DRAFT - DO NOT CITE OR QUOTE
EPA/63 5/R-11/006D
www.epa.gov/iris
c/EPA
TOXICOLOGICAL REVIEW
OF
TETRAHYDROFURAN
(CAS No. 109-99-9)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
July 2011
NOTICE
It
policy.
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.
does not represent and should not be construed to represent any Agency determination or
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DRAFT - DO NOT CITE OR QUOTE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 FOR TETRAHYDROFURAN
(CAS No. 109-99-9)
LIST OF TABLES	v
LIST OF FIGURES	vi
ABBREVIATIONS AM) ACRONYMS	vii
FOREWORD	ix
AUTHORS, CONTRIBUTORS, AND REVIEWERS	x
1.	INTRODUCTION	1
2.	CHEMICAL AM) PHYSICAL INFORMATION	3
3.	PHARMACOKINETICS	5
3.1.	ABSORPTION	5
3.1.1.	Gastrointestinal Absorption	5
3.1.2.	Respiratory Tract Absorption	8
3.1.3.	Dermal Absorption	9
3.2.	DISTRIBUTION	10
3.3.	METABOLISM	14
3.4.	ELIMINATION	17
3.5.	BIO ACCUMULATION	20
3.6.	PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS	20
3.7.	SUMMARY	21
4.	HAZARD IDENTIFICATION	23
4.1.	STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS	23
4.2.	SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AM) INHALATION	25
4.2.1.	Subchronic Studies	25
4.2.2.	Chronic Studies and Cancer Bioassays	32
4.3.	REPRODUCTIVE/DEVELOPMENTAL TOXICITY STUDIES—ORAL AND
INHALATION	35
4.3.1.	Oral	35
4.3.2.	Inhalation	44
4.4.	OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES	46
4.5.	MECHANISTIC DATA AND OTHER STUDIES	47
4.6.	SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS	47
4.6.1.	Oral	47
4.6.2.	Inhalation	50
4.7.	EVALUATION OI CARCINOGENICITY	59
4.7.1.	Summary of Overall Weight of Evidence	59
4.7.2.	Synthesis of Human, Animal, and Other Supporting Evidence	60
4.7.3.	Mode of Action Information	61
in
DRAFT - DO NOT CITE OR QUOTE

-------
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES	70
4.8.1.	Possible Childhood Susceptibility	70
4.8.2.	Possible Gender Differences	72
4.8.3.	Other	73
5.	DOSE-RESPONSE ASSESSMENTS	74
5.1.	ORAL REFERENCE DOSE (RID)	74
5.1.1.	Choice of Principal Study and Candidate Critical Effects—with Rationale and
Justification	74
5.1.2.	Methods of Analysis	75
5.1.3.	RfD Derivation—Including Application of Uncertainty Factors (UFs)	77
5.1.4.	Previous RfD Assessment	79
5.2.	INHALATION REFERENCE CONCENTRATION (RfC)	79
5.2.1.	Choice of Principal Study and Critical Effect—with Rationale and Justification79
5.2.2.	Methods of Analysis	82
5.2.3.	RfC Derivation—Including Application of Uncertainty Factors (UFs)	85
5.2.4.	Previous RfC Assessment	87
5.3.	CANCER ASSESSMENT	87
5.3.1.	Choice of Study/Data—with Rationale and Justification	87
5.3.2.	Exposure Adjustments and Extrapolation Method	89
5.3.3.	Inhalation Unit Risk	91
5.3.4.	Previous Cancer Assessment	92
6.	MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD_AND DOSE
RESPONSE	93
6.1. HUMAN HAZARD POTENTIAL	93
6.1.1.	Oral Noncancer	93
6.1.2.	Inhalation Noncancer	93
6.1.3.	Cancer	95
7.	REFERENCES	97
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS
AND DISPOSITION	A-l
APPENDIX B. BMD MODELING	B-l
APPENDIX C. SUPPLEMENTAL INFORMATION	C-l
DRAFT - DO NOT CITE OR QUOTE

-------
LIST OF TABLES
Table 2-1. Chemical and physical properties of THF	3
Table 3-1. Pharmacokinetic parameters in rat and mouse plasma following a single gavage
administration of [14C]-THF	6
Table 3-2. Overall percent recovery of radioactivity at 168 hours following gavage
administration of [14C]-THF	8
Table 3-3. Radiolabel concentration in tissues of rats and mice at 168 hours following gavage
administration of [14C]-THF	11
Table 4-1. Changes in absolute and relative thymus and liver weights of F344/N rats and
B6C3Fi mice following subchronic inhalation exposure to THFa	27
Table 4-2. Incidences of selected nonneoplastic lesions in B6C3Fi mice following subchronic
inhalation exposure to THFa	29
Table 4-3. Renal findings in male F344/N rats exposed to THF for 2 years	33
Table 4-4. Liver findings in female B6C3Fi mice exposed to THF for 2 years	34
Table 4-5. Selected findings from one-generation reproductive toxicity study in Wistar rats
exposed to THF in drinking water	36
Table 4-6. Selected findings from a two-generation reproductive toxicity study in Wistar rats
exposed to THF in drinking water	37
Table 4-7. Correlations between decreased pup body weight gain and each of three independent
variables, maternal water intake, THF intake, and number of pups in each litter	43
Table 4-8. Summary of effect levels observed in the two-generation reproduction study in
Wistar rats exposed to THF in drinking water	44
Table 4-9. Summary of effects observed in drinking water toxicity studies with THFa	49
Table 4-10. Summary of findings in developmental, subchronic, and chronic inhalation studies
with THF	56
Table 5-1. F1 and F2 Pup body weight gain changes for RfD derivation from the two-generation
reproductive toxicity study in Wistar rats exposed to THF in drinking water3	76
v
DRAFT - DO NOT CITE OR QUOTE

-------
Table 5-2. BMD modeling results for pup body weight gain in the Wistar rat two-generation
reproductive toxicity study	77
Table 5-3. Measures of liver toxicity in B6C3Fi male mice following subchronic inhalation
exposure to THFa	84
Table 5-4. BMCa modeling results for noncancer effects in male mice, resulting from subchronic
inhalation exposure to THF	85
Table 5-5. Incidences of neoplastic lesions of the livers of female B6C3Fi mice and kidneys of
male F344/N rats exposed to THF 6 hours/day, 5 days/week for 105 weeks	88
Table 5-6. Cancer Multistage modeling results for THF	91
Table B-l. BMD modeling results for pup body weight gain in the Wistar rat two-generation
reproductive toxicity study	B-2
Table B-2. BMCa modeling results for noncancer effects resulting from subchronic inhalation
exposure to THF	B-l5
Table B-3. Summary of model selection and modeling results for best-fitting multistage models
for cancer effects resulting from chronic inhalation exposure to THF	B-21
Table C-l. Comparison of target organ toxicity for THF and its metabolites	C-l 1
Table C-2. Comparative effects of single and multiple daily dosing of GHB	C-13
Table C-3. Mode of action study findings in male F344 rat kidneys following exposure to THF
by inhalation	C-l6
Table C-4. BrdU labeling and MI as a measure of cell proliferation in female B6C3Fi mouse
livers following exposure to THF by inhalation	C-18
Table C-5. Summary of studies on the direct mutagenicity/genotoxicity of THF	C-22
LIST OF FIGURES
3-1. Possible metabolic pathways of THF	15
vi
DRAFT - DO NOT CITE OR QUOTE

-------
ABBREVIATIONS AND ACRONYMS
ABT
1 -aminobenzotriazole
AIC
Akaike Information Criterion
ALT
alanine aminotransferase
AST
aspartate aminotransferase
ATH
atypical tubule hyperplasia
ATPase
adenosine triphosphatase
AUC
area under the curve
BASF
Badische Anilin- und Sodafabrik
BMC
benchmark concentration
BMCL
95% lower bound on the BMC
BMD
benchmark dose
BMDL
95% lower bound on the BMD
BMDS
BMD software
BMR
benchmark response
BPE
benzo[a]pyrene-trans-7,8-dihydrodiol-9,10-epoxide
BrdU
5-bromo-2-deoxyuridine
CASRN
Chemical Abstracts Service Registry Number
Cmax
maximum plasma concentration following administration of a chemical
CNS
central nervous system
C02
carbon dioxide
CPN
chronic progressive nephropathy
CYP450
cytochrome P450
dUTP
deoxyuridine triphosphate
EEG
el ectroencephal ogram
EPA
U.S. Environmental Protection Agency
EROD
ethoxyresorufin-O-deethylase
FOB
functional observational battery
GABA
y-aminobutyric acid
GBL
y-butyrolactone
GGT
y-glutamyl transferase
GHB
y-hydroxybutyric acid
GI
gastrointestinal
GJIC
gap junctional intercellular communication
HEC
human equivalent concentration
i.p.
intraperitoneal
IRIS
Integrated Risk Information System
LA
labeled area
LC
labeled cell
LC50
median lethal concentration
LD50
median lethal dose
LI
labeling index
LOAEL
lowest-observed-adverse-effect level
LOD
limit of detection
vii
DRAFT - DO NOT CITE OR QUOTE

-------
LOEL
MI
NAS
NIOSH
NOAEL
NOEL
NPH
NRC
NTP
PBPK
PCNA
PIso
PND
POD
PON
PROD
RBC
RfC
RfD
RGDR
SD
SSA
Tl/2
THF
TUNEL
UF
VOC
lowest-observed-effect level
mitotic index
National Academy of Sciences
National Institute for Occupational Safety and Health
no-ob served-adverse-effect level
no-ob served-effect level
nitrophenol hydroxylase
National Research Council
National Toxicology Program
physiologically based pharmacokinetic
proliferating cell nuclear antigen
50% reduction of cell protein content
postnatal day
point of departure
paraoxonase
pentoxyresorufin-O-depentylase
red blood cell
reference concentration
reference dose
regional gas dose ratio
standard deviation
succinic semialdehyde
half-life
tetrahydrofuran
the time after administration of a chemical when the maximum plasma
concentration is reached; when the rate of absorption equals the rate of
elimination
terminal deoxynucleotidyl dUTP nick-end-labeling staining
uncertainty factor
volatile organic compound
viii
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
tetrahydrofuran. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of tetrahydrofuran.
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).
IX
DRAFT - DO NOT CITE OR QUOTE

-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER
Ghazi A. Dannan, Ph.D.
National Center for Environmental Assessment - Washington Division
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
Ghazi A. Dannan, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. EPA
David Lai, Ph.D., DABT
Office of Pollution Prevention and Toxics
U.S. EPA
Elizabeth Margosches, Ph.D.
Office of Pollution Prevention and Toxics
U.S. EPA
Jamie B. Strong, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. EPA
CONTRIBUTORS
Karen Hogan, M.S.
National Center for Environmental Assessment
Office of Research and Development
U.S. EPA
Babasaheb Sonawane, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. EPA
CONTRACTOR SUPPORT
Andrew Maier, Ph.D, C.I.H.
Joan Strawson, M.S., M.T.S.C., J.D.
Andrea Wullenweber, M.S.
Jay Zhao, Ph.D.
Toxicology Excellence for Risk Assessment
x
DRAFT - DO NOT CITE OR QUOTE

-------
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 scientists external to EPA. A summary and EPA's disposition of the comments
received from the independent external peer reviewers and from the public is included in
Appendix A.
INTERNAL EPA REVIEWERS
Michael Beringer, M.S.
Region 7
U.S. EPA
J. Michael Davis, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. EPA
Joyce Donohue, Ph.D.
Office of Water
U.S. EPA
EXTERNAL PEER REVIEWERS
John Christopher, Ph.D.
California Environmental Protection Agency
George Corcoran, Ph.D.
Wayne State University
David William Gaylor, Ph.D.
Gaylor and Associates, LLC
Nancy Kerkvliet, Ph.D.
Oregon State University
Lisa Peterson, Ph.D.
The Cancer Center, University of Minnesota
Karl Rozman, M.D.
The University of Kansas Medical Center
XI
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
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
tetrahydrofuran (THF). 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
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) 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, an inhalation unit risk is a
plausible upper bound on the estimate of risk per (j,g/m3 air breathed.
Development of these hazard identification and dose-response assessments for
tetrahydrofuran has followed the general guidelines for risk assessment as set forth by the
National Research Council (NRC, 1983). U.S. Environmental Protection Agency (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, 1986a), Guidelines for Mutagenicity Risk
Assessment (U.S. EPA, 1986b), Recommendations for and Documentation of Biological Values
1
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
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 (U.S. EPA, 1994a), Methods for 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, 2000a), Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b),
Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures (U.S.
EPA, 2000c), 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 fi'om Early-Life Exposure to Carcinogens (U.S. EPA,
2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A Framework
for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA, 2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through January
2011. It should be noted that references have been added to the Toxicological Review after the
external peer review in response to public comments and for the sake of completeness. These
references have not changed the overall qualitative and quantitative conclusions. See Section 7
for a list of the references added after peer review.
2
DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2. CHEMICAL AND PHYSICAL INFORMATION
Tetrahydrofuran (THF) is a synthesized organic compound that is not found in the natural
environment (ACGIH, 2001). It is a colorless, volatile liquid with an ethereal or acetone-like
smell and is miscible in water and most organic solvents. Table 2-1 summarizes the physical and
chemical properties of THF. THF is highly flammable. Upon contact with air, THF can
decompose into explosive peroxides and carbon monoxide.
Table 2-1. Chemical and physical properties of THF
CAS Registry Number
109-99-9
Verschueren (2001)
Synonym(s)
THF; diethyleneoxide; tetramethyleneoxide;
1,4 -epoxy butane; furanidine;
oxacyclopentane
Verschueren (2001)
Melting point, °C
-108.5
Verschueren (2001)
Boiling point, °C
65/66
Verschueren (2001)
Vapor pressure, atm at 20°C
0.173
Verschueren (2001)
Density, at 20°C relative to the
density of H20 at 4°C
0.89
Verschueren (2001)
Flashpoint (closed cup)
-1 to -21.5°C
BASF (1993)
Water solubility
Miscible
NIOSH (1997)
Log Kow
0.46
SRC (2001)
Odor threshold
2-7.4 ppm
60-150 mg/m3
ACGIH (2001);
RIVM (2001)
Molecular weight
72.10
Verschueren (2001)
Conversion factors
1 ppm =2.95 mg/m3
NIOSH (1997)
Empirical formula
c4h8o
Verschueren (2001)
Chemical structure
o
Verschueren (2001)
THF is used as a solvent for polyvinyl chlorides, vinylidene chloride polymers, and
natural and synthetic resins (particularly vinyls), and in topcoating solutions, polymer coatings,
cellophane, protective coatings, adhesives, magnetic strips, and printing inks. It is also used for
Grignard and metal hydride reactions. THF is used as an intermediate in chemical synthesis.
For example, it is used in the preparation of chemicals, including adipic acid, butadiene, acrylic
acid, butyrolactone, succinic acid, 1,4-butanediol diacetate, motor fuels, vitamins, hormones,
pharmaceuticals, synthetic perfumes, organometallic compounds, and insecticides. It is also
3
DRAFT - DO NOT CITE OR QUOTE

-------
1	used in the manufacture of polytetramethylene ether glycol, polyurethane elastomers, and elastic
2	polymers. THF can be used in the fabrication of materials for food packaging, transport, and
3	storage. When THF is used in food processing, it can be an indirect food additive (National
4	Toxicology Program [NTP], 1998).
5	Potential exposures to humans result from anthropogenic sources, primarily from
6	occupational exposures related to THF's use as a solvent for resins, adhesives, printers' ink, and
7	coatings. Exposure to THF is primarily through inhalation or dermal absorption in the
8	workplace. Nonoccupational exposure is uncommon, but may occur via inhalation and oral
9	routes from contamination of the environment (air and water) (NTP, 1998).
10
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
3. PHARMACOKINETICS
3.1. ABSORPTION
3.1.1. Gastrointestinal Absorption
No information on THF absorption from the human gastrointestinal (GI) tract is
available. However, blood and tissue concentration data from a pharmacokinetic study in rats
and mice conducted by DuPont Haskell Laboratory (1998) have demonstrated that THF is
readily absorbed from the GI tract. In this study, single gavage doses of approximately 50 or
500 mg/kg [14C]-THF dissolved in water were administered to male and female F344 rats and
B6C3Fi mice, and the level of THF-associated radioactivity in plasma was monitored for up to
168 hours. The mean values of selected pharmacokinetic parameters for plasma identified in this
study are presented in Table 3-1. In both rats and mice, radioactivity appeared in the plasma
soon after the THF treatment, demonstrating the rapid absorption of THF from the GI tract. In
rats, detectable levels of radioactivity were present in the plasma as early as 15 minutes after
dosing (the earliest time point measured). Maximum plasma concentrations were reached after
approximately 4 hours in the low-dose rats and after 4-8 hours in the high-dose rats. In the low-
dose group, the plasma concentration reached a maximum (Cmax) of 19.8 [j,g THF equivalents/g
in males at 4 hours and 13.8 [j,g THF equivalents/g in females at 3 hours. In the high-dose group,
the Cmax was 71.6 [j,g THF equivalents/g plasma in males at 8.0 hours and 89.2 [j,g THF
equivalents/g plasma in females at 3.2 hours. The Tmax (the time after administration of a
chemical when the maximum plasma concentration is reached; when the rate of absorption
equals the rate of elimination) in females was highly variable. Maximum plasma concentrations
were not proportional to the administered dose, since Cmax values differed by approximately
fourfold for males and sevenfold for females between dose groups, while the administered dose
differed by 10-fold. A similar evaluation of the plasma area under the curve (AUC) data
revealed the same pattern of nonproportionality with dose. This phenomenon could reflect the
saturability of absorption processes at high doses. Also, independent of absorption, dose-
dependent changes in first-pass metabolism could possibly explain this result. Since GI tract
absorption rates have not been measured directly, the data are not adequate to attribute the
nonlinearity in maximum plasma concentrations or AUCs to absorption kinetics. As the values
of many of the kinetic parameters are highly variable (Table 3-1), the study authors (DuPont
Haskell Laboratory, 1998) indicated that there were no gender differences for any of the kinetic
parameters in the rat (statistical significance not reported by the study authors).
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
Table 3-1. Pharmacokinetic parameters in rat and mouse plasma following
a single gavage administration of [14C]-THF

