DRAFT - DO NOT CITE OR QUOTE EPA/635/R-11/005D
www. ep a. gov/iris
rnji
43rCnrV
TOXICOLOGICAL REVIEW
OF
BIPHENYL
(CAS No. 92-52-4)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
April 2013
NOTICE
This document is a Final Agency Review/Interagency Science Discussion draft. This
information is distributed solely for the purpose of pre-dissemination peer review under
applicable information quality guidelines. It has not been formally disseminated by EPA. It
does not represent and should not be construed to represent any Agency determination or policy.
It is being circulated for review of its technical accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
11
DRAFT - DO NOT CITE OR QUOTE
-------
CONTENTS—TOXICOLOGICAL REVIEW OF BIPHENYL (CAS No. 92-52-4)
LIST OF TABLES vi
LIST OF FIGURES ix
LIST 01 ABBREVIATIONS AM) ACRONYMS x
FOREWORD xii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiii
1. INTRODUCTION 1
2. CHEMICAL AM) PHYSICAL INFORMATION 3
3. TOXICOKINETICS 5
3.1. ABSORPTION 5
3.2. DISTRIBUTION 7
3.3. METABOLISM 7
3.3.1. Identification of Metabolites 7
3.3.1.1. Results from In Vivo Animal Studies 7
3.3.1.2. Results from In Vitro Studies with Animal and Human Cells or Tissues 8
3.3.2. Metabolic Pathways 10
3.3.2.1. Description of Metabolic Scheme and Enzymes Involved 10
3.3.3. Regulation of Metabolism and Sites of Metabolism 12
3.3.3.1. Evidence for Induction of Phase I and II Enzymes 12
3.3.3.2. Demonstrated Tissue Sites of Metabolism 14
3.4. ELIMINATION 14
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS 15
4. HAZARD IDENTIFICATION 16
4.1. STUDIES IN HUMANS 16
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
AMY1AI.S ORAL AM) INHALATION 20
4.2.1. Oral Exposure 21
4.2.1.1. Sub chronic Toxicity 21
4.2.1.2. Chronic Toxicity and Carcinogenicity 23
4.2.2. Inhalation Studies 37
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION... 39
4.3.1. Oral Exposure 39
4.3.2. Inhalation Exposure 42
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES 42
4.4.1. Acute and Short-term Toxicity Data 42
4.4.2. Kidney/Urinary Tract Endpoint Studies 43
4.4.3. Biphenyl as a Tumor Promoter 47
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
ACTION 49
4.5.1. Effects on the Urinary Bladder of Rats 49
4.5.2. Genotoxicity 49
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS 50
4.6.1. Oral 56
4.6.2. Inhalation 59
4.6.3. Mode-of-Action Information 59
iii DRAFT - DO NOT CITE OR QUOTE
-------
4.7. EVALUATION 01 CARCINOGENICITY 60
4.7.1. Summary of Overall Weight of Evidence 60
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence 62
4.7.3. Mode-of-Action Information 63
4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats 63
4.7.3.2. Mode-of-Action Information for Liver Tumors in Female Mice 69
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 72
4.8.1. Possible Childhood Susceptibility 72
4.8.2. Possible Gender Differences 72
5. DOSE-RESPONSE ASSESSMENTS 74
5.1. ORAL REFERENCE DOSE (RID) 74
5.1.1. Choice of Candidate Principal Studies and Candidate Critical Effects—with
Rationale and Justification 74
5.1.2. Methods of Analysis—Including Models (e.g., PBPK, BMD) 76
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs) 84
5.1.4. Previous RfD Assessment 85
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 86
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification 86
5.2.2. Previous RfC Assessment 87
5.3. UNCERTAINTIES IN THE RID AM) RfC 87
5.4. CANCER ASSESSMENT 88
5.4.1. Choice of Study/Data—with Rationale and Justification 89
5.4.2. Dose-Response Data 89
5.4.3. Dose Adjustments and Extrapolation Method(s) 90
5.4.3.1. Liver Tumors in Female Mice 90
5.4.3.2. Bladder Tumors in Male Rats 93
5.4.4. Oral Slope Factor and Inhalation Unit Risk 93
5.4.5. Uncertainties in Cancer Risk Values 94
5.4.5.1. Oral Slope Factor 94
5.4.5.2. Inhalation Unit Risk 96
5.4.6. Previous Cancer Assessment 96
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE 97
6.1. HUMAN HAZARD POTENTIAL 97
6.1.1. Noncancer 97
6.1.2. Cancer 98
6.2. DOSE RESPONSE 99
6.2.1. Noncancer/Oral 99
6.2.2. Noncancer/Inhalation 99
6.2.3. Cancer/Oral 99
6.2.4. Cancer/Inhalation 100
7. REFERENCES 101
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AM) DISPOSITION A-1
APPENDIX B. LITERATURE SEARCH STRATEGY AND STUDY SELECTION B-l
iv DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX C. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
MODE OF ACTION C-l
C.l. EFFECTS ON THE URINARY BLADDER OF RATS C-l
C.2. EFFECTS ON THE LIVER 01 MICE C-2
C.3. ESTROGENIC EFFECTS C-3
C.4. EFFECTS ON APOPTOSIS C-4
C.5. MITOCHONDRIAL EFFECTS C-5
C.6. GENOTOXICITY C-5
APPENDIX D. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE DOSE D-l
APPENDIX E. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR E-l
v
DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF TABLES
Table 2-1. Physicochemical properties of biphenyl 4
Table 3-1. Metabolites of biphenyl identified in urine, feces, and bile of male albino rats 8
Table 4-1. Biphenyl concentrations in the air of a Finnish paper mill producing biphenyl-
impregnated fruit wrapping paper 17
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl: comparison with 60
unexposed males 18
Table 4-3. Exposure data and clinical features for five Parkinson's disease patients with
occupational exposure to biphenyl 20
Table 4-4. Incidences of urinary bladder lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 25
Table 4-5. Incidences of ureter and kidney lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 26
Table 4-6. Body and organ weight data for male and female rats administered biphenyl in the
diet for 2 years 30
Table 4-7. Dose-related changes in selected clinical chemistry values from male and female
BDFi mice exposed to biphenyl via the diet for 2 years 33
Table 4-8. Incidences of gross and histopathological findings in male and female BDFi mice fed
diets containing biphenyl for 2 years 34
Table 4-9. Incidences of selected tumor types among controls and mice administered biphenyl
orally for 18 months 37
Table 4-10. Incidences of selected histopathological lesions in tissues of CD-I mice exposed to
biphenyl vapors 7 hours/day, 5 days/week for 13 weeks 39
Table 4-11. Prenatal effects following oral administration of biphenyl to pregnant Wistar rats on
GDs 6-15 40
Table 4-12. Summary of reproductive data in albino rats exposed to dietary biphenyl 42
Table 4-13. Change in kidney weight and cellular architecture in Wistar rats exposed to biphenyl
46
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in
rats and mice 51
Table 4-15. Summary of major studies evaluating effects of biphenyl after inhalation exposure
in rats, mice and rabbits 55
Table 5-1. Datasets employed in the dose-response modeling of nonneoplastic effects in the
urinary tract of male and female F344 rats exposed to biphenyl in the diet for 2 years... 77
Table 5-2. Datasets employed in dose-response modeling of body weight, selected clinical
chemistry results, and histopathological kidney effects in male and female BDFi mice
exposed to biphenyl in the diet for 2 years 78
Table 5-3. Dataset for dose-response modeling of incidence of fetuses with missing or
unossified sternebrae, from Wistar rat dams administered biphenyl by gavage on GDs 6-
15 78
Table 5-4. Summary of candidate PODs for selected nonneoplastic effects following oral
exposure of rats and mice to biphenyl 81
vi
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-6. Incidence data for tumors in the urinary bladder of male and female F344 rats
exposed to biphenyl in the diet for 2 years 90
Table 5-7. Incidence data for liver tumors in male and female BDFi mice fed diets containing
biphenyl for 2 years 90
Table 5-8. Scaling factors for determining HEDs to use for BMD modeling of female BDFi
mouse liver tumor incidence data from Umeda et al. (2005) 91
Table 5-9. Incidence of liver adenomas or carcinomas in female BDFi mice fed diets containing
biphenyl for 2 years 92
Table 5-10. POD and oral slope factor derived from liver tumor incidence data from BDFi
female mice exposed to biphenyl in the diet for 2 years 94
Table C-l. Content of biphenyl sulphate conjugates in urine and urinary crystals from male
F344 rats treated with biphenyl and potassium bicarbonate (to elevate the pH and K+
concentration of the urine) 1
Table C-2. Comparison of the physicochemical characteristics of urinary calculi in male and
female F344 rats 2
Table C-3. Genotoxicity test results for biphenyl 7
Table C-4. Genotoxicity test results for biphenyl metabolites 12
Table D-l. BMD modeling datasets for incidences of nonneoplastic effects in the urinary tract of
male and female F344 rats exposed to biphenyl in the diet for 2 years 1
Table D-2. BMD modeling datasets for body weight, selected clinical chemistry results, and
histopathological kidney effects in male and female BDFi mice exposed to biphenyl in
the diet for 2 years 2
Table D-3. BMD modeling dataset for incidence of fetuses with missing or unossified sternebrae
from Wistar rat dams administered biphenyl by gavage on GDs 6-15 2
Table D-4. Summary of BMD modeling results for incidence of renal nodular transitional cell
hyperplasia in male F344 rats exposed to biphenyl in the diet for 2 years 3
Table D-5. Summary of BMD modeling results for incidence of renal nodular transitional cell
hyperplasia in female F344 rats exposed to biphenyl in the diet for 2 years 5
Table D-6. Summary of BMD modeling results for incidence of renal simple transitional cell
hyperplasia in male F344 rats exposed to biphenyl in the diet for 2 years 7
Table D-7. Summary of BMD modeling results for incidence of renal simple transitional cell
hyperplasia in female F344 rats exposed to biphenyl in the diet for 2 years 9
Table D-8. Summary of BMD modeling results for incidence of mineralization in renal pelvis of
male F344 rats exposed to biphenyl in the diet for 2 years 11
Table D-9. Summary of BMD modeling results for incidence of mineralization in renal pelvis of
female F344 rats exposed to biphenyl in the diet for 2 years 13
Table D-10. Summary of BMD modeling results for incidence of hemosiderin deposits in the
kidney of female F344 rats exposed to biphenyl in the diet for 2 years 15
Table D-l 1. Summary of BMD modeling results for incidence of papillary mineralization in the
kidney of male F344 rats exposed to biphenyl in the diet for 2 years 17
Table D-12. Summary of BMD modeling results for incidence of papillary mineralization in the
kidney of female F344 rats exposed to biphenyl in the diet for 2 years 19
Table D-13. Summary of BMD modeling results for incidence of combined transitional cell
hyperplasia in the bladder of male F344 rats exposed to biphenyl in the diet for 2 years 21
vii
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-14. Summary of BMD modeling results for incidence of mineralization in the kidney
(inner stripe outer medulla) of male BDFi mice exposed to biphenyl in the diet for 2
years 23
Table D-15. Summary of BMD modeling results for incidence of mineralization in the kidney
(inner stripe outer medulla) of female BDFi mice exposed to biphenyl in the diet for 2
years 25
Table D-16. BMD model results for serum LDH activity in female BDFi mice exposed to
biphenyl in the diet for 2 years 27
Table D-17. BMD modeling results for serum AST activity in female BDFi mice exposed to
biphenyl in the diet for 2 years 28
Table D-18. BMD modeling results for serum ALT activity in female BDFi mice exposed to
biphenyl in the diet for 2 years 31
Table D-19. BMD modeling results for serum AP activity in female BDFi mice exposed to
biphenyl in the diet for 2 years 32
Table D-20. BMD modeling results for changes in BUN levels (mg/dL) in male BDFi mice
exposed to biphenyl in the diet for 2 years 33
Table D-21. BMD modeling results for changes in BUN levels (mg/dL) in female BDFi mice
exposed to biphenyl in the diet for 2 years 34
Table D-22. BMD modeling results for changes in mean terminal body weight in male BDFi
mice exposed to biphenyl in the diet for 2 years 35
Table D-23. BMD modeling results for changes in mean terminal body weight in female BDFi
mice exposed to biphenyl in the diet for 2 years 36
Table D-24. Summary of BMD modeling results for fetal incidence of missing or unossified
sternebrae from Wistar rat dams administered biphenyl by gavage on GDs 6-15. (The
highest dose was not included because of maternal toxicity) 38
Table E-l. Incidences of liver adenomas or carcinomas in female BDFi mice fed diets
containing biphenyl for 2 years 1
Table E-2. Model predictions for liver tumors (adenomas or carcinomas) in female BDFi mice
exposed to biphenyl in the diet for 2 years 2
Table E-3. Incidences of urinary bladder transitional cell papilloma or carcinoma in male F344
rats fed diets containing biphenyl for 2 years 4
Table E-4. Model predictions for urinary bladder tumors (papillomas or carcinomas) in male
F344 rats exposed to biphenyl in the diet for 2 years 5
viii
DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
Figure 3-1. Schematic presentation of the metabolic pathways of biphenyl 11
Figure 5-1. Candidate PODs for selected noncancer effects in rats and mice from repeated oral
exposure to biphenyl 82
Figure B-l. Study selection strategy 3
IX
DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS
ACGIH
American Conference of Governmental Industrial Hygienists
AIC
Akaike's Information Criterion
ALT
alanine aminotransferase
AP
alkaline phosphatase
AST
aspartate aminotransferase
BBN
N-butyl -N-(4 -hy droxybutyl)nitrosamine
BMD
benchmark dose
BMDL
95% lower confidence limit on the BMD
BMR
benchmark response
BMDS
Benchmark Dose Software
BrdU
5-bromo-2-deoxyuridine
BUN
blood urea nitrogen
CA
chromosomal aberration
CASRN
Chemical Abstracts Service Registry Number
CHL
Chinese hamster lung
CHO
Chinese hamster ovary
CVSF
conduction velocity of the slowest motor fibers
CYP
cytochrome P-450
DNA
deoxyribonucleic acid
EEG
el ectroencephal ography
EHEN
N-ethyl-N-hydroxyethylnitrosamine
EMG
el ectromy ographi c
ENMG
el ectroneuromy ography
GC
gas chromatography
GD
gestation day
GOT
glutamate oxaloacetate transaminase
GPT
glutamate pyruvate transaminase
HED
human equivalent doses
HGPRT
hypoxanthine guanine phosphoribosyl transferase
HPLC
high-performance liquid chromatography
IARC
International Agency for Research on Cancer
i.p.
intraperitoneal or intraperitoneally
IRIS
Integrated Risk Information System
Kow
octanol/water partition coefficient
Km
Michaelis constant
KP
permeability coefficient
LD50
median lethal dose
LDH
lactate dehydrogenase
LOAEL
lowest-observed-adverse-effect level
MCV
motor conduction velocity
MS
mass spectrometry
NOAEL
no-ob served-adverse-effect level
NRC
National Research Council
PBPK
physiologically based pharmacokinetic
PCB
polychlorinated biphenyl
POD
point of departure
x
DRAFT - DO NOT CITE OR QUOTE
-------
PPAR
peroxisome proliferator activated receptors
RD
relative deviation
RfC
reference concentration
RfD
reference dose
ROS
reactive oxygen species
RR
relative risk
SCE
sister chromatid exchange
SD
standard deviation
SULT
sulphotransferase
TLV
threshold limit value
TMS
trimethylsilyl
TWA
time-weighted average
UDS
unscheduled DNA synthesis
UF
uncertainty factors
UGT
uridine diphosphate glucuronosyl transferase
U.S. EPA
U.S. Environmental Protection Agency
XI
DRAFT - DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to biphenyl.
It is not intended to be a comprehensive treatise on the chemical or toxicological nature of
biphenyl.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of data and related uncertainties. The discussion is intended to convey the limitations
of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
Xll
DRAFT - DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER/AUTHOR
Zheng (Jenny) Li, Ph.D., DABT
U.S. EPA, ORD/NCEA
Washington, DC
CONTRIBUTORS
James Ball, Ph.D.
U.S. EPA, ORD/NCEA
Washington, DC
Christine Cai, MS, PMP
U.S. EPA, ORD/NCEA
Washington, DC
Catherine Gibbons, Ph.D.
U.S. EPA, ORD/NCEA
Washington, DC
Karen Hogan, MS
U.S. EPA, ORD/NCEA
Washington, DC
J. Connie Kang-Sickel, Ph.D.
U.S. EPA, ORD/NCEA
Washington, DC
CONTRACTOR SUPPORT
George Holdsworth, Ph.D.
Lutz W. Weber, Ph.D., DABT
Oak Ridge Institute for Science and Education
Oak Ridge, TN
David Wohlers, Ph.D.
Joan Garey, Ph.D.
Peter McClure, Ph D, DABT
SRC, Inc.
Syracuse, NY
REVIEWERS
This document was provided for review to EPA scientists, interagency reviewers from
other federal agencies and White House offices, and the public, and peer reviewed by independent
Xlll
DRAFT - DO NOT CITE OR QUOTE
-------
scientists external to EPA. A summary and EPA's disposition of the comments received from the
independent external peer reviewers and the public is included in Appendix A.
INTERNAL EPA REVIEWERS
Jane Caldwell, Ph.D.
U.S. EPA, ORD/NCEA
Research Triangle Park, NC
Glinda Cooper, Ph.D.
U.S. EPA, ORD/NCEA
Washington, DC
Maureen Gwinn, Ph.D., DABT
U.S. EPA, ORD/NCEA
Washington, DC
Susan Makris
U.S. EPA, ORD/NCEA
Washington, DC
Margaret Pratt, Ph.D.
U.S. EPA, ORD/NCEA
Washington, DC
EXTERNAL PEER REVIEWERS
Scott M. Bartell, Ph.D.
University of California, Irvine
John M. Cullen, Ph.D., V.M.D.
North Carolina State University
Brant A. Inman, M.D., M.Sc., FRCS(C)
Duke University Medical Center
Frederick J. Miller, Ph.D., Fellow ATS
Fred J. Miller & Associates LLC
Ricardo Saban, D.V.M., Ph.D.
University of Oklahoma
Mary Alice Smith, Ph.D.
University of Georgia
Paul W. Snyder, D.V.M., Ph.D., DACVP
Purdue University
xiv
DRAFT - DO NOT CITE OR QUOTE
-------
Lauren Zeise, Ph.D.
California Environmental Protection Agency
xv
DRAFT - DO NOT CITE OR QUOTE
-------
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of biphenyl.
IRIS Summaries may include oral reference dose (RfD) and inhalation reference concentration
(RfC) values for chronic and other exposure durations, and a carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
"3
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m ) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal-of-entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, a plausible inhalation unit risk is
-3
an upper bound on the estimate of risk per (j,g/m air breathed.
Development of these hazard identification and dose-response assessments for biphenyl
has followed the general guidelines for risk assessment as set forth by the National Research
Council (NR.C. 1983). EPA Guidelines and Risk Assessment Forum Technical Panel Reports
that may have been used in the development of this assessment include the following:
Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA. 1986b). Guidelines
for Mutagenicity Risk Assessment (U.S. EPA. 1986a). Recommendations for and Documentation
of Biological Values for Use in Risk Assessment (U.S. EPA. 1988). Guidelines for
Developmental Toxicity Risk Assessment (U.S. EPA. 1991). Interim Policy for Particle Size and
Limit Concentration Issues in Inhalation Toxicity Studies (U.S. EPA. 1994a). Methods for
1
DRAFT - DO NOT CITE OR QUOTE
-------
Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA. 1994b). Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA.
1995). Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA. 1996). Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA. 1998). Science Policy Council Handbook. Risk
Characterization (U.S. EPA. 2000). Supplementary Guidance for Conducting Health Risk
Assessment of Chemical Mixtures {U.S. EPA, 2000, 4421}, A Review of the Reference Dose and
Reference Concentration Processes (U.S. EPA. 2002). Guidelines for Carcinogen Risk
Assessment (U.S. EPA. 2005a), Supplemental Guidance for Assessing Susceptibility from Early-
Life Exposure to Carcinogens (U.S. EPA. 2005b). Science Policy Council Handbook: Peer
Review (U.S. EPA. 2006b). A Framework for Assessing Health Risk of Environmental Exposures
to Children (U.S. EPA. 2006a). Recommended Use of Body Weight3/4 as the Default Method in
Derivation of the Oral Reference Dose (U.S. EPA. 2011). and Benchmark Dose Technical
Guidance Document (U.S. EPA, 2012).
This Toxicological Review is based on a review and evaluation of the primary, peer-
reviewed literature pertaining to biphenyl. The search strategy used to identify this literature,
including databases and keywords, and the results of the literature search are described in
Appendix B. References from health assessments developed by other national and international
health agencies were also examined. Other peer-reviewed information, including review articles,
literature necessary for the interpretation of biphenyl-induced health effects, and independent
analyses of the health effects data were retrieved and included in the assessment where
appropriate. EPA requested public submissions of additional information on biphenyl in
December 2007 (U.S. EPA, 2007); no submissions in response to the data call-in were received.
A comprehensive literature search was last conducted in September 2012. No major
epidemiology studies or subchronic and chronic animal studies on biphenyl were identified since
the draft Toxicological Review (dated September 2011) was released for external peer review
and public comment.
2
DRAFT - DO NOT CITE OR QUOTE
-------
2. CHEMICAL AND PHYSICAL INFORMATION
Pure biphenyl is a white or colorless crystalline solid that usually forms leaflets or scales;
commercial preparations may be yellowish or slightly tan (HSDB. 2005). Biphenyl is said to
have a pleasant odor that is variably described as peculiar, butter-like, or resembling geraniums
(HSDB. 2005; Boehncke et al.. 1999). Biphenyl melts at 69°C and has a vapor pressure of
8.93 x 10" mm Hg at 25°C, making it likely to enter the environment in its vaporized form
(HSDB. 2005). If particle-bound biphenyl is precipitated to the ground, it is likely to be
reintroduced to the atmosphere by volatilization. The water solubility of biphenyl is 7.48 mg/L
at 25°C. The logarithm of the octanol/water partition coefficient (Kow) of biphenyl of 3.98
suggests a potential for bioaccumulation (HSDB. 2005). Because it is biodegraded with an
estimated half-life of 2 and 3 days in air and water, respectively (HSDB. 2005). and is
metabolized rapidly by humans and animals (see Section 3), bioaccumulation does not occur
(Boehncke et al.. 1999). Biphenyl is ubiquitous in the environment, with reported indoor air
3 3
concentrations of 0.16-1 |ig/m and outdoor levels of approximately 0.03 |ig/m (Boehncke et
al.. 1999). The physicochemical properties of biphenyl are summarized in Table 2-1.
3
DRAFT - DO NOT CITE OR QUOTE
-------
Table 2-1. Physicochemical properties of biphenyl
Synonyms
Diphenyl, l,l'-biphenyl, 1,1'-diphenyl, bibenzene, phenylbenzene,
lemonene, Carolid AL, Phenador-X, Tetrosine LY
CASRN
92-52-4
Chemical structure
0{)
Chemical formula
C12H10
Molecular weight
154.2
Melting point
69°C
Boiling point
256°C
Specific gravity
1.041 g/cm3 at 20°C
Vapor pressure
8.93 x 10"3 mm Hg at 25°C
Log Kow
4.01; 4.11a; 4.17 or 5.27-5.46b
Water solubility
7.48 mg/L at 25°C
Henry's law constant
3.08 x 10"4 atm-m3/mol at 25°C
Conversion factors
1 ppm = 6.31 mg/m3; 1 mg/m3 = 0.159 ppm
aMonsanto (19461.
bEstimated by different methods: Dow Chemical Co. (1983).
Source: HSDB (20051.
Biphenyl exists naturally as a component of crude oil or coal tar. The current major uses
of biphenyl are as chemical synthesis intermediates (among them, the sodium salt of
2-hydroxybiphenyl, a pesticide known as Dowicide 1), as dye carriers in polyester dyeing, and as
components in heat transfer fluids (in particular Dowtherm A or Therminol® VP-1, consisting of
26.5% biphenyl and 73.5% diphenyl oxide). Biphenyl is currently not registered for use as a
pesticide in the United States, but is still used in other countries as a fungistat, most commonly to
preserve packaged citrus fruits or in plant disease control (HSDB. 2005).
Biphenyl is primarily produced by debromination/dimerization of bromobenzene, is
isolated as a byproduct of the hydrodealkylation of toluene (yield approximately 1%), or is
synthesized by catalytic dehydrocondensation of benzene. The purity of technical biphenyl
ranges from 93 to 99.9%. The prevalent impurities in technical preparations are terphenyls, a
side product from the dehydrocondensation of benzene. Biphenyl is rated as a high-volume
production chemical. Annual U.S. production in 1990 was approximately 1.6 x 104 metric tons
(HSDB. 20051
4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
3. TOXICOKINETICS
Summary. Animal studies indicate that biphenyl is rapidly and readily absorbed
following oral exposure. An in vitro study suggests that biphenyl can also be absorbed via
dermal exposure. Absorbed biphenyl is not preferentially stored in tissues and is rapidly
excreted, principally through the urine. Phase I metabolism by CYP enzymes, including
CYP1A2 and CYP3A4, in the liver converts biphenyl to a range of hydroxylated metabolites,
with 4-dihydroxybiphenyl, 4,4'-dihydroxybiphenyl and 3,4-dihydroxybiphenyl being the major
metabolites. Phase II metabolism catalyzing the conjugation of hydroxylated biphenyl
metabolites to sulphate or glucuronic acid occurs mostly in the liver, followed by the intestine
and kidney. Absorbed biphenyl is rapidly eliminated from the body, principally as conjugated
hydroxylated metabolites in the urine. The toxicokinetic properties of biphenyl are described in
more detail in the remainder of this section.
3.1. ABSORPTION
No quantitative studies on the absorption of biphenyl have been conducted in humans.
Animal studies in rats, rabbits, guinea pigs, and pigs indicate that biphenyl is rapidly and readily
absorbed following oral exposure, as evidenced by the detection of metabolites in urine and bile
(Meyer. 1977; Meyer and Scheline. 1976; Meyer etal.. 1976b; Meyer et al.. 1976a). Results
from a study with rats administered radiolabeled biphenyl indicate extensive oral absorption
(Meyer et al.. 1976a) (see below), whereas results from studies of rabbits, guinea pigs, and pigs
administered nonlabeled biphenyl indicate less extensive oral absorption in the range of 28-49%
of the administered dose (Meyer. 1977; Meyer etal.. 1976b).
Male albino rats (n = 3; body weight = 200-300 g) given an oral dose of 100 mg/kg (0.7-
1.0 (j,Ci) of [14C]-biphenyl (in soy oil) excreted 75-80% of the radioactivity in their urine within
the first 24 hours, with a total average urinary excretion of 84.8% and fecal excretion of 7.3%
during the 96-hour postdosing period (Meyer etal.. 1976a). Only a trace of [14C]-C02 was
detected in expired air and <1% of the radioactivity was recovered from tissues obtained at the
96-hour sacrifice of the rats. These results indicate that at least 85% of the administered dose
was absorbed and excreted from rats through urine or feces.
Male White Land rabbits and SffPIR guinea pigs were given biphenyl (100 mg/kg) by
gavage in soy oil, and urine and feces were collected at 24-hour intervals, up to 96 hours after
administration (Meyer. 1977). The phenolic metabolites of biphenyl were analyzed as
trimethylsilyl (TMS) ethers by combined gas chromatography (GC)/mass spectrometry (MS)
(guinea pigs) or GC (rabbits). The biphenyl was hydroxylated to monohydroxylated biphenyls
and minor amounts of dihydroxylated derivatives, with the main route of excretion being through
the urine in both species and the major metabolite being 4-hydroxybiphenyl. In guinea pigs
5
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
(n = 3), the mass of identified metabolites in urine collected at 24 or 96 hours post-exposure
accounted for 29.5 or 32.9% of the administered dose, respectively. In the first 24 hours,
biphenyl and biphenyl metabolites in feces accounted for 20.3% of the dose; most of this
(14.3%)) was biphenyl, presumably unabsorbed. Bile was collected for 24 hours from another
group of two bile-cannulated guinea pigs dosed with 100 mg/kg biphenyl. No unchanged
biphenyl was detected in the collected bile, but conjugated mono- and dihydroxy metabolites
accounted for about 3% of the administered dose. The results with guinea pigs indicate that at
least 33%o of the administered dose was absorbed. In rabbits, urinary metabolites accounted for
49.1% of the dose, with most of this (25.4% on the first day and 15.9% on the second day)
eliminated as conjugates. In the first 24 hours, biphenyl and metabolites in feces accounted for
1.6% of the dose with 1.4% being biphenyl. These results indicate that at least 49% of the
administered dose was absorbed in rabbits.
Absorption of single oral 100 mg/kg doses of biphenyl (in soy oil or propylene glycol)
has also been demonstrated in male and female Danish Landrace pigs weighing 31-35 kg (Meyer
et al.. 1976a). Metabolites identified in urine collected at four 24-hour intervals after dose
administration included mono-, di-, and trihydroxybiphenyls, detected as TMS ethers by GC/MS
after enzyme hydrolysis of the samples by P-glucuronidase and sulphatase. Metabolites
identified and quantified in 24-hour urine samples accounted for averages of 17.5 and 26.5% of
the dose administered in soy oil to two female pigs and in propylene glycol to two male pigs,
respectively. Unchanged biphenyl was not detected in the urine samples. Metabolites in urine
collected for 96 hours accounted for averages of 27.6 and 44.8% of the doses administered to
female and male pigs, respectively. No phenolic metabolites of biphenyl were detected in feces
collected for 96 hours. Unchanged biphenyl was not detected in the feces collected from male
pigs, but the amount of unchanged biphenyl in feces from the two female pigs accounted for
18.4 and 5% of the administered dose. These results indicate that at least about 28 and 45% of
oral 100 mg/kg doses of biphenyl were absorbed in female and male pigs, respectively. It is
uncertain if the gender difference was due to vehicle differences or actual gender differences in
absorption efficiency.
Dermal absorption by human skin was measured in an in vitro static diffusion cell model
(Fasano. 2005). Epidermis (-0.64 cm ) was mounted onto an in vitro static diffusion cell,
stratum corneum uppermost. An infinite dose (100 |iL/cm for permeability experiment,
20 |iL/cm for exposure rate experiment) of biphenyl in isopropyl myristate vehicle was applied
to the epidermal surface, via the donor chamber. Fluid in the receptor chamber was analyzed
after different time periods. The study reported a permeability coefficient (Kp) of 6.12 x 10"5
cm/h, and short-term exposure rates of 258.3 jag equiv/cm Ih (10-minute exposure) and 59.1 jag
equiv/cm /h (60-minute exposure).
No animal studies were located examining quantitative aspects of absorption of biphenyl
by the respiratory tract.
6
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
3.2. DISTRIBUTION
No information was located regarding distribution of absorbed biphenyl in humans and
limited animal data are available. Meyer et al. (1976b) orally administered 100 mg/kg
[14C]-biphenyl to male albino rats and measured radioactivity in the lung, heart, kidney, brain,
spleen, liver, skeletal muscles, peritoneal fat, genital tract, and gastrointestinal tract at 96 hours
after dosing. Most of the radioactivity was excreted in urine (84.8%) and feces (7.3%) over the
96-hour period, and only 0.6% of the administered radioactivity remained in the animals at
96 hours: 0.1% was found in peritoneal fat, 0.3% in the gastrointestinal tract (including its
contents), 0.1% in skeletal muscles, and 0.1% in the genital tract. Levels of radioactivity in other
examined tissues were very low. The results indicate that absorbed biphenyl is not preferentially
stored in tissues and is rapidly excreted, principally through the urine.
3.3. METABOLISM
3.3.1. Identification of Metabolites
3.3.1.1. Results from In Vivo Animal Studies
No human studies on the in vivo metabolism of biphenyl have been identified. However,
the in vivo metabolism of biphenyl has been studied extensively in laboratory animals. These
studies have determined that in rats, rabbits, pigs, dogs, mice, and guinea pigs, biphenyl is
converted into a range of hydroxylated metabolites (Halpaap-Wood et al.. 1981b; Meyer. 1977;
Meyer and Scheline. 1976; Meyer et al.. 1976b; Meyer et al.. 1976a). These metabolites have
been detected in urine as both nonconjugated compounds and acidic conjugates.
The derivation of urinary metabolites and their subsequent analysis with GC has resulted
in the identification of >10 mono-, di-, and trihydroxybiphenyl metabolites from the urine of rats,
pigs, guinea pigs, and rabbits (Meyer. 1977; Meyer and Scheline. 1976; Meyer et al.. 1976b;
Meyer etal.. 1976a). These metabolites have been found as mercapturic acid conjugates and
glucuronide conjugates (Millburn et al.. 1967). Comparable metabolites have been identified
among mammalian species tested, although quantitative differences in metabolite formation are
evident among species. A major metabolite in the rat, mouse, guinea pig, rabbit, and pig was
reportedly 4-hydroxybiphenyl (Halpaap-Wood et al.. 1981b; Meyer. 1977; Meyer and Scheline.
1976). 4,4'-Dihydroxybiphenyl was identified as a major metabolite in the pig (Meyer et al..
1976a) and the rat (Halpaap-Wood et al.. 1981b; Meyer and Scheline. 1976). while 3,4-di-
hydroxybiphenyl was a major urinary metabolite in two strains of mice (Halpaap-Wood et al..
1981b). Table 3-1 reviews the metabolites that have been identified in the excreta and bile of
male albino rats given single doses of 100 mg biphenyl/kg, as reported by Meyer and Scheline
(1976).
7
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Table 3-1. Metabolites of biphenyl identified in urine, feces, and bile of male
albino rats
Metabolite"
Urine
Feces
Bile
Day 1
Day 2
Days 3 + 4
Days 1-4
Day 1
Day 1
Biphenyl
0.1
0.1
NDb
0.2
ND
ND
2-Hydroxybiphenyl
0.4
0.5
0.1
1.0
0.3
0.1
3 -Hydro xybiphenyl
0.9
0.4
0.3
1.6
0.5
0.5
4-Hydroxybiphenyl
6.8
0.7
0.2
7.7
1.0
1.5
3,4-Dihydro xybiphenyl
0.6
0.2
ND
0.8
ND
0.1
3,4' -Dihydro xybiphenyl
1.5
0.3
0.8
2.6
ND
0.3
4,4'-Dihydro xybiphenyl
9.6
1.7
0.1
11.4
1.8
1.9
2,5 -Dihydro xybiphenyl
Trace
ND
ND
Trace
ND
ND
Methoxy-hydroxybiphenyls
0.1
ND
ND
0.1
ND
0.1
Methoxy-dihydroxybiphenyls
0.5
0.3
0.1
0.9
ND
ND
3,4,4' -Trihydro xybiphenyl
1.8
0.9
0.5
3.2
1.1
0.7
Total
22.3
5.1
2.1
29.5
4.7
5.2
aValues are percent of administered dose.
bND = not detected.
Source: Meyer and Scheline (19761.
The hydroxylation of biphenyl to produce 2-hydroxybiphenyl is a minor pathway in rats
and mice, but is more easily detected in mice than rats (Halpaap-Wood et al.. 1981a. b).
Following intraperitoneal (i.p.) injection of [14C]-labeled biphenyl (30 mg/kg), the pattern of
percentages of radioactivity detected in urinary metabolites showed a relatively greater ability to
produce 2-hydroxybiphenyl in mice than rats. In Sprague-Dawley rats, metabolites identified in
order of abundance were (with percentage of total urinary radioactivity noted in parentheses):
4,4'-dihydroxybiphenyl (44.5%); 4-hydroxybiphenyl (28.5%); 3,4,4'-trihydroxybiphenyl (8.8%);
3,4'-dihydroxybiphenyl (8.5%); 3,4-dihydroxybiphenyl (5.1%); 3-hydroxybiphenyl (1.8%); and
2-hydroxybiphenyl (1.5%). In DBA/2Tex mice, major identified metabolites were: 4-hydroxy-
biphenyl (39.5%); 3,4-dihydroxybiphenyl (30.3%); 4,4'-dihydroxybiphenyl (10.2%);
3,4,4'-trihydroxybiphenyl (6.2%); 3-hydroxybiphenyl (4.3%); and 2-hydroxybiphenyl (4.2%).
In rats, 2,3-, 2,4-, and 2,5-dihyroxybiphenyl were detected at trace levels (<0.1%), whereas in
mice, these metabolites were detected at levels of 0.3, 0.8, and 0.7%, respectively (Halpaap-
Wood et al.. 1981b). No in vivo studies have been identified that directly investigate differential
metabolism of biphenyl between males and females of any species.
3.3.1.2. Results from In Vitro Studies with Animal and Human Cells or Tissues
The metabolism of biphenyl in vitro has been investigated using tissues of human origin,
resulting in evidence that the human metabolism of biphenyl is qualitatively similar to, but may
8
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
be quantitatively different from, rat metabolism. Benford et al. (1981) measured 2-, 3-, and
4-hydroxylation of biphenyl in microsomes prepared from the livers of five rats (sex not
identified) and four humans (sex not identified). The reaction products, after solvent extraction
and high-performance liquid chromatography (HPLC) quantitation, revealed that 2-hydroxylase
in the rat was 35 times higher than in humans, while 3- and 4-hydroxylases in humans were
1.5 and 1.2 times higher than in rats.
The evidence from studies of human tissue samples exposed to biphenyl metabolites in
vitro suggests differential Phase II metabolism contingent upon tissue origin. Powis et al. (1988)
have shown that/>hydroxybiphenyl is conjugated with glucuronic acid and sulphate in human
liver and kidney tissue slices. In the liver, glucuronidation was the favored conjugation pathway,
while sulphation was favored in the kidney. Powis et al. (1989) also compared Phase I biphenyl
metabolism in human (from surgery), dog (mongrel), and rat (male F344) liver slices and
primary hepatocytes. It was found that liver slices from all three species had a similar capacity
to metabolize biphenyl, -3.5 nmol biphenyl/minute per g tissue, while hepatocyte preparations
from rats had about 4 times the metabolic capacity of dog hepatocytes and about 20 times that of
human hepatocytes. Powis et al. (1989) speculated that hepatocytes from dog and human liver
slices may have experienced more damage during isolation than rat hepatocytes.
A study of the sulphation of biphenyl metabolites in human surgical tissue samples was
conducted by Pacifici et al. (1991). Tissue samples of various types (liver, intestinal mucosa,
lung, kidney, bladder, and brain) were obtained from surgeries of patients of both sexes between
the ages of 49 and 76 years of age (each patient contributed only one tissue type, so that within-
patient organ comparisons were not made). The tissues were homogenized, filtered, and
centrifuged at 12,000 and 105,000 g to obtain supernatants to study sulphation of biphenyl
metabolites, specifically 2-, 3-, and 4-hydroxybiphenyl. Sulphotransferase activity for each of
these substrates was detected in all tissues studied, although marked tissue dependence was
observed, with the highest activity found in the liver and the lowest in the brain. The Michaelis
constant (Km) of sulphotransferase was dependent on the substrate, but not on tissue type, with
Km varying over a 500-fold range. The highest values of Km were found with 4-hydroxybiphenyl
and the lowest were found with 3-hydroxybiphenyl.
Several studies of biphenyl metabolism with in vitro animal systems support the findings
from the in vivo urinary metabolite investigations that: (1) a range of hydroxylated biphenyl
metabolites are formed, (2) 4-hydroxybiphenyl is a major metabolite, and (3) hydroxylated
biphenyl metabolites are conjugated to glucuronic acid or sulphate. Wiebkin et al. (1984; 1976)
reported that isolated rat and hamster hepatocytes metabolized biphenyl primarily to
4-hydroxybiphenyl and also to 4,4'-hydroxybiphenyl, both of which were then conjugated. A
small amount of 2-hydroxybiphenyl was produced. When 4-hydroxybiphenyl was incubated
with the hepatocytes, it was hydroxylated to 4,4'-dihydroxybiphenyl. Pretreatment of the
animals with either 5,6-benzoflavone or phenobarbital had little effect on the conjugate
9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
formation rate in the in vitro experiment. Bianco et al. (1979) reported that rat hepatic
microsomes metabolize biphenyl to 4-, 2-, and 3-hydroxybiphenyl, which are conjugated to form
glucuronides and sulphates. The 4-hydroxybiphenyl isomer was the major metabolite. The
formation of 4-hydroxybiphenyl as a major metabolite in the hamster, mouse, and rabbit was
confirmed by Billings and McMahon (1978) 2-Hydroxybiphenyl and 3-hydroxybiphenyl were
detected in a lower amount in a ratio of 2:1 by hamster and rabbit microsomes, and in a 1:1 ratio
by mouse microsomes. In contrast, almost all hydroxylation of biphenyl in rat microsomes gave
rise to 4-hydroxybiphenyl.
3.3.2. Metabolic Pathways
3.3.2.1. Description of Metabolic Scheme and Enzymes Involved
Burke and Bridges (1975) suggested that biphenyl metabolism is mediated by
cytochrome P-450 (CYP) monooxygenases. Evidence of an arene oxide intermediate, which
may participate in binding to cellular macromolecules, was reported by Billings and McMahon
(1978). Support for CYP metabolism of biphenyl was provided by Halpaap-Wood et al. (1981a.
b), who reported that greater amounts of hydroxybiphenyls were obtained in in vitro assays using
liver homogenates when rats were treated first with P-naphthoflavone, 3-methylcholanthrene, or
Aroclor 1254, which are known CYP inducers. In C57BL/6Tex mice, CYP induction with
P-naphthoflavone led to relatively greater amounts of urinary excretion of 2-hydroxybiphenyl,
compared with uninduced mice, whereas pretreatment with P-naphthoflavone led to increases in
urinary excretion of 2-, 3-, and 4-hydroxybiphenyl in Sprague-Dawley rats and was without
influence on the pattern of hydroxybiphenyl metabolites in DBA/2Tex mice (Halpaap-Wood et
al.. 1981b).
Figure 3-1 details combined evidence from the Halpaap-Wood et al. (1981a. b) and
Meyer and Scheline (1976a) studies on the metabolic pathways of biphenyl. While sulphates
and glucuronides are formed on all three metabolic levels illustrated, only monosulphates and
monoglucuronides are identified. Monomethyl ethers are formed from dihydroxy and trihydroxy
metabolites alone. Glucuronides at the dihydroxy and trihydroxy levels are additionally labeled
with a question mark to suggest that, while these metabolites are likely, they have not been
identified.
10
DRAFT - DO NOT CITE OR QUOTE
-------
Bi phenyl
CYP
3-Hydroxybi phenyl
2-Hydroxybiphenyl
HO'
4-Hydroxybiphenyl
ar,ar'-Dihydroxybiphenyl
a r, a r- Di hyd roxy bi ph e ny I
ar,ar,ar'-Trihydroxybi phenyl
HO
^ //~^ //
Sulphono-
transferase Monohydroxysulfate
-glucuronide
HO.
HO.
UGT
\ /r^\
OH
\ /r~\
OH
OH >
COMT
Dihydroxy-monosulfate
-monoglycuronide (?)
-monomethyl ether
Trihydroxy-monosulfate
-monoglucuronide (?)
-monomethyl ether
ar = aryl group; COMT = catechol-O-methyltransferase; UGT = uridine diphosphate
glucuronosyl transferase; question marks denote tentative metabolites (see text).
Sources: Halpaap-Wood et al. (1981a. b); Meyer and Scheline (1976a).
Figure 3-1. Schematic presentation of the metabolic pathways of biphenyl.
11
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
The metabolic scheme in Figure 3-1 does not include the possible redox cycling of
2,5-dihydroxybiphenyl (also known as phenylhydroquinone), which involves CYP-mediated
cycling between phenylhydroquinone and phenylbenzoquinone leading to the generation of
reactive oxygen species (ROS) (Balakrishnan et al.. 2002; Kwok et al.. 1999). This pathway is
thought to play a role in the carcinogenic effect of 2-hydroxybiphenyl (also known as
ortho-phenylphenol), a broad spectrum fungicide that, like biphenyl, induces urinary bladder
tumors in chronically exposed male rats with a nonlinear dose-response relationship (i.e.,
incidence of bladder tumors of 96% at 1.25% in diet, but no tumors at the concentrations of
0.625%) or lower) (Kwok et al.. 1999; Hiraga and Fujii, 1984). Free 2,5-dihydroxybiphenyl and
its glucuronide or sulphate conjugates are readily detected in the urine of rats exposed to
2-hydroxybiphenyl, and the formation of 2,5-dihydoxybiphenyl and phenylbenzoquinone is the
principal metabolic pathway for 2-hydroxybiphenyl in the rat, especially at high exposure levels
associated with urinary bladder tumor formation (Kwok et al.. 1999; Morimoto et al.. 1989;
Nakao et al.. 1983; Reitz et al.. 1983; Meyer and Scheline. 1976). In contrast, the formation of
4-hydroxybiphenyl and 4,4'-dihydroxybiphenyl is the principal metabolic pathway for biphenyl
in rats and mice, and 2,5-dihydroxybiphenyl was not detected, or only detected at trace levels, in
the urine of rats exposed to 100 mg biphenyl/kg (Meyer and Scheline, 1976; Meyer et al., 1976a)
(see Table 3-1). In mice exposed to i.p. doses of [14C]-biphenyl (30 mg/kg), radioactivity in
2-hydroxybiphenyl and 2,5-dihydroxybiphenyl in the urine accounted for only about 5%> of the
total radioactivity detected in urinary metabolites (Halpaap-Wood et al.. 1981b).
3.3.3. Regulation of Metabolism and Sites of Metabolism
3.3.3.1. Evidence for Induction of Phase I and II Enzymes
No studies of Phase I or II enzyme induction using liver microsomes of human origin
were identified. However, a number of studies have been conducted in rodents to investigate the
induction of Phase I enzymes that catalyze biphenyl hydroxylation. For example, Creaven and
Parke (1966) reported that pretreatment of weanling Wistar rats or ICI mice with phenobarbital
[an inducer of CYP3 A4, 2B6, and 2C8 as reported by Parkinson and Ogilvie (2008)1 or
3-methylcholanthrene [an inducer of CYP1A2 as reported by Parkinson and Ogilvie (2008)1
increased NADPH-dependent activities of liver microsomes to produce 2-hydroxybiphenyl and
4-hydroxybiphenyl from biphenyl to varying degrees depending on the inducer. Haugen (1981)
reported that pretreatment of male CD rats with phenobarbital or 3-methylcholanthrene increased
NADPH-dependent activities of liver microsomes to produce 2-, 3-, and 4-hydroxybiphenyl from
biphenyl, again to varying degrees depending on the inducer. Stuehmeier et al. (1982) reported
that phenobarbital pretreatment of male C57BL/6JHan mice induced liver microsomal activities
to produce 4-hydroxybiphenyl, but not 2-hydroxybiphenyl, from biphenyl, whereas
3-methylcholanthrene induced activities for both 4- and 2-hydroxylation of biphenyl. Halpaap-
Wood et al. (1981b) reported that pretreatment of male Sprague-Dawley rats with
12
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
P-naphthoflavone [an inducer of CYP1A2 as reported by Parkinson and Ogilvie (2008); also
known as 5,6-benzoflavone] enhanced the urinary excretion of 2-, 3-, and 4-hydroxybiphenyl,
3,4-dihydroxybiphenyl, and 3,4,4'-trihydroxybiphenyl following i.p. administration of 30 mg
biphenyl/kg body weight. In contrast, pretreatment of male C57BL/6Tex mice with
P-naphthoflavone did not increase the overall urinary excretion of biphenyl metabolites
following i.p. administration of 60 mg biphenyl/kg, but shifted the principal metabolite from
4-hydroxybiphenyl to 2-hydroxybiphenyl and 2,5-dihydroxybiphenyl (Halpaap-Wood et al..
1981b). Wiebkin et al. (1984) reported that P-naphthoflavone pretreatment of male Lewis rats or
male Syrian golden hamsters induced biphenyl hydroxylation activities in freshly isolated
pancreatic acinar cells or hepatocytes. From these observations and examination of patterns of
inhibition of biphenyl hydroxylation activities by CYP inhibitors (e.g., a-naphthoflavone and
1-benzyl-imidazole) under non-induced and induced conditions (Haugen. 1981). it is apparent
that multiple CYP enzymes (e.g., CYP1A2 and CYP3A4) are likely involved in biphenyl
hydroxylation. However, no studies were located that used more modern techniques (such as
CYP knockout mice) to identify the principal CYP enzymes involved in the initial hydroxylation
of biphenyl or the formation of the dihydroxy- or trihydroxybiphenyl metabolites.
Several animal studies were located examining the possible coordinated induction of
Phase I enzymes with Phase II enzymes catalyzing the conjugation of hydroxylated biphenyl
metabolites to sulphate or glucuronic acid. Hepatocytes from rats (strain and sex were not noted)
pretreated with the CYP inducers, phenobarbital or 3-methylcholanthrene, produced glucuronide
and sulphate conjugates of 4-hydroxybiphenyl when incubated with biphenyl (Wiebkin et al..
1978). Glucuronide conjugates were predominant under these "CYP-induced" conditions,
whereas hepatocytes from non-induced control rats produced predominant sulphate conjugates of
4-hydroxybiphenyl. These results suggest that induction (or possibly activation) of
glucuronidation enzymes may be coordinated with the induction of CYP enzymes. In contrast,
pretreatment of male Lewis rats with P-naphthoflavone (an inducer of CYP1A2) did not enhance
activities of freshly isolated pancreatic acinar cells to conjugate 4-hydroxybiphenyl with sulphate
or glucuronic acid, but the influence of this pretreatment on the conjugation capacity of
hepatocytes was not examined in this study (Wiebkin et al.. 1984). In another study, uridine
diphosphate glucuronosyl transferase (UGT) activities with 1-naphthol or 3-hydroxy-
benzo[a]pyrene as substrates were higher in liver microsomes from male Wistar rats pretreated
with Aroclor 1254 (an inducer of several CYP enzymes) or phenobarbital, respectively,
compared with microsomes from control rats without pretreatment with CYP inducers (Bock et
al.. 1980). Although Bock et al. (1980) measured UGT activities in microsomes from several
tissues from non-induced rats with 4-hydroxybiphenyl as a substrate, no comparisons between
induced and non-induced conditions were made using 4-hydroxybiphenyl as substrate. Paterson
and Fry (1985) reported that hepatocytes or liver slices from male Wistar rats pretreated with
P-naphthoflavone showed decreased rates of glucuronidation of 4-hydroxybiphenyl, compared
13
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
with hepatocytes or liver slices from rats without P-naphthoflavone pretreatment. Results from
this database provide equivocal evidence that the induction of Phase I enzymes catalyzing the
hydroxylation of biphenyl may be coordinated with induction of Phase II enzymes catalyzing
glucuronidation of hydroxylated biphenyl metabolites.
3.3.3.2. Demonstrated Tissue Sites of Metabolism
CYP enzymes catalyzing hydroxylation of biphenyl and other substrates are present in
most, if not all, mammalian tissues, but the highest levels of activities are normally found in liver
(Parkinson and Ogilvie. 2008). In a study of male Sprague-Dawley rats, CYP content was 20-
40-fold higher in the microsomes from liver than from lung, although biphenyl-4-hydrolase
activity was only 1.7-fold higher in the microsomes from liver than from lung (Matsubara et al..
1974). Wiebkin et al. (1984) observed 200- and 1,000-fold higher rates of biphenyl metabolism
in 5,6-benzoflavone-pretreated hepatocytes compared to similarly treated pancreatic acinar cells
from male Lewis rats and Syrian golden hamsters, respectively.
Activities for enzymes catalyzing the conjugation of hydroxybiphenyls and other
hydroxylated aromatic compounds with glucuronic acid or sulphate have been detected in a
number of mammalian tissues, and, similar to CYP, the highest levels are found in the liver
(Parkinson and Ogilvie. 2008). Available data for conjugation activities with hydroxybiphenyls
in various mammalian tissues are consistent with this concept. Sulphotransferase activities with
2-, 3-, or 4-hydroxybiphenyl as substrates in microsomes from several human tissues showed an
approximate 100- to 500-fold range with the following order: liver > ileum > lung > colon >
kidney > bladder > brain (Pacifici et al.. 1991). UGT activities with 4-hydroxybiphenyl as
substrate in microsomes from several male Wistar rat tissues showed the following order: liver >
intestine > kidney > testes ~ lung; activities were below the limit of detection in microsomes
from skin and spleen (Bock et al.. 1980).
3.4. ELIMINATION
No studies were located on the route or rate of elimination of biphenyl in humans, but
results from studies of orally exposed animals indicate that absorbed biphenyl is rapidly
eliminated from the body, principally as conjugated hydroxylated metabolites in the urine.
The most quantitative data on the routes and rates of elimination come from a study of
rats following administration of radiolabeled biphenyl (Meyer etal.. 1976a). Urine collected for
24 hours after the oral administration of 100 mg/kg [14C]-labeled biphenyl in soy oil to male
albino rats contained 75.8% of the administered radioactivity, compared with 5.8% detected in
feces collected in the same period. Ninety-six hours after dose administration, <1% of the
administered radioactivity remained in tissues, 84.8% was in collected urine, 7.3% was in feces,
and O.P/o was in collected expired air (Meyer etal.. 1976b). Although chemical identity analysis
of fecal radioactivity was not conducted by Meyer et al. (1976b). results from GC/MS analyses
14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
of bile collected from bile-cannulated rats given single 100 mg/kg doses of unlabeled biphenyl
indicate that biliary excretion of metabolites represents a minor pathway of elimination (Meyer
and Scheline. 1976). In bile collected for 24 hours, unchanged biphenyl was not detected and
conjugated metabolites accounted for 5.2% of the administered dose; in contrast, conjugated
metabolites of biphenyl in 24-hour urine accounted for 22.3% of the dose (Meyer and Scheline.
1976).
Supporting evidence for the importance of urinary elimination of conjugated metabolites
is provided by the results of other studies, which analyzed biphenyl and biphenyl metabolites by
GC/MS or GC in urine and feces collected from rabbits (Meyer. 1977). guinea pigs (Meyer.
1977). and pigs (Meyer et al.. 1976a) following oral administration of 100 mg/kg doses of
unlabeled biphenyl. In 24-hour urine samples, unchanged biphenyl was not detected, and total
metabolites accounted for averages of 25.4% of the administered dose in rabbits, 31.3% in
guinea pigs, 17.5% in female pigs, and 26.4% in male pigs. As in rats, biliary excretion
represents a minor elimination pathway in guinea pigs and rabbits; metabolites detected in bile
collected for 24 hours from bile-cannulated guinea pigs accounted for 3.3% of the administered
dose, but for only 0.3% of the dose in bile collected for 7 hours from a rabbit given 100 mg/kg
biphenyl (Meyer. 1977). Neither unchanged biphenyl nor hydroxylated biphenyl metabolites
were detected in bile collected from a bile-cannulated pig for 24 hours after administration of
100 mg/kg biphenyl (Meyer et al.. 1976a).
No studies were located examining quantitative aspects of elimination in animals
following inhalation or dermal exposure to biphenyl.
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS
No studies were located on the development of PBPK models for biphenyl in animals or
humans.
15
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS
Summary. Human studies of biphenyl include assessments of workers exposed to
biphenyl during production of biphenyl-impregnated fruit wrapping paper at one mill in Finland
and another mill in Sweden. The study of Finish mill workers provided evidence of abnormal
electroencephalography (EEG), nerve conduction velocity, and electromyographic (EMG) test
results in workers exposed to biphenyl at levels in excess of occupational exposure limits
(Seppalainen and Hakkinen, 1975; Hakkinen et al., 1973). Similar neurological findings were
not reported in the study of Swedish mill workers whose exposures were likely to have exceeded
the occupational exposure limit (Wastensson et al.. 2006); however, an increased relative risk of
Parkinson's disease was reported.
A case report of a 46-year-old female who worked at a fruit-packing facility in Italy over
a 25-year period where biphenyl-impregnated paper was used presented with hepatomegaly,
neutrophilic leukocytosis, clinical chemistry findings indicative of hepatic perturbation, and liver
biopsy indicative of chronic hepatitis (Carella and Bettolo.(T994). Following cessation of work
in citrus packing, serum enzymes returned to normal, suggesting that occupational exposure to
biphenyl may have been the principal etiological factor.
Hakkinen and colleagues assessed the health of paper mill workers exposed to biphenyl
during the production of biphenyl-impregnated paper used to wrap citrus fruits. In 1959,
workers complained about a strong odor and irritation to the throat and eyes. Air measurements
made at various locations within the facility in June of 1959 resulted in estimated average
"3
biphenyl concentrations of 4.4-128 mg/m (Table 4-1). In 1969, a 32-year-old worker at the
facility, who had worked for 11 years in the oil room where biphenyl levels were particularly
high, became ill. Despite aggressive medical intervention, the patient grew worse and died. Key
features at autopsy included necrosis of most liver cells, severe, but unspecified changes in the
kidneys, degeneration of the heart muscles, hyperactive bone marrow, and edematous changes in
the brain (Hakkinen et al.. 1973; 1971). Subsequent measurements of biphenyl in the workplace
"3
air (January 1970) resulted in estimated average concentrations ranging from 0.6 to 123 mg/m
(Table 4-1). Measurements taken in both 1959 and 1971 indicated that biphenyl air
concentrations at multiple work areas greatly exceeded the current American Conference of
Governmental Industrial Hygienists (ACGIH. 2001) threshold limit value (TLV) of 0.2 ppm
"3
(1.3 mg/m ). In the location where biphenyl was mixed with paraffin oil (the oil room), biphenyl
occurred both as a vapor and as a dust, suggesting the possibility of both dermal and inhalation
exposures.
16
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Table 4-1. Biphenyl concentrations in the air of a Finnish paper mill
producing biphenyl-impregnated fruit wrapping paper
Sampling center locations
Average concentrations (mg/m3)
June 1959
January 1970
Paper mill hall
In front of paper reel
17.9
7.2
Behind impregnating roller
128.0
64.0
Near paper machine
7.2
1.5
Near rolling machine
4.4
0.6
Oil-room
Near measuring container
19.5
3.5
Above measuring container (lid open)
No data
123.0
Near mixing container
No data
15.5
During addition of biphenyl to mixing container
No data
74.5
Source: Hakkinen et al. (19731.
Thirty-one male workers engaged in the biphenyl-impregnation process and two other
workers exposed to biphenyl elsewhere in the facility were included in the study. Common
complaints among these workers included fatigue, headache, gastrointestinal discomfort,
numbness and aching of the limbs, and general fatigue; laboratory tests revealed elevated serum
aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (which can indicate
inflammation or damage to liver cells) in 10 of the 33 workers (Hakkinen et al.. 1973). Eight of
the 33 workers were admitted to the hospital for further examination, including liver biopsy.
