson L, Bjerkemo M. Mortensen I, Lanke J, Holmquist B. 1981.
Toxicology 21: 105-115.
Kjellstrand P, Holmquist B, Mandal N, Bjerkemo M. 1983 (a). Effects of continuous
trichloroethylene inhalation on different strains of mice. Acta Pharmacol et Toxicol 53: 369-374.
Kjellstrand P, Holmquist B, Aim P, Kanje M, Romare S, Jonsson I, MannsonL, Bjerkemo M.
1983 (b). Trichloroethylene: further studies of the effects on body and organ weights and plasma
butyrylcholinesterast activity in mice. Acta Pharmacol et Toxicol 53: 375-384.
Klaunig JE, Babich MA, Baetcke KP, Cook JC, Corton JC, David RM, et al. 2003. PPARa
agonist-induced rodent tumors: modes of action and human relevance. Crit Rev Toxicol 33:655-
780.
Kobayashi M, Ikeda K, Hosaka T, Sezaki H, Someya T, Akuta N, Suzuki F, Suzuki Y, Saitoh S,
Arase Y, Kumada H. 2006. Dysplastic nodules frequently develop into hepatocellular carcinoma
in patients with chronic viral hepatitis and Cirrhosis. Cancer 106:636-647.
Kraupp-Grasl B, Huber W, Putz Bm Gerbracht U, Schulte-Herman R. 1990. Tumor promotion
by peroxisome proliferator nafenopin involving specific subtype of altered foci in rat liver.
Cancer Research 50:37-1-3706.
Kudryavtsev BN, Kudryavtseva MV, Sakuta GA, Stein GI. 1993. Human hepatocyte
polyploidization kinetics in the course of life cycle. Virchows Arch B Cell Pathol Incl Mol Pathol
64 (6):387-393.
Kumar P, Prasad AK, Maji BK, Mani U, Dutta KK. 2001. Hepatotoxic alterations induced by
inhalation of trichloroethylene (TCE) in rats. Biomedical and Environmental Sciences 14 : 325-
332.
02/12/09
E-369
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Lake BG, Evans JG, Cunninghame ME, Price RJ. 1993. Comparison of the hepatic effects of
nafenopin and WY-14,643 on peroxisome proliferation and cell replication in the rat and Syrian
hamster. Environ Health Perspect 101:241-248.
Lalwani ND, Dethloff LA, Hasking JR, Robertson DG, DE LA Iglesia FA. 1997. Increased
nuclear ploidy, not cell proliferation, is sustained in the peroxisome proliferated-treated rat liver.
Toxicologic Pathology 25(2)165-176.
Lantendresse JR, PereiraMA. 1997. Dissimilar characteristics of N-methyl-N-nitrosurea-initiated
foci and tumors promoted by dichloroacetic acid or trichloroacetic acid in the liver of female
B6C3F1 mice. Toxicologic Pathology 25(5): 433-440.
Lapis K, Zalatnai A, Timar F, Thorgeirsson UP. 1995. Quantitative evaluation of lysozome- and
CD68 -positive Kupffer cells in diethylnitrosoamine-induced hepatocellular carcinomas in
monkeys. Carcinogenesis 16(12): 3083-3085.
Larson JL, Bull RJ. 1992. Metabolism of lipoperoxidative activity of trichloroacetate and
dichloroacetate in rats and mice. Toxicology and Applied Pharmacology 115L 268-277.
Laughter AR, Dunn CS, Swanson CL, Howroyd P, Cattley RC, Corton JC. 2004. Role of the
peroxisome proliferator-activated receptor a (PPARa) in responses to trichloroethylene and
metabolites, trichloroacetate and dichloroacetate in mouse liver. Toxicology 203:83-98.
LaVecchia C, Negri E, Franceschi S, D'Avanzo B, Boyle P. 1994. A case-control study of
diabetes mellitus and cancer risk. Br J Cancer 70:950-953.
Lawrence T, Hagemann T, Balkwill F. 2007. Sex, cytokines, and cancer. Science 317: 51-52.
Leakey JEA, Seng JE, Lantendresse JR, Hussain N, Allen LJ, Allaben WT. 2003a. Dietary
controlled carcinogenicity study of chloral hydrate in male B6C3F1 mice. Toxicology and
Applied Pharmacology 193: 266-280.
Leakey JEA, Seng JE, Allaben WT. 2003b. Body weight considerations in the B6C3F1 mouse
and the use of dietary control to standardize background tumor incidence in chronic bioassays.
Toxicology and Applied Pharmacology 193: 237-265.
Ledda-Columbano GM, Columbano A, Curto M, Ennas MG, Coni P, Sarma DSR, Pani P. 1989.
Further evidence that mitogen-induced cell proliferation does not support the formation of
enzyme-altered islands in rat liver by carcinogens. Carcinogenesis 10: 847-850.
Ledda-Columbano GM, Coni P, Simbula G, Zeddam and Columbano A 1993. Compensatory
regeneration, mitogen induced liver growth, and multistage chemical carcinogenesis.
02/12/09
E-370
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Ledda-Columbano GM, Curto M, Piga R, Zedda AI, Menegazzi M, Sartori C, Shinozuka H,
Bluethmann H, Poli V, Ciliberto G, Columbano A. 1998. In vivo hepatocyte proliferation is
inducible through a TNF, IL-6-independent pathway. Oncogene 17:1039-1044.
Ledda-Columbano GM, Perra A, Concas D, Cossu C, Molotzu F, Sartori C, Shinozuka H,
Columbano A. 2003. Different effects of the liver mitogens triiodo-thyronine and ciprofibrate on
the development of rat hepatocellular carcinoma. Toxicol Pathol 31:113-120.
Lee SST, Pineau T, Drago J, Lee EJ, Owens JW, Kroetz DL, Fernandez-Salguero PM, Westphal
H, Gonzalez FJ. 1995. Targeted disruption of the a isoform of the peroxisome proliferator-
activated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome
proliferators. Molecular and Cellular Biology. 15(6):3012-3022.
Lee S, Lee HJ, Kim JH, Lee HS, Jang JJ, Kang GH. 2003. Aberrant CpG island
hypermethylation along multistep hepatocarcinogenesis. Am J Pathol 163: 1371-1378/
Lee JS, Chu IS, Heo J, Calvisi DF, Sun Z, Roskams T, Durnez A, Demetrius AJ, Thorgeirsson
SS. 2004(a). Classification of prediction of survival in hepatocellular carcinoma by gene
expression profiling. Hepatology 40(3): 667-676.
Lee JS, Chu IS, Mikaelyan A, Calvisi DF, Heo J, Reddy JK, Thorgeirsson SS. 2004(b).
Application of comparative functional genomics to identify best-fit mouse models to study human
cancer. Nature Genetics.36(12): 1306-1311.
Librecht L. 2006. Hepatic progenitor cells in human liver tumor development. World J
Gastroenterol 12 (39: 6261-6265.
Leuschner J, Beuscher N. 1998. Studies on the mutagenic and carcinogenic potential of chloral
hydrate. Arzneim-Forsch./Drug Res. 48(11): 961-968.