50 mg/kg
500 mg/kg
Male
Female
Male
Female
Rat
Actual dose (mg/kg)
40.3
45.9
428.7
478.3
Tmax (hrs)
4.0
3.0
8.0
3.2
Cmax (|ig equivalents/g)
19.8
13.8
71.6
89.2
T1/2 (hrs)
52.1
50.5
48.0
59.0
AUC (|ig'hr/g)
535.8
319.6
2,825.5
1,998.0
Clearance (g/hr«kg)
75.2
143.6
151.7
239.4
Mouse
Actual dose (mg/kg)
44.3
38.0
490.3
495.9
Tmax (hrs)
0.5
0.4
0.8
1.0
Cmax (|ig equivalents/g)
27.7
19.4
149.4
106.0
T1/2 (hrs)
56.9
51.4
57.3
98.5
AUC (|ig'hr/g)
207.4
157.3
3,237.9
1,904.4
Clearance (g/hr«kg)
213.6
241.6
151.4
260.4
Tic = half-life
Source: Adapted from data in DuPont Haskell Laboratory (1998); data are expressed as mean values.
Similar to the observations in the rat, THF-associated radioactivity appeared rapidly in
mouse plasma after gavage dosing. Fifteen minutes following the 50 mg/kg treatment, a mean
value of 17.4 [j,g THF equivalents/g plasma was observed in females while no radiolabel was
detected in males at this sampling time. Following the 500 mg/kg treatment, the mean values at
15 minutes were 84.8 and 56.8 [^g THF equivalents/g plasma for males and females,
respectively. In the 50 mg/kg dose group, plasma radioactivity reached the Cmax of 27.7 and
19.4 [j,g THF equivalents/g at approximately 30 minutes after dosing in males and females,
respectively. In the 500 mg/kg group, the plasma radioactivity reached Cmax values of 149.4 and
106.0 [j,g THF equivalents/g at approximately 1 hour after dosing in males and females,
respectively. No gender differences were observed for the mouse Tmax values (statistical
significance not reported by the study authors). The mouse Tmax values were shorter than for the
parallel dose-groups in rats, suggesting that the absorption of THF is more rapid in mice than in
rats. As was observed in rats, the Cmax values in mice were not proportional to the administered
dose. However, evaluation of the plasma AUC data for mice suggested that the total absorbed
dose was more than proportional to the administered doses; the AUC was 12-fold higher at the
high dose in females and 16-fold higher at the high dose in males as compared to the AUC in the
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
corresponding low-dose groups. The lack of proportionality of the Cmax and AUC is consistent
with an effect of dose on absorption rate. However, effects of other kinetic parameters such as
metabolism could explain these observations, and therefore, the apparent nonlinearity in plasma
kinetics cannot be attributed only to absorption.
The oral bioavailability of THF has not been assessed directly. However, measurement
of THF-associated radioactivity in the excreta of the rats and mice in the pharmacokinetics study
by DuPont Haskell Laboratory (1998) suggests that most (if not all) of orally administered doses
of THF can be absorbed. In rats and mice, the total radioactivity recovered in urine, feces,
expired air (carbon dioxide [CO2] or volatile organics), tissues, cage wash, and residual feed was
measured over a period of 168 hours after gavage dosing (Table 3-2). The total recovery of
radioactivity (i.e., mass balance) was low in both dose groups of rats and the high-dose group of
mice, which was attributed by the study authors to saturation in the CO2 capture system at early
time points after dosing and limited performance of the solvent used to capture volatile organics.
However, changes in the apparatus for collection of CO2 and volatile organics employed for the
low-dose mice yielded much better recovery of the administered radioactivity. Analysis of data
from the low-dose mice shows that little THF remains unabsorbed from the GI tract, since
recovery of radioactivity in the feces did not account for more than 1.4% of the administered
dose. The amount of THF-associated radioactivity recovered in the feces in these treatment
groups was similar to the low-dose mice, suggesting that THF is nearly completely absorbed
following oral dosing of up to 500 mg/kg in rats and mice.
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
Table 3-2. Overall percent recovery of radioactivity at 168 hours following
gavage administration of [14C]-THF
Sample"
50 mg/kg
500 mg/kg
Rat
Mouse
Rat
Mouse
Male
Female
Male
Female
Male
Female
Male
Female
Urine
4.4
3.5
2.7
5.3
2.2
2.2
3.8
3.6
Feces
1.1
1.0
1.4
0.9
1.0
0.4
1.3
0.8
C02
47.8
47.5
58.2
74.6
21.9
18.8
51.1
36.2
Volatile organics
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
72.7% for females during normal breathing and 78.4% for males and 81.3% for females during
deep breathing. No consistent concentration-related effects on uptake were apparent. These
results suggested that as much as 81.3% of the THF was absorbed or retained in the lung under
acute exposure conditions. In a second experiment, five male subjects were exposed for 3 hours
to mean concentrations of 56 ppm THF, followed by a 1-hour recovery period and then a second
3-hour exposure. Exhaled air was monitored throughout the first 3-hour exposure period. The
percentage of THF in expired air relative to inhaled air was reported as 40% during normal
breathing and 27% during deep breathing. These results correspond to uptake ratios of 60 and
73%, respectively. The same results were observed for five male subjects exposed for a single
3-hour exposure period to a mean THF concentration of 193 ppm THF (experiment 3). The
authors also exposed five male volunteers to approximately 200 ppm (207 ppm for first exposure
and 178 ppm for second exposure) THF for sequential 3-hour exposure periods with a 1-hour
recovery period in between (experiment 4). Blood samples were collected for several of the
exposure protocols (experiments 2, 3, and 4). THF kinetics in blood were highly variable among
individuals. However, the appearance of THF in the blood demonstrates the systemic absorption
of THF from the lungs in exposed humans.
Wagner (1974) also reported on the respiratory tract absorption of THF in four
volunteers. The volunteers were exposed to 100 ppm THF for 20 minutes. The absorption rate
of THF was reported to be 60%. The author suggested that the reported absorption rate
represented 80% of the steady-state absorption rate normally reached over a period of several
hours. This value is similar to reports in other human volunteer studies (Teramoto et al., 1989;
Kageyama, 1988).
Tissue distribution studies in animals also provide evidence for absorption of THF
through the respiratory tract, since measurable levels of THF were found in a variety of tissues in
rats exposed through the inhalation route (Elovaara et al., 1984; Kawata and Ito, 1984).
3.1.3. Dermal Absorption
Limited information is available on the dermal absorption of THF in either humans or
animals. Systemic toxicity observed in acute dermal toxicity studies (Stasenkova and
Kochetkova, 1963) showed that THF can be absorbed through the skin. Brooke et al. (1998)
demonstrated that uptake of vapor of industrial solvents across the skin can also occur in
humans, but the degree of dermal uptake appears to be negligible (compared to inhalation).
Under the conditions of the study in which four volunteers, two with and two without masks,
were exposed to 150 ppm THF vapor for 4 hours, dermal uptake of THF vapor (in volunteers
with masks) was found to contribute around 1-2% of the body burden received following whole-
body (including inhalation) exposure (in volunteers without masks).
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
3.2. DISTRIBUTION
No tissue distribution studies have been conducted for humans exposed to THF by any
route of exposure. However, Ong et al. (1991) reported that occupational exposures (potentially
inhalation and dermal) to THF resulted in measurable blood and urine THF levels. Kageyama
(1988) and Droz et al. (1999) reported measurable blood concentrations of THF in volunteers
exposed by the inhalation route. These results demonstrate the potential for wide tissue
distribution of THF.
Tissue distribution of THF has been studied comprehensively in rats and mice following
oral dosing (DuPont Haskell Laboratory, 1998). Single gavage doses of [14C]-THF at target
concentrations of 50 or 500 mg/kg were administered to male and female F344 rats or B6C3Fi
mice, and radioactive residues were measured in the plasma, red blood cells (RBCs), skin, whole
blood, bone marrow, brain, fat, heart, lungs, spleen, liver, kidney, GI tract and GI tract contents,
ovaries, testes, adrenals, plasma, uterus, muscle, bone, and carcass.
For rats, plasma and RBCs were collected at multiple time points, and, at 168 hours after
dosing, the animals were sacrificed and tissues were harvested for analysis of THF-associated
radioactivity. The presence of radioactivity in plasma demonstrates that THF or its metabolites
are available for systemic distribution. Comparison of kinetic data for plasma and RBCs
provides information on partitioning of THF (or its metabolites) in the blood compartment. The
Cmax values for plasma were consistently higher than Cmax values for RBCs, ranging from 2.7- to
4.8-fold among both dose groups in males and females. When the AUC data are compared for
plasma versus RBCs, the opposite relationship was observed (i.e., AUC values were higher in
RBCs than in plasma), consistent with the longer biological half-life (Ti/2) in RBCs as compared
to plasma (see Table 3-1). No data on protein binding in the plasma were available. These data
suggest that THF-associated radioactivity partitions rapidly to the plasma, resulting in higher
peak concentrations in the plasma than in RBCs.
Total recovery of the administered dose in tissues was minimal, ranging from 3.7 to
10.3% among the two dose groups in male and female rats. The highest percent recovery was in
the carcass, indicating that THF or its metabolites are widely distributed. Tissue-specific data on
a concentration basis (|ig equivalent THF/g tissue) are shown in Table 3-3. These data indicate
that the liver has the highest concentrations of radioactivity, followed by the fat and adrenal
glands. Both male and female rats had similar patterns in the tissue distribution of THF-
associated radioactivity at the two treatment doses, suggesting that at doses between 50 and
500 mg/kg, no significant shift in relative target tissue doses would be expected.
10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
Table 3-3. Radiolabel concentration in tissues of rats and mice at 168 hours
following gavage administration of [14C]-THF
Tissue
Rat
Mouse
Rat
Mouse
Male
Female
Male
Female
Male
Female
Male
Female
50 mg/kg
500 mg/kg
Tissue concentration (jig equivalent/g)
Carcass
2.0
1.5
1.4
0.9
11.9
00
00
14.2
12.4
Skin
2.4
1.6
1.5
0.9
14.7
7.4
18.1
14.6
Whole blood
1.0
0.7
0.8
0.5
6.1
5.1
8.6
5.5
Bone marrow
3.7
2.9
1.1
2.4
17.0
9.4
0.2
9.9
Brain
2.1
1.3
1.4
1.0
8.3
7.7
12.3
10.0
Fat
4.1
3.0
3.1
2.2
31.3
14.0
35.7
20.5
RBCs (terminal)
1.8
1.2
1.2
0.9
8.5
8.1
12.8
00
00
Heart
1.7
1.4
1.0
0.8
10.1
7.8
11.6
9.0
Lungs
2.1
1.4
1.1
0.6
11.9
7.9
11.6
8.6
Spleen
2.2
1.1
1.0
0.7
9.5
6.6
12.9
9.1
Liver
15.4
11.9
1.4
0.9
60.5
38.3
17.9
12.9
Kidney
2.7
2.0
1.7
1.1
15.8
12.2
22.8
14.1
GI tract
1.8
1.0
0.9
0.6
8.4
6.0
11.3
8.0
GI contents
0.5
0.2
0.2
0.1
1.3
0.9
1.5
1.2
Ovaries
-
1.4
-
1.1
-
8.4
-
13.0
Testes
1.8
-
1.4
-
7.3
-
12.5
-
Adrenals
5.4
3.9
3.0
1.4
30.2
18.5
27.1
23.5
Plasma (terminal)
0.6
0.3
0.3
0.2
3.4
2.2
3.1
4.3
Uterus
-
1.1
-
0.8
-
7.8
-
8.5
Muscle
2.0
1.7
1.3
1.0
11.5
10.3
12.5
9.7
Bone
1.8
1.2
1.2
0.6
10.6
7.5
8.3
6.3
Source: Adapted from DuPont Haskell Laboratory (1998).
Similar to rats, THF-associated radioactivity appeared rapidly in the plasma of mice after
oral exposure. Evaluation of kinetic parameters for blood compartments showed that peak
concentrations were higher, but total integrated doses (AUC) were lower in plasma compared to
RBCs. In mice, the total percent of the administered dose recovered within 168 hours after oral
dosing in these tissues ranged from 3.1 to 4.0%. The highest percent of the dose was recovered
in the carcass, indicating that THF or its metabolites were widely distributed. Tissue-specific
data on a concentration basis (jag equivalent THF/g tissue) at 168 hours are shown in Table 3-3.
Tissue distribution of THF-associated radioactivity was reported for male mice at multiple time
points until terminal sacrifice at 168 hours after dosing. In the high-dose males, peak
concentrations were reached within 4 hours after dosing for all of the tissues studied, with peak
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
concentrations notably higher in the adrenal glands, liver, and kidney. The rate of decrease in
the levels of radioactivity was tissue dependent. Most notably, at longer time points, fat had
higher levels of radioactivity than liver. At the low dose, the peak concentrations of radioactivity
in the liver and kidney, but not adrenal glands, were higher than in other tissues. As in the high-
dose group, the concentration of radioactivity in the fat of the low-dose group at 168 hours was
higher than in other tissues measured.
Hara et al. (1987) investigated the distribution of THF by giving 300 and 700 mg/kg THF
orally to male Wistar rats and rabbits (strain unspecified), respectively. Blood and tissue
samples were collected for analysis of THF concentrations from groups of three rats at 10 and
30 minutes and at 1, 2, 3, and 5 hours and from two rabbits at 7 or 8.5 hours after administration.
No significant differences were observed between the two species. Ratios of tissue levels to
blood levels were approximately 1.5-2.0 in adipose tissue and kidney and about 1.0 in the brain,
liver, spleen, and muscle.
The distribution of THF has also been studied following inhalation exposures in animals.
Elovaara et al. (1984) measured the distribution of THF into the brain and fat tissue of rats
exposed to 0, 200, 1,000, or 2,000 ppm (0, 590, 2,950, and 5,900 mg/m3) THF 6 hours/day,
5 days/week for 2-18 weeks. The exposed rats were sacrificed at 2, 8, 13, or 18 weeks, and THF
concentrations were measured in the brain and perirenal fat. At all of the time points, THF
concentrations in the fat were consistently higher than in the brain by a factor of approximately
two- to threefold. THF in both tissues increased with THF exposure concentration. As the
treatment extended from 2 to 18 weeks, the THF concentrations in both tissues gradually
decreased. The authors suggested that the decrease in tissue levels with longer exposure duration
was due to induction of the oxidative metabolism of THF, as evidenced by increases in liver and
kidney 7-ethoxycoumarin O-deethylase activity (as a marker for metabolic enzyme activity) in
THF-exposed animals beginning at 2 weeks (not duration-dependent). However, the observed
statistically significant increases in enzymatic activity appeared to reflect a decrease in the
activity in control animals rather than an increase in activity in the treated animals. No changes
in liver cytochrome P450 (CYP450) content were observed at the end of the study. Comparison
of tissue levels of THF revealed, at the highest exposure concentration, that tissue levels were
greater than the 10-fold difference in dose. This result is consistent with the greater partitioning
of THF as the parent compound into fatty tissues as discussed above for the oral dosing study in
mice.
Kawata and Ito (1984) compared the distribution of THF following several different
inhalation exposure regimens. Male Wistar rats (5/control group and 25/experimental group)
were exposed to 15,000 ppm (44,250 mg/m3) THF for a single 30-minute exposure or for seven
daily 30-minute exposures. In addition, rats were exposed to 3,000 ppm (8,850 mg/m3) THF
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
vapor for 1 hour/day, 5 days/week for 12 weeks. THF concentration was determined in tissues
immediately and 1, 3, 6, and 12 hours following the last exposure. Tissues evaluated in the study
were the brain, thymus, lung, heart, liver, kidney, spleen, and blood. For the single exposure
group, immediately after exposure, the pattern of THF distribution in organs was: blood > brain
= kidneys = heart > liver = spleen = thymus = lungs. Within 1 hour, differences among the
tissue levels began to decrease, with only the lung levels being significantly lower and blood
levels being significantly higher than the other tissues. No significant difference in THF levels
was observed among the tissues within 3 hours postexposure. The study authors suggested that
lower levels of THF in the lung reflected elimination of unmetabolized THF. Lower levels of
THF in the liver and kidney would be consistent with the metabolic capacity of these organs,
since THF was measured as the parent compound in this study. Repeated exposure to 15,000
ppm resulted in a similar pattern of tissue level, except that immediately after exposure only the
lung (significantly lower) and blood (significantly higher) levels were different from the other
tissues.
In the rats exposed to 3,000 ppm THF for 12 weeks, a different pattern of distribution
was observed. Immediately after the last exposure, THF tissue levels were greatest in the
thymus, followed by spleen > brain = heart > lung > blood > liver = kidney. The concentration
of THF in thymus was significantly higher than THF concentration in other tissues and remained
higher for up to 12 hours postexposure. Tissue levels of THF measured immediately after the
last exposure for the 1-day and the 6- or 12-week 3,000 ppm exposure regimens were compared.
THF levels were proportionally higher with increasing duration of exposure from 1 day to
6 weeks, although for many tissues, THF levels at 6 weeks were similar to those observed at
12 weeks. Daily tissue accumulation was most apparent for the thymus, in which tissue
concentrations were nearly twice as high as for the other tissues immediately after the last
exposure at 12 weeks. Beginning at 6 weeks of exposure, THF concentrations were also notably
higher in the spleen than in other tissues. Taken together, these data show that THF is taken up
in the blood and is widely distributed following exposure by the inhalation route. Longer
duration exposures may generate daily accumulation in some organs, although tissue levels
decrease to background rapidly after cessation of exposure. THF distributed preferably to the
thymus and spleen following subchronic exposures. The study authors suggested that higher
THF concentrations in the thymus after longer-term exposures might reflect increased age-
associated fattening of the thymus periphery, which seems to coincide with the normal age-
related atrophy in the parenchyma of this organ. However, the spleen was also noted as an organ
with high tissue concentrations, suggesting to the study authors (Kawata and Ito, 1984) the
possibility of THF distribution through the lymph system.
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
Pellizzari et al. (1982) reported the presence of THF in the milk from mothers who were
living in one of four urban areas in the United States. THF was found in one of eight samples
that were analyzed. This study did not provide quantitative data on the concentrations of THF
that were present or information on mothers' exposure.
No data on placental transfer of THF or fetal distribution is available in humans or in
animal studies.
3.3. METABOLISM
Several lines of evidence suggest that THF undergoes oxidative metabolism by liver
microsomal CYP450 enzymes followed by further hydrolysis catalyzed by lactonase (also
known as paraoxonasel or PON1) and additional oxidation by cytosolic dehydrogenases. Based
on the available in vivo and in vitro data, the ultimate metabolite of THF is CO2 and the
proposed metabolic pathway for this conversion is presented in Figure 3-1 (Couper and
Marinetti, 2002; DuPont Haskell Laboratory, 2000). According to this pathway, THF undergoes
oxidative metabolism to form the intermediates 5-hydroxy-THF and 4-hydroxybutanal which
may undergo further oxidation to y-butyrolactone (GBL), y-hydroxybutyric acid (GHB), and
succinaldehyde.
In vivo studies on THF metabolism indicate that CO2 is the major terminal metabolite, as
shown in Table 3-2 (DuPont Haskell Laboratory, 1998). In mice administered a single gavage
dose of 50 mg/kg 14C-THF, the percent of the radioactivity recovered as CO2 was 58.2% in
males and 74.6% in females. Volatile organics (possibly as unmetabolized THF) accounted for
17.8% of the administered dose in males and 24.5% of the administered dose in females. In mice
administered a single dose of 500 mg/kg 14C-THF, the percent of the administered dose
recovered as CO2 was 51.1 and 36.2% for males and females, respectively. Rat metabolism
studies also demonstrated that oxidative metabolism of THF to CO2 is an important pathway. In
rats given a single gavage dose of 50 mg/kg of 14C-THF, 47.8 and 47.5% of 14C-THF in males
and females, respectively, was recovered in the form of CO2. In rats given 500 mg/kg of
radiolabeled THF, these percentages were 21.9% in males and 18.8% in females.
In both sexes of mice and rats, metabolism of THF to CO2 was greater at the low dose,
suggesting that metabolism may be saturated at higher doses. Although the data suggest that
there might be species differences in the contribution of CO2 to THF metabolism, potential
saturation of the CO2 trap and therefore loss of CO2 in the rat study make comparison of the rat
and mice data unreliable.
The metabolism of GBL and GHB has also been studied extensively (NSF, 2003). GBL
may readily convert to GHB, as lactones are known to readily equilibrate in aqueous media
between their closed (lactone) and open (hydroxyl acid) forms, a process that may be influenced
14
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
by pH and structural features of the specific lactone (Teiber et al., 2003; Roth and Giarman,
1966). Hydrolysis of lactones to the corresponding organic acids as well as the reverse reaction,
namely formation of lactones from hydroxy acids, have recently been shown to be catalyzed by
liver and serum enzymes known as paraoxonases (PON) (Draganov et al., 2005; Teiber et al.,
2003; Billecke et al., 2000). In these studies, several lactone and hydroxy acid substrates,
including GBL and GHB, were converted to the corresponding hydroxy acids and lactones by a
specific human serum PON isoenzyme (PON1).
[O]
-OH
O
o
THF
5-hydroxyTHF
A
[O]
Os,
o °
GBL
[O]
hydrolysis
HO—' ^=0
,	- 4-hydroxybutanal
reduction/ ' '
reduction
HO ' x OH
1,4-butanediol
HO
hydrolysis
-JO
OH
GHB
oxidation	reduction
oxidation
[O] = oxidation
o=/ \=o
succinaldehyde
o= y=o
HO
succinic semialdehyde
O
HO
OH
O
succinic acid
CO,
Figure 3-1. Possible metabolic pathways of THF.
Source: Modified from Couper and Marinetti (2002) and DuPont Haskell Laboratory (2000).
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
It is well established that, in the absence of exposure to THF, normal brain and peripheral
tissues from several mammalian species, including humans, have built in metabolic machinery to
produce and process GHB. High concentrations of GHB have been found in normal brain and in
peripheral tissues including brown fat, liver, heart, spleen, and kidneys from human and other
species where endogenous formation of brain GHB is thought to come from the neurotransmitter
y-aminobutyric acid (GABA) and possibly 1,4-butanediol (Nelson et al., 1981; Doherty et al.,
1978; Roth and Giarman, 1968). More recently, a GHB receptor from a human brain frontal
cortex cDNA library has also been cloned and characterized (Andriamampandry et al., 2007).
GHB can be oxidized to succinic semialdehyde (SSA) by a cytosolic NADP+ dependent
GHB dehydrogenase commonly found in brain as well as several other tissues including brown
fat, liver, heart, spleen, and kidneys (Kaufman and Nelson, 1987; Kaufman et al., 1979). An
enzyme known as succinic semialdehyde dehydrogenase then oxidizes SSA to succinic acid
(Kaufman and Nelson, 1987; Gibson et al., 1983) which is an intermediate in the citric acid cycle
that ultimately generates CO2, water, and usable energy. As discussed earlier, the in vivo
metabolism studies of THF have shown that C02 is the predominant metabolite.
In an in vitro experiment with hepatic microsomal preparations from rats, mice, or
humans, the only metabolite of THF identified was y-hydroxybutyric acid (GHB) (DuPont
Haskell Laboratory, 2000). The T 1/2 for disappearance of THF in these reactions was 40 hours
for rat microsomes, 28 hours for human microsomes, and 9 hours for mouse microsomes. The
data suggest that liver microsomes in mice may have a greater capacity to metabolize THF than
do human or rat microsomes. No data are available to confirm whether these relative rates of
metabolism by microsomes are predictive of THF metabolism among species in vivo. Further,
though no attempt was made to characterize the role of specific metabolizing enzymes, the fact
that microsomes were used, in the presence of an NADPH-generating system (DuPont Haskell
Laboratory, 2000), strongly suggests that one or more of the CYP450 isoenzymes were involved.
The metabolism of THF to GBL is further supported by metabolic studies of p-dioxane, a
structural analogue of THF. p-Dioxane-2-one, a lactone with a six-member ring analogous to
GBL, has actually been identified as the major urinary metabolite of p-dioxane in rats (Woo et
al., 1977). In addition, in vitro studies of structurally related compounds with a THF ring or
similar ring structures indicate that there are a number of possible pathways (see Figure 3-1) for
the metabolism of THF to GHB, including (1) a-hydroxylation (by microsomal CYP450
enzymes) to 5-hydroxy-THF, which can be rapidly converted to GBL and GHB (Woo et al.,
1977; Fujita and Suzuoki, 1973); (2) oxidation of THF (by cytosolic enzymes) to
4-hydroxybutanal, followed by immediate oxidation to GHB and GBL or reversibly reduced to
1,4-butanediol (El Sayed and Sadee, 1983; Roth and Giarman, 1968); and (3) direct oxidation of
THF to succinaldehyde (by microsomal CYP450 enzymes)—not shown in Figure 3-1, followed
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
by reversible reduction to 4-hydroxybutanal and oxidation to GBL or GHB in the presence of
cytosolic soluble enzymes. The formation of GBL or GHB from succinaldehyde by soluble
enzymes could also occur by oxidation to SSA, followed by reversible reduction (El Sayed and
Sadee, 1983).
The implication of these metabolic intermediates to the overall toxicity of THF is unclear.
Many of these intermediates (i.e., 5-hydroxy-THF, 4-hydroxybutanal, 1,4-butanediol,
succinaldehyde) are expected to be unstable and rapidly undergo further metabolism to GHB.
Studies in rats have shown that 1,4-butanediol is metabolized in the blood and brain to GHB and
that GHB is the active intermediate responsible for the central nervous system (CNS) effects of
1,4-butanediol (Roth and Giarman, 1968). In fact, in vitro and in vivo studies have shown that
GHB can be converted to the neurotransmitter, GABA (Vayer et al., 1985; DeFeudis and Collier,
1970), which provides a possible mechanistic link between THF and its potential for causing
CNS effects. Appreciable amounts of radioactive-labeled GABA were detected in the brains of
mice 60, 120, and 180 minutes after intraperitoneal (i.p.) injection of 1-14C-GHB (DeFeudis and
Collier, 1970). Increased tissue level of GABA and putrescine (the primary source of GABA in
many tissues) may also be hypothesized to play a role in the THF-induced cell proliferation and
carcinogenicity in the liver (see Section 4.7.3.2).
3.4. ELIMINATION
The available human data suggest that expiration is an important route of excretion for
THF. In a human occupational study (Ong et al., 1991), workers exposed to THF by the
inhalation and dermal routes excreted THF in exhaled air and in the urine. Kageyama (1988)
measured exhaled air concentrations of THF in volunteers exposed by the inhalation route. THF
was present in the exhaled air for several hours after exposure to a concentration of 200 ppm,
suggesting that THF is excreted in exhaled air. Droz et al. (1999) summarized the results from
several additional human volunteer studies that support the conclusion that THF is rapidly
excreted from the body via exhaled air and urine. Exposure periods were for as long as 8 hours
to concentrations as high as 200 ppm. In all cases, THF levels in breath, blood, or urine declined
rapidly and reached background levels within a period of approximately 12 hours.
Oral dosing studies in animals provide further evidence for the important role that
exhaled air plays as a route of excretion for THF. In rats exposed to an oral dose of 50 mg/kg
THF, 47% of the oral dose was recovered in the expired air as CO2, while only about 4% of the
radioactivity was detected in the urine and 1% in the feces. In the mice exposed to the same dose
of THF, 58-75% of the oral dose was recovered in expired air as C02 and 18-25% as volatile
organic compounds (VOCs), while 3-5% of the radioactivity was detected in the urine and 1%
was detected in the feces. A similar pattern was observed in the animals exposed to the high
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
dose of 500 mg/kg, but relatively less radioactivity, 19-22% as CO2 in the rats and 36-51% as
CO2 in mice, was recovered in the expired air. Because of some technical difficulties in recovery
of VOCs from the expired air, significant losses of trapped VOCs occurred in most of the
measurements. Among all the data available for VOCs, the only reliable data were from the
mice exposed to the low dose of THF. Nevertheless, the available data indicate that expiration
was the major route of excretion of absorbed THF, and CO2 was the major final product. The
study authors suggested that the VOCs in the exhaled air were likely to be parent THF. Urine
and feces were relatively minor routes of THF excretion (DuPont Haskell Laboratory, 1998).
In the same study (DuPont Haskell Laboratory, 1998), the time course of THF in the
plasma of exposed rats and mice was also studied. The results are summarized in Table 3-1. In
the rats exposed to the low dose (50 mg/kg), the Ty2 of the radioactivity in the plasma was
52 hours in the males and 51 hours in the females. Following exposure to the high dose
(500 mg/kg) THF, the plasma T1/2 was estimated to be 48 and 59 hours, respectively. In the mice
exposed to the low dose, the plasma T1/2 was 57 hours in the males and 51 hours in the females.
Following exposure to the high dose (500 mg/kg) THF, the serum Ty2 was 57 and 99 hours,
respectively. Based on these data, there were no apparent differences in the plasma Ti/2 between
rats and mice. At the 50 mg/kg dose level, male and female animals had a comparable Ti/2,
while at 500 mg/kg THF the males had shorter plasma half-lives than the females. The half-lives
reported in this study are not the biological half-lives of THF but only represent radioactivity
measured in plasma and serum. The radioactivity present is likely derivatives of THF that are
either covalently bound to cellular macromolecules or have been incorporated into the primary
carbon pool. Available data indicate that the biological Ty2 of THF is about 5-7 hours. Hara et
al. (1987) reported a Ty2 of 5.2 hours in rats, following oral administration of 300 mg/kg, and a
T1/2 of 5.1 hours in rabbits at a dose of 700 mg/kg.
The AUCs for the THF-associated radioactivity in the plasma were estimated for the
exposed rats and mice in the study conducted by the DuPont Haskell Laboratory (1998). In the
rats exposed to 50 mg/kg THF, the plasma AUC in males and females was 536 and 320 [j,g THF
equivalents-hour/g plasma, respectively. In rats exposed to 500 mg/kg THF, the plasma AUC in
males and females was 2,826 and 1,998 [j,g THF equivalents-hour/g plasma, respectively (see
Table 3-1). At either the low or high doses, the AUC was always higher in the male rats than in
female rats. A similar gender difference was observed in mice. In the 50 mg/kg dose group, the
plasma AUC was 207 and 157 (j,g THF equivalents-hour/g plasma. The plasma AUC in males
and females was 3,238 and 1,904 in the high-dose group (500 mg/kg), respectively. Based on
these findings, the same oral dose of THF results in a higher internal dose of THF and/or its
metabolites in male rats or mice than in females of the corresponding species. However, the
toxicological implications of this result are difficult to interpret since the AUC reflects 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
combination of THF and its metabolites, while the toxic moiety has not been clearly identified.
Nevertheless, in general, the greater AUC for males would be consistent with a greater degree of
systemic dose in males versus females.
The AUC data from this study can be used to estimate the body clearance of THF. The
clearance was calculated based on the ratio of administered dose/AUC. All the relevant kinetic
parameters and estimated clearance values are summarized in Table 3-1. In both rats and mice,
females had a higher clearance rate than males. The more rapid clearance (i.e., due to lower
AUC values) observed in females might reflect differences in excretion kinetics or alternatively
might reflect differences in the degree of THF absorption, since the administered dose was used
for this calculation rather than the absorbed dose. The clearance rates in the rats of the low-dose
group were lower than the high-dose group, while there were no such differences in the mice.
Kawata and Ito (1984) compared the blood and tissue distribution and elimination of
THF, following several different inhalation exposure regimens. In male Wistar rats exposed to
15,000 ppm (44,250 mg/m3) THF for a single 30-minute exposure, 70-80% of the THF was
eliminated from the organs within 1 hour following exposure. After 1 hour, concentration of
THF decreased slowly and was almost completely eliminated by 12-13 hours following
exposure. In animals that received seven exposures of 15,000 ppm, only 18-39% of THF was
eliminated from the organs in 1 hour following exposure, indicating some saturability in the
elimination kinetics for these organs at very high concentrations. In these animals, the rate of
THF decrease was 31% at 3 hours following last exposure and 68% at 6 hours following last
exposure; by 12 hours THF was almost completely eliminated. Similar to the acute dosing
studies, THF was nearly completely eliminated from blood and tissues within 12 hours after the
last exposure in the 12-week exposure protocol. These data indicate that, for exposure
concentrations as high as 15,000 ppm, THF is rapidly eliminated from blood and other tissues.
3.5. BIO AC CUMULATION
Two pharmacokinetic studies employed longer-term exposure regimens that provide
information useful for assessing the potential for bioaccumulation of THF in tissues. Kawata
and Ito (1984) measured tissue levels of THF immediately after the last exposure period
following daily inhalation exposures to 3,000 ppm THF for 1 day, 6 weeks, or 12 weeks. Daily
levels increased in some tissues, particularly from 1 day to 6 weeks. In the thymus and spleen,
tissue levels continued to increase through the 12-week exposure period. These data suggest
some potential for tissue accumulation with repeated daily exposure. However, it is notable that
even in animals exposed for 12 weeks, tissue levels declined rapidly after the end of the last
exposure period (within hours). These data suggest that the rate of uptake of THF is more rapid
than the rate of excretion. Therefore, during periods of continuous exposure, there is some
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
potential for tissue levels of THF to accumulate. However, periods of intermittent exposure
would allow for clearance of the THF body burden and thus limit the potential bioaccumulation.
Elovaara et al. (1984) measured the distribution of THF into the brain and fat tissue of
rats exposed to 0, 200, 1,000, or 2,000 ppm (0, 590, 2,950, and 5,900 mg/m3) THF 6 hours/day,
5 days/week for 2-18 weeks. As the treatment extended from 2 to 18 weeks, the THF
concentrations in both tissues of the exposed rats gradually decreased. The observed decline in
brain and fat THF levels suggests that THF may not bioaccumulate in these tissues.
Evaluation of human volunteer studies to derive a physiologically based pharmacokinetic
model for THF revealed rapid elimination of THF from the body (Droz et al., 1999). The
resulting model predicted that no significant accumulation of THF would be expected over the
workweek or across workweeks. THF elimination rates observed in inhalation (Elovaara et al.,
1984; Kawata and Ito, 1984) and oral studies (DuPont Haskell Laboratory, 1998) in animals
support this conclusion. Taken together, the data support the general conclusion that THF is not
likely to bioaccumulate.
3.6. PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS
A human physiologically based pharmacokinetic (PBPK) model has been developed by
Droz et al. (1999) to estimate THF concentrations in the blood, breath, and urine, following an
inhalation exposure for the purpose of determining biological exposure indices in these media
that would equate to an occupational exposure level of 200 ppm THF. The PBPK model was
constructed with seven compartments: lungs, muscles and skin, fatty tissue, liver, kidneys, brain,
and other tissues. Physiological parameters (tissue volumes, blood flow rates, etc.) were
calculated from body weight and height and from physical workload by using formulas
previously developed by the author (Droz et al., 1989). Blood-air and tissue-air partition
coefficients were estimated from in vitro experiments. THF metabolism was assumed to follow
first order kinetics. Urinary excretions were calculated assuming a urine flow of 1 mL/minute
and a creatinine excretion rate of 1.4 g/day. The model was validated by using four discrete sets
of human exposure data from workers or human volunteer studies. The model provided an
adequate fit to the data from three out of four sets of data. The reason for the lack of fit for one
of these data sets was not determined. Based on the model predictions, repeated inhalation
exposures to 200 ppm THF would yield end-of-the-work-shift levels of THF in biological
samples of 5.1 ppm in breath, 57 [j,mol/L in the blood, and 100 [j,mol/L in the urine. However,
this model does not account for the pharmacokinetic and pharmacodynamic variability in
humans and no PBPK models have been developed in animals. Also, there are no comparative
pharmacokinetic or pharmacodynamic studies following exposure to THF by the oral route in
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
humans and animals. Therefore, this model is not adequate for calculating human equivalent
exposure concentrations from the available rodent study data.
3.7. SUMMARY
Overall, the available data demonstrate that THF is readily absorbed through multiple
routes, is systemically distributed, and is rapidly metabolized and excreted.
THF is readily absorbed from the respiratory tract, based on the observed rapid increase
of THF in biological samples or calculated uptake rates in human studies (Droz et al., 1999; Ong
et al., 1991; Kageyama, 1988; Wagner, 1974). Although no human data are available to evaluate
the rate or degree of absorption of THF following exposure through the oral route, oral dosing
studies in rats and mice show that radiolabeled THF is readily absorbed from the GI tract with
wide tissue distribution; however, total recovery of radioactivity in tissues represented only a
small fraction of the administered dose (DuPont Haskell Laboratory, 1998). No studies on
dermal absorption were identified, but the observed systemic toxicity in a dermal toxicity study
in mice and rabbits (Stasenkova and Kochetkova, 1963) demonstrated that THF can be absorbed
through the skin.
A metabolic pathway has been proposed in which THF is oxidatively metabolized to
succinic acid, which being an intermediate in the citric acid cycle, undergoes a series of reactions
ultimately leading to the release of C02 from the parent molecule. In addition, several
intermediate metabolites are expected to be unstable and rapidly undergo further metabolism to
GHB which can be converted to the neurotransmitter GAB A. Several enzymes, including
CYP450, PON1, and dehydrogenases, may be involved in metabolizing THF and some of its
intermediate metabolites (see Section 3.3 and Figure 3.1).
The available human data suggest that THF is rapidly excreted. Excretion in exhaled air
and urine were correlated with exposure concentration in an occupational study (Ong et al.,
1991). Human volunteer studies demonstrate that THF is rapidly excreted in exhaled air and
urine, with concentrations of THF in these tissues generally returning to background levels
within hours of cessation of exposure (Droz et al., 1999; Kageyama, 1988). The rapid excretion
of THF observed in human studies is supported by an inhalation study in rats (Kawata and Ito,
1984) in which tissue levels of THF decline rapidly during the postexposure period. THF is also
rapidly cleared from the body following oral dosing, with exhaled air serving as the primary
route of excretion (DuPont Haskell Laboratory, 1998). Analysis of the mass balance of
radioactivity in the exhaled air, excreta, and tissues showed that nearly the entire administered
dose was excreted in the exhaled air as C02 or volatile organics (possibly unmetabolized THF).
The rate of excretion was rapid. The half4ives in the plasma were approximately 50 hours for
most groups, although blood and tissue levels of radioactivity decreased rapidly, and tissue levels
21
DRAFT - DO NOT CITE OR QUOTE
-------
1	of radioactivity represented only a small percentage of the administered dose within 168 hours of
2	exposure. Available data indicate that the biological Ti/2 of THF is about 5-7 hours (Hara et al.,
3	1987).
4
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
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS
There are a number of human occupational exposure studies and case reports on humans
exposed to THF. These human studies identify effects on the nervous system and liver. Most of
these studies do not identify THF exposure levels. Also, all of the human studies report
coexposures to other chemicals, including solvents that are neurotoxic.
Gamier et al. (1989) reported two cases of occupational exposure to THF. In both cases,
the men (ages 35 and 55) worked as plumbers repairing pipes in confined spaces with a glue that
contained THF. No exposure information was provided. Symptoms included nausea, headache,
dizziness, chest pain, cough, dyspnea, and epigastric pain. In both men, blood count and renal
function were normal. However, the serum liver enzymes aspartate aminotransferase (AST),
alanine aminotransferase (ALT), and y-glutamyl transferase (GGT) were elevated several times
above the normal range. Clinical symptoms resolved in about 2 days and liver enzymes returned
to normal within 2 weeks. The authors suggested that THF exposure may result in irritation,
CNS effects, and transient liver toxicity in humans.
Emmett (1976) reported the case of a 41-year-old pipe fitter exposed for about 3 months
to a mixture of THF and other solvents in a pipe cleaning solution and a pipe glue. Other
solvents present in the solution included acetone and cyclohexanone. No information was
provided on exposure concentrations. The only effects reported by the patient were a slight
rhinorrhea (runny nose) during exposure and a gradual onset, over 10 weeks, of a constant
unpleasant smell or loss of sense of smell. No other clinical signs were reported. A neurological
exam, radiography of skull and sinuses, and hematological exam were all normal. Within 6
weeks after cessation of exposure, some sense of smell returned. However, by 7 months after the
initial diagnosis, sense of smell was still diminished.
Edling (1982) reported the occupational exposure of a shoemaker to a mixture of solvents
that included THF, acetone, chloroform, and trichloroethylene. No information on exposure
concentrations was provided. In addition, the patient had concurrent exposure to acetylsalicylic
acid to treat lumbago (back pain). Clinical chemistry results revealed increased liver enzymes
including GGT and ALT. Liver biopsy showed centriacinar fatty change and siderosis.
Juntunen et al. (1984) reported cerebral convulsions in a patient following occupational
exposure to both THF and enfluran anesthesia. The patient was a 45-year-old man who worked
as a plumber, using a solvent containing THF to insulate the inside of a water piping system. For
2 weeks, the patient had been working with THF in enclosed spaces with no ventilation. No
information was provided on the resulting exposure concentration. The patient reported that he
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
33
34
35
had felt unusually tired and had a headache in the week before he was admitted to the hospital
for an appendectomy. On awakening from the enfluran anesthesia, the patient had several
convulsions. In addition, liver enzymes were slightly elevated following the surgery. The
authors concluded that THF exposure was the main contributing factor for the convulsions
because the patient was exposed to high concentrations of THF for 2 weeks before the surgery.
In addition he had never had epilepsy or neurological disease and his clinical status and
computed tomography results were normal.
Albrecht et al. (1987) reported a case of autoimmune glomerulonephritis in a plumber
working with pipe cement containing THF. The 28-year-old male plumber had been working
with pipe cement for over 9 years. The initial symptom was gross hematuria. A needle biopsy
of the kidney revealed segmental proliferative glomerulonephritis with immunoglobulin A
deposits, capillary adhesions to the Bowman's capsule, and fibrin in the glomerular mesangial
deposits. Industrial hygiene monitoring identified 15-minute exposures to THF, ranging from
389-757 ppm (1,148-2,233 mg/m3) during periods that pipe cement was in use.
The National Institute for Occupational Safety and Health (NIOSH) (1991) investigated
reports of adverse health effects at a plant that manufactured flexible hose. Environmental
monitoring was conducted for respirable particulates, respirable silica, THF, total dust, metals,
nitrosamines, and other organic compounds. Approximately 35-40 employees were interviewed
by NIOSH investigators. In addition, the medical records of six employees who had sought
medical attention for a work-related health problem and the death certificates of nine employees
who were thought to have had work-related health problems were reviewed by NIOSH
investigators. THF was detected in five air samples collected during a sealing operation. The
concentrations ranged from 20 to 83 ppm (59-245 mg/m3), but none of the sampling results were
above the Occupational Safety and Health Administration standard of 200 ppm. However, the
backup sections on the sampling apparatus also contained THF, indicating that breakthrough had
occurred and suggesting that the THF exposure concentrations may have been higher. In
addition to THF, other organic solvents detected in the air monitoring samples included acetone,
toluene, methyl ethyl ketone, and 1,1,1-trichloroethane. The interviewed employees reported a
variety of symptoms, including eye and respiratory tract irritation, headaches, lightheadedness,
and drowsiness. The authors suggested that these symptoms may be related to solvent exposure
but could not associate specific symptoms with individual chemicals.
Horiuchi et al. (1967) evaluated the health of workers employed in a vinyl chloride hose-
manufacturing facility where THF was used as an adhesive. THF was detected in workplace air
samples at concentrations as high as 1,000 ppm (2,950 mg/m3). Workers who handled THF
reported fatigue in the lower extremities. Clinical findings included decreased specific gravity of
24
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
whole blood (more predominant in females), decreased white blood cell count, increased serum
ALT activity, palpable liver, and hypotension.
Two human dermal THF exposure studies were identified. A study by BASF (1938) did
not observe contact dermatitis or sensitization in dermal tests in 196 volunteers exposed to THF
(exposure concentration not reported by study authors). Hofmann and Oettel (1954) reported
that THF applied to the skin of six people produced irritation that was more severe when THF
was allowed to evaporate. The authors concluded that THF itself was nonirritating, and the
irritation was caused by impurities that remained after THF had evaporated away. No additional
information was provided to evaluate the adequacy of this study.
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
4.2.1. Subchronic Studies
4.2.1.1.	Oral Studies
No subchronic studies in animals by the oral route of exposure were identified.
4.2.1.2.	Inhalation Studies
Horiguchi et al. (1984) evaluated the subchronic inhalation toxicity of THF in rats. Male
Sprague-Dawley rats (11-12/group) were exposed to THF vapors 5 days/week, 4 hours/day for
12 weeks. Two experiments using different concentrations were conducted. THF concentrations
for the first experiment were 0, 200, or 1,000 ppm (0, 590, or 2,950 mg/m3) and for the second
experiment were 0, 100, or 5,000 ppm (0, 295, or 14,750 mg/m3). Body weights and clinical
signs of intoxication were observed daily during the exposure period. Rats were sacrificed on
the second day following termination of exposure. Blood was drawn for hematological and
serum chemistry evaluation. Major organs were weighed and evaluated histopathologically.
Body weight in rats exposed to 5,000 ppm was significantly lower than controls for the entire
exposure period; no differences from controls were observed in the other treated groups.
Animals in the 5,000 ppm group displayed signs of local irritation and CNS effects, which were
described by the study authors as similar to those observed for the acute study (Horiguchi et al.,
1984). These local irritation and CNS effects were reported as moderating with continued
exposure. Serum AST was statistically significantly increased above controls (by 18-50%) at
exposures >200 ppm; however, the magnitude of the increase was minimal and was not
dependent on the exposure level (the highest increase was 50% greater than controls at 1,000
ppm while at 5,000 ppm it only increased by 18%). Compared to the control values, the
following parameters were also changed in the 1,000 and/or 5,000 ppm exposure groups. At
1,000 and 5,000 ppm, cholinesterase was slightly but statistically significantly increased by 8 and
25
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
15%, respectively, while blood sugar was statistically significantly decreased by 20 and 39%,
respectively. Serum ALT, cholesterol, and bilirubin were statistically significantly increased
only in the 5,000 ppm group (by 100, 44, and 46%, respectively). White blood cell count was
significantly decreased (by about 24%) in the 5,000 ppm group compared with controls.
Relative organ weights were significantly increased (by 7-28%) only in the 5,000 ppm group,
including brain, lung, liver, pancreas, and kidney, while the relative spleen weight was decreased
(by 13%). All histopathological findings were comparable between treated and control groups.
Based on body weight, organ weight changes, local irritation and CNS effects, and serum
chemistry parameter changes, EPA identified 5,000 ppm (14,750 mg/m3) as the study lowest-
observed-adverse-effect level (LOAEL) and the no-observed-adverse-effect level (NOAEL) as
1,000 ppm (2,950 mg/m3). The results of Horiguchi et al. (1984) were also reported in an earlier
Japanese publication from the same laboratory (Katahira et al., 1982).
In an NTP subchronic inhalation study (NTP, 1998; Chhabra et al., 1990), F344/N rats
and B6C3Fi mice (10/sex/group) were exposed to target concentrations of 0, 66, 200, 600, 1,800,
or 5,000 ppm THF vapor (0, 195, 590, 1,770, 5,310, or 14,750 mg/m3) 6 hours/day, 5 days/week
for 90 days. Animals were observed for morbidity and mortality, body weight, and clinical
observations. Within 24 hours after last exposure, animals were euthanized, and blood and
tissues were collected. All major tissues were fixed in formalin and processed.
Histopathological examination was performed on all tissues from the high-dose group and
controls and on all gross lesions and target tissues from all dose groups. Organ weights were
measured for heart, liver, lung, right kidney, spleen, and thymus. Standard hematology and
clinical parameters were evaluated in rats only. Thymus and liver weights and relative weights
are summarized in Table 4-1.
26
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-1. Changes in absolute and relative thymus and liver weights of
F344/N rats and B6C3Fi mice following subchronic inhalation exposure to
THFa