Twenty two of the 33 workers (including the 8 who were hospitalized for testing) were subjected
to neurophysiological examinations, including electroencephalography (EEG) and
electroneuromyography (ENMG, consisting of nerve conduction velocity and electromyographic
[EMG] tests). Fifteen of these 22 workers displayed abnormal findings and four displayed
borderline findings on one or both of theses tests. Exposure to biphenyl was terminated
immediately following the initial neurophysiological examinations, and 11 and 7 of these
subjects were retested 1 and 2 years later, respectively. Seppalainen and Hakkinen (1975)
reported more detailed information about these examinations, and included results for two
additional workers for a total of 24, as summarized below.
EEG results. At initial examination, 10 of the 24 workers had abnormal EEGs, which
included diffuse slow wave abnormalities (6 cases), lateral spike and slow wave discharges
(2 cases), posterior slowing only (1 case), and mild slow wave abnormality in the right temporal
area (1 case). Six subjects exhibited unusual distribution of alpha rhythm, with alpha activity
also prominent in the frontal areas. Four of the subjects exhibited no EEG abnormalities. In
general, the EEG results observed at initial examination were qualitatively similar in the 11
subjects reexamined 1 year later. Exceptions included additional diffuse slow wave
17
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
abnormalities in the two subjects initially exhibiting only spike and wave discharges and the
disappearance of the one case of mild temporal local abnormality. There was no discernable
improvement in the EEGs of the seven subjects reexamined after 2 years.
ENMG results. As shown in Table 4-2, the 24 biphenyl-exposed workers exhibited no
significant differences in mean maximal motor conduction velocity (MCV) relative to those of a
control group consisting of 60 healthy Finnish males, but significantly (p < 0.001) slower mean
conduction velocity of the slowest motor fibers (CVSF) of the ulnar nerves. Results at the 1-year
follow up of 11 of the biphenyl-exposed workers revealed no significant changes in initial
conduction velocity measures, but at the 2-year reexamination of 7 of the 11 subjects, the MCVs
of the median and deep peroneal nerves were significantly slower (p < 0.02 and p < 0.01,
respectively) compared to the initial measurements. Abnormal EMGs among the biphenyl-
exposed workers included diminished numbers of motor units on maximal muscle contraction
(10 subjects) and fibrillations in some muscles (7 subjects). Workers exhibiting abnormal EMGs
typically displayed slowing of some nerve conduction velocities as well. Of those 11 subjects
undergoing repeat ENMG examination after 1 year, 5 subjects showed an increased level of
ENMG abnormality, while 4 remained unchanged and 2 had diminished abnormalities. At the
end of 2 years, three of seven subjects displayed diminished ENMG abnormalities, three of seven
were unchanged, and one of seven had the abnormality increased.
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl:
comparison with 60 unexposed males
Nerve
Biphenyl group (mean ± SD)
Control group (mean ± SD)
Median
MCV
57.7 ±6.3
58.0 ±3.8
Ulnar
MCV
56.3 ±4.6
56.6 ±4.0
CVSF
41.4 ±5.2*
45.5 ±3.2
Deep peroneal
MCV
50.2 ±5.4
50.3 ±3.5
CVSF
37.7 ±3.9
38.2 ±5.6
Posterior tibial
MCV
43.4 ±3.9
42.4 ±4.7
Statistically significant (/-test, p < 0.05) as reported by study authors.
Source: Scppalaincn and Hakkincn (1975).
Seppalainen and Hakkinen (1975) noted that subjects often exhibited signs of dysfunction
in both the peripheral nervous system, as evidenced by abnormal ENMGs, and the central
nervous system, as evidenced by abnormal EEGs and abnormal distribution of alpha activity.
Only five subjects (four men and the only woman in the biphenyl-exposed group) were found to
18
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
have completely normal neurophysiological records. The authors interpreted their data to
indicate that biphenyl can attack the nervous system at different levels, the sites of greatest
vulnerability being the brain and peripheral nerves. Anomalies in nerve conduction, EEG, and
ENMG signals, while small, were consistent with the persistence of incapacity and the incidence
of subjective symptoms.
Another study examined the prevalence and incidence of Parkinson's disease among
workers at a facility manufacturing biphenyl-impregnated paper in Sweden (Wastensson et al.,
(2006). The study was prompted by the recognition that three cases seen at a neurological clinic
shared a history of work at this workplace. The investigators used company and union records to
identify 506 people who had worked in this production process between 1954 and 1970. Vital
status was traced through the Swedish National Population registry; 222 had died and 284 were
still alive in Sweden in August 2002. The files were missing data for 4 years (1965-1968), and
the investigators estimated that this resulted in approximately 30 missing individuals from the at
risk pool. Prevalent cases were identified among those still alive through review of medical
records as well as a second examination by a study neurologist. Case definition was based on the
presence of at least two signs (tremor, rigidity, hypokinesia) and positive response to levodopa (a
treatment for Parkinson's disease). The National Hospital Discharge Register, Cause of Death
Register, and medical records were examined to determine presence of Parkinson's disease
among those who had died. Comparison rates for prevalence of Parkinson's disease was based
on age- and sex-specific prevalence rates from a study in eastern Sweden; prevalence risk ratios
were calculated for ages <80 years because of the larger variation seen among studies in rates at
older ages. The data from the deceased group was not included in these calculations, but were
included in analyses of lifetime risk, with comparison rates based on age- and sex-specific data
from a study in Olmsted County, Minnesota (a population served by the Mayo Clinic).
Wastensson et al. ((2006) identified 5 prevalent cases among the 255 workers ages
<80 years compared with 0.9 cases expected, for a relative risk [RR] of 5.6 [95% confidence
interval 1.9-13], The mean age at onset of symptoms was 51 years (range 45-55), considerably
lower than the mean of 66 years seen in the comparison population. Nine cases were identified
among the 222 deceased workers, compared to 4.3 expected (RR 2.1, 95% CI 0.96, 4.0). The
clinical features and exposure data for the five living subjects, all of whom were diagnosed with
Parkinson's disease by a neurologist at a local hospital, are summarized in Table 4-3. With one
exception, the patients were in comparatively good health on initial diagnosis. The exception
was a 53-year-old male who had diabetes mellitus and withdrew from the study before his
neurological condition could be confirmed.
19
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Table 4-3. Exposure data and clinical features for five Parkinson's disease
patients with occupational exposure to biphenyl
Case
1
2
3
4
5
Exposure data
Age
63
63
58
54
63
Workplace
PM3
PM3
PM4
PM3
PM3
Years of exposure3
12
4
9
4
2
Age at onset of exposure
19
26
17
18
21
Age at onset of symptoms
52
55
44
51
55
Clinical features
Resting tremor
+
+
+
+
+
Cogwheel rigidity
+
+
+
-
+
Bradykinesia
+
+
+
+
-
Positive response to levodopab
+
+
+
+
+
"Exposure to biphenyl about one-third of each year.
bAll five patients improved with levodopa..
PM = paper mill
Source: Wastensson et al. (20061
Four of the five prevalent cases worked in the vicinity of a rewinder/dryer, while the fifth
attended to another rewinder. Although no ambient biphenyl levels were available for the
subjects' work space, it was thought likely that the level of biphenyl in air would be greater
"3
(more than two times higher) than the existing TLV of 1.3 mg/m (0.2 ppm) based on
measurements at a Finnish paper mill with similar production practices (Hakkinen et al.. 1973).
Two subjects may have been exposed to higher levels of biphenyl than the others when they
created the paraffin oil/biphenyl mixture.
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
Summary. Available oral data for biphenyl include two well-designed two-year chronic
toxicity and carcinogenicity studies, one in F344 rats (TJmeda et al.. 2002) and one in BDFi mice
(TJmeda et al.. 2005). Increased incidence of urinary bladder transitional cell papillomas and
carcinomas, associated with the formation of urinary bladder calculi, occurred in male, but not
female, F344 rats only at the highest tested dietary concentration, 4,500 ppm; neither the
neoplasia nor the calculi were found at lower exposure levels of 1,500 or 500 ppm.
Nonneoplastic kidney lesions were found in F344 rats at biphenyl dietary concentrations
>1,500 ppm (TJmeda et al.. 2002). Several other rat studies provide supporting evidence that the
kidney and other urinary tract regions are critical targets for biphenyl in rats (Shiraiwa et al..
20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
1989; Ambrose et al.. 1960; Pecchiai and Saffiotti. 1957; Dow Chemical Co. 1953). In BDFi
mice, increased incidence of liver tumors (hepatocellular adenomas and carcinomas) and
noncancer effects on the kidney (mineralization) and liver (increased activities of plasma ALT
and AST) were found in females exposed to biphenyl dietary concentrations of 2,000 or
6,000 ppm (Umeda et al.. 2005). There was a small increase in reticular cell sarcomas in female
B6C3Fi mice exposed to 517 ppm biphenyl in the diet for 18 months, but not in similarly
exposed male B6C3Fi mice or either sex of B6AK Fi mice (Innes et al.. 1969; NCI. 1968). In
contrast, no carcinogenic responses or adverse noncancer effects were found in female ddY mice
exposed to 5,000 ppm biphenyl in the diet for 2 years (Tmai et al.. 1983).
No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
3 3
biphenyl were reported in mice exposed to 5 mg/m and in rats exposed to 300 mg/m , but not in
rabbits exposed to 300 mg/m (Deichmann et al.. 1947; Monsanto. 1946). Congestion or edema
of the lung, kidney, and liver, accompanied by hyperplasia with inflammation of the trachea, was
found in CD-I mice exposed to biphenyl vapors at 158 or 315 mg/m for 13 weeks (Sun. 1977a).
Study descriptions for all available subchronic and chronic toxicity and carcinogenicity
studies follow.
4.2.1. Oral Exposure
4.2.1.1. Subchronic Toxicity
Twenty-one-day-old female Long-Evans rats (8/group) were exposed to 0, 0.01, 0.03, or
0.1% biphenyl in the diet for 90 days (Dow Chemical Co., 1953). Based on U.S. EPA (1988)
subchronic reference values for body weight and food consumption in female Long-Evans rats,
these dietary levels corresponded to doses of 10, 30, and 100 mg/kg-day, respectively. Body
weights were monitored 3 times/week, and the weights of the liver, kidneys, adrenals, and spleen
were recorded at necropsy. Heart, liver, kidney, spleen, adrenals, pancreas, ovary, uterus,
stomach, small and large intestine, voluntary muscle, lung, thyroid, and pituitary from each rat
were examined histopathologically (2 rats/group).
There were no significant treatment-related effects on body weight, food consumption, or
organ weights. Results of histopathologic examinations were unremarkable. Biphenyl-exposed
groups exhibited lower average plasma blood urea nitrogen (BUN) levels than controls (28.2,
25.7, and 26.3 mg percent for low-, mid-, and high-dose groups, respectively, compared to
35.3 mg percent for controls, based on measurements in 4 rats/group). The biological
significance of these decreases in BUN is unclear.
Six-week-old BDFi mice (10/sex/group) were exposed to biphenyl at dietary
concentrations of 0, 500, 2,000, 4,000, 8,000, 10,000, or 16,000 ppm for 13 weeks (Umeda et al.
(2004a). Based on U.S. EPA (1988) subchronic reference values for body weight and food
consumption (average values for combined sexes), these dietary concentrations corresponded to
21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
doses of 93, 374, 747, 1,495, 1,868, and 2,989 mg/kg-day, respectively1. Animals were checked
daily for clinical signs; body weight and food consumption were recorded weekly; organ weights
were noted at term; and liver sections were processed for light microscopic examination.
Electron microscopy was carried out on liver tissue from one control and one 16,000 ppm
female.
A single 16,000 ppm female mouse died during the study; all other mice survived until
terminal sacrifice. Final body weights of mice of both sexes in the 8,000, 10,000, and
16,000 ppm groups were decreased by more than 10% compared to controls (for males: 83.3,
84.9, and 75.1% of controls; for females: 93.7, 91.6, and 85.8% of controls, respectively).
Umeda et al. (2004b) noted that absolute liver weights were significantly higher in 8,000 and
16,000 ppm female mice, but did not include the extent of these increases in the study report.
Light microscopic examination of liver specimens from all 16,000 ppm female mice revealed
enlarged centrilobular hepatocytes, the cytoplasm of which was filled with numerous
eosinophilic fine granules. Upon electron microscopic examination, these eosinophilic granules
were identified as peroxisomes, indicative of a peroxisome proliferative effect in the liver of the
16,000 ppm female mice. Evidence of histopathologic liver lesions was not found in females of
the 8,000 or 10,000 ppm groups. There were no signs of treatment-related increased liver weight
or histopathologic evidence of clearly enlarged hepatocytes in any of the biphenyl-treated groups
of male mice.
Mongrel dogs (two males and one female/group) were administered 0, 2.5, or 25 mg/kg
biphenyl in corn oil by capsule 5 days/week for 1 year (Monsanto, (1946). Dogs were examined
daily for clinical signs and weighed weekly. Blood samples were drawn at 3-month intervals to
measure hematological and clinical chemistry parameters. Urine samples were obtained at
similar intervals to measure specific gravity, sugar, protein, bile pigments, occult blood, and
microscopic sediment. Samples of urine from the high-dose dogs were collected during week
18, pooled, and analyzed for the presence of biphenyl and metabolites. At termination, gross
necropsies were performed, and sections of large and small intestine, pancreas, ovary or testis,
adrenal, urinary bladder, stomach, lung, thyroid, brain, heart, spleen, and liver were prepared for
histopathologic examination. Although slight fluctuations were seen in body weight during the
study, the dogs generally exhibited a net weight gain. Fluctuations in hematological parameters
and urine analysis were inconsistent and not considered compound-related. Gross pathological
examination of the dogs showed no obviously compound-related effects. Histopathologic
examinations revealed lung congestion consistent with bronchial pneumonia in one high-dose
dog; histopathology was unremarkable for each of the other dogs in the study.
1 To overcome possible problems with taste aversion, mice assigned to the 8,000 and 10,000 ppm groups were fed
4,000 ppm dietary biphenyl for the first week and 8,000 or 10,000 ppm for the remaining 12 weeks. Mice
designated to receive 16,000 ppm were fed 4,000 ppm dietary biphenyl for the first week, 8,000 ppm for the second
week, and 16,000 ppm for the remaining 11 weeks.
22
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Dow Chemical Co. (1953) described a biphenyl feeding experiment in which four groups
of Rhesus monkeys (two males and one female/group) were exposed to 0, 0.01, 0.1, or 1%
biphenyl in chow for 1 year, during which time most of the animals experienced ill health not
related to biphenyl exposure. Hematological parameters and BUN were within normal limits in
all groups of animals, and no dose-related effects on final body weight or weights of the lung,
kidney, heart, or spleens were observed. The authors considered an increase in relative liver
weight in high-dose monkeys (4.65 g/100 g body weight versus 3.90 g/100 g body weight in
controls) to possibly be compound-related.
4.2.1.2. Chronic Toxicity and Carcinogenicity
4.2.1.2.1. Chronic rat studies
In a chronic toxicity and carcinogenicity study of F344 rats (50/sex/group) conducted by
the Japan Bioassay Research Center (JBRC), biphenyl was administered in the diet for 2 years at
concentrations of 0, 500, 1,500, or 4,500 ppm (Umeda et al., 2002). Based on time-weighted
average (TWA) body weights estimated from the graphically-depicted data (Umeda et al., 2002;
Figure 1) and chronic reference values for food consumption in F344 rats (U.S. EPA, 1988),
these dietary concentrations corresponded to doses of 36.4, 110, and 378 mg/kg-day,
respectively, for males and 42.7, 128, and 438 mg/kg-day, respectively, for females. All animals
were examined daily for clinical signs; body weights and food intake were determined
once/week for the first 14 weeks and every 4 weeks thereafter. Urinalysis was performed on all
surviving rats at week 105. Upon necropsy, all major organs were weighed and tissue samples
were subjected to histopathologic examination.
Mean body weights of 4,500 ppm male and female rats were lower than those of controls
throughout most of the study period and were approximately 20% lower than respective controls
at terminal sacrifice. There was no statistically significant effect on mean body weights of 500
or 1,500 ppm males or females. Survival of low- and mid-dose male and female rats was
reported not to differ statistically significantly from controls.
The study authors reported that 3/50 of the 4,500 ppm female rats died after 13-26 weeks
of biphenyl exposure and attributed the deaths to marked mineralization of the kidneys and heart.
However, they also indicated that survival of this group was not adversely affected thereafter.
Significantly decreased survival was noted only for the group of 4,500 ppm male rats, 19/50 of
which died prior to terminal sacrifice. The first death occurred around treatment week 36; this
rat exhibited urinary bladder calculi. Survival data for the other groups were not provided.
Evidence of hematuria (blood in the urine) was first noted in 4,500 ppm male rats around week
40 and was observed in a total of 32/50 of the 4,500 ppm males during the remainder of the
treatment period; 14 of these rats appeared anemic. Hematuria and bladder tumors were
considered as primary causes of death among the 4,500 ppm males (n = 19) that died prior to
terminal sacrifice.
23
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Urinalysis performed during the final treatment week revealed statistically significantly
increased urinary pH in the 31 remaining 4,500 ppm male rats (pH of 7.97 versus 7.66 for
controls; p < 0.05), with occult blood noted in the urine of 23 of these males. Urine samples in
10/37 surviving 4,500 ppm females tested positive for occult blood. Relative kidney weights of
1,500 and 4,500 ppm males and females and absolute kidney weights of 4,500 ppm males were
statistically significantly increased (actual data were not reported).
Gross pathologic examinations at premature death or terminal sacrifice revealed the
presence of calculi in the bladder of 43/50 of the 4,500 ppm males and 8/50 of the 4,500 ppm
females, but not in the other dose groups (Table 4-4). The bladder calculi in the male rats were
white, yellow, brown, gray, and black in color, ranged from 0.3 to 1.0 cm in size, and exhibited
triangular, pyramidal, cuboidal, and spherical shapes. The bladder calculi in the female rats were
white and yellow in color, of uniform spheroidal shape, and similar in size to those of the male
rats. Polyp-like or papillary nodules protruding into the lumen from the bladder wall were found
in 41 of the 4,500 ppm male rats; bladder calculi were noted in 38 of these males. Four of the
eight calculi-bearing 4,500 ppm female rats also exhibited thickening of the bladder wall. It was
noted that 30/32 of the 4,500 ppm male rats with hematuria also exhibited kidney or urinary
bladder calculi.
Histopathologic examinations at death or terminal sacrifice revealed no indications of
biphenyl-induced tumors or tumor-related lesions in organs or tissues other than those associated
with the urinary tract. As shown in Table 4-4, neoplastic and nonneoplastic lesions of the
urinary bladder were essentially limited to the 4,500 ppm rats and predominantly the males.
Only 4,500 ppm male rats exhibited papilloma (10/50) or carcinoma (24/50) of transitional cell
epithelium; three rats exhibited both papilloma and carcinoma. Most of the transitional cell
carcinomas (20/24) projected into the lumen, and the tumor cells invaded the entire body wall.
Bladder calculi were found in all 24 males with transitional cell carcinoma and 8/10 of the males
with transitional cell papilloma. Simple, nodular, and papillary hyperplasias that developed in
the focal area of the bladder epithelium were evident in 4,500 ppm animals. Ten of the
4,500 ppm males had polyps in the bladder epithelium, which were composed of spindle fibers
proliferated around transitional epithelial cells accompanied by inflammatory infiltration of
submucosal bladder epithelium. Squamous metaplasia was noted on the surface of the polyps,
which were found at different loci than the bladder tumors.
2 Blood that presents in such small quantities that it is detectible only by chemical tests or by spectroscopic or
microscopic examination.
DRAFT - DO NOT CITE OR QUOTE
24
-------
Table 4-4. Incidences of urinary bladder lesions in male and female F344
rats exposed to biphenyl in the diet for 2 years
Males (n = 50)
Females (n = 50)
Dietary concentration (ppm)
0
500
1,500
4,500
0
500
1,500
4,500
TWA body weight (kg)a
0.411
0.412
0.408
0.357
0.251
0.246
0.246
0.216
Calculated dose (mg/kg-d)b
0
36.4
110
378
0
42.7
128
438
Lesion
Transitional cell
Simple hyperplasia0
0
0
0
12*
0
0
1
1
Nodular hyperplasia0
0
0
0
40*
1
0
0
5
Papillary hyperplasia0
0
0
0
17*
0
0
0
4
Combined hyperplasia
0
0
0
45**
1
0
1
10**
Papilloma
0
0
0
10*
0
0
0
0
Carcinoma
0
0
0
24*
0
0
0
0
Papilloma or carcinoma (combined)
0
0
0
31**
0
0
0
0
Squamous cell
Metaplasia0
0
0
0
19*
0
0
0
4
Hyperplasia0
0
0
0
13*
0
0
0
1
Papilloma or carcinoma (combined)
0
0
0
1
0
0
0
0
Inflammatory polyp0
0
0
0
10*
0
0
0
0
Calculi
0
0
0
43**
0
0
0
8*.
aTWA body weight calculated using graphically-presented body weight data in Umeda et al. (2002).
Calculated doses based on calculated TWA body weights and chronic reference food consumption values for
F344 rats (0.030 and 0.021 kg/day for males and females, respectively; taken from Table 1-6 of U.S. EPA (1988).
The number is the sum of animals with severity grades of slight, moderate, marked, or severe.
Statistically significant (Fisher's exact test, p < 0.05) as reported by study authors.
"Statistically significant (Fisher's exact test, p < 0.05) as determined by EPA.
Source: Umeda et al. (2002)
1
2 Table 4-5 summarizes the incidences of lesions of the ureter and kidney in the male and
3 female rats. The incidence of simple transitional cell hyperplasia in the ureter was greater in the
4 4,500 ppm males than the 4,500 ppm females. Other responses, such as mineralization of the
5 corticomedullary junction, were increased over controls to a greater extent in males compared to
6 females. In the renal pelvis, the incidence of simple and nodular hyperplasia showed a dose-
7 related increase in males and females. Treatment-related increases in the incidence of papillary
8 necrosis, infarct, and hemosiderin deposition occurred predominantly in exposed females.
9
25
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 4-5. Incidences of ureter and kidney lesions in male and female
F344 rats exposed to biphenyl in the diet for 2 years
Males (n = 50)
Females (n = 50)
Dietary concentration (ppm)
0
500
1,500
4,500
0
500
1,500
4,500
Calculated dose (mg/kg-d)
0
36.4
110
378
0
42.7
128
438
Lesion
Ureter
Transitional cell simple hyperplasia
1
0
0
8*
0
0
0
2
Transitional cell nodular hyperplasia
0
0
0
1
0
0
0
0
Dilatation
0
0
0
14*
0
0
0
6*.
Kidney
Renal pelvis
Transitional cell simple hyperplasia
6
8
5
19*
3
5
12*
25*
Transitional cell nodular hyperplasia
0
1
1
21*
0
0
1
12*
Squamous metaplasia
0
0
0
2
0
0
0
0
Mineralization
9
6
10
18
12
12
18
27*
Desquamation
1
0
0
11*
0
0
0
2
Calculi
0
0
0
13*
0
0
0
3
Other
Mineralization of corticomedullary
junction
0
0
0
10*
21
2*.
26
18
Mineralization of papilla
9
9
14
23*
2
6
3
12*
Papillary necrosis
0
0
0
0
0
0
23*
Infarct
0
0
0
0
1
0
0
8*
Hemosiderin deposits
0
0
0
0
4
8
22*
25*
Chronic nephropathy
45
45
43
34
33
35
30
26
Statistically significant (x2 or Fisher's exact test, p < 0.05) as reported by study authors.
"Statistically significant (Fisher's exact test, p < 0.05) as determined by EPA.
Source: Umeda et al. (2002).
In summary, the chronic toxicity and carcinogenicity study of male and female F344 rats
administered biphenyl in the diet for 2 years (Umeda et al.. 2002) provides evidence for
biphenyl-induced bladder tumors in males, but not females, based on the development of
transitional cell papillomas and carcinomas in the 4,500 ppm (378 mg/kg-day) males (Table 4-4).
This study identified a no-observed-adverse-effect level (NOAEL) of 500 ppm (42.7 mg/kg-day)
and a lowest-observed-adverse-effect level (LOAEL) of 1,500 ppm (128 mg/kg-day) for
nonneoplastic kidney lesions in female F344 rats exposed to biphenyl in the diet for 2 years.
The chronic toxicity of biphenyl was assessed in Wistar rats (50/sex/group) administered
the chemical at 0, 2,500, or 5,000 ppm in the diet for up to 75 weeks (Shiraiwa et al..(1989). The
rats were observed daily for clinical signs. Body weight and food consumption were measured
weekly. At death or scheduled sacrifice, gross pathologic examinations were performed and all
26
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
organs were removed and preserved. Other than body weight and biphenyl consumption data,
the published results of this study were limited to kidney weight data and findings related to
urinary calculi formation. Based on reported values for mean daily biphenyl intake (mg
biphenyl/rat) and mean initial and final body weights for each study group, doses of biphenyl at
the 2,500 and 5,000 ppm dietary levels are estimated to have been 165 and 353 mg/kg-day for
males, respectively, and 178 and 370 mg/kg-day for females, respectively.
Mean final body weights in both 2,500 and 5,000 ppm groups of biphenyl-exposed male
and female rats were significantly lower (by approximately 15 and 25 %; p<0 .01) than their
respective controls. Absolute and relative kidney weights of control and biphenyl-exposed rats
were similar, with the exception of significantly increased (p < 0.001) mean relative kidney
weight in 2,500 ppm female rats. The study authors reported the occurrence of hematuria in both
the 2,500 and 5,000 ppm groups as early as week 16 and stated that it was more recognizable at
60 weeks (Shiraiwa et al., 1989). Kidney stone formation was reported in 6/46 and 1/43 of the
2,500 ppm males and females, respectively, and in 19/47 and 20/39 of the 5,000 ppm males and
females, respectively. Detection of stones in other regions of the urinary tract was essentially
limited to the 5,000 ppm groups and included the ureter (2/47 males and 2/39 females) and
urinary bladder (13/47 males and 6/39 females). Kidney stones were hard, black, and located
from the pelvic area to the medullary region. Investigators described the stones in the ureter as
hard, black, and composed of protein. Stones in the urinary bladder were described as hard,
yellowish-white, round to oval in shape, and composed of ammonium magnesium phosphate.
Kidneys with stones exhibited obstructive pyelonephritis accompanied by hemorrhage,
lymphocytic infiltration, tubular atrophy, cystic changes of tubules, and fibrosis. Urinary
bladders with stones exhibited simple or diffuse hyperplasia and papillomatosis of the mucosa;
neoplastic lesions were not seen following 75 weeks of exposure. No control rats (44 males and
43 females) showed stones in the kidney, ureter, or urinary bladder. The lowest exposure level
in this study, 2,500 ppm in the diet for 75 weeks, was a LOAEL for formation of kidney stones
associated with pyelonephritis in Wistar rats (dose levels of 165 and 178 mg/kg-day for males
and females, respectively).
Shiraiwa et al. (1989) also reported the results of an initiation-promotion study in male
Wistar rats (25/group) that included three groups administered a basal diet for 2 weeks followed
by diets containing 0, 1,250, or 5,000 ppm biphenyl for 34 weeks. Three other groups received
diets containing 0.1% N-ethyl-N-hydroxyethylnitrosamine (EHEN, an initiator of kidney tumors
in rats) for 2 weeks followed by diets containing 0, 1,250, or 5,000 ppm biphenyl for 34 weeks.
Initial and final body weights were recorded. At terminal sacrifice, gross pathologic
examinations were performed. The study report included information regarding kidney weights,
but did not indicate whether weights of other organs were measured. Kidney and urinary bladder
were fixed; kidneys were sectioned transversely (10-12 serial slices) and urinary bladders were
cut into 4-6 serial slices. The authors used a computer-linked image analyzer to determine the
27
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
incidence of kidney lesions and dysplastic foci. The presence of stones in the kidney and urinary
bladder was assessed qualitatively using an infrared spectrophotometer. Based on reported
values for mean daily biphenyl intake (mg biphenyl/rat) and average body weight (mean initial
body weight + one-half the difference between mean initial and mean final body weight) for each
study group, doses of biphenyl at the 1,250 and 5,000 ppm dietary levels are estimated to have
been 59.3 and 248.3 mg/kg-day, respectively, for rats on basal diet alone for the first 2 weeks
and 62.0 and 248.2 mg/kg-day, respectively, for rats receiving EHEN in the diet for the first 2
weeks.
The mean final body weight of the rats receiving basal diet followed by diet containing
5,000 ppm biphenyl was significantly lower (p < 0.001) than that of controls (0.389 ± 22 versus
0.432 ± 30 kg). Relative kidney weights were increased in this group of biphenyl-exposed rats
compared to the basal diet control group (actual data were not presented). Stones were detected
only in the rats receiving 5,000 ppm biphenyl in the diet; incidences were 4/25 (kidney), 1/25
(ureter), and 3/25 (urinary bladder) in rats that had received that basal diet for the first 2 weeks.
Similar results regarding final body weight and the detection of stones in the urinary tract were
reported for the rats that had received EHEN in the diet prior to the administration of biphenyl.
Incidences of dysplastic foci and renal cell tumors were determined in the kidneys of all groups
of rats. Only rats that had received EHEN during the initial 2 weeks exhibited neoplastic kidney
lesions (dysplastic foci, renal cell tumors). For the EHEN + 0 ppm biphenyl, EHEN + 1,250
ppm biphenyl, and EHEN + 5,000 ppm biphenyl groups, incidences of rats with dysplastic foci
were 25/25, 21/25, and 25/25, respectively, and incidences of rats with renal cell tumors were
13/25, 12/25, and 7/25, respectively. Under the conditions of this study, biphenyl did not exhibit
tumor promoting characteristics for the kidney tumor initiator, EHEN.
Weanling albino rats (15/sex/group) were administered biphenyl in the diet at
concentrations of 0, 10, 50, 100, 500, 1,000, 5,000, or 10,000 ppm for 2 years (Ambrose et al.,
(1960). Based on U.S. EPA (1988) reference values for body weight and food consumption in
F344 rats (averages of values for males and females), these concentrations corresponded to
"3
estimated doses of 1, 4, 8, 42, 84, 420, and 840 mg/kg-day, respectively . Body weights were
monitored every week during the period of active growth and then at 50-day intervals.
Hemoglobin was monitored every 100 days in control and high-dose rats; at 500, 600, and
700 days in rats receiving 5,000 ppm biphenyl, and at 500 and 600 days in rats receiving 1,000
ppm biphenyl. A 98-day paired-feeding experiment was conducted in which control rats were
provided the same amount of food that rats of the 5,000 and 10,000 ppm biphenyl groups
consumed to assess whether possible differences in growth would indicate a biphenyl exposure-
related toxicological response or decreased palatability. At necropsy, liver, kidney, heart, and
testes weights were recorded for all groups except those receiving 10,000 ppm biphenyl in the
3There is greater uncertainty in the dose estimates at the two highest exposure levels because the magnitude of
reported decreased food consumption in these groups was not specified in the study report.
28
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
diet. Tissues from major organs (heart, lung, liver, kidney, adrenal, spleen, pancreas, stomach,
intestine, bladder, thyroid, brain, pituitary, and gonads) were examined histopathologically. In
some cases, bone marrow smears were prepared. Except for one rat sacrificed prior to
termination, necropies were performed only on terminal sacrifice animals (males: n = 2-13
rats/group; females: n = 2-11 rats/group).
Survival was decreased in male and female rats of the 5,000 and 10,000 ppm biphenyl
exposure groups, but was not evident at lower exposure levels. Growth rates appeared similar
among controls and groups exposed to biphenyl levels <1,000 ppm. At the two highest exposure
levels, decreased growth ranged from 8% to 48% compared to control, but was attributable to
decreased food consumption and indicative of decreased palatability based on results of the
paired-feeding experiment. Decreased hemoglobin levels were reported in male and female rats
of the two highest exposure levels after 300-400 and 500-600 days, respectively, but were
considered at least partially related to lower food consumption in these groups relative to
controls. Selected organ weights are summarized in Table 4-6. There were no statistically
significant treatment-related effects on organ weights at dietary levels <1,000 ppm, levels below
those associated with decreases in food consumption, body weight, and survival (i.e., 5,000 and
10,000 ppm). Relative liver and kidney weights of female rats of the 5,000 ppm biphenyl
exposure group were significantly (p < 0.05) increased, approximately 45 and 215% higher than
those of respective controls. The only significant compound-related histopathological change
occurred in the kidneys, which, in all rats of the two highest exposure groups, showed irregular
scarring, lymphocytic infiltration, tubular atrophy, and tubular dilation associated with cyst
formation. Some evidence of hemorrhage was present, and calculi were frequently noted in the
renal pelvis. The authors concluded that there was no compound-related increase in tumor
incidence. Bladder tumors were reported in male rats in most groups (controls-2/9; 10 ppm-2/8;
100 ppm-1/9; 1,000 ppm-1/9; 5,000 ppm-1/2; and 10,000 ppm-1/2) and female control rats
(1/9). However, because histopathological examination was limited to terminal sacrifice animals
and survival was especially low in the two highest dose groups at 13—33%, this study was not
adequate to evaluate the potential for biphenyl to induce tumors. The study identified a NOAEL
of 1,000 ppm biphenyl in the diet (84 mg/kg-day) and a LOAEL of 5,000 ppm (420 mg/kg-day)
for kidney effects including tubular atrophy and dilation associated with cyst formation and
calculi formation in the renal pelvis of albino rats of both sexes.
29
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 4-6. Body and organ weight data for male and female rats
administered biphenyl in the diet for 2 years
Biphenyl in diet
(ppm)
Days on
diets
Number
of rats
Mean body weight
(g)±SE
Mean relative organ weight (g) ± SE
Liver
Kidneys
Heart
Testes
Males
0
745
9
396 ±24.6
2.89 ±0.16
0.75 ± 0.02
0.32 ±0.015
0.72 ±0.03
10
744
8
424 ±5.1
2.66 ± 0.06
0.70 ±0.03
0.28 ±0.008
0.62 ± 0.07
50
747
10
383 ± 19.8
2.84 ±0.15
0.73 ± 0.02
0.30 ±0.01
0.56 ±0.06
100
752
11
394 ± 14.2
2.47 ± 0.07
0.72 ±0.01
0.31 ±0.008
0.67 ± 0.07
500
730
13
371 ± 15.8
3.03 ±0.12
0.74 ± 0.02
0.31 ±0.007
0.65 ± 0.06
1,000
746
10
366 ±23.7
2.98 ±0.19
0.83 ±0.05
0.34 ±0.012
0.60 ±0.08
5,000
746
2
345
3.12
1.17
0.36
0.36
Females
0
745
9
333 ±9.4
3.11 ± 0.15
0.65 ±0.01
0.33 ±0.01
NA
10
744
6
369 ± 13.4
3.21 ±0.17
0.62 ± 0.02
0.28 ± 0.07
NA
50
747
5
335 ± 16.6
2.81 ±0.28
0.64 ± 0.02
0.31 ±0.03
NA
100
752
11
341 ±9.1
3.46 ±0.74
0.62 ± 0.02
0.30 ±0.01
NA
500
730
5
306 ± 12.5
3.51 ±0.12
0.68 ± 0.02
0.31 ±0.01
NA
1,000
746
5
327 ±6.8
3.18 ± 0.10
0.65 ±0.01
0.32 ±0.01
NA
5,000
746
5
226 ±25.8
4.52 ±0.20*
1.39 ±0.14*
0.46 ± 0.04
NA
Statistically significant (Student's t-test, p < 0.05) as reported by study authors.
NA = not applicable; SE = standard error of the mean
Source: Ambrose et al. (19601.
Male albino rats (8/group; strain not stated) were given biphenyl in the diet for up to
13 months at concentrations resulting in estimated doses of 250 or 450 mg/kg-day (Pecchiai and
Saffiotti, (1957). Upon sacrifice, liver, kidney, spleen, heart, lung, thyroid, parathyroid, adrenal,
pancreas, testis, stomach, and intestine were processed for histopathological examination. At 2-
month interim sacrifices, moderate degenerative changes in liver and kidney were observed at
both dose levels. Liver effects consisted of moderate degeneration and hypertrophy of the
Kupffer cells with a generally well-preserved structure. Renal glomeruli were undamaged, but
tubuli showed mild signs of degeneration. The liver and kidney effects did not appear to
increase in severity in rats treated for up to 13 months. Other histopathologic effects noted in the
biphenyl-treated rats included hypertrophied splenic reticular cells, small follicles with sparse
colloid and desquamation of follicular epithelium in the thyroid, and hyperplastic and
hyperkeratinized forestomach epithelium with occasional desquamation. The study authors
reported neoplastic lesions in the forestomach of three biphenyl-treated rats. Two of the rats
exhibited papillomas of the forestomach epithelium (one after 7 weeks and one after 7 months of
treatment); a squamous cell carcinoma was diagnosed in the other rat after 1 year of treatment.
The study authors noted two sequential responses to chronic biphenyl exposure: degenerative
30
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
changes of nuclei and cytoplasm in the parenchyma of liver and kidney, spleen, thyroid, and
adrenals within 2 months followed within 1 month or more by functional-regenerative changes
that resulted in hyperplasia and nuclear hypertrophy of liver and kidney parenchyma as well as
functional hyperactivity of the thyroid and parathyroid. Irritation and hyperplasia were evident
in the lower urinary tract. The lowest dose, 250 mg/kg-day biphenyl, was an apparent LOAEL
for nonneoplastic degenerative changes in the liver, kidney, thyroid, and parathyroid of male
albino rats resulting in hyperplasia of liver, kidney, and thyroid. Overall, this study was too
limited in duration (13-month exposure) and group size for use in evaluating the carcinogenicity
of biphenyl in rats.
Sprague-Dawley rats (12/sex/group) were exposed to biphenyl in the diet for 2 years at
exposure levels of 0, 100, 1,000, or 10,000 ppm (Dow Chemical Co., (1953). Based on U.S.
EPA (1988) chronic reference values for body weight and food consumption in Sprague-Dawley
rats (average values for combined sexes), these dietary levels are estimated to correspond to
doses of 7, 73, and 732 mg/kg-day, respectively. Body weights were monitored twice weekly for
3 months, then weekly. Blood samples were taken from all animals at the start of the
experiment, approximately every 3 months thereafter, and at term. Hemoglobin levels, red and
white blood cell counts and differential cell counts, and BUN concentrations were recorded. At
death or scheduled necropsy, organ weights were recorded for liver, lung, kidneys, heart, and
spleen. Sections from heart, liver, kidney, spleen, adrenals, pancreas, gonads, stomach, small
and large intestine, voluntary muscle, lung, bladder, and brain were fixed and stained for
histopathologic examination. An outbreak of pneumonia affected the colony during the course
of the experiment.
Survival was poor in control males, all of which had died by 18 months. Only two of the
females receiving 1,000 ppm biphenyl in the diet survived to the end of the 21st month, and none
rd
had survived by the end of the 23 month. The authors considered the decreased survival in this
group of females to have been compound-related. Eight to 30% of biphenyl concentration-
related reductions in body weight gain were observed among the groups, although, in monitoring
food efficiency (data not provided in report), the authors indicated that the reduced growth was
likely due to a lower daily consumption of food rather than to biphenyl toxicity. There were no
clear indications of exposure-related changes in hematological parameters. The authors reported
significant (p < 0.05) increases in average (combined sexes) relative liver and kidney weights at
the highest exposure level, compared with control values (4.71 versus 3.05 g/100 g and 1.68
versus 1.00 g/100 g, respectively). Tubular dilatation was evident in controls as well as treated
animals, but increased in severity with dose (measured on a scale of 0-4). Among the controls,
low-, mid, and high-dose rats, incidences for tubular dilatation with severity scores >2 were 1/12,
6/12, 7/12, and 11/12 for males and 1/12, 3/12, 4/12, and 11/12 for females, respectively.
Incidences of tubular dilatation with severity scores >3 were 0/12, 1/12, 2/12, and 9/12 for males
and 1/12, 2/12, 2/12, and 11/12 for females, respectively. Calcification and intratubular
31
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
inflammation were frequently observed in high-dose rats. The study identified a LOAEL of
1,000 ppm in the diet (732 mg/kg-day) for renal effects (renal tubular dilatation with a severity
score >3) in Sprague-Dawley rats and a NOAEL of 100 ppm biphenyl (73 mg/kg-day). The
small number of rats in the exposure groups and the decreased survival at the highest exposure
level may have impaired the ability to detect late-developing tumors in this study.
4.2.1.2.2. Chronic mouse studies
In a chronic toxicity and carcinogenicity study of BDFi mice (50/sex/group) conducted
by JBRC, biphenyl was administered in the diet for 2 years at concentrations of 0, 667, 2,000 or
6,000 ppm corresponding to doses of 97, 291, and 1,050 mg/kg-day in the males and 134, 414,
and 1,420 mg/kg-day in the females (Umeda et al., 2005). All animals were observed daily for
clinical signs and mortality. Body weights and food consumption were recorded weekly for the
first 14 weeks and every 4 weeks thereafter. Hematological and clinical chemistry parameters
were measured in blood samples drawn from all 2-year survivors just prior to terminal sacrifice.
At death or terminal sacrifice, gross pathological examinations were performed and organs were
removed and weighed. Specific tissues prepared for microscopic examination were not listed in
the study report, but included liver and kidney.
There were no overt clinical signs or effects on food consumption or survival among
biphenyl-exposed mice of either sex compared to controls. Mean terminal body weights showed
a dose-related decrease; body weights were significantly less than those of controls at 2,000 and
6,000 ppm (males: 46.9, 43.1, 42.9, and 32.4 g; females: 34.0, 32.5, 30.5, and 25.5 g, at 0, 667,
2,000, and 6,000 ppm, respectively).
Although there were no compound-related changes in hematological parameters, some
clinical chemistry parameters showed marked changes in relation to dose, including a biphenyl
dose-related increase in BUN that achieved statistical significance in 6,000 ppm males and
females and 2,000 ppm males. In female mice, dose-related increases in activities of the plasma
enzymes AP, lactate dehydrogenase (LDH), glutamate oxaloacetate transaminase (GOT; also
referred to as AST), and glutamate pyruvate transaminase (GPT; also referred to as ALT) (see
Table 4-7) suggested effects of biphenyl on the liver. Umeda et al. (2005) noted that females
with malignant liver tumors exhibited extremely high AST, ALT, and LDH activities. In
general, biphenyl did not induce dose-related changes in liver enzymes in male mice, although
AP activity was significantly greater than controls in 6,000 ppm males (Table 4-7).
32
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 4-7. Dose-related changes in selected clinical chemistry values from
male and female BDFi mice exposed to biphenyl via the diet for 2 years
Males
Biphenyl dietary
concentration (ppm)
0
667
2,000
6,000
Dose (mg/kg-d)
0
97
291
1,050
Endpoint (mean ± SD)
n = 34
n = 39
r-
II
a
r-
II
a
AST (IU/L)
85 ±92
58 ±38
69 ±60
88 ± 151
ALT (IU/L)
73 ± 113
34 ±31
36 ±49
43 ±80
AP (IU/L)
178± 111
155 ±30
169 ±36
261 ± 102*
LDH (IU/L)
321 ±230
252 ±126
432 ± 868
283 ± 200
BUN (mg/dL)
20.2 ±3.6
22.0 ±4.0
23.2 ±4.4*
22.9 ±2.7*
Females
Biphenyl dietary
concentration (ppm)
0
667
2,000
6,000
Dose (mg/kg-d)
0
134
414
1,420
Endpoint (mean ± SD)
n = 28
o
II
a
II
a
n = 31
AST (IU/L)
75 ±27
120 ±110
211 ± 373*
325 ± 448*
ALT (IU/L)
32 ± 18
56 ±46
134 ±231*
206 ± 280*
AP (IU/L)
242 ± 90
256±121
428 ± 499
556 ± 228*
LDH (IU/L)
268 ± 98
461 ±452
838 ± 2,000
1,416 ±4,161*
BUN (mg/dL)
14.9 ±2.0
14.8 ±3.4
21.0 ±20.5
23.8 ± 11.7*
Statistically significant (Dunnett's test, p < 0.05) as reported by study authors.
Source: Umeda et al. (20051.
The only apparent exposure-related effect on organ weights was 1.3-, 1.4-, and 1.6-fold
increases in relative liver weights of 667, 2,000, and 6,000 ppm female mice, respectively (the
liver weight data were not presented in Umeda et al. (2005)1). Gross pathologic examinations
revealed biphenyl dose-related increased incidences of liver nodules in females, but not males
(Table 4-8). The nodules were round- or oval-shaped cystic or solid masses (~ 3-23 mm in
diameter). Histopathological examinations revealed that 5, 16, and 19 of the nodule-bearing 667,
2,000, and 6,000 ppm female mice also exhibited proliferative lesions of hepatocellular origin
(Table 4-8). Significantly increased incidences of basophilic cell foci were observed in 2,000
and 6,000 ppm female mice. The incidence of basophilic cell foci was significantly increased in
667 ppm male mice but not in 2,000 or 6,000 ppm males compared to controls. Peto's trend tests
confirmed significant positive trends for dose-related increased incidences of hepatocellular
adenomas (p < 0.05) and combined incidences of hepatocellular adenomas or carcinomas (p <
0.01). Incidences of hepatocellular carcinomas were significantly increased in 2,000 ppm
females, but not 667 or 6,000 ppm females. However, Umeda et al. (2005) noted that the
incidences of hepatocellular carcinomas (-5/50 or 10%) in each of the 667 and 6,000 ppm
groups of females exceeded the range of historical control data for that laboratory
33
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(26 hepatocellular carcinomas in 1,048 female mice [2.5% incidence in 21 bioassays; maximum
incidence of 8%]). Liver tumor incidences in male mice showed a statistically significant
decrease with increasing dose; however, the incidences were within the range of historical
control data for adenomas or carcinomas in male mice (10-68%; see Table 4-8), and may reflect
the dose-related decrease in body weight in this study (e.g., Leakey et al., 2003). Investigators
reported statistically significantly increased incidences of desquamation of the urothelium in the
renal pelvis in 6,000 ppm male and female mice, and mineralization in the inner stripe of the
outer medulla of the kidney in 2,000 and 6,000 ppm female mice.
Table 4-8. Incidences of gross and histopathological findings in male and
female BDFi mice fed diets containing biphenyl for 2 years
Parameter
Dietary concentration of biphenyl (ppm)
Males
Females
0
667
2,000
6,000
0
667
2,000
6,000
Average dose (mg/kg-d)
0
97
291
1,050
0
134
414
1,420
Necropsy
Liver nodules
20/50
16/49
14/50
11/50
7/50
13/50
24/50**
26/49**
Histopathology
Liver3
Adenoma
8/50
6/49
7/50
3/50
2/50
3/50
12/50*
10/49*
Carcinoma
8/50
8/49
5/50
4/50
1/50
5/50
7/50*
5/49
Adenoma or carcinoma
(combined)
16/50
12/49
9/50
7/50**
3/50
8/50
16/50*
14/49*
Basophilic cell foci
0/50
6/49*
1/50
2/50
1/50
1/50
12/50*
6/49*
Clear cell foci
0/50
6/49*
2/50
0/50
2/50
1/50
3/50
2/49
Eosinophilic cell foci
0/50
0/49
0/50
0/50
0/50
1/50
0/50
0/49
Kidney
Desquamation: pelvis
0/50
0/49
0/50
10/50*
4/50
0/50
0/50
15/49*
Mineralization inner stripe-
outer medulla
9/50
8/49
14/50
14/50
3/50
5/50
12/50*
26/49*
"Historical control data for hepatocellular tumors: Male BDF, mouse: adenoma—17.2% (4-34%), carcinoma—
18.8% (2-42%), adenoma/carcinoma—32.2% (10-68%). Female BDF, mouse: adenoma—4.8% (0-10%),
carcinoma—2.5% (0-8%), adenoma/carcinoma—7.1% (2-14%). Source: email dated July 25, 2011, from Yumi
Umeda, JBRC, to Connie Kang, NCEA, ORD, U.S. EPA.
Statistically significant (Fisher's exact test, p < 0.05) as reported by study authors.
"Statistically significant (Fisher's exact test, p < 0.05) as determined by EPA.
Source: Umeda et al. (2005).
In summary, the chronic toxicity and carcinogenicity study of male and female BDFi
mice administered biphenyl in the diet for 2 years (Umeda et al.. 2005) provides evidence for
biphenyl-induced liver tumors in females, but not males, based on significantly increased
incidences of hepatocellular adenomas and combined carcinomas or adenomas in the female
mice receiving biphenyl from the diet (Table 4-8). This study identified a NOAEL of 134
34
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
mg/kg-day and a LOAEL of 414 mg/kg-day for nonneoplastic effects (mineralization in the
kidney and significantly increased plasma ALT and AST activities) in female BDFi mice
exposed to biphenyl in the diet for 2 years.
Groups of female ddY mice (n = 60) were fed diets containing 0 or 5,000 ppm biphenyl
in the diet for 2 years (Imai et al., 1983) . Food consumption, body weights, and survival were
assessed at intervals throughout exposure. At terminal sacrifice, several organs were weighed
(9-11/group). The following organs were examined for histopathological changes, in 34-37
mice/group: brain, pituitary, thymus, liver, spleen, pancreas, lung, heart, adrenal, kidney, ovaries,
uterus, thyroid, stomach, small intestine, and large intestine. Urine and blood samples were
collected from mice (6-12/group) at terminal sacrifice and were analyzed for urinalysis,
hematological, and serum chemistry endpoints. Based on estimated food consumption rates
(U.S. EPA, 1988) and reported average terminal body weight (0.037 kg), the dose corresponding
to a diet of 5,000 ppm is estimated to be 855 mg/kg-day.
Exposure to biphenyl did not influence survival, food consumption, or growth compared
with controls. No marked exposure-related effects were found on terminal organ and body
weights or on the urinalytic, hematologic, or serum chemistry endpoints. Histological
examination revealed no increased incidence of nonneoplastic lesions in examined tissues in the
5,000 ppm biphenyl group, compared with the control group. The only tissues showing tumors
at elevated incidence in the 5,000 ppm mice, compared with the control group, were the lung
(11/34 [32.4%] versus 9/37 [24.3%] in controls) and lymphatic tissues (lymphomas: 5/34
[14.7%>] versus 4/37 [10.8%>]; leukemia: 3/34 [8.8%>] versus 2/37 [5.4%>];p > 0.05 by Fisher's
exact test). At the same time, the lack of histopathological information concerning
approximately 40% of the animals on test increases the uncertainty of these results. In summary,
5,000 ppm biphenyl in the diet of female ddY mice for 2 years was a NOAEL for non-neoplastic
lesions, survival, body and organ weight changes, and changes in urinalytic, hematologic, and
serum chemistry endpoints. No carcinogenic response was associated with exposure to
5,000 ppm biphenyl in the diet (estimated dose of 855 mg/kg-day) for 2 years in female ddY
mice (Imai et al.. 1983).
The carcinogenic potentials of 130 chemicals, including biphenyl, were assessed in a
protocol that exposed groups of two strains of F1 hybrid mice (18/sex/strain/group), produced by
mating female C57BL/6 mice to either male C3H/Anf mice (F1 generation: strain B6C3F1,
designated by study authors as strain A) or male AKR mice (F1 generation: strain B6AKF1,
designated as strain B) to individual chemicals by the oral route for 18 months (NCI, 1968).
(The study was subsequently published as Innes et al. [1969], but detailed results for biphenyl
were not included in that publication.) Four groups of untreated controls and a group of gelatin
vehicle controls (18/sex/strain/group) were included in the study. In the case of biphenyl, the
chemical was administered via gavage to mice for 3 weeks, starting at the age of 7 days at 215
mg biphenyl/kg body weight in 0.5% gelatin. Thereafter, and for the rest of the experimental
35
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
period, biphenyl was mixed with chow to a final concentration of 517 ppm. The gavage dose
level and food concentration of biphenyl were selected to reflect the maximum tolerated dose
identified in preliminary range-finding, single-dose subcutaneous injection and single- and
repeated-dose oral administration studies. Initial gavage dose and dietary levels of biphenyl
were not adjusted for weight gain during the 18-month study. Based on U.S. EPA (1988)
chronic reference values for body weight and food consumption in strain A mice (average values
for combined sexes), a time-weighted average oral dose of 91 mg/kg-day is estimated from the
dietary exposure. Blood smears were prepared from mice that showed splenomegaly, liver
enlargement, or lymph adenopathy at necropsy. At term, mice were examined for any gross
pathological features. Major organs were processed for histopathologic examination (including
total chest contents, liver, spleen, kidneys with adrenals, stomach, and genital organs).
Incidences of hepatomas, pulmonary tumors, and sarcomas in control mice and biphenyl -
treated mice are summarized in Table 4-9. Although, there were no statistically significant
increases in hepatoma or pulmonary tumor incidence, it should be noted that the study duration
of 18 months would tend to underestimate incidences associated with 24-month exposures. EPA
found only the reticular cell sarcoma incidence was significantly elevated in strain B female mice
but not in male mice of this strain or strain A mice of either sex. The origin of this kind of
neoplasm is uncertain as three different stromal cells (follicular dendritic cells, interdigitating
reticular cells, and interfollicular fibroblastic reticular cells) could give rise to reticular cell
sarcoma, and special staining is needed to differentiate (Jones et al., 2001). This pathology term
is not considered specific because no information on differential diagnosis was provided in the
NCI (1968) report. Interpretation of the biological significance of this tumor type may also be
influenced by the early-life exposure in this study, starting at 1 week of age.
36
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 4-9. Incidences of selected tumor types among controls and mice
administered biphenyl orally for 18 months
Group
Incidences of selected tumor types"
Hepatoma
Pulmonary tumors
Reticular cell sarcoma
C57BL/6 x C3H/Anf (B6C3F1 or "strain A") male mice
Controls
8/79 (10.1%)
5/79 (6.3%)
5/79 (6.3%)
Biphenyl-treated
2/17 (11.8%)
3/17 (17.7%)
1/17 (5.9%)
C57BL/6 x C3H/Anf (B6C3F1 or "strain A") female mice
Controls
0/87 (0%)
3/87 (3.4%)
4/87 (4.6%)
Biphenyl-treated
0/18 (0%)
1/18 (5.6%)
0/18 (0%)
C57BL/6 x AKR (B6AKF1 or "strain B") male mice
Controls
5/90 (5.6%)
10/90(11.1%)
1/90(1.1%)
Biphenyl-treated
3/17 (17.6%)
1/17 (5.9%)
0/17 (0%)
C57BL/6 x AKR (B6AKF1 or "strain B") female mice
Controls
1/82 (1.2%)
3/82 (3.7%)
4/82 (4.9%)
Biphenyl-treated
0/17 (0%)
0/17 (0%)
4/17 (23.5%)*
aTumor incidences were tallied from those mice for which histopathologic^ examinations were performed.
Statistically significant (Fisher's exact test, p < 0.05) as determined by EPA.
Source: NCI (19681.
4.2.2. Inhalation Studies
In three separate experiments, albino rabbits (sex and strain not stated), Sprague-Dawley
rats (sex not stated), and mice (sex and strain not stated) were repeatedly exposed to dusts
composed of 50% biphenyl attached to celite for 7 hours/day, 5 days/week (Deichmann et al.,
(1947); Monsanto,(1946). In the first experiment, 3 rabbits and 10 rats were exposed to an
"3
average concentration of 300 mg/m on each of 64 days over a period of 94 days. The rats
exhibited irritation of the nasal mucosa accompanied by serosanguineous discharge. Five of the
rats died prior to term, and the survivors lost weight. The rabbits exhibited no exposure-related
adverse signs. In the second experiment, three rabbits and six rats were exposed to an average
"3
concentration of 40 mg/m on each of 46 days over a total period of 68 days. One rat died prior
to term. The surviving rats showed signs of mucous membrane irritation, but appeared to gain
weight at a normal rate. The rabbits exhibited no exposure-related adverse signs. In the third
"3
experiment, 12 mice and 4 rats were exposed to an average concentration of 5 mg/m on each of
62 days over a total period of 92 days. While the rats were unaffected at this concentration, all
of the mice showed signs of irritation of the upper respiratory tract and two died prior to term.
Bronchopulmonary lesions (including acute emphysema, congestion, edema, bronchitis,
widespread lobular pneumonia, and multiple pulmonary abscesses) were reported in rats from
experiments 1 and 2 and in mice from experiment 3. Some unspecified minor liver and kidney
37
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
"3
lesions were also noted. Based on the results of these three experiments, a LOAEL of 5 mg/m
"3
in mice and a LOAEL of 40 mg/m in rats for upper respiratory tract irritation were identified.
Groups of CD-I mice (50/sex/group) were exposed to airborne biphenyl at vapor
"3
concentrations of 0, 25, or 50 ppm (0, 157.7, and 315.3 mg/m , respectively) for 7 hours/day,
5 days/week for 13 weeks (Sun Company Inc., 1977a). Mice were maintained and exposed to
biphenyl in groups of 5 (for a total of 10 groups/sex/exposure group). All animals were checked
daily for clinical signs and mortality, and body weight data were collected. Upon completion of
the 13-week exposure period, surviving mice were placed in metabolic cages for 12-hour
collection of urine for urinalysis. Blood samples were collected for blood chemistry and
hematology assessments. Gross and histopathologic examinations were performed on all mice.
Ten surviving mice/sex/group were held for a 30-day recovery period prior to terminal sacrifice.
During the first few days of biphenyl exposure, some of the test material crystallized in
the delivery system; analysis of biphenyl exposure levels was not performed on these days.
Daily measured biphenyl exposure concentrations were highly variable during the first half of
the 13-week exposure period, whereas subsequently measured concentrations were closer to
target concentrations. For example, during the first 45 exposure sessions, measured daily
biphenyl concentrations in the 50 ppm target groups ranged from as low as 5 ppm to as high as
102 ppm and subsequent measurements ranged from 48 to 55 ppm. Mean biphenyl
concentrations (± 1 SD) calculated for the entire 13 weeks of exposure were 25 ± 7 and
50 ± 16 ppm for the 25 and 50 ppm target groups, respectively. The authors reported the loss of
46 mice (40 males and 1 female at 25 ppm and 5 males at 50 ppm) due to overheating and
cannibalization. Since the overheating event occurred after 46 exposures, the overall study
duration ran for 117 days to ensure that replacement mice received a total of 65 exposures as
called for in the protocol. Body weights and results of urinalysis, hematology, and clinical
chemistry did not indicate any clear exposure-related changes that could be attributed to biphenyl
toxicity. Gross and histopathological examinations revealed congested and hemorrhagic lungs,
hyperplasia of the trachea with inflammation accompanied by a high incidence of pneumonia,
and congestion and edema in liver and kidney of biphenyl-exposed mice (Table 4-10). The
pathologist considered the congestion in the lung, liver, and kidney a likely effect of the
anesthetic used for killing the mice, although control mice did not exhibit these effects at 13-
week sacrifice. The hemorrhagic lungs and tracheal hyperplasia were considered effects of
biphenyl exposure. Results from the 30-day recovery groups suggest that the biphenyl exposure-
related pulmonary effects were reversible. This study identified a LOAEL of 25 ppm for
histopathologic lung, liver, and kidney lesions in male and female CD-I mice exposed to
biphenyl by inhalation for 7 hours/day, 5 days/week for 13 weeks.