Lewitt MS, Brismar K, Wang J, Wivall-Helleryd IL, Sindelar P,Gonzalez FJ, et al. 2001.
Responses of insulin-like growth factor (IGF)-I and IGF-binding proteins to nutritional status in
peroxisome proliferator-activated receptor-a knockout mice. Growth Horm IGF Res 11:303—313.
Lin YZ, Brunt EM, Bowling W, Hafenrichter DG, Kennedy SC, Flye MW, Ponder KP. 1995.
Ras-transduced diethylnitrosamine-treated hepatocytes develop into cancers of mixed phenotype
in vivo. Cancer Res 55: 5242-5250.
Lin LW, Yin XY, He YM, Gao HD, Lin XD. 2003. Biological characteristics of HCC by
ultrasound-guided aspiration biopsy and it clinical application. World J Gastroenterol 9(5): 941-
945.
02/12/09
E-371
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Lingohr MK, Bull RJ, Kato-Weinstein J, Thrall BD. 2002. Dichloroacetate stimulates glycogen
accumulation in primary hepatocytes through an insulin-independent mechanism. Toxicol Sci
68(2):508-515.
Lingohr MK, Thrall BD, Bull RJ. 2001. Effects of dichloroacetate (DCA) on serum insulin levels
and insulin-controlled signaling proteins in livers of male B6C3F1 mice. Toxicol Sci 59(1): 178-
184.
Lorimore SA, Coates PJ, Wright EG. 2003. Radiation-induced genomic instability and bystander
effects: inter-related nontargeted effects of exposure to ionizing radiation. Oncogene 22: 7058-
7069.
Luedde T, Beraza N, Kotsikoris V, van Loo G, Nenci A, De Vos R, Roskams T, Trautwein C,
Pasparakis M. 2007. Deletion of NEMO/IKKgamma in liver parenchymal cells causes
steatohepatitis and hepatocellular carcinoma. Cancer Cell 11: 119-132.
Lundgren B, DePeirre JW. 1989. Proliferation of peroxisomes and induction of cytosolic and
microsomal epoxide hydrolase in different strains of mice and rats after dietary treatment with
clofibrate. Xenobiotica 19: 876-881.
Lundgren B, Meijer J, DePeirre JW. 1987. Induction of cytosolic and microsomal epoxide
hydrolases and proliferation of peroxisomes and mitochondria in mouse liver after dietary
exposure to p-chlorophenoxyacetic acid, 2,4-Dichlorophenoxyacetic acid and 2,4,5-
trichlorophenoxy acetic acid. Biochemical Pharmacology 36(6):815-821.
Macdonald N, Holden RR, Roberts RA. 1999. Additional peroxisome proliferator-activated
receptor alpha to guinea pig hepatocytes confers increased responsiveness to peroxisome
proliferators. Cancer Res 59: 4776-4780.
Maeda S, Kamata H, Luo JL, Leffert H, Karin M. 2005. IKKp couples hepatocyte death to
cytokine-driven compensatory proliferation that promotes chemical hepatocarcinogenesis. Cell.
121:977-990.
Malarkey DE, Johnson K, Ryan L, Boorman G, Maronpot RR. 2005. New insights into
functional aspects of liver morphology. Toxicol Pathol 33: 27-34.
Maltoni C, Lefemine G, Cotti G. 1986. Experimental research on trichloroethylene
carcinogenesis. Archive of Research on Industrial Carcinogenesis Vol 5 Princton Scientific
Publishing. Maltoni and Mehlman editors. Pages 45- 153.
Maltoni C, Lefemine G, Cotti G, Perino G. 1988. Long-term carcinogenicity bioassays on
trichloroethylene administered by inhalation to Sprague-Dawley rats and swiss and B6C3F1 mice.
Annals of the New York Academy of Sciences 534: 316-343.
02/12/09
E-372
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Maronpot RR. 2007. Biological basis of differential susceptibility to hepatocarcinogenesis
among mouse strains. Background paper prepared for the National Academy of Sciences for use
by the Standing Committee on Risk Analysis Issues & Reviews. Workshop #3: Mouse Liver
Tumors, Benefits and Constraints on Use in Human Health Risk Assessment, Qualitative and
Quantitative Aspects. Oct 14, 2007.
Maronpot RR, Fox T„ Malarkey DE, Goldsworthy TL. 1995. Mutations in the ras proto-
oncogene clues to etiology and molecular pathogenesis of mouse liver tumors. Toxicology 101:
125-156.
Maronpot RR, Flake G, Huff J. 2004. Relevance of animals carcinogenesis findings to human
cancer predictions and prevention. Toxicol Pathol 32(Suppl. 1): 40-48.
Marsman DS, Cattley RC, Conway JG, Popp JA. 1988. Relationship of hepatic peroxisome
proliferation and replicative DNA synthesis to the hepatocarcinogenicity of the peroxisome
proliferators di(2-ethylhexyl)phthalate and [4-chloro-6-(2,3-xylidino)-2-pyrimidinylthio]acetic
acid (Wy- 14,643) in rats. Cancer Res 48:6739-6744.
Marsman DS, Goldsworthy TL, Popp JA. 1992. Contrasting hepatocyte peroxisome
proliferation, lipofuscin accumulation and cell turnover for the hepatocarcinogen WY-14,643 and
clofibric acid.. Carcinogenesis 13(6): 1011-1017.
Martinez-Larranaga MR, Anadon A, Diaz MJ, Fernandez B, Sevil B, Fernandez-Cruz ML,
Fernandez MC, Martinez MA, Anton R. 1996. Induction of cytochrome P4501A1 and P4504A1
activities and peroxisomal proliferation by fumonisin Bl. Toxicol and Appl Pharmacol 141: 185-
194.
Mather GG, Exon JH, Koller LD. 1990. Subchronic 90 day toxicity of dichloroacetic and
trichloroacetic acid in rats. Toxicology 64: 71-80.
McEntee GP, Batts KA, Katzmann JA, Ilstrup DM Nargorney DM. 1991. Relationship of
nuclear DNA content to clinical and pathologic findings in patients with primary hepatic
malignancy. Surgery . 111(4): 376-379.
McLean AJ, Cogger VC, Chong GC, Warrant A, Markus AM, Dahlstrom JE, Le Couter DG.
2003. Age-related pseudocapillarization of the human liver. J Pathol 200: 112-117.
Melchiorri C, Chieco P, Zedda AI, Coni P, Ledda-Columbano GM, Columbano A. 1993. Ploidy
and nuclearity of rat hepatocytes after compensatory regeneration or mitogen-induced liver
growth. Carcinogenesis 14(9): 1825-1830.
02/12/09
E-373
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Melchiorri C, Bolondi L, Chieco P, Pagnoni M, Gramantieri L, Barbara L. 1994. Diagnostic and
prognostic value of DNA ploidy and cell nuclearity in ultrasound-guided liver biopsies. Cancer
74(6): 1713-1719.
Melnick RL, Jameson CW, Goehl TJ, Maronpot RR, Collins BJ, Greenwell A, Harrington FW,
Wilson RE, Tomaszewski KE, Agarwal DK. 1987. Toxicity of microencapsulated
trichloroethylene in Fischer 344 rats. Fundamental and Applied Toxicology 8: 432-442.