Concentration (ppm)
0
66
200
600
1,800
5,000
Male rats
Body weight (g)
361 ±6
353 ±7
368 ± 11
364 ±6
372 ±9
343 ±7
Thymus weight (g)
Relative weight (mg/g)
0.36 ±0.02
1.00 ±0.03
0.35 ±0.01
1.00 ±0.03
0.33 ±0.01
0.92 ±0.015
0.35 ±0.01
0.95 ±0.04
0.33 ±0.02
0.88 ± 0.03b
0.28 ± 0.02°
0.81 ± 0.04°
Liver weight (g)
Relative weight (mg/g)
12.65 ±0.65
34.92 ± 1.34
11.50 ±0.49
32.53 ±0.85
12.46 ±0.41
33.84 ±0.29
12.40 ±0.42
34.04 ±0.79
12.91 ±0.38
34.72 ±0.45
12.80 ±0.31
37.28 ±0.57b
Female rats
Body weight (g)
205 ±5
207 ±5
205 ±4
210 ±3
209 ±4
214 ±3
Thymus weight (g)
Relative weight (mg/g)
0.27 ±0.01
1.29 ±0.04
0.26 ±0.02
1.26 ±0.06
0.26 ±0.01
1.26 ±0.04
0.25 ±0.02
1.17 ±0.06
0.26 ±0.01
1.25 ±0.04
0.21 ±0.01c
0.99 ± 0.03°
Liver weight (g)
Relative weight (mg/g)
6.62 ±0.13
32.36 ±0.81
6.43 ±0.17
31.05 ±0.69
6.32 ±0.19
30.76 ±0.59
6.63 ±0.22
31.52 ± 1.08
6.71 ±0.19
32.02 ±0.54
7.78 ±0.17°
36.41 ±0.87°
Male mice
Body weight (g)
36.7 ±0.8
36.9± 0.4
35.8 ±0.7
36.3 ±0.7
36.6 ±0.8
32.7 ± 1.0°
Thymus weight (g)
Relative weight (mg/g)
0.047 ± 0.003
1.27 ±0.06
0.045 ± 0.003
1.23 ±0.08
0.042 ± 0.002
1.17 ±0.05
0.039 ±0.001b
1.08 ± 0.05b
0.036 ±0.003c
0.99 ± 0.07°
0.027 ± 0.002°
0.81 ± 0.05°
Liver weight (g)
Relative weight (mg/g)
1.613 ±0.037
44.00 ±0.57
1.667 ±0.022
45.24 ±0.27
1.695 ±0.037
47.28 ±0.37°
1.722 ±0.031b
47.52 ±0.60°
1.789 ±0.035°
48.94 ±0.81°
1.964 ±0.060°
60.03 ±0.33°
Female mice
Body weight (g)
32.4 ± 1.0
32.2 ±0.6
33.3 ± 1.1
32.5 ±0.7
33.1 ± 1.1
33.3 ± 1.1
Thymus weight (g)
Relative weight (mg/g)
0.051 ±0.003
1.57 ±0.09
0.055 ±0.003
1.71 ±0.08
0.056 ± 0.002
1.71 ±0.10
0.053 ± 0.002
1.64 ±0.06
0.052 ±0.003
1.59 ± 0.11
0.046 ± 0.003
1.36 ±0.08
Liver weight (g)
Relative weight (mg/g)
1.592 ±0.036
49.38 ±0.94
1.574 ±0.035
48.95 ±0.92
1.609 ±0.034
48.66 ± 1.30
1.551 ±0.034
47.79 ±0.60
1.733 ±0.045b
52.51 ± 1.22b
1.814 ±0.074°
54.42 ± 0.96°
1
2
3
4
5
27	DRAFT - DO NOT CITE OR QUOTE
Organ weights and relative organ weights are, respectively, in g and mg organ weight/g BW (mean ± standard
error). All group sizes are 10 animals/group except for male mice in the 5,000 ppm group where N = 7.
hp< 0.05.
><0.01.
Source: Adapted from NTP (1998).
In F344/N rats, body weight and survival were not affected by THF exposure.
Immediately after exposure, clinical signs of ataxia, described as irregular movement with lack
of coordination, were observed in both male and female rats at 5,000 ppm only. In male and
female rats at 5,000 ppm, absolute and relative thymus (Table 4-1) and spleen weights were
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
statistically significantly decreased. In the 5,000 ppm exposure group, there were statistically
significant increases in absolute and relative liver weights of female rats (by 17 and 13%,
respectively) and in relative weights of male rat liver (by 7%), kidney (by 8%) and lung (by
15%). Several hematological parameters in both male and female rats were significantly
increased at 5,000 ppm, including RBC counts, hemoglobin, volume of packed red cells, mean
corpuscular volume, mean corpuscular hemoglobin (males only), segmented neutrophil count
(males only), and platelet counts (females only). In the 5,000 ppm exposure group, male and
female rats had increased levels of serum bile acids (by 70 and 80%, respectively) but the
increase was statistically significant only in females; blood urea nitrogen and creatinine were
also significantly decreased (by about 20%) in females. In the absence of cholestatic injury or
hepatocellular necrosis (both alkaline phosphatase and ALT were normal) the change in bile
acids was considered consistent with decreased or altered hepatocellular function (NTP, 1998).
The only histopathological lesions observed in rats occurred in the forestomach at 5,000 ppm.
Acanthosis (increased thickness) was found in 5/10 males and 8/10 females, and suppurative
inflammation of the forestomach was found in 2/10 males and 4/10 females. However, the
authors concluded that forestomach lesions were minimal inflammatory changes resulting from
direct contact of THF ingested during the exposure period, rather than a systemic effect of
inhaled THF. Based on observation of clinical signs, changes in organ weights, hematological
effects, and clinical chemistry findings, EPA identified a concentration of 5,000 ppm
(14,750 mg/m3) as a LOAEL and 1,800 ppm (5,310 mg/m3) as aNOAEL in F344/N rats.
In B6C3Fi mice, body weights were similar across groups, except for an 11% decrease in
high dose males. Survival in female mice was not affected by THF exposure for 14 weeks, while
three high-dose males died in weeks 2, 4, or 8 (NTP, 1998). Two male deaths were attributed to
suppurative pyelonephritis, while the third (in week 4) was not explained. Male and female mice
at both 1,800 and 5,000 ppm showed clinical signs of CNS toxicity characterized as narcosis
during exposure. At 5,000 ppm, mice were in a stupor for 2 hours following the exposure
period; at 1,800 ppm, mice were fully awake when chamber doors were opened following
exposure. However, no incidence data were reported for CNS effects. In male mice,
concentration-related trends included increasing relative liver weight starting at concentrations of
200 ppm (7.5% above control, p<0.05) and both absolute and relative liver weights were
statistically significantly increased by 7-36% at concentrations of >600 ppm (Table 4-1). In
addition, absolute and relative thymus weights were dose-dependently decreased (by 15-36%) at
concentrations of >600 ppm. Absolute and relative spleen weights were significantly decreased
(by 31-38%) at 5,000 ppm only (not shown). In female mice, absolute and relative liver weights
were statistically significantly increased (by 6-14%) at 1,800 and 5,000 ppm. Absolute and
relative weights of spleen, lung, and heart were all significantly decreased at 5,000 ppm (not
28
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
shown). Histopathological lesions in mice were observed in liver, uterus, and adrenal gland
(Table 4-2).
Table 4-2. Incidences of selected nonneoplastic lesions in B6C3Fi mice
following subchronic inhalation exposure to THFa