38
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Table 4-10. Incidences of selected histopathological lesions in tissues of CD-
1 mice exposed to biphenyl vapors 7 hours/day, 5 days/week for 13 weeks
Effect
13-Week exposure groups"
0 ppm
25 ppm
50 ppm
Pulmonary congestion, edema
0/80
95/98
71/71
Pneumonia
0/80
15/98
20/71
Tracheal hyperplasia
0/80
80/98
70/71
Hepatic congestion, edema
0/80
87/98
71/71
Renal congestion, edema
0/80
87/98
71/71
aThe study report presented incidences of histopathological lesions for combined male and female mice only; no
statistical analyses were conducted.
Source: Sun Company Inc. (1977a).
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
4.3.1. Oral Exposure
Pregnant female Wistar rats (18-20/ group) were administered 0, 125, 250, 500, or
1,000 mg/kg-day biphenyl in corn oil by gavage on gestation days (GDs) 6-15 (with GD 1
defined as the the day the evidence of copulation was observed) (Khera et al., (1979). Body
weights of dams were recorded on GDs 1, 6-15, and 22, at which point all dams were sacrificed.
Parameters evaluated at necropsy included the number of corpora lutea, fetal weights and
viability, and resorptions; fetal sex was apparently not determined. Two-thirds of the live
fetuses/litter were examined for skeletal development and the rest were examined for the
presence of visceral abnormalities.
At 1,000 mg/kg-day, five of 20 high-dose dams died prior to sacrifice, and there was a
10% decrease from control in body weight in the remaining dams in that group (data not shown).
Doses <500 mg/kg-day produced no clinical signs of maternal toxicity or evidence of treatment-
related effects on maternal weight gain. The number of dams without live fetuses was
significantly increased at 1,000 mg/kg-day; of the surviving dams, five were found not pregnant
and one had seven resorption sites but no live fetuses (Table 4-11). Mean numbers of corpora
lutea and live fetuses per pregnancy in the remaining pregnant 1,000 mg/kg-day dams were
similar to those of controls and dams of other dose levels.
The incidence of anomalous fetuses and litters bearing anomalous fetuses, including
wavy ribs, extra ribs, missing and unossified sternebrae or delayed calvarium ossification,
generally increased with dose. When data from the high-dose (1,000 mg/kg-day) group were
dropped because of frank maternal toxicity at that dose, missing or unossified sternebrae was the
only endpoint that showed a statistically significant increasing trend with dose (Cochran-
Armitage test).
As noted in EPA's Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA,
39
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1991), a significant, dose-related increase in a variation (e.g., delayed ossification) should be
evaluated as a possible indication of developmental toxicity, although an assessment of the
biological significance of such variations should take into consideration knowledge of the
developmental stage, background incidence of certain variations, other strain- or species-specific
factors, and maternal toxicity. Other information that would help in interpreting the biological
significance of anomalies in Khera et al. (1979), however, were not available. In light of the
finding of a statistically significant increasing trend of missing or unossified sternebrae with dose
and consideration of this anomaly as more severe than the other anomalies identified, EPA
identified a LOAEL of 500 mg/kg-day for increased incidence of fetuses with missing and
unossified sternebrae and a NOAEL of 250 mg/kg-day.
Table 4-11. Prenatal effects following oral administration of biphenyl to
pregnant Wistar rats on GDs 6-15
Effect
Dose (mg/kg-d)
0
125
250
500
1,000
Rats without live fetuses at term/number mated
2/18
0/20
1/19
2/20
1 l/20a
Corpora lutea/pregnancy (mean ± SE)
12.6 ±0.4
12.9 ±0.4
13.7 ±0.5
13.3 ±0.4
12.5 ±0.7
Live fetuses/pregnancy (mean ± SE)
11.3 ±0.7
11.8 ± 0.6
11.9 ±0.6
11.2 ± 0.5
10.7 ± 1.3
Dead or resorbed fetuses (%)
4.8
3.3
6.1
7.8
13.7b
Fetal weight (g mean ± SE)
5.1 ±0.1
5.3 ±0.1
5.2 ±0.1
5.2 ±0.1
4.5 ±0.3
Anomalous fetuses/number examined
17/176
(9.7%)
22/236
(9.3%)
22/213
(10.3%)
35/199
(17.6%)
25/107
(23.4%)
Anomalous litters/number examined"
8/16 (50%)
11/20 (55%)
13/18 (72%)
15/18 (83%)
6/9 (67%)
Anomalies, number (percent) of fetuses affected
Wavy ribs, uni- and bilateral
3 (1.7%)
7 (3.0%)
9 (4.2%)
8 (4.0%)
5 (4.7%)
Extra ribs, uni- and bilateral
9(5.1%)
12 (5.1%)
9 (4.2%)
15 (7.5%)
6 (5.6%)
13th rib, small sized
1 (0.6%)
1 (0.4%)
2 (0.9%)
1 (0.5%)
0 (0.0%)
Sternebrae, missing or unossified"
4 (2.3%)
3 (1.3%)
4 (1.9%)
16 (8.0%)
17 (15.9%)
Calvarium, delayed ossification
0 (0.0%)
2 (0.8%)
0 (0.0%)
0 (0.0%)
8 (7.5%)
Miscellaneous
1 (0.6%)
1 (0.4%)
1 (0.5%)
0 (0.0%)
0 (0.0%)
aFive dams—died prior to scheduled sacrifice; 5—not pregnant at term; one—7 resorption sites and no live fetuses.
bDerived from nine pregnant dams with live fetuses and one dam with seven resorptions and no live fetuses.
"Statistically significant trend (Cochran-Armitage trend test, p < 0.05) as determined by EPA, after dropping the
highest dose because of frank maternal toxicity.
Source: Khera et al. (19791.
Dow Chemical Co. (1953) reported the results of a multigenerational study conducted by
the Stanford Research Institute in which groups of 4-month-old male and female Long Evans rats
(three males and nine females/group) were fed diets containing 0, 100, 1,000, or 10,000 ppm
biphenyl. Based on U.S. EPA (1988) subchronic reference values for body weight and food
consumption in male and female Long Evans rats, these dietary concentrations are estimated to
40
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
correspond to doses of 9, 89, and 887 mg/kg-day, respectively, for the males and 10, 101, and
1,006 mg/kg-day, respectively, for the females. For breeding, three females were placed together
with one male. Following the breeding phase, females were separated and number of litters cast,
number of days between mating and delivery, and number of pups/litter at delivery were
recorded. F1 pups were weighed and culled to seven/litter at 2 days of age and weaned at 3
weeks of age, and weights were recorded weekly for postnatal weeks 3-6. The F1 rats were
continued on the same diets as their parents, and, at 10 weeks of age, nine F1 females and three
F1 males were mated to produce an F2 generation of pups. F2 pups were selected (by the same
procedure) for mating and production of an F3 generation that were sacrificed at 3 weeks of age;
12 F3 pups from each dose group were subjected to gross pathologic examinations.
There were no significant differences between controls and 100 and 1,000 ppm biphenyl
groups regarding litters cast, gestation length, or average number or weight of pups/litter at birth
or at 3 or 6 weeks of age. Decreased fertility in the 10,000 ppm biphenyl group of F0 and F1
females was observed (6/9, 7/9, and 8/9 confirmed pregnancies for the three successive
generations of 10,000 ppm biphenyl groups versus 8/9, 9/9, and 8/9 confirmed pregnancies for
controls). Averaged for Fl, F2, and F3 pups combined, the 10,000 ppm biphenyl group
exhibited a significantly (p < 0.05) decreased number of pups/litter at birth (6.2/litter versus
8.6/litter for controls) and lower average body weight at 3 weeks of age (34 versus 48 g for
controls) and 6 weeks of age (78 versus 113 g for controls). Gross pathologic evaluations of F3
weanlings revealed no signs of biphenyl treatment-related effects. There was no reported
evidence of a cumulative effect over the three generations. The study authors suggested the
possibility that the decreased fertility, smaller litter size, and reduced rate of growth in the
10,000 ppm biphenyl group may have been associated with unpalatability and resultant
decreased food intake; however, food consumption data were not reported. Further, palatability
is unlikely to have been the cause of all observed effects since gavage dosing at a similar dose
level produced maternal and fetal toxicity in the Khera et al. (1979) study. Overall, this report
did not provide sufficient information to support a thorough evaluation of reproductive toxicity
with biphenyl exposure.
Ambrose et al. (1960) examined the reproductive toxicity of biphenyl in two
experimental series. In the first experiment, weanling albino rats were administered 0 or 1,000
ppm biphenyl (5 males and 10 females/group) or 5,000 ppm biphenyl (3 males and 9 females) in
the diet for 60 days prior to mating. In the second experiment, groups of 90-day-old albino rats
were administered 0 or 1,000 ppm biphenyl (4 males and 8 females/group) or 5,000 ppm
biphenyl (3 males and 9 females) in the diet for 11 days prior to mating. Based on U.S. EPA
(1988) subchronic reference values for body weight and food consumption in rats of unspecified
strain (average values for combined sexes), these dietary levels correspond to estimated doses of
105 and 525 mg/kg-day, respectively. All rats were maintained on their respective diets
throughout mating and until the progeny of all litters were weaned. Although the authors
41
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
concluded that the compound had no significant effect on reproduction, the reported data for
number of rats casting litters, total born, and range of litter size (Table 4-12) were insufficient to
support a full evaluation of the association between dietary exposure to biphenyl and
reproductive deficits.
Table 4-12. Summary of reproductive data in albino rats exposed to dietary
biphenyl
Experimental series
Diet (ppm)c
Dams with litters
Total offspring
Litter size (range)
First3
Control
9/10
59
3-9
1,000
10/10
67
2-10
5,000
8/9
53
3-9
Secondb
Control
8/8
64
5-13
1,000
6/8
63
3-10
5,000
8/9
48
3-9
"Weanling rats on diets for 60 days before mating.
b90-Day-old rats on diets for 11 days before mating.
°1,000 ppm = 105 mg/kg-day and 5,000 ppm = 525 mg/kg-day
Source: Ambrose et al. (19601.
4.3.2. Inhalation Exposure
No studies were identified that examined the reproductive/developmental toxicity of
biphenyl via the inhalation route.
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
4.4.1. Acute and Short-term Toxicity Data
Acute oral toxicity studies of biphenyl provide median lethal dose (LD50) values ranging
from 2,180 to 5,040 mg/kg for rats (Pecchiai and Saffiotti. 1957; Union Carbide. 1949;
Deichmann et al.. 1947; Monsanto. 1946) and an LD50 value of 2,410 mg/kg for rabbits
(Deichmann et al.. 1947). Dow Chemical Co. (1939) reported 100% survival and 100% lethal
doses of 1,600 and 3,000 mg/kg, respectively, in rats. Clinical signs commonly observed
following single oral dosing in these studies included increased respiration, lacrimation, loss of
appetite and body weight, and muscular weakness. Deaths occurred in the first few days
following dosing. Typical targets of histopathologic lesions were lungs, liver, and upper
gastrointestinal tract.
Groups of mice (10/sex of unspecified strain) were exposed to biphenyl by inhalation for
4 hours at average analytical concentrations of 14.11, 38.40, or 42.80 ppm (89.0, 242.2, and
-3
270.0 mg/m , respectively) and observed for up to 14 days following exposure (Sun Company
Inc., (1977a. b). Clinical signs of hyperactivity and mild respiratory discomfort were noted
42
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
during exposure, but resolved during postexposure observation. One male mouse of the
42.80 ppm group died after 2 hours of exposure, but this death was not attributed to biphenyl
exposure. All other mice survived throughout the 14-day postexposure observation period.
Slight lung congestion was noted in most mice upon gross pathological examination.
In a study by Sun Company Inc. (1977b). mice (10/sex of unspecified strain) were
exposed to biphenyl for 7 hours/day, 5 days/week for 2 weeks at average analytical
concentrations of 0, 24.8, or 54.75 ppm (0, 156.4, and 345.5 mg/m , respectively). Five
animals/group were sacrificed immediately after exposure; the remaining animals were sacrificed
following a 14-day recovery period. Clinical signs were monitored daily. Gross pathologic
examinations at necropsy included assessment of lungs, trachea, heart, spleen, liver, kidneys,
stomach, and intestines. Histopathologic examinations included tissues from lung, trachea,
kidney, spleen, and liver. The study authors reported signs of hyperactivity in some mice during
the first few exposure periods. One female mouse of the 24.8 ppm exposure group died prior to
the third exposure session and one control female mouse died prior the final exposure session.
No abnormal clinical signs were seen during the 14-day recovery period. Gross and
histopathologic examinations revealed no signs of exposure-related adverse effects.
Four rabbits (sex and strain unspecified) received up to 20 daily doses of 500 mg/kg
"purified" biphenyl to the skin; the compound was applied as a 25% preparation in olive oil.
Three rabbits received the same concentration of technical biphenyl (Deichmann et al., (1947);
Monsanto, (1946). The compound was left on the skin for 2 hours and then washed off with
soap and water. Some biphenyl derivatives were similarly assessed. One rabbit receiving
purified biphenyl died after eight applications, and the rest of the animals survived to term.
Average weight loss for the rabbits receiving purified and technical biphenyl was 45 and 172 g,
respectively.
4.4.2. Kidney/Urinary Tract Endpoint Studies
Endpoint-specific studies of biphenyl-induced urinary tract effects in rats (Shibata et al..
1989b; Shibata et al.. 1989a; Kluwe. 1982; S0ndergaard and Blom. 1979; Booth et al.. 1961)
support findings of the chronic oral rat studies described in Section 4.2.1.2 (Chronic Toxicity and
Carcinogenicity).
In a preliminary study, five adult rats (sex and strain unspecified) were administered
10,000 ppm biphenyl in the diet for 26 days followed by a 29-day postexposure recovery period
for a total study period of 55 days (Booth et al., (1961). Total urine volume and the volume of
sulfosalicylic acid-precipitable sediment were recorded from urine collected from all five rats on
study days 4, 8, 18, 20, and 26 (exposure days), and study days 28, 32, 35, and 54 (recovery
period). Volumes of both urine and sulfosalicylic acid-precipitable sediment increased from 7
and 0.56 mL, respectively, on exposure day 4 to 32 and 2.24 mL, respectively, on exposure day
20. Both values remained relatively high (approximately 27 and 2.2 mL, respectively) on
43
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
exposure day 26 and decreased to approximately 14 and 0.8 mL, respectively, by the end of the
recovery period. Fractionation and analysis of the precipitate suggested the presence of p-
hydroxybiphenyl and its glucuronide. Similar effects were observed in male and female rats
receiving 5,000 ppm biphenyl in the diet, but not 500 ppm.
A follow-up study employed 42 rats/sex/group and biphenyl dietary levels of 1,000,
2,500, or 5,000 ppm. Biphenyl doses are estimated to be 83.7, 209, and 419 mg/kg-day,
respectively, based on U.S. EPA (1988) chronic reference values for body weight and food
consumption in F344 rats (averages of values for males and females). Rats were exposed for up
to 165 days and followed for 0, 30, or 60 days of recovery. Urine samples were collected
periodically from five rats/sex/exposure group. Interim sacrifices of five rats/sex/exposure group
were performed after 30, 60, and 120 days on the diet in order to assess the progression of
biphenyl-induced histopathological effects on the kidney. Consistent with the preliminary study
findings, the rats of the 5,000 ppm group in the follow-up study exhibited gradual increases in
urine volume and sulfosalicylic acid-precipitable sediment and decreases in both parameters
during postexposure recovery. These effects were less pronounced in the 2,500 ppm group and
absent in the 1,000 ppm group. At 5,000 ppm, kidney lesions were noted in 1/5 males (several
small cysts and dilated tubules in the medulla and inner cortex) and 2/5 females (mild local
tubular dilation with some epithelial flattening) following 30 days of exposure. Similar, but
more extensive, kidney lesions were noted in 3/5 males and 5/5 females following 60 days of
exposure. The kidney lesions were even more prominent following 120 days of exposure.
Reported histopathologic findings in the kidneys of rats from the 2,500 ppm group were limited
to a single instance of an unspecified "prominent kidney lesion" at 60 days, and one small
calculus in the pelvis of one rat and a small calcareous deposit in the renal pyramid of another rat
following 120 days of exposure. Urinary and histopathologic renal effects were not assessed at
the end of the 165-day treatment period; however, during the 60-day postexposure recovery
period, rats of the 5,000 ppm biphenyl group exhibited a regression of kidney lesions and
improvement in urine quality.
Kluwe (1982) examined changes in urine composition and kidney morphology in F344
rats exposed to biphenyl. Groups of male F344 rats were administered biphenyl (in corn oil) by
single gavage dosing at 0, 250, 500, or 1,000 mg/kg and observed for 15 days following
treatment. Body weights were recorded, and urine was collected on days 1, 2, 3, 4, 8, and
15 following treatment for urinalysis. Interim sacrifices were performed on eight control and
eight high-dose rats on posttreatment days 1, 2, 3, 8, and 15 for assessment of weight and kidney
histopathology. There were no significant effects on body weight in the low-dose group. Mean
body weight gains of mid- and high-dose groups were consistently 6-10% lower than control
values (p < 0.05), beginning as early as day 2 following the initiation of dosing and continuing
through day 15. Dose-related increases in polyuria, proteinuria, and glucosuria were observed on
day 1; polyuria and glucosuria were no longer apparent by day 4 and proteinuria resolved
44
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
between days 8 and 15. Histopathologic examinations of kidneys revealed renal papillary
necrosis in 8/32 high-dose rats; this effect was observed as early as day 1 and persisted during
the 15-day posttreatment period.
Kluwe et al. (1982) conducted a similar experiment in which groups of male F344 rats
received biphenyl at doses of 0, 250, or 500 mg/kg-day by gavage for 14 days. In this
experiment, polyuria persisted throughout the treatment period; glucosuria was no longer
apparent by day 4 and proteinuria resolved between treatment days 8 and 15. Relative kidney
weight of high-dose rats was significantly increased during the second half of the treatment
period, but the magnitude of this effect was small and considered by the study authors to be of
little biological significance. There was some indication of tubular dilatation in focal areas of
kidneys from the high-dose rats.
Groups of male and female SPF-Wistar rats were administered diets consisting of
semisynthetic chow and biphenyl at concentrations resulting in biphenyl doses of 0, 50, 150, 300,
or 450 mg/kg-day (S0ndergaard and Blom, (1979). Other groups were administered diets
consisting of commercial chow and biphenyl at concentrations resulting in biphenyl doses of 0,
50, 150, 300, 500, or 1,000 mg/kg-day. The treatment period lasted for up to 21 days. The
numbers of male and female rats in each treatment group are specified in Table 4-13. Urine was
collected on days 4, 10, and 17 for urinalysis. At terminal sacrifice, absolute and relative kidney
weights were determined and kidney tissues were prepared for light and electron microscopic
assessment. Interim sacrifices (days 1, 2, 4, and 10) were performed in order to assess the
activity of AP in proximal tubules. Table 4-13 presents semiquantitative study results, which
include increases in urine volume/specific gravity and relative kidney weight, as well as
polycystic kidney changes. No changes in AP levels were seen as a result of biphenyl exposure.
The kidney effects of biphenyl appeared to be more pronounced when added to the semisynthetic
diet versus the commercial diet, with 50 mg/kg-day as a LOAEL for the onset of kidney changes.
45
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 4-13. Change in kidney weight and cellular architecture in Wistar
rats exposed to biphenyl
Exposure
(mg/kg-d)
Number of animals
(male/female)
Relative kidney weight
increases
Cystic change
Increases of urine
volume/specific gravity
Semisynthetic diet
0
3/14
-
-
-/-
50
4/3
+
-
150
0/10
+
*
•/•
300
14/14
+++
***
450
4/4
+++
***
Commercial chow
0
10/20
-
-
-/-
50
10/10
-
-
150
10/10
-
-
300
10/10
-
-
500a
0/10
+b
-
•/•
l,000a
0/10
+++b
**
•/•
aDose for 14 days.
bAbsolute organ weight.
+ = statistically significant compared with controls (p < 0.05), as calculated by the authors (Student's t-test);
+++ = statistically significant compared with controls (p < 0.001), as calculated by the authors (Student's t-test);
* = less than one-third of the area; ** = less than two-thirds of the area; *** = greater than two-thirds of the area;
• = effect; - = no effect
Source: Sondergaard and Blom (1979).
Male F344 rats (20/group) were exposed to 0 or 5,000 ppm biphenyl in the diet for
24 weeks (Shibata et al.. 1989a). After 4 weeks, 5 rats/group were injected with 100 mg/kg
5-bromo-2-deoxyuridine (BrdU) and sacrificed 1 hour later. One kidney from each rat was
processed for immune-histopathologic identification of BrdU as an index of cell proliferation,
while the second kidney was processed for light and scanning electron microscopic examination.
The remaining rats were sacrificed after 8, 16, and 24 weeks to monitor further development of
morphological alterations in the renal papilla and pelvis. Survival was unaffected by treatment
and biphenyl-treated animals showed no adverse clinical signs. Treatment was associated with
significantly lower mean body weight compared to controls; food consumption was unaffected
and water consumption was slightly higher than that of controls. There were no significant
treatment-related effects on labeling indices of cell proliferation (BrdU incorporation) in renal
papilla or pelvic epithelia, and no histopathologic lesions of the renal papilla and pelvis were
evident. Focal calcification of the renal medulla was observed in the majority of the biphenyl -
treated rats. The study authors stated that urinalysis demonstrated an association between
biphenyl exposure and microcalculi formation, but provided no additional information regarding
urinalysis results.
46
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
In a similar study (Shibata et al.. 1989b). a group of 10 male F344 rats received 5,000
ppm biphenyl in the diet for up to 8 weeks. Based on U.S. EPA (1988) subchronic reference
values for body weight and food consumption in male F344 rats, the dose was estimated at
500 mg/kg-day. At 4 weeks, five rats/group were processed as described by Shibata et al.
(1989b) for assessment of BrdU incorporation, but in the urinary bladder rather than in the
kidney. During week 4, urine samples were taken for urinalysis. At terminal sacrifice, urinary
bladder tissues were processed for scanning electron microscopic examinations. There were no
treatment-related deaths or adverse clinical signs. Although food and water consumption were
similar to controls, biphenyl-treated rats showed a consistent reduction in average body weight
(229 versus 247 g after 4 weeks and 300 versus 327 g after 8 weeks, for treated versus controls,
respectively \p < 0.01]). A greater than fourfold increase in the BrdU labeling index was
observed in urinary bladder epithelium of the biphenyl-fed rats (mean percent labeling index of
0.58 ± 0.31 compared to 0.13 ± 0.09 in controls; p < 0.05). Urinalysis revealed numerous
microcalculi in the urinary sediment of the biphenyl-treated rats. This condition, designated as
"severe" by the authors, was associated with histopathological lesions of the epithelium of the
urinary bladder that included simple hyperplasia with moderate severity (5/5 rats), moderate
pleomorphic microvilli (5/5 rats), moderate uniform microvilli (5/5 rats), and the occurrence of
ropey or leafy microridges (5/5 rats), the latter condition designated as severe. Scanning electron
microscope images of the luminal surface of bladder epithelial cells showed pleomorphic
microvilli that varied in size and shape and the formation of microridges.
4.4.3. Biphenyl as a Tumor Promoter
Male B6C3Fi mice (10-20/group) received the bladder carcinogen N-butyl-
N (4-hydroxybutyl)nitrosamine (BBN) at 0 or 0.05% in the drinking water for 4 weeks followed
by 0 or 10,000 ppm biphenyl in the diet for 32 weeks (Tamano et al., (1993). The mice were
observed for clinical signs, and body weight and food consumption were monitored. At 37-week
terminal sacrifice, kidneys and urinary bladders were prepared for histopathological examination.
No treatment-related clinical signs were observed. Mean body weight of the BBN + 10,000 ppm
biphenyl-treated mice was significantly (p < 0.01) lower than that of mice receiving BBN
treatment only (32.2 ±1.8 versus 38.4 ± 2.6 g). Biphenyl treatment did not result in increased
incidences of simple hyperplasia or papillary or nodular dysplasia in the BBN-initiated mice.
Administration of 10,000 ppm biphenyl in the diet to eight mice for 8 weeks did not significantly
affect indices of cell proliferation (BrdU incorporation) in urinary bladder epithelium.
In the initiation-promotion portion of a chronic toxicity study designed to assess the
ability of biphenyl to promote carcinogenesis by EHEN in the kidney (see Section 4.2.1.2.1 for a
detailed study description), male Wistar rats (25/group) received a basal diet with either 0 or
0.1% dietary EHEN for 2 weeks, followed by a basal diet containing either 0, 1,250 or 5,000
ppm biphenyl for 34 weeks (Shiraiwa et al.. 1989). Based on reported values for mean daily
47
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
biphenyl intake (mg biphenyl/rat) and average body weight (mean initial body weight + one-half
the difference between mean initial and mean final body weight) for each study group,
corresponding doses are estimated to have been approximately 0, 60, and 248 mg/kg-day,
respectively. At terminal sacrifice, gross pathologic examinations were performed. Kidney and
urinary bladder were fixed; kidneys were sectioned transversely (10-12 serial slices) and urinary
bladders were cut into 4-6 serial slices. The authors used a computer-linked image analyzer to
determine the incidence of kidney lesions and dysplastic foci. The presence of stones in the
kidney and urinary bladder was assessed qualitatively using an infrared spectrophotometer.
Stones were present in the kidney, ureter, and urinary bladder of high-dose rats
irrespective of whether animals were initially exposed to the basal or EHEN-containing diet
(combined incidences of 6/25 and 8/25, respectively). The incidence of rats with renal cell
tumors after EHEN and subsequent biphenyl administration was lower than that of rats receiving
EHEN followed by basal diet (7/25 and 13/25, respectively). This finding indicates that biphenyl
was not a promoter of renal cell tumors in male Wistar rats under the conditions of the study.
Male F344 rats (25/group) were exposed to 0.05% BBN (a bladder carcinogen) in the
drinking water for 4 weeks followed by diets containing either 0 or 5,000 ppm biphenyl for
32 weeks (Kurata et al., (1986). One group of five rats received biphenyl without pretreatment
with BBN. The rats receiving biphenyl either with or without pretreatment with BBN gained
less weight than control rats or those receiving only BBN. Incidences of urinary bladder
hyperplasia, papilloma, and carcinoma were 17/18 (94%), 15/18 (83%), and 11/18 (61%),
respectively, in the group of rats that survived treatment of BBN followed by biphenyl,
compared to 6/24 (25%), 3/24 (12%), and 0/24 (0%), respectively, in the rats receiving BBN
only. These urinary bladder lesions were not seen in any of the five rats receiving biphenyl
without BBN pretreatment. Urinary bladder calculi were found in 25% of the rats receiving
BBN followed by biphenyl and in 12% of the rats receiving BBN only. Biphenyl was
considered a urinary bladder tumor promoter in male F344 rats under the conditions of the study.
Biphenyl was negative for tumor promotion in a skin-painting experiment in which the
initiator was 0.3% 9,10-dimethyl-l,2-benzanthracene in benzene (Boutwell and Bosch. 1959). In
the 16/20 mice that survived the topical application of 20% biphenyl for 16 weeks, none
developed papillomas or carcinomas as a result of treatment.
Six-week-old male F344 rats (20-30/group) were exposed to BBN in drinking water at
0.01 or 0.05%) for 4 weeks, followed by 5,000 ppm biphenyl in the diet for 32 weeks (Ito et al.,
(1984). Controls receiving only BBN and controls receiving only biphenyl were included. After
sacrifice, urinary bladders were prepared for light microscopic assessment of neoplastic and
cancerous lesions. The study authors reported that biphenyl exhibited moderate bladder cancer-
promoting activity, but data to support this finding were not included in the study report.
48
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
ACTION
Studies have been conducted to investigate the mechanisms by which biphenyl induces
effects on the urinary bladder, liver, and endocrine system. Other studies have looked at the
potential for biphenyl to induce apoptosis, to affect mitochondrial activity, and to induce genetic
changes. This literature is summarized in Appendix C. Mechanistic studies of biphenyl effects
on the urinary bladder, a principal target of biphenyl toxicity, and genotoxic potential are briefly
discussed below.
4.5.1. Effects on the Urinary Bladder of Rats
Mechanistic studies have been performed in F344 rats to investigate the relationship
between calculi formation in the urinary bladder and bladder tumor induction in male rats.
Ohnishi et al. (2001; 2000a; 2000b) proposed that gender differences in urinary conditions
(including pH and potassium concentrations) and kidney sulphatase activity may be responsible
for the gender differences in urinary calculi composition and formation and the subsequent
development of urinary bladder tumors in male, but not female, F344 rats. Information from
available mechanistic studies is summarized in Appendix C.
4.5.2. Genotoxicity
The genotoxicity studies of biphenyl and its metabolites are summarized in Appendix C,
Tables C-2 and C-3. A review of the available data suggests biphenyl may have some capability
of inducing genetic damage under certain conditions. Bacterial mutagenicity assays are
uniformly negative, even with metabolic activation; however, several in vitro mammalian cell
assays were able to detect weak evidence of mutagenicity with activation (Glatt et al., 1992;
Wangenheim and Bolcsfoldi, 1988). Indications of the ability to induce chromosomal
aberrations were also observed with the addition of metabolic activation (Sofuni et al., 1985),
although this was accompanied by cytotoxicity in one study without metabolic activation
(Rencuzogullari et al., 2008). In addition, evidence of DNA strand breaks was observed in mice
in several organs, including the stomach, blood, liver, bone marrow, kidney, bladder, lung, and
brain (Sasaki et al., 2002, 1997). Micronuclei were observed in primary human lymphocytes
(Rencuzogullari et al., 2008), but were not found in another study in mouse bone marrow
(Gollapudi et al., 2007). Chromosomal aberrations (CAs) were not observed following
inhalation exposures in rats (Johnston et al., 1976).
There are indications that the metabolites of biphenyl may be more genotoxic than the
parent compound when the metabolites are directly tested in assay systems. Genotoxicity results
for the major metabolite, 4-hydroxybiphenyl, and a minor metabolite, 2-hydroxybiphenyl (i.e., o-
phenylphenol, or OPP), can be found in Appendix C, Table C-3. Thus, it is possible that the
49
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
genotoxic potential in any given system or organism is directly related to the proportion these
metabolites are formed in that system.
It is unknown if reports of DNA damage following exposure to biphenyl are caused by a
direct reaction of metabolites with DNA or by indirect damage from cytotoxicity, reactive
oxygen species (ROS) generated from redox cycling of hydroquinone metabolites, or some
combination of these mechanisms. Biphenyl in an activated system was not investigated for its
ability to form DNA-reactive metabolites, but in studies of DNA adduct formation using the
metabolites, most were negative (Kwok et al., 1999; Smith et al., 1998) except for one study of
high doses applied to skin (Pathak and Roy, 1993). However, several reports indicate that
genetic damage often occurred only after high doses that were accompanied by decreased cell
survival or was concurrent with redox cycling following metabolism of 2-hydroxybiphenyl, a
minor metabolite of biphenyl (see Appendix C). One study that directly tested the mutagenicity
of the major metabolite, 4-hydroxyquinone, in the Salmonella Ames assay was positive
(Narbonne et al., 1987), but no other investigations of this metabolite were located. In summary,
there is not enough evidence to conclude that biphenyl is mutagenic or can react directly with
DNA. The available genotoxicity database suggests that most indications of genotoxicity
following biphenyl exposure are likely to be secondary responses resulting from oxidative
damage and cytotoxicity.
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS
Tables 4-14 and 4-15 include the major studies and the observed effects for oral and
inhalation exposure to biphenyl, respectively.
50
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure
route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Subchronic studies
Rat, Long-Evans
(female, 8/group)
Diet
0, 10, 30, or 100
90 d
ND
ND
Lower average plasma BUN
levels in all exposed groups
(biological significance is
uncertain).
Dow Chemical
Co. (1953)a
Mouse, BDF,
(10/sex/group)
Diet
0, 93, 347, 747, 1,495,
1,868, or 2,989
13 wks
M: 747
F: 747
M: 1,495
F: 1,495
M: Decreased body weight.
F: Decreased body weight
To overcome possible
problems with taste aversion,
animals in the 3 highest dose
groups received lower doses
for exposure weeks 1-2,
followed by the final dose for
the remaining time.
Umeda et al.
(2004a)
Chronic studies
Rat, F344
(50/sex/group)
Diet
M: 0,36.4, 110, or
378
F: 0, 42.7, 128, or
438
2yrs
M: 110
F: 42.7
M: 378
F: 128
M: Bladder tumors and
transitional cell hyperplasia.
F: Nonneoplastic kidney lesions
(transitional cell hyperplasia in
the renal pelvis and hemosiderin
deposits).
Umeda et al.
(2002)
Rat, Wistar
(50/sex/group)
Diet
M: 0, 165, or 353
F: 0, 178, or 370
75 wks
M: ND
F: ND
M: 165
F: 178
Formation of kidney stones
associated with pyelonephritis in
both sexes.
Shiraiwa et al.
(1989)
51
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure
route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Rat, Wistar
(male, 25/group)
Diet
Control groups: basal
diet for 2 wks
followed by exposure
at 0,59.28, or 248.3
for 34 wks
Exposure groups: diet
containing 0.1%
EHEN for 2 wks
followed by 0, 62, or
248.2 for 34 wks
Control:
59.28
Exposure: 62
Control:
248.3
Exposure:
248.2
Formation of kidney stones
associated with pyelonephritis in
both groups.
Biphenyl did not exhibit
tumor promoting
characteristics for the kidney
tumor initiator, EHEN, under
the conditions of this study.
Shiraiwa et al.
(1989)
Rat, albino
(weanling,
15/sex/group)
Diet
0, 1, 4, 8, 42, 84, 420,
and 840
2yrs
84
420
Kidney effects including tubular
atrophy and dilation associated
with cyst formation and calculi
formation in the renal pelvis of
both sexes.
Necropsies were performed
on terminal sacrifice animals
only (n= 2-13
animals/group)
Ambrose et al.
(1960)
Rat, albino
(male, 8/group)
Diet
0, 250, or 450
13 mo
ND
250
Nonneoplastic degenerative
changes in the liver, kidney,
thyroid, and parathyroid
resulting in hyperplasia of liver,
kidney, and thyroid.
Pecchiai and
Saffiotti (1957)
Rat, Sprague-
Dawley
(12/sex/group)
Diet
0, 7, 73, or 732
2yrs
73
732
Renal effects (tubular dilatation,
calcification, and intratubular
inflammation).
Decreased survival and small
number of animals/group
may have impaired ability to
detect late-developing
tumors.
Dow Chemical
Co. (1953)a
52
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure
route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Mouse, BDFi
(50/sex/group)
Diet
M: 0,97, 291, or
1,050
F: 0, 134, 414, or
1,420
2yrs
M: 97
F: 134
M: 291
F: 414
M: Decreased body weight.
F: Nonneoplastic effects
(mineralization in the kidney
and significantly increased
plasma ALT and AST activities)
in female mice.
Umeda et al.
(2005)
Mouse, ddY
(female, 60/group)
Diet
0 or 855
2yrs
855
ND
No adverse effects observed at
the dose tested
Results were reported for
34-37/group.
Imai et al. (1983)
Mouse, hybrid
(two strains,
18/sex/strain/group)
Gavage (215
mg/kg body
weight in
0.5% gelatin)
for the first 3
wks, followed
by dietary
exposure for
the remaining
time
0 or 91
18 mo
91
ND
Reticular cell sarcoma incidence
significantly elevated in strain B
female mice, but not in male
mice of this strain or strain A
mice of either sex
Two strains of F1 hybrid
mice were produced by
mating female C57BL/6
mice with either male
C3H/Anf mice (strain A) or
male AKR mice (strain B)
Innes et al.
(1969); NCI
(1968)
Dog, mongrel
(males/group; 1
female/group)
Capsule in
corn oil
0,2.5, or 25
5 d/wk for 1 yr
ND
ND
ND
Monsanto
(1946)a
Monkey, Rhesus
(2 males/group;
lfemale/group)
Diet
0,0.01, 0.1, or 1% for
1 yr
ND
ND
ND
Author considered an
increase in relative liver
weight in high-dose monkeys
to be possibly compound-
related
Dow Chemical
Co. (1953)a
53
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure
route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Reproductive and developmental studies
Rat, Wistar
(18-20 pregnant
females/group)
Gavage in
corn oil
0, 125, 250, 500, or
1,000 on GDs 6-15
Dam: 500
Offspring:
250
Dam: 1,000
Offspring:
500
Dam: Maternal toxicity
(increased mortality), increased
dead or resorbed fetuses.
Offspring: Increased incidence
of fetuses with missing and
unossified sternebrae
Khera et al.
(1979)
Rat, Long Evans
(3 males/group;
9 females/group)
Diet
M: 9, 89, or 887
F: 10, 101, or 1,006
Continuous breeding
M: ND
F: 101
M: ND
F: 1,006
M: ND
F: Decreased fertility and litter
size; reduced offspring body
weight.
The authors suggested that
effects seen in the high-dose
group were associated with
unpalatability and resultant
decreased food intake;
however, food consumption
data were not provided to
support this interpretation.
Dow Chemical
Co. (1953)a
Rat, albino
Experiment 1:3-5
males/group; 9-10
females/group.
Experiment 2:3-4
males/group; 8-9
females/group.
Diet
0, 105, or 525
Experiment 1: 60 days
prior to mating
Experiment 2: 11 days
prior to mating
ND
ND
ND
Authors presented tabulated
data and concluded that the
compound had no significant
effect on reproduction.
Ambrose et al.
(1960)
aReport was not peer reviewed.
F = female; M = male; ND = not determined
Note: Other studies of subchronic duration that examined the effects of biphenyl on the urinary tract only (Shibata et al.. 1989a: Shibata et al.. 1989b) are summarized in
Section 4.4.2. Because these studies were designed to investigate the effects of biphenyl on the kidney and urinary bladder and the mode of action by which biphenyl
induces these effects, the studies were not useful for identifying NOAELs and LOAELs, and were not included in this table.
54
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-15. Summary of major studies evaluating effects of biphenyl after inhalation exposure in rats, mice and
rabbits
Species, strain
Dose (mg/m3), duration
NOAEL
(mg/m3)
LOAEL
(mg/m3)
Effect(s) at the LOAEL
References
Rabbit, albino
(3/group)
Rat, Sprague-Dawley
(10/group)
300 mg/m3 (7 hrs/d, 5 d/wk)
64 d over 94-d period
Rabbit: ND
Rat: ND
Rabbit: ND
Rat: 300
Rabbit: ND
Rat: Mortality (5/10), acute emphysema,
congestion, edema, bronchitis, lobular pneumonia,
and multiple pulmonary abscesses
Deichmann et al.
(1947); Monsanto
(1946)
Rabbit, albino
(3/group)
Rat, Sprague-Dawley
(6/group)
40 mg/m3 (7 hrs/d, 5 d/wk)
46 d over 68-d period
Rabbit: ND
Rat: ND
Rabbit: ND
Rat: 40
Rabbit: ND
Rat: Mortality (1/6), acute emphysema, congestion,
edema, bronchitis, lobular pneumonia, and multiple
pulmonary abscesses
Mice (12/group)
Rat, Sprague-Dawley
(4/group)
5 mg/m3 (7 hrs/d, 5 d/wk)
62 d over 92-d period
Mouse: ND
Rat: ND
Mouse: 5
Rat: ND
Mouse: Mortality (2/12); upper respiratory tract
irritation (acute emphysema, congestion, edema,
bronchitis, lobular pneumonia, and multiple
pulmonary abscesses)
Rat: ND
Mouse, CDI
(50/sex/group)
0, 157.7, or 315.3 mg/m3 (7 hrs/d,
5 d/wk), 13 wks
ND
157.7
Histopathological lung, liver, and kidney lesions
(congested and hemorrhagic lungs, tracheal
hyperplasia, and congestion and edema in the liver
and kidney) in both sexes.
Sun Company Inc.
(1977b)a
"Report was not published.
ND = not determined
55
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4.6.1. Oral
The primary targets of toxicity of ingested biphenyl in experimental animals are the
kidney, urinary bladder, liver, and developing fetus. Decreased body weight has also been
associated with oral biphenyl exposure. No information was located regarding possible
associations between oral exposure to biphenyl and health outcomes in humans.
Chronic oral studies identified the kidney as one of the noncancer targets of biphenyl in
both rats and mice. Exposure to biphenyl in the diet for 2 years produced a range of
histopathological changes in the kidney in F344 rats (TJmeda et al.. 2002). Mineralization of the
papilla (part of the renal medulla) showed a dose-related increase in both male and female rats;
papillary necrosis was observed in both sexes of rats at the high dose only. Papillary
mineralization can be found in association with papillary necrosis (Bach and Thanh, 1998), and
the histopathologic changes in the medulla overall suggest a continuum of increasing severity of
damage with increasing biphenyl dose. Effects in the papillary region of the medulla were
supported by dose-related histopathologic changes in the renal pelvis of male and female rats in
the Umeda et al. (2002) bioassay, including mineralization, transitional cell hyperplasia (simple
and nodular), desquamation, and calculus formation. A dose-related increase in the incidence of
hemosiderin deposits was observed in female rats, but not in male rats at any dose level.
Hemosiderin, an iron-protein complex that may be present as a product of hemoglobin
degradation, can arise from various conditions (Jennette et al., 2007). Without information in
Umeda et al. (2002) on severity and location of hemosiderin within the kidney, the biological
significance of this endpoint is unclear. Kidney findings were consistently observed in other
studies in rats, including tubular dilation or mild tubuli degeneration in albino and Sprague-
Dawley rats (Ambrose et al. 1960; Pecchiai and Saffiotti. 1957; Dow Chemical Co. 1953) and
calculi formation in the renal pelvis in Wistar and albino rats (Shiraiwa et al.. 1989; Ambrose et
al., 1960). Dose-related pathological changes in the kidney in BDFi mice following 2-year
dietary exposure to biphenyl included desquamation of the renal pelvis and mineralization of the
medulla (Umeda et al.. 2005). A dose-related increase in BUN levels in mice in this study
(Umeda et al.. 2005) provides evidence of biphenyl-induced functional disruption of the kidney.
Imai et al. (1983) did not find histopathological changes in the kidney of ddY mice exposed to
biphenyl in diet for 2 years (-60% of the animals were subjected to pathological examination in
this study). There is a hazard potential for kidney toxicity based on consistent evidence of
biphenyl-induced kidney toxicity in studies in rats and some support from studies in mice.
Urinary bladder toxicity associated with oral exposure to biphenyl was observed in rats
only. Increased incidences of urinary bladder hyperplasia and calculi or stones were observed in
male and female F344 rats exposed to biphenyl in the diet (378 and 438 mg/kg-day, respectively)
for 2 years (Umeda et al.. 2002) and in male and female Wistar rats exposed to biphenyl in the
diet (353 and 370 mg/kg-day, respectively) for up to 75 weeks (Shiraiwa et al.. 1989). In a
subchronic study by Shibata et al. (1989b), increases in BrdU labeling index and simple
56
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
hyperplasia in urinary bladder epithelium were observed in male F344 rats given biphenyl in the
diet (500 mg/kg-day) for 4 weeks. Ambrose et al. (1960) and Dow Chemical Co. (1953) did not
find lesions in urinary bladder in albino and SD rats exposed to biphenyl in the diet for two
years; however, both studies used relatively small group sizes and provided limited necropsy
data. Biphenyl did not induce changes in the urinary bladder in mice (Imai et al., 1983; Umeda
et al., 2005). There is a hazard potential for urinary bladder toxicity from biphenyl exposure
based on evidence of calculi formation and epithelial lesions in the urinary bladder of rats.
Because urinary bladder toxicity was not found in a second species, the evidence for hazard
potential is weaker than for the kidneys.
Liver toxicity, including histopathological changes and increased liver weight and serum
liver enzymes, were observed in studies of mice and rats. Relative liver weight was increased by
more than 10% in female albino and Sprague-Dawley rats exposed to 420 and 732 mg/kg-day
biphenyl for 2 years, respectively (Ambrose et al., 1960; Dow Chemical Co. 1953). and in rhesus
monkeys exposed to 1% biphenyl in the diet for one year (Dow Chemical Co. 1953). The only
histopathological change observed in rats was moderate degeneration of parenchymal
hepatocytes within 2 months followed by regenerative hyperplasia and nuclear hypertrophy that
persisted to 13 months in male albino rats exposed to >250 mg/kg-day biphenyl (Pecchiai and
Saffiotti. 1957). Liver toxicity was not reported in F344 rats exposed to biphenyl in diet up to
438 mg/kg-day for 2 years (Umeda et al., 2002). Differences in response in the two studies may
be due to differences in strain susceptibility. In BDFi mice, relative liver weight of female mice
exposed to 134 -1,420 mg/kg-day biphenyl in the diet for 2 years was increased by 1.3-1.6-fold
(Umeda et al., 2005); biphenyl exposure did not affect liver weight in male mice.
Histopathological changes included enlarged centrilobular hepatocytes filled with eosinophilic
granules identified as peroxisomes in BDFi mice exposed to 2,989 mg/kg-day biphenyl in diet
for 13 weeks (Umeda et al., 2004), and basophilic foci in female BDFi mice exposed to biphenyl
in the diet (>414 mg/kg-day) for two years (Umeda et al.. 2005). Significantly increased plasma
enzyme levels (AST, ALT, AP, and LDH) were observed primarly in female BDFi mice exposed
to biphenyl in the diet for 2 years (Umeda et al.. 2005). No liver toxicity was found in female
ddY mice exposed to 855 mg/kg-day biphenyl for 2 years (Imai et al., 1983) based on
histopathological examination of-60% of the animals (34 of 60 exposed). In summary,
biphenyl exposure resulted in increased liver weight and histopathological changes of the liver in
mice and rats, and increased liver weight in monkeys; however, liver toxicity was not observed
consistently across different strains of rats and mice or across sexes. Based on these findings,
there may be a hazard potential for liver toxicity from biphenyl exposure.
In the only available oral developmental toxicity study of biphenyl (Khera et al.. 1979).
the incidence of anomalous fetuses and litters bearing anomalous fetuses (including wavy ribs,
extra ribs, missing and unossified sternebrae, or delayed calvarium ossification) generally
increased with dose. When the anomalies were considered individually, only the incidence of
57
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
missing or unossified sternebrae exhibited an increasing trend with dose. As noted in EPA's
Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA, 1991), a significant, dose-
related increase in a variation (e.g., delayed ossification) should be evaluated as a possible
indication of developmental toxicity, although an assessment of the biological significance of
such variations should take into consideration knowledge of the developmental stage,
background incidence of certain variations, other strain- or species-specific factors, and maternal
toxicity. Carney and Kimmel (2007) observed that the biological significance of skeletal
variations that seem to be readily repairable via postnatal skeletal remodeling should be
interpreted in the context of other maternal and fetal findings, information on normal
skeletogenesis patterns, mode of action of the agent, and historical control incidence. The Khera
et al. (1979) study showed a 10% decrease in body weight gain and frank maternal toxicity in
dams at the high dose of 1,000 mg/kg-day (increased mortality [5/20 versus 0/18 in controls]) but
not at doses of 125, 250, or 500 mg/kg-day. Therefore, the increasing trend of fetuses with
missing or unossifed sternebrae at or below 500 mg/kg-day cannot be attributed to maternal
toxicity. In summary, findings from a single developmental toxicity study (Khera et al., 1979)
provide evidence that biphenyl may directly target skeletal development in Wistar rats
independent of maternal toxicity; however, no other developmental toxicity studies are available
to confirm these findings. Based on these findings, there may be a hazard potential for
developmental toxicity from biphenyl exposure.
Reproductive effects of biphenyl were evaluated in two multigeneration studies
(Ambrose et al., 1960; Dow Chemical Company, 1953). There was some indication in Dow
Chemical Co. (1953) that oral doses similar to those observed to be maternally and
developmentally toxic following administration during gestation (Khera et al., 1979) resulted in
evidence of reduced fertility and decreased pup growth. Ambrose et al. (1960) reported limited
findings and concluded that biphenyl had no significant effect on reproduction in albino rats
exposed to biphenyl in the diet at doses up to 525 mg/kg-day. Overall, the available
multigeneration studies in rats (Ambrose et al., 1960; Dow Chemical Company, 1953) were
inadequate to fully evaluate effects of biphenyl exposure on reproductive function, and a
determination of reproductive hazard cannot be made.
Decreased body weight gain associated with biphenyl exposure was observed in both rats
and mice. Following a 2-year dietary exposure to biphenyl, more than a 10% decrease in body
weight relative to controls was reported in F344 rats of both sexes (males—378 mg/kg-day;
females—438 mg/kg-day) (Umeda et al.. 2002) and in BDFi mice in both sexes (males—291
mg/kg-day; females—>414 mg/kg-day) (Umeda et al.. 2005). A 75-week study in Wistar rats
also found more than a 10% body weight decrease in males at doses >165 mg/kg-day and in
females at doses >178 mg/kg-day (Shiraiwa et al.. 1989). Shorter-duration oral exposure
(13 weeks) of mice to biphenyl at higher dietary concentrations (estimated doses >1,500 mg/kg-
day) was also associated with >17% decreased body weight (Umeda et al.. 2004a). Ambrose et
58
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
al. (1960) and Dow Chemical Co. (1953) reported more than 10% reduced body weight gain,
but the authors attributed low body weight to low palatability of the feed. In summary,
decreased body weight gain appears to be associated with oral exposure to biphenyl.
4.6.2. Inhalation
The toxicity of inhaled biphenyl has received less investigation than ingested biphenyl.
An epidemiological study of workers engaged in the production of biphenyl-impregnated paper
(Seppalainen and Hakkinen, 1975; Hakkinen et al., 1973, 1971) provides some evidence of liver
damage (including elevated levels of serum AST and ALT) and effects on the central and
peripheral nervous (including abnormal EEGs and ENMGs). In a study of a different facility
manufacturing biphenyl-impregnated paper prompted by the finding of 3 cases of Parkinson's
disease at that facility, an elevated RR of Parkinson's disease among biphenyl workers was
reported (Wastensson et al.. 2006). The workplace conditions reported for these studies
(Wastensson et al., 2006; Seppalainen and Hakkinen, 1975; Hakkinen et al., 1973, 1971)
suggested that inhalation represented the predominant route of exposure and that existing
occupational exposure limits had been exceeded, but dermal absorption as well as oral uptake
(hand to mouth) might have occurred at a significant level.
In mice, short-term biphenyl inhalation at concentrations as high as 55 ppm
(345.5 mg/m ) appeared to cause no observable clinical toxicity (Sun. 1977b). In ano ther study,
groups of rabbits, rats, or mice were exposed to biphenyl by inhalation for 7-13 weeks at
concentrations ranging from 5 to 300 mg/m (Deichmann et al.. 1947). No adverse effects were
observed in rabbits, while rats and mice showed irritation of mucous membranes and succumbed
at high concentrations. Mice were more sensitive than rats in these experiments, additionally
showing congestion and hemorrhage of the lungs (Deichmann et al.. 1947). High incidences of
pneumonia and tracheal hyperplasia, and congestion and edema in the lungs, liver, and kidney
were reported in a 13-week inhalation study of biphenyl in mice that was limited by study
methodology and reporting issues (Sun. 1977a). Reproductive or developmental studies using
the inhalation route of exposure were not identified.
4.6.3. Mode-of-Action Information
The urinary bladder is a target of biphenyl toxicity in the rat, and histopathological
lesions in this organ appear to be related to the formation of urinary bladder calculi induced by
biphenyl exposure. Mode of action information related to the role of calculi formation in the
induction of urinary bladder toxicity is described in Section 4.7.3. The mode of action for
biphenyl-induced toxicity in the kidney, another organ in the urinary system, has not been
investigated. Bioassay data suggest that a mode of action involving calculi formation does not
fully explain kidney lesions induced by biphenyl; kidney lesions were found in mice exposed to
biphenyl in the diet for 104 weeks without calculi formation (Umeda et al., 2005). Further, the
59
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
incidences of kidney histopathologic lesions in male and female rats exposed to biphenyl in the
diet for 104 weeks were similar (Umeda et al., 2002), whereas the incidence of calculi in the
kidney was lower in females than males (i.e., 3/50 versus 13/50 in the high-dose groups,
respectively).
Mode of action information related to biphenyl-induced liver toxicity is limited to the
proposed involvement of peroxisome proliferation-activated receptors (PPARs). Evaluation of
the evidence for a proposed PPAR mode of action is provided in Section 4.7.3.2.
Mechanistic studies provide some information on the induction of decreased body weight
gain by biphenyl. A possible mode of action is suggested by an in vitro study, where biphenyl
can act as an uncoupler of respiration (Nishihara, (1985).
There is no mode of action information on the toxicity of biphenyl to the developing fetus
or reproductive system.
4.7. EVALUATION OF CARCINOGENICITY
4.7.1. Summary of Overall Weight of Evidence
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the
database for biphenyl provides "suggestive evidence of carcinogenic potential" based on an
increased incidence of urinary bladder tumors (transitional cell papillomas and carcinomas) in
male F344 rats (Umeda et al.. 2002) and liver tumors (hepatocellular adenomas and carcinomas)
in female BDFi mice (Umeda et al.. 2005) exposed to biphenyl in the diet for 104 weeks, as well
as information on mode of carcinogenic action. The carcinogenic potential of biphenyl in
humans has not been investigated.
As emphasized in the Cancer Guidelines (U.S. EPA. 2005a). selection of the cancer
descriptor followed a full evaluation of the available evidence. The biphenyl case could be
considered a borderline case between two cancer descriptors—"likely to be carcinogenic to
humans" and "suggestive evidence of carcinogenic potential." In particular, biphenyl tested
positive at more than one site (urinary bladder and liver) and in more than one species (rat and
mouse), corresponding most closely to one of the examples in the Cancer Guidelines (U.S. EPA,
2005a) for the descriptor "likely to be carcinogenic to humans": "an agent that has tested positive
in animal experiments in more than one species, sex, strain, site, or exposure route, with or
without evidence of carcinogenicity in humans."
In contrast, the Cancer Guidelines indicate that the descriptor "suggestive evidence of
carcinogenic potential" is appropriate when a concern for potential carcinogenic effects in
humans is raised, but the data are judged not sufficient for a stronger conclusion, given "an
extensive database that includes negative studies in other species," and that "additional studies
may or may not provide further insights." The database for biphenyl includes studies in rats and
mice that did not show clear evidence of carcinogenicity (Shiraiwa et al., 1989; Imai et al., 1983;
NCI, 1968; Ambrose et al., 1960; Dow Chemical Co., 1950), but that were also limited in large
60
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
part in design, conduct, or reporting of results and therefore considered less informative for
evaluating the carcinogenicity of biphenyl than the studies by Umeda et al. (2005, 2002). The
range of evidence regarding each tumor type is described further in Section 4.7.2.
Mode of action information indicates that the induction of urinary bladder tumors in F344
male rats by dietary biphenyl exposure is a high-dose phenomenon closely related to the
formation of urinary bladder calculi. As discussed in more detail in Section 4.7.3.1, the mode of
action information is sufficient to conclude that urinary bladder tumors in male F344 rats will not
occur without the development of calculi, and that the induction of these tumors by biphenyl is
specific to male rats. Gender-specific differences in urinary conditions such as pH and
potassium concentrations appear to play a role in the differences in calculi formation and
composition. While the proposed mode of action for urinary bladder tumors in male rats is
assumed to be relevant to humans, the available evidence suggests that humans would be less
susceptible to these tumors than rats (see discussion in Section 4.7.3.1.4). Overall, the mode of
action analysis supports the conclusion that biphenyl should not pose a risk of urinary bladder
tumors in humans at exposure levels that do not cause calculi formation.
Mechanistic data to support a mode of action for biphenyl-induced liver tumors in the
mouse are not available (see Section 4.7.3.2). In the absence of information to indicate
otherwise, the development of liver tumors in female BDFi mice with chronic exposure to
biphenyl (Umeda et al., 2005) is assumed to be relevant to humans. EPA acknowledges that
some mouse strains are relatively susceptible to liver tumors and the background incidence of
this tumor can be high, and that the use of mouse liver tumor data in risk assessment has been a
subject of controversy (e.g., King-Herbert and Thayer. 2006). According to historical control
data from JBRC, the institute that conducted the mouse bioassay published by Umeda et al.
(2005), the mean incidences of liver tumors (hepatocellular adenoma or carcinoma) in male and
female control BDFi mice are 32.2 and 7.1%, respectively. These incidences are consistent with
the concurrent controls in the mouse bioassay of biphenyl. The relatively low background
incidence of liver tumors in female control mice from Umeda et al. (2005) minimizes the
possible confounding of compound-related liver tumors in this sex.
While the cancer descriptor "likely to be carcinogenic to humans" is plausible and the
positive evidence of tumors at two sites in two species raises a concern for carcinogenic effects
in humans, this assessment attaches some weight to (1) the lack of evidence for either tumor type
in a second study, strain, or species and (2) a mode of action for urinary bladder tumors specific
in experimental animal studies to the male rat and consistent with these tumors as a high-dose
phenomenon closely related to the formation of urinary bladder calculi. Recognizing that each
cancer descriptor covers a continuum of evidence, this assessment concludes that biphenyl shows
"suggestive evidence of carcinogenic potential."
EPA's Cancer Guidelines (U.S. EPA. 2005a) indicate that for tumors occurring at a site
other than the initial point of contact, the cancer descriptor may apply to all routes of exposure
61
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
that have not been adequately tested at sufficient doses. An exception occurs when there is
convincing toxicokinetic data that absorption does not occur by other routes. Information
available on the carcinogenic effects of biphenyl demonstrates that tumors occur in tissues
remote from the site of absorption following chronic oral exposure (urinary bladder in male rats
and liver in female mice). No information on the carcinogenic effects of biphenyl via the
inhalation or dermal routes in humans and animals is available. Studies in rats, rabbits, and
guinea pigs demonstrate that biphenyl is rapidly and extensively absorbed by the oral route of
exposure, and an in vitro model using human skin provides evidence of dermal absorption of
biphenyl (Fasano, 2005). Qualitative evidence for absorption of inhaled biphenyl comes from
inhalation toxicity studies in rats and mice that reported systemic (liver and kidney) effects
following inhalation exposure to biphenyl for 46-90 days (Sun Company Inc., 1977a;
Deichmann et al., (1947); Monsanto, (1946)). A case report of hepatic toxicity produced by a
probable combination of inhalation and dermal exposures in a worker in a biphenyl-impregnated
fruit wrapping paper production facility (Hakkinen et al.. 1973) provides qualitative evidence of
human absorption by these routes. Therefore, based on the observation of systemic tumors
following oral exposure and limited qualitative evidence for inhalation and dermal absorption, it
is assumed that an internal dose will be achieved regardless of the route of exposure. In the
absence of information to indicate otherwise, the database for biphenyl provides "suggestive
evidence of carcinogenic potential" by all routes of exposure.