Melnick RL, Huff J, Barrett JC, Maronpot RR, Lucier G, Portier CJ. 1993. Cell proliferation and
chemical carcinogenesis: Symposium Overview. Environ Health Persp 101(Suppl 5): 3-8.
Melnick RL, Kohn MC, Portier CJ. 1996. Implications for risk assessment of suggested
nongenotoxic mechanisms of chemical carcinogenesis. Environmental Health Perspectives
104(1): 123-134.
Melnick RL. 2001. Is peroxisome proliferation an obligatory precursor step in the carcinogenicity
of di(2-ethylhexyl)phthalate (DEHP)? EnvironHealth Perspect 109(5): 437-442.
Melnick RL. 2002. The IARC evaluation of di(2-ethylhexyl)phthalate (DEHP): a flawed decision
based on an untested hypothesis. Int J Occup Environ Health 8(3): 284-286.
Melnick RL. 2003. Suppression of crucial information in the IARC evaluation of DEHP. Int J
Occup Environ Health 9(1): 84-85.
Melnick RL, Thayer KA, Bucher JR. 2007. Conflicting views on chemical carcinogenesis arising
from the design and evaluation of rodent carcinogenecity studies. Environ Health Perspectives.
Doi:10.1289/ehp.9989 Online November 7, 2007.
Menegazzi M, Carcereri-De Prati A, Suzuki H, SHinozuka H, Pibiri Mo, Piga R, Columbano A,
and Ledda-Columbano GM. 1997. Liver cell proliferation induced by nafenopin and cyproterone
acetate is not associated with increases in activation of transcription factors NF-kB and AP-1 or
with expression of tumor necrosis factor a. Hepatology. 25(3):585-591.
Merdink JL, Robison LM, Stevens DK, Hu M, Parker JC, Bull RJ. 2008. Kinetics of chloral
hydrate and its metabolites in male human volunteers. Toxicology 245: 130-140.
Merrick BA, Robinson M, Condi LW. 1989. Differing hepatotoxicity and lethality after subacute
trichloroethylene exposure in aqueous or corn oil gavage vehicles in B6C3F1 mice. Journal of
Applied Toxicology. 9(1): 15-21.
Meyer K, Lee JS, Dyck PA, Cao WQ, Rao MS, Thorgeirsson SS, Reddy JK. 2003. Moleculare
profiling of hepatocellular carcinomas developing spontaneously in acyl-CoA deficient mice:
02/12/09
E-374
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
comparison with liver tumors induced by wild-type mice by a peroxisome proliferator and a
genotoxic carcinogen. Carcinogenesis 24 (5): 975-84.
Michel C, Desdouets C, Slaoui M, Isaacs KR, Roberts RA, Boitier E. 2007. Diethylnitrosamine
inititiation does not alter clofibric acid-induced hepatocarcinogenesis in the rat. Toxicological
Sciences. 99(1): 58-69.
Miller RT, Shah RS, Cattley RC, Popp JA. 1996. The peroxisome proliferator WY-14,643 and
methylclofenapate induce hepatocellular ploidy alterations and ploidy-specific DNA synthesis in
F344 rats. Toxicology and Applied Pharmacology 138: 317-323.
Minamishima YA, NakayamaK, NakayamaKI. 2002. Recovery of liver mass without
proliferation of hepatocytes after partial hepatectomy in Skpl-deficient mice. Cancer Research
62; 995-999.
Mirsalis JC, Tyson CK, Steinmetz KL, Loh EK, Hamilton CM, Bakke JP, Spalding JW. 1989.
Measurement of unscheduled DNA synthesis and S-phase synthesis in rodent hepatocytes
following in vivo treatment: testing pf 24 compounds. Environmental and Molecular Mutagenesis
14:155-164.
Miura N Miura N, Horikawa I, Nishimoto A, Ohmura H, Ito H, Hirohashi S, Shay JW, Oshimura
M. 1997. Progressive telomere shortening and telomerase reactivation during hepatocellular
carcinogenesis. Cancer Genet Cytogenet 93(1): 56-62.
Moinzadeh P, Breuhahn K, Stutzer H, Schirmacher P. 2005. Chromosome alterations in human
hepatocellular carcinomas correlated with aetiology and histological grade-results of an
exploratocyr CGH meta-analysis. British Journal of Cancer 92:935-941.
Morimura K, Cheung C, Ward JM, Reddy JK, Gonzalez FJ. 2006. Differential susceptibility of
mice humanized for peroxisome proliferator-activated receptor {alpha} to Wy-14,643 induced
liver tumorigenesis. Carcinogenesis. 27:1074-1080.
Nakajima T, Kamino Y, Usuda N, Liang Y, Fukushima Y, Kametani K, Gonzalez FJ, Aoyama T.
2000. Sex-dependent regulation of hepatic peroxisome proliferation in mice by trichloroethylene
via peroxisome proliferator-activated receptor a (PPARa). Carcinogenesis 21(4): 677-682.
Nakajima T, Moriguchi M, Mitsumoto Y, Sekoguchi S, Nishikawa T, Takshima H, Watanabe T,
Katagishi T, Kimura H, Okanoue T, Kagawa K. 2004. Centrosome aberration accompanied with
p53 mutation can induce genetic instability in hepatocellular carcinoma. Modern Pathology 17:
722-727.
02/12/09
E-375
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
NAS (National Academy of Sciences). 2006. Assessing the Human Health Risks of
Trichloroethylene: Key Scientific Issues. BEST-K-03-06-A. Available: http://www8.
nationalacademies.org/cp/projectview.aspx?key=127 [accessed 14 March 2006],
National Cancer Institute (NCI) Technical Report No. 2. Carcinogenesis bioassay of
trichloroethylene. (Cas no. 79-01-6). NCI-CG-TR-2. February 1976.
National Toxicology Program (NTP) Technical Report No. 273. Toxicology and Carcinogenesis
studies of trichloroethylene (Cas No. 79-01-6) in four strains of rats (ACI, August, Mashal,
Osborne-Mendel) (gavage studies). April 1988.
National Toxicology Program (NTP) Technical Report No. 243. Carcinogenesis studies of
trichloroethylene (without epichlorhydrin) (Cas No. 79-01-6) in F344/N Rats and B6C3F1 mice
(gavage studies). May 1990.
Naugler WE, Sakurai T, Kim S, Maeda S, Kim K, Elsharkawy AM, Karin M. 2007. Gender
disparity in liver cancer due to sex differences in MyD88-dependent IL-6 production. Science
317: 121-124.
Neghab M, Stacey NH. 1997. Toluene-induced elevation of serum bile acids: relationship to bile
acid transport. J Toxicol and Environ Health. 52: 249-268.
Neghab M, Qu Sm, Bai CL, Caples J, Stacey NH. 1997. Raised concentration of serum bile acids
following occupational exposure to halogenated solvents, l,l,2-trichloro-l,2,2-trifluoroethane and
trichloroethylene. Int Arch Occup Environ Health 70: 187-194.