Concentration (ppm)
0
66
200
600
1,800
5,000
Male mice
Liver
Cytomegaly,
Centrilobular
0
_b
—
—
1 (1.0)°
7" (2.0)
Female mice
Adrenal Cortex
Degeneration,
X-zone
0
—
—
—
0
10" (2.0)
Liver
Cytomegaly,
Centrilobular
0
—
—
—
0
10" (1.0)
Uterus
Atrophy
0
-
-
-
0
10" (2.0)
aAll examined group sizes are 10 animals/group.
bTissue not examined.
cAverage severity grade of lesions in affected animals: l=minimal, 2=mild, 3=moderate, 4=marked.
" Significantly different (p < 0.01) from the control group by the Fisher exact test.
Source: Adapted from NTP (1998).
Liver centrilobular cytomegaly was observed in 7/10 male mice (graded mild) and
10/10 female mice (graded minimal) at 5,000 ppm (statistically significant) and 1/10 male mice
(graded minimal) at 1,800 ppm. In addition, 10/10 female mice at 5,000 ppm demonstrated
uterine atrophy and degenerative changes of the adrenal cortex. EPA identified the LOAEL for
this study as 1,800 ppm (5,310 mg/m3) based on statistically significant liver effects and clinical
signs of toxicity (narcosis); the NOAEL is 600 ppm (1,770 mg/m3).
BASF (Gamer et al., 2002; BASF, 2001a) evaluated a series of endpoints in male F344
rats (6/group plus 5/group) and female B6C3Fi mice (10/group plus 5) in tissues for which THF-
treated animals developed tumors in the NTP cancer bioassay (NTP, 1998). Animals were
placed in one of three groups that were exposed 6 hours/day for either 5 consecutive days,
5 consecutive days followed by a 21-day observation period, or 20 consecutive days over a
period of approximately 28 days. Test animals were exposed nose only to average THF
29
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
concentrations of 0, 598, 1,811, or 5,382 mg/m3 (0, 199, 604, or 1,794 ppm), corresponding to
the concentrations used in the NTP (1998) cancer bioassay. Concentrations adjusted for
continuous exposure were 0, 107, 323, or 961 mg/m3. For the animals in each of the four
concentration groups, a full necropsy was done, including histopathological evaluation of the
kidney (rat), liver (mouse), and uterus (mouse). No clinical effects, body weight changes, kidney
weight changes, or gross pathology related to THF exposures were reported for male rats. In the
low-concentration group, no gross or histopathological effects were observed. No clinical effects
or gross pathology changes related to the THF exposures were reported for female mice. In mice
exposed for 5 days, absolute and relative uterus weights were decreased in the high-dose group.
In mice exposed for 5 days and followed for a 21-day recovery period, relative uterus weights
were decreased (up to 21%) and appeared to decrease in a concentration-dependent manner,
although this decrease was not statistically significant. In mice exposed for 20 days, statistically
significant increases in absolute body weight (5%), absolute liver weight (11%), and relative
liver weight (6%) were reported. The absolute and relative uterus weights were decreased by 11
and 15%, respectively. None of the uterus weight changes for any of the groups were
statistically significant. No treatment-related histopathological effects were observed in the
uterus at any concentration. Histopathological effects in the form of fatty phanerosis (unmasking
of previously invisible fat in the cytoplasm), especially in zones 3 (centrilobular) and 2
(midzonal), were observed in the livers of mice exposed to THF for 5 days or 20 days, but not in
mice that had 5 days of THF exposure followed by a 21-days recovery period. Specifically, the
study authors reported that fatty phanerosis was present in 5/10 and 10/10 animals exposed for 5
consecutive days at the mid and high THF concentration, respectively. Similar fatty changes
were also seen in livers from all mice that were exposed for 20 days to the high THF
concentration. It should be noted that "fatty phanerosis" is an obsolete term (Popjak, 1945) and
that "fatty infiltration," "fatty degeneration," or "fatty change" may be more appropriate to
describe the morphological manifestation of altered fat metabolism of the parenchyma cells. The
report indicated that there were no additional liver changes including cloudy swelling, vacuolar
degeneration, or necrosis. Other histopathological changes in the high-concentration 5-day
exposure group included a change in the hepatocyte cytoplasm to a more homogeneously
eosinophilic appearance as compared with hepatocytes in control livers.
Kawata and Ito (1984) evaluated the health effects of THF following several different
inhalation exposure regimens. Male Wistar rats (5/control group and 25/experimental group)
were exposed to 15,000 ppm (44,250 mg/m3) THF for a single 30-minute exposure or for seven
30-minute exposures. In addition, rats were exposed to 3,000 ppm (8,850 mg/m3) THF vapor for
1 hour/day, 5 days/week for 12 weeks. Animals were observed for clinical signs and body
weight. Blood was collected for serum chemistry analysis from animals exposed to 3,000 ppm
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
only. The following tissues were collected for histopathology: brain, thymus, lung, heart, liver,
kidney, and spleen. Animals exposed to 15,000 ppm developed clinical signs of face-washing,
shaking head, and rubbing face with paws. These behaviors were weaker and had shorter
duration compared with those observed in rats that received repeated exposures (either seven
30-minute exposures to 15,000 or 3,000 ppm for 12 weeks). In addition, rats receiving seven
exposures to 15,000 ppm developed irritation of skin and mucous membranes as evidenced by
severe salivation and nasal discharge. Rats exposed to 3,000 ppm for 12 weeks also developed
irritation symptoms that were milder than those observed at 15,000 ppm. No effects on body
weight were observed after either single or multiple exposures to 15,000 ppm. However, by the
fourth week of exposure, rats exposed to 3,000 ppm had significantly reduced body weight
compared with controls. Serum chemistry parameters were comparable between treated and
control animals. No histopathological lesions were observed in either of the groups exposed to
15,000 ppm. In the animals exposed to 3,000 ppm, histopathological lesions were observed in
both lungs and kidney. Papillary hyperplasia and catarrhal (inflammation of mucus membranes)
degeneration were observed in lungs and bronchial epithelium. Protein casts and hyaline droplet
degeneration were observed in the kidney tubule lumen epithelium in kidneys. Based on lung
and kidney histopathological lesions, EPA identified 3,000 ppm (8,850 mg/m3) as a LOAEL; a
NOAEL was not established.
BASF (1938) investigated the subchronic effects of THF exposure in dogs. Four dogs
(strain and sex not specified) were exposed by inhalation to THF vapor at a concentration of
200 ppm (590 mg/m3) 6 hours/day, 5 days/week for 9 weeks, followed by exposure to a
concentration of 366 ppm (1,080 mg/m3) 6 hours/day, 5 days/week for 3 weeks. At the end of
the 12 weeks, two of the four dogs were exposed on 2 successive days to a THF concentration of
approximately 2,100 ppm (5,250 mg/m3). Blood pressure was measured in dogs in the morning
and afternoon for a 4-week control period and then before and after each daily exposure during
the 12-week exposure period. Hematology, urinalysis, and limited pathological evaluations were
also completed. Pulse pressure was decreased in 3/4 dogs following exposure to 200 ppm during
weeks 3-4 of the study. In addition, increasing the THF concentration to 366 ppm resulted in a
decrease in blood pressure compared to the control period in 3/4 dogs. In the two dogs exposed
to 2,100 ppm THF, a "sharp drop" in systolic, diastolic, and pulse pressure was reported by the
study authors after the second day of exposure. No signs of narcosis or eye or respiratory tract
irritation were observed in these two dogs. In one dog, hemoglobin decreased and white blood
cells increased compared to the control levels. However, examination of the urine did not reveal
any abnormality in kidney function. No gross or microscopic pathology was observed in the
heart, lungs, spleen, pancreas, or kidneys of any of the dogs. Based on alterations in blood
pressure, the study authors (BASF, 1938) reported a LOAEL of 200 ppm (590 mg/m3).
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
34
35
4.2.2. Chronic Studies and Cancer Bioassays
4.2.2.1.	Oral
No chronic studies in animals by the oral route of exposure were identified.
4.2.2.2.	Inhalation
Stasenkova and Kochetkova (1963) evaluated the effects of a 6-month inhalation
exposure on rats. Male rats (20/group, strain not specified) in a single exposure group were
exposed to air concentrations of 1-2 mg/L (1,000-2,000 mg/m3) 4 hours/day, 7 days/week for
6 months. Endpoints evaluated included clinical signs, body weight changes, blood cell count,
blood pressure, and functional condition of the neurovascular system, liver, and kidney. At the
end of the 6-month treatment period, animals were sacrificed and histopathological examination
of major organs was conducted. No effects were observed on behavior, body weight, liver and
kidney function, or neuromuscular irritability of treated rats compared with controls. Within 2-
3	months of treatment, exposed rats developed increased numbers of leukocytes, which remained
elevated compared with controls for the remainder of the experimental period. After 3-
4	months, blood pressure in treated rats was reduced compared to controls, and this observation
continued for the remainder of the treatment period. Histopathological lesions included mild
hypertrophy in the muscle fibers of the bronchi walls and arteries of lungs and spleen. Because
of poor reporting of this study, no NOAEL orLOAEL can be identified.
NTP (1998) reported on the chronic toxicity and carcinogenicity of THF inhalation
exposure in rats and mice. In the 2-year study, groups of F344/N rats and B6C3Fi mice
(50/sex/group) were exposed to 0, 200, 600, or 1,800 ppm (0, 590, 1,770, or 5,310 mg/m3) THF
6 hours/day, 5 days/week for 105 weeks. Survival of treated rats was comparable to chamber
controls at all exposure levels. Neither mean body weight differences nor clinical findings
related to THF exposure were reported for either male or female rats. Pathology noted at
sacrifice in male rats included apparent increases of renal tubular epithelial adenoma (at 600 and
1,800 ppm) and two renal tubular epithelial carcinomas (at 1,800 ppm), which, when combined
with the adenomas, suggested a treatment-related trend. . The incidences of adenoma or
carcinoma in the 600 and 1,800 ppm males exceeded the historical range for chamber controls in
the 2-year NTP (1998) inhalation studies, and the overall trend was statistically significant
(p=0.037). Table 4-3 summarizes the incidence of neoplastic and nonneoplastic changes in the
kidney of male rats. No treatment-related changes in the incidence of neoplastic or
nonneoplastic lesions in other tissues in the male or female rats were observed.
32
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-3. Renal findings in male F344/N rats exposed to THF for 2 years

Control
200 ppm
600 ppm
1,800 ppm
Number of animals examined
50
50
50
50
Nephropathy, chronic
48a (3.0)b
50 (2.9)
50 (3.1)
50 (3.0)
Hyperplasia
7 (3.4)b
5 (3.6)
6 (2.5)
7(3.3)
Mineralization
8 (16%)°
7(14%)
2 (4%)
5 (10%)
Adenoma
1/50 (2%)
1/50 (2%)
4/50 (8%)
3/50 (6%)
Carcinoma
0/50 (0%)
0/50 (0%)
0/50 (0%)
2/50 (4%)
Adenoma or carcinomad
1/50 (2%)
1/50 (2%)
4/50 (8%)
5/50 (10%)
Adjusted rate6
8.3 %
16.7%
18.8%
38.3%
First incidence (days)
733 (T)f
733 (T)
631
668
Logistic regression test8
p = 0.037
p = 0.602
p = 0.159
p = 0.065
aNumber of animals with lesions.
bAverage severity of lesions in affected animals: 1 = minimal; 2 = mild; 3 = moderate; 4 = marked.
°Percent affected.
historical incidence for 2-year inhalation studies with chamber controls: 6/652 (0.9 ± 1.3%); historical control
range, 0-4%.
eKaplan-Meier estimated tumor incidence at the end of the study, incorporating an adjustment for intercurrent
mortality.
fT = terminal sacrifice.
gIn the control column are the p values associated with the trend test. In the exposed group column are the p values
corresponding to the pair-wise comparison between the controls and the exposed group.
Sources: Adapted from Chhabra et al. (1998); NTP (1998).
1
2	No treatment-related effects on survival or clinical observations were noted in female
3	mice. Several statistically significant pathological changes were reported. These included
4	concentration-related trends in hepatocellular adenoma or carcinoma (p < 0.001). An increase in
5	liver necrosis was also observed in females exposed to 1,800 ppm THF. Table 4-4 summarizes
6	the incidence of neoplastic and nonneoplastic changes in the livers of female mice.
7
33
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Table 4-4. Liver findings in female B6C3Fi mice exposed to THF for 2 years

Control
200 ppm
600 pm
1,800 ppm
Number of animals examined
50
50
50
48
Eosinophilic focus
T
9
7
11
Necrosis
3 (2.0)b
0
0
7(1.9)
Adenoma
12/50 (24%)°
17/50 (34%)
18/50 (36%)
31/48 (65%)
Logistic regression testd
p < 0.001
p = 0.249
p = 0.188
p < 0.001
Carcinoma
6/50 (12%)
10/50 (20%)
10/50 (20%)
16/48 (33%)
Adenoma or carcinoma6
17/50 (34%)
24/50 (48%)
26/50 (52%)
41/48 (85%)
Adjusted ratef
46.3%
61.3%
69.1%
93.0%
First incidence (days)
478
552
469
399
Logistic regression test
p < 0.001
p = 0.188
p = 0.086
p < 0.001
aNumber of animals with lesion.
bAverage severity of lesions in affected animals: 1, minimal; 2, mild; 3, moderate; 4, marked.
°Percent affected.
dIn the control column are the p values associated with the trend test. In the exposed group column are the p values
corresponding to the pair-wise comparison between the controls and that of the exposed group.
eHistorical incidence: 200/937 (21.3% ± 11.9%); range, 3-54%.
fKaplan-Meier estimated tumor incidence at the end of the study, incorporating an adjustment for intercurrent
mortality.
Sources: Adapted from Chhabra et al. (1998); NTP (1998).
In male mice, mean survival of the 1,800 ppm exposed group was significantly less than
chamber controls (average life span of 456 versus 689 days). As a result, the number of male
mice available for evaluation of neoplastic changes at the termination of the study was small
(12 animals compared to 32 animals in the control group). The only clinical observation was
narcosis in male mice exposed to THF at 1,800 ppm that lasted up to 1 hour following exposure.
During periods of narcosis, the preputial fur was wet with urine, a condition that was thought to
increase urogenital tract lesions and possibly lead to decreased survival. The lower survival rate
and pathology findings, including bone marrow and lymph node hyperplasia, hematopoetic
proliferation of the spleen, and thymic atrophy, were considered by the study authors (NTP,
1998) to be secondary to the urogenital tract inflammation. Although the number of male mice
surviving to termination was small, statistical analyses for early mortality by NTP (1998) did not
indicate that there was a treatment-related effect of THF on the incidence of liver tumors in male
mice. Overall the only effect observed was clinical signs of toxicity (narcosis) in male mice at
1,800 ppm (5,310 mg/m3).
Under the conditions of this 2-year bioassay, the NTP (1998) concluded that there was
some evidence of carcinogenic activity of THF in male F344/N rats due to increased incidences
of adenoma or carcinoma of the kidney. There was clear evidence of carcinogenic activity of
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
THF in female B6C3Fi mice due to increased incidences of hepatocellular adenomas or
carcinomas.
4.3. REPRODUCTIVE/DEVELOPMENTAL TOXICITY STUDIES—ORAL AND
INHALATION
4.3.1. Oral
BASF (1994) reported the results of a one-generation reproductive toxicity range-finding
study in rats given THF in drinking water. Male and female Wistar rats (10/sex/dose) were
given THF at concentrations of 0, 4,000, 8,000, or 12,000 ppm in the drinking water for 7 weeks
prior to mating and throughout cohabitation, gestation, and lactation. THF intake values
estimated from measured water consumption and body weights are shown in Table 4-5. The F0
females were allowed to litter and rear pups (F1 generation) for 4 days postpartum, at which time
the litters were culled to eight pups/litter (ideally four of each sex). Culled pups were sacrificed
and examined for gross pathologic lesions, and the surviving F1 pups were sacrificed after
weaning on postnatal day (PND) 21. Clinical chemistry, hematology, and urinalysis parameters
were measured in the F0 animals near the end of the study (approximately 12 weeks from
initiating exposure), after which the F0 animals were sacrificed and assessed for gross pathology.
Key treatment-related findings are also summarized in Table 4-5.
35
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-5. Selected findings from one-generation reproductive toxicity study
in Wistar rats exposed to THF in drinking water
Generation,
sex

Concentration (ppm)
Parameter3
0
4,000
8,000
12,000
F0 Generation
Males
THF intake (mg/kg-day)
0
444
795
1,107
Females
THF intake (mg/kg-day):
Premating
Gestation
Lactation
All periods
O O O O
467
434
714
503
798
758
1264
890
1,088
1,139
1,847
1,240
Males
Food consumption (g/day)
28.3 ± 1.81
28.1 ± 1.87
27.0 ± 1.57
25.9 ± 1.79b
Females
Food consumption (g/day)
19.9 ±0.54
20.5 ±0.72
18.8 ±0.67
19.6 ±0.62
Males
Water consumption (g/day)
28.2 ± 1.80
26.8 ± 1.91
23.7 ± 1.60b
21.5 ± 1.94b
Females
Water consumption (g/day)
21.1 ±0.92
19.8 ± 1.09
16.2 ± 0.73b
15.1 ± 0.87b
Males
Body weight (g)
355.4 ±31.61
356.7 ±32.09
342.0 ±46.72
327.0 ±34.32
Females
Body weight gain (g)
104.6 ± 14.62
115.7 ± 15.75
100.9 ±9.94
104.4 ± 12.42
Males
Absolute kidney weight (g)
3.071 ±0.178
3.032 ±0.223
3.101 ±0.289
3.141 ±0.302
Females
Absolute kidney weight (g)
2.012 ±0.157
2.115 ±0.202
2.036 ±0.12
2.153 ±0.167
Males
Relative kidney weight
(%BW)
0.654 ±0.047
0.647 ±0.021
0.680 ±0.036
0.705 ± 0.049b
Females
Relative kidney weight
(%BW)
0.717 ±0.034
0.735 ±0.035
0.775 ± 0.04b
0.783 ± 0.048b
F1 Generation (pups)
Males
Body weight gain (g) PND
4-21
44.0 ±3.16
42.4 ±3.52
40.6 ± 3.18b
37.6 ± 5.33b
Females
Body weight gain (g) PND
4-21
42.7 ±3.50
40.3 ±2.60
38.0 ± 3.26b
36.2 ± 4.44
aAll values except for THF intake are shown as mean ± standard deviation (SD); THF intake shown as means.
Statistically different (p < 0.05) from controls.
Source: BASF (1994).
1
2	Food consumption was statistically significantly reduced in the high-dose F0 males and
3	in the mid-dose F0 females. Water consumption was statistically significantly decreased in both
4	sexes at the mid- and high-doses. No mortalities were recorded in either the F0 or F1 rats at any
5	exposure concentration. No effects were observed for any measured reproductive endpoint.
6	However, relative kidney weights were statistically significantly increased in high-dose F0
7	males and in mid- and high-dose F0 females. In the F1 generation, numbers of pups, sex ratio,
8	and viability/mortality were comparable to controls. Mean body weight gains of both male and
9	female F1 pups were statistically significantly decreased in both the mid- and high-dose groups.
10	The NOAEL for this study was 571 mg/kg-day and the LOAEL is 1005 based on decreased pup
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
20
21
22
23
24
25
26
27
body weight gain and using time-weighted average maternal THF intake during gestation and
lactation.
The results from this range-finding study were used to select dose levels for a two-
generation developmental and reproductive toxicity study of THF administered to rats in
drinking water (Hellwig et al., 2002; BASF, 1996). Wistar rats (25/sex/group) received THF in
their drinking water at concentrations of 0, 1,000, 3,000, or 9,000 ppm for 70 days prior to
mating and throughout cohabitation, gestation, and lactation. THF intake values estimated from
measured water consumption and body weights are shown in Table 4-6. Before weaning,
25 F1 pups/sex/group were randomly selected to be the F1 parental animals. The remaining
F1 pups were sacrificed. After the F1 generation pups were weaned, the F0 animals were
sacrificed. The F1 animals were exposed continuously to THF at the same concentrations as
their parents from weaning and throughout cohabitation, gestation, and lactation. THF intake
values estimated from measured water consumption and body weights are shown in Table 4-6.
Endpoints evaluated in F0 and F1 parental animals included food and water consumption,
body weight, mortality, and clinical signs. In addition, necropsy was performed on all parental
animals at sacrifice, and organ weights were obtained for kidney, liver, testes, and epididymis.
Histopathology was performed on all gross lesions, liver, kidney, reproductive organs, and GI
organs of sacrificed parental animals. Reproductive endpoints evaluated include mating index,
fertility index, gestation index, and live birth index. Litter/delivery endpoints for both F1 and F2
generations included total number of pups, number of live and stillborn pups, sex ratio, clinical
signs, body weight, viability index, and lactation index. In addition, pups were evaluated for
developmental stages (pinna unfolding, opening of auditory canal, opening of eyes) and
behavioral tests (grip reflex, acoustic startle, pupil constriction). Culled pups, surplus pups, and
all pups that died before weaning were assessed macroscopically and, if abnormalities were
found, were evaluated by skeletal staining and histological processing of the head. Key
treatment-related findings are summarized in Table 4-6.
Table 4-6. Selected findings from a two-generation reproductive toxicity
study in Wistar rats exposed to THF in drinking water


Concentration (ppm)
Sex
Parameter"
0
1,000
3,000
9,000
F0 Generation
Males
THF intake (mg/kg-day)
0
91
268
714
37
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-6. Selected findings from a two-generation reproductive toxicity
study in Wistar rats exposed to THF in drinking water