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence
Available human studies were not designed to evaluate associations between exposure to
biphenyl and occurrence of cancer (see Section 4.1). As discussed in Section 4.2,
carcinogenicity studies in animals are limited to the oral exposure route.
Urinary bladder tumors were found in F344 male rats in a well-designed 2-year cancer
bioassay by Umeda et al. (2002). This is a rare tumor type, not having been observed in
historical control male F344 rats of the JBRC or the NTP—1,148 and 1,858 rats, respectively, as
reported by Umeda et al. (2002). Although the other available bioassays evaluated exposure
ranges comparable to those used by Umeda et al. (2002), they did not report increased urinary
bladder tumors. It is plausible that these other studies could not confirm or contradict these
findings due either to smaller group sizes and shorter effective exposure durations. In the 75-
week dietary study in Wistar rats (Shiraiwa et al., 1989), some of the male rats exhibited urinary
bladder calculi and simple or diffuse hyperplasia and papillomatosis of the urinary bladder
mucosa in the absence of neoplastic lesions. The duration, being much shorter than the standard
104-week bioassay, may not have been long enough to observe later occurring tumors. Ambrose
et al. (1960) exposed albino rats to biphenyl in the diet at concentrations ranging from 10 to
10,000 ppm for 2 years; urinary bladder tumors occurred in most groups. Because of decreased
survival in rats exposed to 5,000 or 10,000 ppm and the evaluation of histopathology only for
62
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
rats surviving to study termination (as few as 2 per group at the higher doses), however, this
study was not adequate for evaluation of the tumorigenic potential of biphenyl. In the 2-year
dietary study of biphenyl conducted by Dow Chemical Co. (1953) in Sprague-Dawley rats
(12/sex/group), a pneumonia outbreak (resulting in deaths of all control male rats by the end of
one year), relatively small group sizes, and decreased survival may have impaired the ability to
detect late-developing tumors. In addition, these studies were conducted in other rat strains (i.e.,
Wistar, Sprague-Dawley) that might not demonstrate the same response as F344 rats.
Evidence concerning liver tumors includes positive findings in one sex of one species
(i.e., female BDFi mice) from a well-conducted 2-year dietary study in by Umeda et al. (2005).
Male mice in this study showed decreases in liver tumors with increasing dose, but within the
range of historical controls for the laboratory. There was no liver tumor response in either sex of
B6C3Fi mice or B6AKF1 mice (NCI, 1968), but these evaluations were carried out at a lower
exposure than those used by Umeda et al. (2005) and for a shorter duration (18 months rather
than 24 months). There was no observed liver tumor response in female ddY mice (Imai et al.,
1983)—males were not tested—with exposure at a level intermediate to the higher exposures
tested by Umeda et al. (2005). Umeda et al. (2005) suggested that the difference in reponse
between the two studies might be due to differences in susceptibility between the two mouse
strains, but specific support for this hypothesis is not available.
In an 18-month NCI (1968) bioassay that used just one biphenyl dose group, the
incidence of reticular cell sarcoma was significantly elevated in one strain of female mice, but
not in male mice of the same strain or in either sex of mice of a second strain. In light of the
inconsistency in this finding across mouse strains and sexes in NCI (1968) and lack of
confirmation in other studies in mice, the biological significance of the elevated incidence of
reticular cell sarcoma in female mice of one strain is unclear. On the other hand, it is notable
that this study started exposure during early life at one week of age, while the other available
studies in mice started later (i.e., 6 weeks for Umeda et al., 2005).
The evidence for genotoxicity of biphenyl and its metabolites is reviewed in Appendix C,
Tables C-2 and C-3, and is summarized in Section 4.5.2. The in vitro evidence does not indicate
that biphenyl is mutagenic; however, in vivo data suggest that biphenyl metabolites that are
capable of redox cycling may induce genetic damage resulting from oxidative damage and
cytotoxicity.
4.7.3. Mode-of-Action Information
4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats
4.7.3.1.1. Hypothesized mode of action. The best-supported hypothesis proposes a mode of
action whereby the formation of urinary bladder calculi (from the precipitation of
4-hydroxybiphenyl-O-sulphate) is a key event in the development of urinary bladder tumors in
male rats fed high levels of biphenyl in the diet for 2 years. According to this hypothesis, the
63
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
calculi (occurring in association with increased urinary pH and potassium, and predominantly
composed of 4-hydroxybiphenyl-O-sulphate) cause irritation to transitional epithelial cells of the
urinary bladder leading to sustained cell proliferation, which promotes the development of
initiated cells in the urinary bladder with progression to papillomas and carcinomas.
4.7.3.1.2. Experimental support for the hypothesized mode of action
Strength. consistency, and specificity of association. including support for the
hypothesized mode of action in male rats. The formation of urinary bladder calculi,
predominantly composed of potassium 4-hydroxybiphenyl-O-sulphate, is strongly, consistently,
and specifically associated with the formation of urinary bladder tumors in male rats chronically
exposed to high dietary concentrations of biphenyl. Several findings support this association.
Urinary bladder calculi were formed at a high prevalence (43/50; 86%) in a group of male rats
exposed to biphenyl in the diet at a concentration of 4,500 ppm, but were absent in male rats
receiving diets containing 0, 500, or 1,500 ppm biphenyl (TJmeda et al.. 2002). These
observations were consistent with the detection of urinary bladder transitional cell papilloma
(10/50; 20%), carcinoma (24/50; 48%), and papilloma or carcinoma (31/50; 62%) in the
4,500 ppm group of male rats and total absence of urinary bladder papilloma or carcinoma in the
control, 500, or 1,500 ppm groups of male rats. Bladder calculi were found in all 24 of the male
rats with urinary bladder transitional cell carcinoma and in 8/10 of the male rats with transitional
cell papilloma.
The association between urinary bladder calculus formation and development of urinary
bladder tumors is supported by the species and gender specificity of calculi and tumor
development. Urinary bladder calculi were observed in female rats only at 4,500-ppm biphenyl
in the diet and at a lower incidence (8/50; 16%) than in male rats; no urinary bladder transitional
cell papillomas or carcinomas were observed in any female rats (Umeda et al.,2002). The
available evidence suggests that differences in physical properties and chemical composition of
calculi in male and female rats account for the gender difference in development of urinary
bladder tumors (Umeda et al.. 2002; Ohnishi et al.. 2000b). Urinary bladder calculi in male rats
are formed by irreversible chemical reactions; these calculi have been described as triangular,
pyramidal, or cubical in shape, 0.3-1 cm in size, and composed primarily of potassium
4-hydroxybiphenyl-O-sulphate. In contrast, urinary bladder calculi in female rats are of
homogeneous size, spheroidal in shape, and primarily composed of 4-hydroxybiphenyl and
potassium bisulphate (which are hydrolysis products of potassium 4-hydroxybiphenyl-O-
sulphate) (Umeda et al.. 2002; Ohnishi et al.. 2000b). The calculi formed in female rats may
undergo reversible hydroxylation reaction and are less stable than those formed in males
(Ohnishi et al., 2000b). Umeda et al. (2005) suggested that the physical characteristics of the
calculi in male rats lead to mechanical damage to the urinary bladder epithelium not induced by
calculi in female rats and, hence, to tumor formation. There was no evidence of biphenyl-
64
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
induced urinary bladder calculi or bladder tumors in male or female BDFi mice receiving dietary
biphenyl at concentrations as high as 6,000 ppm for 2 years (TJmeda et al.. 2005).
Gender differences in urinary conditions of the rat (including pH and potassium
concentrations) and sulphatase activities in kidneys may be responsible for the gender
differences in urinary calculi composition and formation and the subsequent development of
urinary bladder tumors in male, but not female, F344 rats (Ohnishi et al., 2001, 2000a, 2000b).
Urinary bladder calculi in male rats were associated with significantly increased urinary pH
(average pH of 7.97 in the 4,500 ppm group at the final week of exposure compared to 7.66 in
controls) (TJmeda et al.. 2002). The urine pH of female rats exposed to 4,500 ppm for 104 weeks
(pH = 7.26) was not elevated compared with controls (pH = 7.29) (TJmeda et al.. 2002). Ohnishi
et al. (2000b) fed biphenyl, biphenyl and potassium chloride (KC1), biphenyl and sodium
bicarbonate (NaHCCb), or biphenyl and potassium bicarbonate (KHCO3) to male F344 rats for
13 weeks. Urine crystals were found only in rats coadministered biphenyl and KHCO3. These
observations suggest that the formation of the calculi results from the precipitation of the
potassium salt of the sulphate conjugate of 4-hydroxybiphenyl under the elevated pH conditions
of the male rat urine. The mechanism responsible for increased urinary pH in 4,500-ppm male is
not known.
Relatively strong, consistent, and specific associations between calculi formation and
transitional cell hyperplasia and between transitional cell hyperplasia and the development of
transitional cell tumors in the urinary bladder have been shown in male F344 rats chronically
exposed to high concentrations of biphenyl in the diet. Urinary bladder transitional cell
hyperplasia (simple, nodular, papillary) occurred in 45/50 (90%) male rats receiving biphenyl in
the diet for 2 years at the same dietary concentration (4,500 ppm) that induced urinary bladder
calculi formation (43/50; 86%) and transitional cell tumors (31/50; 62%) (Umeda et al.. 2002).
Forty-two of the 45 male rats with urinary bladder transitional cell hyperplasia also exhibited
urinary bladder calculi. In another study, evidence of biphenyl-induced calculi formation
(microcalculi in the urine) and increased indices of urinary bladder transitional cell proliferation
(greater than fourfold increase in BrdU incorporation) in male F344 rats was reported following
as little as 4-8 weeks of dietary exposure to 5,000 ppm biphenyl (Shibata et al.. 1989b).
A mode of action involving calculi formation, ulcerations or inflammation, subsequent
hyperplasia, and urinary bladder tumor induction has been proposed for other chemicals,
including melamine, uracil, and the sodium salt of 2-hydroxybiphenyl, that induce urinary
bladder tumors in rodents (Capen et al.. 1999; IARC. 1999a); (IARC. 1999b; Cohen. 1998.
1995). These findings provide further evidence that calculi formation and subsequent
degenerative changes are involved in the etiology of rodent urinary bladder tumors. It is not
unusual to see extensive proliferation or hyperplasia in bladder epithelium in response to urinary
calculi from other rodent bladder tumorigens without an associated ulceration or intense
inflammatory response. In male rats exposed to 4,500 ppm biphenyl, increasing numbers of rats
65
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
with clinical hematuria were observed beginning at about the 40th week of exposure, and
histologic examinations at study termination revealed focal hyperplasia in 45/50 rats, providing
some evidence of calculi-induced bladder epithelial damage followed by cell proliferation
(TJmeda et al.. 2002). Over the course of the study, 94% of male rats with hematuria had bladder
or kidney calculi. In addition, with 8 weeks, but not 4 weeks, of exposure to 5,000 ppm biphenyl
in the diet, moderate urinary bladder epithelial hyperplasia and microcalculi in urine were
observed in 5/5 male F344 rats, but no descriptions of degenerative changes were provided; these
observations are consistent with a rapid repair response to epithelial damage from biphenyl-
induced urinary tract calculi (Shibata et al.. 1989b).
The ability of repeated biphenyl exposure to promote previously initiated urinary bladder
cells to bladder tumors is supported by results of a bladder tumor initiation-promotion study
(Kurata et al.. 1986). Incidences of urinary bladder hyperplasia, papilloma, and carcinoma were
significantly increased in male F344 rats initiated with dietary BBN for 4 weeks followed by
5,000 ppm biphenyl in the diet for 32 weeks, compared with rats receiving BBN only for
4 weeks. For example, 94 and 83% of rats treated with BBN followed by biphenyl developed
urinary bladder hyperplasia and papillomas, respectively, compared with 25 and 12% of rats
exposed to BBN alone.
Dose-resyonse concordance. Dose-response relationships for urinary bladder calculi formation,
transitional cell hyperplasia, and transitional cell tumor development show concordance in the
Umeda et al. (2002) rat bioassay. In male rats, urinary calculi, nonneoplastic lesions (epithelial
hyperplasia), and neoplastic lesions (papillomas and carcinomas) of the urinary bladder were
observed only at the highest exposure level (4,500 ppm); no urinary bladder calculi, transitional
cell hyperplasia, or transitional cell tumors were found in control, 500, or 1,500 ppm male rats.
Furthermore, urinary bladder calculi were found in 43/45 high-dose male rats, in all 24 male rats
with transitional cell carcinoma, and in 8/10 male rats with transitional cell papilloma.
Temporal relationship. Results from the 2-year oral study in rats (Umeda et al.. 2002) provide
some evidence of a progression from urinary bladder calculi formation to the development of
bladder tumors. Urinary bladder calculi were observed in the first 4,500 ppm male rat that died
(week 36), evidence of blood in the urine was observed in 4,500 ppm male rats by week 40, and
incidences of bladder calculi and bloody urine that paralleled increases in mortality and tumor
formation were observed throughout the remainder of the study. In addition, results of a short-
term oral study demonstrate that microcalculi can be detected in the urine of male rats after as
little as 4 weeks of dietary exposure to 5,000 ppm biphenyl and that hyperplasia of urinary
bladder epithelium can be detected at least by week 8 (Shibata et al.. 1989b). Presumably, the
development of biphenyl-induced urinary bladder tumors requires a longer exposure period to
66
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
urinary calculi of sufficient size, shape, and composition to induce urinary bladder epithelial
damage and a sustained proliferative response.
Biological plausibility and coherence. The proposed mode of action is consistent with the
current understanding of cancer biology and is supported by the body of evidence that other
chemicals with primarily nongenotoxic profiles produce urinary bladder tumors in rodents at
high exposure levels by a mode of action involving calculi formation, ulceration or
inflammation, and regenerative cell proliferation (Capen et al.. 1999; IARC. 1999a. b; Cohen.
1998. 1995). Additional information could strengthen the plausibility and coherence of the
proposed mode of action to explain the occurrence of biphenyl-induced urinary bladder tumors
in male rats. These additional data include results from investigations of earlier time points in
the proposed temporal progression from calculi formation to epithelial damage, regenerative cell
proliferation, and tumor development and further investigations into the factors underlying
gender-specific differences in precipitation of 4-hydroxybiphenyl-O-sulphate to form bladder
calculi in rats.
4.7.3.1.3. Other possible modes of action for bladder tumors in male rats. The available data
suggest there may be some ability of biphenyl or its metabolites to induce genetic damage.
Genotoxicity testing of 2-hydroxybiphenyl, which is associated with the development of urinary
bladder tumors in male rats, provides mixed results. The induction of genotoxic effects by 2-
hydroxybiphenyl in the urinary bladder epithelium leading to tumor initiation is proposed to
occur via redox cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone generating
reactive oxygen species resulting in oxidative DNA damage (Balakrishnan et al., 2002; Pathak
and Roy, 1993; Morimoto et al., 1980). However, no DNA adducts or DNA binding in urinary
bladder epithelial tissue was found in rats following short-term (Kwok et al.. 1999) or subchronic
(Smith et al.. 1998) oral exposure to 2-hydroxybiphenyl at high doses associated with the
formation of urinary bladder tumors. 2-Hydroxybiphenyl is a minor urinary metabolite of
biphenyl, constituting only a small fraction (0.1-1.0%, Meyer and Scheline, 1976) of the
metabolites produced. The metabolite of 2-hydroxybiphenyl responsible for the redox cycling,
2,5-dihydroxybiphenyl, was generally not detected (or detected in trace amounts) in the urine of
biphenyl-exposed rats (Meyer and Scheline, 1976). Overall, key mutational events consistent
with a mutagenic mode of action for urinary bladder tumors (e.g., mutations in urinary bladder
epithelial tissue leading to initiation of tumor cells) are not supported by the available data.
Support for a proposed mutagenic mode of action caused by oxidative DNA damage would come
from studies showing, for example, formation of 2,5-dihydroxybiphenyl and
phenylbenzoquinone in the urinary bladder epithelium of rats exposed to low doses of biphenyl.
67
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4.7.3.1.4. Conclusions about the hypothesized mode of action for bladder tumors in male rats
Support for the hypothesized mode of action in rats. There is strong evidence that urinary
bladder tumors in male rats chronically exposed to biphenyl in the diet is a high-dose
phenomenon involving sustained occurrence of calculi in the urinary bladder leading to
transitional cell damage, sustained regenerative cell proliferation, and eventual promotion of
spontaneously initiated tumor cells in the urinary bladder epithelium.
To summarize, chronic exposure of male rats to a high dietary concentration of biphenyl
(4,500 ppm) caused increased urinary pH and high prevalence of urinary bladder calculi (from
the precipitation of 4-hydroxybiphenyl-O-sulphate in the urine), transitional cell hyperplasia, and
transitional cell tumors. Incidences of male rats with calculi and those with bladder tumors were
strongly correlated, and chronic exposure of male rats to lower dietary concentrations of
biphenyl (500 and 1,500 ppm) did not increase urinary pH and did not cause calculi formation,
transitional cell hyperplasia, or bladder tumor development. There were relatively strong
associations between incidences of rats with calculi and those with transitional cell hyperplasia
and between incidences of rats with transitional cell hyperplasia and bladder tumors. In contrast,
high concentrations of biphenyl in the diet of female rats had no effect on urinary pH, caused a
much lower prevalence of urinary bladder calculi of a different composition, and resulted in no
urinary bladder tumors. The urinary bladder calculi in the male rats were mainly composed of
the conjugated biphenyl metabolite, potassium 4-hydroxybiphenyl-O-sulphate, whereas those of
the female rats were predominantly composed of 4-hydroxybiphenyl and potassium bisulphate
(which are hydrolysis products of potassium 4-hydroxybiphenyl-O-sulphate). There was no
evidence of urinary bladder calculi formation or tumor development in male and female mice
exposed to similar dietary concentrations of biphenyl. Results of a tumor initiation-promotion
study in male rats support the proposal that biphenyl-induced sustained cell proliferation
promotes initiated tumor cells in the urinary bladder.
Relevance of the hypothesized mode of action to humans. The proposed mode of action is
expected to be relevant to humans at exposure levels sufficient to cause urinary bladder calculi in
human because calculi resulting from human exposure to other substances have been associated
with urinary bladder irritation, regeneration, and cancer (Capen et al.. 1999; Cohen. 1998. 1995).
Four case-control studies of urinary bladder cancer in white human populations found RRs for an
association between a history of urinary tract stones and bladder carcinomas ranging from about
1.0 to 2.5 (Capen et al.. 1999). In addition, sulphate conjugation of hydroxylated biphenyl
metabolites has been demonstrated in human tissues (see Section 3.3), suggesting that humans
have the potential to develop calculi.
The underlying physiological factors determining the precipitation of 4-hydroxybiphenyl-
O-sulphate in urine to form calculi in male rats, but not female rats, exposed to high dietary
biphenyl concentrations are unknown. Elevated urine pH appears to play a role in the induction
68
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
of urinary bladder tumors by biphenyl in the male rat (Umeda et al.. 2002). Because humans on
average have a slightly more acidic urine than the rat (Cohen, 19951 it is possible that humans
might be less susceptible than the rat to the development of urinary bladder calculi. Another
physiological factor potentially contributing to reduced susceptibility of humans is the difference
in posture between rodents and humans. Based on the anatomy of the urinary tract in humans
and their upright, bipedal stature, calculi are either quickly excreted in urine or cause obstruction,
leading to pain and subsequent therapeutic removal of the calculi (Cohen. 1998. 1995). In
contrast, the rodent horizontal quadruped stature is expected to promote calculi residency time in
the bladder without causing obstruction (Cohen. 1998. 1995). Given the lack of understanding
of physiological factors that influence susceptibility in rats and the absence of specific human
data on biphenyl-induced calculi or urinary stones, there is uncertainty in extrapolation of the
dose-response relationship for biphenyl-induced calculi formation in male rats to humans.
Populations or lifestages particularly susceptible to the hypothesized mode of action.
Increased risks for bladder carcinoma in humans have been associated with cigarette smoking,
occupational exposure to polycyclic aromatic hydrocarbons, exposure to Shistosoma
haematobium that causes urinary tract inflammation, and a history for urinary tract infections in
general (Pelucchi et al., 2006; IARC, 1999). As such, people with these types of exposure or
history may be susceptible to urinary bladder irritation leading to bladder cancer, but evidence
supporting this inference is lacking. People with kidney failure, kidney tubular acidosis, urinary
tract infection, and vomitting are found to have alkaline urine (Israni and Kasiske, 2011), and
therefore could be susceptible to biphenyl-induced calculi formation. In addition, there are
conditions (bladder diverticuli, neurogenic bladder, and staghorn renal pelvic calculi) that can
increase the residency time of calculi in humans; thus, individuals with these conditions may also
be particularly susceptible to biphenyl-induced bladder tumors under the hypothesized mode of
action. Specific evidence supporting these potential susceptibilities is lacking.
4.7.3.2. Mode-of-A ction Information for Liver Tumors in Female Mice
Evidence that chronic oral exposure to biphenyl can cause liver tumors comes from the
2-year BDFi mouse bioassay by Umeda et al. (2005). Exposure to 2,000 or 6,000 ppm biphenyl
in the diet, but not to 667 ppm, produced increased incidences of hepatocellular adenomas or
carcinomas in female mice, but no carcinogenic response in male BDFi mice. Earlier studies
found no liver carcinogenic response in B6C3Fi or B6AkFi mice exposed to 517 ppm biphenyl
in the diet for 18 months (NCI, 1968) or in ddY female mice exposed to 5,000 ppm biphenyl in
the diet for 2 years (Imai et al.. 1983). The only investigations into the mode of action for
biphenyl-induced liver tumors in mice involve examinations of indicators of peroxisome
proliferation following biphenyl exposure (Umeda et al.. 2004a; Sunouchi et al.. 1999). Thus, an
69
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
evaluation of a mode of action involving peroxisome proliferation-activated receptors (PPARs)
follows.
4.7.3.2.1. Hypothesized mode of action for liver tumors in female mice. Proliferation of
peroxisomes is regulated by a class of ligand-activated transcription factors known as PPARs.
Peroxisome proliferators (PPARa agonists) are a structurally diverse group of non- or weakly
mutagenic chemicals that activate the PPARs and induce peroxisome proliferation as well as a
suite of responses including the induction of tumors in rats and mice. A mode of action for
PPARa agonists involving the following key events has been proposed: PPARa agonists activate
PPARa to transcribe genes involved in peroxisome proliferation, cell cycling/apoptosis, and lipid
metabolism. The changes in gene expression lead to changes in cell proliferation and apoptosis,
and to peroxisome proliferation. Suppression of apoptosis coupled with increased cell
proliferation allows transformed cells to persist and proliferate, resulting in preneoplastic hepatic
foci and ultimately promotion of tumor growth via selective clonal expansion (Klaunig et al..
2003).
Peroxisome proliferation was once thought to be the sole mode of action for
hepatocarcinogenesis induced by PPARa agonists; however, new information in PPARa-null
mice (Ito et al., 2007) and in transgenic mouse strains (Yang et al., 2007) have shown that
peroxisome proliferation may be neither required nor adequate for hepatocarcinogenicity, and
many molecular pathways in different cell types in the liver may contribute to liver cancer
development (Guyton et al., 2009). Nonetheless, the remainder of this section considers the extent
to which the available experimental data provide support for biphenyl as a PPARa agonist.
4.7.3.2.2. Experimental support for the hypothesized mode of action for liver tumors in female
mice
Data for a possible association between biphenyl-induced proliferation of peroxisomes
and liver tumors is limited to findings in BDFi mice exposed to biphenyl in the diet for 13 weeks
(Umeda et al., 2004). Identification of peroxisomes was based on light microscopy, with
electron microscopic confirmationy performed for liver tissue samples from 2 control group and
2 high-dose (16,000 ppm) female mice; no specific staining for peroxisome (e.g., using 3,3'-
diaminobenzidene) was performed. Umeda et al. (2004) reported hepatocellular peroxisome
proliferation in the livers of female BDFi mice exposed to biphenyl in diet for 13 weeks, but not
in male mice. In female mice, evidence of peroxisome proliferation was limited to the 16,000-
ppm dose group; no peroxisome proliferation was induced in female mice fed biphenyl at dietary
concentrations of 500, 2,000, 4,000, 8,000, or 10,000 ppm. Importantly, Umeda et al. (2004) did
not observe peroxisome proliferation at concentrations (2,000 and 6,000 ppm) that produced
statistically significantly increased incidences of liver tumors in the two-year bioassay in female
BDFi mice (Umeda et al., 2005). Although peroxisome proliferation was examined in female
70
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
mice exposed to biphenyl for 13 weeks (Umeda et al., 2004), whereas liver tumors were
observed after two years of exposure (Umeda et al., 2005), a 13-week exposure to biphenyl
should have been sufficient to demonstrate induction of peroxisome proliferation. Other studies
of PPARa agonists suggest that peroxisome proliferation in the mouse liver (as confirmed by
electron microscopy) could occur as early as 10-14 days after treament (Nakajima et al., 2000;
DeAngelo et al., 1989; Elcombe et al., 1985).
As reported in an abstract only, activities of 2 enzymes associated with PPARa
activation—potassium cyanide-insensitive palmitoyl CoA oxidase (PCO) in liver homogenate
and lauric acid 12-hydroxylation in liver microsomes—were significantly increased (up to 1.9-
and 3.8-fold, respectively) in female BDFi mice given oral doses up to 5.2 mmol/kg-day
biphenyl (800 mg/kg-day) for 3 days (Sunouchi et al.. 1999). Because PCO activity can vary
greatly in both baseline measure and response to chemical exposure, it is not necessarily a
consistent indicator of peroxisome proliferation (Laughter et al., 2004; Parrish et al., 1996;
Goldsworthy and Popp, 1987; Melnick et al., 1987).
In summary, the available data are not adequate to demonstrate that biphenyl acts as a
PPARa agonist or that PPARa agonism is involved in the mode of action for biphenyl-induced
liver tumors. In particular, the biphenyl dose associated with peroxisome proliferation in female
BDFi mice as reported by Umeda et al. (2004) is not concordant with doses associated with liver
tumor induction in Umeda et al. (2005).
4.7.3.2.3. Other possible modes of action for liver tumors in mice. As discussed in
Section 4.5.6, the available data suggest there may be some ability of biphenyl to induce genetic
damage. A genotoxic mode of action for biphenyl-induced liver tumors in mice could be
proposed based on the large metabolic capacity of the mouse liver to convert biphenyl to
hydroxylated metabolites and evidence that metabolites of 2-hydroxybiphenyl (2,5-
dihydroxybiphenyl and 2,5'-benzoquinone) can produce DNA damage (Tani et al.. 2007;
Balakrishnan et al.. 2002; Sasaki et al.. 2002; Sasaki et al.. 1997; Pathak and Roy. 1993;
Morimoto et al.. 1989). However, hydroxylation of biphenyl to produce 2-hydroxybiphenyl
appears to be a minor metabolic pathway in mice administered single i.p. doses of 30 mg
biphenyl/kg (Halpaap-Wood et al.. 1981b). and the available data are inadequate to establish that
this genotoxic mode of action operates in the biphenyl induction of liver tumors in mice. There
have been no in vitro or in vivo investigations of biphenyl-induced DNA adducts or ROS
generation in mouse liver cells or of possible gender differences in the production of biphenyl-
induced DNA adducts or other genotoxic events.
4.7.3.2.4. Conclusions about the hypothesized mode of action for liver tumors in mice. A
PPARa agonism mode of action for liver tumors in female mice exposed to 2,000 or 6,000 ppm
biphenyl in the diet for 2 years is not supported by the experimental data. This is based on the
71
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
limited investigation of biphenyl as a PPARa agonist and, in the one available subchronic study,
lack of concordance between dose-response relationships for biphenyl-induced liver tumors and
proliferation of hepatocellular peroxisomes in female mice. Available data are inadequate to
support alternative modes of action that propose direct or indirect genotoxic events from reactive
biphenyl metabolites or ROS, respectively, as key events.
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.8.1. Possible Childhood Susceptibility
No information was identified that would specifically suggest an early childhood
susceptibility for biphenyl toxicity. However, the developmental profiles of superoxide
dismutase and catalase in humans that were reported by McElroy et al. (1992) indicate that the
activities of both enzymes may be comparatively low before and at birth, placing humans in the
perinatal period at an increased risk of adverse effects elicited by quinoid metabolites of
biphenyl. Specifically, Buonocore et al. (2001) drew attention to the fact that the human brain
has relatively low superoxide dismutase activity at birth. Given the limited data on age-specific
ROS scavenging enzymes, any suggestions of childhood susceptibility to biphenyl is speculative.
Studies in animals provide evidence that biphenyl metabolism is mediated by CYP1A2
and CYP3A4 (Haugen. 1981). Phase II enzymes, such as sulphotransferases (SULTs) and
UGTs, may be involved in conjugation activities with hydroxybiphenyls in mammalian tissues
(Pacifici et al.. 1991; Bock et al.. 1980). CYP1A2 expression is negligible in the early neonatal
period, but is significantly increased to 50% of adult levels by 1 year of age (Sonnier and
Cresteil. 1998). In general, SULTs and UGTs, depending on the isoforms, also exhibit
differential expression during human development (Duanmu et al.. 2006; Strassburg et al.. 2002).
To the extent that metabolism increases or reduces the toxicity of biphenyl, changes in the
expression of Phase I and II enzymes during development can influence susceptibility to
biphenyl toxicity. Specific isoforms of CYPs and Phase II enzymes have not been identified as
the principal catalyzers involved in biphenyl metabolism and the effect of differences in enzyme
expression on childhood susceptibility to biphenyl has not been established.
4.8.2. Possible Gender Differences
Benford and Bridges (1983) evaluated the sex- and tissue-specific induction of biphenyl
2-, 3-, and 4-hydroxylase activities in microsomal preparations or primary hepatocyte cultures
from male and female Wistar rats. No differences in biphenyl hydroxylase activities were
observed between the sexes. However, there were some sex differences in the way tissues
responded to the action of enzyme inducers. For example, the CYP1A inducer a-naphthoflavone
strongly induced 2-hydroxylase in male liver but had no effect on female liver. Betamethasone
induced 2-hydroxylase activity in female liver but inhibited it in male liver. The available
limited human data do not suggest that gender differences exist in the response to biphenyl
72
DRAFT - DO NOT CITE OR QUOTE
-------
1 exposure. However, available animal data suggest gender-related differences in susceptibility to
2 tumors (i.e., bladder tumors in male, but not female, F344 rats and increased incidences of liver
3 tumors in female, but not male, BDFi mice administered biphenyl in the diet for a lifetime).
73
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
The RfD (expressed in units of mg/kg-day) is defined as an estimate (with uncertainty
spanning perhaps an order of magnitude) of a daily exposure to the human population (including
sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a
lifetime. It can be derived from a no-observed-adverse-effect level (NOAEL), lowest-observed-
adverse-effect level (LOAEL), or the 95 percent lower bound on the benchmark dose (BMDL),
with uncertainty factors (UFs) generally applied to reflect limitations of the data used.
5.1.1. Choice of Candidate Principal Studies and Candidate Critical Effects—with
Rationale and Justification
Human studies are preferred over animal studies when quantitative measures of exposure
are reported and the reported effects are determined to be associated with exposure (U.S. EPA.
2002); however, no information was located regarding possible associations between oral
exposure to biphenyl and health outcomes in humans. In experimental animals, kidney, urinary
bladder, liver, developmental toxicities and decreased body weight were identified as the major
effects of biphenyl exposure by the oral route (see Section 4.6.1).
Studies that reported these effects were evaluated using general study quality
considerations described in EPA guidance (U.S. EPA. 2002. 1994b). Among the chronic studies
that observed effects on the kidney, urinary bladder, and liver and on body weight, the studies by
Umeda et al. (2002) in the rat and Umeda et al. (2005) in the mouse were selected as candidate
principal studies for dose-response analysis. These were well-conducted studies performed in
accordance with Organisation for Economic Co-operation and Development (OECD) test
guidelines and Good Laboratory Practice (GLP). Both studies used three biphenyl dose groups
plus a control, 50 animals/sex/group, and comprehensive measurement of endpoints. Other
chronic studies that evaluated noncancer endpoints (Shiraiwa et al.. 1989; Ambrose et al.. 1960;
Pecchiai and Saffiotti. 1957; Dow Chemical Co. 1953) reported effects on the kidney and liver,
but the Umeda studies (Umeda et al.. 2005; Umeda et al.. 2002) were more comprehensive in the
outcomes evaluated and used larger group sizes, supporting the selection of these studies as
candidate principal studies.
Other subchronic and chronic studies were less informative as evaluations of the
noncancer toxicity of biphenyl, and were judged less suitable as candidate principal studies.
Endpoints evaluated by Shiraiwa et al. (1989) were limited to body weight, kidney weight, and
urinary calculi formation. The studies by (Ambrose et al.. 1960). Pecchiai and Saffiotti (1957).
and Dow Chemical Co. (1953) were conducted before the implementation of GLPs and used
smaller numbers of animals (8-15/sex/group), which reduced the power of the studies to identify
74
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
treatment-related effects. Neither Ambrose et al. (1960) nor Pecchiai and Saffiotti (1957)
identified the strain of rat used. The Dow Chemical Co. (1953) study was compromised by an
outbreak of pneumonia, causing death of all the control animals. Other chronic studies in mice
(Imai et al.. 1983; NCI. 1968) reported tumor data only.
Regarding kidney toxicity, the study by Umeda et al. (2002) showed the most sensitive,
dose-related measures of kidney effects in the F344 rat to be histopathological changes: renal
pelvis transitional cell nodular and simple hyperplasia (males and females), renal pelvis
mineralization (males and females), hemosiderin deposits (females only), and papillary
mineralization (males and females). These endpoints were selected as candidate critical effects
(see Table 5-1). Increased incidences of other histopathologic changes in the kidney (including
renal pelvis desquamation in male rats, renal pelvis calculi in male rats, mineralization of the
cortico-medullary junction in male rats, papillary necrosis in male and female rats, and infarct in
female rats) were observed in high-dose animals only, supporting a continuum of kidney effects
increasing in severity with higher exposure that could not be evaluated more comprehensively
without individual joint incidence data. While the latter endpoints were not selected for dose-
response analysis (see Table 4-5), they were taken into account qualitatively in interpreting the
results. In the male and female mouse (Umeda et al.. 2005). the most sensitive measures of
kidney toxicity were a dose-related increase in the incidence of mineralization in inner stripe of
the outer medulla of the kidney and increased urine BUN levels (see Tables 4-7 and 4-8). These
endpoints were selected as candidate critical effects.
Evidence of urinary bladder toxicity is limited to the rat. Umeda et al. (2002) reported
histopathologic changes of the bladder in high-dose F344 rats only, with incidences of lesions
higher in males than females (see Table 4-4). Histopathological examination showed that the
highest incidence of bladder lesions was for transitional cell hyperplasia (simple, nodular, and
papillary combined) in male rats; this histopathologic finding was selected as a candidate critical
effect. Because the response was more robust in males than that in females, dose-response data
for this endpoint in female rats was not modeled.
Liver toxicity associated with biphenyl exposure has been observed primarily in the
mouse. Increases in serum liver enzymes (i.e., AST, ALT, LDH, and AP) in female BDFi mice
observed by Umeda et al. (2005) (see Table 4-7) were the most sensitive measures of biphenyl -
related liver toxicity and were selected as candidate critical effects. In general, liver enzyme
levels in the male mouse did not show treatment-related changes and were not considered for
dose-response analysis.
In the 2-year Umeda studies (Umeda et al.. 2005; Umeda et al.. 2002). body weights at
terminal sacrifice were approximately 20% lower in high-dose F344 rats (males—378 mg/kg-
day; females—438 mg/kg-day) than controls and approximately 25-31% lower in high-dose
BDFi mice (males—1,050 mg/kg-day; females—1,420 mg/kg-day) compared to control. In rats,
depression of body weight gain throughout the majority of the study was apparent in high-dose
75
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
group male and female animals only, whereas biphenyl-related effects on body weight gain in
mice were observed to some extent in all dose groups. Therefore, body weight relative to the
control at terminal sacrifice in mice from Umeda et al. (2005) was selected as a candidate critical
effect.
In the only developmental toxicity study of biphenyl (Khera et al.. 1979). the incidence of
fetuses with missing or unossified sternebrae showed an increasing trend with dose that was
judged to be biologically significant below the exposure level associated with maternal toxicity.
Therefore Khera et al. (1979) was selected as a candidate principal study and incidence of
missing or unossified sternebrae in fetuses was selected as a candidate critical effect.
5.1.2. Methods of Analysis—Including Models (e.g., PBPK, BMD)
No biologically-based dose-response models are available for biphenyl. In this situation,
EPA evaluates a range of empirical dose-response models thought to be consistent with
underlying biological processes to model the dose-response relationship in the range of the
observed data. Consistent with this approach, all standard models available as part of EPA's
Benchmark Dose Software (BMDS, version 2.1.2) were evaluated.
Datasets modeled included selected nonneoplastic lesions in the urinary system of F344
rats exposed to biphenyl in the diet for 2 years (Umeda et al. (2002); see Table 5-1);
mineralization in the kidney, clinical chemistry parameters, and body weight of BDFi mice
exposed to biphenyl in the diet for 2 years (Umeda et al. (2005); see Table 5-2); and fetuses with
missing or unossified sternebrae from Wistar rat dams administered biphenyl by gavage on GDs
6-15 (Khera et al. (1979); see Table 5-3).
76
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-1. Datasets employed in the dose-response modeling of nonneoplastic
effects in the urinary tract of male and female F344 rats exposed to biphenyl in the
diet for 2 years
Males (n = 50)
Females (n = 50)
Biphenyl dietary concentration (ppm)
0
500
1,500
4,500
0
500
1,500
4,500
Calculated dose (mg/kg-d)
0
36.4
110
378
0
42.7
128
438
Effect
Renal pelvis
Nodular transitional cell hyperplasia
0
1
1
21
0
0
1
12
Simple transitional cell hyperplasia
6
8
5
19
3
5
12
25
Mineralization
9
6
10
18
12
12
18
27
Other kidney effects
Hemosiderin deposit3
0
0
0
0
4
8
22
25
Papillary mineralization
9
9
14
23
2
6
3
12
Bladder
Combined transitional cell hyperplasia13
0
0
0
45
1
0
1
10
aMale data for incidences of hemosiderin deposits not selected for quantitative analysis.
bFemale data for incidences of combined transitional cell hyperplasia not selected for quantitative analysis.
Source: Umeda et al. (20021.
77
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-2. Datasets employed in dose-response modeling of body weight,
selected clinical chemistry results, and histopathological kidney effects in
male and female BDFi mice exposed to biphenyl in the diet for 2 years
Endpoint
Biphenyl concentration in the diet (ppm)
0
667
2,000
6,000
Males
Dose (mg/kg-d)
0
97
291
1,050
Kidney histopathology
n= 50
n= 49
n= 50
n= 50
Mineralization inner stripe-outer medulla
9
8
14
14
Clinical chemistry parameter
n= 34
n= 39
n= 37
n= 37
BUN (mg/dL)
20.2 ±3.6
22.0 ±4.0
23.2 ±4.4
22.9 ±2.7
Body weight
n= 35
n= 41
n= 41
n= 39
Mean terminal body weight (g)
46.9 ±4.9
43.1 ±7.9
42.9 ±6.0
32.4 ±3.6
Females
Dose (mg/kg-d)
0
134
414
1,420
Kidney histopathology
n= 50
n= 50
n= 50
n= 49
Mineralization inner stripe-outer medulla
3
5
12
26
Clinical chemistry parameter
n= 28
n= 20
n= 22
n= 31
AST (IU/L)
75 ±27
120 ±110
211 ±373
325 ± 448
ALT (IU/L)
32 ± 18
56 ±46
134 ±231
206 ± 280
AP (IU/L)
242 ± 90
256±121
428 ± 499
556 ± 228
LDH (IU/L)
268 ± 98
461 ±452
838 ± 2,000
1,416 ±4,161
BUN (mg/dL)
14.9 ±2.0
14.8 ±3.4
21.0 ±20.5
23.8 ± 11.7
Body weight
n= 31
n= 22
n= 25
n= 32
Mean terminal body weight (g)
34.0 ±4.0
32.5 ±3.3
30.5 ±3.1
25.5 ±3.0
Source: Umeda et al. (2005).
1
Table 5-3. Dataset for dose-response modeling of incidence of fetuses with
missing or unossified sternebrae, from Wistar rat dams administered
biphenyl by gavage on GDs 6-15
Effect
Dose (mg/kg-d)
0
125
250
500
Fetuses with missing or unossified
4/176
3/236
4/213
16/199
sternebrae7animals examined
(number of litters examined)
(16)
(20)
(18)
(18)
aData from the 1000 mg/kg-day dose group was not included here because of frank maternal toxicity.
Source: Khera et al. (19791.
2
3 Consistent with EPA's Benchmark Dose Technical Guidance (U.S. EPA. 2012).
4 benchmark responses (BMRs) characterizing minimally biologically significant responses for
5 each endpoint were identified where possible. BMDs and BMDLs for body weight decrease
78
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
were calculated for a BMR of 10% decrease from the control (i.e., 10% relative deviation [RD])
because a 10% decrease in body weight is generally considered to represent a minimally
biologically significant effect (e.g., in determining maximum tolerated doses). For serum
enzyme activities (AST, ALT, AP, LDH), BMDs and BMDLs were calculated for a BMR of
100%) increase from the control (i.e., equivalent to a twofold increase, or a relative deviation of 1
(1 RD); denoted BMDird and BMDLird). Several expert organizations, particularly those
concerned with early signs of drug-induced hepatotoxicity, have identified an increase in liver
enzymes (AST, ALT, AP) compared with concurrent controls of two- to fivefold as an indicator
of concern for hepatic injury (EMEA. 2006; Boone et al.. 2005). Because LDH, like liver
enzymes, is one of the more specific indicators of hepatocellular damage in most animal species
and generally parallels changes in liver enzymes in toxicity studies where liver injury occurs, a
similar twofold increase in LDH is considered to indicate liver injury in experimental animals.
For reproductive and developmental studies with nested designs, a BMR of 5% extra risk
in individual offspring has been used analogously to 10%> extra risk in adults, to reflect greater
susceptibility during this critical window of development. To be able to use nested models, the
numbers of affected and total fetuses within each litter are required, which were not included in
the Khera et al. (1979) study report. An approach that uses dichotomous models to approximate
the result of nested models was used, as follows. First, note that although the BMD
corresponding to a particular fetal risk (e.g., 5% extra risk) can be estimated correctly using the
incidence of affected fetuses among the total number of live fetuses (Williams and Ryan. 1997;
Haseman and Kupper. 1979; Haseman and Hogan. 1975). it is the BMDL that cannot be
estimated correctly without the numbers of both affected and total fetuses within each litter to
calculate the variance. The correct variance estimate lies between the variance with total litters
as sample size and the variance with total fetuses as sample size (Rao and Scott 1992).
Consequently, the dichotomous models in BMDS were fit to the proportions of fetuses affected
in two separate analyses—one with the number of litters in each dose group as sample sizes, and
one with the total number of fetuses in each dose group as sample sizes (Table 5-3). These two
sets of modeling results bracket the BMDL that would result from nested modeling.
In the absence of information regarding what level of change is considered biologically
significant, the BMD and BMDL were estimated using a BMR of 10%> extra risk for
dichotomous data (e.g., hyperplasia), or a BMR of 1 standard deviation (SD) from the control
mean for continuous data (e.g., BUN). For all endpoints, these latter BMRs (a BMR of 1 SD for
continuous data or 10%> extra risk for dichotomous data) were also used to facilitate a consistent
basis of comparison across endpoints, studies, and assessments.
In general, adequate model fit was judged by the chi-square goodness-of-fit p-value (p >
0.1), visual inspection of the fit of the dose-response curve to the data points, scaled residuals,
and fit in the low-dose region and in the vicinity of the BMR. For continuous data, the
assumption of constant variance in the responses across each set of dose groups was tested. If
79
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
the assumption was met (p> 0.1), the fit of continuous models to the mean was evaluated while
assuming constant variance; if not, all models were evaluated while applying the power model
integrated into BMDS to account for nonhomogeneous variance.
If standard models failed to provide adequate fit to the data, modifications of these
standard models (i.e., parameter restriction adjustments) or use of alternative models were
considered in an effort to achieve adequate fit. Then if adequate fit could not be achieved, the
highest dose was dropped, and the entire modeling procedure was repeated. If no adequate fit
could be achieved after dropping the highest dose, then the dataset was regarded as not amenable
for BMD modeling.
Among all of the models providing adequate fit to a dataset, the model with the lowest
Akaike's Information Criterion (AIC) was chosen as the best-fitting model when the difference
between the BMDLs estimated from a set of models was less than threefold. Otherwise, the
model with the lowest BMDL was selected as the best-fitting model for a dataset (U.S. EPA.
2012). If datasets could be adequately modeled, the BMDLs from the selected models were used
as candidate PODs. If not, NOAEL or LOAEL values were considered as candidate PODs.
Summary modeling results are presented in Table 5-4 and Figure 5-1; more detailed
modeling results are presented in Appendix D (Tables D-4 through D-24 and respective model
output files). The BMDs and BMDLs shown in Table 5-4 and Figure 5-1 are those from the
best-fitting models for each endpoint. BMD and BMDL for serum AST levels in female mice
were derived after dropping the data from the highest dose groups.
80
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-4. Summary of candidate PODs for selected nonneoplastic effects
following oral exposure of rats and mice to biphenyl
Males
Females
Best fitting
model
BMR
Benchmark
result (mg/kg-d)
Best fitting
model
BMR
Benchmark result
(mg/kg-d)
BMD
BMDL
BMD
BMDL
F344 rats (limeda et al., 2002); biphenvl in the diet for 2 vrs
Kidney
Renal pelvis
Transitional cell
nodular hyperplasia
Logistic
10%
234
192
Multistage
2-degree
10%
274
212
Transitional cell
simple hyperplasia
Gamma
10%
314
113
Gamma
10%
71
52
Mineralization
Log-probit
10%
208
138
Multistage
1-degree
10%
88
56
Kidney - other
Hemosiderin deposit
NA
Dichotomous-
Hill
10%
45
23
Papillary
mineralization
Multistage
1-degree
10%
92
58
Logistic
10%
292
219
Bladder
Transitional cell
hyperplasia
Gamma
10%
205
147
NA
BDF, mice (Umcda et al.. 2005); biphenvl in the diet for 2 vrs
Kidney
Mineralization
Log-logistic
10%
721
276
Log-logistic
10%
233
122
Clinical chemistry
AST
NA
Power
1RD
1903
1223
ALT
NA
No adequate fit3
1RD
-
-
LDH
NA
No adequate fit3
1RD
-
-
AP
NA
No adequate fit3
1RD
-
-
BUN
No adequate fit3
1SD
-
-
No adequate fit3
1SD
-
-
Body weight
Terminal body wt.
No adequate fit3
0.1RD
-
-
Linear
0.1RD
583
511
Wistar rats (Khcra et al.. 1979); biphenvl bv savase to dams on GDs 6-15
Fetuses with missing or unossified sternebrae, sample size = number
of litters in each dose group
Log-logisticb
5%
477
173
Fetuses with missing or unossified sternebrae, sample size = number
of fetuses in each dose group
Multistage
3-degreeb
5%
460
382
""No adequate fit" indicates that none of the models in BMDS provided an adequate fit to the data. Where
BMD/BMDL values could not be derived, NOAELs were used as the POD. NOAELs for male mice: BUN-97
mg/kg-day; body weight-291 mg/kg-day. NOAELs for female mice: AP-414 mg/kg-day; ALT, LDH, andBUN-134
mg/kg-day.
bData from the 1,000 mg/kg-day dose group was not included because of frank maternal toxicity.
RD = relative deviation; SD = standard deviation.
81
DRAFT - DO NOT CITE OR QUOTE
-------
800
700
600
500
400
200
100
o. .E
CCL >.
Body weight
Bladder
Kidney
Developmental
Liver
¦ BMD ~BMDL • NOAEL
(1) = Umeda et al., 2005
(2) = Umeda et al., 2002
(3) = Khera et al., 1979
M = male
F = female
Figure 5-1. Candidate PODs for selected noncancer effects in rats and mice from repeated oral exposure to biphenyl.
82
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Selection of the Critical Effect
Based on the results of dose-response modeling presented in Table 5-4 and Figure 5-1,
the kidney of rats exposed to biphenyl in the diet for 2 years appears to be the most sensitive
target of biphenyl toxicity in both male and female F344 rats, with the lowest BMDio values
obtained. These results ranged from 45-92 mg/kg-day, corresponding to renal pelvis simple
transitional cell hyperplasia and mineralization (females), renal papillary mineralization (males),
and hemosiderin deposition (females). As discussed in Section 4.6.1, in the kidney medulla,
papillary mineralization falls on a continuum of effects progressing (at higher doses) to papillary
necrosis, and is consistent with a functional change in the kidney. Papillary mineralization was a
more sensitive endpoint among male rats than female rats, with BMDi0s of 92 and 292 mg/kg-
day, respectively. At the same time, the female rats showed more sensitive results than the males
for renal pelvis simple transitional cell hyperplasia and and mineralization, with BMDioS of 71-
88 mg/kg-day, compared with 208-314 mg/kg-day in the males. Although the BMDio for
hemosiderin deposits in the female rat was lower (by about twofold) than the value associated
with papillary mineralization, the biological relevance of hemosiderin deposits as reported in
Umeda et al. (2002) is unclear (see Section 4.6.1). Papillary mineralization in male rats was
selected as the critical effect and the basis for derivation of the RfD because it was judged to be
the more serious outcome in this range of BMDi0s, given its likely progression to necrosis at
higher exposures. Similar results for the other kidney histopathology outcomes support this
selection.
Derivation of Human Equivalent Doses
Human equivalent doses (i.e., HEDs) for oral exposures were derived from the PODs
estimated from the laboratory animal data, as described in EPA's Recommended Use of Body
Weight3/4 as the Default Method in Derivation of the Oral Reference Dose (U.S. EPA. 2011). In
this guidance, EPA advocates a hierarchy of approaches for deriving HEDs from data in
laboratory animals, with the preferred approach being physiologically-based toxicokinetic
modeling. Other approaches can include using chemical-specific information in the absence of a
complete physiologically-based toxicokinetic model. Since a validated human PBPK model for
biphenyl for extrapolating doses from animals to humans is not available, in lieu of either
chemical-specific models or data to inform the derivation of human equivalent oral exposures, a
body weight scaling to the 3/4 power (i.e., BW3/4) approach was applied to extrapolate
toxicologically equivalent doses of orally administered biphenyl from adult laboratory animals to
adult humans for the purpose of deriving an oral RfD. Consistent with EPA guidance (U.S.
EPA. 2011). the PODs estimated based on effects in adult animals was converted to HEDs
employing a standard dosimetric adjustment factor (DAF) derived as follows:
DAF = (BWa14/ BWh1 4),
Where BWa = animal body weight and BWh = human body weight
83
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Using a BWa of 0.25 kg for rats and a BWh of 70 kg for humans (U.S. EPA. 1988). the
resulting DAF for rats was 0.24, respectively. Applying this DAF to the POD identified for
effects in adult rats yields a PODhed as follows:
PODhed = laboratory animal dose (mg/kg-day) x DAF
The POD for deriving the RfD for biphenyl, i.e., the BMDLio for papillary mineralization
in male rats, was converted to a PODhed as follows:
PODhed = BMDLio (mg/kg-day) x DAF
= 58 mg/kg-day x 0.24
= 13.9 mg/kg-day
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs)
Consideration of available dose-reponse data led to the selection of the 2-year bioassay of
biphenyl in the F344 rat (Umeda et al.. 2002) and papillary mineralization as the principal study
and critical effect, respectively, for RfD derivation. The uncertainty factors (UFs), selected
based on EPA's A Review of the Reference Dose and Reference Concentration Processes (U.S.
EPA (2002): Section 4.4.5), addressed five areas of uncertainty resulting in a composite UF of
30. This composite UF was applied to the selected POD to derive an RfD.
• An UF of 3 (10°5 = 3.16, rounded to 3) was applied to account for uncertainty in
characterizing toxicodynamic differences between rodents and humans. Toxicokinetic
differences between rodents and humans were addressed through the use of BW3 4
scaling; an HED was calculated using a standard DAF according to EPA guidance (U.S.
EPA. 2011).
• An UF of 10 was applied to account for intraspecies variability in susceptibility to
biphenyl, as quantitative information for evaluating toxicokinetic and toxicodynamic
differences among humans are not available.
• An UF of 1 was applied for subchronic to chronic extrapolation in this assessment
because the candidate principal study was chronic in duration.
• An UF of 1 was applied for LOAEL to NOAEL extrapolation because the current
approach is to address this factor as one of the considerations in selecting a BMR for
BMD modeling. In this case, a BMR of 10% increased incidence of papillary
mineralization in the rat kidney was selected under the assumption that it represents a
minimal biologically significant change.
84
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
• An UF of 1 to account for database deficiencies was applied. The biphenyl database
includes chronic toxicity studies in rats (Umeda et al.. 2002; Shiraiwa et al.. 1989;
Ambrose et al.. 1960; Pecchiai and Saffiotti. 1957; Dow Chemical Co. 1953) and mice
(Umeda et al.. 2005; Imai et al.. 1983); subchronic toxicity studies in rats (Shibata et al..
1989b; Shibata et al.. 1989a; Kluwe. 1982; S0ndergaard and Blom. 1979; Booth et al..
1961) and mice (Umeda et al.. 2004a); a developmental toxicity study in rats (Kheraet
al.. 1979); and one- and three-generation reproductive toxicity studies in rats (Ambrose et
al.. 1960; Dow Chemical Co. 1953) that did not fully evaluate effects of biphenyl
exposure on reproductive function as would studies conducted using current study
protocols. Epidemiological studies provide some evidence that biphenyl may induce
functional changes in the nervous system at concentrations in excess of occupational
exposure limits. Seppalainen and Hakkinen (1975) reported abnormal EEG and ENMG
findings and increases in clinical signs in workers exposed to biphenyl during the
production of biphenyl-impregnated paper at concentrations that exceeded the
occupational limit by up to 100-fold, and Wastensson et al. (2006) reported an increased
prevalence of Parkinson's disease in a Swedish factory manufacturing biphenyl-
impregnated paper where exposures were likely to have exceeded the TLV of 1.3 mg/m .
Wastensson et al. (2006) acknowledged that chance is an alternative explanation for the
cases identified in the Swedish factory workers. Animal studies did not include
examination of sensitive measures of neurotoxicity. The 2-year oral bioassays in rats and
mice (Umeda et al.. 2005; Umeda et al.. 2002) did, however, include daily observations
for clinical signs and histopathological examination of nervous system tissues. No
nervous system effects were reported, suggesting that the nervous system is not a
sensitive target of oral biphenyl toxicity. Overall, the findings from studies of
occupational (predominantly inhalation) exposure to biphenyl introduce some
uncertainties in the characterization of biphenyl hazard, but were not considered a data
gap sufficient to warrant a database UF.
The RfD for biphenyl was calculated as follows:
RfD = PODhed UF
= 13.9 mg/kg-day ^ 30
= 0.46 mg/kg-day, or 0.5 mg/kg-day rounded to one significant figure
5.1.4. Previous RfD Assessment
The previous IRIS assessment for biphenyl, posted to the IRIS database in 1987, derived
an oral RfD of 0.05 mg/kg-day based on kidney damage in albino rats administered biphenyl for
2 years at dietary levels >0.5% (Ambrose et al.. 1960). U.S. EPA considered the dietary level of
0.1% (50 mg/kg-day using a food factor of 0.05/day) to represent a NOAEL due to the
85
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
following: (1) uncertainty in the significance of effects observed at lower doses as compared to
the more certain adverse effect level of 0.5% in the diet and (2) supportive findings of 0.1%
biphenyl as a NOAEL in an unpublished report of a subchronic rat feeding study and a three-
generation rat reproduction study performed by Stanford Research Institute (Dow Chemical Co.
1953). The NOAEL of 50 mg/kg-day was divided by a total UF of 1,000 (10 for extrapolation
from animals to humans, 10 for protection of sensitive human subpopulations, and a modifying
factor of 10 to account for intraspecies variability demonstrated in the threshold suggested by the
data in the chronic animal study).
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
"3
The RfC (expressed in units of mg/m ) is defined as an estimate (with uncertainty
spanning perhaps an order of magnitude) of a continuous inhalation exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. It can be derived from a NOAEL, LOAEL, or the 95
percent lower bound on the benchmark concentration (BMCL), with UFs generally applied to
reflect limitations of the data used.
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
Human studies are preferred over animal studies when quantitative measures of exposure
are reported and the reported effects are determined to be associated with exposure (U.S. EPA.
2002). The available human data for biphenyl are limited to two occupational epidemiology
studies and a case report of workers engaged in the production of biphenyl-impregnated fruit
wrapping paper (Carella and Bettolo. 1994; Seppalainen and Hakkinen. 1975; Hakkinen et al..
1973; Hakkinen et al.. 1971). None of these studies provided air monitoring data adequate to
characterize workplace exposures to biphenyl. Therefore, data from the available human studies
could not be used for dose-response analysis and derivation of an RfC.
Limited information is available regarding the effects of inhaled biphenyl in laboratory
animals. These studies were evaluated using general study quality considerations described in
EPA guidance (U.S. EPA. 2002. 1994b). In three separate studies that included repeated
"3
inhalation exposure of rabbits, rats, and mice to air containing 300, 40, or 5 mg/m of biphenyl,
respectively, for periods of 68-94 days (Deichmann et al.. 1947; Monsanto. 1946). rabbits
"3
exhibited no signs of exposure-related adverse effects at concentrations as high as 300 mg/m .
"3
Irritation of mucous membranes was observed in rats at concentrations of 40 and 300 mg/m .
Mice were the most sensitive to inhaled biphenyl; irritation of the upper respiratory tract was
noted at a concentration of 5 mg/m (Deichmann et al.. 1947; Monsanto. 1946). Limitations in
study design, including lack of control animals and use of a single exposure level, as well as
poorly reported study details preclude the use of these studies for RfC derivation.