Nelson MA, Bull RJ. 1988. Induction of strand breakes in DNA by trichloroethylene and
metabolites in rat and mouse liver in vivo. Toxicology and Applied Pharmacol 94: 45-54.
Nelson MA, Lansing AJ, Sanches AM, Bull RJ. 1989. Dichloroacetic acid and trichloroacetic
acid-induced DNA strand breaks are independent of peroxisome proliferation. Toxicology
58:239-248.
Nelson MA, Sanchez, Bull RJ, Sylvester SR. 1990. Increased expression of c-myc and c-H-ras
in dichloroacetate and trichloroacetate-induced liver tumors in B6C3F1 mice. Toxicology 64: 47-
57.
Nickoloff BJ, Ben-Neriah, Pikarsky E. 2005. Inflammation and cancer: is the link as simple as
we think? J Invest Dermatol 124(6): x-xiv.
National Institute of Environmental Health Science. 1998. Tumor incidence in control animals
by route and vehicle of administration, B6C3F1 mice, Research Triangle ParkNC, NIEHS
Toxicology Data Management Systems.
02/12/09
E-376
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Nissen SE, Nicholls SJ, Wolski K, Howey DC, McErlean E, Want MD, Gomez EV, Russo JM.
2007. Effects of a potent and selective PPAR-a agonist in patients with atherogenic dyslipidemia
or hypercholesterolemia: two randomized controlled trials. JAMA 297 (12): 1362-1373.
Nohammer C, Brunner F, Wolkart G, Staber PB, Steyrer E, Gonzalez FJ, Zechner R, Hoefler G.
2003. Myocardial dysfunction and male mortality in peroxisome proliferator-activated receptor
alpha knockout mice overexpressing lipoprotein lipase in muscle. Lab Invest. 83:259-69.
Nunes J, Ehrich M, Robertson J. 2001. Toxicosis associated with dual oral exposure of rats to
lead and trichloroethylene. Toxicologic Pathology. 29(4): 451-457.
Okino T, Nakajima T, Nakano M. 1991. Morphological and biochemical analyses of
trichloroethylene hepatoxicity: difference in ethanol- and Phenobarbital-pretreated rats.
Toxicology and Applied Pharmacology 108: 379-389.
Olivier P, Plancke MO, Marzin D, Clavey V, Sauzieres J, Fruchart JC. 1988. Effects of
fenofibrate, gemfibrozil, and nicotinic acid on plasma lipoprotein levels in normal and
hyperlipidemic mice: A proposed model for drug screening. Athersclerosis. 70(1-2): 107-114.
Oneta CM, Lieber CS, Li J, Ruttimann S, Schmid B, Lattmann J, Rosman AS, Seitz HK. 2002.
Dynamics of cytochrome P4502E1 activity in man: Induction by ethanol and disappearance
during withdrawal phase. J Hepatol 36:47-52.
Parrish JM, Austin EW, Stevens DK, Kinder DH, Bull RJ. 1996. Haloacetate-induced oxidative
damage to DNA in the liver of male B6C3F1 mice. Toxicology 110:103-111.
Parnel MJ, Koller LD, Exon JH, Arnzen JM. 1986. Trichloroacetic acid effects on rat liver
peroxisome and enzyme-altered foci. Environmental Health Perspectives 69: 73-79.
Pereira MA. 1996. Carcinogenic activity of dichloroacetic acide and trichloroacetic acid in the
liver of female B6C3F1 mice. Fundamental and Applied Toxicology 31: 192-199.
Pereira MA, Phelps BJ. 1996. Promotion by dichloroacetic acid and trichloroacetic acid of N-
methyl-N-nitrosurea-initiated cancer in the liver of female B6C3F1 mice. Cancer Letters 102:
133-141.
Pereira MA, Li K, Kramer PM. 1997. Promotion by mixtures of dichloroacetic acid and and
trichloroacetic acid of N-methyl-N-nitrosourea-initiated cancer in the liver of female B6C3F1
mice. Cancer Letters 115:15-23.
Pereira MA, Kramer PM, Conran PB, Tao L. 2001. Effect of chloroform on dichloroacetic acid
and trichloroacetic acid-induced hypomethylation and expression of the c-myc gene and on their
promotion of liver and kidney tumors in mice. Carcinogenesis 22(9): 1511-1519.
02/12/09
E-377
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Pereira MA, Wang W, Kramer PM, Tao L. 2004. Prevention by methionine of dichloroacetic
acid-induced liver cancer and DNA hypomethylation in mice. Toxicological Sciences 77: 243-
248.
Perkins ND. 2004. NF-KB:tumor promoter or suppressor? Trends Cell Biol 14: 64-69.
Perkins ND, Gilmore TD. 2006. Good cop, bad cop: the different faces of NF-kB. Cell death and
Differentiation. 13:759-772.
Peters JM, Cattley RC, Gonzalez FJ. 1997. Role of PPAR alpha in the mechanism of action of the
nongenotoxic carcinogen and peroxisome proliferator Wy-14,643. Carcinogenesis. 18:2029-33.
Pikarsky E, Porat RM, Stein I, Abramovitch R, Amit S, Kasem s. Gutkovich-Pyest E, Urieli-
Shoval S, Galun E, Ven-Neriah Y. 2004. NF-kB functions as a tumour promoter in
inflammation-associated cancer. Nature 431: 461-466.
Pogribny IP, Tryndyak VP, Woods CG, Witt SE, Rusyn I. 2007. Epigenetic effects of the
continuous exposure to peroxisome proliferator WY-14,643 in mouse liver are dependent on
peroxisome proliferator activated receptor a. Mutation Research 625:62-71.
Popper H. 1986. Aging and the liver. In: Popper H, Schaffner F, Editors. Progress in liver
disease. Vol. III. Orlando (FL): Grune and Stratton: 659-683.
Rake JP, Visser G, Labrune P, Leonard JV, Ullrich K, Smit GP. 2002. Glycogen storage disease
type I: diagnosis, management, clinical course and outcome. Results of the European Study on
Glycogen Storage Disease Type I (ESGSD I). Eur J Pediatr 161(Suppl 1):S20-S34.
Rakoff-Nahoum S, Medzhitov R. 2007. Regulation of spontaneous intestinal tumorigenesis
through the adaptor protein MyD88. Science 317: 124 - 127.
Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E. 2002. Metastable epialleles in
mammals. Trends Genet 18: 348-351.
Rakyan VK, Chong S, Champ ME, Cuthbert PC, Morgan HD, Luu KV, Whitelaw E. 2003.
Transgenerational inheritance of epigenetic states at the muring AxinFu allele occurs after
maternal and paternal transmission. Proc Natl Acad Sci USA 100: 2538-2543.
Ramdhan DH, Kamijima M, Yamada N, Ito Y, Yanagiba Y, Nakamura D, Okamura A, Ichihara
G, Aoyama T, Gonzalez FJ, Nakajima T. 2008. Molecular mechanisms of trichloroethylene-
induced hepatoxicity mediated by CYP2E1. Toxicol and Appl Pharmacol 231: 300-307.