Concentration (ppm)
Sex
Parameter3
0
1,000
3,000
9,000
Females
THF intake (mg/kg-day):
Premating
Gestation
Lactation
All periods
0
0
0
0
104
104
166
112
301
288
478
322
742
790
1,365
835
Males
Food consumption (g/day)
27.3 ± 1.35
27.2 ± 1.39
27.0 ± 1.48
26.5 ± 1.42
Females
Food consumption (g/day):
Premating
Gestation
Lactation
19.9 ±0.61
25.0 ± 1.03
47.8 ± 12.34
20.0	±0.79
25.1	± 1.28
47.4 ± 10.08
19.6 ±0.73
24.4 ± 1.26
46.6 ± 10.17
18.3	±0.71d
23.4	± 1.38d
46.0 ± 9.46d
Males
Water consumption (g/day)
26.9 ± 1.35
25.8 ± 1.12
25.1 ± 1.03
22.0 ± 0.99d
Females
Water consumption (g/day):
Premating
Gestation
Lactation
20.6 ± 1.21
32.2	±6.63
57.3	± 16.06
19.7 ± 1.15
29.7 ±6.10
52.4 ± 12.70d
19.1 ± 0.95d
27.9 ± 5.75d
50.7 ± 12.05d
15.1 ± 1.00d
24.3 ± 5.81d
45.9 ± 11.27d
Males
Body weight (g)
379.3 ±53.52
378.3 ±36.69
374.3 ±40.73
364.2 ± 39.41
Females
Body weight gain (g):
Premating
Gestation
Lactation
138.3 ± 17.30
129.7 ± 15.46
9.7 ± 14.02
138.9 ± 16.31
127.0 ± 14.01
3.7 ± 12.10
141.7 ± 13.59
124.7 ±21.03
9.9 ±9.71
128.5 ± 14.25
128.3 ± 15.82
7.9 ±9.53
Males
Absolute kidney weight (g)
3.244 ±0.301
3.203 ±0.284
3.104 ±0.272
3.438 ±0.27b
Females
Absolute kidney weight (g)
2.092 ±0.113
2.126 ±0.142
2.159 ±0.146
2.123 ±0.133
Males
Relative kidney weight (%BW)
0.665 ±0.052
0.662 ±0.057
0.641 ±0.059
0.719 ±0.059b
Females
Relative kidney weight (%BW)
0.749 ±0.039
0.774 ±0.05
0.774 ±0.054
0.785 ±0.033b
F1 Generation (Pups)
Maternal THF intake (mg/kg-day), TWA of
FO gestation and lactation periods3:
0
134
381
1071
Male
pups
Body weight gain (g):
PND 4-21
PND 1-4
PND4-7
PND 7-14
PND 14-21
45.4 ±3.04
3.0	±0.57
6.1	±0.57
17.8 ± 1.15
21.4 ±2.37
46.3 ±3.23
3.3 ±0.96
6.0 ±0.71
17.5 ± 1.55
22.7 ± 1.80
44.8	±3.63
2.7	±0.80
5.8	±0.74
17.2 ± 1.43
21.9	±2.03
41.7 ± 3.38b
2.6 ±0.53
5.5 ± 0.75b
15.7 ± 1.65b
20.5 ± 1.84
Female
pups
Body weight gain (g):
PND 4-21
PND 1-4
PND 4-7
PND 7-14
PND 14-21
43.3 ±2.72
2.8	±0.60
5.9	±0.50
17.3 ± 1.47
20.1 ± 1.97
44.0 ±3.45
3.1 ±0.85
5.6 ±.10
17.4 ± 1.72
20.7 ± 1.86
42.3 ±2.61
2.7 ±0.80
5.5 ±0.52
16.9 ± 1.66
19.9 ± 1.42
40.1	±3.46b
2.6 ±0.51
5.3 ± 0.65b
15.6 ± 1.56b
19.2	± 1.84
F1 Generation
Males
THF intake (mg/kg-day)
0
98
293
788
3 8	DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-6. Selected findings from a two-generation reproductive toxicity
study in Wistar rats exposed to THF in drinking water
Sex
Parameter"
Concentration (ppm)
0
1,000
3,000
9,000
Females
THF intake (mg/kg-day):
Premating
Gestation
Lactation
All periods
0
0
0
0
125
107
152
125
358
318
455
362
882
792
1,165
898
Males
Food consumption (mg/kg-day)
28.0 ± 1.90
28.3 ± 1.77
28.1 ± 1.98
26.3 ± 1.99b
Females
Food consumption (mg/kg-day)
Premating
Gestation
Lactation
21.1 ±0.50
26.6 ± 1.53
47.0 ± 13.63
21.4 ±0.44
26.7	± 1.46
44.8	± 11.93
21.0 ±0.44
26.6 ± 1.33
44.6 ± 12.69
20.9 ±0.68
26.0 ± 1.42
40.5 ± 11.55b
Males
Water consumption (g/day)
27.9 ±2.07
29.2 ±2.23
28.8 ±2.65
24.2 ± 2.39b
Females
Water consumption (g/day):
Premating
Gestation
Lactation
23.5 ± 1.28
32.3 ±7.74
57.0 ± 15.32
25.9 ± 1.63
33.8 ±8.00
52.5 ± 10.82
24.0	± 1.07
33.1	±6.54
52.1 ± 11.64
19.5	± 0.89b
27.7 ± 6.53b
43.6	± 10.93b
Males
BW (g)
453.4 ±40.49
456.6 ±35.27
458.4 ±53.88
426.1 ±37.39
Females
BWgain(g): Premating
Gestation
Lactation
198.0	± 19.39
127.1	± 17.23
10.9 ± 13.44
201.0 ±22.92
128.0 ± 14.22
4.6 ± 12.86
204.5 ±23.68
125.0 ± 19.18
7.6 ± 10.99
208.0 ± 24.49
112.6 ± 17.79b
9.4 ± 14.05
Males
Absolute kidney weight (g)
3.233 ±0.455
3.208 ±0.192
3.201 ±0.348
3.181 ±0.338
Females
Absolute kidney weight (g)
2.347 ±0.144
2.364 ±0.201
2.365 ±0.2
2.411 ±0.153
Males
Relative kidney weight (%BW)
0.62 ±0.099
0.606 ±0.041
0.608 ±0.05
0.642 ±0.058
Females
Relative kidney weight (%BW)
0.805 ±0.048
0.8 ±0.066
0.812 ±0.043
0.826 ±0.059
F2 Generation
Maternal THF intake (mg/kg-day), TWA of
F1 gestation and lactation periods3:
0
129
385
974
Male
pups
BW gain (g): PND 4-21
PND 1-4
PND 4-7
PND 7-14
PND 14-21
42.6 ±3.55
2.7 ±0.85
5.7 ±0.95
17.4 ± 1.56
19.4 ±2.23
43.8	±4.67
3.0 ± 1.22
5.8 ±0.82
17.9	± 1.98
20.2 ±2.63
41.5 ±4.64
2.7 ± 1.00
5.3 ± 1.15
17.0 ± 1.94
19.2 ±2.07
39.5	± 3.13b
3.0 ±0.75
5.0 ± 0.63b
15.6	± 1.67b
18.9 ± 1.71
Female
pups
BW gain (g): PND 4-21
PND 1-4
PND 4-7
PND 7-14
PND 14-21
40.7 ±3.67
2.7 ±0.71
5.6 ±0.75
17.2 ± 1.50
17.9 ±2.26
41.2 ±3.35
2.7 ± 1.10
5.2 ± 1.32
17.1 ± 1.62
18.6 ± 1.83
38.7	±4.67
2.4 ± 1.11
5.0 ± 1.12
16.0 ±2.41
17.8	±2.69
38.1 ±3.67
2.9 ±0.75
5.0 ±0.64
15.4 ± 1.84b
17.6 ±2.15
All pups
% with eyes open on PND 15
89.9 ±22.73
98.7 ±4.13
94.0 ± 12.43
79.2 ± 31.18b
aAll values except for THF intake are shown as mean ± SD; THF intake shown as mean. TWA= time weighted
average using 22 days for gestation and 21 days for lactation. PND = postnatal day.
Statistically significantly different (p < 0.05) from controls.
Sources: Hellwig et al. (2002); BASF (1996).
39
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
In the F0 generation, food consumption of the high-dose females was statistically
significantly reduced during selected weekly measurements compared with controls during the
premating period, gestation, and lactation. Water consumption for males in the high-dose group
was statistically significantly decreased during the premating period, and for the mid- and high-
dose females it was statistically significantly decreased during the premating period, gestation,
and lactation. In high-dose females, body weights were statistically significantly decreased
compared with controls during selected periods during premating, gestation, and throughout
lactation, but no significant change in body weight gain was observed. No clinical signs related
to THF were observed in either F0 males or females at any dose. In F0 males, the mating index
and fertility index were comparable among the controls and treated groups. Similarly, the
mating and fertility indices for F0 females were comparable among control and treated groups.
The mean duration of gestation was similar in all groups and the gestation index was 100% for
all groups. Absolute kidney weight was increased in high-dose males, and relative kidney
weight was significantly increased in both high-dose male and female F0 rats. No treatment-
related gross lesions or microscopic findings were observed in either males or females.
The total number of F1 pups delivered, the number of live and stillborn pups, and the sex
ratio were comparable among the groups. In the low-dose group, nine F1 pups from a single
litter died between days 1 and 10. Also, two dams in the mid-dose group cannibalized pups.
Therefore, the lactation index for these dose groups is statistically significantly decreased
compared to controls. However, the authors concluded that this decrease is not related to
administration of THF, because there was no dose-response relationship. The mean body
weights and body weight gains of the F1 pups in the high-dose group were significantly
decreased during PNDs 4-7 and PNDs 7-14. The treated F1 pups did not demonstrate any
clinical signs, changes in developmental stages, changes on behavioral tests, or findings on
necropsy compared with controls.
Food consumption was significantly decreased in high-dose F1 male adult rats during the
premating period and in high-dose F1 female rats during lactation. Water consumption was
significantly decreased in high-dose F1 male rats during the premating period and in the high-
dose F1 females during the premating period, gestation, and lactation. In high-dose F1 males,
slight but significant decreases in body weight were observed throughout the study, but no effect
on body weight gain was observed. No effects on body weight or body weight gain were
observed in F1 female adults. No clinical signs related to THF were observed in either F1 males
or females at any dose. In F1 males and females, the mating and fertility indices were
comparable among the controls and treated groups. The mean duration of gestation was similar
in all groups, and the gestation index was 100% for all groups. No treatment-related effects on
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
organ weight, gross lesions, or microscopic findings were observed in either the male or female
F1 adult rats at any exposure concentration.
The mean number of delivered F2 pups/litter was decreased 16% in the high-dose group
compared with control (12.4, 13.0, 12.9, and 10.4 in the 0, 1,000, 3,000, and 9,000 ppm dose
groups, respectively) and was outside the range of historical control values of 11.1-16.4 (BASF,
1996). The study authors concluded that this was a spontaneous finding and was not related to
treatment since it was not seen in the F0 generation or in the range-finding study; the decrease
was limited to a few litters with <6 pups/litter (BASF, 1996). Data on the number of
implantations and resorptions were not reported. Also, one F1 parental male rat in the high dose
group was found to be infertile which, in the absence of corroborating histopathology findings,
was considered a spontaneous finding (BASF, 1996). The number of stillborn pups was
statistically significantly increased in the two lower dose groups, but not in the high-dose group.
Based on the lack of dose-response relationship, the authors concluded that all of these findings
were spontaneous and not related to THF administration. In the low- and mid-dose groups, there
was an increase in the number of pups cannibalized or dead before scheduled sacrifice. As a
result, the viability index was statistically significantly decreased in the low-dose group; the
viability indices for the mid- and high-dose groups were comparable to controls. Body weight
gain was statistically significantly reduced in the high-dose male and female F2 pups during
PNDs 7-14. A significant number of F2 pups/litter in the high dose group had delayed opening
of eyes (% with eyes open on PND 15); 79 compared to 90% in controls, historical control range
85-100%). Also, there was an increase in the number of sloped incisors in the high dose F2
litters (mean 1.5% of pups/litter compared to 0% in controls; historical control range 0-2.9%).
The study authors considered this finding to be consistent with a slight developmental delay
(Hellwig et al., 2002). The mean percentage of F2 pups/litter with open auditory canal was 96.4,
100, 88.9, and 98.9% in the 0, 1,000, 3,000, and 9,000 ppm dose groups, respectively. This
finding was discounted by the study authors because it was not dose-related and the statistically
significantly different value of 88.9% in the mid-dose group fell within the historical control
range of 81-100%. Values for lactation index, sex ratio, clinical signs, behavioral tests, and
necropsy findings were comparable between controls and treated animals.
In the high concentration groups, general toxicity was indicated by slight to marginal
decreased food consumption, decreased body weight, and increased kidney weight in F0 adults
and decreased food consumption and body weight gain in F1 adults. However, decreased adult
body weights were only observed during selected periods during the study, were of minimal
severity, and were not generally reflected by changes in body weight gain. Therefore, the adult
body weight changes were not considered to be of sufficient magnitude to identify an adverse
effect level.
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
26
27
28
29
30
31
32
33
34
35
36
No clinical signs (in the one- or two-generation studies) or clinical chemistry changes
(only measured in the one-generation study) consistent with dehydration were observed,
suggesting that the decrease in water consumption was not inducing changes in maternal health.
The study authors stated that the reduced water consumption observed in the mid- and high-dose
F0 and high-dose F1 parental rats was most likely due to reduced palatability of the THF in
drinking water. The reduction in water intake averaged 7% during premating and 12-14%
during gestation and lactation following exposure to 3,000 ppm THF. There were no
corresponding decreases in food consumption at this dose during these time periods. Thus,
Hellwig et al. (2002) concluded that the reduction in water consumption was biologically
insignificant and that the NOAEL for systemic toxicity (increased relative kidney weight, body
weight gain, and food consumption) in F0 and F1 parental rats was 3,000 ppm.
Pup weight gain was reduced at the high dose during PNDs 4-7 and 7-14 in both F1 and
F2 pups. This reduction in weight gain may be due to reduced maternal milk production, but the
study authors indicated that it was not related to maternal body weight or water consumption.
Specifically, maternal body weight was reduced significantly in the F0 dams and not the F1 dams
during lactation. Data on the possible relationship between decreased water intake in dams and
decreased production of milk was not provided in this study. Hellwig et al. (2002) stated that
decreased pup weight gain could be related to direct exposure to THF during lactation.
Specifically, the study authors suggested that given that THF is slightly more soluble in lipid
than water, THF may have been more concentrated in the dam's milk fat than in the maternal
water compartment. Based on the developmental effects observed (decreased pup weight gain,
delayed eye opening, and increased incidence of sloped incisors) the study authors designated
3,000 ppm as the NOAEL. The finding of decreased mean number of F2 pups delivered/litter in
the high-dose group (10.4 vs 12.4 in control) is also supportive.
While the two-generation study demonstrated a decrease in pup body weight gain in both
the F1 and F2 generations following THF exposure, the contribution of other potential
confounding factors, such as dam water consumption and litter size (which may influence the
milk availability to each pup), were considered further using multivariable regression analyses.
The regression analyses included pup body weight gain during PNDs 7-14 as the dependent
variable and four independent variables: average THF intake, maternal water intake during
lactation, number of pups in each litter (during the affected postnatal period), and a categorical
variable for the dose group. Since the response data from F1 and F2 generation are independent,
these data were analyzed separately. Preliminary regression analyses suggested that there was a
high degree of colinearity among the independent variables, as indicated by the high variance
inflation factors, and the dose group is the most significantly affected factor. Removal of this
factor diminishes the colinearity in the regression. Therefore, in a second series of regression
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
analyses, dose group was not included as an independent variable. The results from this
regression analysis are summarized in Table 4-7.
Table 4-7. Correlations between decreased pup body weight gain and each
of three independent variables, maternal water intake, THF intake, and
number of pups in each litter

Coefficient
/7-Value
F1 pup body weight gain (adjusted r2 = 0.36)
Average water intake
9.09 x 10"2
o.ooor
Average THF intake
-3.98 x 10"4
0.1458
Number of pups
-4.23 x 10"1
0.03353
F2 pup body weight gain (adjusted r2 = 0.24)
Average water intake
5.90 x 10"2
0.00153
Average THF intake
-8.51 x 10"4
0.02183
Number of pups
-5.04 x 10"1
0.00553
aStatistically significant correlation at p < 0.05.
Based on the results from multiple regression analyses, the dependent variable (pup body
weight gain) can be predicted from a linear combination of the independent variables of average
water intake, THF intake, and number of pups in each litter. For F1 pups, there is no evidence to
suggest a statistically significant correlation (p = 0.1458) between maternal THF intake and pup
BW gain when controlling the other confounding factors, such as maternal water intake and
number of pups in each litter. However, the similar analysis for the F2 pup data indicates that
there is a significant correlation (p = 0.0218) between pup body weight gain and maternal THF
intake after controlling for the other confounding factors. The study authors concluded that the
high concentration effects reflect general toxicity of THF, while noting that decreased water (and
food) intake could have contributed to the observed decrease in body weights.
Based on these analyses for parental (increased kidney weight and decreased body
weight) and developmental effects (decreased pup body weight gain and delayed eye opening),
the NOAEL is 3,000 ppm and the LOAEL is 9,000 ppm for this study. The best value to use for
estimating the corresponding doses (mg/kg-day) differs for each generation, based on THF
intake values over the relevant period of exposure. For parental effects, time-weighted average
(TWA) THF intakes over the entire study period are appropriate for use in assigning effect
levels. For developmental effects, the TWA THF intake during the gestation and lactation period
of the parent females was used to estimate the effective dose. Table 4-8 summarizes the
corresponding effect level doses across all endpoints that showed a treatment-related effect.
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
Table 4-8. Summary of effect levels observed in the two-generation
reproduction study in Wistar rats exposed to THF in drinking water
Effect
NOAEL
(mg/kg-day)
LOAEL
(mg/kg-day)
F0 Males—increased kidney weight
268
714
F0 Females—decreased body weight, increased kidney weight
322
835
F1 Adult males—decreased body weight gain
268
788
F1 Adult females—decreased body weight gain
362
898
F1 Pups—decreased body weight gain
381
1,071
F2 Pups—decreased pup body weight gain and delayed eye opening
385
974
Sources: Hellwig et al. (2002); BASF (1996).
4.3.2. Inhalation
Mast et al. (1992) assessed developmental toxicity of THF in mice and rats. Female
CD-I mice (10 virgin and 30 mated animals/group) were exposed to 0, 600, 1,800, or 5,000 ppm
(0, 1,770, 5,310, or 14,750 mg/m3) THF vapor for 6 hours/day, 7 days/week on gestation days 6-
17. Female mice in the 5,000 ppm group demonstrated a high toxicity, with >25% mortality
observed after only 6 days of exposure. Consequently, mice in this group were removed from
exposure at this time and placed in a chamber with fresh air until time of scheduled sacrifice.
Developmental evaluations were conducted on pregnant mice euthanized on gestation day 18.
Developmental endpoints included gross maternal toxicity and number, position, and status of
implantation sites. Live fetuses were weighed, sexed, and examined for gross defects. Half of
the live fetuses and any fetus with gross defects were examined for visceral defects, and the
heads were examined for soft-tissue craniofacial abnormalities. All fetal carcasses were
examined for gross changes in cartilage and ossified bone. Maternal deaths occurred in the high-
concentration group. Other statistically significant maternal effects that were observed at
concentrations of >1,800 ppm included narcosis, decreased terminal body weight, reduced
adjusted maternal weight gain (adjusted for uterine weight), and reduced gravid uterine weight.
A reduction in the percent live pups/litter and delayed ossification of the sternum were observed
at concentrations of >1,800 ppm. Surviving pregnant mice in the high concentration group had
litters with a 95% resorption incidence; however, if the conceptus survived, development
continued normally. There were no effects on the number of implantations, the fetal sex ratio, or
the incidence of abnormalities in fetuses. Based on decreased gravid uterine weight in dams and
reduced fetal survival, EPA identified the LOAEL as 1,800 ppm (5,310 mg/m3) and the NOAEL
as 600 ppm (1,770 mg/m3) in mice.
Pregnant Sprague-Dawley rats (10 virgin and 30 mated animals/group) were exposed to
0, 600, 1,800, or 5,000 ppm (0, 17, 70, 5,310, or 14,750 mg/m3) THF vapor for 6 hours/day,
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
27
28
29
30
31
32
33
34
35
36
7 days/week on gestation days 6-19 (Mast et al., 1992). Developmental evaluations were
conducted on pregnant rats euthanized on gestation day 20. Developmental endpoints included
gross maternal toxicity and the number, position, and status of implantation sites. Live fetuses
were weighed, sexed, and examined for gross defects. Half of the live fetuses and any fetus with
gross defects were examined for visceral defects, and the heads were examined for soft-tissue
craniofacial abnormalities. All fetal carcasses were examined for cartilage and ossified bone. In
dams, the cumulative BWs were significantly reduced in the high concentration group
throughout the exposure period. In addition, nonsignificant reductions of gravid uterine weight
and extragestational weight gain (adjusted for uterine weight) were observed in the high
concentration group. Fetal rat weights were significantly reduced at 5,000 ppm. There were no
effects on the number of implantations, fetal sex ratio, or incidence of fetal abnormalities. Based
on decreased maternal and fetal weight, EPA identified the LOAEL as 5,000 ppm (14,750
mg/m3) and the NOAEL as 1,800 ppm (5,310 mg/m3) in rats.
DuPont Haskell Laboratory (1980) investigated the effects of inhaled THF on the
developing fetus. The authors first performed a range-finding study in which Crl:CD® rats (7-
14/group) were exposed to 0, 590, 1,475, 7,375, or 14,750 mg/m3 6 hours/day on gestation days
6-15. In a follow-up study, Crl:CD® rats (29/group) were exposed to 0, 2,950, or 14,750 mg/m3
THF 6 hours/day on gestation days 6-15. Body weight, clinical signs, and feed consumption
were observed in dams during the exposure period. Dams were sacrificed on gestation day 21
and were examined for gross pathologic changes, liver weight, and reproductive status. The
number of corpora lutea, implantation sites, and live and dead fetuses were recorded. Live
fetuses were weighed, sexed, and examined for external alterations. One-third of all fetuses and
all stunted or malformed fetuses were examined for visceral alterations, and the heads were fixed
for evaluation of eye malformations. Remaining fetuses were fixed and stained for examination
of skeletal alterations. The same endpoints were examined in both parts of the study.
No mortality was observed in dams in either study. In both studies, dams in the high-
concentration group demonstrated decreased response to noise stimulus, reduced muscle tone,
and staggering gait that persisted for about 1 hour following each daily exposure period. In
addition, dams in the lower concentration group (7,375 mg/m3 in the range-finding study and
2,950 mg/m3 in the main study) had a diminished response to noise stimulus. Food consumption
in the main study high-concentration group was significantly reduced compared to controls. In
both studies, dams in the high-concentration group had significantly reduced body weight gain
compared to controls. The number of implants/dam and mean fetal body weight both were
significantly decreased with increasing exposure (although no information is provided on which
dose-level significance was first observed). In addition, fetuses in the high-concentration group
exhibited a significantly decreased incidence of sternal ossification. Based on decreased fetal
45
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
weight and skeletal alterations, EPA identified the developmental LOAEL as 14,750 mg/m3 and
the NOAEL as 7,375 mg/m3. Based on clinical signs of sedation (diminished response to noise
stimulus), the maternal LOAEL is 2,950 mg/m3 and the NOAEL is 1,475 mg/m3.
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
Several short-term oral studies in animals are available (see Appendix C for study
descriptions). In rats treated with a total of six gavage doses of THF in distilled water, increased
mortality was observed at doses >2,000 mg/kg (Stasenkova and Kochetkova, 1963). Toxicity
observed in this study included CNS toxicity (immobility, drowsiness, reduced response to
external stimuli) and necrosis, edema, and hemorrhage of stomach, brain, liver, heart, spleen, and
kidneys. However, it is not possible to more fully characterize the specific histopathology
endpoints in the study by Stasenkova and Kochetkova (1963). In a 4-week study of THF in
drinking water administered to rats (Komsta et al., 1988), doses as high as 96 mg/kg-day had no
effect on mortality and did not produce clinical signs of CNS toxicity in rats. Histopathologic
lesions in liver (increased cytoplasmic homogeneity and anisokaryosis) and kidney (tubular
cytoplasmic inclusions) were observed in the high-dose group males and females.
Several acute inhalation studies in animals suggest that the primary effects observed
following single exposures to THF, ranging from 30 minutes to several hours, are CNS toxicity
and respiratory tract irritation. Symptoms of CNS toxicity, including sedation, coma, altered
respiration, and decreased response to external stimuli, were observed in dogs (Stoughton and
Robbins, 1936), mice (Stasenkova and Kochetkova, 1963; Stoughton and Robbins, 1936), and
rats (Horiguchi et al., 1984; DuPont Haskell Laboratory, 1979; Stasenkova and Kochetkova,
1963). Clinical signs of respiratory tract irritation, observed only in studies in rats, included
scratching, head shaking, face washing, tearing, salivation, and bleeding from the nose
(Horiguchi et al., 1984; DuPont Haskell Laboratory, 1979). In addition, several other acute
studies observed structural or functional changes in respiratory tissue (suggesting respiratory
tract irritation), including congested mottled lungs in rats (Henderson and Smith, 1936), edema
and hemorrhage in lungs and bronchi of rats (Stasenkova and Kochetkova, 1963), and decreased
ciliary beat frequency and vacuolation/degeneration of both nasal mucosa (Ohashi et al., 1983)
and tracheal mucosa (Ikeoka et al., 1988) in rabbits, and nasal and tracheal histopathology
changes in rats (Horiguchi et al., 1984). Two studies report histopathological lesions in other
organs such as liver (Stasenkova and Kochetkova, 1963; Henderson and Smith, 1936), kidney,
brain, and spleen (Stasenkova and Kochetkova, 1963). However, Hofmann and Oettel (1954)
specifically examined the liver and kidney and found no effects. These studies are further
described in Appendix C.
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
4.5.	MECHANISTIC DATA AND OTHER STUDIES
Genotoxicity Studies
Only one study that evaluated genotoxicity endpoints in humans was identified. Funes-
Cravioto et al. (1977) reported increased chromosome breaks in peripheral lymphocytes from
solvent-exposed versus nonexposed adults. However, of the seven occupational groups that were
pooled for the statistical analysis, only one was identified as having used THF in the workplace
(no exposure information was provided by the study authors), thus suggesting that agents other
than THF likely played a greater role in the observed genotoxicity.
NTP (1998) presented the results of a battery of mutagenicity/genotoxicity tests of THF.
The in vitro tests included the Salmonella typhimurium bacterial mutagenicity assay (with and
without S9 microsomal activation), induction of sister chromatid exchange and chromosomal
aberrations in the Chinese hamster ovary cell system, and in vivo in mouse bone marrow cells.
Micronuclei frequency in peripheral blood erythrocytes following 14-day inhalation exposure of
mice to THF was also evaluated. NTP (1998) concluded that there was little evidence of
mutagenic activity, with most data determined to be conclusively negative.
In summary, the genotoxic potential of THF has been evaluated in a variety of in vitro
and in vivo assays. Nearly all the results are conclusively negative, with equivocal findings
reported in a small number of assays that have been conducted. The genotoxicity data are
summarized in Table C-5 and discussed in more detail Appendix C.2. Taken together, these data
support the conclusion that THF is not likely genotoxic.
4.6.	SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS
A summary and synthesis of the major noncancer effects observed following oral and
inhalation exposure to THF are described below. The modes of action for the noncancer effects
are not known; however, mechanistic data relating to the potential modes of action for the
noncancer effects are further described in Appendix C.
4.6.1. Oral
No human studies of THF following oral exposure are available and the oral database for
animal studies is limited. A summary of the effects observed in the subchronic oral studies is
presented in Table 4-9. In a one-generation reproductive toxicity study (BASF, 1994) of THF
administered to rats in drinking water, symptoms of general toxicity, including decreased food
(males) and water consumption (males and females) and increased kidney weight (males and
females), were observed in parental generation rats administered 8,000 ppm THF (795 mg/kg-
day for males and 890 mg/kg-day for females). At this concentration, male and female pups had
significantly decreased body weight gain compared with controls. A follow-up two-generation
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
reproductive toxicity study (BASF, 1996) of THF administered to rats in drinking water
demonstrated similar results as the one-generation study in the parental generation rats, including
decreased food consumption (F0 females, F0/F1 males), decreased water consumption (F0/F1
both sexes), decreased body weight (F0/F1 both sexes), and increased kidney weight (F0 both
sexes) at 9,000 ppm THF (714 mg/kg-day for F0 males, 788 mg/kg-day for F1 males,
835 mg/kg-day for F0 females, 898 mg/kg-day for F1 females). At these same concentrations,
the F1 and F2 pups had significantly reduced body weight gain compared with controls, and the
F2 pups also demonstrated delayed eye opening and increased incidence of sloped incisors
compared with controls (see Table 4-8). Histopathology examination on parental rats included
liver, kidney, reproductive organs, and digestive tract organs and demonstrated no observed
effects on these organs. Exposure at the high concentration of THF in drinking water may have a
subtle effect on male rat fertility/fecundity based on a 16% decrease in the mean number of
delivered F2 pups (not statistically significant but below the range of historical control values)
and a finding of one infertile F1 parental male rat in the high dose group. In both studies, no
effects were observed on any other reproductive parameters measured.
Some similar effects were noted in short-term studies, (Stasenkova and Kochetkova,
1963; Komsta et al., 1988, described in Appendix C). Increased mortality and effects including
CNS toxicity (immobility, drowsiness, reduced response to external stimuli), and necrosis,
edema, and hemorrhage of stomach, brain, liver, heart, spleen, and kidneys were observed in rats
administered THF in distilled water via gavage (Stasenkova and Kochetkova, 1963). Another
short-term study of lower doses of THF administered to rats in drinking water (Komsta et al.,
1988), had no effect on mortality and did not produce clinical signs of CNS toxicity.
Histopathologic lesions in liver (increased cytoplasmic homogeneity and anisokaryosis) and
kidney (tubular cytoplasmic inclusions) were observed in the male and female rats.
48
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-9. Summary of effects observed in drinking water toxicity studies with THFa
Study
Species,
number,
sex
Route, duration, doses
Observed effects
NOAEL (mg/kg-day)
LOAEL (mg/kg-day)
Comments
BASF (1996);
Hellwig et al.
(2002)
Wistar rat
25/sex/dose
Drinking water,
two-generation reproductive
0, 1,000, 3,000, 9,000 ppm
General toxicity (descreased
BW and increased kidney
weight in F0 and F1 adults,
decreased BW gain in F1/F2
pups,
delayed eye opening in male
F2 pups
F0 males: 268
F0 females: 322
F1 Adult males: 268
F1 Adult females: 362
F1 Pups: 381
F2 Pups: 385b
F0 males: 714
F0 females: 835
F1 Adult males: 788
F1 Adult females: 898
F1 Pups: 1,071
F2 Pups: 974b
Each generation
treated 70 days prior
to mating through
cohabitation,
gestation lactation
BASF (1994)
Wistar rat
10/sex/dose
Drinking water,
one-generation reproductive
0, 4,000, 8,000, 12,000 ppm
Increased kidney weight (F0
males—high dose, F0
females—mid dose)
Decreased pup BW (mid
dose)
503
546
890
960