86
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Repeated exposure of mice to biphenyl at vapor concentrations of 25 or 50 ppm
-3
(157.75 or 315.5 mg/m ) for 13 weeks resulted in high incidences of pneumonia and tracheal
hyperplasia, and high incidences of congestion and edema in the lungs, liver, and kidney (Sun.
1977a). Study limitations and lack of supporting data preclude the use of this study for deriving
an RfC for biphenyl. Measured biphenyl exposure concentrations varied greatly during the first
half of the 13-week exposure period; for example, in the high concentration group (target
concentration of 50 ppm), the measured concentrations ranged from 5 to 102 ppm during the first
45 exposure sessions. High mortality after 46 exposures (as a result of accidental overheating of
the chambers) necessitated the use of 46 replacement animals. Histopathological findings were
reported only for males and females combined. Reports of lung congestion and hemorrhagic
lungs in some control mice were not confirmed histopathologically, and congestion in the lung,
liver, and kidney were considered by the study pathologist a likely effect of the anesthetic used
for killing the mice. The severity of reported histopathologic lesions was not specified.
Given these deficiencies, the Sun Company Inc. (1977a) 13-week inhalation mouse
study, the only available study that employed at least subchronic-duration exposure and multiple
biphenyl exposure levels, is considered inadequate for RfC derivation. An RfC was not derived
due to the significant uncertainty associated with the inhalation database for biphenyl, and route-
to-route extrapolation was not supported in the absence of a physiologically based
pharmacokinetic (PBPK) model. Although an RfC cannot be derived, it should be noted that the
available inhalation data provides some evidence that inhalation exposure to biphenyl could
induce respiratory or systemic lesions.
5.2.2. Previous RfC Assessment
No RfC was derived in the previous (1985) IRIS assessment.
5.3. UNCERTAINTIES IN THE RfD AND RfC
This section provides a discussion of uncertainties associated with the derived toxicity
values. To derive the oral RfD, the UF approach (U.S. EPA. 2002. 1994b) was applied to a POD
of 13.9 mg/kg-day (see Section 5.1). Uncertainty factors were applied to the POD to account for
extrapolating from responses observed in an animal bioassay to a diverse human population of
varying susceptibilities. Uncertainties associated with the data set used to derive the biphenyl
RfD are more fully described below. The available database was determined to be inadequate
for deriving a chronic inhalation RfC for biphenyl (see Section 5.2).
Selection of the critical effect for RfD determination. The critical endpoint selected for
derivation of the RfD is increased incidence of kidney papillary mineralization in F344 rats as
reported by Umeda et al. (2002). The fact that kidney effects have been consistently associated
with biphenyl exposure in multiple oral studies in male and female rats (Umeda et al.. 2002;
Shiraiwa et al.. 1989; Ambrose et al.. 1960; Pecchiai and Saffiotti. 1957; Dow Chemical Co.
87 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
1953) and in one study in male and female mice (TJmeda et al.. 2005) provides a measure of
confidence that the kidney is a target of biphenyl toxicity. Kidney effects have not been reported
in populations exposed to biphenyl in the workplace, however, and there is some degree of
uncertainty associated with extrapolation of kidney effects in experimental animals to humans.
As discussed in Section 4.7.3.1.4 (in the context of the relevance of rat urinary bladder tumors to
humans), physiological factors such as urine pH appear to play a role in the formation of calculi
by biphenyl. To the extent that these physiological factors influence the renal response to
biphenyl, the response in humans and rodents to biphenyl could differ. The lack of
understanding of physiological factors that influence susceptibility to biphenyl exposure
introduces uncertainty in the RfD.
Dose-response modeling. BMD modeling was used to estimate the POD for the biphenyl
RfD. BMD modeling has advantages over a POD based on a NOAEL or LOAEL because, in
part, the latter are a reflection of the particular exposure concentration or dose at which a study
was conducted. A NOAEL or LOAEL lacks characterization of the entire dose-response curve,
and for this reason, is less informative than a POD obtained from BMD modeling. Although the
selected model (i.e., multistage model) provided the best mathematical fit to the papillary
mineralization data in the male rat, (as determined by the criteria described in Section 5.1.2), this
model does not necessarily have greater biological support over the various other models that
were available. Some BMDS models yielded estimates of the POD that were similar to the
selected POD, and other models yielded values for the POD approximately twofold higher than
the best fitting model.
Inadequate data to support RfC derivation. The available data do not support RfC
derivation (see Section 5.2.1). Nevertheless, limited findings from human reports and from
inhalation toxicity studies in experimental animals suggest that exposure to sufficiently high
concentrations of biphenyl can potentially result in effects on the lungs or other systemic targets.
The lack of adequate data to derive an RfC represents a significant data gap.
5.4. CANCER ASSESSMENT
As noted in Section 4.7.1, EPA concluded that there is "suggestive evidence of
carcinogenic potential" for biphenyl. The Guidelines for Carcinogen Risk Assessment (U.S.
EPA. 2005a) state: "When there is suggestive evidence, the Agency generally would not attempt
a dose-response assessment, as the nature of the data generally would not support one; however,
when the evidence includes a well-conducted study, quantitative analyses may be useful for
some purposes, for example, providing a sense of the magnitude and uncertainty of potential
risks, ranking potential hazards, or setting research priorities. In each case, the rationale for the
quantitative analysis is explained, considering the uncertainty in the data and the suggestive
nature of the weight of evidence. These analyses generally would not be considered Agency
consensus estimates."
88
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In this case, the carcinogenicity of biphenyl has been evaluated in two well-conducted 2-
year bioassays in rats and mice (TJmeda et al.. 2005; Umeda et al.. 2002) that provide evidence of
increased incidences of liver tumors in female BDFi mice and urinary bladder tumors in male
F344 rats. Considering these data and uncertainty associated with the suggestive nature of the
tumorigenic response, EPA concluded that quantitative analyses may be useful for providing a
sense of the magnitude of potential carcinogenic risk. Based on the weight of evidence, a dose-
response assessment of the carcinogenicity of biphenyl is deemed appropriate.
5.4.1. Choice of Study/Data—with Rationale and Justification
No information was located regarding possible associations between oral exposure to
biphenyl and cancer in humans. A review of the available chronic animal bioassays of biphenyl,
including strenghts and limitations, is provided in Section 4.7.2. The two most recent and well-
conducted animal bioassays found statistically significant associations between lifetime oral
exposure to biphenyl and tumor development. Biphenyl was associated with urinary bladder
tumors in male, but not female, F344 rats (Umeda et al.. 2002) and liver tumors in female, but
not male, BDFi mice (Umeda et al.. 2005). Tumor data for these two sites were selected for
dose-response analysis.
No studies were identified that examined the association between inhalation exposure to
biphenyl and cancer in humans or animals.
5.4.2. Dose-Response Data
The dose-response data for urinary bladder tumor formation resulting from lifetime oral
exposure of male and female F344 rats (Umeda et al.. 2002) are shown in Table 5-6. The dose-
response data for liver tumor formation resulting from lifetime oral exposure of male and female
BDFi mice (Umeda et al.. 2005) are shown in Table 5-7. The datasets selected for dose-response
analysis include urinary bladder transitional cell papilloma or carcinoma in male F344 rats and
liver adenoma or carcinoma in female BDFi mice. In both the urinary bladder and liver, benign
and malignant tumors were considered together because benign and malignant tumors in both of
these organs develop from the same cell lines and benign tumors can progress to carcinomas
(McConnell et al.. 1986) (U.S. EPA. 2005a).
89
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-6. Incidence data for tumors in the urinary bladder of male and
female F344 rats exposed to biphenyl in the diet for 2 years
Males
Females
Biphenyl dietary concentration (ppm)
0
500
1,500
4,500
0
500
1,500
4,500
Calculated dose (mg/kg-d)
0
36.4
110
378
0
42.7
128
438
Tumor incidence3
Transitional cell
Papilloma
0/50
0/50
0/50
10/49*
0/50
0/50
0/50
0/50
Carcinoma
0/50
0/50
0/50
24/49*
0/50
0/50
0/50
0/50
Papilloma or carcinoma
050
0/50
0/50
31/49**
0/50
0/50
0/50
0/50
aOne high-dose male rat was excluded from the denominator because it died prior to week 52. It is assumed that
this rat did not have a tumor and was not exposed for a sufficient time to be at risk for developing a tumor. Umeda
et al. (20021 did not specify the time of appearance of the first tumor.
Statistically significant (Fisher's exact test, p < 0.05) as reported by study authors.
"Statistically significant (Fisher's exact test, p < 0.05) as determined by EPA.
Source: Umeda et al. (2002).
1
Table 5-7. Incidence data for liver tumors in male and female BDFi mice
fed diets containing biphenyl for 2 years
Males
Females
Biphenyl dietary
concentration (ppm)
0
667
2,000
6,000
0
667
2,000
6,000
Reported dose (mg/kg-d)
0
97
291
1,050
0
134
414
1,420
Tumor incidence"
Adenoma
8/50
6/49
7/49
3/50
2/48
3/50
12/49*
10/48*
Carcinoma
8/50
8/49
5/49
4/50
1/48
5/50
7/49*
5/48
Adenoma or carcinoma
16/50
12/49
9/49
7/50
3/48
8/50
16/49*
14/48*
'One low-dose, one mid-dose male, two control, one mid-dose, and two high-dose female mice were excluded from
the denominators because they died prior to week 52. It is assumed that they did not have tumors and were not
exposed for a sufficient time to be at risk for developing a tumor. Umeda et al. (2005) did not specify the time of
appearance of the first tumor.
Statistically significant (Fisher's exact test, p < 0.05) as reported by study authors.
Source: Umeda et al. (2005).
2
3 5.4.3. Dose Adjustments and Extrapolation Method(s)
4 5.4.3.1. Liver Tumors in Female Mice
5 A scaling approach based on BW3 4 was used to extrapolate toxicologically equivalent
6 doses of orally administered dose from laboratory animals to humans. Mouse body weights from
7 Umeda et al. (2005) were estimated from data provided on average daily food consumption and
90
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
intake.4 Scaling factors were calculated using the EPA (1988) reference body weight for humans
(70 kg) and the average body weight for each dose group of female mice: (average body
weight/70)0 25 = scaling factor (U.S. EPA. 1992). The HED was calculated as: HED = scaling
factor x reported dose (Table 5-8).
Table 5-8. Scaling factors for determining HEDs to use for BMD modeling
of female BDFi mouse liver tumor incidence data from Umeda et al. (2005)
Biphenyl dietary concentration (mg/kg food)
667
2,000
6,000
Reported dose (mg/kg-d)
134
414
1,420
Reported average food consumption (kg/d)
0.0058
0.0059
0.0059
Average mouse body weight (kg)a
0.0289
0.0285
0.0249
Scaling factorb
0.143
0.142
0.137
HED (mg/kg-d)c
19
59
195
a(Biphenyl concentration in food [mg/kg food] x reported average food consumption [kg/day]) ^ reported average
daily dose of biphenyl (mg/kg-day) = calculated average mouse body weight (kg).
Calculated using reference body weight for humans (70 kg) (U.S. EPA. 1988). and the average body weights for
each dose group: mouse-to-human scaling factor = (average mouse body weight/70)0 25.
°HED = reported dose x scaling factor.
EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a) recommend that
when the weight of evidence evaluation of all available data are insufficient to establish the
mode of action for a tumor site and when scientifically plausible based on the available data,
linear extrapolation is used as a default approach. A linear approach to low-dose extrapolation
for biphenyl-induced liver tumors in female mice was selected because the mode of action for
this tumor site has not been established (see Section 4.7.3.2).
Incidence data for liver adenoma or carcinoma in the female mouse used to derive the
oral slope factor are presented in Table 5-9. Tumor incidence data were adjusted to account for
mortalities before 52 weeks; it was assumed that animals dying before 52 weeks were not
exposed for sufficient time to be at risk for developing tumors.
4 Umeda et al. (2005) provided average food consumption and biphenyl dose estimates for each exposure group
[Table 1 of (Umeda et al.. 2005)1. The study report did not include average body weights for the exposure groups.
Therefore, the biphenyl concentration in the food was multiplied by the corresponding average daily food
consumption value to determine the average daily biphenyl intake. Dividing this average daily biphenyl intake by
the author-calculated daily dose yielded the average body weight that would have been used by the study authors to
calculate the average daily biphenyl dose.
91
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Table 5-9. Incidence of liver adenomas or carcinomas in female BDFi mice
fed diets containing biphenyl for 2 years
Biphenyl dietary concentration (ppm)
0
667
2,000
6,000
HED (mg/kg-d)
0
19
59
195
Tumor incidence
Adenoma or carcinoma (combined)
3/48a
8/50
16/493*
14/483'*
aTwo control, one mid-dose, and two high-dose female mice were excluded from the denominators because they
died prior to week 52. It is assumed that they did not have tumors and were not exposed for a sufficient time to be
at risk for developing a tumor. Umeda et al. (20051 did not specify the time of appearance of the first tumor.
Statistically significant (Fisher's exact test, p < 0.05) as reported by study authors.
Source: Umeda et al. (20051.
The multistage-cancer model in the EPA BMDS (version 2.1.2), using the extra risk
option, was fit to the female mouse liver tumor incidence data. The multistage model5 has been
used by EPA in the vast majority of quantitative cancer assessments because it is thought to
reflect the multistage carcinogenic process and it fits a broad array of dose-response patterns.
The multistage model was run for all polynomial degrees up to n-1 (where n is the number of
dose groups including control). An extra risk of 10% tumor incidence was selected as the BMR,
consistent with EPA guidance (U.S. EPA. 2005a). as a 10% response corresponded to a POD
near the lower end of the observed range in the Umeda et al. (2005) bioassay data. Adequate
model fit was judged by the same three criteria used for noncancer modeling. If an adequate fit
to the data was not achieved with the multistage models, then the other dichotomous models
were fit to the data. If none of the models achieved an adequate fit for the full dataset, then the
highest dose was dropped and the entire modeling procedure was repeated.
When liver tumor incidence data for all dose groups were modeled, none of the models in
BMDS, including the multistage model, provided an adequate fit of the data (see Appendix E,
Table E-2). The incidence of liver tumors showed a plateau in animals in the two highest dose
groups. The lack of a monotonic increase in liver tumor incidence in the high-dose group could
not be attributed to higher mortality, as the survival rate in the high-dose group was comparable
to the control, low and medium dose groups. To better estimate responses in the low-dose
region, the high-dose group was excluded as a means of improving the fit of the model in the
region of interest. When the high-dose group was dropped, the multistage model provided an
adequate fit to the data (see Appendix E, Table E-2). The BMDhedio and BMDLhedio using this
latter dataset were 18.7 and 12.2 mg/kg-day, respectively. See Appendix E for more
information.
5 Multistage model is mathematically identical to multistage-cancer model.
92
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
5.4.3.2. Bladder Tumors in Male Rats
There is strong evidence that the occurrence of urinary bladder tumors in male rats
chronically exposed to biphenyl in the diet is a high-dose phenomenon involving occurrence of
calculi in the urinary bladder leading to transitional cell damage, sustained regenerative cell
proliferation, and eventual promotion of spontaneously initiated tumor cells in the urinary
bladder epithelium (see Section 4.7.3.1 for a detailed discussion of the hypothetized mode of
action for urinary bladder tumors in biphenyl-exposed male rats). Based on the proposed mode
of action, exposure to biphenyl at doses that would not result in calculi formation and subsequent
key events would not be associated with bladder tumors. As noted in the EPA Guidelines for
Carcinogen Risk Assessment (U.S. EPA. 2005a). a nonlinear approach to dose-response analysis
is used when there are sufficient data to ascertain the mode of action and conclude that it is not
linear at low doses and the agent does not demonstrate mutagenic or other activity consistent
with linearity at low doses. Therefore, consistent with the Cancer Guidelines, a nonlinear
extrapolation approach for biphenyl-induced urinary bladder tumors was selected.
Bladder calculi, the formation of which is a key event in the mode of action for urinary
bladder tumors, were observed in male rats in the Umeda et al. (2002) bioassay at a dose of 378
mg/kg-day; the NOAEL for this effect was 110 mg/kg-day. The HED for this NOAEL is 26
mg/kg-day, derived by application of a DAF of 0.24 (see Section 5.1.2 for discussion of the
DAF). A candidate RfD for bladder calculi of 0.9 mg/kg-day is derived by applying a composite
UF of 30 to this HED (see Section 5.1.3 for discussion of UFs). The RfD of 0.5 mg/kg-day
based on papillary mineralization in kidney is approximately twofold below the candidate RfD
for bladder calculi induction. Based on the proposed mode of action, it is anticipated that
exposure to biphenyl at doses that would not result in calculi formation would not be associated
with an increased risk of bladder tumors.
5.4.4. Oral Slope Factor and Inhalation Unit Risk
A low-dose linear extrapolation approach results in calculation of an oral slope factor that
describes the cancer risk per unit dose of the chemical at low doses. The oral slope factor was
calculated by dividing the extra risk (i.e., BMR of 10% extra risk) at the POD by the
corresponding BMDL (i.e, 0.1/BMDLhedio). Using linear extrapolation from the BMDLhedio,
3 1
the human equivalent oral slope factor of 8.2 x 10" (mg/kg-d)" (rounded to one significant
3 1
figure, 8 x 10" (mg/kg-d)" ) was derived for liver tumors in female BDFi mice (Table 5-10).
93
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 5-10. POD and oral slope factor derived from liver tumor incidence
data from BDFi female mice exposed to biphenyl in the diet for 2 years
Species/tissue site
BMDhedio
(mg/kg-d)
BMDLhedio
(mg/kg-d)
Slope factor" (risk per
[mg/kg-d])
Female mouse liver tumors
18.7
12.2
8 x 10"3
aHuman equivalent slope factor = 0.1/BMDLiOHed; see Appendix E for details of modeling results.
This slope factor should not be used with exposures >12.2 mg/kg-day (the POD for this
dataset), because above the POD, the fitted dose-response model better characterizes what is
known about the carcinogenicity of biphenyl (i.e., the slope factor may not approximate the
observed dose-response relationship adequately at exposure exceeding 12.2 mg/kg-day).
An inhalation unit risk for biphenyl was not derived in this assessment. The potential
carcinogenicity of inhaled biphenyl has not been evaluated in human or animal studies, and
route-to-route extrapolation was not possible in the absence of a PBPK model.
5.4.5. Uncertainties in Cancer Risk Values
5.4.5.1. Oral Slope Factor
A number of uncertainties underlie the cancer unit risk for biphenyl. Table 5-11
summarizes the impact on the assessment of issues such as the use of models and extrapolation
approaches (particularly those underlying the Guidelines for Carcinogen Risk Assessment (U.S.
EPA. 2005a). the effect of reasonable alternatives, the decision concerning the preferred
approach, and its justification.
Table 5-11. Summary of uncertainties in the biphenyl cancer slope factor
Consideration/
approach
Impact on slope
factor
Decision
Justification
Selection of data
set
No other studies or
tumor data sets with
MOA information
Umeda et al. (2005)
study was selected.
The bioassav bv Umeda et al. (2005) was a well
conducted experiment with sufficient dose
groups (four dose groups, including control) and
animal numbers (50 animals/sex/group).
Cross-species
scaling
Alternatives could t
or I slope factor (e.g.,
7-fold i [scaling by
BW] or twofold t
[scaling by B W23] for
mouse liver tumor)
Administered dose was
scaled to humans on
the basis of
equivalence of
mg/kg3/4-day (default
approach)
There are no data to support alternatives. Use of
[body weight]3 4 for cross-species scaling is
consistent with data that allow comparison of
potencies in humans and animals, and it is
supported by analysis of the allometric variation
of key physiological parameters across
mammalian species. No PBPK model is
available to derive internal doses.
94
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Table 5-11. Summary of uncertainties in the biphenyl cancer slope factor
Consideration/
approach
Impact on slope
factor
Decision
Justification
Extrapolation
procedure for rat
urinary bladder
tumors
No impact on the
slope factor because
the MOA for male rat
bladder tumors does
not support low-dose
linear extrapolation.
Nonlinear
extrapolation. The
RfD of 0.5 mg/kg-day
is based on a POD of
58 mg/kg-day, which
is -twofold lower than
the NOAEL for
bladder calculi
induction.
Available MOA data for urinary bladder tumors
support nonlinearity (i.e., that bladder tumor is a
high-dose phenomenon, and is closely related to
calculi formation in the urinary bladder of male
rats). An uncertainty analysis based on the
assumption that another mode of action for
urinary bladder tumors might be operative;
under this assumption, a linear extrapolation
approach was performed. See text of this
section.
Extrapolation
procedure for
mouse liver tumors
Departure from EPA's
Guidelines for
Carcinogen Risk
Assessment POD
paradigm, if justified,
could | or | slope
factor by an unknown
extent
Multistage model to
determine the POD,
linear low-dose
extrapolation from
POD (default
approach)
Available MOA data do not inform selection of
a dose-response model. Linear approach in the
absence of clear support for an alternative is
generally consistent with scientific deliberations
supporting EPA's Guidelines for Carcinogen
Risk Assessment.
Human relevance
of female mouse
liver tumor data
Human risk could for
I, depending on
relative sensitivity
Liver tumors in female
mice are relevant to
human exposure
It was assumed that humans are as sensitive as
the most sensitive rodent gender/species tested;
true correspondence is unknown.
Model uncertainty
For poorly fitting liver
tumors dataset,
alternatives could j or
t slope factor by an
unknown extent
Drop highest dose of
the liver tumor dataset.
Model options explored with the full liver tumor
dataset did not generate ap> 0.05. Dropping the
highest dose allowed a better fit to the low-dose
region of the data set.
Statistical
uncertainty at POD
i slope factor 1.5-fold
if BMD10 used rather
thanBMDL10
BMDL (default
approach for
calculating plausible
upper bound)
Limited size of bioassay results in sampling
variability; lower bound is 95% confidence limit
on dose.
Human population
variability /
sensitive
subpopulations
Low-dose risk f to an
unknown extent
Considered
qualitatively
No data to support range of human
variability/sensitivity in metabolism or response,
including whether children are more sensitive.
Two members of the peer review panel offered the views that the data do not prove that
bladder stones are required for carcinogenesis and that an alternative mode of carcinogenic
action was not adequately investigated. To explore the situation where the MOA is unknown, a
3 1
linear extrapolation approach was performed. A slope factor of 2 x 10" (mg/kg-day)" was
derived from a BMDLhedio of 41.2 mg/kg-day based on incidence of bladder tumors in male rats
and linear low-dose extrapolation from the BMDLhedio (see Appendix E for BMD modeling
documentation). This slope factor is lower than the slope factor derived from mouse liver
tumors, indicating that urinary bladder tumors are less likely than liver tumors at a given
exposure under the assumption of low-dose linearity. Because the available data support calculi
formation as a key event in the mode of action for male rat urinary bladder tumors, EPA does not
95
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
consider linear low-dose extrapolation to be supported for this tumor type.
The uncertainties presented in Table 5-11 have a varied impact on risk estimates. Some
suggest risks could be higher than was estimated, while others would decrease risk estimates or
have an impact of an uncertain direction. Several uncertainties are quantitatively characterized
for the significantly increased rodent tumors. These include the statistical uncertainty in the
multistage modeling estimate. Due to limitations in the data, particularly regarding the MOA
and relative human sensitivity and variability, the quantitative impact of other uncertainties of
potentially equal or greater impact has not been explored. As a result, an integrated quantitative
analysis that considers all of these factors was not undertaken.
5.4.5.2. Inhalation Unit Risk
The potential carcinogenicity of inhaled biphenyl has not been assessed. Therefore, a
quantitative cancer assessment for biphenyl by the inhalation pathway was not performed.
5.4.6. Previous Cancer Assessment
In the previous IRIS cancer assessment posted to the IRIS database in 1991, biphenyl was
listed in Group D; not classifiable as to human carcinogenicity based on no human data and
inadequate studies in mice and rats. Neither an oral slope factor nor inhalation unit risk was
derived in the previous cancer assessment.
96
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
6. MAJOR CONCLUSIONS IN Till CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE
6.1. HUMAN HAZARD POTENTIAL
6.1.1. Noncancer
Toxicokinetic studies of animals indicate that orally administered biphenyl is rapidly and
readily absorbed, distributed widely to tissues following absorption, and rapidly eliminated from
the body, principally as conjugated hydroxylated metabolites in the urine (Meyer. 1977; Meyer
and Scheline. 1976; Meyer etal.. 1976b; Meyer et al.. 1976a). Limited data show that biphenyl
can be absorbed by human skin (Fasano, 2005). Data for absorption, distribution, and
elimination are not available for inhaled biphenyl. Metabolism to a range of hydroxylated
metabolites has been demonstrated in in vitro systems with rat and human cells and tissues.
Human metabolism of biphenyl appears to be qualitatively similar to metabolism in the rat,
although some reports of quantitative differences are available (Powis et al.. 1989; Powis et al..
1988; Benford et al.. 1981V
Available human health hazard data consist of limited assessments of workers exposed to
biphenyl during the production or use of biphenyl-impregnated fruit wrapping paper in which
signs of hepatic and nervous system effects were observed.
Chronic oral studies in rats and mice identify the liver and urinary system as principal
targets of biphenyl toxicity. In rats exposed to biphenyl in the diet for two years, nonneoplastic
kidney lesions (including histopathological changes in the renal pelvis and papilla of the
medulla) were found at dietary concentrations >1,500 ppm (>128 mg/kg-day). Several other rat
studies provide supporting evidence that the kidney and other urinary tract regions are sensitive
targets for biphenyl in rats (Shiraiwa et al.. 1989; Ambrose et al.. 1960; Pecchiai and Saffiotti.
1957; Dow Chemical Co. 1953). In chronically exposed BDFi mice, increased incidence of
nonneoplastic effects on the kidney (mineralization) and liver (increased activities of plasma
ALT and AST) were found in females exposed to >2,000 ppm biphenyl in the diet (>414 mg/kg-
day) (Umeda et al.. 2005). In the only available developmental toxicity study for biphenyl, the
incidence of fetal skeletal anomalies (mainly missing or unossified sternebrae) showed a
significantly increasing trend with exposure to biphenyl on GDs 6-15 (Khera et al.. 1979).
Biphenyl effects on reproductive function in rats have been reported at exposure levels
higher than those associated with effects on the urinary tract, liver, or developing fetus. No
exposure-related effect on the number of dams with litters was found following exposure of male
and female albino rats to up to 5,000 ppm biphenyl in the diet (525 mg/kg-day) for 11 or 60 days
prior to mating (Ambrose et al.. 1960). In a three-generation rat study, decreased fertility,
97
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
decreased number of pups/litter, and decreased pup body weight were observed at 10,000 ppm
biphenyl in the diet (947 mg/kg-day), but not at <1,000 ppm (Dow Chemical Co. 1953).
No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
biphenyl (7 hours/day, 5 days/week for up to about 90 days) were reported in mice exposed to
3 3 3
5 mg/m and in rats exposed to 300 mg/m , but not in rabbits exposed to 300 mg/m (Deichmann
et al.. 1947; Monsanto. 1946). Congestion or edema of the lung, kidney, and liver, accompanied
by hyperplasia with inflammation of the trachea, was reported in CD-I mice exposed to biphenyl
vapors at 25 or 50 ppm (158 or 315 mg/m ) for 13 weeks (Sun. 1977a). In general, the toxicity
of inhaled biphenyl is poortly characterized because the available inhalation studies are limited
by study methodology and reporting issues.
6.1.2. Cancer
No assessments are available regarding possible associations between exposure to
biphenyl and increased risk of cancer in humans.
In a 2-year study of F344 rats administered biphenyl in the diet (Umeda et al.. 2002).
significantly increased incidences of urinary bladder tumors in males were observed at the
highest dose level (378 mg/kg-day). There is strong evidence that the occurrence of urinary
bladder tumors in male rats is a high-dose phenomenon involving occurrence of calculi in the
urinary bladder leading to transitional cell damage, sustained regenerative cell proliferation, and
eventual promotion of spontaneously initiated tumor cells in the urinary bladder epithelium.
Urinary bladder calculi in high-dose (438 mg/kg-day) female rats were observed at lower
incidence and were different in physical appearance and chemical composition; furthermore,
there were no urinary bladder tumors in any biphenyl-exposed female rats.
In a 2-year study of BDFi mice administered biphenyl in the diet (Umeda et al.. 2005).
the incidence of liver tumors in female mice was significantly increased at doses >414 mg/kg-
day, but not in males at doses up to and including 1,050 mg/kg-day. Available data are
insufficient to establish a mode of action for liver tumors in female mice.
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the
database for biphenyl provides "suggestive evidence of carcinogenic potential." This cancer
descriptor is based on an increase in the incidence of urinary bladder tumors (transitional cell
papillomas and carcinomas) in male F344 rats (Umeda et al.. 2002) and liver tumors
(hepatocellular adenomas and carcinomas) in female BDFi mice (Umeda et al.. 2005) exposed to
biphenyl in the diet for 104 weeks, as well as information on mode of carcinogenic action.
98
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
6.2. DOSE RESPONSE
6.2.1. Noncancer/Oral
The RfD of 0.5 mg/kg-day was based on an increased incidence of renal papillary
mineralization (Umeda et al., 2002). To derive the RfD, the PODHed was divided by a
composite UF of 30 (3 for animal-to-human extrapolation and 10 for human interindividual
variability in susceptibility). The interspecies uncertainty factor was applied to account for the
lack of quantitative information to assess toxicodynamic differences between animals and
humans. The intraspecies UF was applied to account for the lack of information regarding the
range of responses to biphenyl in the human population.
The overall confidence in the RfD assessment is medium to high. Confidence in the
principal study (Umeda et al., 2002) is high. Umeda et al. (2002) is a well-conducted study
performed in accordance with OECD test guidelines and GLPs. Confidence in the database is
medium to high. The database is robust in that it includes well-conducted chronic oral exposure
studies in the rat and mouse, other supporting repeat-dose studies in multiple species, a
developmental toxicity study in Wistar rats, and one- and three-generation reproductive toxicity
studies in rats. Confidence in the database is reduced because the reproductive toxicity studies
come from the older toxicological literature (1953 and 1960) and do not fully evaluate effects of
biphenyl exposure on reproductive function as would studies conducted using current study
protocols.
6.2.2. Noncancer/Inhalation
No inhalation RfC was derived due to the lack of inhalation studies of biphenyl toxicity
following chronic exposure and studies involving subchronic exposure that were inadequate for
RfC derivation. Repeated exposure of mice to biphenyl vapors for 13 weeks resulted in high
incidences of pneumonia and tracheal hyperplasia, and high incidences of congestion and edema
in the lungs, liver, and kidney (Sun. 1977a); however, study limitations and lack of supporting
data preclude the use of this study for deriving an RfC for biphenyl. Study limitations include
highly variable biphenyl exposure concentrations during the first half of the study, high mortality
after 46 exposures in one group of biphenyl-exposed mice due to an overheating event and
cannibalization that necessitated the use of replacement animals, and limitations in the reporting
of histopathological findings.
6.2.3. Cancer/Oral
"3
The oral slope factor of 8 x 10" per mg/kg-day is based on the tumor response in the liver
of female BDFi mice exposed to biphenyl in the diet for 2 years (Umeda et al.. 2005). The slope
factor was derived by linear extrapolation from a human equivalent BMDLio of 12.2 mg/kg-day
for liver adenomas or carcinomas.
99
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
A nonlinear extrapolation approach for biphenyl-induced urinary bladder tumors in male
rats was used because the available mode of action information indicates that the induction of
urinary bladder tumors is a high-dose phenomenon involving occurrence of calculi in the urinary
bladder leading to transitional cell damage, sustained regenerative cell proliferation, and eventual
promotion of spontaneously initiated tumor cells in the urinary bladder epithelium. Bladder
calculi were observed in male rats in the Umeda et al. (2002) bioassay at a dose of 378 mg/kg-
day; the NOAEL for this effect was 110 mg/kg-day. The HED for this NOAEL is 26 mg/kg-day,
derived by application of a DAF of 0.24 (see Section 5.1.2 for discussion of the DAF). A
candidate RfD for bladder calculi of 0.9 mg/kg-day is derived by applying a composite UF of 30
30 (3 for interspecies toxicodynamic differences, 10 for intraspecies variability in susceptibility)
to this HED. The RfD of 0.5 mg/kg-day based on papillary mineralization in kidney is
approximately twofold below the candidate RfD for bladder calculi induction. Based on the
proposed mode of action, it is anticipated that exposure to biphenyl at doses that would not result
in calculi formation would not be associated with an increased risk of bladder tumors.
6.2.4. Cancer/Inhalation
No human or animal data on the potential carcinogenicity of inhaled biphenyl are
available. Therefore, a quantitative cancer assessment for biphenyl by the inhalation pathway
was not performed.
100
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
7. REFERENCES
Abe. S; Sasaki. M. (1977). Chromosome aberrations and sister chromatid exchanges in Chinese
hamster cells exposed to various chemicals. J Natl Cancer Inst 58: 1635-1641.
ACGIH (American Conference of Governmental Industrial Hygienists). (2001). 1,1-Biphenyl. In
Documentation of the threshold limit values and biological exposure indices (7th ed.).
Cincinnati, OH.
Ambrose. AM; Booth. AN; DeEds. F; Cox. AJ. Jr. (1960). A toxicological study of biphenyl, a
citrus fungistat. Food Res 25: 328-336. http://dx.doi.org/10.Ill 1/j. 1365-
2621.1960.tb00338.x
Balakrishnan. S; Uppala. PT; Rupa. DS; Hasegawa. L; Eastmond. DA. (2002). Detection of
micronuclei, cell proliferation and hyperdiploidy in bladder epithelial cells of rats treated
with o-phenylphenol. Mutagenesis 17: 89-93. http://dx.doi.Org/10.1093/mutage/17.l.89
Benford. DJ; Bridges. JW. (1983). Tissue and sex differences in the activation of aromatic
hydrocarbon hydroxylases in rats. Biochem Pharmacol 32: 309-313.
http://dx.doi.org/10.1016/0006-2952(83^)90560-9
Benford. DJ; Bridges. JW; Boobis. AR; Kahn. GC; Brodie. MJ; Davies. DS. (1981). The
selective activation of cytochrome P-450 dependent microsomal hydroxylases in human
and rat liver microsomes. Biochem Pharmacol 30: 1702-1703.
http://dx.doi. org/10.1016/0006-2952(8 n90401-9
Bianco. PJ; Jones. RS; Parke. DV. (1979). Effects of carcinogens on biphenyl hydroxylation in
isolated rat hepatocytes. Biochem Soc Trans 7: 639-641.
http://dx.doi.org/10.1042/bst007Q639
Billings. RE; McMahon. RE. (1978). Microsomal biphenyl hydroxylation: The formation of 3-
hydroxybiphenyl and biphenyl catechol. Mol Pharmacol 14: 145-154.
Bock. KW; von Clausbruch. UC; Kaufmann. R; Lilienblum. W; Oesch. F; Pfeil. H; Piatt. KL.
(1980). Functional heterogeneity of UDP-glucuronyltransferase in rat tissues. Biochem
Pharmacol 29: 495-500. http://dx.doi.org/10.1016/0006-2952(80^)90368-8
Boehncke. A; Koennecker. G; Mangelsdorf. I; Wibbertmann. A. (1999). Concise international
chemical assessment document 6: Biphenyl. Geneva, Switzerland: World Health
Organization, http://www.who.int/ipcs/publications/cicad/en/cicad06.pdf
Boone. L; Meyer. D; Cusick. P; Ennulat. D; Bolliger. AP; Everds. N; Meador. V; Elliott. G;
Honor. D; Bounous. D; Jordan. H. (2005). Selection and interpretation of clinical
pathology indicators of hepatic injury in preclinical studies [Review], Vet Clin Pathol 34:
182-188. http://dx.doi.Org/10.llll/i.1939-165X.2005.tb00041.x
Booth. AN; Ambrose. AM; DeEds. F; Cox. AJ. Jr. (1961). The reversible nephrotoxic effects of
biphenyl. Toxicol Appl Pharmacol 3: 560-567. http://dx.doi.org/10.1016/0Q41-
008X(61N)90046-1
Bos. RP; Theuws. JL; Jongeneelen. FJ; Henderson. PT. (1988). Mutagenicity of bi-, tri- and
tetra-cyclic aromatic hydrocarbons in the "taped-plate assay" and in the conventional
salmonella mutagenicity assay. MutatRes 204: 203-206. http://dx.doi.org/10.1016/Q165-
1218(88^)90090-0
Boutwell. RK; Bosch. DK. (1959). The tumor-promoting action of phenol and related
compounds for mouse skin. Cancer Res 19: 413-424.
Brams. A; Buchet. JP; Crutzen-Favt. MC; De Meester. C; Lauwerys. R; Leonard. A. (1987). A
comparative study, with 40 chemicals, of the efficiency of the Salmonella assay and the
SOS chromotest (kit procedure). Toxicol Lett 38: 123-133.
Brouns. RE; Poot. M; de Vrind. R; von Hoek-Kon. T; Henderson. PT; Kuvper. CMA. (1979).
Measurement of DNA-excision repair in suspensions of freshly isolated rat hepatocytes
after exposure to some carcinogenic compounds: Its possible use in carcinogenicity
101
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
screening. Mutat Res Environ Mutagen Relat Subj 64: 425-432.
http://dx.doi.org/10.1016/0165-1161(79^)90112-2
Buonocore. G: Perrone. S: Bracci. R. (2001). Free radicals and brain damage in the newborn
[Review], Biol Neonate 79: 180-186. http://dx.doi.Org/10.l 159/000047088
Burke. MP: Bridges. JW. (1975). Biphenyl hydroxylations and spectrally apparent interactions
with liver microsomes from hamsters pre-treated with phenobarbitone and 3-
methylcholanthrene. Xenobiotica 5: 357-376.
http://dx.doi.org/10.3109/004982575090561Q6
Capen. CC: Dvbing. E; Rice. JM; Wilbourn. JD. (1999). Species differences in thyroid, kidney
and urinary bladder carcinogenesis. In CC Capen; E Dybing; JM Rice; JD Wilbourn
(Eds.), (pp. 1-225). Lyon, France: International Agency for Research on Cancer.
Carella. G: Bettolo. PM. (1994). Reversible hepatotoxic effects of diphenyl: Report of a case and
a review of the literature [Review], J Occup Med 36: 575-576.
Chung. KT; Adris. P. (2002). Growth inhibition of intestinal bacteria and mutagenicity of
aminobiphenyls, biphenyl and benzidine [Abstract], Abstracts of the General Meeting of
the American Society for Microbiology 102: 10.
Chung. KT: Adris. P. (2003). Growth inhibition of intestinal bacteria and mutagenicity of 2-, 3-,
4-aminobiphenyls, benzidine, and biphenyl [Review], Toxicol In Vitro 17: 145-152.
http://dx.doi.org/10.1016/S0887-2333(02X)0131-5
Cline. JC: McMahon. RE. (1977). Detection of chemical mutagens: Use of concentration
gradient plates in a high capacity screen. Res Comm Chem Pathol Pharmacol 16: 523-
533.
Cohen. SM. (1995). Cell proliferation in the bladder and implications for cancer risk assessment
[Review], Toxicology 102: 149-159. http://dx.doi.org/10.1016/0300-483X(95N)03044-G
Cohen. SM. (1998). Cell proliferation and carcinogenesis [Review], Drug Metab Rev 30: 339-
357. http://dx.doi.org/10.3109/036025398Q8996317
Deichmann. WB; Kitzmiller. KV; Dierker. M; Witherup. S. (1947). Observations on the effects
of diphenyl, O- and P-aminodiphenyl, O- and P- nitrodiphenyl and
dihydroxyoctachlorodiphenyl upon experimental animals. J Ind Hyg Toxicol 29: 1-13.
Dow Chemical Co (Dow Chemical Company). (1939). Toxicity of diphenyl and diphenyl oxide
(sanitized). (EPA Document No. 86-890001205S). Midland, MI.
http://www.ntis. gov/search/product.aspx?ABBR=QTS0520717
Dow Chemical Co (Dow Chemical Company). (1953). Toxicological study of diphenyl in citrus
wraps with cover letter. Menlo Park, CA: Stanford Research Institute.
http://www.ntis.gov/search/product.aspx?ABBR=QTS0206456
Dow Chemical Co (Dow Chemical Company). (1983). Partition coefficients of biphenyl,
diphenyl oxide and dowtherm a between 1-octanol and wateranother look. (EPA/OTS
Doc No. 878213735). Midland, MI.
http://www.ntis.gov/search/product.aspx?ABBR=QTS0206456
Duanmu. Z; Weckle. A: Koukouritaki. SB: Hines. RN: Falany. JL; Falany. CN: Kocarek. TA;
Runge-Morris. M. (2006). Developmental expression of aryl, estrogen, and
hydroxysteroid sulfotransferases in pre- and postnatal human liver. J Pharmacol Exp Ther
316: 1310-1317. http://dx.doi.org/10.1124/ipet.105.093633
EMEA (European Medicines Agency). (2006). Draft guidelines on detection of early signals of
drug-induced hepatoxicity in non-clinical studies. London, United Kingdom: Committee
for Medicinal Products for Human Use.
Fasano. WJ. (2005). Biphenyl: In vitro dermal absorption rate testing. (15667). Washington, DC:
Biphenyl Working Group.
102
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Fuiita. H; Koiima. A; Sasaki. M; Higara. K. (1985). [Mutagenicity test of antioxidants and
fungicides with Salmonella typhimurium TA97a, TA102], Tokyo-toritsu Eisei
Kenkyusho KenkyuNenpo 36: 413-417.
Garberg. P; A kerb loin. EL; Bolcsfoldi. G. (1988). Evaluation of a genotoxicity test measuring
DNA-strand breaks in mouse lymphoma cells by alkaline unwinding and hydroxyapatite
elution. MutatRes 203: 155-176. http://dx.doi.org/10.1016/Q165-l 161(88)90101-X
Garrett. NE; Stack. HF; Waters. MP. (1986). Evaluation of the genetic activity profiles of 65
pesticides. MutatRes 168: 301-325.
Glatt. H; Anklam. E; Robertson. LW. (1992). Biphenyl and fluorinated derivatives: Liver
enzyme-mediated mutagenicity detected in Salmonella typhimurium and Chinese hamster
V79 cells. MutatRes 281: 151-156.
Gollapudi. BB; Schisler. MR: Golden. RM. (2007). Evaluation of biphenyl fp in the mouse bone
marrow micronucleus test. Midland, MI: Dow Chemical Company.
Hakkinen. I; Siltanen. E; Hernberg. S: Seppalainen. AM: Karli. P; Vikkula. E. (1973). Diphenyl
poisoning in fruit paper production: A new health hazard. Arch Environ Health 26: 70-
74.
Hakkinen. I; Vikkula. E; Hernberg. S. (1971). The clinical picture of diphenyl poisoning
[Abstract], Scand J Clin Lab Invest 27: 53.
Halpaap-Wood. K; Horning. EC: Horning. MG. (1981a). The effect of 3-methylcholanthrene,
Aroclor 1254, and phenobarbital induction on the metabolism of biphenyl by rat and
mouse 9000g supernatant liver fractions. Drug Metab Dispos 9: 103-107.
Halpaap-Wood. K; Horning. EC: Horning. MG. (1981b). The effect of phenobarbital and beta-
naphthoflavone induction on the metabolism of biphenyl in the rat and mouse. Drug
Metab Dispos 9: 97-102.
Hanada. S. (1977). Studies on food additives, diphenyl (biphenyl) and o-phenyl phenol from the
view point of public health: Part 2. On the toxicities of diphenyl and o-phenyl phenol.
Nagoya-shiritsu Daigaku Igakkai Zasshi 28: 983-995.
Haseman. JK; Hogan. MP. (1975). Selection of the experimental unit in teratology studies.
Teratology 12: 165-171. http://dx.doi.org/10.1002/tera. 1420120209
Haseman. JK: Kupper. LL. (1979). Analysis of dichotomous response data from certain
toxicological experiments. Biometrics 35: 281-293.
Haugen. DA. (1981). Biphenyl metabolism by rat liver microsomes: Regioselective effects of
inducers, inhibitors, and solvents. Drug Metab Dispos 9: 212-218.
Haworth. S: Lawlor. T; Mortelmans. K; Speck. W: Zeiger. E. (1983). Salmonella mutagenicity
test results for 250 chemicals. Environ Mutagen 5: 3-142.
http://dx.doi.org/10.1002/em.28600507Q3
Hellmer. L; Bolcsfoldi. G. (1992). An evaluation of the E. coli K-12 uvrB/recA DNA repair
host-mediated assay: I. In vitro sensitivity of the bacteria to 61 compounds. Mutat Res
272: 145-160. http://dx.doi.org/10.1016/0165-1161(92)90043-L
Houk. VS: Schalkowsky. S: Claxton. LP. (1989). Development and validation of the spiral
Salmonella assay: An automated approach to bacterial mutagenicity testing. Mutat Res
223: 49-64. http://dx.doi.org/10.1016/0165-1218(89)90062-1
HSPB (Hazardous Substances Pata Bank). (2005). Biphenyl - CASRN 92-52-4 [Patabase],
Bethesda, MP: National Library of Medicine. Retrieved from
http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSPB
Hsia. MTS: Kreamer. BL; Polara. P. (1983a). A rapid and simple method to quantitate
chemically induced unscheduled PNA synthesis in freshly isolated rat hepatocytes
facilitated by PNA retention of membrane filters. Mutat Res 122: 177-185.
http://dx.doi. org/10.1016/0165-7992(83)90057-X
103
PRAFT - PO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Hsia. MTS; Kreamer. BL; Dolara. P. (1983b). Quantitation of chemically induced DNA damage
and repair in isolated rat hepatocytes by a filter elution method. In AW Hayes; RC
Schnell; TS Miya (Eds.), Developments in the science and practice of toxicology:
Proceedings of the Third International Congress on Toxicology held in San Diego,
California, USA, August 28-September 3, 1983 (pp. 375-378). New York, NY: Elsevier
Science.
I ARC (International Agency for Research on Cancer). (1999a). Melamine. In Some chemicals
that cause tumours of the kidney or urinary bladder in rodents and some other substances
(pp. 329-338). Lyon, France.
http://monographs.iarc.fr/ENG/Monographs/vol73/mono73.pdf
IARC (International Agency for Research on Cancer). (1999b). ortho-Phenylphenol and its
sodium salt. In Some chemicals that cause tumours of the kidney or urinary bladder in
rodents and some other substances (pp. 451480). Lyon, France.
http://monographs.iarc.fr/ENG/Monographs/vol73/index.php
Imai. S: Morimoto. J: Sekigawa. S. (1983). Additive toxicity test of thiabendazole and diphenyl
in mice. Nara Igaku Zasshi 34: 512-522.
Innes. JRM; Ulland. BM; Valerio. MG: Petrucelli. L; Fishbein. L; Hart. ER; Pallotta. AJ; Bates.
RR; Falk. HL; Gart. JJ; Klein. M; Mitchell. I: Peters. J. (1969). Bioassay of pesticides
and industrial chemicals for tumorigenicity in mice: A preliminary note. J Natl Cancer
Inst 42: 1101-1114.
Inoue. S: Yamamoto. K; Kawanishi. S. (1990). DNA damage induced by metabolites of o-
phenylphenol in the presence of copper(II) ion. Chem Res Toxicol 3: 144-149.
http://dx.doi.org/10.1021/txQ0014a010
Ishidate. M. Jr; Odashima. S. (1977). Chromosome tests with 134 compounds on Chinese
hamster cells in vitro: A screening for chemical carcinogens. Mutat Res 48: 337-353.
http://dx.doi. org/10.1016/0027-5107(77^)90177-4
Ishidate. MJ; Sofuni. T; Yoshikawa. K; Havashi. M; Nohmi. T; Sawada. M; Matsuoka. A.
(1984). Primary mutagenicity screening of food additives currently used in Japan. Food
Chem Toxicol 22: 623-636. http://dx.doi.org/10.1016/0278-6915(84^)90271 -0
Ito. N: Fukushima. S: Shirai. T; Hagiwara. A: Imaida. K. (1984). Drugs, food additives and
natural products as promoters in rat urinary bladder carcinogenesis. In M Borzsonyi; K
Lapis; NE Day; H Yamasaki (Eds.), Models, mechanisms and etiology of tumour
promotion: Proceedings of a symposium held in Budapest on 16-18 May 1983 (pp. 399-
407). Lyon, France: International Agency for Research on Cancer.
Johnston. RV; Schwetz. BA; Middleton. JJ: Lisowe. RW. (1976). Cytogenetic effects of
diphenyl-99 on rat bone marrow cells. Johnston, RV; Schwetz, BA; Middleton, JJ;
Lisowe, RW.
Khera. KS: Whalen. C: Angers. G: Trivett. G. (1979). Assessment of the teratogenic potential of
piperonyl butoxide, biphenyl, and phosalone in the rat. Toxicol Appl Pharmacol 47: 353-
358. http://dx.doi.org/10.1016/0041 -008X(79N)90330-2
Kitamura. S: Sanoh. S: Kohta. R; Suzuki. T; Sugihara. K; Fuiimoto. N: Ohta. S. (2003).
Metabolic activation of proestrogenic diphenyl and related compounds by rat liver
microsomes. J Health Sci 49: 298-310.
Klaunig. JE; Babich. MA: Baetcke. KP; Cook. JC: Corton. JC: David. RM; Deluca. JG: Lai. DY;
Mckee. RH; Peters. JM; Roberts. RA; Fenner-Crisp. PA. (2003). PPARalpha agonist-
induced rodent tumors: Modes of action and human relevance [Review], Crit Rev
Toxicol 33: 655-780. http://dx.doi.org/10.1080/713608372
Kluwe. WM. (1982). Development of resistance to nephrotoxic insult: Changes in urine
composition and kidney morphology on repeated exposures to mercuric chloride or
104
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
biphenyl. J Toxicol Environ Health 9: 619-635.
http://dx.doi.org/10.1080/1528739820953Q191
Koiima. A; Hiraga. K. (1978). Mutagenicity of citrus fungicides in the microbial system. Tokyo-
toritsu Eisei Kenkyusho Kenkyu Nenpo 29: 83-85.
Kokel. D; Xue. D. (2006). A class of benzenoid chemicals suppresses apoptosis in C. elegans.
Chembiochem 7: 2010-2015. http://dx.doi.org/10.1002/cbic.20060Q262
Kurata. Y; Asamoto. M; Hagiwara. A: Masui. T; Fukushima. S. (1986). Promoting effects of
various agents in rat urinary bladder carcinogenesis initiated by N-butyl-N-(4-
hydroxybutyl)nitrosamine. Cancer Lett 32: 125-135.
Kwok. ES: Buchholz. BA; Vogel. JS: Turteltaub. KW: Eastmond. DA. (1999). Dose-dependent
binding of ortho-phenylphenol to protein but not DNA in the urinary bladder of male
F344 rats. Toxicol Appl Pharmacol 159: 18-24. http://dx.doi.org/10.10Q6/taap.1999.8722
Matsubara. T; Prough. RA: Burke. MP: Estabrook. RW. (1974). The preparation of microsomal
fractions of rodent respiratory tract and their characterization. Cancer Res 34: 2196-2203.
McConnell. EE: Solleveld. HA: Swenberg. JA; Boorman. GA. (1986). Guidelines for combining
neoplasms for evaluation of rodent carcinogenesis studies. J Natl Cancer Inst 76: 283-
289.
McElroy. MC: Postle. AD: Kelly. FJ. (1992). Catalase, superoxide dismutase and glutathione
peroxidase activities of lung and liver during human development. Biochim Biophys Acta
1117: 153-158. http://dx.doi.org/10.1016/0304-4165(92)90073-4
Meyer. T. (1977). The metabolism of biphenyl. IV. Phenolic metabolites in the guinea pig and
the rabbit. Acta Pharmacol Toxicol 40: 193-200. http://dx.doi.Org/10.l 11 l/j.1600-
0773.1977.tb02068.x
Meyer. T; Aarbakke. J: Scheline. RR. (1976a). The metabolism of biphenyl. I. Metabolic
disposition of 14C-biphenyl in the rat. Acta Pharmacol Toxicol 39: 412-418.
Meyer. T; JChr. L; Hansen. EV; Scheline. RR. (1976b). The metabolism of biphenyl III Phenolic
metabolites in the pig. Basic Clin Pharmacol Toxicol 39: 433-441.
http://dx.doi.Org/10.llll/i.1600-0773.1976.tb03194.x
Meyer. T; Scheline. RR. (1976). The metabolism of biphenyl II Phenolic metabolites in the rat.
Basic Clin Pharmacol Toxicol 39: 419-432. http://dx.doi.Org/10.l 11 l/j.1600-
0773.1976.tb03193.x
Millburn. P; Smith. RL; Williams. RT. (1967). Biliary excretion of foreign compounds.
Biphenyl, stilboestrol and phenolphthalein in the rat: molecular weight, polarity and
metabolism as factors in biliary excretion. Biochem J 105: 1275-1281.
Monsanto (Monsanto Company). (1946). Final report on the physiological response of
experimental animals to the absorption of diphenyl, and several resins, elastomers and
plastics (pp. 205). St. Louis, MO.
Morimoto. K; Sato. M; Fukuoka. M; Hasegawa. R; Takahashi. T; Tsuchiya. T; Tanaka. A:
Takahashi. A: Havashi. Y. (1989). Correlation between the DNA damage in urinary
bladder epithelium and the urinary 2-phenyl-l,4-benzoquinone levels from F344 rats fed
sodium o-phenylphenate in the diet. Carcinogenesis 10: 1823-1827.
http://dx.doi.org/10.1093/carcin/10.10.1823
Nakao. T; Ushiyama. K; Kabashima. J: Nagai. F; Nakagawa. A: Ohno. T; Ichikawa. H;
Kobavashi. H; Hiraga. K. (1983). The metabolic profile of sodium o-phenylphenate after
subchronic oral administration to rats. Food Chem Toxicol 21: 325-329.
http://dx.doi.org/10.1016/0278-6915(83)90068-6
Narbonne. JF; Cassand. P; Alzieu. P; Grolier. P; Mrlina. G: Calmon. JP. (1987). Structure-
activity relationships of the N-methylcarbamate series in Salmonella typhimurium. Mutat
Res 191: 21-27. http://dx.doi.org/10.1016/0165-7992(87)90165-5
105
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
NCI (National Cancer Institute). (1968). Evaluation of carcinogenic, teratogenic and mutagenic
activities of selected pesticides and industrial chemicals. Volume I. Carcinogenic study.
(NCI-DCCP-CG-1973-1-1). Bethesda, MD: National Institutes of Health.
http://www.ntis.gov/search/product.aspx?ABBR=PB223159
Nishihara. Y. (1985). Comparative study of the effects of biphenyl and Kanechlor-400 on the
respiratory and energy linked activities of rat liver mitochondria. Occup Environ Med 42:
128-132.
NRC (National Research Council). (1983). Risk assessment in the federal government:
Managing the process. Washington, DC: National Academies Press.
http://www.nap.edu/openbook.php7record id=366&page=Rl
Ohnishi. M; . H; Takemura. T; Yamamoto. S: Matsushima. T; Ishii. T. (2000a). Characterization
of hydroxy-biphenyl-O-sulfates in urine and urine crystals induced by biphenyl and
KHC03 administration in rats. J Health Sci 46: 299-303.
Ohnishi. M; Yaiima. H; Takeuchi. T; Saito. M; Yamazaki. K; Kasai. T; Nagano. K; Yamamoto.
S: Matsushima. T; Ishii. T. (2001). Mechanism of urinary tract crystal formation
following biphenyl treatment. Toxicol Appl Pharmacol 174: 122-129.
http://dx.doi.org/10.1006/taap.2Q01.9192
Ohnishi. M; Yaiima. H; Yamamoto. S: Matsushima. T; Ishii. T. (2000b). Sex dependence of the
components and structure of urinary calculi induced by biphenyl administration in rats.
Chem Res Toxicol 13: 727-735. http://dx.doi.org/10.1021/txQ000163
Pacifici. GM; Vannucci. L; Bencini. C: Tusini. G: Mosca. F. (1991). Sulphation of
hydroxybiphenyls in human tissues. Xenobiotica 21: 1113-1118.
Pagano. G: Cipollaro. M; Corsale. G: Delia Morte. R; Esposito. A: Giordano. GG: Mi call o. G:
Ouinto. I: Staiano. N. (1988). Comparative toxicity of diphenyl, diphenyl ester, and some
of their hydroxy derivatives. Medecine Biologie Environnement 16: 291-297.
Pagano. G: Esposito. A: Giordano. GG: Vamvakinos. E; Ouinto. I: Bronzetti. G: Bauer. C: Corsi.
C: Nieri. R; Ciaiolo. A. (1983). Genotoxicity and teratogenicity of diphenyl and diphenyl
ether: A study of sea urchins, yeast, and Salmonella typhimurium. Teratog Carcinog
Mutagen 3: 377-393. http://dx.doi.org/10.1002/1520-6866(1990B:4<377::AID-
TCM1770030407>3.0.CQ:2-6
Parkinson. A: Ogilvie. BW. (2008). Biotransformation of xenobiotics. In CD Klaasen (Ed.),
Casarett & Doulls toxicology: The basic science of poisons (7th ed., pp. 161-295). New
York, NY: McGraw-Hill Companies, Inc.
Paterson. P; Fry. JR. (1985). Influence of cytochrome P-450 type on the pattern of conjugation of
4-hydroxybiphenyl generated from biphenyl or 4-methoxybiphenyl. Xenobiotica 15: 493-
502. http://dx.doi.org/10.3109/00498258509045Q23
Pathak. DN: Roy. D. (1993). In vivo genotoxicity of sodium ortho-phenylphenol:
Phenylbenzoquinone is one of the DNA-binding metabolite(s) of sodium ortho-
phenylphenol. Mutat Res-Fundam Mol Mech Mutagen 286: 309-319.
http://dx.doi.org/10.1016/0027-5107(93^)90196^
Pecchiai. L; Saffiotti. U. (1957). [Study of the toxicity of biphenyl, oxydiphenyl and their
mixture (Dowtherm)]. Med Lav 48: 247-254.
Powis. G: Jardine. I: Van Dyke. R; Weinshilboum. R; Moore. D; Wilke. T; Rhodes. W: Nelson.
R; Benson. L; Szumlanski. C. (1988). Foreign compound metabolism studies with human
liver obtained as surgical waste. Relation to donor characteristics and effects of tissue
storage. DrugMetab Dispos 16: 582-589.
Powis. G: Melder. DC: Wilke. TJ. (1989). Human and dog, but not rat, isolated hepatocytes have
decreased foreign compound-metabolizing activity compared to liver slices. Drug Metab
Dispos 17: 526-531.