02/12/09
E-378
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Rao MS, Tatematsu M, Subbarao V, Ito N, Reddy JK. 1986. Analysis of peroxisome
proliferator-induced preneoplastic and neoplastic lesions of rat liver fro placental form of
glutathione S-transferase and y-glutamyltranspeptidase. Cancer Research 46: 5287-5290.
Reamone -Buettner SM, Borlak J. 2007. A new paradigm in toxicology and teratology; altering
gene activity in the absence of DNA sequence variation. Reproductive Toxicology 24: 20-30.
Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, et al. 2002. DNMT1 and DNMT3b cooperate
to silence genes in human cancer cells. Nature 416(6880):552-556.
Richardson FC, Boucheron JA, Dyroff MC, Popp JA, Swenberg JA. 1986. Biochemical and
morphological studies of heterogeneous lobe responses in hepatocarcinogenesis. Carcinogenesis
7(2): 247-251.
Richmond, RE; Carter, JH; Carter, HW; et al. (1995) Immunohistochemical analysis of
dichloroacetic acid (DCA)-induced hepatocarcinogenesis in male Fischer (F344) rats. Cancer Lett
92:67-76.
Roskams T, Yang SQ, Koteish A, Durnez A, DeVos R, Huang X, Achten R, Verslype C, Diehl
AM. 2003. Oxidative stress and oval cell accumulation in mice and humans with alcoholic and
nonalcoholic fatty liver disease. Am J Pathol 164(4): 1301-1311.
Roskams TA, Theise ND, Balabaud C, Bhagat G, Bhagat PS, Bioulac-Sage P, Brunt EM,
Crawford JM, Crosby HA, Desmert V, Finegold MJ, Geller SA, Gouw ASH, Hytiroglou P,
Knisely AS, Kojiro M, Lefkowitch JH, Nakanuma Y, Olynyk JK, Park YN, Portman B, Saxena
R, Scheur PJ, Strain AJ, Thung SN, Wanless IR, West AB. 2004. Nomenclature of the finer
branches of the biliary tree: canals, ductules, and ductular regions in human livers. Special
Article. Hepatology 39(6): 1739-1745.
Rudolf KL, DePnho RA. 2001. In the liver biology and pathology (eds Arias I, Fausto N, Boyer
J, ChisartF and ShafritzD) 1001-1011 (Lippincott Williams and Wilkins, Philadelphia)
Rusyn I, Bradham CA, Cohn L, Schoonhoven R, Swenberg JA, Brenner DA, Thurman RG.
1999. Corn oil rapidly activates nuclear factor - kB in hepatic Kupffer cells by oxidant-
dependent mechanisms. Carcinogenesis 20 (11): 2095-2100.
Rusyn I, Peters JM, Cunningham ML. 2006. Modes of action and species-specific effects of di-
(2-ethylhexl)phthalate in the liver. Critical Reviews in Toxicology. 36:459-479.
Sakurai T, Maeda S, Chang L, Karin M. 2006. Loss of hepatic NR-kappa B activity enhances
chemical hepatocarcinogenesis through sustained c-Jun N-terminal kinase 1 activation. Proc Natl
Acad Sci USA 103 : 10544-10551.
02/12/09
E-379
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Sanchez IM, Bull RJ. 1990. Early induction of reparative hyperplasia in the liver of B6C3F1
mice treated with dichloroacetate and trichloroacetate. Toxicology 64: 33-46.
Sanz N, Diez-Fernandez C, Cascales M. 1996. Variations of hepatocyte antioxidant and DNA
ploidy in rats aged 2 to 8 months. Biochimica et Biophysica Acta 1315: 123-130.
Sargent LM, 1999. Nonrandom cytogenetic alterations in hepatocellular carcinoma from
transgenic mice overexpressing c-Myc and transforming growth factor-a in the liver. Am J Pathol
154: 1047-1055.
Scharf JG, Dombrowski F, Ramadori G. 2001. The IGF axis and hepatocarcinogenesis. Mol
Pathol 54:138-144.
Schmucker DL. 2001. Liver Function and Phase I Drug Metabolism in the Elderly, a paradox.
Drugs Aging 18(11): 837-851.
Schmucker DL. 2005. Age-related changes in liver structure and function: implications for
disease?. Experimental Gerontology 49: 650-659.
Schmucker DL, Mooney JS, Jones AL. 1977. Age-related changes in the hepatic endoplasmic
reticulum: a quantitative analysis. Science 197:1005-1008.
Schmucker DL, Mooney JS, Jones AL. 1978. Stereological analysis of hepatic fine structure in
the Fischer 344 rat. Influence of sublobular location and animals age. J Cell Biol 78: 319-337.
Schwartz-Arad D, Zajicek G, Bartfeld E. 1989. The streaming liver IV: DNA content of the
hepatocyte increases with its age. Liver 9: 93-99.
Scott CS, Chiu WA. 2006. Trichloroethylene cancer epidemiology: A consideration of select
issues. Environ Health Perspect 114:1471-1478.
Schulte-Hermann R, Timmermann-Trosiener I, Schuppler J. 1986. Facilitated expression of
adaptive responses to Phenobarbital in putative pre-stages of liver cancer. Carcinogenesis 7(10):
1651-1655.
Schultz IR, Merdink JL, Gonzalez-Leon A, Bull RJ. 2002. Dichloroacetate toxicokinetics and
disruption of tyrosine catabolism in B6C3F1 mice: dose-response relationships and age as a
modifying factor. Toxicology 173:229-247
Seigel JH, Kasmin FE. 1997. Biliary tract diseases in the elderly: management and outcomes. Gut
41: 433-435.
02/12/09
E-380
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Seitz CS, Lin Q, Deng H, Khavari PA. 1998. Alterations in NF-kB function in transgenic
epithelial tissue demonstrate a growth inhibitory role for NF-kB. Proc Natl Acad Sci USA 95:
2307-2312.
Seitz HK, Stickel F. 2006. Risk factors and mechanisms of hepatocarcinogenesis with special
emphasis on alcohol and oxidative stress. Biol Chem 387(4): 349-60.
Seng JE, Agrawal N, Horsley ETM, Leakey TI, Scherer EM, Xia S, Allaben WT, Leakey JEA.
2003. Toxicokinetics of chloral hydrate in ad libitum-fed, dietary controlled and calorically
resticted male B6C3F1 mice following short-term exposure. Toxicol and Appl Pharmacol 193(2):
281-292.
Shah YM, Morimura K, Yang Q, Tanabe T, Takagi M, Gonzalez FJ. 2007. Peroxisome
proliferator-activated receptor a regulates a microRNA-mediated signaling cascade responsible
for Hepatocellular proliferation. Molecular and cellular biology 27(12):4238-4247.
Shankar K, Vaidya VS, Corton JJ, Bucci TJ, Liu J, Waalkes MP, Mehendale HM. 2003.
Activation of PPAR-a in streptozotocin-induced diabetes is essential for resistance against
acetaminophen toxicity. FASEB J. 17(12):1748-50
Shearer BG, Hoekstra. 2003. Recent advances in peroxisome proliferator-activated receptor
science. Current Medicinal Chemistry 10:267-280.