aThe best value to use for estimating the corresponding doses (mg/kg-day) differs for each generation based on THF intake values over the relevant period of
exposure. For parental effects, average THF intakes over the entire study period are appropriate for use in assigning effect levels. For developmental effects,
the time-weighted average THF intake during the gestation and lactation periods of the parent females was used to estimate the effective dose. THF intake
estimates are shown in Table 4-5.
bTHF intake estimates corresponding to NOAEL and LOAEL estimates were calculated for a variety of effects and are presented in Table 4-8.
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
23
24
25
26
27
28
29
30
31
32
33
34
35
36
4.6.2. Inhalation
Although no epidemiological studies of THF have been conducted, several case studies in
humans illustrate the potential for health effects following inhalation exposure in an occupational
setting. In almost all of the cases, workers were exposed to THF through activities where THF
was present as a component of solvents or adhesives. In general, workers were exposed for a
period of a few weeks to a few months before symptoms were reported. Target organs in
humans appear to be the CNS, respiratory tract, liver, and kidney. Symptoms of CNS toxicity
included headache, dizziness, fatigue, loss of the sense of smell (Gamier et al., 1989; Emmett,
1976; Horiuchi et al., 1967), and convulsions following enfluran anesthesia in a worker exposed
to THF in the weeks prior to surgery (Juntunen et al., 1984). Symptoms of respiratory tract
irritation included cough, chest pain, rhinorrhea, and dyspnea (Gamier et al., 1989; Emmett,
1976). In three cases, liver enzymes (ALT, AST, and GGT) were elevated above normal values
(Gamier et al., 1989; Edling, 1982; Horiuchi et al., 1967), and in one case a liver biopsy revealed
fatty changes following THF exposure (Edling, 1982). In one study, hematological changes and
decreased white blood cell counts were reported in THF-exposed workers (Horiuchi et al., 1967).
In one case study, autoimmune glomerulonephritis was observed in a man who worked with
THF in adhesives for 9 years (Albrecht et al., 1987). The human case studies suggest that CNS
toxicity, respiratory tract irritation, and liver and kidney toxicity are the potential health effects
following inhalation exposure to THF. An uncertainty associated with all of the reported human
case studies is the fact that workers were exposed to other solvents and chemicals in addition to
THF, so it is not possible to conclusively attribute the observed effects to THF exposure alone.
In addition, in most cases quantitative estimates of exposure were not provided.
In animals, subchronic and chronic studies reported several systemic effects following
inhalation exposure to THF; a summary of these effects is presented in Table 4-10. Decreased
body weight has been observed in rats (Horiguchi et al., 1984; Kawata and Ito, 1984). Decreased
blood pressure was observed in dogs (BASF, 1938) and rats (Stasenkova and Kochetkova, 1963).
Altered hematological parameters were observed in rats (NTP, 1998; Horiguchi et al., 1984),
mice (NTP, 1998; Stasenkova and Kochetkova, 1963), and dogs (BASF, 1938). Following 14
weeks of inhalation exposure, rats of both sexes had significantly increased relative and absolute
liver weight (NTP, 1998). In the same study, mice of both sexes showed increased relative and
absolute liver weight (NTP, 1998). In addition, Horiguchi et al. (1984) observed increased
relative weights of liver. Changes in mice included liver centrilobular cytomegaly in both sexes
following 14 weeks of exposure to THF (NTP, 1998). Increased incidence of hepatocellular
necrosis was also observed in female mice in the 2-year inhalation study (NTP, 1998).
Longer-term inhalation exposure to THF appears to also result in symptoms of CNS
toxicity and respiratory tract irritation. In a subchronic neurotoxicity assay (DuPont Haskell
50
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
Laboratory, 1996b), the only effects observed were transient symptoms of CNS toxicity that
were not observed on mornings prior to the start of the weekly exposures. No permanent
neurotoxic effects were observed on motor activity or in an FOB. Altered brain catecholamine
levels were observed following 8 weeks of inhalation exposure (Kawata et al., 1986), and altered
EEGs were observed following i.p. injection (Marcus et al., 1976). While the clinical
significance of these findings is not clear in terms of assigning adverse effect levels for THF, the
observation that similar brain alterations are induced by the THF metabolites GBL and GHB
(NSF, 2003) suggests that these metabolites may be responsible for the observed neurotoxicity of
THF. In two subchronic studies, authors specifically note that symptoms of CNS toxicity (NTP,
1998; Horiguchi et al., 1984) appeared to moderate with continued exposures. Based on findings
in Elovaara et al. (1984) of decreased concentrations of THF in rat brain and fat tissues with
extended exposure, the authors of the NTP (1998) study considered it likely that the apparent
tolerance to the CNS effects may be due to stimulation by THF of its own metabolism. They
also concluded that it is not possible to ascertain whether the clinical findings of CNS toxicity
(narcosis) were primary (i.e., specific to THF or its metabolites) or secondary (i.e., nonspecific
due to solvent interaction with cell membranes of the nervous system as seen with other
solvents) and that further research is needed to better characterize THF neurotoxicity. However,
support for the THF-induced CNS effects was provided by evidence of these effects in the
subchronic and chronic studies as well as short-term and acute studies. Several acute inhalation
studies in animals suggest that one of the primary effects observed following single exposures to
THF, ranging from 30 minutes to several hours, is CNS toxicity. Symptoms of CNS toxicity,
including sedation, coma, altered respiration, and decreased response to external stimuli, were
observed in dogs (Stoughton and Robbins, 1936), mice (Stasenkova and Kochetkova, 1963;
Stoughton and Robbins, 1936), and rats (Horiguchi et al., 1984; DuPont Haskell Laboratory,
1979; Stasenkova and Kochetkova, 1963).
Additional effects observed include respiratory tract irritation, kidney effects, thymus
weight changes, and effects associated with immunotoxicity and developmental toxicity. The
respiratory effect study that identified the lowest adverse effect level was conducted by
Horiguchi et al. (1984), who reported that rats exposed to 100 ppm THF for 3 weeks had changes
in the nasal mucous membrane that were similar to those observed in the tracheal mucosa.
Changes in the tracheal mucosa in the group exposed to 5,000 ppm were described as occurring
in the cilia, with disorder of the epithelial architecture and darkening of cell bodies. However,
the study authors did not clarify whether the nasal effects at 100 ppm were the same as the
tracheal effects at 100 or 5,000 ppm, although it was presumed that it was the tracheal effects at
5,000 ppm that were being equated to the 100 ppm nasal effects. The authors did not describe
any results for the tracheal mucosa at 100 ppm. A major deficiency in this study is that the
51
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
results represent a single animal per exposure level at each time point. Based on the small
sample size, duration of exposure, absence of clear documentation of the severity of the nasal
histopathology, and uncertainty regarding the concentration at which nasal changes were
observed, this study provided equivocal results regarding respiratory toxicity. In addition,
Stasenkova and Kochetkova (1963) evaluated the effects of THF in mice and rats following 2
months of exposure and in rats following 6 months of exposure. After 2 months of exposure to
6,000-8,000 mg/m3 THF, mice had eye irritation while mice and rats displayed symptoms of
respiratory tract irritation and an increase (in mice) or decrease (in rats) in the threshold of
neuromuscular irritability. These symptoms were not reported in rats following 6 months of
exposure at 1,000-2,000 mg/m3.
Clinical signs of irritation as well as histopathological changes in the respiratory tract
were also observed in one subchronic study at 3,000 ppm (Kawata and Ito, 1984) and in several
acute and short-term studies (Ikeoka et al., 1988; Horiguchi et al., 1984; Ohashi et al., 1983;
Stasenkova and Kochetkova, 1963) with some at relatively low exposure concentrations.
Specifically, Horiguchi et al. (1984) found nasal histopathology after a 3-week exposure to 100
or 5,000 ppm THF but no such effects were reported following exposure to 5,000 ppm for 12
weeks. Also, symptoms of eye and respiratory tract irritation as well as changes in the threshold
of neuromuscular irritability were found in rats and mice following 2 months of exposure
(6,000-8,000 mg/m3), but similar symptoms were not reported following 6 months of exposure
to lower concentrations (1,000-2,000 mg/m3) (Stasenkova and Kochetkova, 1963). These data
demonstrate that the irritation effects induced by THF were not consistently observed with
increasing duration of exposure. These effects were observed at higher exposure concentrations
than those where liver effects were observed. In addition, there are limitations in documentation
and reporting.
Several acute inhalation studies in animals suggest that the primary effects observed
following single exposures to THF, ranging from 30 minutes to several hours Clinical signs of
respiratory tract irritation, observed only in studies in rats, included scratching, head shaking,
face washing, tearing, salivation, and bleeding from the nose (Horiguchi et al., 1984; DuPont
Haskell Laboratory, 1979). In addition, several other acute studies observed structural or
functional changes in respiratory tissue (suggesting respiratory tract irritation), including
congested mottled lungs in rats (Henderson and Smith, 1936), edema and hemorrhage in lungs
and bronchi of rats (Stasenkova and Kochetkova, 1963), and decreased ciliary beat frequency
and vacuolation/degeneration of both nasal mucosa (Ohashi et al., 1983) and tracheal mucosa
(Ikeoka et al., 1988) in rabbits, and nasal and tracheal histopathology changes in rats (Horiguchi
et al., 1984). Two studies report histopathological lesions in other organs such as liver
(Stasenkova and Kochetkova, 1963; Henderson and Smith, 1936), kidney, brain, and spleen
52
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
(Stasenkova and Kochetkova, 1963). However, Hofmann and Oettel (1954) specifically
examined the liver and kidney and found no effects.
Although the data indicate that THF induced an increase in kidney weight in rats, the
severity of the impact on the kidneys appears to be minimal. This conclusion is supported by
several considerations as discussed in detail in Section 4.3.1. THF exposure had no effect on
absolute or relative kidney weight in F1 generation adults. Furthermore, the kidney weight
changes that were observed in the F0 generation were not accompanied by gross kidney
pathology or hematology or clinical chemistry findings consistent with an effect on renal
function (in the one-generation study) or by histopathological examination (in the two-generation
study). Evaluation of the overall database for THF, including inhalation studies, does not
suggest that THF is a potent kidney toxicant. For example, most of the available human case
reports have not identified the kidney as a target of THF exposure. Furthermore, in the
subchronic and chronic inhalation NTP (1998) studies, changes in kidney weight or pathology
were not particularly sensitive to THF exposure.
As reported by NTP (1998), absolute and relative thymus weights were statistically
significantly decreased, beginning at 1,770 mg/m3 in male mice. The thymus weight changes
were not accompanied by histopathological changes in the subchronic study. The study authors
indicated that the significance of the thymus weight changes were unclear and suggested that
these changes might have been due to stress associated with THF administration. However, the
thymus weight changes were concentration-dependent, suggesting that if they were stress related,
this response would have been secondary to the effects of THF. Organ weights were not
reported for the chronic study, and therefore, it is not possible to determine if thymus weight is
similarly affected by long-term exposure. Histopathological analysis of the thymus in the
chronic study revealed an increase in the incidence of thymic atrophy that was statistically
significant in the 5,310 mg/m3 exposure group. This finding was attributed by the authors to be a
secondary response, based on the high incidence of urogenital inflammation observed in the
high-concentration males. However, since the increase in infections occurred in the same group
that had thymic changes, it cannot be determined whether the thymus weight and histopathology
effects increased susceptibility to infection or the inflammation had a stress-related effect on the
thymus.
It is unclear whether the observed effects on the thymus in the subchronic and chronic
studies (NTP, 1998) represent a functional effect on the immune system, and no data are
available to differentiate between mechanisms involving a generalized stress response versus
other mechanisms directly targeting the immune system. Evaluation of the THF database as a
whole provides inconsistent results related to immune effects, with some studies identifying
effects and others showing no effect. Nevertheless, some of the available studies show evidence
53
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
for potential immunotoxicity. For example, decreased white blood cell counts were reported in a
study of workers (Horiuchi et al., 1967) and changes in white blood cell counts were reported in
an oral drinking water study (Pozdnyakova, 1965) and in a subchronic inhalation study
(Horiguchi et al., 1984). Both thymus and spleen weights were reduced in male and female rats
in the subchronic NTP (1998) study. In addition, data for THF metabolites are consistent with
potential immunotoxicity. For example, thymic depletion was reported at 262 mg/kg-day GBL
in mice in a gavage study (NTP, 1992), although this may have been secondary to an
inflammatory response or a factor leading to the susceptibility to inflammation. The
pharmacokinetic information also provides a possible connection between THF exposure and
immune effects, in which the tissue distribution study by Kawata and Ito (1984) reported that the
thymus and spleen had significantly higher THF concentrations than other tissues following
inhalation exposure to 3,000 ppm THF for 12 weeks.
The predictivity of thymus weight changes for functional immune responses has been
studied by Luster et al. (1992) who determined the ability of a variety of common measures of
immune toxicity, including thymus/body weight ratios, to predict the immunotoxicity of a series
of test compounds in mice. When evaluated as a single measure, thymus/body weight ratios
were characterized as an unreliable indicator of immunotoxicity (68% concordance—the ability
to correctly identify compounds of known immunotoxic potential). However, thymus/body
weight ratio was part of several testing configurations that showed 100% concordance with
immunotoxicity, suggesting that this measure can contribute to the immunotoxicity assessment.
In addition, the authors noted that the lack of concordance for most assays was generally due to a
decreased sensitivity (i.e., failure to detect positive immunotoxicants) not a decrease in
specificity (i.e., the ability to correctly identify negative compounds). This suggests that
thymus/body weight ratios might underreport immunotoxicity. In a follow-up publication by
Luster et al. (1993), a good correlation was reported between immune function assays and
changes in host resistance (e.g., increased susceptibility to infection from a challenge agent),
although the predictivity of individual assays varied (the concordance was 76% for thymus/BW
ratios).
In summary, there are no studies of host resistance or data from other types of
immunotoxicity studies following inhalation exposure to THF. Also, it is unclear whether the
observed thymus weight changes had a functional impact on the immune function of mice in the
subchronic study (NTP, 1998). For this reason, the biological significance of the decrease in
thymus weight is questionable. An area of uncertainty exists for the potential effects of THF on
the immune system, specifically with regard to decreased thymus weight.
Developmental studies by the inhalation route have been conducted in both rats (Mast et
al., 1992; DuPont Haskell Laboratory, 1980) and mice (Mast et al., 1992). In both studies and
54
DRAFT - DO NOT CITE OR QUOTE
-------
1	both species, maternal toxicity included significant decreases in body weight accompanied by
2	decreases in gravid uterine weight (Mast et al., 1992) or food consumption (DuPont Haskell
3	Laboratory, 1980). Decreased fetal weight was observed at the same concentration that resulted
4	in maternal toxicity in rats (Mast et al., 1992). In both mice (Mast et al., 1992) and rats (DuPont
5	Haskell Laboratory, 1980), decreased fetal survival also occurred at the same concentrations that
6	resulted in maternal toxicity. With regard to potential teratogenic effects, Mast et al. (1992)
7	noted that in mice that survived the exposure period, no increase was observed in the incidence
8	of fetal abnormalities. However, an increased incidence of incomplete sternal ossification in rat
9	fetuses was observed (DuPont Haskell Laboratory, 1980).
10
55
DRAFT - DO NOT CITE OR QUOTE
-------
1
Table 4-10. Summary of findings in developmental, subchronic, and chronic inhalation studies with THF
Study
Species, sex, number,
concentration
(mg/m3)
Duration
Observed effects
NOAEL/LOAEL'
(mg/m3)
Comments
Developmental toxicity studies
Mast et al. (1992)
CD-I mice, female
(40/group)
0, 1,770, 5,310, 14,750
6	hours/day,
7	days/week,
gestation days 6-17
Decreased dam body
weight and gravid uterine
weight, decreased fetal
survival
1,770/5,310

Sprague-Dawley rat,
female (40/group)
0, 1,770, 5,310, 14,750
6	hours/day,
7	days/week,
gestation days 6-19
Decreased dam body
weight, decreased fetal
body weight
5,310/14,750

DuPont Haskell
Laboratory (1980)
Crl:CD BR rat, female
(29/group)
0, 590, 1,475, 2,950,
7,375, 14,750
6	hours/day,
7	days/week,
gestation days 6-15
Dams: CNS clinical
signs
Fetal: decreased fetal
weight, skeletal
alterations
Dams: 1,475/2,950
Fetal: 7,375/14,750

Subchronic studies
BASF (1938)
Dog, sex and strain not
specified (four/group)
590 mg/m3:
6 hours/day,
5	days/week, 9 weeks
then
1,080 mg/m3:
6	hours/day,
5 days week, 3 weeks
Decreased blood
pressure
NAb/590
No microscopic pathology noted in
heart, lungs, spleen, pancreas, or
kidneys
Horiguchi et al.
(1984)
Sprague-Dawley rat,
male (11-12/group)
0, 295, 590, 2,950,
14,750
4	hours/day,
5	days/week,
12 weeks
Body and organ weight
changes, altered serum
chemistry
2,950/14,750

56
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-10. Summary of findings in developmental, subchronic, and chronic inhalation studies with THF
Study
Species, sex, number,
concentration
(mg/m3)
Duration
Observed effects
NOAEL/LOAEL'
(mg/m3)
Comments
Kawata and Ito
(1984)
Wistar rat, male
(25/group)
0, 8,850
1 hour/day,
5 days/week,
12 weeks
Decreased body weight,
papillary hyperplasia in
lung and bronchial
epithelium, protein
casts/hyaline in kidney
NA/8,850
No information given on incidence
of histopathologic lesions or
statistical significance
DuPont Haskell
Laboratory (1996b);
Malley et al. (2001)
Crl:CD BR rat (12-
18/sex/group)
0, 1,475, 4,425, 8,850
6 hours/day,
5 days/week,
13-14 weeks
CNS clinical signs
1,475/4,425
This was a subchronic neurotoxicity
study. No other neurotoxic effects
were observed (i.e., FOB, motor
activity, or neuropathology)
NTP (1998)
F344/N rat
(10/sex/group) 0, 195,
590, 1,770, 5,310,
14,750
6 hours/day,
5 days/week,
90 days
CNS clinical signs,
organ weight changes,
hematological effects
5,310/14,750

B6C3Fi mouse
(10/sex/group)
0, 195, 590, 1,770,
5,310, 14,750
6 hours/day,
5 days/week,
90 days
CNS clinical signs,
increased liver weight
1,770/5,310
Decreased thymus weight at lower
concentrations and histopathology
of the liver, uterus, adrenal gland
only at the high concentration
Stasenkova and
Kochetkova (1963)
Rat, male, strain not
specified (20/group)
1,000-2,000
4 hours/day,
7 days/week,
6 months
Decreased blood
pressure,
increased leukocyte
count, hypertrophy of
muscle fibers in bronchi
walls and spleen
NA/NA
Air concentration reported as a
range; study judged as not suitable
for dose-response assessment
Chronic studies
NTP (1998)
F344/N rat
(50/sex/group)
0, 590, 1,770, 5,310
6 hours/day,
5 days/week,
2 years
No noncancer effects
observed
5,310/NA

57	DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-10. Summary of findings in developmental, subchronic, and chronic inhalation studies with THF
Study
Species, sex, number,
concentration
(mg/m3)
Duration
Observed effects
NOAEL/LOAEL'
(mg/m3)
Comments