106
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Probst GS; McMahon. RE; Hill LE; Thompson. CZ; Epp. JK; Neal. SB. (1981). Chemically-
induced unscheduled DNA synthesis in primary rat hepatocyte cultures: A comparison
with bacterial mutagenicity using 218 compounds. Environ Mutagen 3: 11-32.
http://dx.doi.org/10.1002/em.28600301Q3
Purchase. IFH; Longstaff. E; Ashbv. J: Styles. JA; Anderson. D; Lefevre. PA: Westwood. FR.
(1978). An evaluation of 6 short-term tests for detecting organic chemical carcinogens.
Br J Cancer 37: 873-903. http://dx.doi.org/10.1038/264624a0
Rao. JN; Scott. AJ. (1992). A simple method for the analysis of clustered binary data. Biometrics
48: 577-585.
Reitz. RH; Fox. TR; Quast. JF; Hermann. EA; Watanabe. PG. (1983). Molecular mechanisms
involved in the toxicity of orthophenylphenol and its sodium salt. Chem Biol Interact 43:
99-119. http://dx.doi.org/10.1016/0009-2797(83)90107-2
Renctizogullari. E; Parlak. S: Ha. HB. (2008). The effects of food protector biphenyl on sister
chromatid exchange, chromosome aberrations, and micronucleus in human lymphocytes.
Drug Chem Toxicol 31: 263-274. http://dx.doi.org/10.1080/014805407Q1873285
Sasaki. YF; Kawaguchi. S: Kamava. A: Ohshita. M; Kabasawa. K; Iwama. K; Taniguchi. K;
Tsuda. S. (2002). The comet assay with 8 mouse organs: results with 39 currently used
food additives. MutatRes 519: 103-119. http://dx.doi.org/10.lQ16/S1383-
5718(02)00128-6
Sasaki. YF: Saga. A: Akasaka. M; Yoshida. K; Nishidate. E; Su. YQ: Matsusaka. N: Tsuda. S.
(1997). In vivo genotoxicity of ortho-phenylphenol, biphenyl, and thiabendazole detected
in multiple mouse organs by the alkaline single cell gel electrophoresis assay. Mutat Res
Genet Toxicol Environ Mutagen 395: 189-198. http://dx.doi.org/10.lQ16/S1383-
5718(97)00168-X
Schultz. TW: Sinks. GD; Cronin. MT. (2002). Structure-activity relationships for gene activation
oestrogenicity: Evaluation of a diverse set of aromatic chemicals. Environ Toxicol 17:
14-23. http://dx.doi.org/10.1002/tox.10Q27
Seppalainen. AM: Hakkinen. I. (1975). Electrophysiological findings in diphenyl poisoning. J
Neurol Neurosurg Psychiatry 38: 248-252.
Shibata. MA: Tanaka. H; Yamada. M; Tamano. S: Fukushima. S. (1989a). Proliferative response
of renal pelvic epithelium in rats to oral administration of ortho-phenylphenol, sodium
ortho-phenylphenate and diphenyl. Cancer Lett 48: 19-28.
http://dx.doi. org/10.1016/0304-3835(89)90198-5
Shibata. MA: Yamada. M; Tanaka. H; Kagawa. M; Fukushima. S. (1989b). Changes in urine
composition, bladder epithelial morphology, and DNA synthesis in male F344 rats in
response to ingestion of bladder tumor promoters. Toxicol Appl Pharmacol 99: 37-49.
http://dx.doi. org/10.1016/0041 -008X(89)90109-9
Shiraiwa. K; Takita. M; Tsutsumi. M; Kinugasa. T; Denda. A: Takahashi. S: Konishi. Y. (1989).
Diphenyl induces urolithiasis does not possess the ability to promote carcinogenesis by
N-ethyl-N-hydroxyethylnitrosamine in kidneys of rats. J Toxicol Pathol 2: 41-48.
Smith. RA; Christenson. WR: Battels. MJ; Arnold. LL; St John. MK; Cano. M; Garland. EM:
Lake. SG: Wahle. BS: McNett. DA: Cohen. SM. (1998). Urinary physiologic and
chemical metabolic effects on the urothelial cytotoxicity and potential DNA adducts of o-
phenylphenol in male rats. Toxicol Appl Pharmacol 150: 402-413.
http://dx.doi.org/10.10Q6/taap.1998.8435
Snyder. RD; Matheson. DW. (1985). Nick translation—a new assay for monitoring DNA damage
and repair in cultured human fibroblasts. Environ Mutagen 7: 267-279.
http://dx.doi.org/10.1002/em.28600703Q4
107
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Sofuni. T; Havashi. M; Matsuoka. A; Sawada. M; Hatanaka. M; Jr. IM. (1985). [Mutagenicity
tests on organic chemical contaminants in city water and related compounds. II.
Chromosome aberration tests in cultured mammalian cells], Kokuritsu Iyakuhin
Shokuhin Eisei Kenkyusho Hokoku 103: 64-75.
S0ndergaard. D; Blom. L. (1979). Polycystic changes in rat kidney induced by biphenyl fed in
different diets. Arch Toxicol 2: 499-502.
Sonnier. M; Cresteil. T. (1998). Delayed ontogenesis of CYP1A2 in the human liver. Eur J
Biochem 251: 893-898.
Strassburg. CP: Strassburg. A: Kneip. S: Barut A: Tukev. RH; Rodeck. B: Manns. MP. (2002).
Developmental aspects of human hepatic drug glucuronidation in young children and
adults. Gut 50: 259-265. http://dx.doi.Org/10.1136/gut.50.2.259
Stuehmeier. G: Legrum. W: Netter. KJ. (1982). Does cobalt pretreatment of mice induce a
phenobarbitone-type cytochrome P-450. Xenobiotica 12: 273-282.
http://dx.doi.org/10.3109/00498258209Q52467
Sun (Sun Company, Inc.). (1977a). 90-day inhalation toxicity study of biphenyl (99 + % purity)
in CD1 mice. Radnor, PA.
Sun (Sun Company, Inc.). (1977b). Acute inhalation toxicity of biphenyl. Radnor, PA.
Sunouchi. M; Miyaiima. A: Ozawa. S: Ohno. Y. (1999). Effects of diphenyl on hepatic
peroxysomal enzyme and drug-metabolizing enzyme activities in BDF 1 mice. J Toxicol
Sci 24: 333.
Tamano. S: Asakawa. E; Boomyaphiphat. P; Masui. T; Fukushima. S. (1993). Lack of promotion
of N-butyl-N-(4-hydroxybutyl)nitrosamine-initiated urinary bladder carcinogenesis in
mice by rat cancer promoters. Teratog Carcinog Mutagen 13: 89-96.
http://dx.doi.org/10.1002/tcm.17701302Q5
Tan. Y; Yamada-Mabuchi. M; Arya. R; St Pierre. S: Tang. W: Tosa. M; Brachmann. C: White.
K. (2011). Coordinated expression of cell death genes regulates neuroblast apoptosis.
Development 138: 2197-2206. http://dx.doi.org/10.1242/dev.058826
Tani. S: Yonezawa. Y; Morisawa. S: Nishioka. H. (2007). Development of a new E. coli strain to
detect oxidative mutation and its application to the fungicide o-phenylphenol and its
metabolites. Mutat Res Genet Toxicol Environ Mutagen 628: 123-128.
http://dx.doi.Org/10.1016/i.mrgentox.2006.12.006
U.S. EPA (U.S. Environmental Protection Agency). (1986a). Guidelines for mutagenicity risk
assessment [EPA Report], (EPA/630/R-98/003). Washington, DC.
http://www.epa.gov/iris/backgrd.html
U.S. EPA (U.S. Environmental Protection Agency). (1986b). Guidelines for the health risk
assessment of chemical mixtures [EPA Report], (EPA/630/R-98/002). Washington, DC.
http://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=22567
U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and
documentation of biological values for use in risk assessment [EPA Report], (EPA/600/6-
87/008). Cincinnati, OH. http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=34855
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental
toxicity risk assessment [EPA Report], (EPA/600/FR-91/001). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/iris/backgrd.html
U.S. EPA (U.S. Environmental Protection Agency). (1992). A cross-species scaling factor for
carcinogen risk assessment based on equivalence of mg/kg3/4/day [EPA Report],
Washington, DC.
108
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
U.S. EPA (U.S. Environmental Protection Agency). (1994a). Interim policy for particle size and
limit concentration issues in inhalation toxicity studies [EPA Report], Washington, DC.
http://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=l 86068
U.S. EPA (U.S. Environmental Protection Agency). (1994b). Methods for derivation of
inhalation reference concentrations and application of inhalation dosimetry [EPA
Report], (EPA/600/8-90/066F). Research Triangle Park, NC.
http://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=71993
U.S. EPA (U.S. Environmental Protection Agency). (1995). The use of the benchmark dose
approach in health risk assessment [EPA Report], (EPA/630/R-94/007). Washington, DC.
http://www.epa.gov/raf/publications/useof-bda-healthrisk.htm
U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity
risk assessment [EPA Report], (EPA/630/R-96/009). Washington, DC.
http://www.epa.gov/raf/publications/pdfs/REPR051 .PDF
U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment [EPA Report], (EPA/630/R-95/001F). Washington, DC.
http://www.epa.gov/raf/publications/pdfs/NEUROTQX.PDF
U.S. EPA (U.S. Environmental Protection Agency). (2000). Science policy council handbook:
Risk characterization [EPA Report], (EPA 100-B-00-002). Washington, D.C.
http://www.epa.gov/osa/spc/pdfs/rchandbk.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes [EPA Report], (EPA/630/P-02/002F). Washington,
DC. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=51717
U.S. EPA (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen risk
assessment [EPA Report], (EPA/630/P-03/001F). Washington, DC.
http://www.epa.gov/cancerguidelines/
U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for
assessing susceptibility from early-life exposure to carcinogens [EPA Report] (pp. 1125-
1133). (EPA/630/R-03/003F). Washington, DC.
http://www.epa.gov/cancerguidelines/guidelines-carcinogen-supplement.htm
U.S. EPA (U.S. Environmental Protection Agency). (2006a). A framework for assessing health
risk of environmental exposures to children [EPA Report], (EPA/600/R-05/093F).
Washington, DC. http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=l58363
U.S. EPA (U.S. Environmental Protection Agency). (2006b). Science policy council handbook:
Peer review. (EPA/100/B-06/002). Washington, DC.
http://www.epa.gov/OSA/spc/2peerrev.htm
U.S. EPA (U.S. Environmental Protection Agency). (2011). Recommended use of body weight
3/4 as the default method in derivation of the oral reference dose [EPA Report],
(EPA/100/R11/0001). Washington, DC.
http://www.epa.gov/raf/publications/interspecies-extrapolation.htm
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance.
(EPA/100/R-12/001). Washington, DC.
http://www.epa.gov/raf/publications/pdfs/benchmark dose guidance.pdf
Umeda. Y; Aiso. S: Arito. H: Nagano. K; Matsushima. T. (2004a). Short communication:
Induction of peroxisome proliferation in the liver of biphenyl-fed female mice. J Occup
Health 46: 486-488.
Umeda. Y; Aiso. S: Yamazaki. K; Ohnishi. M; Arito. H; Nagano. K; Yamamoto. S: Matsushima.
T. (2005). Carcinogenicity of biphenyl in mice by two years feeding. J Vet Med Sci 67:
417-424.
109
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Umeda. Y; Arito. H; Kano. H; Ohnishi. M; Matsumoto. M; Nagano. K; Yamamoto. S;
Matsushima. T. (2002). Two-year study of carcinogenicity and chronic toxicity of
biphenyl in rats. J Occup Health 44: 176-183.
Umeda. Y; Matsumoto. M; Yamazaki. K; Ohnishi. M; Arito. H; Nagano. K; Yamamoto. S;
Matsushima. T. (2004b). Carcinogenicity and chronic toxicity in mice and rats
administered vinyl acetate monomer in drinking water. J Occup Health 46: 87-99.
Union Carbide (Union Carbide Corporation). (1949). Range finding tests on diphenyl tables of
protocols attached with cover letter. (878213680). Danbury, CT: Union Carbide Corp.
http://www.ntis.gov/search/product.aspx?ABBR=QTS0206426
Wangenheim. J: Bolcsfoldi. G. (1986). Mouse lymphoma tk+/- assay of 30 compounds
[Abstract], Environ Mutagen 8: 90.
Wangenheim. J: Bolcsfoldi. G. (1988). Mouse lymphoma L5178Y thymidine kinase locus assay
of 50 compounds. Mutagenesis 3: 193-205. http://dx.doi.Org/10.1093/mutage/3.3.193
Wastensson. G: Hagberg. S: Andersson. E; Johnels. B; Barregard. L. (2006). Parkinson's disease
in diphenyl-exposed workers— A causal association? Parkinsonism Relat Disord 12: 29-
34. http://dx.doi.Org/10.1016/i.parkreldis.2005.06.010
Westinghouse Electric Corporation. (1977). Potential carcinogenicity testing of PCB
replacements using the Ames test with cover letter. (OTS0206616). Pittsburgh, PA.
http://www.ntis.gov/search/product.aspx?ABBR=QTS0206616
Wiebkin. P; Fry. JR; Jones. CA; Lowing. R; Bridges. JW. (1976). The metabolism of biphenyl
by isolated viable rat hepatocytes. Xenobiotica 6: 725-743.
http://dx.doi.org/10.3109/0049825760915139Q
Wiebkin. P; Fry. JR: Jones. CA: Lowing. RK; Bridges. JW. (1978). Biphenyl metabolism in
isolated rat hepatocytes: effect of induction and nature of the conjugates. Biochem
Pharmacol 27: 1899-1907. http://dx.doi.org/10.1016/0006-2952(78^)90003-5
Wiebkin. P; Schaeffer. BK; Longnecker. DS: Curphev. TJ. (1984). Oxidative and conjugative
metabolism of xenobiotics by isolated rat and hamster acinar cells. Drug Metab Dispos
12: 427-431.
Williams. GM; Mori. H; McQueen. CA. (1989). Structure-activity relationships in the rat
hepatocyte DNA-repair test for 300 chemicals [Review], Mutat Res 221: 263-286.
http ://dx. doi. org/10.1016/0165 -1110(89^)9003 9-0
Williams. PL: Ryan. LM. (1997). Dose-response models for developmental toxicology.
(DART/TER/95003988). Williams, PL; Ryan, LM.
110
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
The Toxicological Review of Biphenyl, dated September 2011, has undergone a formal
external peer review performed by scientists in accordance with EPA guidance on peer review
(U.S. EPA, 2006a, 2000a). An external peer-review workshop was held on April 3, 2012. The
external peer reviewers were tasked with providing written answers to general questions on the
overall assessment and on chemical-specific questions in areas of scientific controversy or
uncertainty. A summary of significant comments made by the external reviewers and EPA's
responses to these comments follow. In many cases, the comments of the individual reviewers
have been synthesized and paraphrased in development of Appendix A. EPA also received
scientific comments from the public. These comments and EPA's responses are included in a
separate section of this appendix.
I. External Peer Review Comments
The reviewers made several editorial suggestions to clarify specific portions of the text.
These changes were incorporated in the document as appropriate and are not discussed further.
General Comments
1. Is the Toxicological Review logical, clear and concise? Has EPA clearly presented and
synthesized the scientific evidence for noncancer and cancer health effects of biphenyl?
Comments: To varying degrees, all of the reviewers commented that the draft was well written,
logical, clear, and generally well done. Four reviewers commented that the document was not
concise or that there was some redundancy in the information presented; two reviewers, on the
other hand, specially stated that the document was concise. Several reviewers suggested that
clear and concise conclusions at the end of each section (in particular, the Toxicokinetics
section) or introductory paragraphs at the beginning of major sections would be helpful. One
reviewer identified several statistical issues (e.g., failure to identify a finding as statistically
significantly different from the control in summary tables and questions about the application of
certain statistical tests).
Response: The Toxicological Review was revised throughout to reduce redundancy, and
information of lesser relevance throughout the document was removed to the extent practicable.
Summaries of biphenyl toxicokinetics and human health effects information were added to the
beginning of Sections 3 and 4.1. A summary of animal studies was already included in Section
A-l
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4.2. Section 4.6 was revised to provide a more comprehensive review and synthesis of biphenyl
health effects information. Statistical errors and omissions were corrected.
Comments: One reviewer recommended further discussion of the evaluation of older studies of
cancer and noncancer endpoints, including more details on the strengths and weaknesses of these
studies, and more explanation as to how each study contributed to the final decision making.
Response: Section 4.6.1, Synthesis of Major Noncancer Effects, concerning the noncancer
effects of biphenyl and Sections 4.7.1, Summary of Overall Weight of Evidence, and 4.7.2,
Synthesis of Human, Animal, and Other Supporting Evidence, concerning the carcinogenicity of
biphenyl were revised, as appropriate, to more explicitly take into consideration study quality in
identifying the hazards associated with biphenyl exposure. Section 5.1.1, Choice of Candidate
Principal Studies and Candidate Critical Effects - With Rationale and Justification, was revised
to include a more explicit evaluation of the strengths and weakness of major studies and the
rationale for choosing studies for dose-response analysis.
Comments: One reviewer recommended that a description of the literature search strategy for
locating relevant literature be included.
Response: Documentation of the literature search strategy, including a graphical depiction of the
literature search strategy and search outcomes, was added as Appendix B of the Toxicological
Review; reference to this appendix was added to Section 1. The search strategy documentation
also provides a link to EPA's Health and Environmental Research Online (HERO) database
(www.epa.gov/hero) that contains a web page showing the references that were cited in the
Toxicological Review as well as those references identified in the literature search that were
screened (considered) but not cited.
Comments: One reviewer observed that Section 4, Hazard Identification, was well written, clear,
and concise, but offered suggestions for presentation or clarification beyond those provided in
response to specific charge questions.
• The reviewer noted that the incidence of reticular cell sarcoma in biphenyl-treated female
strain B mice (summarized in Table 4-9) was significantly greater than in controls by
Fisher Exact Test (p < 0.01), and should be noted in Table 4-9 and briefly discussed in
accompanying text and Section 4.7, Evaluation of Carcinogenicity.
• The nonrodent oral studies reported in Section 4.2.1.2.3 are shorter than one-tenth the
lifespan of the animal species and should not be included in the "Chronic toxicity and
carcinogenicity studies" section (i.e., one-year dog and one-year rhesus monkey studies).
The reviewer recommended that these studies be moved to a separate section or included
A-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
in the subchronic study section.
• The reviewer stated that the overall weight of evidence for genotoxicity appears more
equivocal than negative given the clastogenicity in human lymphocytes, the in vivo
findings, and the limited evidence for genotoxicity of metabolites.
• Regarding statements in the MOA section related to lack of concordance for
neurotoxicity between humans and animals, the reviewer observed that the animal studies
were not designed to detect the neurotoxicity seen in human studies.
Response: EPA noted the statistical significance of the increased incidence of reticular cell
sarcoma in strain B female mice (NCI, 1968) in Section 4.2.1.2.2 and Table 4-9. Discussion of
the biological significance of this tumor finding was added to Sections 4.2.1.2.2 and 4.7.1. EPA
agrees that the one-year dog and monkey studies should not be considered chronic duration
studies. Summaries of these nonrodent oral studies in Section 4.2.1.2.3 were moved to Section
4.2.1.1, Subchronic Toxicity. Section 4.5.2, Genotoxicity, and Appendix C were revised to more
precisely characterize the available evidence for the genotoxicity of biphenyl and its metabolites.
The comment related to evidence for neurotoxicity associated with biphenyl exposure is
addressed in responses under Charge Question A.4.
2. Please identify any additional peer-reviewed studies from the primary literature that
should be considered in the assessment of the noncancer and cancer health effects of
biphenyl.
Comments: Seven of the eight reviewers did not identify any additional studies. One reviewer
recommended consideration of an issue of the journal Birth Defects Research that was devoted to
interpreting skeletal malformations and variations (Birth Defects Research, Part B, volume 80
(6), 2007). This reviewer stated that articles in this volume address some of the malformations
found in the Khera et al. (1979) study and may directly impact the consideration of using skeletal
malformations as the endpoint for calculation of the RfD.
Response: EPA agrees that the recommended journal issue is pertinent for this assessment.
Discussion of a particular paper from this issue (Carney and Kimmel, 2007) was added to
Section 4.6.1 to support interpretation of fetal skeletal variations as reported by Khera et al.
(1979).
A. Oral Reference Dose (RfD) for Biphenyl
The first two charge questions in this portion of the review address the selection of the
critical effect and the principal study for developing an RfD. For this database, the critical
endpoint used in the draft assessment and another recommended by reviewers were specific to
A-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
different studies (i.e., skeletal anomalies as reported in a developmental toxicity study by Khera
et al. [1979] and renal endpoints as reported in a chronic bioassay in the rat by Umeda et al.
[2002]). As such, preference for one endpoint also determines the choice of study. For this
reason, the comments and responses to the following two related charge questions were merged.
1. A developmental toxicity study of biphenyl in Wistar rats (Khera et al., 1979) was
selected as the basis for the derivation of the RfD. Please comment on whether the
selection of this study is scientifically supported and clearly described. If a different study
is recommended as the basis for the RfD, please identify this study and provide scientific
support for this choice.
2. A developmental effect in Wistar rats (i.e., fetal skeletal anomalies) was concluded by
EPA to be an adverse effect and was selected as the critical effect for the derivation of the
RfD. Please comment on whether the selection of this critical effect and its
characterization is scientifically supported and clearly described. If a different endpoint is
recommended as the critical effect for deriving the RfD, please identify this effect and
provide scientific support for this choice.
Comments: Several peer reviewers raised concerns about the selection of fetal skeletal anomalies
in Khera et al. (1979) as the critical effect, and proposed as an alternative critical effect renal
lesions as reported in the 2-year rat bioassay of biphenyl by Umeda et al. (2002). More
specifically, three reviewers commented that justification for the selection of fetal skeletal
anomalies as the critical effect needed to be expanded, noting that consideration should be given
to maternal toxicity and whether delayed ossification and extra ribs are adverse effects. One of
these reviewers commented that it is difficult to determine the appropriateness of selecting Khera
et al. (1979) as the principal study without more details on the fetal anomalies—details that were
not provided in the published study. Two reviewers did not support selection of fetal skeletal
anomalies as the critical effect. One of these two reviewers did not consider the skeletal
anomalies to be adverse findings in the absence of other malformations, and concluded that the
anomalies could be attributed to maternal toxicity. Two reviewers expressed concern about the
quality of the developmental study conducted approximately 35 years ago. On the other hand,
three reviewers considered the selection of fetal skeletal anomalies as reported by Khera et al.
(1979) either to be appropriate, consistent with EPA guidelines, or clearly described.
Five reviewers, including one who considered the selection of fetal skeletal anomalies a
reasonable choice and consistent with EPA guidelines, identified renal lesions as reported in the
2-year rat bioassay by Umeda et al. (2002) as an alternative or more scientifically defensible
critical effect. One reviewer specifically recommended hemosiderin deposition in the kidney
(Umeda et al., 2002) as an alternative critical effect, whereas another reviewer considered
A-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
hemosiderin to be a nonspecific effect that "usually is meaningless to humans." The latter
reviewer recommended simple hyperplasia of the kidney, renal pelvis mineralization, or
papillary mineralization as more scientifically defensible as the critical effect.
Response: EPA agrees that the Khera et al. (1979) study may have differed from more current
study designs, but is unaware of any particular study quality issues due to the age of the study
that would decrease confidence in its conduct or reported results. The study design used a
typical number of rats (18-20 dams/dose group), used four dose groups after consideration of the
results of a range-finding study, and evaluated skeletal and visceral anomalies using standard
methods. As it is the only developmental toxicity study of biphenyl available, there is no
corroboration of the findings. Without any indication that the study was designed or conducted
in an inappropriate manner, however, these findings have a place in the hazard evaluation of
biphenyl.
EPA agrees that the uncertainties in the interpretation of fetal skeletal anomalies,
including maternal toxicity and adversity of the anomalies, were not adequately weighed in
selecting this endpoint as the critical effect for the RfD. Discussion of the Khera et al. (1979)
study was revised to more clearly present the following points that influenced interpretation of
the study findings. Maternal toxicity was observed in the highest dose group (1,000 mg/kg-day),
but not at 500 mg/kg-day or lower doses. Skeletal anomalies were found at or below 500 mg/kg-
day, and thus cannot be attributed to maternal toxicity. Among the anomalies listed, missing or
unossified sternebrae was the only endpoint elevated with increasing dose at doses lower than
1,000 mg/kg-day. Consistent with reviewers' advice and the more recent publications they
recommended (e.g., Carney and Kimmel, 2007), anomalies with biological significance were
limited to missing or unossified sternebrae.
The Khera et al. (1979) study was retained as a candidate principal study. In light of the
issues raised by the reviewers, however, EPA clarified the interpretation of the anomalies in this
study in Sections 4.3.1 and 4.6.1. In addition, EPA listed the anomalies observed in "anomalous
litters"—wavy ribs, extra ribs, missing or unossified sternebrae, or delayed ossification of the
calvarium—and included the respective incidences of fetuses in each dose group (see Section
4.3.1). The incidence of missing or unossified sternebrae and the number of litters examined
were repeated in the dose-response section.
Consistent with peer reviewer recommendations, the robust toxicity studies in rats and
mice by Umeda et al. (2005, 2002) were also considered as candidate principal studies, with the
rationale clarified in Section 5.1.1. Also consistent with peer reviewer recommendations, renal
lesions, and in particular renal papillary mineralization in male rats, was selected as the critical
effect. Sections 4.6.1 and 5.1.2 were revised to better characterize the evidence for renal lesions
as a hazard of biphenyl exposure and to provide the rationale for selection of renal papillary
mineralization as the critical effect.
A-5
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
3. Benchmark dose (BMD) modeling was conducted using the incidence of litters with fetal
skeletal anomalies to estimate the point of departure (POD) for derivation of the RfD. Has
the modeling been appropriately conducted and clearly described based on EPA's draft
Benchmark Dose Technical Guidance Document (U.S. EPA, 2000)? Is the choice of the
benchmark response (BMR) for use in deriving the POD (i.e., a BMR of 10% extra risk of
the incidence of litters with any fetal skeletal anomalies) supported and clearly described?
Comments: Four reviewers commented that the modeling was appropriately conducted, clearly
described, or followed EPA guidance. One of these reviewers also commented that EPA's
argument for not applying cross-species scaling to the oral dose for the developmental endpoint
was problematic. Another reviewer emphasized the maternal toxicity at the high dose in the
developmental study, and asked that 1) the assessment be clearer whether or not these data were
included in the modeling and 2) that if included that this be justified. Two reviewers reiterated
that the developmental study was not appropriate for RfD derivation, and the remaining reviewer
noted his lack of familiarity with dose-response modeling.
Regarding BMR selection, two reviewers stated that the reason for using a BMR of 10%
extra risk for incidence of litters with effects versus 5% among fetuses with effects was
adequately explained, while two others commented that this selection should be explained
further. The remaining four reviewers did not comment.
Response: As summarized under the first two charge questions, EPA agrees that the renal effects
reported by Umeda et al. (2002) are more compelling for RfD derivation than the developmental
effects reported by Khera et al. (1979). However, a candidate RfD for developmental toxicity
was retained in the revised assessment in order to provide some perspective on the
developmental hazard of biphenyl exposure. Following the reviewers' evaluation of the Khera et
al. (1979) study, the dose-response analysis focused on missing or unossified sternebrae, the only
anomaly that showed an increasing trend with dose in the absence of maternal toxicity. The
high-dose group was omitted from dose-response modeling because of the demonstrated
maternal toxicity. A modeling approach that approximates the result of nested models (due to
the unavailability of detailed data showing the distribution of fetuses among litters) was
implemented that enabled using a BMR of 5% extra risk among fetuses, precluding the need to
consider an equivalent degree of effect in terms of litter incidence. Briefly, BMD analyses used
the proportions of affected fetuses within each dose group, and alternately used the total number
of fetuses and the total number of litters as the group sizes to bracket the BMD and BMDL
expected to result from a nested analysis of individual data, if they were available (see, e.g., Rao
and Scott, 1992). Section 5.1.2 was revised to reflect this change.
EPA agrees that a body weight scaling to the % power (i.e., BW3 4) approach should be
A-6
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
applied to extrapolate equivalent doses from dams to humans for the purpose of calculating a
human equivalent dose, consistent with EPA guidance (U.S. EPA, 2011). Also consistent with
this guidance, BW3 4 scaling was used to extrapolate to human-equivalent doses for the renal
endpoints. Detailed calculations can be found in Section 5.1.2.
4. Please comment on the rationale for the selection of the uncertainty factors (UFs) applied
to the POD for the derivation of the RfD. Are the UFs appropriate based on the
recommendations described in A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002; Section 4.4.5) and clearly described? If changes to the selected
UFs are proposed, please identify and provide scientific support for the proposed changes.
Comments: Five of the eight reviewers generally agreed with the selection of UFs applied to the
POD for the derivation of the RfD; one of these reviewers further observed that the UFs were
consistent with EPA guidance (U.S. EPA, 2002). Two reviewers did not offer comments
because the topic was outside their area of expertise.
The remaining reviewer agreed with the UFs applied for interspecies and intraspecies
adjustments, but recommended further discussion of the UFs for LOAEL to NOAEL
extrapolation and for database deficiencies. Specifically, more discussion was recommended to
support the justification for a LOAEL to NOAEL UF of 1 based on skeletal anomalies and the
assumption that an effect at the BMDL represented a minimally biologically significant change.
In addition, this reviewer suggested that the database UF of 1 could be raised to 3 or 10 because
some animal studies were limited by small numbers of animals, incomplete histopathology, and
insufficient study length and because the database lacked animal studies examining neurological
effects (which were observed in workers) and developmental neurological effects.
Response: For the LOAEL to NOAEL UF assigned to the POD for developmental effects, EPA
considered (in the draft assessment) an increase of 10% (extra risk) in incidence of litters with
skeletal anomalies to be a change with minimal biological significance, because of its expected
equivalence to a 5% extra risk in incidence of fetuses with skeletal anomalies. The revised
analyses use a 5% extra risk BMR for incidence of missing or unossified sternebrae among
fetuses, and a 10% extra risk BMR for the renal effects, both of which are judged to characterize
minimally biologically significant changes.
The database UF of 1 for the oral RfD is supported, in part, by two chronic oral toxicity
studies in rats and mice by Umeda et al. (2005, 2002) that were conducted according to OECD
testing guidelines and conformed to OECD GLP principles. As noted by one reviewer, some
animal studies were limited by small numbers of animals, incomplete histopathology, or
insufficient study length. Nevertheless, these studies generally support the findings of the more
robust Umeda et al. studies and as such do not represent database deficiencies. Potential
A-7
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
neurological effects of biphenyl were examined in two epidemiological studies of workers in two
factories manufacturing biphenyl-impregnated paper. Information was not available to
characterize biphenyl exposure quantitatively in either study, although workers from both
"3
factories were exposed to biphenyl at levels above the occupational limit of 1.3 mg/m (threshold
limit value [TLV] by ACGIH, 2001), and in one of the two studies, an average air concentration
almost 100 times the TLV was reported in one location in the plant. It is unclear how the
findings from these workplace studies that predominantly involved inhalation exposure would
relate to oral exposure. As noted by one reviewer, animal studies did not include examination of
sensitive measures of neurotoxicity. The 2-year oral bioassays in rats and mice (Umeda et al.
2005, 2002) did, however, include daily observations for clinical signs and histopathologic
examination of nervous system tissues. No nervous system effects were reported, suggesting
that the nervous system is not a sensitive target of oral biphenyl toxicity. In summary, the
findings from studies of occupational (predominantly inhalation) exposure to biphenyl introduce
some uncertainties in the characterization of biphenyl hazard. These uncertainties are discussed
in the justification for the database UF for the oral RfD in Section 5.1.3; however, EPA did not
consider the uncertainties sufficient to warrant a database UF more than 1 in deriving the RfD.
(B) Inhalation Reference Concentration (RfC) for Biphenyl
1. The draft "Toxicological Review of Biphenyl" did not derive an RfC. Has the
justification for not deriving an RfC been clearly described in the document? Are there
available data to support the derivation of an RfC for biphenyl? If so, please identify these
data.
Comments: All reviewers agreed that there are insufficient data to derive an inhalation RfC for
biphenyl and that the justification for not deriving an RfC was clearly and adequately described.
One reviewer specifically recommended against extrapolating from the oral value to derive an
RfC because biphenyl pharmacokinetics may be relatively complicated and no data on route
differences in pharmacokinetics are available. One reviewer disagreed with the text on page 89
stating "The lack of adequate data to derive an RfC represents a significant uncertainty for the
evaluation of risks from exposure to inhaled biphenyl," and recommended that EPA compare
ambient air biphenyl concentrations with the TLV to provide perspective on likely risks from
biphenyl inhalation.
Response: Consistent with the recommendations of the peer reviewers, an RfC for biphenyl was
not derived. With regard to the recommendation to use the TLV as a point of comparison, it
should be noted that this value applies to healthy adult workers and does not take into
consideration effects of the chemical in children and other potentially susceptible lifestages and
A-8
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
"3
populations. Established in 1972, the TLV of 0.2 ppm (1 mg/m ) was based on a subchronic
mouse study conducted in 1947 (Deichmann et al., 1947) that showed respiratory effects at
"3
1 ppm (6 mg/m ). Thus, the TLV was established at a level only fivefold lower than the air
concentration producing effects in the mouse. For the above reasons, the biphenyl TLV is not
considered to be a health-protective value for general population exposures. In light of the
comment, however, the text in Section 5.3 regarding potential risks of inhaled biphenyl was
revised.
(C) Carcinogenicity of Biphenyl
1. Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a;
www.epa.gov/iris/backgrd.htmn, the draft "Toxicological Review of Biphenyl" concludes
that the database for biphenyl provides "suggestive evidence of carcinogenic potential" by
all routes of exposure. Please comment on whether this characterization of the human
cancer potential of biphenyl is scientifically supported and clearly described.
Comments: Three reviewers agreed with the cancer descriptor of "suggestive evidence of
carcinogenic potential" for biphenyl. One of these reviewers characterized the liver tumor
findings in female BDFi mice (Umeda et al., 2005) as robust and as a sufficient basis in and of
itself to support the suggestive descriptor. This reviewer also suggested that studies of durations
not sufficiently long to be informative for carcinogenicity determination (including Dow
Chemical Co., 1953; Monsanto, 1946) be excluded from this discussion and that deficiencies and
limitations of other studies (including Pecchiai and Saffiotti, 1957; Ambrose 1960; Shiraiwa et
al., 1989) be further discussed.
One reviewer commented that the rationale for the cancer characterization should be
more clearly described in Section 4.7, including identifying study limitations of Imai et al.
(1983), strengthening the argument that humans are less susceptible to urinary bladder tumors,
and making more explicit whether or not urinary bladder tumors were excluded in selecting the
descriptor such that the positive tumor findings for biphenyl carcinogenicity apply to only one
species, sex, strain, and site, thereby obviating the "likely to be carcinogenic" category.
One reviewer did not agree with the descriptor and recommended instead the term "some
evidence" of carcinogenicity consistent with the terminology from the National Toxicology
Program (NTP).
Three reviewers did not indicate whether or not they agreed with the selection of the
suggestive descriptor. One of these reviewers observed that there was not enough synthesis of
the data or attention paid to confounding (e.g., palatability, weight loss).
Response: EPA retained the cancer descriptor of "suggestive evidence of carcinogenic potential"
A-9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
for biphenyl and expanded the consideration of factors influencing the weight of evidence for
carcinogenicity in Sections 4.7.1 and 4.7.2. EPA agreed that the absence of a tumor response in
one-year dog (Monsanto, 1946) and monkey (Dow Chemical Co, 1953) studies should not be
considered in evaluating the cancer weight of evidence because the study durations were not
sufficiently long and the group sizes (1-2 animals/sex/group) were too small to allow for
detection of tumors. These two studies were excluded from the discussion. More thorough
characterization of other studies that found no evidence of carcinogenic response (including
study limitations) was added to Sections 4.7.1 and 4.7.2.
EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) recommend that
the narrative that characterizes the cancer weight of evidence for a chemical also assign one of
the five weight-of-evidence descriptors identified in the guidelines to provide some measure of
clarity and consistency across assessments. "Some evidence of carcinogenicity," used by NTP,
is not among the descriptors in EPA's Cancer Guidelines. Therefore, the cancer descriptor of
"suggestive evidence of carcinogenic potential"—one of five descriptors provided in EPA's
cancer guidelines—was retained.
No treatment-related changes in food consumption or palatability that could have
potentially confounded the results of the most informative studies of biphenyl carcinogenicity,
i.e., Umeda et al. (2005, 2002), were identified. For all studies, food consumption and body
weight information, where available, were used in calculating doses in mg/kg body weight-day.
Comments: One reviewer considered EPA's treatment of bladder tumor findings as not
contributing to the positive evidence at "environmentally relevant dose" to be well described, but
also proposed that an alternative approach would be to address the issue of high-dose
carcinogenicity via calculi formation leading to higher overall evidence for carcinogenicity in
this dose region. Two reviewers did not agree with adding the language "at environmentally
relevant exposure levels in humans" in the context of bladder tumors. One reviewer
recommended more discussion of what constitutes "environmentally relevant exposure" in the
context of bladder tumors.
Response: EPA agrees with the peer reviewers that the phrase "environmentally relevant
exposures" is not particularly clear or helpful language. To be more clear and specific, EPA
revised the text in Section 4.7.1 such that the descriptor was not tied to ranges of exposure.
Comment: One reviewer considered the discussion of evidence by route of exposure to have
been well laid out, but suggested noting that Sun (1977a) provides evidence of distal impacts in
the liver and kidney from inhalation exposure. Three reviewers questioned the application of the
suggestive descriptor to all routes of exposure, noting that the data supporting
toxicity/carcinogenicity by routes of exposure other than oral was scanty or absent.
A-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Response: Evidence of distal effects of inhaled biphenyl was reported in studies by Deichmann
et al. (1947) and Sun (1977a); the discussion of carcinogenic potential by other routes of
exposure in Section 4.7.1 was expanded to include these studies as indirect support for
absorption of inhaled biphenyl. Evidence of dermal absorption of biphenyl is provided in an
unpublished in vitro study (Fasano, 2005) submitted during the public comment period. Sections
3.1 and 4.7.1 were revised to include reference to this study. According to the Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005a), "[W]hen tumors occur at a site other than the
point of initial contact, the descriptor generally applies to all exposure routes that have not been
adequately tested at sufficient doses. An exception occurs when there is convincing information,
e.g., toxicokinetic data that absorption does not occur by another route." Given the evidence,
albeit limited, for absorption of biphenyl by inhalation and dermal routes, application of the
descriptor to all routes of exposure was retained consistent with Agency guidance.
Comment: One reviewer commented that justification for the position that certain minor
metabolites of biphenyl do not contribute to tumorigenesis was not adequate.
Response: EPA agrees that evidence for the genotoxicity of the metabolites of biphenyl was not
adequately characterized. The discussions of the evidence for genotoxic vs. mutagenic activity
in Section 4.5.2 and Appendix C were revised, and the evidence for a mutagenic mode of action
based on data for biphenyl and its metabolites was clarified in Section 4.7.3.1.3.
2. EPA has concluded that biphenyl-induced urinary bladder tumors in male rats is a high-
dose phenomenon involving sustained occurrence of calculi in the urinary bladder leading
to transitional cell damage, sustained regenerative cell proliferation, and eventual
promotion of spontaneously initiated tumor cells in the urinary bladder epithelium. Please
comment on whether this determination is scientifically supported and clearly described.
Please comment on data available that may support an alternative mode of action for
biphenyl-induced urinary bladder tumors.
Comments: Six of the eight reviewers agreed that the proposed mode of action for biphenyl-
induced urinary bladder tumors was supported and clearly described. One of these reviewers
observed that a small contribution from genotoxic biphenyl metabolites to urinary bladder
carcinogenesis cannot be ruled out, but that did not preclude a conclusion that the observed
bladder tumors would not have occurred without calculi formation.
Two reviewers did not consider the mode of action to be sufficiently supported. One of
these reviewers commented that data did not prove that bladder stones were required for
carcinogenesis and biphenyl may cause both stones and cancer, not necessarily in any specific
A-ll
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
order. However, this reviewer stated that he was not aware of another proven mode of bladder
carcinogenesis for biphenyl. The second reviewer did not consider the explanations for gender-
and species-specific association between bladder calculi formation and development of bladder
tumors to be clear, questioned whether there had been exploration of alternative mechanisms of
action, and suggested consideration be given to an alternative mode of action based on data for
2-aminobiphenyl for which there is evidence of up-regulation of the expression of COX-2 via
NADPH oxidase-derived ROS-dependent pathways in a bladder cancer cell line.
Response: Regarding the contribution of biphenyl metabolites to a mutagenic mode of action, see
the response under Charge Question C. 1.
EPA retained the hypothesized mode of action for biphenyl-induced urinary bladder
tumors because the available data demonstrated a strong, consistent, and specific association
between calculi formation and urinary bladder tumor occurrence. As discussed in Section
4.7.3.1 and consistent with the cancer mode of action framework provided in EPA's Guidelines
for Carcinogen Risk Assessment (U.S. EPA, 2005a), the key events in the hypothesized mode of
action (i.e., calculi formation followed by irritation of transitional epithelial cells of the urinary
bladder, sustained cell proliferation, and promotion of initiated cells in the urinary bladder with
progression to papillomas and carcinomas) show dose-response concordance, a temporal
relationship, and biological plausibility.
The available information on gender- and species-specific differences in calculi
formation and development of bladder tumors presented in Section 4.7.3.1.2 was revised to
clarify the gender differences in calculi formation and tumor response. As discussed in this
section, the differences in calculi formation (i.e., lack of calculi in mice and the differences in
chemical composition and physical properties of calculi between male and female rats) is
consistent with the lack of urinary bladder tumor response in mice and female rats. An
alternative mode of action for biphenyl based on a mechanistic study of 2-aminobiphenyl was
not included in the Toxicological Review because 2-aminobiphenyl is not a metabolite or
precursor of biphenyl and the relevance of the findings of this study to biphenyl is not clear.
Comments: One reviewer considered the term "transitional cell carcinoma" to be outdated and
recommended using instead the current terminology—"urothelial carcinoma" (Epstein et al.,
1998; I ARC, 2004).
Response: The term "urothelial" specifically refers to a carcinoma of the urothelium, meaning a
transitional cell carcinoma of the urinary system. Currently, "transitional cell carcinoma" and
"urothelial carcinoma" are used interchangeably. To be consistent with the term used by Umeda
et al. (2002), the term "transitional cell carcinoma" was retained.
A-12
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
3. EPA has concluded that there is insufficient information to identify the mode(s) of
carcinogenic action for biphenyl-induced liver tumors in mice. Please comment on whether
this determination is appropriate and clearly described. If it is judged that a mode of
action can be established for biphenyl-induced mouse liver tumors, please identify the
mode of action and its scientific support (i.e., studies that support the key events, and
specific data available to inform the shape of the exposure-response curve at low doses).
Comments: All reviewers agreed that there is insufficient information to identify the mode(s) of
carcinogenic action for biphenyl-induced liver tumors in mice, and that this determination was
appropriate and clearly described.
Response: No response is needed.
Oral Slope Factor (OSF)
4. A two-year cancer bioassay of biphenyl in BDFi mice (Umeda et al., 2005) was selected
as the basis for the derivation of the OSF. Please comment on whether the selection of this
study is scientifically supported and clearly described. If a different study is recommended
as the basis for the OSF, please identify this study and provide scientific support for this
choice.
Comments: Seven reviewers agreed with the selection of the Umeda et al. (2005) study as the
basis for the derivation of the OSF, generally noting that the rationale was clearly described and
scientifically supported. One of these reviewers suggested including more detailed explanation
and evaluation of the strengths and weakness of other studies in mice and other species to assess
the entire set of relevant data. The eighth reviewer recommended selecting the study with the
lowest NOAEL for the derivation of the OSF.
Response: A discussion of strengths and weakness of the available chronic bioassays for
biphenyl was provided in Section 4.7.2; text was added to Section 5.4.1 directing the reader to
that discussion.
An OSF describes the cancer risk per unit dose of the chemical at low doses. Unless a
mode of action consistent with nonlinear extrapolation is established, the assumption is made
that the relationship between risk of cancer and exposure is linear, i.e., there is some risk of
cancer at all exposures to the chemical. Under this assumption, a NOAEL for cancer cannot be
identified. Therefore, selection of the study with the lowest NOAEL was not an appropriate
consideration for deriving the OSF for biphenyl.
A-13
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
5. The incidence of liver tumors (i.e., adenomas or carcinomas) in female mice was selected
to serve as the basis for the derivation of the OSF. Please comment on whether this
selection is scientifically supported and clearly described. If a different cancer endpoint is
recommended for deriving the OSF, please identify this endpoint and provide scientific
support for this choice.
Comments: Two reviewers agreed that liver tumor incidence in female mice was the most
appropriate data set for the derivation of the OSF. Two reviewers considered the rationale for
the selection of female liver tumors to be clearly described. One reviewer commented that the
difference in liver tumor incidence between male and female mice should be discussed in as part
of the consideration of the appropriateness of calculating an OSF and the implication for the
usefulness of the OSF. One reviewer commented that consideration of another study was
warranted given the finding of liver tumors in female mice only and low incidence in the
concurrent control group, but could not identify a more appropriate study.
One reviewer suggested that consideration be given to using urologic toxicity data given
that liver tumors form more easily in mice, liver tumors occurred almost exclusively in female
mice, urinary toxicity has been consistently observed in all studies at high levels, and bladder
tumors were the common cause of animal death. This reviewer also acknowledged that liver
toxicity was the predominant toxic effect in human studies.
Response: The reason for the difference in susceptibility of male and female BDF i mice to
induction of liver tumors by biphenyl is unknown. In the absence of an understanding of the
mode of action of biphenyl hepatocarcinogenicity, it is also unknown whether the human
response to biphenyl might be more similar to the male or female mouse. The health protective
assumption is made that the tumor response from the most sensitive gender is relevant to
humans, and therefore liver tumor incidence in the female mouse served as the basis for the OSF
for biphenyl. The assumption was included in the summary of uncertainties in the biphenyl OSF
in Section 5.4.5.1.
The occurrence of urinary bladder tumors in male rats chronically exposed to biphenyl in
the diet is considered a high-dose phenomenon associated with calculi formation. No increased
risk of bladder tumors is expected as long as exposure to biphenyl is below the dose needed to
form calculi. Because the occurrence of urinary bladder tumors is considered to be nonlinear at
low doses, derivation of an OSF based on urologic toxicity data (in this case bladder tumor
incidence data) is not supported.
Comments: One reviewer did not consider the rationale for combining adenoma and carcinoma
data for the calculation of the OSF to be well described, and suggested that adenoma data alone
would be more appropriate since the carcinoma incidence at the high dose was not statistically
A-14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
different from control.
Response: Data are not available to indicate whether malignant tumors developed specifically
from the progression of benign tumors in biphenyl-exposed female mice; however, etiologically
similar tumor types (i.e., benign and malignant tumors of the same cell type) were combined for
dose-response analyses because of the possibility that the benign tumors could progress to the
malignant form (McConnell et al., 1986). This is consistent with the Guidelines for Carcinogen
Risk Assessment (U.S. EPA, 2005a), which state that "[t]he incidence of benign and malignant
lesions of the same cell type, usually within a single tissue or organ, are considered separately
but may be combined when scientifically defensible." The rationale for combining liver
adenoma and carcinoma incidence for OSF derivation was added to Section 5.4.2.
6. Benchmark dose (BMD) modeling was conducted using the incidence of liver tumors in
female mice in conjunction with dosimetric adjustments for calculating the human
equivalent dose (HED) to estimate the point of departure (POD). A linear low-dose
extrapolation from this POD was performed to derive the OSF. Has the modeling been
appropriately conducted and clearly described based on EPA's draft Benchmark Dose
Technical Guidance Document (U.S. EPA, 2000)? Has the choice of the benchmark
response (BMR) for use in deriving the POD (i.e., a BMR of 10% extra risk of the incidence
of liver tumors in female mice) been supported and clearly described?
Comments: Six reviewers generally considered that the BMD modeling was appropriately
conducted and clearly described. One reviewer stated that the BMD modeling approach was
clearly described, but did not provide a critical assessment because modeling was outside the
reviewer's area of expertise. Two of the reviewers specifically commented that the rationale for
using 10% extra risk of the liver tumor incidence in female mice was well supported. One
reviewer recommended changing the text on page 94 from "the multistage model" to "the
multistage-cancer model." One reviewer offered no comment.
Response: It is technically correct that the "multistage model-cancer" was used for analysis of
cancer data; however, the model is mathematically identical to the multistage model.
Clarification was added as a footnote in Section 5.4.3.1.
7. EPA has concluded that a nonlinear approach is appropriate for extrapolating cancer
risk from male rats to humans because the mode of action analysis suggests that rat
bladder tumors occur only after a series of events that begin with calculi formation. At
exposure levels below the RfD (i.e., below exposure levels needed to form calculi), no
increased risk of cancer is expected. Please comment on whether this approach is
A-15
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
scientifically supported and clearly described. Please identify and provide the rationale for
any other extrapolation approaches that should be selected.
Comments: Six of the eight reviewers agreed with use of a nonlinear approach for extrapolating
cancer risk from male rat bladder tumors to humans. One of these reviewers recommended a
comparison between the RfD and the NOAEL for calculi formation since the RfD was derived
from a developmental endpoint rather than calculi formation. One reviewer stated that modeling
the bladder tumor endpoint is not needed since it was determined that these tumors would not
occur at environmentally relevant doses. One reviewer observed that the data suggest, but do not
prove, a multistep carcinogenic process for bladder tumors and considered bladder stones to be
contributing, but not sufficient, to cause bladder cancer.
Response: EPA agrees with the reviewers who considered a nonlinear extrapolation approach for
male rat bladder tumors to be supported; this approach was retained. A comparison of the
candidate RfD that would be derived from the NOAEL for bladder calculi in the male rat (i.e., a
key event in the mode of action for urinary bladder tumors) and the RfD based on renal toxicity
of biphenyl in rats was added to the discussion of the nonlinear extrapolation approach for
bladder tumors in Section 5.4.3.2. (As noted in response to comments under Charge Questions
A. 1 and A.2, the critical effect for the RfD was changed from a developmental endpoint to renal
toxicity.)
To address concerns raised by two peer reviewers who questioned whether the key events
in the mode of action for biphenyl-induced bladder tumors had been established, EPA added, as
a part of the uncertainty analysis, a linear low-dose extrapolation approach to data for urinary
bladder tumors in male rats. The resulting OSF based on bladder tumors of 2 x 10" (mg/kg-
1 3 1
day)" is fourfold lower than the OSF based on liver tumors of 8 x 10" (mg/kg-day)" . This
analysis, which is presented in Section 5.4.5, Uncertainties in Cancer Risk Values, demonstrates
that the OSF derived from liver tumor data is protective of the OSF that would be derived from
urinary bladder tumor data under the assumption that a linear extrapolation approach for bladder
tumors was supported. The comment related to the role of calculi formation in bladder tumor
carcinogenesis is addressed in a response under Charge Question C.2.
Inhalation Unit Risk (IUR)
8. The draft "Toxicological Review of Biphenyl" did not derive an IUR due to the lack of
available studies. Are there available data to support the derivation of an IUR for
biphenyl? If so, please identify these data.
Comments: None of the reviewers identified studies to support derivation of an IUR. One
A-16
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
reviewer observed that deriving an IUR from the oral slope factor in the absence of inhalation
pharmacokinetics would be uncertain.
Response: EPA agrees that use of route-to-route extrapolation to derive an IUR is not supported.
II. Public Comments
EPA received two sets of public comments. One of these commenters observed that the
draft Toxicological Review ofBiphenyl as well written, concise, and made reasonable assertions
based on the literature. Specific comments and responses are summarized below.
Comments: One public commenter stated that the toxicokinetics section (Section 3) was well
written, but offered two recommendations for providing expanded detail:
(1) Regarding the discussion of 2-hydroxybiphenyl (or ortho-phenylphenol) in Section 3.3.2.1,
emphasis should be given to the fact that urinary bladder tumor formation following 2-
hydroxybiphenyl exposure is dose-dependent and is observed only at high doses. The
commenter provided the following citations: Reitz et al. (1983a), Reitz et al. (1983b), and Smith
etal. (1998).
(2) A description of the potential redox cycling between 4,4'-dihydroxybiphenyl and 4,4'-
dihydroxybiphenylquinone should be included for clarity and completeness.
Response: The sentence related to urinary bladder tumors associated with 2-hydroxybiphenyl
was changed to provide a more accurate description of bladder tumor induction by this chemical,
including the fact that the dose-response relationship is nonlinear, i.e., incidence of bladder
tumors of 96% at 1.25% in diet, but no tumors at the lower concentration of 0.625%) (Kwok et
al., 1999; Hiraga and Fujii, 1984). One of the Reitz et al. studies and the Smith et al. (1998)
study were already cited in the Toxicological Review; the second Reitz et al. (1983) study did
not contribute substantive new information to the Toxicological Review and therefore was not
added. A focused literature search did not locate any studies on metabolism of 4,4'-
dihydroxybiphenyl to the semiquinone and the potential redox cycling between 4,4'-
dihydroxybiphenyl and 4,4'-dihydroxybiphenylquinone. Because this metabolic pathway is
speculative, it was not included in the Toxicological Review.
Comments: One public commenter recommended that limitations of the Sun Company Inc.
(1976) inhalation study be reiterated in the summary of the noncancer endpoints, and that more
clear explanations be added in Section 4.2, Subchronic and Chronic Studies and Cancer
Bioassays in Animals—Oral and Inhalation, to clarify the reasons some studies were considered
more reliable than others. Another public commenter pointed out that the protocols used to
evaluate the studies relied upon in the assessment were not defined.
A-17
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Response: A summary of the limitations of the Sun Company Inc. (1977a) study was included in
Section 4.6.2 and reiterated in Section 5.2.1. The evaluation of study quality was consistent with
EPA guidance, including A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002). Reference to relevant agency guidance was added to Section 5.2.1.
A new appendix, Literature Search Strategy and Study Selection, provides additional information
on study selection strategy and identification of EPA guidance documents used to guide study
evaluation.
Comments: Both public commenters stated that the critical effect selected for derivation of the
RfD, fetal skeletal anomalies (missing, delayed or unossified sternebrae), should be considered
as a non-adverse variation, noting that delayed sternebrae ossification would be expected to fully
ossify within a few days postnatally, and would have no impact on the viability or function of the
offspring. These commenters cited Carney and Kimmel (2007) and Marr et al. (1992) as
support. One of the public commenters also considered the delayed ossification to be secondary
to maternal toxicity (reduced body weight). Therefore, the public commenters argued that
delayed sternebrae ossification should not be the critical effect used to calculate the oral RfD.
Response: As discussed in responses under Charge Questions A.l and A.2, EPA agrees that the
uncertainties in the interpretation of fetal skeletal anomalies reported in Khera et al. (1979) as an
adverse effect were not adequately weighed in selecting this endpoint as the critical effect for the
RfD, and discussion of the issues associated with interpretation of these anomalies was expanded
in Sections 4.3.1 and 4.6.1. Among the anomalies listed, missing or unossified sternebrae was
the only endpoint elevated with increasing dose at doses lower than 1,000 mg/kg-day (i.e., the
dose associated with maternal toxicity), and this endpoint was retained as a candidate critical
effect.
Consistent with peer reviewer recommendations, the robust toxicity studies in rats and
mice by Umeda et al. (2005, 2002) were also considered as candidate principal studies, with the
rationale clarified in Section 5.1.1. Consistent with peer reviewer recommendations, renal
lesions, and in particular renal papillary mineralization in male rats, was selected as the critical
effect. The rationale for selection of this critical effect is provided in Section 5.1.2.
Comments: One public commenter offered the determination that the in vitro genotoxicity
evaluation of biphenyl was negative to slightly equivocal and the in vivo data were negative,
implying that biphenyl is not genotoxic.
Response: EPA disagrees with the conclusion that biphenyl is not genotoxic, although overall
there is not enough evidence to conclude that biphenyl is mutagenic or can react directly with
A-18
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
DNA. The discussions in Section 4.5.2 and Appendix C were revised to more precisely
characterize the evidence for the genotoxicity of biphenyl and its metabolites.
Comments: One public commenter agreed with the overall conclusion that bladder tumors
(Section 4.7) are secondary to calculi formation and are not caused by a genotoxic mode of
action. This commenter recommended that a discussion of the reversibility of calculi formation
as reported by Booth et al. (1961) be added to Section 4.7.
Response: Booth et al. (1961) reported that urine volume, urine turbidity, and histopathological
lesions, including focal tubular dilation and cellular fibrous tissue formation, were increased in
male albino rats exposed to biphenyl in the diet for 120 days compared to control. After
exposure was stopped and the rats were fed a control diet for 30 days, the severity of these
effects decreased. Effects mostly disappeared after being on the control diet for 60 days. The
formation of calculi was not reported in the study. Although reversibility of kidney lesions was
observed, this study did not directly demonstrate calculi formation was reversible. Therefore,
this study was not included in the discussion of mode of action of bladder tumors (Section 4.7).
Comment: One public commenter pointed to chronic studies that provided no evidence that
biphenyl is carcinogenic in rats (Shiraiwa et al., 1989; Ambrose et al., 1960; Pecchiai and
Saffiotti, 1957; Dow Chemical Co, 1953), mice (Imai et al., 1983; Innes et al., 1969; NCI 1968),
dogs (Monsanto, 1946), and Rhesus monkeys (Dow Chemical Co, 1953), and argued that the
total weight of evidence of biphenyl carcinogenicity should be "inadequate information to assess
carcinogenic potential" based on EPA guidance that states that where there is "conflicting
evidence—that is—some studies provide evidence of carcinogenicity but other studies of equal
quality in the same sex and strain are negative." Another public commenter observed that in
light of the susceptibility for liver tumors in female mice, the negative carcinogenicity findings
in male mice in the Umeda et al. (2005) bioassay and in mice in other studies (NCI, 1968; Imai
et al., 1983), and absence of a carcinogenic response in chronic assays in dogs and monkeys, the
cancer descriptor of "suggestive evidence of carcinogenic potential" was not supported. This
commenter also stated that EPA did not define the protocols used to evaluate the studies relied
on in the assessment, in particular with respect to determination that the negative chronic
bioassays in mice, rats, dogs, and monkeys published between 1946 and 1989 were less
informative.
Response: According to the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), a
descriptor of "inadequate information to assess carcinogenic potential" may be appropriate when
there is "conflicting evidence, that is, some studies provide evidence of carcinogenicity but other
studies of equal quality in the same sex and strain are negative." Earlier studies of biphenyl that
A-19
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
provided no evidence of carcinogenicity used more limited study designs, including less-than-
lifetime exposure durations, relatively small numbers of animals, or low doses and therefore
were less informative than the more recent studies by Umeda et al. (2005; 2002). The limitations
of these earlier studies are noted in Section 4.7.1 and summarized in more detail in Section 4.7.2.