Sher T, Yi HF, McBride OW, Gonzalez FJ. 1993. cDNA cloning chromosomal mapping, and
functional characterization of the human peroxisome proliferator activated receptor.
Biochemistry 32: 5598-5604.
Sigal SH, Rajvanshi P, Gorla GR, Sokhi RP, Saxena R, Gebhard DR, Reid LM, Gupta S. 1999.
Partial hepatectomy-induced polyploidy attenuates hepatocyte replication and activates cell aging
events. Am J Physiol 176 (Gastrointest. Liver Physiol. 39): G1260-G1272.
Snyder RD, Pullman J, Carter JH, Carter HW, DeAngelo AB. 1995. In vivo administration of
dichloroacetic acid suppresses spontaneous apoptosis in muring hepatocytes. Cancer Research
55: 3702-3705.
Soni MG, Mangipudy RS, Mumtax MM, Mehendale HM. 1998. Tissue repair response as a
function of dose during trichloroethylene hepatoxicity. Toxiological Sciences 42: 158 - 165.
Soni MB, Ramaiah SK, Mumtaz MM, Clewell H, Mehendale HM. 1999. Toxicant-inflicted
injury and stimulated tissue repair are opposing toxicodynamic forces in predictive toxicology.
Regulatory Toxicology and Pharmacology. 29: 165-174.
02/12/09
E-381
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Soresi M, Giannitrapani L, D'Antona F, Florena AM, La Spada E, Terranova A, Cervello M,
D'Alessandro N, Montalto G. 2006. Interleukin-6 and its soluble receptor in patients with liver
cirrhosis and hepatocellular carcinoma. World J Gastroenterol 12(16): 2563-2568.
Stanley LA, Blackburn DR, Devereaux S, Foley J, Lord PG, Maronpot RR, Orton TC, Anderson
MW. 1994. Ras mutations in methylclofenapate-induced B6C3F1 and C57BL/10J mouse liver
tumors. Carcinogenesis 15(6): 1125-1131.
Stauber AJ, Bull RJ. 1997. Differences in phenotype and cell replicative behavior of hepatic
tumors induced by dichloroacetate (DCA) and trichloroacetate (TCA). Toxicology and Applied
pharmacology 144: 235-246.
Stauber AJ, Bull RJ, Thrall BD. 1998. Dichloroacetate and trichloroacetate promote clonal
expansion of anchorage-independent heaptcytese in vivo and in vitro. Toxicology and Applied
Pharmacology 150: 287-294.
Steup DR, Hall P, McMillan DA, Sipes IG. 1993. Time course of hepatic injury and recovery
following coadministration of carbon tetrachloride and trichloroethylene in Fischer-344 rats.
Toxicol Pathol 21:327-334.
Styles JA, Wyatt I, Coutts C. 1991. Trichloroacetic acid: studies on uptake and effects on hepatic
DNA and liver growth in mouse. Carcinogenesis. 12(9): 1715-1719.
Su Q, Bannasch P. 2003. Relevance of hepatic preneoplasia for human hepatocarcinogenesis.
Toxicol Pathol. 31: 126-133.
Szyf M. 2007. The dynamic epigenome and its implications in toxicology. Toxicological
Sciences. 100(1): 7-23.
Takubo K, Kaminishi M. 2001. Diseases of the digestive tract and telomere lengths: significance
and problems of telomere measurements. Nippon Shokakibyo Gakkai Zasshi 98: 144-150.
Takubo K, Nakamura K, Izumiyama N, Furugori E, Sawabe M, Arai T, Esaki Y, Mafane K,
Kammori M, Fujiwara M, Kato M, Oshimura M, Sasajima K. 2000. Telomere shortening with
aging in human liver. J Gerontol A: Biol Med Sci 55: 533-536.
Takeda S, Kondo M, Kumada T, Koshikawa T, Ueda R, Nishio M, et al. 1996. Allelic-expression
imbalance of the insulin-like growth factor 2 gene in hepatocellular carcinoma and underlying
disease. Oncogene 2(7):1589-1592.
Tanaka K, Smithy PF, Stromberg PC, Eydelloth RS, Herold EG, Grossman SJ, Frank JD,
Hertzog PR, Soper KA, Keenan KP. 1992. Studies of early hepatocellular proliferation and
02/12/09
E-382
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
peroxisomal proliferation in Sprague-Dawley rats treated with tumorigenic doses of clofibrate.
Toxicology and Applied Pharmacology 116: 71-77.
Tanguay RM, Jorquera R, Poudrier J, St-Louis M. 1996. Tyrosine and its catabolites: from
disease to cancer. Acta Biochem Pol 43:209-216.
Tao L, Kramer PM, Ge R, Pereira MA. 1998. Effect of dichloroacetic acid and trichloroacetic
adic on DNA methylation in liver and tumors of female B6C3F1 mice. Toxicological Sciences
43: 139-144.
Tao L, Yan S, Xie M, Kramer PM, Pereira MA. 2000. Effect of trichloroethylene and its
metabolites, dichloroacetic acid and trichloroacetic acid, on the methylation and expression of c-
Jun and c-Myc protooncogenes in mouse liver: prevention by methionine. Toxicological Sciences
54: 399-407.
Tao L, Li Y, Kramer PM, Wang W, Pereira MA. 2004. Hypomethylation of DNA and the
insulin-like growth factor-II gene in dichloroacetic and trichloroacetic acid-promoted mouse liver
tumors. Toxicology 196: 127-136.
Tarsetti F, Lenzi R, Salvi R, Schuler E, Rijhsinghani K, Lenzen R, Tavoloni N. 1993. Liver
carcinogenesis associatged with feeding of ethionine in a choline-free diet: evidence against the
role of oval cells in the emergence of hepatocellular carcinoma. Hepatology 18(3): 596-603.
Thai SF, Allen JW, DeAngelo AB, George MH, Fuscoe JC. 2001. Detection of early gene
expression changes by differential display in the livers of mice exposed to dichloroacetic acid.
Carcinogenesis. 22:1317-1322.
Thai SF, Allen JW, DeAngelo AB, George MH, Fuscoe JC. 2003. Altered gene expression in
mouse livers after dichloroacetic acid exposure. Mutat Res 543:167-180.
Tharappel JC, Cunningham ML Spear BT, GlauertHP. 2001. Differential activation of hepatic
NF-kB in rats and hamsters by the peroxisome proliferators Wy-14,643, Gemfibrozil, and dibutyl
phthalate. Toxicol Sci 62: 20-27.
Thrall KD, Poet TS. 2000. Determination of biokinetic interactions in chemical mixtures using
real-time breath analysis and physiologically based pharmacokinetic modeling. J Toxicol Environ
Health 59:653-70.
Thorgeirsson SS. 2006. See "a tumor progression model for hepatocellualr carcinoma;
bioinformatic analysis of genomic data" by Poon TCW, Wong N, Lai PBS, Rattray M, Johnson
PJ, Sung JJY on pager 1262. Editorial comment. Gastroenterology 131(4): 1344- 1346.