B6C3Fi mouse
(50/sex/group)
0, 590, 1,770, 5,310
6 hours/day,
5 days/week,
2 years
CNS clinical signs
(males); increased liver
necrosis (females)
1,770/5,310
Decreased survival, urogenital tract
inflammation and histopathology
lesions in bone marrow, lymph
nodes, spleen, thymus attributed to
infection secondary to observed
narcosis
1
2	aNOAEL/LOAEL from the study concentrations.
3	bNA indicates that the NOAEL or LOAEL was not identified.
4
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
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 THF provides "suggestive evidence of carcinogenic potential." No human data are
available to assess the carcinogenic potential of THF. A 2-year NTP (1998) inhalation cancer
bioassay reported a marginally increased incidence of renal tubule adenomas and carcinomas in
male F344/N rats (statistically significant exposure-response trend) and an increased incidence of
hepatocellular adenomas and carcinomas in female B6C3Fi mice (statistically significant trend
and increase incidence at the highest concentration tested) following inhalation exposure. No
other treatment-related increases in tumor incidence were observed. NTP (1998) concluded that
the data provided some evidence for THF carcinogenicity in male rats (renal tubular adenomas
and carcinomas) and clear evidence of carcinogenicity in female mice (hepatocellular adenomas
and carcinomas). There was no evidence of carcinogenic activity in female rats. Likewise, in
male mice there was no evidence of carcinogenicity reported by NTP (1998).
There are some data suggesting that the observed renal tumors in the male rats may be
secondary to a,2u-globulin accumulation. A review of the data available for THF indicates that
the data do not support an a2u-globulin-related MOA (Section 4.7.3.1). Another consideration
regarding the renal tumors is the possibility that advanced chronic progressive nephropathy
(CPN) may play a role in the incidence of atypical tubule hyperplasia (ATH) and perhaps the
THF-induced kidney tumors in male rat kidneys). CPN is an age-related renal disease of
laboratory rodents that occurs spontaneously. There was no difference in the incidence or
severity of CPN in the control versus treated male rats of the NTP 2-year carcinogenicity study
on THF. Therefore, although THF did not exacerbate development of CPN, it is possible that it
may have exacerbated the development of proliferative lesions within CPN-affected tissue;
however, there is no direct evidence in support of this. Thus, the kidney tumors observed in
male rats are considered relevant to the assessment of the carcinogenic potential of THF to
humans.
For the liver tumors in mice, some mechanistic data suggest that THF may induce cell
proliferation and lead to a promotion in the growth of pre-initiated cells. However, key
precursor events linked to observed cell proliferation have not been clearly identified and the
available data are insufficient to establish a mode of action for the THF liver tumor induction
(Section 4.7.3.2). Thus, the liver tumors observed in female mice are considered relevant to the
assessment of the carcinogenic potential of THF to humans.
U.S. EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) indicate that
for tumors occurring at a site other than the initial point of contact, the weight of evidence for
carcinogenic potential may apply to all routes of exposure that have not been adequately tested at
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
sufficient doses. An exception occurs when there is convincing information, e.g.,
pharmacokinetic data that absorption does not occur by another route. Information available on
the carcinogenic effects of THF via the inhalation route demonstrates that tumors occur in tissues
remote from the site of absorption. Information on the carcinogenic effects of THF via the oral
and dermal routes in humans or animals is not available. Based on the observance of systemic
tumors following inhalation exposure, and in the absence of information to indicate otherwise, it
is assumed that an internal dose will be achieved regardless of the route of exposure. Therefore,
there is "suggestive evidence of carcinogenic potential" following exposure to THF by all routes
of exposure.
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence
As discussed in Section 4.1, there are no human studies investigating the carcinogenic
effects of THF following inhalation exposure. However, the NTP (1998) chronic inhalation
exposure bioassay in laboratory animals was adequately designed to assess the carcinogenic
potential of lifetime inhalation exposure to THF. This study involved exposure of F344/N rats
(50/sex/group) and B6C3Fi mice (50/sex/group) to 0, 200, 600, and 1,800 ppm (0, 590, 1,770,
and 5,310 mg/m3) THF for 6 hours/day, 5 days/week for 105 weeks. For the male rats, a
statistically significant treatment-related trend was observed for combined incidences of renal
tubular epithelial adenomas or carcinomas (1/50, 1/50, 4/50, and 5/50) (NTP, 1998). The
response was predominantly benign except for two carcinomas present at the high exposure
concentration. The individual incidences of the kidney adenomas or carcinomas in the high
exposure male rats appeared to exceed the incidence of these tumors in F344/N historical
controls (rate: 0.9 ± 1.3%; range: 0-4%) but were not statistically significant when compared
with the concurrent controls (NTP, 1998).
In female mice there was a statistically significant increased incidences of hepatocellular
adenomas or carcinomas at the high concentration (1,800 ppm) and a positive trend for these
hepatocellular neoplasms across exposure to 200, 600, and 1,800 ppm THF compared with
controls (17/50, 24/50, 26/50, and 41/48) (NTP, 1998). The females also showed a statistically
significant positive trend in hepatocellular carcinomas (albeit not a significantly increased
incidence; 6/50, 10/50, 10/50, and 16/48). There was no statistically significantly increased
incidence of hepatocellular adenomas or carcinomas in male mice (35/50, 31/50, 30/50, and
18/50), even after adjustment for differential survival.
A 2-year cancer bioassay by the oral route has been conducted for the THF metabolite
GBL (NTP, 1992), which showed no evidence of carcinogenicity in rats (male and female) or
female mice, although an increased incidence of adrenal medulla pheochromocytomas and
hyperplasia were observed. The authors concluded that there was equivocal evidence of
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
carcinogenic potential. Mechanistic studies for THF following exposure by the inhalation route
also suggest that THF itself rather than a metabolite might be responsible for the observed liver
and kidney responses. Based on mode of action data and the difference in tumor responses for
THF and GBL in NTP (1998, 1992) bioassays, EPA concluded that the cancer bioassay data for
THF metabolites were not relevant for the assessment of THF carcinogenicity in humans.
As discussed in Section 4.5, results from genotoxicity studies for THF are mostly
negative and provide very limited evidence to suggest a genotoxic mode of action. All bacterial
mutation assays were negative for THF genotoxicity. In vitro genotoxicity assays with
eukaryotic cells also proved to be negative with the exception of a slight increase in
chromosomal aberrations in Chinese hamster ovary cells with metabolic activation (Galloway et
al., 1987). In vivo studies suggest that THF is not likely to be mutagenic; however, studies have
not been conducted in target tissues.
4.7.3. Mode of Action Information
Both renal and hepatocellular adenomas and carcinomas are observed following
inhalation exposure to THF (NTP, 1998). There are mechanistic data suggesting that the
induction of kidney tumors in male rats and liver tumors in female mice may involve the
accumulation of a2u-globulin in the kidney and increased cell proliferation in the liver,
respectively. However, an analysis of the data as outlined below indicates that there is
insufficient evidence to establish the roles of a2U-globulin in THF-induced kidney tumors or cell
proliferation in THF-induced liver tumors. THF is not likely to be genotoxic, as the results of
the mutagenicity tests conducted by NTP (1998) provide little evidence of mutagenic activity,
with most data determined to be conclusively negative (Section 4.5). Therefore, the mode of
carcinogenic action of THF has not been established.
4.7.3.1. Kidney Tumors
Description of the Hypothesized Mode of Action
Hypothesized mode of action
Generally, kidney tumors observed in cancer bioassays in laboratory animals are assumed
to be relevant to humans. However, a number of chemicals have been shown to induce renal
tumors as a result of accumulation of a2U-globulin in hyaline droplets. This accumulation
initiates a sequence of events that leads to renal nephropathy and, eventually, renal tubular tumor
formation. The phenomenon is unique to the male rat since female rats and other laboratory
mammals administered the same chemicals do not accumulate a2U-globulin in the kidney and do
not subsequently develop renal tubule tumors (Doi et al., 2007; IARC, 1999; U.S. EPA, 1991b).
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
Some experimental data suggest that the development of kidney tumors in male rats following
exposure to THF may involve an a2U-globulin-mediated mode of action. An analysis of the data
is outlined below.
Identification of key events
For chemicals inducing kidney tumors in male rats involving the a2u-globulin
accumulation mode of action, the following events occur after binding of the chemicals or their
metabolites specifically, but reversibly, to a,2u-globulin:
•	Increased number and size of hyaline droplets in renal proximal tubule cells of treated
male rats
•	Accumulation of hyaline droplets containing a,2U-globulin in renal proximal tubules due
to the resistance of the a2u-globulin chemical complex to hydrolytic degradation by
lysosomal enzymes
•	Induction of typical pathological lesions associated with a,2U-globulin nephropathy (e.g.,
single-cell necrosis, exfoliation of epithelial cells into the proximal tubular lumen,
formation of granular casts, linear mineralization of papillary tubules, and tubule
hyperplasia).
Experimental Support for the Hypothesized Mode of Action
Strength, consistency, specificity of association
Chhabra et al. (1998) published a summary of the NTP (1998) bioassay and presented
data on the accumulation of a,2u-globulin (as indicated by protein droplets) in male rat kidney
following 13 weeks of exposure to 1,800 ppm THF. The NTP reported qualitative differences in
the appearance of protein droplets of the kidneys of control versus male rats exposed to
1,800 ppm THF. Differences in the appearance and location of protein droplets in the male rat
kidneys for control and high-concentration rats were noted. Protein droplets were described as
finer and more densely and diffusely distributed in tubular epithelial cells in the outer cortex for
control rats. In the high-concentration rats, protein droplets were characterized as coarser and
concentrated in scattered foci in the outer cortex. However, the average severity grades for the
accumulation of protein droplets did not differ and no other differences in the incidence of
nonneoplastic lesions in the male rat kidneys were observed. Therefore, no clear signs of
treatment-related pathological lesions in the kidney were found in the NTP (1998) study
(Chhabra et al., 1998).
BASF (1998) reevaluated kidney tissues of male rats to examine the relationship between
cell proliferation responses and an increase in kidney tumors following THF administration in
the NTP (1998) study. Histopathological examination and evaluation of cell proliferation as
measured by PCNA staining was conducted for tissue samples from the 0, 200, 600, and 1,800
ppm (0, 590, 1,770, and 5,310 mg/m3) exposure groups (10/group) from the NTP (1998)
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
subchronic (13 weeks) study. Kidney tissues from the cortex, outer stripe of the outer medulla,
inner stripe of the outer medulla, and the inner medulla were evaluated separately. The
histopathological examination revealed an increased incidence of moderate grade hyaline droplet
accumulation in the male rat kidney tissues of the high-concentration group as compared with
controls, but these changes were not accompanied by evidence of cell degeneration. No increase
in cell proliferation was found in any of the individual kidney compartments or in evaluation of
all compartments combined. Cell proliferation index was statistically significantly decreased in
individual kidney compartments, although these changes did not show a concentration-dependent
pattern. No other differences among controls and exposure groups were noted.
Gamer et al. (2002; BASF, 2001a) conducted a series of mode of action studies for
kidney effects in male F344 rats (6/group) at similar THF-exposure concentrations to those that
were used in the NTP (1998) cancer bioassay. Endpoints, including a,2U-globulin accumulation,
cell proliferation, and apoptosis, were evaluated. Animals were placed in one of three groups
that were exposed 6 hours/day for either 5 consecutive days, 5 consecutive days followed by a
21-day observation period, or 20 consecutive days over a period of approximately 28 days. Test
animals were exposed nose-only to average THF concentrations of 0, 598, 1,811, or 5,382 mg/m3
(0, 199, 604, or 1,794 ppm), corresponding to the concentrations used in the NTP (1998) cancer
bioassay. For the animals in each of the four concentration groups, a full necropsy was done,
including histopathological evaluation of the kidney. Additional evaluations in these same
organs included measurements of cell proliferation (S-phase response by BrdU staining) and
TUNEL apoptosis assay.
Results of the study (Gamer et al., 2002; BASF, 2001a) provide some evidence for
a,2U-globulin accumulation. Specifically, THF exposure induced a,2u-globulin accumulation in
male rats in an exposure-related manner (see Table C-3) after 5- or 20-day exposures (6
hours/day). The accumulation of a,2U-globulin as measured by immunohistochemistry was
supported by histopathological evaluation of hyaline droplets in the kidneys of control and high-
concentration animals exposed to THF for 20 days. The incidence of proximal tubule cells with
grade 2 (slightly increased) staining for hyaline droplets (putatively a2u-globulin) was 5/6 for
exposed animals versus 1/6 for controls. The study also showed that focal areas of a,2U-globulin
accumulation corresponded to areas of increased cell proliferation. Although no significant
increase in labeling index in the renal cortex was determined by standard assessment methods,
focal areas of increased BrdU labeling were noted. Quantitation of these areas revealed
increased cell proliferation in subcapsular proximal tubules (cortex 1) in animals exposed to THF
at the mid and high concentrations for 20 days and at the high concentration for 5 consecutive
days. No increase in labeling was observed in the groups given a 21-day recovery period. An
increase in cell proliferation was also noted in the proximal tubules between the outer stripe of
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
the outer medulla and the subcapsular layer (cortex 2) at the highest concentration following 20
exposures. The number of cells undergoing apoptosis was significantly increased in the high-
concentration groups exposed for 5 days and observed for 21 days or after 20 exposure days.
Marginal increases were observed in the mid-concentration groups for these two exposure
regimens, but the results were not statistically significant (see Table C-3).
Kawata and Ito (1984) reported protein casts and hyaline droplets in the kidneys of THF-
exposed male Wistar rats. No other details of the study are available.
Dose-response concordance
THF exposure induced a2u-globulin accumulation in male rats treated under all three
exposure regimens in the study by Gamer and coworkers (Gamer et al., 2002; BASF, 2001a).
Increases were generally concentration related, with increases at the high concentration ranging
from 175 to 280% of control levels for cortex 1 and from 188 to 324% of control levels for
cortex 2 among the three exposure regimens. When the whole cortex was used as the labeled
area for the analysis, accumulation was significantly elevated beginning at the low concentration
and following 5 consecutive days or 20 days of exposure. Maximum effects observed at the high
concentration ranged from 178 to 299% of controls among the three exposure regimens.
Increased cell proliferation and apoptosis in kidneys of animals exposed to THF for 20 days also
appeared to show a dose-response relationship (see Table C-3).
Temporal relationship
The mode of action data were obtained from short-term exposures (5 or 20 days) of THF.
Except for some qualitative differences in the appearance of protein droplets of the kidneys of
control versus male rats exposed to 1,800 ppm THF, no clear signs of treatment-related
pathological lesions in the kidney were found in the 2-year bioassay of NTP (1998). No increase
in cell proliferation was found in any of the kidney compartments in the 13-week study of BASF
(1998). Therefore, a temporal relationship of the key events to male rat kidney tumor induction
cannot be established.
Biological plausibility and coherence
The concordance between a,2u-globulin accumulation, cell proliferation, and induction of
apoptosis in the renal cortex with exposure concentrations that induced kidney tumors in the
cancer bioassay, lends support to the involvement of these mechanisms in THF-induced rat
kidney tumors. However, no increase in renal tubule hyperplasia or mineralization was observed
in the NTP (1998) study. The detection of a,2u-globulin accumulation only when sensitive
detection methods were used (i.e., immunohistochemical staining as opposed to standard staining
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
for histopathological examination) suggests that the responses are weak (Chhabra et al., 1998;
NTP, 1998). Furthermore, the observed cell proliferation response, which was increased to a
maximum of 298% of controls when selected for focal areas of proliferation, was minimal as
compared with cell proliferation responses induced by other well-characterized inducers of
a,2U-globulin accumulation (Gamer et al., 2002; BASF, 2001a; U.S. EPA, 1991b). There is also
an uncertainty regarding the specificity of the relationship between cell proliferation (a putative
tumor precursor event) and the observed a,2u-globulin accumulation, since the mode of action
study by Gamer and colleagues (Gamer et al., 2002; BASF, 2001a) did not include a similar
analysis of cell proliferation in female rat kidneys. A major area of uncertainty arises from the
absence of detectable histopathological lesions characteristic of this mode of action. No
treatment-related renal histopathology or hyaline or granular casts were noted in the BASF study
(Gamer, et al., 2002; BASF, 2001a). Because of the weak response in a,2u-globulin accumulation
and cell proliferation and the absence of the detectable pathological findings, the evidence for
this mode of action is equivocal.
Other Possible Modes of Action
It is possible that advanced CPN may play a role in kidney toxicity and perhaps THF-
induced kidney tumors in male rat kidneys. Accelerated tubular cell degeneration and
regeneration associated with CPN could be involved in the development of proliferative lesions
observed in the kidneys of THF-exposed rats. These slight increases in cell proliferation may
have contributed to the development of adenomas in male rats exposed to the high dose in the
chronic cancer bioassay.
CPN is an age-related renal disease of laboratory rodents that occurs spontaneously and
generally with high incidence. In a study aimed at discriminating lesions common to advanced
CPN from those that are precursors of renal tubule neoplasia, namely atypical tubule hyperplasia
(ATH), several archived NTP carcinogenicity studies, including THF, were evaluated (Hard and
Seely, 2005). ATH is designated as renal tubule hyperplasia in NTP technical reports. Hard and
Seely (2005) reported foci of ATH were considered synonymous with renal tubule hyperplasia in
NTP reports (Hard and Seely, 2005). Likewise, a comprehensive review that analyzed renal
tubule findings from NTP/National Cancer Institute (NCI) bioassays for 69 chemicals, including
THF, stated that while the NTP criteria differ in descriptive detail from those of the Society of
Toxicologic Pathology, in practice, the actual diagnoses of atypical (focal) tubule hyperplasia,
adenoma, and carcinoma are usually in accord (Hard et al., 1995; Lock and Hard, 2004).
Additionally, in a study that examined the utility of multiple-section kidney sampling in the
histopathologic evaluation of several NTP bioassays, renal tubule hyperplasia, also termed in the
same study as focal renal tubule hyperplasia or focal hyperplasia was differentiated as a
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
potentially preneoplastic lesion that is distinguished from the background regenerative changes
of the tubule epithelium that accompany renal toxicity or the common age-related degenerative
diseases of kidney in rats and mice (Eustis et al., 1994). In the same study by Eustis et al.
(1994), focal hyperplasia, adenoma, and carcinoma of the renal tubule were considered to
constitute a morphological continuum in the development and progression of neoplasia, and that
other hyperplastic lesions, specifically focal oncocytic hyperplasia and oncocytoma, were not
combined with rat renal tubule hyperplasia because their histogenesis were considered uncertain.
The Society of Toxicologic Pathology Hyperplasia Working Group evaluated the
contribution of hyperplastic lesions in two-year rodent carcinogenicity studies to human hazard
identification and risk assessment (Boorman et al., 2003). While acknowledging that ATH is
generally considered a preneoplastic lesion, the Society of Toxicologic Pathology asserted that
the appearance of neoplasms is the only conclusive evidence of a carcinogenic response and that
qualitative evaluation of hyperplastic lesions is more appropriate than statistical analysis. It is
not appropriate to combine hyperplastic and neoplastic lesions for statistical analysis (Boorman
et al., 2003). Additionally, in a comprehensive review that analyzed renal tubule tumor findings
in the NTP/NCI carcinogenicity bioassay database covering 69 chemicals, including THF, the
incidences of renal tubule tumors were separated from the findings of renal tubule hyperplasia
(ATH) although consideration was given, in a qualitative sense, to supporting information from
hyperplasia (Lock and Hard (2004).
There was no difference in the incidence or severity of CPN in male rats in the NTP
(1998) 2-year carcinogenicity study of THF (both the control and high-exposure groups had the
same incidence of end-stage renal CPN). Specifically, against a background of nephropathy that
was uniform across all groups, there were more renal tubular tumors in treated rats than in the
controls, and those in the higher doses were larger in size (NTP, 1998). Although THF did not
exacerbate development of CPN, it was postulated that it may have exacerbated the development
of proliferative lesions within CPN-affected tissue. Taken together, the data are equivocal.
Conclusions about the Hypothesized Modes of Action
Generally, kidney tumors observed in cancer bioassays are assumed to be relevant for
assessment of human carcinogenic potential. However, for male rat kidney tumors, when the
mode of action evidence convincingly demonstrates that the response is secondary to
a,2U-globulin accumulation, the tumor data are not used in the cancer assessment (U.S. EPA,
1991b). There are some data suggesting that male rat kidney tumors, following the inhalation
exposure observed in the NTP (1998) bioassay, may be due to the accumulation of a,2U-globulin.
The criteria for demonstrating this mode of action for risk assessment purposes have been
described (U.S. EPA, 1991b). Three core criteria are considered to be most important: (1)
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
increase in hyaline droplets in the renal proximal tubule cells; (2) determination that the
accumulating protein in the droplets is a2u-globulin; and (3) presence of additional pathological
lesions associated with a,2U-globulin. Review of the mode of action data of THF indicates that
criteria (1) and (2) are met but criterion (3) is not. An area of uncertainty is the absence of
detectable histopathological lesions characteristic of this mode of action (BASF, 2001a; NTP,
1998). The specificity of the response is also difficult to ascertain in the absence of an
evaluation of potential a,2u-globulin accumulation or other potential precursor events (e.g., cell
proliferation) in female rats. However, no increased incidence of kidney tumors was observed in
female rats in the NTP (1998) study. Thus, the mode of carcinogenic action of THF-induced
renal tumors has not been established.
4.7.3.2. Liver Tumors
Description of the Hypothesized Mode of Action
Hypothesized mode of action
Induction of a cell proliferation response in the liver by chemicals is generally considered
a possible mode of action for liver tumorigenesis that can occur in rodents. Sustained increase in
cell proliferation may lead to the promotion of growth of preinitiated cells and subsequently to
tumorigenesis. Changes in cellular apoptosis rates can also impact the net rate of tissue growth.
Key events for this mode of action may include histopathological evidence of
cytotoxicity/necrosis, regenerative growth, and/or apoptosis. Some experimental data suggest
that the development of liver tumors in female mice following exposure to THF may involve a
cell proliferation-related mode of action. An analysis of the data is outlined below.
Experimental Support for the Hypothesized Mode of Action
Strength, consistency, specificity of association
BASF (1998) evaluated the liver tissues from female mice from the NTP (1998) study to
examine the relationship between cell proliferation responses and increase in tumors observed in
these tissues following THF administration. Histopathological examination and evaluation of
cell proliferation as measured by PCNA staining was conducted for tissue samples from the 0,
200, 600, and 1,800 ppm (0, 590, 1,770, and 5,310 mg/m3) exposure groups (10/group) from the
NTP (1998) subchronic (13 weeks) study. No treatment-related histopathology was observed in
the female mouse liver tissues. The cell proliferation index was increased (39% over controls) in
tissues from the high-concentration mice. However, this result was not statistically significant
and was noted as being predominantly based on the results from 2/10 animals. Furthermore, no
clear concentration-response pattern was observed, and a significant decrease in proliferation
index was observed in the mid-concentration group. Based on these results, the study authors
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
concluded that the examination of the tissues from the 13-week NTP (1998) study revealed no
clear increase in cell replication that can be correlated to a tumorigenic mechanism. Gamer and
colleagues (Gamer et al., 2002; BASF, 2001a) evaluated a series of endpoints in female B6C3Fi
mice (10/group plus 5 in the control and high-concentration enzyme assays) in liver tissues in a
short-term, repeated exposure study. Test animals were exposed nose only to average THF
concentrations of 0, 598, 1,811, or 5,382 mg/m3 (0, 199, 604, or 1,794 ppm), corresponding to
the concentrations used in the NTP (1998) cancer bioassay. Concentrations adjusted for
continuous exposure were 0, 107, 323, or 961 mg/m3. For the animals in each of the four
concentration groups, a full necropsy was done, including histopathological evaluation of the
liver in addition to measurements of cell proliferation (S-phase response by BrdU staining) and
TUNEL apoptosis assay in the same organ. Since chemical exposures can have varying affects
in different regions of the liver lobule, cell proliferation was evaluated separately for zone 1
(periportal, the region adjacent to the portal triad), zone 3 (centrilobular, the region adjacent to
the central vein), and zone 2 (midzonal, the area of the lobule intermediate between zones 1 and
3).
THF exposure appeared to induce cell proliferation (see Table C-4) in the female mouse
liver. Increased cell proliferation was observed in zones 2 and 3 of the liver of the high-exposure
mice following THF exposure for 5 days and in zone 3 following 20 exposures. Coincident with
the increase in BrdU labeling, the mitotic index was increased in zone 3 after 5 or 20 exposures
in the high-concentration groups. No concentration-dependent increase in BrdU labeling was
observed in the animals given a 21-day recovery period, suggesting that the increases in cell
proliferation may be an adaptive effect. No treatment-related change in the number of liver cells
undergoing apoptosis was observed.
Dose-response concordance
Gamer and colleagues (Gamer et al., 2002; BASF, 2001a) reported increased cell
proliferation following short-term inhalation exposures at concentrations corresponding to those
that were tumorigenic in the NTP (1998) bioassay. Therefore, this event appeared consistent
with the expected dose response as compared to the tumor outcome.
Temporal relationship
Gamer and colleagues (Gamer et al., 2002; BASF, 2001a) reported increased cell
proliferation in the liver of the high-exposure female mice following short-term inhalation
exposures (5 or 20 days) of THF. However, no concentration-dependent increase in BrdU
labeling was observed in the animals given a 21-day recovery period.
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
Biological plausibility and coherence
Although increased cell proliferation was noted in short-term mode of action studies, the
data are not adequate to identify key events that precede this effect. In the earlier of these two
mode of action studies (Gamer et al., 2002; BASF, 2001a) it was not clear if the lower degree of
BrdU staining after 20 exposures as compared to 5 exposures (see Table C-3) represented
fluctuation around an average increase in cell proliferation or a decrease in the rate of
proliferation with continued exposure. While the observation that the mitotic index did not
similarly decrease after 20 exposures supports the former conclusion, the absence of a significant
increase in cell proliferation in tissues obtained from the subchronic NTP (1998) study as
reported by BASF (1998) suggests that cell proliferation might not be a sustained response even
with continued dosing and fails to explain the late onset of tumors. In the NTP (1998) bioassay,
no clear concentration-dependent increase in necrosis was observed, although the incidence of
necrosis appeared slightly elevated at the high concentration. Gamer and colleagues (Gamer et
al., 2002; BASF, 2001a) reported no histopathological evidence of cell degeneration at
concentrations that induced cell proliferation. Other in vitro studies did not suggest that THF is