In light of the overall weight of evidence, EPA retained the cancer descriptor of "suggestive
evidence of carcinogenic potential" for biphenyl.
EPA guidelines for evaluation of study quality are discussed in responses under General
Charge Question 1 and Charge Question C.l.
Comment: One public commenter noted that in light of the limited inhalation and dermal
exposure data in animals and humans, the cancer descriptor was not justified for all routes of
exposure. A second public commenter submitted an unpublished reported performed by E.I. du
Pont de Nemours and Company's Haskell Laboratory for Health and Environmental Sciences
entitled "Biphenyl: In Vitro Dermal Absorption Rate Testing" (Fasano, 2005) and recommended
that this study be considered in the determination of potential carcinogenicity by non-oral routes
of exposure in Section 4.7.1.
Response: The study by Fasano (2005) measured human skin penetration rates of biphenyl using
an in vitro skin culture system. This study was added to Sections 3.1 and 4.7.1 as evidence that
biphenyl can be absorbed by dermal exposure. Inhalation toxicity studies in rats and mice
reported systemic (liver and kidney) effects, and provided qualitative evidence for absorption of
inhaled biphenyl (Deichmann et al., (1947): Monsanto. (1946): Sun Company Inc., 1977a). As
discussed in a response under Charge Question C.l, the discussion of biphenyl's carcinogenic
potential by other routes of exposure (Section 4.7.1) was revised to better support the cancer
descriptor of "suggestive evidence of carcinogenic potential" by all routes of exposure.
Comments: One public commenter recommended using the two-year bioassay by Umeda et al.
(2002) as the principal study for derivation of the RfD, noting that it was conducted according to
OECD Guideline 453 and yielded the lowest NOAEL of 38 mg/kg-day (calculated from the
dietary concentration of 500 ppm) of the five available dietary studies.
Response: Consistent with comments from external peer reviewers and the public, the principal
study was changed from Khera et al. (1979) to Umeda et al. (2002). See response under Charge
Questions A.l and A.2 for additional discussion of the basis for this revision.
Comment: One public commenter supported the decision to use a nonlinear dose-response
analysis for biphenyl-induced urinary bladder tumors.
Response: No response necessary.
A-20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Comment: One reviewer submitted three unpublished studies: (1) Cytogenetic Effects of
3 Diphenyl-99 on Rat Bone Marrow Cells (conducted by Toxicology Research Laboratory,
4 undated), (2) Biphenyl: In Vitro Dermal Absorption Rate Testing (conducted by Haskell
5 Laboratory for Health and Environmental Sciences, 2005), and (3) Evaluation of Biphenyl FP in
6 the Mouse Bone Marrow Micronucleus Test (conducted by Toxicology & Environmental
7 Research and Consulting, Dow Chemical Company, 2007).
8
9 Response: These studies were added to the Toxicological Review.
10
A-21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
APPENDIX B. LITERATURE SEARCH STRATEGY AND STUDY SELECTION
The literature search strategy used to identify primary, peer-reviewed literature pertaining
to biphenyl was conducted using the databases and keywords listed in Table B-l. References
from health assessments developed by other national and international health agencies were also
examined. Other peer reviewed information, including review articles, literature necessary for
interpretation of biphenyl-induced health effects, and independent analyses of health effects data
were retrieved and included in the assessment where appropriate. EPA requested public
submissions of additional information in December 2007; no submission in response to the data
call-in were received. A comprehensive literature search was last conducted in September 2012.
Figure B-l depicts the literature search, study selection strategy, and the number of
references obtained at each stage of literature screening for all searches. A total of 3,682
references were obtained from the literature searches. A more detailed manual review of titles,
abstracts, and/or papers was then conducted. Selection of studies for inclusion in the
Toxicological Review was based on consideration of the extent to which the study was
informative and relevant to the assessment and general study quality considerations. In general,
relevance and study quality was evaluated as outlined in EPA guidance, including A Review of
the Reference Dose and Reference Concentration Processes (U.S. EPA, 2002) and Methods for
Derivation of Inhalation Reference Concentrations and Application of Inhaled Dosimetry (U.S.
EPA, 1994b). The reasons for excluding references identified by the search are provided in
Figure B-l. A preliminary manual screening of titles and abstracts determined that 3,398 studies
were not relevant to the toxicity of biphenyl. Based on evaluation of the abstracts and full papers
for the 284 considered references, 126 additional references were further eliminated.
The available studies examining the health effects of biphenyl exposure in humans and
laboratory animals are discussed and evaluated in the hazard identification sections of the
assessment (Section 4), with specific limitations of individual studies and of the collection of
studies noted.
The references considered and cited in this document, including bibliographic
information and abstracts, can be found on the Health and Environmental Research Online
(HERO) website6 (http://hero.epa.gov/biphenvn.
6HERO is a database of scientific studies and other references used to develop EPA's risk assessments aimed at
understanding the health and environmental effects of pollutants and chemicals. It is developed and managed in
EPA's Office of Research and Development (ORD) by the National Center for Environmental Assessment (NCEA).
The database includes more than 300,000 scientific articles from the peer-reviewed literature. New studies are
added continuously to HERO.
B-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-l. Details of the literature search strategy employed
Database
Keywords
2007 Search
Pubmed
Toxline
Biosis
Embase
Chemical CASRN: 92-52-4
Synonyms: Biphenyl, Diphenyl, l,l'-Biphenyl, l,l'-Diphenyl, Bibenzene, Biphenyl,
Lemonene, Phenylbenzene, Xenene
Chemical CASRN: 8004-13-5
Synonyms: therminol vp-1, dowtherm A, dinil, dinyl, or diphyl
PubMed: toxic*
Toxline: standard terms such as toxic, eenotoxic. developmental, etc.
Biosis and Embase: toxic, toxico?. toxicit?. chronic, subchronic. acute, oral, inhale?,
inhalation, dermal, intravenous, cancer?, carcinog?, carcinoma?, oncogene?, tumor?,
neoplasm?, mutag?, mutat?, genotox?, fetotox?, embryotox?, teratology?, teratogen?,
reproductive, developmental, neurotox?, immunotox?, pharmacokinetic?,
pharmacodynamic?, PBPK, metabolism, epidemiol?, human study, and human studies
2008 and 2012 Searches
Pubmed
Toxcenter
Toxline
Current Contents
Chemical CASRN: 92-52-4
Synonyms: Biphenyl, Diphenyl, l,l'-Biphenyl, l,l'-Diphenyl, Bibenzene, Biphenyl,
Lemonene, Phenylbenzene, Xenene
Standard toxicoloev fall databases)
Toxicity (including duration, effects to children and occupational exposure); development;
reproduction; teratogenicity; exposure routes; pharmacokinetics; toxicokinetics; metabolism;
body fluids; endocrinology; carcinogenicity; genotoxicity; antagonists; inhibitors
Chemical specific Call databases)
Further limited searches as needed to remove terms related to large classes of chemicals
(PCB, PBDD, etc.) especially when searching for synonyms
TSCATS
Searched by chemical names (including synonyms) and CASRNs
ChemID
Chemfinder
CCRIS
HSDB
GENETOX
RTECS
HERO
Downloaded items already tagged to biphenyl
B-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
References excluded based on preliminary
manual screen of titles/abstracts: 3398
references
Not relevant to biphenyl toxicity in
mammals
Site specific risk assessments
Chemical analytical methods
References excluded based on further manual
review of papers/abstracts: 126 references
Not relevant to biphenyl toxicity in mammals
Inadequate basis to infer exposure
Inadequate reporting of study methods or
results
Animal toxicity studies with mixtures of
chemicals
Abstracts
Not available in English and, based on
abstract, judged not to be informative
Duplicates
References identified based on initial keyword search (see Table B-1): 3682 references
References considered for
inclusion in the Toxicological
Review; references evaluated
based on U.S. EPA (2002, 1994b):
284 references
• References identified in
peer review comments
and through public
submissions
• Literature identified to
support interpretation of
biphenyl toxicity
Other references, including:
References cited in the Toxicological Review: 173 references
References cited in specific sections of the Toxicological Review:
Note: References may be cited in more than one subsection; therefore, the
sum of subsection citations may be higher than the number of references
cited in that section.
• Human studies 6
• Animal studies 28
Oral subchronic and chronic 13
Inhalation 3
Reproductive and developmental 4
Acute and short-term 7
Urinary tract endpoint studies 6
Tumor promotion studies 5
• Other studies 96
Physical and chemical properties 5
Dose response assessment 33
Mechanistic and genotoxic studies 61
Studies supporting mode-of-action 14
• Toxicokinetics 30
Figure B-1. Study selection strategy.
B-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
APPENDIX C. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
MODE OF ACTION
C.l. EFFECTS ON THE URINARY BLADDER OF RATS
Urinary bladder effects in male rats chronically exposed to biphenyl in the diet are
associated with the formation of urinary bladder calculi. Mechanistic studies performed by
Ohnishi and coworkers (2001; 2000a; 2000b) were designed to identify urinary metabolites of
biphenyl, to assess conditions leading to calculi formation, and to determine the composition of
urinary crystals and calculi. Ohnishi et al. (2000a) identified sulphate conjugates of mono- and
dihydroxy biphenyl metabolites in urine and urinary crystals from F344 rats treated with
biphenyl and KHCO3 (to elevate the pH and K+ concentration of the urine). Male F344 rats (five
per group) were administered a diet containing 16,000 ppm biphenyl and 5% potassium
bicarbonate for 7 days (Ohnishi et al.. 2000a). Urine was collected on days 6 and 7 and pooled.
Urinary crystals (i.e., precipitates) were collected, dissolved in acetonitrile, and analyzed by
HPLC to identify metabolites or by inductively coupled plasma spectroscopy to identify
inorganic elements. As shown in Table C-l, biphenyl sulphate conjugates in the urine consisted
primarily of 3,4-dihydroxybiphenyl-3-0-sulphate (40.9% of the total biphenyl sulphate
conjugates) and 3-hydroxybiphenyl (23.4%). No bisulphates were observed (Ohnishi et al..
2000a). In contrast, about 90% of sulphate conjugates in urinary crystals were 4-hydroxy-
biphenyl-O-sulphate, and only 3.9 and 1.06% were 3,4-dihydroxybiphenyl-3-0-sulphate and
3-hydroxybiphenyl, respectively.
Table C-l. Content of biphenyl sulphate conjugates in urine and urinary
crystals from male F344 rats treated with biphenyl and potassium
bicarbonate (to elevate the pH and concentration of the urine)
Biphenyl sulphate conjugates
Urine (%)
Urine crystals (%)
2-Hydroxybiphenyl-O-sulphate
3.32a
0.06
3 -Hydroxybiphenyl-O-sulphate
23.37
1.06
4-Hydroxybiphenyl-O-sulphate
11.94
89.45
4,4'-Dihydroxybiphenyl-0-sulphate
7.17
3.11
2,5 -Dihydroxybiphenyl-O-sulphate
5.62
0.02
3,4-Dihydroxybiphenyl-3 -O-sulphate
40.88
3.90
3,4- Dihydroxybiphenyl-4-O-sulphate
2.27
2.28
2,3- Dihydroxybiphenyl-3-O-sulphate
5.43
0.12
aThe component fraction (%) for each of the sulphate conjugates was estimated from the ratio of the liquid
chromatography tandem MS peak area of the sulfate to the total area.
Source: Ohnishi et al. (2000a').
C-l
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
In a follow-up study, Ohnishi et al. (2000b) evaluated the composition of urinary calculi
in male and female rats exposed to 4,500 ppm biphenyl in the diet for 104 weeks. Urinary
calculi in chronically exposed male rats were composed mainly of 4-hydroxybiphenyl-O-
sulphate, whereas calculi in female rats were composed primarily of 4-hydroxybiphenyl and
potassium sulphate, the hydrolysis products of 4-hydroxybiphenyl-O-sulphate (Ohnishi et al..
2000b). In addition to differences in chemical composition, Ohnishi et al. (2001) observed that
the physical appearance of calculi, including shape, size, and color, differed between male and
female rats. Table C-2 compares the physical characteristics and major chemical constituents of
calculi from male and female rats.
Table C-2. Comparison of the physicochemical characteristics of urinary
calculi in male and female F344 rats
Property
Male
Female
Shape
spheroid, triangular pyramidal, cubical
spheroid
Size
0.3-1.0 cm
homogeneous
Color
white, yellow, brown, gray, black
white, yellow
Main constituent
potassium 4-hydroxybiphenyl-O-sulphate
4-hydroxybiphenyl and potassium sulfate
Source: Umeda et al., (2005); Ohnishi et al. (2000b).
In the Ohnishi et al. (2000b) study, the pH of the urine of treated male rats was in the
range of 7.5-8.5 during the last week of exposure, whereas in female rats it was in the range of
6.5-8.0; there was no difference in urine pH between male and female controls (the range for
both was 6.5-8.0). To investigate if pH of the urine was the only factor associated with calculi
formation, Ohnishi et al. (2001) added potassium bicarbonate (5%), potassium chloride (5%), or
sodium bicarbonate (8%) to the diet for 13 weeks, and reported hydronephrosis and blood in the
urine only in those animals receiving biphenyl plus potassium bicarbonate. Feed consumption
was not affected by the dietary additions, while water intake was greatly increased in all groups
of animals that received biphenyl and/or salts. Neither high urinary potassium levels alone, as
induced by co-feeding of potassium chloride, nor high urinary pH alone, as induced by co-
feeding of sodium bicarbonate, were sufficient to cause kidney damage. It was concluded that a
combination of high urinary pH and high potassium levels was necessary to cause precipitation
of biphenyl sulphate. It was proposed that the crystalline precipitate caused obstruction that led
to hydronephrosis or damaged the transitional epithelium in the bladder causing hyperplasia.
C.2. EFFECTS ON THE LIVER OF MICE
Based on findings of biphenyl-induced liver tumors in female BDFi mice administered
high dietary concentrations of biphenyl for 2 years (Umeda et al.. 2005) (see Section 4.2.1.2.2), a
13-week oral study was performed to assess whether peroxisome proliferation might be induced
(Umeda et al.. 2004a). Groups of male and female BDFi mice (10/sex/group) were administered
C-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
biphenyl in the diet at six different concentrations ranging from 500 to 16,000 ppm. Biphenyl
concentrations >8,000 ppm resulted in significantly decreased final body weights of males and
females. Significantly increased liver weights were noted in the 8,000 and 16,000 ppm groups of
female mice. Evidence of peroxisome proliferation was restricted to the 16,000 ppm group of
female mice. Identification of peroxisomes was based on light microscopy findings of clearly
enlarged hepatocytes filled with eosinophilic fine granules and electron microscopy confirmation
that the granules corresponded to increased numbers of peroxisomes. Electon microscopy was
limited to tissues from 2 female mice in the control and 16,000-ppm groups. Light microscopy
of livers from rats exposed to concentrations <8,000 ppm showed no indications of proliferation
of peroxisomes. There were no indications of other biphenyl-induced liver effects in any of the
groups of mice.
To examine the effects of biphenyl on hepatic peroxisomal enzyme and drug-
metabolizing enzyme activities, Sunouchi et al. (1999) administered biphenyl to BDFi mice at
oral doses of 1.3, 2.6 and 5.2 mmol/kg for 3 days. In female mice, biphenyl administration was
associated with increases in potassium cyanide-insensitive palmitoyl CoA (PCO) oxidation in
liver homogenates (up to 1.9-fold), lauric acid (LA) 12-hydroxylation in liver microsomes (up to
3.8-fold), and cytochrome P450 protein level (as determined by immunochemical analysis).
PCO oxidation and LA 12-hydroxylation were not affected in biphenyl-exposed male mice.
Administration of biphenyl (5.2 mmol/kg) increased pentoxyresorufin O-depentylation (PROD)
(1.8-fold in females; 2.3-fold in males) and P450 protein level (as determined by
immunochemical analysis). Relative liver weights were not affected. This study was reported as
an abstract only; additional study details were not provided.
C.3. ESTROGENIC EFFECTS
Several biphenyl derivatives display estrogenic activity. Schultz et al. (2002) used the
Saccharomyces cerevisiae/LacZ reporter assay to study the estrogenic activity of 120 chemicals
to identify chemical structures that impart estrogenic activity to a molecule. Chemicals without a
hydroxy group, among them biphenyl, were inactive in this assay. The estrogenic activities of
biphenyl metabolites in this assay were 4,4'-dihydroxybiphenyl (median effective concentration
= (2.6 x 10"7 M) > 4-hydroxybiphenyl (1.2 x 10"6 M) > 3-hydroxybiphenyl (9.2 x 10"6 M)
> 2-hydroxybiphenyl (1.8 x 10"5 M). Estrogenic activities of the corresponding hydroxylated di-,
tri-, or tetrachlorobiphenyl metabolites were approximately two orders of magnitude higher,
provided there were no chlorines and hydroxy groups on the same ring.
Kitamura et al. (2003) used MCF-7 cells transfected with an estrogen receptor-luciferase
reporter construct to test biphenyl and its metabolites for estrogenic activity. The starting point
for this investigation was the structural similarity between hydroxylated metabolites of biphenyl
and of 2,2-diphenyl propane, the 4,4'-dihydroxy metabolite of which is bisphenol A, a known
endocrine disrupter. Biphenyl per se displayed no estrogenic activity in this assay. Metabolites
C-3
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
of biphenyl formed by liver microsome preparations were identified after solvent extraction from
reaction media by HPLC-MS. The compounds were also tested in an in vitro competitive
estrogen receptor binding assay. The biphenyl metabolites, 2-, 3-, 4-hydroxybiphenyl, and
4,4'-dihydroxybiphenyl, all exhibited estrogenic activity when the cell culture contained
microsomes from 3-methylcholanthrene-induced rat livers and to a lesser extent, phenobarbital-
induced rat livers, in the presence of NADPH. In the competitive estrogen receptor binding
assay, 4,4'-dihydroxybiphenyl displayed weak binding affinity, while biphenyl and its
monohydroxy metabolites did not show any activity. 4,4'-Dihydroxybiphenyl is one of two
major biphenyl metabolites in rats and mice (Halpaap-Wood et al.. 1981a. b; Meyer and
Scheline. 1976). suggesting that high doses of biphenyl, in the form of this metabolite, might
induce some minor estrogenic effect.
C.4. EFFECTS ON APOPTOSIS
Kokel and Xue (2006) tested a series of benzenoid chemicals (including mesitylene,
cyclohexane, benzene, toluene, and biphenyl) for their ability to suppress apoptosis in the
nematode, Caenorhabditis elegans, a model suitable for the characterization of carcinogens that
act by way of apoptosis inhibition. The study included wild type and three strains of C. elegans
mutants; the ced-3(n2438) mutant (which carries a partial loss-of-function mutation in the ced-
3 gene), the ced-3(n2273) mutant (also partly defective in cell death), and the ced-(n2433)
mutant (a strong loss-of-function ced-3 mutant). Effects on apoptosis were assessed by counting
the numbers of cells that should have died during embryogenesis, but inappropriately survived.
The results indicated that these chemicals did not significantly affect apoptosis in wild type
C. elegans. However, inhibition of apoptosis was apparent in mutant strains ced-3(n2438) and
ced-3(n2273) exposed to benzene, toluene, or biphenyl. The study authors interpreted these
results as indicative of apoptosis-inhibitory activity that does not depend on mutations in a
specific cell-death gene. A lack of apparent apoptosis-inhibitory activity in the strong loss-of-
function ced-3(n2433) mutant was interpreted as indicative that inhibition of apoptosis, rather
than transformation of cell fates, caused the increase in extra cell observed in the other two
mutant strains. All three chemicals also displayed embryotoxicity. Biphenyl and naphthalene
were both shown to suppress apoptosis in C. elegans mutant strain ced-3(n2438) by causing
overexpression of the CED-3 caspase. The authors proposed that benzenoid chemicals that can
form quinones suppress apoptosis in C. elegans via this reactive intermediate, although this was
proven only for benzene, toluene, and naphthalene.
Regulation of apoptosis during embryogenesis is critical, and a recent study by Tan et al.
(2011) showed that inhibition of apoptosis during this stage of development may have
detrimental effects on the nervous system. No literature was identified, however, that
specifically supports an association between inhibition of apoptosis by biphenyl and effects on
embryogenesis.
C-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
C.5. MITOCHONDRIAL EFFECTS
Nishihara (1985) assessed the effects of biphenyl on the respiratory and energy linked
activities of rat liver mitochondria that had been isolated from male Wistar rats. Biphenyl (5-
60 ng/mL in acetone solvent) was added to liver mitochondria and effects on rates of succinate
oxidation and a-ketoglutarate/malate oxidation were assessed by measuring oxygen
consumption. Solvent controls were included in the study. Biphenyl significantly inhibited
state 3 respiration at concentrations >20 [j,g/mL. The inhibition was greater for
a-ketoglutarate/malate oxidation than for succinate oxidation. State 4 respiration was
significantly stimulated by biphenyl; the effect was greater in magnitude for succinate than for
a-ketoglutarate/malate oxidation. Biphenyl also altered mitochondrial membrane permeability,
as evidenced by the instantaneous release of endogenous K+, leading to instantaneous dissipation
of the mitochondrial membrane potential. Inhibition of state 3 respiration is generally considered
to reflect an interference with electron transport. The study author suggested that the biphenyl-
induced stimulation of state 4 respiration may be explained by an uncoupling action on
respiration.
C.6. GENOTOXICITY
Biphenyl. The results of genotoxicity studies of biphenyl are summarized in Table C-3.
In bacterial systems, reverse mutation assays using Salmonella typhimurium and Escherichia coli
provide consistently negative results both with and without the addition of a mammalian
metabolic activation system (S9 rat liver microsomal fraction). Biphenyl did not appear to
induce DNA repair in the SOS chromotest in E. coli (Brams et al., (1987). in the host-mediated
assay in E. coli (Hellmer and Bolcsfoldi. 1992). or in the recombinational repair assay in Bacillus
subtilis (Garrett et al.. 1986; Koiima and Hiraga. 1978). with or without the presence of S9. In
yeasts, biphenyl did induce mitotic recombination and gene conversion both with and without S9
in Saccharomyces cerevisiae strain D7 (Pagano et al, 1988).
Assays for gene and chromosomal mutations of biphenyl-exposed cultured mammalian
cells demonstrate some ability of biphenyl to induce mutagenicity in these systems. Glatt et al.
(Glattetal.. 1992) observed hprt mutations in Chinese hamster V79 cells, but only when cultured
with NADPH-fortified S9 mix. Biphenyl also induced forward mutations in mouse
L5178Y/TK+/" lymphoma cells (Wangenheim and Bolcsfoldi, 1988). The mutation frequency
was increased two- to fourfold in the 10-20% total growth range only, leading the authors to
consider biphenyl to be weakly mutagenic, even though this result was still within study
guidelines for a positive result (p<0.001).
A study of human primary peripheral blood cells reported significant increases in
chromosomal aberrations (CAs) (two- to fourfold higher than solvent controls), micronuclei
(approximately 2.5-fold higher than solvent controls), and sister chromatid exchanges (SCEs)
C-5
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
(less than twofold higher than solvent controls) without S9 that increased with dose
(Rencuzogullari et al.. 2008). These results, however, were accompanied by dose-dependent
cytotoxicity (measured as a reduction in cell replication indices) that was significant at the two
highest doses. Abe and Sasaki (1977) showed a nearly twofold increase in CAs at 30 ng/mL and
a statistically significant increase in SCEs at 15 ng/mL (pairwise t-test) in Chinese hamster lung
cells without activation, but these responses did not meet the authors' criteria for a positive result
due to a lack of dose response. Ishidate et al. (1984) did not find an increase in chromosomal
aberrations up to 125 |ag/m L in the same cell line, in agreement with other studies (Abe and
Sasaki, 1977) and their own past results (Ishidate and Odashima, 1977). However, the same
group subsequently performed the same analysis in the presence of S9 and obtained positive
results that increased with dose (Sofuni et al., 1985).
In the only study to quantify DNA strand breaks, Garberg et al. (1988) found a significant
increase in DNA breakage as detected by the alkaline elution assay in mouse lymphoma cells.
None of the studies for the detection of unscheduled DNA synthesis (UDS) (Hsia et al., 1983a, b;
Probst et al., 1981; Brouns et al., 1979; Williams 1978) in biphenyl-treated rat liver cells
reported positive results, however, indicating that no DNA excision repair was taking place. A
report of the cell transformation assay in human and hamster cells was also negative (Purchase et
al., 1978).
Evaluations of the potential genotoxicity of biphenyl in vivo have been performed in rats
and mice. Two investigations of chromosomal mutations found no evidence of an increase in
CAs in rats following inhalation exposure to biphenyl dust (Johnston et al., 1976) or of
micronuclei in mouse bone marrow after a single gavage dose (Gollapudi et al., 2007). One
group, however, did find evidence of DNA strand breaks in mice using the comet assay. Positive
results were reported for DNA damage in stomach, blood, liver, bone marrow, kidney, bladder,
lung, and brain cells of CD-I mice administered single doses of 2,000 mg biphenyl/kg (Sasaki et
al.. 2002; Sasaki et al.. 1997).
C-6
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-3. Genotoxicity test results for biphenyl
Dose/
Results
Reference
Endpoint
Strain or test system
concentration"
+S9
-S9
Prokaryotic organisms
Reverse
Salmonella typhimurium TA98, 100
2 ng/plate
-
-
Houketal. M989^
mutation
S. typhimurium TA98, 100
25 ng/plate
-
NT
Bos et al. (1988)
S. typhimurium TA98, 100, 1535, 1537,
1538, 1978
77 |ig/platc
-
-
Westinshouse (1977)
S. typhimurium TA98, 100, 1535, 1537
100 ng/plate
-
-
Haworthetal. (1983)
S. typhimurium TA97, 98, 100
100 ng/plate
-
-
Brams et al. (1987)
S. typhimurium TA98, 100, YG1041
250 ng/plate
-
-
Chung and Adris
(2003. 2002)
S. typhimurium TA98, 100, 1532, 1535,
1537, 1538, 2636
500 ng/plate
-
-
Paeano et al. (1988;
1983)
S. typhimurium TA98, 100
800 ng/plate
-
-
Glatt et al. (1992)
S. typhimurium TA98, 1535
1,000 ng/plate
b
NT
Narbonne et al. (1987)
S. typhimurium TA98, 100
1,000 ng/plate
-
-
Koiima and Hiraga
(1978)
S. typhimurium TA98, 100, 1535, 1537
2,500 ng/plate
-
NT
Purchase et al. (1978)
S. typhimurium TA92, 94, 98, 100, 1535,
1537, 2637
5,000 (ig/plate
-
-
Ishidate et al. (1984)
S. typhimurium TA98, 100, 1535,1537,
1538, C3076, D3052, G46
1,000 ng/mL
-
-
Cline and McMahon
(1977)
S. typhimurium C3076, D3052, G46,
TA98, 1000, 1535, 1537, 1538
104-fold range
-
-
Probst et al. (1981)
Escherichia coli WP2, WPluvrA
1,000 ng/mL
-
-
Cline and McMahon
(1977)
E. coli WP2, WP2uvrA
104-fold range
-
-
Probst et al. (1981)
E. coli WP2
1,000 ng/mL
-
-
Koiima and Hiraga
(1978)
DNA repair
E. coli PQ37
SOS chromotest
154 ng/mL
-
-
Brams et al. (1987)
Differential
DNA repair
E. coli K-12 uvrB/recA+,
K-12 uvrB/rccA
Host-mediated assay
25,000 ng/mL
-
-
Hellmer and Bolcsfoldi
(1992)
DNA
recombination/
repair
Bacillus subtilis rec assay
H17 (rec+), M45 (reel
10,000 (ig
-
-
Koiima and Hiraga
(1978)
Non-mammalian eukaryotic organisms
Mitotic
recombination
S. cerevisiae D7
1.5 ng/mL°
+
(DR)
+
(DR)
Paeano et al. (1988)
Gene
conversion
S. cerevisiae D7
1.5 (ig/mL°
+
(DR)
+
(DR)
Mammalian cells in vitro
Mutation
Chinese hamster V79 cells
hprt locus
25 iig/mL'1
+
(DR)
-
Glatt et al. (1992)
C-7 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-3. Genotoxicity test results for biphenyl
Endpoint
Strain or test system
Dose/
concentration"
Results
Reference
+S9
-S9
Mouse lymphoma L5178Y cells
tk locus
3.1 |ig/mL
±e
(T)
(T)
Wangenheim and
Bolcsfoldi (1988. 1986)
Micronuclei
Human primary peripheral blood
lymphocytes
30 |ig/mLr
NT
+
(DR)
(T)
Renciizogullari et al.
(2008)
CAs
Human primary peripheral blood
lymphocytes
50 |ig/mL
NT
+
(DR)
(T)
Renciizogullari et al.
(2008)
Chinese hamster lung (CHL) fibroblasts
15 iig/mL8
+
(DR)
-
Sofuni et al. (1985)
Chinese hamster lung (CHL) fibroblasts
60 |ig/mL
NT
-
Ishidate and Odashima
(1977)
Chinese hamster lung (CHL) fibroblasts
125 |ig/mL
NT
-
Ishidate et al. (1984)
Chinese hamster lung (Don) cells
150 (ig/mL
NT
-
Abe and Sasaki (1977)
DNA strand
breaks
Mouse lymphoma L5178Y cells
DNA alkaline elution assay
7.7 (ig/mL
+
(DR)
(T)
Garbers et al. (1988)
SCEs
Human primary peripheral blood
lymphocytes
50 |ig/mLh
NT
+
(DR)
(T)
Renciizogullari et al.
(2008)
Chinese hamster lung (Don) cells
150 (ig/mL
NT
i1
Abe and Sasaki (1977)
DNA repair
Human diploid lung fibroblasts (HSBP)
15 (ig/mL
NT
-
Snyder and Matheson
(1985)
UDS
Rat primary hepatocytes
15 (ig/mL
NA
-
Hsia et al. (1983b. a)
Rat primary hepatocytes
15 (ig/mL
NA
-
Probst et al. (1981)
Rat primary hepatocytes
15 (ig/mL
NA
-
Williams (1978)
(1989)(1989)(1989)(19
89)(1989)(1989)(1989)
(1989)(1989)(1989)(19
89)( 1989)
Rat primary hepatocytes
150 (ig/mL
NA
-
Brouns et al. (1979)
Cell
transformation
Human diploid lung fibroblasts (WI-38)
OR liver-derived cells (Chang)
250 |ig/mL
-
NT
Purchase et al. (1978)
Syrian hamster kidney cells
BHK 21/cl 13
250 |ig/mL
-
NT
Mammalian systems in vivo
CAs
Rat, Sprague-Dawley, 5 M/dose, 20
inhalation exposures to biphenyl dust 7
hr/d, 5 d/wk; bone marrow after 30 d
50 ppm
NA
-
Johnston et al. (1976)
Micronuclei
Mouse (CD-I), 6 M&F/dose, single oral
gavage; bone marrow at 24 h
800 mg/kg
NA
-
Gollaoudi et al. (2007)
DNA strand
breaks
Mouse (ddY), 4 M/single oral dose;
comet assay on stomach, colon, liver,
kidney, bladder, lung, brain, and bone
marrow at 3 and 24 h
100 mg/kg
NA
+J-k
(DR)
Sasaki et al. (2002)
C-8 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-3. Genotoxicity test results for biphenyl
Endpoint
Strain or test system
Dose/
concentration"
Results
Reference
+S9
-S9
DNA strand
breaks
Mouse (CD-I), 4 M, single oral dose;
comet assay on stomach, liver, kidney,
bladder, lung, brain, and bone marrow at
3, 8, and 24 h
2,000 mg/kg
NA
+k
Sasaki et al. (1997)
aLowest effective dose for positive results; highest dose for negative results.
bTested range of S9 concentrations up to 100 jxl/plate.
°80-85% survival at this dose. Positive results required test compound to be dissolved in DMSO.
dPrecipitation of test compound occurred at 100 (ig/mL.
Positive by 2 to 4-fold only at 10-20% total growth range; this was still within study guidelines for a positive result
(p<0.001).
Positive (p<0.05; pairwise t-test) at >30 (ig/mL after a 24-h incubation but only at 70 (ig/mL after a 48-h incubation
8No information on cytotoxicity provided.
hPositive (p<0.05; pairwise t-test) at 70 (ig/mL after a 24-h incubation and >50 (ig/mL after a 48-h incubation.
'Positive results at 15, 75, and 150 (ig/mL by pairwise t-test, but overall results considered negative by authors due to
lack of dose response.
JPositive (p<0.05; Dunnett test) at 100 mg/kg in colon only; all other organs positive at 1,000 mg/kg.
kPositive results at 24 h only.
± = weakly positive or equivocal result; CA = chromosomal aberration; DR = dose response observed; HPRT =
Hypoxanthine-guanine phosphoribosyltransferase; NA = not applicable; NT = not tested; SCE = sister chromatid
exchange; T = cytotoxicity observed; UDS = unscheduled DNA synthesis
C-9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Biphenyl metabolites. Table C-4 summarizes results from genotoxicity tests of several
biphenyl metabolites, including 2-hydroxybiphenyl (also known as o-phenylphenol, or OPP), 4-
hydroxybiphenyl (the principal metabolite of biphenyl), and 2,5-dihydroxybiphenyl. 2-
Hydroxybiphenyl and its sodium salt have received the most research attention because they are
used as fungicides and anti-bacterial agents and have been found to cause urinary bladder tumors
in male F344 rats with chronic exposure to high concentrations in the diet (seefBalakrishnan et
al.. 2002; Kwok et al.. 1999).
Limited evidence from bacterial assays specifically designed to detect oxidative DNA
damage suggests that 2-hydroxybiphenyl may be mutagenic due to the formation of reactive
oxygen species resulting in the oxidation of DNA bases. This metabolite was positive in two
bacterial strains developed to detect oxidative DNA damage: S. typhimurium strain TA102 and
E. coli strain WP2katEGsodAB (Tani et al., 2007; Fujita et al., 1985). S. typhimurium strain
TA102 was developed with an A:T base pair at the site of mutation and its sensitivity was
increased by the addition of some 30 copies of a plasmid containing the mutant gene that are
available for back mutation. This strain is sensitive to many oxidative mutagenic compounds
including quinones (Levin et al., 1982). E. coli strain WP2katEGsodAB is sensitive to reactive
oxygen species because this strain lacks the detoxification enzymes superoxide dismutase and
catalase (Tani et al., 2007). In other bacterial mutagenicity tests 2-hydroxybiphenyl showed
mixed results. This metabolite was weakly mutagenic in a study of coded chemicals using S.
typhimurium strains TA1535 without the addition of S9 from rat or hamster livers (Haworth et
al., 1983). In this study, strain TA1535 showed a clear monotonic increase in mutagenicity up to
100 |ig/plate; however, this response was slightly less than threefold of control levels, the
criterion for considering a result positive in this strain. Another study using strain TA1535 for
exposures up to 500 |ig/plate did not replicate these results (Ishidate et al., 1984). Exposure of B.
subtilis to 2-hydroxybiphenyl both with and without S9 in the recombinational repair assay
yielded equivocal responses (Koiima and Hiraga. 1978; Hanada. 1977). In an in vivo
mammalian cell assay, 2-Hydroxybiphenyl did not induce chromosomal aberrations (without S9)
in CHL fibroblasts (Ishidate et al.. 1984).
In animal studies, 2-hydroxybiphenyl induced micronuclei (about threefold increase over
controls) and increased cell proliferation (>200-fold increased incorporation of BrdU in DNA) in
the bladder epithelium of male F344 rats exposed to 2% (20,000 ppm) in the diet for 2 weeks,
without evidence of aneuploidy or polyploidy as assayed by fluorescence in situ hybridization
with a DNA probe for rat chromosome 4 (Balakrishnan et al.. 2002). Similar exposure to 2%
NaCl or 2% 2-hydroxybiphenyl + 2% NaCl produced about two- or sixfold increases of
micronuclei in the bladder epithelium, respectively, but neither treatment stimulated bladder
epithelium cell proliferation to the same degree as 2% 2-hydroxybiphenyl in the diet
(Balakrishnan et al., 2002).
C-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
DNA damage was detected by the comet assay after 24-hour exposures in the urinary
bladder of CD-I mice administered single oral doses of 2,000 mg 2-hydroxybiphenyl/kg (Sasaki
et al., 2002). This was the only organ to show evidence of DNA damage at 24 hours; after 3
hours of exposure, the colon (at 100 mg/kg doses), stomach, liver, kidney, and lung also showed
signs of damage. The bone marrow and brain did not show DNA damage occurring at any
timepoint. Another study of DNA strand breaks compared 2-hydroxybiphenyl with its
metabolites, 2,5-dihydroxybiphenyl and phenylbenzoquinone. Using the alkaline elution assay,
DNA strand breaks were detected in the urinary bladder of male or female rats intravesically
injected with 0.05 or 0.1% phenylbenzoquinone, but not with injections of 0.05% 2-
hydroxybiphenyl or 2,5-dihydroxybiphenyl, although DNA damage was found in urinary
bladders from male F344 rats fed the sodium salt of 2-hydroxybiphenyl in the diet for 3 months
at 10,000 or 20,000 ppm, but not at 5,000 or 2,500 ppm (Morimoto et al.. 1989).
Several investigators sought to determine whether 2-hydroxybiphenyl or its metabolites
32
were capable of interacting directly with DNA. Using [ P]-postlabeling to detect DNA adducts
following topical application of 10 or 20 mg of the sodium salt of 2-hydroxybiphenyl or 5 mg of
2,5-dihydroxybiphenyl to the skin of female CD-I mice, several DNA adducts in the skin were
detected (Pathak and Roy. 1993). Similar adducts were formed in vitro when DNA was
incubated with 2-hydroxybiphenyl (170 |ig/mL) or 2,5-dihydroxybiphenyl (186 |ig/mL) in the
presence of metabolic activation from rat skin homogenates (providing cytochrome P450
activation) or a prostaglandin synthase system (Pathak and Roy. 1993). In contrast, Smith et al.
(1998). using a similar technique to that used by Pathak and Roy (1993). were unable to detect
exposure-related DNA adducts in bladder epithelial tissue from male F344 rats fed 800, 4,000,
8,000, or 12,500 ppm 2-hydroxybiphenyl in the diet for 13 weeks. In this experiment, increased
bladder cell epithelium proliferation (i.e., increased BrdU incorporation) was observed at 8,000
and 12,500 ppm, dietary concentrations associated with the development of urinary bladder
tumors in chronically exposed rats (Smith et al.. 1998). Kwok et al. (1999) found no evidence of
binding of radioactivity to DNA extracted from the bladder epithelium of male F344 rats given
single gavage doses of [14C]-labeled 2-hydroxybiphenyl at 15, 50, 250, 500, or 1,000 mg/kg, but
increased protein binding occurred with increasing doses of 250, 500, and 1,000 mg/kg. Kwok
et al. (1999) noted that protein binding increased with increasing dose levels of 250, 500, and
1,000 mg/kg, in parallel with increasing incidence of bladder epithelial lesions (hyperplasia,
papillomas, and carcinomas) in rats chronically exposed to 2-hydroxybiphenyl in the diet at 0,
269, and 531 mg/kg. The authors have also reported a 50- to 70-fold increase in the rate of cell
division in the bladder epithelium of rats treated with 2% OPP in diet in previous studies (Tadi-
Uppala et al., 1996).
Bacterial mutation assays of the major biphenyl metabolite, 4-hydroxybiphenyl, were
positive (threefold increase) in TA98 through 10 ug/plate using 20, 50 or 100 |iL of S9; the
response declined at higher concentrations presumably due to toxicity. 4-Hydroxybiphenyl was
C-ll
DRAFT - DO NOT CITE OR QUOTE
-------
1 marginally mutagenic in TA1535 (twofold increase) but only at the lowest concentration of S9
2 used (20 |iL) (Narbonne et al.. 1987). 2,5-Dihydroxybiphenyl (i.e., phenylhydroquinone) caused
3 in vitro damage to human DNA from plasmid pbcNI in the presence of Cu(II) (Inoue et al..
4 1990).
5
Table C-4. Genotoxicity test results for biphenyl metabolites
Dose/
Results
Endpoint
Strain or test system
concentration"
+S9
-S9
Reference
2-Hydroxybiphenyl in vitro tests
Reverse
mutation
Salmonella typhimurium
TA98, 100, 1537
200 ng/plate
-
-
Haworthetal. (1983)
TA1535
100 ng/plate
-
±
S. typhimurium TA98, 100
100 ng/plate
-
-
Koiima and Hiraga (1978)
S. typhimurium TA97a, 102
10 ng/plate
+
-
Fujita et al. (1985)
S. typhimurium TA92, 94, 98, 100,
1535, 1537, 2637
500 (ig/plate
-
-
Ishidate et al. (1984)
Escherichia coli WP2
100 iig/mL
-
-
Koiima and Hiraga (1978)
E. coli WPlkatEGsodAB, lacking
catalase and superoxide dismutase
0.85 (ig/mL
NT
+
Tani et al. (2007)
DNA
recombination/
Bacillus subtilis rec assay
H17 (rec+), M45 (reel
10,000 (ig/plate
±
±
Koiima and Hirasa (1978):
repair
CAs
Chinese hamster lung (CHL)
fibroblasts
50 (ig/mL
NT
-
Ishidate et al. (1984)
DNA adducts
Rat Liver DNA
[32P]-post labeling method, in
presence of skin homogenate or
prostaglandin synthase activation
systems
170 (ig/mL
+b
NT
Pathak and Rov (1993)
2-Hydroxybiphenyl in vivo tests
Micronuclei
Rat (F344) bladder epithelial cells,
exposure in diet for 14 d, 5-9
M/group. Significant cell
proliferation was induced, but no
ploidy changes were observed.
Cytotoxicity not measured.
20,000 mg/kg
NA
+
Balakrishnan et al. (2002)
C-12
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-4. Genotoxicity test results for biphenyl metabolites
Dose/
Results
Endpoint
Strain or test system
concentration"
+S9
-S9
Reference
DNA strand
breaks
Rat (F344) bladder epithelial cells,
exposure in diet for 3-5 mos, 5-10
M&F/group
Alkaline elution assay in bladder
epithelial cells
10,000 mg/kg,
sodium salt in diet
NA
+
Morimoto et al. (1989)
Rat (F344), 6 M&F, 10 min
exposures, alkaline elution assay in
bladder epithelial cells
0.05% injected
intravesically into
bladder
NA
C
Mouse (ddY), 4 M/single oral dose,
comet assay
Sasaki et al. (2002)
Colon (3 h)
100 mg/kg
NA
+
Stomach, colon, bladder, and lung
(3 h)
1,000 mg/kg
NA
+
Stomach, colon, liver, kidney, and
lung (3 h); bladder (24 h)
2,000 mg/kg
NA
+
Bone marrow and brain (3 and 24
h)
2,000 mg/kg
NA
-
Mouse (CD-I), 4 M, single oral
dose, comet assay
2,000 mg/kg
Sasaki et al. (1997)
Stomach (3 and 8 h), liver (3 h),
kidney (3 and 8 h), bladder (8 and
24 h), and lung (3 h)
NA
+
Bone marrow and brain (3, 8, and
24 h)
NA
-
DNA adducts
Mouse (CD-I), 6 F/dose;
[32P]-postlabeling of DNA isolated
from skin
10 or 20 mg applied
to skin
NA
+
Pathak and Rov (1993)
Rat (F344) bladder epithelial cells,
20-40/group, in diet for 13 weeks
[32P]-postlabeling method
Cytotoxicity and cell proliferation
observed >8,000 mg/kg.
12,500 mg/kg
NA
-
Smith etal. (1998)
Rat (F344) bladder epithelial cells,
4 M/group, single dose by oral
gavage, labeled with
|' 4C] -2-hydroxy-biphenyl
(uniformly labeled in phenol ring).
Protein adducts were observed.
1,000 mg/kg
NA
-
Kwok et al. (1999)
4-Hydroxybiphenyl in vitro tests
Reverse
mutation
S. typhimurium TA98
TA1535
10 ng/plate
+
±
NT
NT
Narbonne et al. (1987)
2,5-Dihydroxybiphenyl in vitro or in vivo tests
DNA strand
breaks
Human DNA fragments from
plasmid pbcNI measured by gel
electrophoresis
18.6 (ig/mL
NT
+d
Inoue et al. (1990)
Rat (F344), 6 M&F, 10 min
exposures, alkaline elution assay in
bladder epithelial cells
0.05% injected
intravesically into
bladder
NA
C
Morimoto et al. (1989)
C-13 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Table C-4. Genotoxicity test results for biphenyl metabolites
Endpoint
Strain or test system
Dose/
concentration3
Results
Reference
+S9
-S9
DNA adducts
Rat Liver DNA
[32P]-post labeling method, in
presence of skin homogenate or
prostaglandin synthase activation
systems
186 (ig/mL
+b
NT
Pathak and Rov (1993)
Mouse (CD-I), 6 F/dose;
[32P]-postlabeling of DNA isolated
from skin
5 mg applied to skin
NA
+
Pathak and Rov (1993)
aLowest effective dose for positive results; highest dose for negative results.
bSkin homogenate used as source of cytochrome P450 activation system.
Injection with 0.05% or 0.1% phenylbenzoquinone, a metabolite of 2,5-dihydroxybiphenyl, produced DNA
damage at concentrations of 0.05 or 0.1%, but not at 0.005 or 0.0005%.
dPositive response only in the presence of Cu(II).
± = weakly positive or equivocal result; CA = chromosomal aberration; DR = dose response observed; NA = not
applicable; NT = not tested; UDS = unscheduled DNA synthesis
Synthesis of genotoxicity evidence for biphenyl and its metabolites. A review of the
evidence for the genotoxic potential of biphenyl suggests there may be some ability of this
compound to induce genetic damage. Although bacterial mutagenicity assays are uniformly
negative, even with metabolic activation, several in vitro assays were able to detect weak
evidence of mutagenicity with activation (Glatt et al., 1992; Wangenheim and Bolcsfoldi, 1988).
Indications of the ability to induce chromosomal aberrations were also observed (Sofuni et al.,
1985), although this was accompanied by cytotoxicity in one study (Rencuzogullari et al., 2008).
In addition, evidence of DNA strand breaks was observed in mice in several organs, including
the stomach, blood, liver, bone marrow, kidney, bladder, lung, and brain (Sasaki et al., 2002,
1997). Micronuclei were observed in primary human lymphocytes (Rencuzogullari et al., 2008),
but were not found in another study in mouse bone marrow (Gollapudi et al., 2007), and CAs
were not observed following inhalation exposures in rats (Johnston et al., 1976).
There are indications that the metabolites of biphenyl may be more genotoxic than the
parent compound. Genotoxicity results for the major metabolite, 4-hydroxybiphenyl, and a
minor metabolite, 2-hydroxybiphenyl (i.e., o-phenylphenol, or OPP), can be found in Table C-4.
Metabolism of 2-hydroxybiphenyl may induce oxidative DNA damage resulting from redox
cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone (Sasaki et al.. 2002; Sasaki et
al.. 1997; Pathak and Roy. 1993; Morimoto et al.. 1989). Limited evidence for this can be found
in positive results in two bacterial strains developed to be sensitive to oxidative DNA damage
(Tani et al., 2007; Fujita et al., 1985).
Other investigations in vivo appear to corroborate these findings. Balakrishnan et al.
(2002) reported that 2-hydroxybiphenyl induced micronuclei and increased cell proliferation in
C-14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
the bladder epithelium of male F344 rats. The mechanism of 2-hydroxybiphenyl-induced
micronuclei is not understood, but, as discussed by Balakrishan et al. (2002). possible
mechanisms include: (1) DNA damage from reactive oxygen species (ROS) from redox cycling
between 2,5-dihydroxybiphenyl and phenylbenzoquinone, (2) interference of the mitotic spindle
through covalent modification of proteins, (3) inhibition of enzymes regulating DNA replication,
or (4) micronuclei generation as a secondary response to cytotoxicity or regenerative
hyperplasia.
Finding evidence that biphenyl can react directly with DNA when matebolized would
provide evidence that not only oxidative damage and subsequent cytotoxicity and regenerative
cell proliferation could solely be responsible for findings of genotoxicity; these processes are not
mutually exclusive. No investigations of the DNA binding potential of biphenyl either in vivo or
in activated systems have been reported, but several studies reported on tests performed
specifically on the metabolites. One such study, Pathak and Roy (1993), reported finding DNA
adducts with rat DNA in vitro and from mouse skin treated with 2-hydroxybiphenyl and 2,5-
dihydroxybiphenyl in vivo. However, these results could not be reproduced by other groups
specifically looking at the rat bladder, the target organ for carcinogenicity, following oral
exposures (Kwok et al., 1999; Smith et al., 1998). Although topical application to mouse skin
does not represent the primary route of exposure or target organ for biphenyl, such contradictory
reports do not rule out the possibility that biphenyl metabolites may be able in some
circumstances to bind DNA. However, the Smith and Kwok studies also reported significant
cytotoxicity and cell proliferation, providing more evidence of a secondary source for DNA
damage following biphenyl exposures.
Sasaki et al. (2002, 1997), who reported DNA strand breaks in several mouse organs
following oral exposure to biphenyl, also reported similar damage following oral exposure to 2-
hydroxybiphenyl. However, the timing and the pattern of organs affected was slightly different.
The DNA damage was only detected early (3 hours) after initial exposure in several organs
(stomach, liver, kidney and lung) and began to disappear 8 hours after exposure. The exception
was the bladder, in which damage was first detected at 8 hours and persisted 24 hours after
exposure. A reasonable explanation for these results is that DNA damage was repaired over time
in most organs but was increased in the bladder where this compound becomes concentrated due
to its excretion in the urine.
To summarize, it is unknown if reports of DNA damage following exposures to biphenyl
are caused by a direct reaction with DNA or by indirect damage from cytotoxicity, or ROS
generated from redox cycling of hydroquinone metabolites, or some combination of these
mechanisms. Biphenyl in an activated system was not investigated for its ability to form DNA-
reactive metabolites, but in studies of DNA adduct formation using the metabolites, most were
negative (Kwok et al., 1999; Smith et al., 1998) save for one study of very high doses applied to
skin (Pathak and Roy, 1993). However, several reports outlined above indicate that genetic
C-15
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
damage induced by biphenyl or its metabolites often occurred only after very high doses that
were accompanied by decreased cell survival or was concurrent with redox cycling following
metabolism of 2-hydroxybiphenyl, a minor metabolite. One study that directly tested the
mutagenicity of the major metabolite, 4-hydroxyquinone, in the Salmonella Ames assay was
positive (Narbonne et al., 1987), but no other investigations of this metabolite were located. In
addition, since the relative production of these metabolites is unknown in humans, damage
occurring due to 2-hydroxybiphenyl may still be important for understanding genotoxic risk
following biphenyl exposures. In summary, there is not enough evidence to conclude that
biphenyl is mutagenic or can react with DNA, but the overall implication is that most indications
of genotoxicity following biphenyl exposures are likely to be secondary response resulting from
oxidative damage and cytotoxicity.
C-16
DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX D. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE
DOSE
Datasets used for modeling incidences of nonneoplastic effects in the urinary tract of
male and female F344 rats exposed to biphenyl in the diet for 2 years (Umeda et al.. 2002) are
shown in Table D-l. Datasets used for modeling body weight data, selected clinical chemistry
results, and histopathological kidney effects in male and female BDFi mice exposed to biphenyl
in the diet for 2 years (Umeda et al.. 2005) are shown in Table D-2. The dataset for incidence of
fetuses with missing or unossified sternebrae from Wistar rat dams administered biphenyl by
gavage on GDs 6-15 (Khera et al.. 1979) is shown in Table D-3.
Table D-l. BMD modeling datasets for incidences of nonneoplastic effects in
the urinary tract of male and female F344 rats exposed to biphenyl in the
diet for 2 years
Males (n = 50)
Females (n = 50)
Biphenyl dietary concentration (ppm)
0
500
1,500
4,500
0
500
1,500
4,500
Calculated dose (mg/kg-d)
0
36.4
110
378
0
42.7
128
438
Effect
Renal pelvis
Nodular transitional cell hyperplasia
0
1
1
21
0
0
1
12
Simple transitional cell hyperplasia
6
8
5
19
3
5
12
25
Mineralization
9
6
10
18
12
12
18
27
Other kidney effects
Hemosiderin deposit3
0
0
0
0
4
8
22
25
Papillary mineralization
9
9
14
23
2
6
3
12
Bladder
Combined transitional cell hyperplasia13
0
0
0
45
1
0
1
10
aMale data for incidences of hemosiderin deposits not selected for quantitative analysis..
bFemale data for incidences of combined transitional cell hyperplasia not selected for quantitative analysis.
Source: Umeda et al. (20021.
D-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-2. BMD modeling datasets for body weight, selected clinical
chemistry results, and histopathological kidney effects in male and female
BDFi mice exposed to biphenyl in the diet for 2 years
Endpoint
Biphenyl concentration in the diet (ppm)
0
667
2,000
6,000
Males
Dose (mg/kg-d)
0
97
291
1,050
Histopathological kidney effect
n= 50
n= 49
n= 50
n= 50
Mineralization inner stripe-outer medulla
9
8
14
14
Clinical chemistry parameter
n= 34
n= 39
n= 37
n= 37
BUN (mg/dL)
20.2 ±3.6
22.0 ±4.0
23.2 ±4.4
22.9 ±2.7
Body weight
n= 35
n= 41
n= 41
n= 39
Mean terminal body weight (g)
46.9 ±4.9
43.1 ±7.9
42.9 ±6.0
32.4 ±3.6
Females
Dose (mg/kg-d)
0
134
414
1,420
Histopathological kidney effect
n= 50
n= 50
n= 50
n= 49
Mineralization inner stripe-outer medulla
3
5
12
26
Clinical chemistry parameter
n= 28
n= 20
n= 22
n= 31
AST (IU/L)
75 ±27
120 ±110
211 ±373
325 ± 448
ALT (IU/L)
32 ± 18
56 ±46
134 ±231
206 ± 280
AP (IU/L)
242 ± 90
256±121
428 ± 499
556 ± 228
LDH (IU/L)
268 ± 98
461 ±452
838 ± 2,000
1,416 ±4,161
BUN (mg/dL)
14.9 ±2.0
14.8 ±3.4
21.0 ±20.5
23.8 ± 11.7
Body weight
n= 31
n= 22
n= 25
n= 32
Mean terminal body weight (g)
34.0 ±4.0
32.5 ±3.3
30.5 ±3.1
25.5 ±3.0
Source: Umeda et al. (2005).
Table D-3. BMD modeling dataset for incidence of fetuses with missing or
unossified sternebrae from Wistar rat dams administered biphenyl by
gavage on GDs 6-15
Dose (mg/kg-d)
Effect
0
125
250
500
Fetuses with missing or unossified
4/176
3/236
4/213
16/199
stcrncbracVanimals examined
(number of litters examined)
(16)
(20)
(18)
(18)
"Data from the 1,000 mg/kg-day dose group was not included here because of frank maternal toxicity at that dose.
Source: Khera et al. (19791.
Goodness-of-fit statistics and benchmark results for each of the modeled biphenyl-
induced nonneoplastic effects from the chronically-exposed rats (Umeda et al.. 2002) and mice
(Umeda et al.. 2005) and the gestationally-exposed rats (Khera et al.. 1979) are summarized in
D-2
DRAFT - DO NOT CITE OR QUOTE
-------
Tables D-4 through D-24. Each table of modeled results for a particular effect is followed by the
information from the output file of the best-fitting model for that effect.
Table D-4. Summary of BMD modeling results for incidence of renal
nodular transitional cell hyperplasia in male F344 rats exposed to biphenyl
in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
X2/>-valuea
Residual with
the largest
absolute value
AIC
BMDS
BMDLS
BMD10
BMDL10
Gammab
0.31
0.73
95.02
169.71
74.44
212.00
120.62
Logisticc
0.64
0.74
92.72
178.92
133.35
233.81
192.35
Log-Logisticb
0.31
0.74
95.01
172.40
75.93
216.08
120.70
Log-Probitb
0.31
0.71
95.03
163.38
89.50
202.25
128.71
Multistage (2-degree)d
0.39
-0.99
93.56
109.09
64.15
162.37
116.56
Probit
0.59
0.84
92.76
157.59
117.53
212.09
173.76
Weibullb
0.31
0.75
95.00
175.08
73.08
221.75
121.01
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
Source: Umeda et al. (20021.
Logistic Model with 0.95 Confidence Level
Logistic
BMDL BMD
0 50 100 150 200 250 300 350
dose
14:42 07/05 2012
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
D-3
DRAFT - DO NOT CITE OR QUOTE
-------
Input Data File:
C:\USEPA\BMDS212\Data\Biphenyl\RenalNodularTransCellHyperPlasia_Umeda2 0 02\-Umeda 2 0 02-Renal
Nodular Transitional Cell Hyperplasia F Rat-Logistic-10%.(d)
Gnuplot Plotting File:
C:\USEPA\BMDS212\Data\Biphenyl\RenalNodularTransCellHyperPlasia_Umeda2 0 02\-Umeda 2 0 02-Renal
Nodular Transitional Cell Hyperplasia F Rat-Logistic-10%.pit
Thu Jul 05 14:42:08 2012
BMDS_Model_Run
The form of the probability function is:
P[response] = 1/[1+EXP(-intercept-slope*dose)]
Dependent variable = Incidence
Independent variable = Dose
Slope parameter is not restricted
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0 Specified
intercept = -4.37631
slope = 0.0106422
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
intercept slope
intercept 1 -0.95
slope -0.95 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
intercept -5.07619 0.879668 -6.8003 -3.35207
slope 0.0125723 0.00249823 0.00767588 0.0174688
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-43.8185
-44.3579
-71.3686
92.7157
# Param's
4
2
1
Deviance Test d.f.
1.07873
55.1002
P-value
0.5831
<.0001
Dose
Est. Prob.
Goodness of Fit
Expected Observed Size
Scaled
Residual
0.0000
36.4000
110.0000
378.0000
ChiA2
0.89
0.0062
0.0098
0.0243
0.4197
d.f.
0 .310
0.489
1.214
20.987
0 .000
1.000
1.000
21.000
50
50
50
50
-0.559
0 .735
-0.197
0 .004
P-value
0.6403
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 233.809
BMDL = 192.347
D-4
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-5. Summary of BMD modeling results for incidence of renal
nodular transitional cell hyperplasia in female F344 rats exposed to biphenyl
in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-valuc11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.96
-0.24
69.04
200.54
118.95
276.46
198.73
Logistic
0.69
0.63
69.93
277.38
211.02
343.52
289.03
Log-Logisticb
0.96
-0.26
69.07
203.45
118.10
279.78
196.91
Log-Probitb
0.99
-0.15
68.96
188.92
134.61
261.35
193.58
Multistage (2-degree)c'd
0.99
-0.36
67.19
191.47
121.69
274.42
211.52
Probit
0.76
0.54
69.69
253.65
190.94
324.08
268.17
Weibullb
0.95
-0.27
69.08
207.16
119.11
285.37
201.63
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
Source: Umeda et al. (2002).