02/12/09
E-383
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Toraason M, Clark J, Dankovic D, Mathias P, Skaggs S, Walker C, Werren D. 1999. Oxidative
stress and DNA damage in Fischer rats following acute exposure to trichloroethylene or
perchloroethylene. Toxicology 138: 43-53.
Tryndyak VP, Muskhelishvili L, Kovalchuk O, Rodriguez-Juarez R, Montgomery B, Churchwell
MI, Ross SA, Beland FA, Pogribny IP. 2006. Effect of long-term tamoxifen exposure on
genotoxic and epigenetic changes in rat liver; implications for tamoxifen-induced
hepatocarcingenesis. Carcinogenesis 27(8) 1713-1720.
Tucker AN, Sanders VM, Barnes DW, Bradshaw TJ, White KL, Sain LE, Borzelleca JF, Munson
AE. 1982. Toxicology of trichloroethylene in the mouse. Toxicology and Applied
Pharmacology 62: 351-357.
Tzeng HF, Blackburn AC, Board PG, Anders MW. 2000. Polymorphism- and species-dependent
inactivation of glutathione transferase zetaby dichloroacetate. Chem Res Toxicol 13(4):231-236.
Ulrich RG. 2003. The toxicogenomics of nuclear receptor agonists. Current Opinion in Chemical
Biology 7:505-510.
Ulrich RG, Rockett JC, Gibson GG, Pettit SD. 2004. Overview of an interlaboratory collaboration
on evaluating the effects of model hepatotoxicants on hepatic gene expression. Envrion Health
Perspect 112:423-427.
Urabe Y Nouso K, Higashi T, Nakatsukasa H, Hino N, Ashida K, Kinugasa N, Yoshida K,
Uematsu S, Tsuji T. 1996. Telomere length in human liver diseases. Liver 16: 293-297.
U.S. Environmental Protection Agency. 1997. Cumulative risk assessment guidance: Phase I
Planning and Scoping. Memorandum from Carol M. Browner, Administrator, and Fred Hansen,
Deputy Administrator, U.S. Environmental Protection Agency, Washington, DC.
U.S. EPA (Environmental Protection Agency). 2001. Trichloroethylene Health Risk Assessment:
Synthesis and Characterization. U.S. Environmental Protection Agency. Office of Research and
Development. External Review Draft. Available:
http://oaspub.epa.gov/eims/eimscomm.getfile?p_dowload_id=4580 [accessed 3/3/06],
U.S. EPA (Environmental Protection Agency). (2005) Guidelines for Carcinogen Risk
Assessment. EPA/630/P-03/001F. Washington DC: Risk Assessment Forum, U.S. Environmental
Protection Agency. Available:
http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=439797[Accessed 8/05],
Vainio H, Nickels J, Linnainmaa K. 1982. Phenoxy acid herbicides cause peroxisome
proliferation in Chinese hamsters. Scand J work environ health 8:70-73.
02/12/09
E-384
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Vainio H, Linnainmaa K, Kahonen M, Nickels J, Hietanen E, Marniemi J, Peltonen P. 1983.
Hypolipidemia and peroxisome proliferation induced by phenoxyacetic acid and herbicides in
rats. Biochemical Pharmacology 32(18): 2775-2779.
Van Duuren BL, Goldschmidt BM, Loewengart G, Smith AC, Meichionne S, Seidman I, Roth D.
1979. Carcinogenicity of halogenated olefinic and aliphatic hydrocarbons in mice. JNCI 63(6):
1433-1439.
Vickers AEM, Lucier GW. 1996. Estrogen receptor levels and occupancy in hepatic sinusoidal
endothelial and Kupffer cells are enhanced by initiation with diethylnitroamine and promotion
with 17 a-ethinylestradiol in rats. Carcinogenesis 17(6): 1235-1242.
Vielhauer V, Sarafoff M, Gais P, Rabes HM. 2001. Cell type-specific induction of 06-
alkylguanine-DNA alkyltransferase mRNA expression in rat liver during regeneration,
inflammation and preneoplasia. J Cancer Res Clin Oncol 127: 591-602.
Vig P, Russo FP, Edwards RJ, Tadrous PJ, Wright NA, Thomas HC, Alison MR, Forbes SJ.
2006. The source of parenchymal regeneration after chronic hepatocellular liver injury in mice.
Hepatology 43(2): 316-323.
Vogelstein G, Kinzler KW. 2004. Cancer genes and the pathways they control. Nature Medicine
10(8): 789-799.
Vollmar B, Pradarutti S, Richter S, Menger MD. 2002. In vivo quantification of ageing changes
in rat liver from early juvenile to senescent life. Liver 22: 330-341.
Voss C, Zerban H, Bannasch P, Berger MR. 2005. Lifelong exposure to di-(2-ethylhexyl)-
phthalated induced tumors in liver and testes of Sprague-Dawley rats. Toxicology 2006: 359-371.
Voss KA, Riley R, Dunn C, Corton JC. 2006. The role of tumor necrosis factor alpha and the
peroxisome proliferator-activated receptor alpha in modulating the effects of fumonisin in mouse
liver. Toxicology. 222:165-74.
Wada N, Marsman DS, Popp JA. 1992. Dose-related effects of the hepatocarcinogen, WY-
14,643, on peroxisomes and cell replication. Fundamental and Applied Toxicology 18: 149-154.
Wang G, Stacey NH. 1990. Elevation of individual serum bile acids on exposure to
trichloroethylene or a- naphthylisothioncyanate. Toxicol and Appl Pharmacol 105: 209-215.
Wang RS, Nakajima T, Honma T. 1999. Trichloroethylene inhibits aldehyde dehydrogenase only
for aliphatic aldehydes of short chains in rats. Toxicology 132:9-18.
02/12/09
E-385
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Wanson JC, Bernaert D, Penasse W, Mosselmans R, Bannasch P. 1980. Separation in distinct
subpopulations by Eltriation of liver cells following exposure of rats to N-Nitrosomorpholine.
Cancer Research 40: 459-471.
Wantanabe T, Shimada H, Tanaka Y. 1978. Human hepatocytes and aging: a cytophotometrical
analysis in 35 sudden-death cases. Vichows Arch B Cell Pathol. 27(4): 307-316.
Watson RE, Goodman JI. 2002. Effects of Phenobarbital on DNA methylation in GC-rich
regions of hepatic DNA from mice that exhibit different levels of susceptibility to liver
tumorigenesis. Toxicol Sci 68:51-58.
Waxman S, Wurmbach E. 2007. De-regulation of common housekeeping genes in hepatocellular
carcinoma. BMC genomics 8:243 doi:10.1186/1471-2164-8-243.
Weibel ER, Staubli W, Gnage HR, Hess F. 1964. Correlated morphometric and biochemical
studies on the liver cell. I. Morphometric model, stereological methods, and normal morphometric
data for rat liver. J Cell Biol. 21: 68-91.
Weidman JR, Dolinoy DC, Murphy SK, Jirtle RL. 2007. Cancer susceptibility: epigenetic
manifestation of environmental exposures. The Cancer Journal 13(1): 9-16.