Multistage Model with 0.95 Confidence Level
0.2J
o.:
BMDL
BMD
0
50
100
150
200
250
300
350
400
450
11:48 01/13 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalnodularhyper/female/mst_nodhypFrev_MS_2.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalnodularhyper/female/mst_nodhypFrev_MS_2.pit
Thu Jan 13 11:48:49 2011
BHDS Model Run
D-5
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl-beta2*doseA2)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0
Beta (1) = 0
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Beta(l) have been estimated at a boundary point, or
have been specified by the user, and do not appear in the correlation matrix )
Beta (2)
Beta (2) 1
Parameter Estimates
Variable
Background
Beta(1)
Beta(2)
Estimate
0
0
1.39908e-006
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
- Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood)
-32.456
-32.5947
-48.1018
# Param"
4
1
1
Deviance Test d.f.
0.277585
31.2917
P-value
0.9642
<.0001
AIC:
67.1895
Dose
Est. Prob.
Goodness of Fit
Expected Observed Size
Scaled
Residual
0.0000
42.7000
128.0000
438.0000
0.0000
0.0025
0.0227
0.2354
0.000 0.000
0.127 0.000
1.133 1.000
11.770 12.000
50
50
50
50
0 .000
-0.357
-0 . 126
0 . 077
ChiA2
0 .15
d.f.
P-value
0.9853
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 274.422
BMDL = 211.518
BMDU = 351.444
Taken together, (211.518, 351.444)
90% two-sided confidence interval for the BMD
D-6
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-6. Summary of BMD modeling results for incidence of renal simple
transitional cell hyperplasia in male F344 rats exposed to biphenyl in the
diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-valuc11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab'c
0.66
0.71
184.41
284.70
55.27
313.76
113.22
Logistic
0.35
-1.18
185.78
96.07
73.33
171.37
131.76
Log-Logisticb
0.36
0.71
186.41
320.26
58.80
340.21
115.09
Log-Probitb
0.36
0.71
186.41
284.12
100.23
312.44
144.14
Multistage (3-degree)d
0.60
0.74
184.59
201.02
52.30
255.53
107.40
Probit
0.33
-1.22
185.92
90.26
68.00
164.29
124.13
Weibullb
0.36
0.71
186.41
324.89
55.27
344.08
113.14
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit differed by less than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2002).
Gamma Multi-Hit Model with 0.95 Confidence Level
Gamma Multi-Hit
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
dose
11:55 01/13 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Gamma Model. (Version: 2.15; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalsimplehyper/male/gam_rensimphypMrev_gamma.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalsimplehyper/male/gam_rensimphypMrev_gamma.pit
Thu Jan 13 11:55:07 2011
BMDS Model Run
D-7 DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response]= background+(1-
background)*CumGamma[slope*dose,power], where CumGamma(.) is the cummulative Gamma distribution
function
Dependent variable = incidence
Independent variable = dose
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.134615
Slope = 0.00398471
Power = 2.55235
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Power have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
Background Slope
Background 1 -0.27
Slope -0.27 1
Parameter Estimates
Variable Estimate
Background 0.126666
Slope 0.040 8652
Power 18
NA - Indicates that this parameter has hit
has no standard error.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.0734404 0.179892
0.0361236 0.0456068
Std. Err.
0.0271566
0.00241924
NA
bound implied by some inequality constraint and thus
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -89.7871 4
Fitted model -90.2033 2 0.832451 2 0.6595
Reduced model -97.2446 1 14.915 3 0.001891
AIC: 184.407
Goodness of Fit
Scaled
Dose Est._Prob. Expected Observed Size Residual
0.0000
36.4000
110.0000
378.0000
0.1267
0.1267
0.1267
0.3800
6.333 6.000
6.333 8.000
6.333 5.000
19.000 19.000
50
50
50
50
-0.142
0.709
-0.567
0.000
Chi A2
0.84
d.f.
P-value
0.6558
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 313.755
BMDL = 113.219
D-8
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-7. Summary of BMD modeling results for incidence of renal simple
transitional cell hyperplasia in female F344 rats exposed to biphenyl in the
diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab, Weibullb,
Multistage (l-degree)c'd
0.89
0.34
183.87
34.63
25.35
71.12
52.08
Logistic
0.28
1.29
186.14
83.08
66.43
145.87
119.22
Log-Logisticb
0.71
-0.26
185.77
37.52
18.90
71.51
39.91
Log-Probitb
0.41
1.00
185.39
84.12
62.52
120.97
89.91
Probit
0.33
1.22
185.77
75.68
60.94
135.30
110.85
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the gamma and Weibull models took the form of a 1-degree polynomial multistage model and
produced identical goodness of fit statistics and BMD values; the model with the lowest AIC was selected because
BMDL values for models providing adequate fit differed by less than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2002).
Multistage Model with 0.95 Confidence Level
0.7
Multistage
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
400
450
dose
14:01 01/13 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalsimplehyper/female/mst_simplehypFrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalsimplehyper/female/mst_simplehypFrev_MS_l.pit
Thu Jan 13 14:01:13 2011
BHDS Model Run
D-9
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response]
betal*doseAl) ]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00E
Parameter Convergence has been set to: le-008
background + (1-background)*[1-EXP(-
Default Initial Parameter Values
Background = 0.0607741
Beta(1) = 0.00145231
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background
Beta (1)
1
-0 . 61
-0 . 61
1
Parameter Estimates
Variable Estimate Std. Err.
Background 0.057038 *
Beta(1) 0.00148135 *
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -89.8139 4
Fitted model -89.9369 2 0.246113 2 0.8842
Reduced model -106.633 1 33.6378 3 <.0001
AIC:
183.874
Dose
Est. Prob.
Goodness of Fit
Expected
Observed
Size
Scaled
Residual
0.0000
42.7000
128.0000
438.0000
0.0570
0.1148
0.2199
0.5072
2 . 852
5.742
10.995
25.358
3 .000
5.000
12.000
25.000
50
50
50
50
0.090
-0.329
0 .343
-0.101
ChiA2
0 .24
d.f.
2
P-value
0.8850
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level
BMD
BMDL
BMDU
0 . 95
71.1248
52.0766
105.072
Taken together, (52.0766, 105.072) is
90% two-sided confidence interval for the BMD
D-10
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-8. Summary of BMD modeling results for incidence of
mineralization in renal pelvis of male F344 rats exposed to biphenyl in the
diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.35
-0.75
206.13
130.11
42.91
201.71
88.15
Logistic
0.58
-0.79
204.33
98.62
70.79
181.36
130.04
Log-Logisticb
0.34
-0.75
206.14
128.13
36.96
199.42
78.03
Log-Probitb'c
0.64
-0.74
204.13
144.55
96.05
207.88
138.13
Multistage (l-degree)d
0.51
-0.84
204.60
70.84
41.20
145.51
84.62
Probit
0.57
-0.80
204.35
94.16
66.44
175.86
123.70
Weibullb
0.34
-0.75
206.15
131.37
42.84
205.20
88.00
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2002).
Log Pro bit Model with 0.95 Confidence Level
Log Pro bit
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
dose
15:38 01/13 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Probit Model. (Version: 3.2; Date: 10/28/2009)
Input Eiata File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalmineral/male/lnp_minpelvMrev_logprobit.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalmineral/male/lnp_minpelvMrev_logprobit.pit
Thu Jan 13 15:38:28 2011~
BHDS Model Run
D-ll
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response] = Background + (1-Background) *
CumNorm(Intercept+Slope*Log(Dose)), where CumNorm(.) is the cumulative normal distribution
function
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0.18
intercept = -6.59931
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
background intercept
background 1 -0.46
intercept -0.46 1
Parameter Estimates
Variable
background
intercept
slope
Estimate
0.157045
-6.61851
1
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.0932095 0.22088
-7.17111 -6.0659
NA - Indicates that this parameter has hit
has no standard error.
Std. Err.
0.0325697
0.281947
NA
bound implied by some inequality constraint and thus
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -99.607 4
Fitted model -100.063 2 0.91202 2 0.6338
Reduced model -104.101 1 8.98864 3 0.02944
AIC: 204.126
Goodness of Fit
Scaled
Dose Est._Prob. Expected Observed Size Residual
0.0000
36.4000
110.0000
378.0000
0.1570
0.1581
0.1803
0.3653
7.852 9.000
7.905 6.000
9.014 10.000
18.267 18.000
50
50
50
50
0.446
-0.738
0.363
-0.079
Chi A2
0.88
d.f.
2
P-value
0.6434
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 2 07.879
BMDL = 138.127
D-12
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-9. Summary of BMD modeling results for incidence of mineralization
in renal pelvis of female F344 rats exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.57
-0.43
250.89
44.66
27.40
90.32
56.28
Logistic
0.76
0.59
249.10
64.48
48.11
123.84
92.31
Log-Logisticb
<0.001
2.90
263.72
1.33 x 1015
NA
1.58 x 1015
NA
Log-Probitb
<0.001
2.90
263.72
1.54 x 1014
NA
2.21 x 1014
NA
Multistage (l-degree)c'd
0.85
-0.44
248.89
42.68
27.40
87.67
56.28
Probit
0.77
0.57
249.08
62.20
46.34
120.41
89.56
Weibullb
0.56
-0.44
250.89
43.32
27.40
88.56
56.28
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
'Tower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
Source: Umeda et al. (2002).
Multistage Model with 0.95 Confidence Level
Multistage
0.7
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
400
450
dose
16:24 01/13 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalmineral/female/mst_minpelvlFrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/renalmineral/female/mst_minpelvlFrev_MS_l.pit
Thu Jan 13 16:24:18 2011
BHDS Model Run
D-13
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response]
betal*doseAl) ]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00E
Parameter Convergence has been set to: le-008
background + (1-background)*[1-EXP(-
Default Initial Parameter Values
Background = 0.230737
Beta(1) = 0.00118679
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background
Beta (1)
1
-0 . 62
-0 . 62
1
Parameter Estimates
Variable Estimate Std. Err.
Background 0.228898 *
Beta (1) 0.0012018 *
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -122.276 4
Fitted model -122.443 2 0.334544 2 0.846
Reduced model -128.859 1 13.1664 3 0.00429
AIC:
248.887
Dose
Est. Prob.
Goodness of Fit
Expected
Observed
Size
Scaled
Residual
0.0000
42.7000
128.0000
438.0000
0.2289
0.2675
0.3388
0.5445
11.445
13.374
16.942
27.224
12.000
12.000
18.000
27.000
50
50
50
50
0 . 187
-0.439
0 .316
-0.064
ChiA2
0 . 33
d.f.
2
P-value
0.8473
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 87.669
BMDL = 56.2773
BMDU = 172.188
Taken together, (56.2773, 172.188)
90% two-sided confidence
D-14
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-10. Summary of BMD modeling results for incidence of
hemosiderin deposits in the kidney of female F344 rats exposed to biphenyl
in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
X2
/>-valuca
Residual with
the largest
absolute value
AIC
BMDS
BMDLS
BMD10
BMDL10
Gammab, Weibullb,
Multistage (1-degree)0
0.022
2.36
220.99
29.64
21.20
60.87
43.54
Logistic
0.002
2.92
225.98
66.06
52.04
123.37
97.71
Log-Logisticb
0.093
1.75
218.35
19.21
12.74
40.56
26.89
Log-Probitb
0.002
2.82
225.97
74.77
52.43
107.53
75.40
Probit
0.002
2.90
225.57
61.90
49.07
116.90
92.96
Dichotomous-Hilld'e
0.9997
0.026
213.75
34.28
12.76
45.32
23.29
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the only model with an adequate fit (j2 /j-valuc > 0.1).
ev = 0.5 (specified), g = 0.16 (specified), intercept = 0.08 (initialized), slope = 1 (initialized).
Source: Umeda et al. (2002).
Dichotomous-Hill Model with 0.95 Confidence Level
0.7
Dichotomous-Hill
0.5
0.4
0.3
0.2
BMDL BMD
0
100
200
300
400
dose
09:1401/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Dichotomous Hill Model. (Version: 1.2; Date: 12/11/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/hemosiderin/female/dhl_hemosidFrev_dichotomous
hill.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/hemosiderin/female/dhl_hemosidFrev_dichotomous
hill.pit
Fri Jan 14 09:14:35 2011
D-15
DRAFT - DO NOT CITE OR QUOTE
-------
BMDS Model Run
The form of the probability function is: P[response] = v*g +(v-v*g)/[1+EXP(-intercept-
slope*Log(dose))] where: 0 <= g < 1, 0 < v <= lv is the maximum probability of response predicted
by the model, and v*g is the background estimate of that probability.
Dependent variable = incidence
Independent variable = dose
Parameter v is set to 0.5
Parameter g is set to 0.16
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial
v =
g =
intercept =
slope =
Parameter Values
-9999 Specified
-9999 Specified
0.08
1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -v -g have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
intercept slope
intercept 1 -0.99
slope -0.99 1
Variable
intercept
slope
Parameter Estimates
Estimate Std. Err.
-12.5334 5.83724
2.95297 1.43635
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-23.9742 -1.09265
0.137773 5.76817
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
Log(likelihood)
-104 . 87 6
-104 . 87 6
-121.314
# Param's Deviance Test d.f. P-value
4
2 0.000679954 2 0.9997
1 32.8756 3 <.0001
AIC: 213.752
Goodness of Fit
Scaled
Dose Est._Prob. Expected Observed Size Residual
0.0000 0.0800 4.000 4.000 50 0.000
42.7000 0.1600 7.998 8.000 50 0.001
128.0000 0.4401 22.007 22.000 50 -0.002
438.0000 0.4982 24.908 25.000 50 0.026
ChiA2 = 0.00 d.f. = 2 P-value = 0.9997
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 4 5.3249
BMDL = 23.2881
D-16
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-ll. Summary of BMD modeling results for incidence of papillary
mineralization in the kidney of male F344 rats exposed to biphenyl in the diet
for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.63
-0.37
228.81
51.08
28.48
99.83
58.49
Logistic
0.81
0.51
226.99
70.07
52.70
131.45
98.95
Log-Logisticb
<0.001
2.93
241.27
5.64 x 1014
NA
6.68 x 1014
NA
Log-Probitb
0.001
2.93
239.27
5.13 x 1013
NA
7.38 x 1013
NA
Multistage (l-degree)c'd
0.88
-0.40
226.82
44.66
28.45
91.74
58.44
Probit
0.82
0.48
226.96
66.59
49.79
126.42
94.42
Weibullb
0.63
-0.37
228.81
49.89
28.47
98.66
58.48
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
'Tower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
Source: Umeda et al. (2002).
Multistage Model with 0.95 Confidence Level
Multistage
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
dose
11:25 01/14 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/pappmineral/male/mst_papminMrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/rat/pappmineral/male/mst_papminMrev_MS_l.pit
Fri Jan 14 11:25:01 2011
BHDS Model Run
D-17
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response]
betal*doseAl) ]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00E
Parameter Convergence has been set to: le-008
background + (1-background)*[1-EXP(-
Default Initial Parameter Values
Background = 0.168963
Beta(1) = 0.00114658
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background
Beta (1)
1
-0 . 62
-0 . 62
1
Parameter Estimates
Variable Estimate Std. Err.
Background 0.168634 *
Beta(1) 0.00114846 *
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -111.284 4
Fitted model -111.409 2 0.250221 2
Reduced model -117.634 1 12.6991 3
P-value
0.8824
0.005335
AIC:
226.819
Dose
Est. Prob.
Goodness of Fit
Expected
Observed
Size
Scaled
Residual
0.0000
36.4000
110.0000
378.0000
0.1686
0.2027
0.2673
0.4614
8 .432
10.134
13.365
23.071
9.000
9.000
14.000
23.000
50
50
50
50
0 .215
-0.399
0 .203
-0.020
ChiA2
0 . 25
d.f.
2
P-value
0.8839
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 91.741
BMDL = 58.4361
BMDU = 182 . 915
Taken together, (58.4361, 182.915)
90% two-sided confidence interval for the BMD
D-18
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-12. Summary of BMD modeling results for incidence of papillary
mineralization in the kidney of female F344 rats exposed to biphenyl in the
diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
x2
/j-value"
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.11
1.27
139.76
360.00
68.91
397.57
141.55
Logistic0
0.23
1.37
138.04
175.24
129.91
292.33
219.17
Log-Logisticb
0.11
1.27
139.76
388.83
61.62
413.84
130.08
Log-Probitb
0.11
1.27
139.76
356.94
150.95
395.27
217.08
Multistage (l-degree)d
0.21
1.28
138.38
113.15
65.01
232.43
133.53
Probit
0.23
1.36
138.08
164.88
119.64
282.98
206.34
Weibullb
0.11
1.27
139.76
391.23
68.91
415.47
141.55
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2002).
Logistic Model with 0.95 Confidence Level
0.06
BMDL
BMD
0
50
100
150
200
250
300
350
400
450
dose
13:00 01/14 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/female/log_papmineralFrev_logistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/female/log_papmineralFrev_logistic.pit
Fri Jan 14 13:00:44 2011
BMDS Model Run
The form of the probability function is: P[response] = 1/[1+EXP(-intercept-slope*dose)]
D-19 DRAFT - DO NOT CITE OR QUOTE
-------
Dependent variable = incidence
Independent variable = dose
Slope parameter is not restricted
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0 Specified
intercept = -2.67819
slope = 0.00343504
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
intercept slope
intercept 1 -0.78
slope -0.78 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
intercept -2.72974 0.364791 -3.44472 -2.01477
slope 0.00353956 0.00119641 0.00119464 0.00588449
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -65.6458 4
Fitted model -67.0198 2 2.74796 2 0.2531
Reduced model -71.3686 1 11.4455 3 0.009545
AIC: 138.04
Goodness of Fit
Scaled
Dose Est._Prob. Expected Observed Size Residual
0.0000
42.7000
128.0000
438.0000
0.0612
0.0705
0.0931
0.2352
3 . 062
3 . 526
4 . 654
11.758
2 .000
6.000
3 .000
12.000
50
50
50
50
-0 . 626
1.366
-0.805
0 .081
ChiA2
2 . 91
d.f.
2
P-value
0.2330
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 292.331
BMDL = 219.166
D-20
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-13. Summary of BMD modeling results for incidence of combined
transitional cell hyperplasia in the bladder of male F344 rats exposed to
biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab'c
1.00
-0.12
34.54
186.38
125.23
205.40
146.73
Logistic
1.00
0.00
36.51
314.74
151.02
323.93
182.76
Log-Logisticb
1.00
0.00
36.51
283.35
126.46
295.47
147.96
Log-Probitb
1.00
0.00
36.51
227.03
122.78
241.87
140.96
Multistage (3-degree)d
0.39
-1.63
40.12
109.67
93.51
139.41
123.14
Probit
1.00
0.00
36.51
266.72
137.23
280.54
166.54
Weibullb
1.00
0.00
36.51
300.36
131.93
313.72
160.88
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2002).
Gamma Multi-Hit Model with 0.95 Confidence Level
Gamma Multi-Hit
0.4
0.2
BMDL
BMD
0
50
100
150
200
250
300
350
dose
14:1501/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Gamma Model. (Version: 2.15; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/bladdercombinedhyper/male/gam_bladcomhypMrev_gamma.(d
)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/bladdercombinedhyper/male/gam_bladcomhypMrev_gamma.pi
t
Fri Jan 14 14:15:19 2011
D-21
DRAFT - DO NOT CITE OR QUOTE
-------
BMDS Model Run
The form of the probability function is: P[response]= background+(1-
background)*CumGamma[slope*dose,power], where CumGamma(.) is the cummulative Gamma distribution
function
Dependent variable = incidence
Independent variable = dose
Power parameter is restricted as power >=1
Total number of observations = 4Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0192308
Slope = 0.0320399
Power = 8.564 62
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Power have been estimated at a boundary point, or
have been specified by the user, and do not appear in the correlation matrix )
Slope
Slope 1
Variable
Background
Slope
Power
Estimate
0
0.0624215
18
Parameter Estimates
Std. Err.
NA
0.00323795
NA
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.0560752
0.0687677
NA - Indicates that this parameter has hit a bound implied by some inequality constraint and thus
has no standard error.
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood)
-16.2541
-16.2687
-106.633
# Param'
4
1
1
Deviance Test d.f.
0.0290112
180.757
P-value
0.9987
<.0001
AIC:
34.5373
Dose
Est. Prob.
Goodness of Fit
Expected Observed Size
Scaled
Residual
0.0000 0.0000 0.000 0.000 50 0.000
36.4000 0.0000 0.000 0.000 50 -0.000
110.0000 0.0003 0.014 0.000 50 -0.120
378.0000 0.8996 44.981 45.000 50 0.009
ChiA2 = 0.01 d.f. = 3 P-value = 0.9995
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 2 05.404
BMDL = 14 6.7 33
D-22
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-14. Summary of BMD modeling results for incidence of
mineralization in the kidney (inner stripe outer medulla) of male BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab, Weibullb,
Multistage (1-degree)0
0.46
1.03
214.84
369.24
155.65
758.45
319.71
Logistic
0.43
1.07
214.97
454.16
238.75
856.07
446.12
Log-Logisticb'd
0.48
1.01
214.79
341.66
130.84
721.28
276.22
Log-Probitb
0.33
1.24
215.51
710.74
377.36
1,022.10
542.66
Probit
0.44
1.07
214.95
442.78
227.50
844.26
430.21
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
'Tower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
Source: Umeda et al. (2005).
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic —
0.4J
0.2J
o.:
BMDl
BMD
0.05
0
200
400
600
800
1000
dose
12:57 01/17 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/minmedulla/male/lnl_minmedullM_loglogistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/minmedulla/male/lnl_minmedullM_loglogistic.pit
Mon Jan 17 12:57:13 2011
BMDS Model Run
The form of the probability function is: P[response] = background+(1-background)/[1+EXP(-
intercept-slope*Log(dose) ) ]
Dependent variable = incidence
D-23
DRAFT - DO NOT CITE OR QUOTE
-------
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.18
intercept = -8.98323
slope = 1.06986
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
background intercept
background 1 -0.64
intercept -0.64 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
background 0.185925 * * *
intercept -8.77824 * * *
slope 1 * * *
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -104.672 4
Fitted model -105.397 2 1.44976 2 0.4844
Reduced model -106.377 1 3.40987 3 0.3326
AIC: 214.794
Goodness of Fit
Dose Est._Prob. Expected Observed Size
Scaled
Residual
0.0000
97.0000
291.0000
1050.0000
0.1859
0.1979
0.2209
0.2993
9.296 9.000
9.698 8.000
11.043 14.000
14.963 14.000
50
49
50
50
-0.108
-0.609
1.008
-0.298
ChiA2
1.49
d.f.
2
P-value
0.4754
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 721.275
BMDL = 27 6.216
D-24
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-15. Summary of BMD modeling results for incidence of
mineralization in the kidney (inner stripe outer medulla) of female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
//j-value11
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
Gammab
0.70
-0.27
184.21
116.20
76.96
229.86
158.09
Logistic
0.31
1.22
184.34
257.38
205.80
451.19
369.40
Log-Logisticb'e
0.80
-0.18
184.12
127.12
57.98
233.39
122.40
Log-Probitb
0.53
0.80
183.33
253.31
189.78
364.28
272.92
Multistage (l-degree)d
0.92
-0.34
182.23
104.00
76.86
213.63
157.88
Probit
0.38
1.14
183.96
234.00
188.80
417.63
343.46
Weibullb
0.69
-0.28
184.22
113.82
76.94
227.40
158.04
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
Selected model; the model with the lowest BMDL10 was selected because BMDL values for models providing
adequate fit differed by more than threefold.
dBetas restricted to >0.
Source: Umeda et al. (2005).
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 200 400 600 800 1000 1200 1400
dose
13:27 01/17 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/female/lnl_minmedullF_loglogistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/female/lnl_minmedullF_loglogistic.pit
Mon Jan 17 13:27:41 2011
BMDS Model Run
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
BMDL
BMD
D-25
DRAFT - DO NOT CITE OR QUOTE
-------
The form of the probability function is: P[response] = background+(1-background)/[1+EXP(-
intercept-slope*Log(dose))]
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.06
intercept = -9.5037
slope = 1.31777
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 -0.48 0.44
intercept -0.48 1 -0.99
slope 0.44 -0.99 1
Variable
background
intercept
slope
Parameter Estimates
Estimate
0.05773
-8. 90345
1.22989
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -89.0288 4
Fitted model -89.0609 3 0.0641982 1 0.8
Reduced model -107.593 1 37.1286 3 <.0001
AIC: 184.122
Goodness of Fit
Scaled
Dose Est._Prob. Expected Observed Size Residual
0.0000 0.0577 2.887 3.000 50 0.069
134.0000 0.1078 5.391 5.000 50 -0.178
414.0000 0.2307 11.535 12.000 50 0.156
1420.0000 0.5344 26.187 26.000 49 -0.053
ChiA2 = 0.06 d.f. = 1 P-value = 0.8006
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 233.39
BMDL = 122.401
D-26
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-16. BMD model results for serum LDH activity in female BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/>-valuc"
Means
model
/j-valuc11
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDird
BMDLird
All doses
Constant variance
Hillb
<0.0001
NA
0.00
1,687.59
CF
CF
182.66
0.0000
Linear0
<0.0001
0.38
0.34
1,685.52
2,914.91
1,491.53
465.81
0.0026
Polynomial (2-degree)0
<0.0001
0.30
0.34
1,686.01
2,882.07
1,450.54
465.80
0.0011
Polynomial (3-degree)0
<0.0001
0.93
0.31
1,683.73
3,194.19
1,595.47
465.86
1.1 x 10"8
Powerd
<0.0001
0.93
0.31
1,683.73
3,193.16
1,449.38
465.81
0.0036
Non constant variance
Hill
0.91
NA
-0.22
1,461.52
72.34
CF
161.83
107.12
Linearb
0.91
<0.0001
5.08
1,544.20
-9,999.00
720.55
53.40
19.49
Polynomial (2-degree)b
0.91
<0.0001
1.86
1,537.72
554.86
25.81
42.35
6.96
Polynomial (3-degree)b
0.91
<0.0001
5.08
1,544.20
-9,999.00
1,947.93
53.40
0.88
Powerd
0.91
<0.0001
1.33
1,486.07
60.83
41.31
107.91
81.24
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
hRestrict n> 1.
Coefficients restricted to be positive.
dRestrict power >1.
CF = computation failed; NA = not applicable (degrees of freedom for the test of mean fit are <0, the /_2 test for fit is not
valid)
Source: Umeda et al. (20051.
The constant variance models did not fit the variance data. The nonconstant variance
models did not fit the means data. Therefore, none of the models provided an adequate fit to the
data on serum LDH activity in female mice exposed to biphenyl in the diet for 2 years.
D-27
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-17. BMD modeling results for serum AST activity in female BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/>-value"
Means
model
/>-valuc"
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDird
BMDLird
All doses
Constant variance
Hillb
<0.0001
NA
-5.69 x 107
1,264.30
6,722.40
566.24
213.62
0.00
Linear0, Polynomial
(2-degree)0, Powerd
<0.0001
0.72
0.68
1,260.96
1,826.88
1,205.47
595.87
135.74
Non constant variance
Hillb
0.52
NA
0.82
1,121.84
83.86
CF
154.69
114.05
Linear0
0.52
<0.0001
5.04
1,219.20
CF
90.71
21.60
2.76
Polynomial (2-degree)0
0.52
<0.0001
-2.55 x 109
8.00
0.00
CF
185.08
CF
Powerd
0.52
<0.0001
-2.13
1,164.51
106.70
69.43
150.64
110.24
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0, Polynomial
(2-degree)0, Power
<0.0001
0.99
0.01
826.48
648.56
372.37
229.54
33.18
Non constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0
0.78
<0.0001
3.24 x 108
6
0
CF
228.57
CF
Polynomial (2-degree)0
0.78
<0.0001
-2.20 x 109
8
0
CF
219.67
CF
Powerde
0.78
0.28
-0.29
709.33
72.36
44.29
190.33
121.53
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrict n > 1.
Coefficients restricted to be positive.
dRestrict power >1.
"Selected model; only model providing adequate fit to modeled variance and means.
CF = computation failed; NA = not applicable (degrees of freedom for the test of mean fit are <0, the %2 test for fit is not
valid)
Source: Umeda et al. (20051.
D-28
DRAFT - DO NOT CITE OR QUOTE
-------
Power Model with 0.95 Confidence Level
Power
BMDL
BMD
10:47 01/18 2011
BMD and BMDL indicated are associated with a 100% increase from control (1RD), and are in units of mg/kg-day.
Power Model. (Version: 2.16; Date: 10/28/2009)
Input Data File: C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.pit
Tue Jan 18 10:47:11 2011
BMDS Model Run
The form of the response function is: Y[dose] = control + slope * doseApower
Dependent variable = mean
Independent variable = dose
The power is restricted to be greater than or equal to 1
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
lalpha
rho
control
slope
power
10.765
0
75
0.369536
0.980467
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -power have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
lalpha
rho
control
slope
lalpha
1
-1
-0.43
0.85
rho
-1
1
0.37
-0.89
control
-0.43
0.37
1
-0.17
slope
0.85
-0.89
-0.17
1
Variable
lalpha
rho
control
slope
power
Parameter Estimates
Estimate Std. Err.
-12.9059 4.06805
4.54893 0.905641
74.0253 5.21212
0.38893 0.113823
1 NA
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-20.8791 -4.93268
2.7739 6.32395
63.8097 84.2409
0.165841 0.61202
NA - Indicates that this parameter has hit
has no standard error.
bound implied by some inequality constraint and thus
D-29
DRAFT - DO NOT CITE OR QUOTE
-------
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 28 75 74 27 28.1 0.183
134 20 120 126 110 94.6 -0.29
414 22 211 235 373 390 -0.289
Model Descriptions for likelihoods calculated
Model R1: Yij = Mu(i) + e(ij) Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij) Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij) Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)))
Model A3 uses any fixed variance parameters that were specified by the user
Model R: Yi = Mu + e(i) Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) # Param's AIC
A1 -410.240404 4 828.480807
A2 -350.033965 6 712.067929
A3 -350.072753 5 710.145506
fitted -350.666161 4 709.332321
R -412.701435 2 829.402870
Explanation of Tests
Do responses and/or variances differ among Dose levels? (A2 vs. R)
Are Variances Homogeneous? (A1 vs A2)
Are variances adequately modeled? (A2 vs. A3)
Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df p-value
Test 1 125.335 4 <.0001
Test 2 120.413 2 <.0001
Test 3 0.0775771 1 0.7806
Test 4 1.18681 1 0.276
The p-value for Test 1 is less than .05. There appears to be a difference between response
and/or variances among the dose levels. It seems appropriate to model the data
The p-value for Test 2 is less than .1. A non-homogeneous variance model appears to be
appropriate
The p-value for Test 3 is greater than .1. The modeled variance appears to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems to adequately describe the
data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Relative risk
Confidence level = 0.95
BMD = 190.33
BMDL = 121.53 4
D-30
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-18. BMD modeling results for serum ALT activity in female BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Varianc
e model
/>-valuc"
Means
model
7?-valuea
Residual with
the largest
absolute value
AIC
bmd1sd
BMDL1sd
BMD1rd
BMDL1rd
All doses
Constant variance
Hillb
<0.0001
NA
9.61 x 10"7
1,167.39
3,911.09
436.97
160.82
0.00
Linear0, Polynomial
(2-degree)0, Powerd
<0.0001
0.55
0.94
1,164.57
1,613.62
1,106.30
412.90
38.31
Non constant variance
Hillb
0.78
NA
-0.49
1,013.25
116.28
CF
148.75
121.30
Linear0
0.78
<0.0001
1.69 x 1010
6
0
CF
419.08
CF
Polynomial (2-degree)0
0.78
<0.0001
-1.39 x 1011
8
0
CF
87.64
CF
Powerd
0.78
<0.0001
-1.88
1,047.49
90.73
62.72
108.55
77.76
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0,
<0.0001
0.79
-0.22
756.72
518.80
324.41
116.10
0.00
Polynomial (2-degree)0
<0.0001
NA
4.25 x 10"7
758.65
488.92
325.96
170.36
0.00
Powerd
<0.0001
NA
-3.00 x 10"9
758.65
497.95
325.96
167.69
0.00
Non constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0
0.89
<0.0001
-2.59 x 109
6
0
CF
111.13
CF
Polynomial (2-degree)0
0.89
<0.0001
-5.85 x 107
8
0
CF
169.57
CF
Powerd
0.89
NA
0.10
631.43
110.52
67.61
172.25
117.98
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
hRestrict n> 1.
Coefficients restricted to be positive.
dRestrict power >1.
CF = computation failed; NA = not applicable
Source: Umeda et al. (20051.
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models did not fit the means data. Therefore, none of the models provided
an adequate fit to the data on serum ALT activity in female mice exposed to biphenyl in the diet
for 2 years.
D-31
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-19. BMD modeling results for serum AP activity in female BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/j-valuc11
Means
model
/>-value"
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDird
BMDLird
All doses
Constant variance
Hillb
<0.0001
NA
-4.74 x 10"8
1,240.81
642.90
320.63
540.57
180.68
Linear0, Polynomial
(2-degree)0, Powerd
<0.0001
0.31
1.32
1,239.14
1,253.51
919.17
1,208.38
720.75
Non constant variance
Hillb
0.006
NA
-0.93
1,180.07
147.47
CF
177.26
CF
Linear0
0.006
<0.0001
5.04
1,334.76
-9,999.00
244.46
28.02
0.05
Polynomial (2-degree)0
0.006
<0.0001
-2.57 x 1011
8
0
CF
390.64
CF
Polynomial (3-degree)0
0.006
<0.0001
1.89
1,242.58
1,495.81
213.20
1,506.34
333.91
Powerd
0.006
<0.0001
1.41
1,236.21
665.13
345.69
815.01
482.17
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0,
<0.0001
0.55
-0.51
868.21
617.91
361.78
487.67
201.11
Polynomial (2-degree)0
<0.0001
0.95
-0.05
867.85
510.80
393.46
467.69
315.45
Powerd
<0.0001
NA
1.09E-8
869.84
499.45
372.60
464.35
213.97
Non constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0
0.77
<0.0001
4.52 x 109
6
0
CF
465.02
CF
Polynomial (2-degree)0
0.77
NA
0.13
794.19
287.55
183.20
480.63
334.12
Powerd
0.77
NA
-0.21
794.19
285.46
179.35
482.75
333.04
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrict n > 1.
Coefficients restricted to be positive.
dRestrict power >1.
CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models fit the variance data, but did not fit the means data. Therefore,
none of the models provided an adequate fit to the data on serum AP activity in female mice
exposed to biphenyl in the diet for 2 years.
D-32
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-20. BMD modeling results for changes in BUN levels (mg/dL) in male
BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/>-valuc"
Means
model
/>-value"
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDird
BMDLird
Males
All doses
Constant variance
Hillb
0.03
NA
0.25
540.50
CF
CF
CF
CF
Linearod, Polynomial
(2-degree)0, Power
0.03
0.01
-2.00
545.04
2,254.69
1,288.77
12,777.10
7,154.72
Non constant variance
Hillb
0.01
NA
0.25
542.49
CF
CF
CF
CF
Linear0
0.01
0.28
-1.95
540.78
3,134.77
1,690.32
15,745.20
8,512.03
Polynomial (2-degree)0
0.01
0.13
-2.23
542.57
2,029.81
1,459.55
4,649.85
3,312.21
Polynomial (3-degree)0
0.01
0.13
-2.25
542.52
1,688.06
1,324.21
2,974.25
2,291.81
Powerd
0.01
0.13
-2.32
542.51
1,170.31
1,092.10
1,334.64
1,196.80
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0, Polynomial (2-
degree)0, Powerd
0.49
0.32
0.77
420.23
414.78
266.77
2,140.93
1,335.54
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrict n > 1.
Coefficients restricted to be positive.
dRestrict power >1.
CF = computation failed; NA = not applicable
Source: Umeda et al. (20051.
The constant variance models did not fit the variance data. The nonconstant variance models
fit the variance data, but failed to fit the means data. When the data from the highest dose group were
dropped, the constant variance models fit both the variance and means; however, BMDs at the selected
BMRs, both 1SD and 1RD, were higher than the highest observed dose in the model. Therefore,
modeling was not adequate or suitable for the data on BUN level in male mice exposed to biphenyl in
the diet for 2 years.
D-33
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-21. BMD modeling results for changes in BUN levels (mg/dL) in female
BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
7?-valuea
Means
model
/>-valuc"
Residual with
the largest
absolute value
AIC
bmd1sd
BMDL1sd
BMD1rd
BMDL1rd
All doses
Constant variance
Hillb
<0.0001
NA
-3.45 x 10"8
603.61
CF
CF
CF
CF
Linear0, Polynomial
(2-degree)0, Powerd
<0.0001
0.38
1.18
601.53
1,869.01
1,224.15
2,507.85
1,434.76
Non constant variance
Hillb
0.08
NA
-1.21
493.48
141.72
CF
CF
CF
Linear0, Polynomial
(2-degree)0, Powerd
0.08
<0.0001
-1.63
590.70
519.60
216.41
1,191.69
683.73
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0,
<0.0001
0.50
-0.57
417.59
744.99
403.07
921.79
410.67
Polynomial (2-degree)0
<0.0001
0.82
-0.18
417.19
555.48
413.38
627.58
432.73
Powerd
<0.0001
NA
-2.11 x 10"10
419.13
430.03
414.77
436.97
417.75
Non constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0
0.23
0.07
-1.38
300.36
180.70
114.17
1,416.07
916.09
Polynomial (2-degree)0
0.23
NA
-0.93
299.05
263.22
152.60
842.06
495.16
Powerd
0.23
<0.0001
-0.93
297.05
256.90
151.17
925.84
490.39
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
hRestrict n> 1.
Coefficients restricted to be positive.
dRestrict power >1.
CF = computation failed; NA = not applicable
Source: Umeda et al. (20051.
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models fit the variance data, but did not fit the means data. Therefore,
none of the models provided an adequate fit to the data on BUN levels in female mice exposed to
biphenyl in the diet for 2 years.
D-34
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-22. BMD modeling results for changes in mean terminal body weight in
male BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/>-value"
Means
model
/>-valuc"
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDq.ird
BMDLq.ird
All doses
Constant variance
Hillb
<0.0001
0.03
-1.68
716.95
459.61
390.85
358.30
316.09
Linear0, Powerd
<0.0001
0.10
-1.68
714.95
460.46
391.75
359.04
316.87
Polynomial (3-degree)0
<0.0001
0.03
-1.66
716.89
498.04
392.48
390.52
317.33
Non constant variance
Hillb
0.002
NA
-1.52
704.84
600.48
CF
421.46
325.00
Linear0,
0.002
0.59
-1.52
701.13
541.68
460.24
357.54
326.02
Polynomial (3-degree)0
0.002
0.44
-1.42
702.64
643.20
467.09
450.96
328.74
Powerd
0.002
0.38
-1.51
702.84
600.89
464.26
421.53
327.62
Highest dose dropped
Constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0, Polynomial
(2-degree)0, Powerd
0.01
0.05
-1.49
560.11
566.99
328.79
400.33
238.24
Non constant variance
Hillb
Not modeled; number of dose groups less than number of model parameters
Linear0, Polynomial
(2-degree)0, Powerd
0.18
0.001
-1.5
562.10
561.56
308.43
398.66
235.32
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrict n > 1.
Coefficients restricted to be negative.
dRestrict power >1.
CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit either the variance data or the means data. The
nonconstant variance models failed to fit the variance data. When the data from the highest dose
group were dropped, the constant variance models did not fit either the variance data or the
means data. The nonconstant variance models did not fit the means data. Therefore, none of the
models provided an adequate fit to the data on mean terminal body weight in male mice exposed
to biphenyl in the diet for 2 years.
D-35
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-23. BMD modeling results for changes in mean terminal body weight in
female BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
Variance
model
/j-value"
Means
model
/j-value"
Residual with
the largest
absolute value
AIC
BMD1sd
BMDLisd
BMDq.ird
BMDLq.ird
All doses
Constant variance
Hillb
0.36
0.80
-0.21
382.59
387.90
230.17
397.06
243.57
Linearc'd, Polynomial
(2-degree)c, Power6
0.36
0.42
-0.93
382.26
584.12
489.94
583.33
510.85
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrict n > 1.
Coefficients restricted to be negative.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate fit did
not differ by more than threefold.
eRestrict power >1.
Source: Umeda et al. (2005).
Linear Model with 0.95 Confidence Level
Linear
36
34
32
30
28
26
24
BMDL
BMD
0
200
400
600
800
1000
1200
1400
dose
09:20 01/20 2011
BMD and BMDL indicated are associated with a 10% decrease from control (0.1 RD), and are in units of mg/kg-
day.
Polynomial Model. (Version: 2.16; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/termbdwt/female/lin_termbdwtF_linear.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/termbdwt/female/lin_termbdwtF_linear.pit
Thu Jan 20 09:20:01 2011
BHDS Model Run
The form of the response function is: Y[dose] = beta_0 + beta_l'idose + beta_2J'dose,",2 + ...
Dependent variable = mean
D-36
DRAFT - DO NOT CITE OR QUOTE
-------
Independent variable = dose
rho is set to 0
The polynomial coefficients are restricted to be negative
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 11.4937
rho = 0 Specified
beta_0 = 33.4391
beta 1 = -0.00571961
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
alpha beta_0 beta_l
alpha 1 -9.6e-009 9.1e-009
beta_0 -9.6e-009 1 -0.67
beta 1 9.le-0 0 9 -0.67 1
Variable
alpha
beta_0
beta 1
Estimate
11.2518
33.4983
-0.00574262
Parameter Estimates
Std. Err.
1.5172
0.432523
0.000545303
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
8.27818 14.2255
32.6505 34.346
-0.0068114 -0.00467385
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev
0 31
134 22
414 25
1420 32
34
32 . 5
30 . 5
25.5
33 . 5
32 . 7
31.1
25.3
4
3 . 3
3 . 1
3
3 . 35
3 . 35
3 . 35
3 . 35
Scaled Res.
0 . 833
-0 . 32
-0.925
0.264
Model Descriptions for likelihoods calculated
Model A1: Yij = Mu(i) + e(ij) Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij) Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij) Var{e(ij)} = SigmaA2
Model A3 uses any fixed variance parameters that were specified by the user
Model R: Yi = Mu + e(i) Var{e(i)} = SigmaA2
Likelihoods of
Interest
Model
Log(likelihood)
# Param's
AIC
A1
-187.261579
5
384.523158
A2
-185.643849
8
387.287698
A3
-187.261579
5
384.523158
fitted
-188.129218
3
382.258435
R
-226.477701
2
456.955401
Explanation of Tests
Do responses and/or variances differ among Dose levels? (A2 vs. R)
Are Variances Homogeneous? (A1 vs A2)
Are variances adequately modeled? (A2 vs. A3)
Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df p-value
Test 1 81.6677 6 <.0001
Test 2 3.23546 3 0.3567
Test 3 3.23546 3 0.3567
Test 4 1.73528 2 0.4199
The p-value for Test 1 is less than .05. There appears to be a difference between response
and/or variances among the dose levels. It seems appropriate to model the data
The p-value for Test 2 is greater than .1. A homogeneous variance model appears to be
appropriate here
D-37
DRAFT - DO NOT CITE OR QUOTE
-------
The p-value for Test 3 is greater than .1. The modeled variance appears to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems to adequately describe the
data
Benchma
Specified effect =
Risk Type =
Confidence level =
BMD =
BMDL =
k Dose Computation
0.1
Relative risk
0. 95
583.327
510.848
Table D-24. Summary of BMD modeling results for fetal incidence of missing or
unossified sternebrae from Wistar rat dams administered biphenyl by gavage on
GDs 6-15. (The highest dose was not included because of maternal toxicity)
Model
Goodness of fit
Benchmark result (mg/kg-d)
x2
/>-valuc"
Residual with
the largest
absolute value
AIC
BMDS
BMDLj
BMD10
BMDL10
BMDS modeling with sample size = total number of fetuses examined
Gammab
0.44
0.58
227.97
472.48
386.02
554.43
497.84
Logistic
0.18
1.46
228.47
447.48
371.37
614.46
502.98
Log-Logisticb
0.44
0.58
227.97
476.11
388.23
545.44
498.10
Log-Probitb
0.44
0.59
227.97
469.56
379.56
562.13
497.60
Multistage (3-degree)c d
0.37
1.38
204.28
460.22
382.38
585.02
502.28
Probit
0.15
1.48
228.89
448.57
361.27
645.350
510.69
Weibullb
0.44
0.58
227.97
476.62
389.54
543.82
498.17
BMDS modeling with sample size = total number of litters examined
Gammab
0.82
0.17
25.67
473.31
177.26
553.58
349.07
Logistic
0.86
0.42
23.89
447.38
264.21
615.71
379.80
Log-Logisticb'd
0.82
0.17
25.67
476.95
173.39
544.46
348.52
Log-Probitb
0.82
0.17
25.67
470.45
below zero
561.16
340.36
Multistage (2-degree)0
CF
CF
10.27
542.30
243.88
503.58
260.59
Probit
0.85
0.43
23.93
448.31
248.01
646.43
366.98
Weibullb
0.82
0.17
25.67
477.45
177.25
542.86
350.99
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
CF = computation failed
Source: Khera et al. (19791.
D-38
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
Multistage
BMDL
BMD
0
100
200
300
400
500
dose
16:06 09/28 2012
BMD and BMDL indicated are associated with an extra risk of 5%, with total fetuses examined in each dose group
as the sample size and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:\USEPA\BMDS212\Data\Biphenyl\Sternebrae_Kheral97 9\Sternebrae_fetal%_fetalN\mst_Sternebrae_fetal
%_fetalN_M3.(d)
Gnuplot Plotting File:
C:\USEPA\BMDS212\Data\Biphenyl\Sternebrae_Kheral97 9\Sternebrae_fetal%_fetalN\mst_Sternebrae_fetal
%_fetalN_M3.pit
Thu Sep 27 16:41:03 2012
BMDS_Model_Run
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*doseAl-beta2*doseA2-beta3*doseA3) ]
The parameter betas are restricted to be positive
Dependent variable = fetal_pct
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 4
Total number of specified parameters = 0
Degree of polynomial = 3
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.011904
Beta (1) = 0
Beta (2) = 0
Beta (3) = 5.52452e-010
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Beta(l) -Beta(2)
have been estimated at a boundary point, or have been specified by the user,
D-39
DRAFT - DO NOT CITE OR QUOTE
-------
and do not appear in the correlation matrix )
Background Beta(3)
Background 1 -0.51
Beta (3) -0.51 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0.0115907 * * *
Beta (1) 0 * * *
Beta (2) 0 * * *
Beta(3) 5.26214e-GlG * * *
* - Indicates that this value is not calculated.
Warning: Likelihood for the fitted model larger than the Likelihood for the full model.
Error in computing chi-square; returning 2
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -110.686 4
Fitted model -100.14 2 -21.0916 2 2
Reduced model -118.836 1 16.2989 3 0.0009847
AIC: 204.281
Goodness of Fit
Scaled
Dose Est. Prob.
Expected
Observed Size
Residual
0.0000 0.0116
2.040
3.995 176
1.377
125.0000 0.0126
2 . 975
2.997 236
0 .013
250.0000 0.0197
4 .193
4.004 213
-0.093
500.0000 0.0745
14.828
16.000 199
0 .316
ChiA2 = 2.00 d.f.
= 2 P-
value = 0.3670
Benchmark Dose Computation
Specified effect =
0 . 05
Risk Type =
Extra risk
Confidence level =
0 . 95
BMD =
460.221
BMDL =
382.382
BMDU =
576.027
Taken together, (382.382
, 576.027) is
a 90 % two-sided
confidence
interval for the BMD
D-40 DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
BMDL
200 300
dose
16:48 10/26 2012
BMD and BMDL indicated are associated with an extra risk of 5%, with total litters examined in each dose group
as the sample size and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:\USEPA\BMDS212\Data\Biphenyl\Sternebrae_Kheral97 9\Sternebrae_Fetal%_LitterN\lnl_Sternebrae_Feta
l%_LitterN_LogLogistic.(d)
Gnuplot Plotting File:
C:\USEPA\BMDS212\Data\Biphenyl\Sternebrae_Kheral97 9\Sternebrae_Fetal%_LitterN\lnl_Sternebrae_Feta
l%_LitterN_LogLogistic.pit
Thu Sep 27 15:46:15 2012
BMDS Model Run
The form of the probability function is:
P[response] = background+(1-background)/[1+EXP(-intercept-slope*Log(dose))i
Dependent variable = FetalPct
Independent variable = Dose
Slope parameter is restricted as slope
1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.0227
intercept = -8.88585
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background
intercept
slope
1
-0.54
0.54
-0.54
1
-1
0.54
-1
1
Variable
background
intercept
slope
Estimate
0.0172336
-37.7537
5.64407
Parameter Estimates
Std. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
D-41
DRAFT - DO NOT CITE OR QUOTE
-------
* - Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) #
Param's Deviance Test
d.
f. P-value
Full
model
-9.81066
4
Fitted
model
-9.83702
3 0.0527153
1
0.8184
Reduced
model
-10.5318
1 1.44237
3
0.6956
AIC:
25.674
Goodness of Fit
Scaled
Dose
Est. Prob
Expected
Observed Size
Residual
0.0000
0.0172
0.276
0.363 16
0 .168
125.0000
0.0173
0 .345
0.254 20
-0.157
250.0000
0.0186
0 . 334
0.338 18
0 .007
500.0000
0.0804
1.447
1.447 18
-0.000
ChiA2 = 0
.05 d.f.
= 1 P-
-value = 0.8183
Benchmark Dose Computation
Specified
effect =
0 . 05
Risk Type
=
Extra risk
Confidence
level =
0 . 95
BMD =
476.945
BMDL =
173.393
D-42
DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX E. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR
The mouse liver tumor dataset from Umeda et al. (2005) for which dose-response
modeling was performed is shown in Table E-l.
Table E-l. Incidences of liver adenomas or carcinomas in female BDFi mice
fed diets containing biphenyl for 2 years
Biphenyl dietary concentration (ppm)
0
667
2,000
6,000
Reported dose (mg/kg-d)
0
134
414
1,420
HED (mg/kg-d)
0
19
59
195
Tumor incidence
Adenoma or carcinoma
3/48a
8/50
16/493
14/483
aTwo control, one mid-dose, and two high-dose female mice were excluded from denominators because they died
prior to week 52. It is assumed that they did not have tumors and were not exposed for a sufficient time to be at
risk for developing a tumor. Umeda et al. (2005) did not specify the time of appearance of the first tumor.
Source: Umeda et al. (20051.
Summaries of the BMDs, BMDLs, and derived oral slope factors for the modeled mouse
data are presented in Table E-2, followed by the plot and model output file from the best-fitting
model. The incidence of liver tumors exhibited a plateau in animals in two highest dose groups.
To better estimate responses in the low-dose region, the high-dose group was excluded as a
means of improving the fit of the model in the region of interest.
E-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-2. Model predictions for liver tumors (adenomas or carcinomas) in
female BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
r
/j-value"
Residual with
the largest
absolute value
AIC
BIVI D| iidio
BIVID L| | iDio
Cancer slope
factor (risk
per mg/kg-d)
All doses
Multistage (1-, 2-, 3-degree )b.
Gamma0, Weibulf
0.03
2.14
197.37
64.76
37.29
3 x 10"3
Logistic
0.01
2.31
198.96
104.91
71.27
1 x 10"3
Log-Logistic0
0.04
1.97
196.62
50.68
26.80
4 x 10"3
Log-Probit0
0.005
2.58
201.06
128.52
74.43
1 x 10"3
Probit
0.01
2.30
198.80
100.16
67.23
1 x 10"3
Highest dose dropped
Multistage (l-degree)M
0.96
0.04
132.32
18.72
12.15
8 x 10 3
Multistage (2-degree)b
0.96
0.04
132.32
18.72
12.15
8 x 10"3
aValues <0.05 fail to meet conventional goodness-of-fit criteria.
bBetas restricted to >0.
°Power restricted to >1.
dSelected model.
Source: Umeda et al. (2005).
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
0.5
0.4
0.3
0.2
BMDL
BMD
0
50
100
150
200
14:01 09/19 2011
The BMDS graph of multistage (1-degree) model that includes data from the highest dose group.
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-
day.
E-2
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
0.
0.
0.
0.
0.
BMDL
BMD
0
10
20
30
40
50
60
dose
09:33 02/03 2011
The BMDS graph of multistage (1-degree) model with the highest dose dropped. BMD and BMDL
indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Cancer Model. (Version: 1.9; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l.
(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2 011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l.
pit
Thu Feb 03 09:33:34 2011
BMDS Model Run
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: 2.22045e-016
Parameter Convergence has been set to: 1.49012e-008
We are sorry but Relative Function and Parameter Convergence are currently unavailable in
this model. Please keep checking the web site for model updates which will eventually
incorporate these convergence criterion. Default values used.
Default Initial Parameter Values
Background = 0.0638384
Beta (1) = 0.00559363
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.7
Beta (1) -0.7 1
Parameter Estimates
Variable Estimate Std. Err.
Background 0.0630397 *
Beta (1) 0.00562948 *
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
E-3
DRAFT - DO NOT CITE OR QUOTE
-------
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
Log(likelihood)
-64.1585
-64.1595
-70.107
# Param's
3
2
1
Deviance Test d.f.
0.0019921
11.8969
P-value
0.9644
0.00261
AIC:
132.319
Dose
Est. Prob.
Goodness of Fit
Expected Observed Size
Scaled
Residual
0.0000
19.0000
59.0000
0.0630
0.1581
0.3278
3.026
7 . 904
16.064
3 .000
8 .000
16.000
48
50
49
-0.015
0 . 037
-0.019
ChiA2
0 .00
d.f.
1
P-value
0.9644
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 18.7158
BMDL = 12 . 1518
BMDU = 3 6.3895
Taken together, (12.1518, 36.3895)
Multistage Cancer Slope Factor =
is a 90% two-sided confidence interval for the BMD
0.00822924
The urinary bladder tumor dataset from Umeda et al. (2002) for which dose-response
modeling was performed is shown in Table E-3.
Table E-3. Incidences of urinary bladder transitional cell papilloma or
carcinoma in male F344 rats fed diets containing biphenyl for 2 years
Biphenyl dietary concentration (ppm)
0
500
1,500
4,500
Reported dose (mg/kg-d)
0
36.4
110
378
HED (mg/kg-d)
0
10
30
101
Tumor incidence
Papilloma or carcinoma
0/50
0/50
0/50
3 l/49a
'One high-dose male rat was excluded from denominators because of death prior to week 52. It is assumed that this
rat did not have tumors and was not exposed for a sufficient time to be at risk for developing a tumor. Umeda et al.
(20021 did not specify the time of appearance of the first tumor.
Source: Umeda et al. (20021.
Summaries of the BMDs, BMDLs, and a derived oral slope factors for the modeled
mouse data are presented in Table E-4, followed by the plot and model output file from the best-
fitting model.
E-4
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-4. Model predictions for urinary bladder tumors (papillomas or
carcinomas) in male F344 rats exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Benchmark result (mg/kg-d)
x2
/j-value"
Residual with
the largest
absolute value
AIC
BIVI D| iidio
BIVID Lf | iDio
Cancer slope
factor (risk per
mg/kg-d)
Multistage (l-degree)b
0.0002
-3.120
96.71
17.77
13.34
8 x 10"3
Multistage (2-degree)b
0.1713
-1.980
75.50
35.44
30.44
3 x 10"3
Multistage (3-degree)b
0.7113
-1.126
69.10
48.42
41.21
2 x 10 3
aValues <0.05 fail to meet conventional goodness-of-fit criteria.
''Betas restricted to >0.
°Power restricted to >1.
Source: Umeda et al. (2002).
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
0.7
0.6
0.5
0.4
0.3
0.2
BMDL
BMD
0
20
40
60
80
100
22:01 01/30 2011
The BMDS graph of multistage (3-degree) model.
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-
day.
Multistage Cancer Model. (Version: 1.9; Date: 05/26/2010)
Input Data File:
C: /USEPA/IRIS/biphenyl/2 011/rat/bladdertumor/revised/msc_bladturrMrev_MS_3 .(d)
Gnuplot Plotting File:
C: /USEPA/IRIS/biphenyl/2 011/rat/bladdertumor/revised/msc_bladturrMrev_MS_3 . pit
Sun Jan 30 22:01:35 2011
BMDS_Model_Run
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*doseAl-beta2*doseA2-beta3*doseA3)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 4
Total number of specified parameters = 0
Degree of polynomial = 3
E-5
DRAFT - DO NOT CITE OR QUOTE
-------
Maximum number of iterations = 250
Relative Function Convergence has been set to: 2.22045e-016
Parameter Convergence has been set to: 1.49012e-008
**** We are sorry but Relative Function and Parameter Convergence ****
**** are currently unavailable in this model. Please keep checking ****
**** the web sight for model updates which will eventually ****
**** incorporate these convergence criterion. Default values used. ****
Default Initial Parameter Values
Background = 0
Beta (1) = 0
Beta(2) = 0
Beta(3) = 9.80294e-007
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Beta(l)
have been estimated at a boundary point, or have
and do not appear in the correlation matrix )
Beta (3)
-Beta(2)
been specified by the user,
Beta (3)
1
Variable
Background
Beta(1)
Beta(2)
Beta (3)
Parameter Estimates
Estimate
0
0
0
9.27909e-007
Std. Err.
- Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-32.2189
-33.5483
-86.0881
6 9.0966
# Param"
4
1
1
Deviance Test d.f.
2.65884
107.738
P-value
0.4473
<.0001
Dose
Est. Prob.
Goodness of Fit
Expected Observed Size
Scaled
Residual
0.0000 0.0000
10.0000 0.0009
30.0000 0.0247
101.0000 0.6156
ChiA2 = 1.38 d.f.
0.000 0.000 50
0.046 0.000 50
1.237 0.000 50
30.164 31.000 49
3 P-value = 0.7113
0 .000
-0.215
-1.126
0.246
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 48.4236
BMDL = 41.2077
BMDU = 53.8 91
Taken together, (41.2077, 53.891 )
interval for the BMD
is a 90
two-sided confidence
Multistage Cancer Slope Factor
0.00242673
E-6
DRAFT - DO NOT CITE OR QUOTE
------- |