Weinberg HS, Krasner SW, Richardson SD, Thruston AD. 2002. The occurrence of disinfection
by-products (DBPs) of health concern in drinking water. Results of a nationwide DBP occurrence
study. U.S. Environmental Protection Agency. EPA/600/R-02/068.
Weisel CP, Kim H, Haltmeier P, Klotz JB. 1999. Exposure estimates to disinfection by-products
of chlorinated drinking water. Environ Health Perspect 107:103-110.
Weiss PA. 1971. A cell is not an island entire of itself. Perspect Biol Med 14: 182-205.
Werner H, Le Roith D. 2000. New concepts in regulation and function of the insulin-like growth
factors: implications for understanding normal growth and neoplasia. Cell Mol Life Sci 57:932-
942.
Wheeler MD, Smutney OM, Check JF, Rusyn I, Schulte-Hermann R, Thurman RG. 2003.
Impaired Ras membrane association and activation in PPARa knockout mice after partial
hepatectomy. Am J Physiol Gastointest Liver Physiol. 284: G302-G312.
Wideroff L, Gridley G, Mellemkjaer L, Chow WH, Linet M, Keehn S, et al. 1997. Cancer
incidence in a population-based cohort of patients hospitalized with diabetes mellitus in Denmark.
J Natl Cancer Inst 89:1360-1365.
02/12/09
E-386
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Wilkens L, Flemming P, Gebel M, Bleck J, Terkamp C, Wingen L, Kreipe H, Schlegelberger B,
2004. Induction of aneuploidy by increasing chromosomal instability during dedifferentiation of
hepatocellular carcinoma. Proc Natl Acad Sci USA 101: 1309-1314.
Woodruff TJ, Axelrad DA, Caldwell J, Morello-Frosch R, Rosenbauam A. 1998. Public health
implications of 1990 air toxics concentrations across the United States. Environ Health Perspect
106:245-251.
Woods CG, Burns AM, Maki A, Bradford BU, Cunningham ML, Connor HD, Kadiiska MB,
Mason RP, Peters JM, Rusyn I. 2006. Sustained formation of alpha-(4-pyridyl-l-oxide)-N-tert-
butylnitrone radical adducts in mouse liver by peroxisome proliferators is dependent upon
peroxisome proliferator activated receptor-alpha, but not NADPH oxidase. Free Radic Biol Med.
42(3):335-42.
Woods CG, Vanden Heuvel JP, Rusyn I. 2007a. Genomic profiling in nuclear receptor-mediated
toxicity. Toxicologic Pathol 35:474-494.
Woods CG, Burns AM, Bradford BU, Ross PK, Kosyk O, Swenburg JA, Cunningham ML,
Rusyn I. 2007b. Wy-14,643-induced cell proliferation and oxidative stress in mouse liver are
independent of NADPH oxidase. Toxicological Sciences 98(2): 366-374.
Woods CG, Burns AM, Maki A, Bradford BU, Cunningham ML, Connor HD, Kadiiska MB,
Mason RP, Peters JM, Rusyn I. 2007c. Sustained formation of a-(4-pyridyl-l-oxide) -N-tert-
butylnitrone radical adducts in mouse liver by peroxisome proliferators is dependent upon
perosixome proliferator-activated receptor a, but not NADPH oxidase. Free Radical Biology &
Medicine 42L 335-342.
Wood LD, Parsons DW, Jones S, Lin J, Sjoblom T, Leary RJ, Shen D, Boca SM, Barber T, Ptak
J, Silliman N, Szabo S, Dezso Z, Ustyanksky V, Niolaskaya T, Nikolsky Y, Karchin R, Wilson
PA, Kaminker JS, Zhang Z, Croshaw R, Willis R, Dawson D, Shipitsin M, Wilson JKV,
Sukumar S, Polyak K, Park BH, Pethiyagoda CL, Pant PVK, Ballinger DG, Sparks AD, Hartigan
J, Smith DR, Suh E, Papadopoulos N, Buckhaults P, Markowitz SD, Parmigiani G, Kinzler KW,
Velculescu VE, Vogelstein B. 2007d. The genomic landscapes of human breast and colorectal
cancers. Science 318 : 1108-1113.
Woolhiser MR, Krieger SM, Thomas J, Hotchkiss JA. 2006. Report from the Dow Chemical
Company to ATSDR and EPA, Study ID # 031020.Trichloroethylene (TCE): immunotoxicity
potential in CD rats following a 4-week vapor inhalation exposure.
WuXZ, ChenD. 2006. Origin of hepatocellular carcinoma: role of stem cells. J Gastroenterol
Hepatol 21(7): 1093-8.
02/12/09
E-387
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
INTER-AGENCY REVIEW DRAFT—DO NOT CITE OR QUOTE
Wu C, Schaum J. 2000. Exposure assessment of trichloroethylene. Environ Health Perspect
108(Suppl 2):359-363.
Yang Q, Ito S, Gonzalez FJ. 2007a. Hepatocyte-restricted constitutieve activation of PPARi
induced hepatoproliferation but not hepatocarcinogenesis. Carcinogenesis. 28(6): 1171 - 1177.
Yang Q, Nagano T, Shah Y, Cheung C, Ito S, Gonzalez FJ. 2007b. The PPARa-humanized
mouse: a model to investigate the species differences in liver toxicity mediated by PPARa.
Toxicol Sciences. Published on line August 9, 2007.
Yeh MM, Larson AM, Campbell JS, Fausto N, Rulyak SJ, Swanson PE. 2007. The expression of
transforming growth factor -a in cirrhosis, dysplastic nodules, and hepatocellular carcinoma; an
immunohistochemical study in 70 cases. Am J Surg Pathol 31(5): 681-689.
Yeldani AV, Milano M, Subbarao V, Reddy JK, Rao MS. 1989. Evaluation of liver cell
proliferation during ciprofibrate-induced hepatocarcinogenesis. Cancer Letters 47: 21-27.
Zajicek G, Schwartz-Arad D, Bartfeld E. 1989. The streaming liver. V: Time and age-dependent
changes of hepatocyte DNA content, following partial hepatectomy. Liver 9:164-171.
Zajicek G, Schwartz-Arad D. 1990. Streaming liver VII: DNA turnover in acinus zone-3. Liver
10:137-140.
Zajicek G, Arber N, Schwartz-Arad. 1991. Streaming liver VIII: Cell production rates following
partial hepatectomy. Liver 11: 347-351.
Zeppa P, Benincasa G, Troncone G, Lucariello A, Zabatta A, Cochand-Priollet B, Fulciniti F,
Vetrani A, De Rosa G, Palombini L. 1998. Reptrospective evaluation of DNA ploidy of
hepatocarcinoma on cytological samples. Diagnostic Cytopathology 19(5): 323-329.
Zhang B, Pan Z, Cobb GP, Anderson TA. 2007. microRNAs as oncogenes and tumor
suppressors. Developmental Biology 302: 1-12.
Zhou H, Randers-Pehrson G, Geard CR, Brenner DJ, Hall EJ, Hei TK. 2003. Interaction between
radiation-induced adaptive response and bystander mutagenesis in mammalian cells. Radiat Res
160: 512-516.
02/12/09
E-388
------- |