cytotoxic (Matthews et al., 1993; Dierickx, 1989; Curvall et al., 1984). Taken together, these
data indicate that THF-induced cell proliferation is not secondary to regenerative hyperplasia.
Changes in cellular apoptosis rates can also impact the net rate of tissue growth. However, the
single study that evaluated this endpoint (Gamer et al., 2002; BASF, 2001a) suggested that THF
exposure has little impact on apoptosis in the livers of female mice. Therefore, the available data
are not sufficient to determine key events associated with cell proliferation that would likely be
involved in carcinogenesis.
Other Possible Modes of Action
One possible mode of carcinogenic action of THF is the ability of THF to inhibit GJIC.
In a study by Chen et al. (1984), co-cultures of 6-thioguanine-sensitive and resistant Chinese
hamster V79 fibroblast cells were treated with THF, and the degree of metabolic cooperation
was determined by the survival of the resistant cells. The killing of resistant cells serves as an
indicator of metabolic cooperation because the toxic 6-thioguanine metabolite that is formed
only in the sensitive cells can be passed on to normally resistant cells when gap junctions are
intact. Therefore, robust growth of the resistant cells in this assay system would suggest that
GJIC is inhibited. THF was judged to be positive (as defined by at least a doubling in recovery
of resistant colonies) in the metabolic cooperation assays, suggesting that THF can inhibit GJIC.
The recovery rate of resistant cells increased with increasing concentration (up to 100 [xL of
THF 5 mL of medium). Although there appears to be a correlation between inhibition of GJIC
and mouse liver carcinogenesis by some nongenotoxic carcinogens, the mechanism is unclear
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
(Klaunig et al., 1998). The data on GJIC presented by Chen et al. (1984) are too limited to
establish that this is the mode of action for the liver tumor induction of THF.
As the major metabolite of THF, GHB, can be converted to GAB A, and it has been
hypothesized that the production of GABA from THF may perturb the cellular level of
putrescine (1,4-diaminobutane), since putrescine is the primary source of GABA in many tissues.
Putrescine is required for proper functioning of the cell cycle and for cell growth (Lopez et al.,
1999) and has been shown to induce cell transformation and stimulate the expression of c-fos, a
proto-oncogene (Tabib and Bachrach, 1999). Therefore, it is possible that THF exposure would
increase tissue levels of GABA and putrescine, which in turn might promote cell growth and
carcinogenesis. However, the link between GABA and putrescine has not been investigated.
While this mode of action provides a possible basis for THF-induced cell proliferation and
subsequent carcinogenesis, it has not been investigated directly for THF.
Conclusions about the Hypothesized Mode of Action
Although increased cell proliferation was noted in short-term studies, the data are not
adequate to support the hypothesized mode of action. The absence of a significant increase in
cell proliferation in tissues obtained from the subchronic NTP (1998) study suggests that cell
proliferation might not be a sustained response even with continued dosing. Therefore, while the
cell proliferation event meets the requirement of showing the expected temporal relationship at
early time points, it is not clear that the effect is sustained for a sufficient duration to adequately
explain the late onset of tumors. Furthermore, key precursor events linked to observed cell
proliferation have not been identified. The data on other potential modes of action are too
limited to establish a mode of action for the THF-induced liver tumors.
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.8.1. Possible Childhood Susceptibility
No adequate studies on the potential reproductive or developmental toxicity of THF in
humans were available. However, these endpoints have been evaluated following oral and
inhalation exposures to THF in animal studies and oral studies with THF metabolites. A one-
generation screening assay (BASF, 1994) and a more comprehensive two-generation assay
(BASF, 1996) were conducted for THF administered in the drinking water of rats. Decreased
BWs in both male and female pups and delayed eye opening and increased incidence of sloped
incisors in F2 pups were observed. There are no data that indicate why developmental delays in
eye opening are observed in male pups but not female pups. These developmental effects were
observed at doses that also induced maternal effects (although the maternal effects were only of
minimal severity). For the THF metabolite GBL, no maternal or developmental effects were
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
observed in rats (Kronevi et al., 1988); since no effects were observed, this study is not
informative in comparing relative susceptibility of adult and young animals. Decreased
testicular weight was reported in a short-term reproductive study for GBL (Debeljuk et al.,
1983), but no impairment of fertility was reported in the oral two-generation study for THF
(BASF, 1996). Developmental studies by the inhalation route have been conducted in both rats
(Mast et al., 1992; DuPont Haskell Laboratory, 1980) and mice (Mast et al., 1992). Mast et al.
(1992) reported decreased fetal survival and incidence of sternal ossification in mice and
decreased fetal BW in rats. DuPont Haskell Laboratory (1980) reported decreased fetal weight
and skeletal alterations. In these inhalation studies, developmental effects were only observed at
concentrations that also induced maternal toxicity.
Comparisons of maternal to developmental effect levels can be useful for evaluating the
susceptibility of young animals. The inhalation data for THF suggest that fetuses are not likely
to be more susceptible than adult animals. This conclusion is supported by the observation that
in the inhalation toxicity database (see Table 4-10), the LOAELs for systemic toxicity in adult
animals are significantly lower than the LOAELs for developmental toxicity. However, the
inhalation developmental studies are limited, since they did not provide an evaluation of
postnatal development. In the only available multigeneration study for THF, postnatal
development (decreased pup BW gain, delayed eye opening, and increased incidence of sloped
incisors) was affected at drinking water concentrations that had minimal effects on the dams.
The results from the two-generation study indicate that the early postnatal period is a period of
increased susceptibility, but this conclusion is uncertain since the changes in pup BW may be
explained by effects on maternal water intake. Furthermore, the related measure of fetal weight
at the end of the prenatal period was not affected in the inhalation developmental studies or in
the oral developmental study for GBL.
Only one study was identified that specifically evaluated the effect of age on toxicity of
THF. Kimura et al. (1971) estimated oral LD50 values for a variety of solvents, including THF,
for newborn, 14-day-old, young adult, and older adult rats. The oral LD50 values for THF were
estimated as 2.3 mL/kg for 14-day-old rats, 3.6 mL/kg for young adult rats, and 3.2 mL/kg for
older adult rats; none of these values were statistically different. However, the authors report
that the newborn animals were much more susceptible than the other age groups, in which doses
of 1 mL/kg of all the solvents tested, including THF, were generally fatal. Since sensitivity was
increased in newborns for all the solvents tested, it is not clear whether the increased sensitivity
to THF was due to its inherent toxicity to newborn rats or whether some other aspect of the study
protocol was responsible. The study results suggest that young animals are at best marginally
more susceptible to oral THF exposure than adult animals to high-dose effects.
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
No pharmacokinetic data are available to evaluate potential childhood susceptibility. As
a result, the role of age-dependent differences in THF metabolism could not be evaluated. It is
important to note, however, that in addition to possible genetic variability (polymorphism) as
discussed in Section 4.8.3, age-dependent variability may also exist among key THF-
metabolizing enzymes including CYP450 and laconase (PON1).
The overall data are not sufficient to conclude with certainty whether children are likely
to be more susceptible to THF toxicity than adults. Adequate studies directly testing the
systemic effects of THF in animals of different ages, as well as data on relevant metabolic
parameters are lacking. However, the occurrence of developmental toxicity only at maternally
toxic doses suggests that children may not be more susceptible to THF than adults.
4.8.2. Possible Gender Differences
No adequate human studies on gender-based differences in THF toxicity are available.
Several toxicity studies of acute, subchronic, or chronic duration in animals have evaluated the
toxicity of THF in both males and females administered similar doses. In general, a similar
spectrum of noncancer endpoints and effect levels has been observed in both sexes for oral
(BASF, 1996; Komsta et al., 1988) and inhalation (NTP, 1998; DuPont Haskell Laboratory,
1996b) exposure studies. However, in the NTP (1998) subchronic study, uterine histopathology
changes were observed in mice, but no histopathological effects on the uterus were noted in the
companion chronic bioassay (NTP, 1998) or in a short-term inhalation study that evaluated
histopathology of the uterus (BASF, 2001a). Changes in uterine weight (not statistically
significant) were reported in the short-term study (BASF, 2001a). None of the available studies
that evaluated reproductive capacity (BASF, 1996, 1994) suggested that either male or female
fertility are targets for THF toxicity.
In addition, a comprehensive pharmacokinetics study of THF following oral dosing of
rats and mice of both sexes was conducted by DuPont Haskell Laboratory (1998). The AUC was
higher in males, and the corresponding clearance of THF-associated radioactivity from the blood
was lower in males of both species. This result might suggest that there are gender differences in
THF metabolism, since absorption and distribution of THF were similar for males and females.
The available data suggest that THF metabolism is extensive and that oxidative metabolism is
due to CYP450 isozymes. However, the identities of the isozymes responsible for THF
metabolism have not been elucidated. In vitro evidence suggests that there are species
differences in THF metabolism (DuPont Haskell Laboratory, 2000), and, therefore, the
differences in THF metabolism between male and female rodents cannot be used to infer the
relationship in THF metabolism between sexes in humans. As noted above, whether THF or one
of its metabolites is responsible for each of the observed toxic effects has not been demonstrated.
72
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
As a result of these considerations, the implications of sex-based differences in metabolism
cannot be determined.
A significant gender difference in response observed following exposure to THF is the
sex-specific induction of kidney tumors in male rats and liver tumors in female mice (see
Section 4.7.2), although the absence of an effect in male mice may be due to the apparently
higher susceptibility to narcosis (and resulting mortality) in male mice in the chronic inhalation
bioassay (NTP, 1998).
The overall similarity in noncancer toxicity between male and female rodents in a variety
of bioassays and the absence of functional effects on male or female fertility suggest that gender-
based differences in susceptibility to THF are likely to be limited. However, a number of
findings raise questions about the potential for increased susceptibility based on gender,
including potential effects in the uterus of mice, apparent sex-specific tumor formation, and
pharmacokinetic differences between male and female rodents.
4.8.3. Other
Possible genetic variability (polymorphism) and/or age-dependent variability in key THF
metabolizing enzymes may contribute to interindividual variability in pharmacokinetics and
possibly to increased sensitivity to THF among ceratain individuals within the population. As
discussed in Section 3.3, the oxidative metabolism of THF to GBL may be catalyzed by one or
more of the liver microsomal CYP450 isoenzymes which may be subject to interindividual
variation due to genetic polymorphism. GBL may undergo further metabolism to GHB by the
lactonase PON1 enzyme, which also has been known to have genetic variability (polymorphism)
in expression and activity (van Himbergen et al., 2006) including a possible link to
cardiovascular risk (Bhattacharyya et al., 2008). It is not clear if and to what extent genetic
variability in CYP450 and PON1 may influence the respective oxidative metabolism of THF to
GBL or the metabolism of GBL to GHB, and how, in turn, such variability might influence
human risk to THF exposure. In addition to possible variability in THF pharmacokinetics due to
genetic polymorphism of key metabolizing enzymes, other variables could also contribute to the
degree of interindividual variability including hepatic blood flow and compensating metabolic
pathways (Ginsberg et al., 2009).
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
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Principal Study and Candidate Critical Effects—with Rationale and
Justification
A number of human occupational exposure and case report studies following exposure to
THF are available (see Section 4.1). These human studies identified effects on both the CNS and
liver. However, these studies are unsuitable for the derivation of the RfD because they do not
report levels of exposure to THF. In addition, all of these studies report concomitant exposures
to other chemicals including solvents that are potentially neurotoxic.
The oral database for characterizing the potential hazards posed by THF in laboratory
animals is limited. A one-generation reproductive toxicity (dose range-finding) study (BASF,
1994) and a two-generation reproductive toxicity study (Hellwig et al., 2002; BASF, 1996) in
rats (both in drinking water) exist. Both of these studies identified increased kidney weight,
decreased pup body weight gain, and delayed eye opening in F2 pups as the sensitive effects.
The two-generation study is considered to be more appropriate for use as the principal study
because it used a narrower range of exposure concentrations and larger group sizes, and is the
more comprehensive of the two studies. The one-generation study was considered supportive.
Regarding kidney weight effects, increased relative kidney weight was observed at
similar doses in the F0 males and females in both the one- and two-generation studies (less than
10% of the control mean). Treatment-related effects on absolute kidney weight were not as
pronounced. For example, the only group for which both relative and absolute kidney weights
were significantly increased (p < 0.05) was F0 males in the two-generation study, although
smaller increases (that were not statistically significant) were noted in other groups. The
observation that, at least for one group, both absolute and relative kidney weights were increased
indicate that these changes reflect the effects of THF on the kidney itself and are not due solely
to body weight changes. This conclusion is supported by the general absence of an effect of
THF on body weight gain in adult animals. Kidney weight changes that were observed in the F0
generation were not accompanied by gross kidney pathology, or clinical chemistry findings
consistent with an effect on renal function (in the one-generation study) or by histopathological
examination (in the two-generation reproductive toxicity study). In addition, exposure to THF
had no effect on absolute or relative kidney weight in F1 generation adults. Thus, the kidney
data were not considered further in the derivation of the RfD.
Decreases in pup body weight gain in F1 and F2 and delayed eye opening in F2 pups
observed in rats of the two-generation reproductive toxicity study were considered candidate
critical effects for RfD derivation. The decreases were consistently observed in both the F1 and
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
F2 generation pups, and were most pronounced during PND 7-14. In F2 pups these changes
were accompanied with other developmental delays (i.e., delayed eye opening). These changes
occurred in the absence of significant maternal body weight changes or other overt signs of
systemic toxicity.
Alternative approaches for deriving the RfD were considered, including the use of the
inhalation data and application of a route-to-route extrapolation approach or use of the oral data
for metabolites of THF. A human PBPK model has been developed by Droz et al. (1999) to
estimate the THF concentrations in the blood, breath, and urine following an inhalation exposure
for the purpose of determining biological exposure indices that would equate to an occupational
exposure level of 200 ppm THF. Human PBPK models with both oral and inhalation portals of
entry have not been developed, and no PBPK models have been developed in animals. Also,
there are no comparative toxicokinetic or toxicodynamic studies following exposure to THF by
the oral route in humans and animals. In the absence of PBPK models that include oral and
inhalation routes of exposure, and lacking inhalation absorption efficiency data in humans and
rats, a route-to-route extrapolation from inhalation to oral exposure for THF would be highly
uncertain and was not considered further for development of the RfD.
The use of metabolite data to calculate a reference value may be appropriate when there
are no adequate data for the parent compound or when the data indicate that the active form that
induces the critical effect is a metabolite derived from the parent compound. In both cases, the
kinetics of metabolism would need to be sufficiently understood in order to calculate the
administered dose of parent compound from the target tissue dose of the active metabolite. A
basic requirement for using the data on metabolites in a quantitative fashion for the dose-
response assessment is a demonstration that the critical effects following THF administration can
be attributed to the toxicity of metabolites. While THF metabolites also induce CNS toxicity
(narcosis), and may be more potent than THF, it is not known if this is true for other target tissue
toxicity, such as liver or kidney, as well as effects on postnatal development since evidence is
lacking that these effects are due to the action of THF metabolites. Additionally, it is not known
whether first pass hepatic metabolism of THF is or is not a detoxifying event in the absence of
information on the roles that the intermediate metabolites may play. The available data suggest
that the parent compound may be responsible for THF-induced toxicity. Therefore, the oral data
for THF, and not a metabolite, are most appropriate to serve as the basis for deriving the RfD.
5.1.2. Methods of Analysis
The candidate critical effects from the two-generation reproductive toxicity study
(Hellwig et al., 2002; BASF, 1996) considered for benchmark dose (BMD) modeling were the
F1 and F2 pup body weight gains during PND 7-14, as well as F2 delayed eye opening.
75
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
However, visual inspection of the data set for delayed eye opening in F2 pups suggested that the
results were not amenable to modeling. Therefore, this endpoint is represented by a NOAEL of
3000 ppm (385 mg/kg-day). The F1 and F2 pup weight gain data were deemed suitable for
BMD modeling. Table 5-1 summarizes the pup body weight gain data that were considered for
modeling and deriving the chronic RfD.
Table 5-1. F1 and F2 Pup body weight gain changes for RfD derivation from
the two-generation reproductive toxicity study in Wistar rats exposed to
THF in drinking water"
Generation,
sex
Parameter3
Concentration (ppm)
0
1,000
3,000
9,000
F0 Generation/Fl Pups
F0 Females
TWA THF intake during
gestation and lactation
(mg/kg-day)
0
134
381
1071
F1 Male
pups
Pup body weight gain (g)
PND 7-14
17.8 ± 1.15
17.5 ± 1.55
17.2 ± 1.43
15.7 ± 1.65b
F1 Female
pups
Pup body weight gain (g)
PND 7-14
17.3 ± 1.47
17.4 ± 1.72
16.9 ± 1.66
15.6 ± 1.56b
F1 Generation/F2 Pups
F1 Females
TWA THF intake during
gestation and lactation
(mg/kg-day)
0
129
385
974
F2 Male
pups
Pup body weight gain (g)
PND 7-14
17.4 ± 1.56
17.9 ± 1.98
17.0 ± 1.94
15.6 ± 1.67b
F2 Female
pups
Pup body weight gain (g)
PND 7-14
17.2 ± 1.50
17.1 ± 1.62
16.0 ±2.41
15.4 ± 1.84b
aSee Table 4-6 for additional details.
Statistically significantly different (p < 0.05) from controls.
TWA = time-weighted average.
Sources: Hellwig et al. (2002); BASF (1996).
Details of the BMD modeling conducted for each endpoint are presented in Table 5-2 and
in Appendix B. The modeling was conducted following EPA's draft BMD Technical Guidance
Document (U.S. EPA, 2000b) using Benchmark Dose Software (BMDS) version 2.0 (U.S. EPA,
2008). EPA's BMD technical guidance (U.S. EPA, 2000b) recommends selecting a benchmark
response (BMR) based on the biological considerations for defining an adverse effect. A 5%
reduction in pup body weight gain as a percent of the control mean is consistent with
recommendations described by Kavlock et al. (1995). Decreased pup body weight gain as low as
5% relative to controls was in the experimental range of the data. In addition, a BMR of 1
76
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
standard deviation (SD) was also estimated for each endpoint for comparison purposes (see
Appendix B), as recommended by technical guidance (U.S. EPA, 2000b).
In general, model fit was assessed by a chi-square goodness-of-fit test (i.e., models with p
<0.1 failed to meet the goodness-of-fit criterion), visual fit, and the Akaike Information
Criterion (AIC) value (i.e., a measure of the deviance of the model fit that allows for comparison
across models for a particular endpoint). Of the models exhibiting adequate fit, the model
yielding the lowest AIC value for a data set was selected as the best-fit model (U.S. EPA,
2000b); modeling details are provided in Appendix B.
Table 5-2. BMD modeling results for pup body weight gain in the Wistar
rat two-generation reproductive toxicity study
Dataset
Selected Model
BMD o.os
(mg/kg-day)
BMDL0.os
(mg/kg-day)
F1 males, days 7-14
Linear
457
355
F1 females, days 7-14
Linear
513
376
F2 males, days 7-14
Linear
417
306
F2 females, days 7-14
Linear
440
303
aAIC = Akaike Information Criterion (see Appendix B).
''BIVIDL = 95% lower bound of the BMD. Subscript denotes the specified benchmark response (BMR) level, 0.05
x (control mean).
Sources: Hellwig et al. (2002); BASF (1996).
All of the data sets for pup body weight gain during days 7-14 showed adequate visual
and statistical fit by at least one of the models considered. The dose-response pattern was
generally similar across the data sets, with linear models providing the best fit in each case. For
pup body weight gain decreases induced by THF, data corresponding to the F2 males and
females provided the lowest BMDLo.os (95% lower bound on the BMDo.os), as described by a
linear model, of 303 and 306 mg/kg-day, respectively. The outputs for these results, including
the 1 SD results for general reporting purposes, are presented in Appendix B.
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs)
The BMDL05 of 303 mg/kg-day for reduced pup weight gain in F2 female Wistar rats
exposed throughout gestation and lactation was selected as the POD in the derivation of the
chronic RfD (Hellwig et al., 2002; BASF, 1996). A composite UF of 1000 was applied to the
POD.
77
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
A default UF of 10 was applied for inter-individual variability (UFH) to account for
human-to-human variability in susceptibility in the absence of quantitative information to assess
the toxicokinetics and toxicodynamics of THF in humans. Although a human PBPK model
based on inhalation exposure of volunteers (Droz et al., 1999) is available, information on the
human variability in response to THF exposure in humans is not available.
A default UF of 10 was applied for interspecies extrapolation (UFA) to account for
uncertainty in extrapolating from laboratory animals to humans (i.e., interspecies variability)
because information was unavailable to quantitatively assess toxicokinetic or toxicodynamic
differences between animals and humans for THF.
An UF of 1 was applied to account for subchronic to chronic extrapolation (UFS) because
developmental toxicity resulting from a narrow period of exposure was used as the critical effect.
The developmental period is recognized as a susceptible life stage when exposure during a time
window of development is more relevant to the induction of developmental effects than lifetime
exposure (U.S. EPA, 1991a).
An UF of 1 was applied for LOAEL-to-NOAEL extrapolation (UFL) because the current
approach is to address this factor as one of the considerations in selecting a BMR for benchmark
dose modeling. In this case, a BMR of 5% change in pup body weight gain in F2 female rats
was selected under an assumption that it represents a minimal biologically significant change.
An UF of 10 was selected to account for deficiencies in the oral database (UFD). The oral
database for THF contains a two-generation reproductive toxicity study and a range-finding one-
generation reproductive study (Hellwig et al., 2002; BASF, 1996, 1994). The one-generation
study did not include a histopathological examination of tissues and the two-generation study
provided the results of histopathologic examinations of the liver, kidney, digestive, and
reproductive organs in male and female rats. There are no available human occupational or
epidemiological studies or standard toxicity studies, including developmental toxicity studies, in
animals via the oral route of exposure. Following inhalation exposure, there are developmental
toxicity studies (no two-generation reproductive toxicity studies are available) and chronic and
subchronic studies available in rats and mice (NTP, 1998; Mast et al., 1992; DuPont Haskell
Laboratory, 1980) which may be informative with respect to the potential oral toxicity of THF.
The inhalation developmental studies provided evidence of effects on the fetus, although these
studies are limited as they did not provide an evaluation of postnatal development. The
subchronic and chronic studies reported systemic toxicity (CNS effects and liver weight changes)
at exposure concentrations lower than those inducing developmental toxicity; suggesting that
fetuses and weanling animals may not be more sensitive than adult animals. Thus, the lack of
studies examining endpoints other than reproductive and developmental toxicity following oral
exposure is a database deficiency. Therefore, due to the absence of a developmental toxicity
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
study and other toxicity studies examining a comprehensive array of endpoints following oral
exposure to THF, a 10-fold UF was applied.
The RfD based on the BMDL0s for decreased pup body weight gain (Hellwig et al., 2002;
BASF, 1996) was derived as follows:
RfD = BMDLos - (UFH x UFA x UFD)
= 303 mg/kg-day (10 x 10 x 10)
= 303 mg/kg-day ^ 1,000
= 0.3 mg/kg-day
5.1.4. Previous RfD Assessment
This is the first IRIS assessment for THF; thus, no oral RfD was previously available on
IRIS.
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
Human occupational exposure studies and case reports have investigated the health
effects following exposure to THF. These studies indicate that the nervous system and liver may
be targets of toxicity of THF. However, all of the published human studies contain insufficient
data on the duration and/or concentration of THF exposure. In addition, the human exposure
studies indicate the potential for coexposure to other solvents. For these reasons, the available
human studies are not considered to be suitable for use in the derivation of an RfC.
Animal studies are available that examine inhalation effects of THF following subchronic
exposure in rats and mice (NTP, 1998; DuPont Haskell Laboratory, 1996b; Horiguchi et al.,
1984; Kawata and Ito, 1984; Stasenkova and Kochetkova, 1963) and 2-year exposure in rats and
mice (NTP, 1998), in addition to developmental toxicity studies in both mice (Mast et al., 1992)
and rats (Mast et al., 1992; DuPont Haskell Laboratory, 1980). Several of these studies reported
portal-of-entry findings, including irritation of the nasal and respiratory tracts (Horiguchi et al.,
1984; Kawata and Ito, 1984; Stasenkova and Kochetkova, 1963) but were not considered
suitable for RfC derivation due to concerns about lack of consistency among study findings,
reporting of these effects, and study design (see Section 4.6.2).
Following chronic exposure, no effects or clinical findings were observed in female mice,
except for a slight increase in liver necrosis in the 5,310 mg/m3exposure group (from 3/50 in
controls to 7/48) (NTP, 1998). Clinical signs of CNS toxicity (narcosis) were the only effects
observed in male mice during and up to 1 hour after cessation of exposure to THF at 5,310
mg/m3. Similar effects were observed following subchronic exposure to THF in which CNS
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
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
toxicity (narcosis) was reported in both male and female rats at 14,750 mg/m3 THF and mice at >
5,310 mg/m3, respectively. Immediately after exposure, both male and female rats in the high
exposure group showed ataxia (irregular movement with lack of coordination). Male and female
mice exposed to 5,310, and 14,750 mg/m3 were in a state of narcosis (stupor) during exposure,
but were alert and fully awake immediately after exposure to 5,310 mg/m3 while mice in the
14,750 mg/m3 group required up to 2 hours for recovery. It should be noted that it is possible
that the rats and mice may have developed a tolerance to THF exposure considering the effects
were observed at similar concentrations (5,310 mg/m3) in the subchronic and chronic studies.
However, this cannot be determined due to the lack of reporting of incidence data for these
effects and because the chronic study did not include the higher exposure group (14,750 mg/m3)
for comparison.
Further support for THF-induced CNS effects is provided by neurotoxicity,
developmental, acute, and short-term studies. The only findings in a neurotoxicity study were
sedative effects in male and female rats at 4,425 and 8,850 mg/m3 (DuPont Haskell Laboratory,
1996b; Malley et al., 2001). Developmental studies conducted in both rats and mice reported
maternal toxicity including CNS effects (Mast et al., 1992). Following acute and short-term
exposure, symptoms of CNS toxicity, including sedation, coma, altered respiration, and
decreased response to external stimuli, were observed in dogs (Stoughton and Robbins, 1936),
mice (Stasenkova and Kochetkova, 1963; Stoughton and Robbins, 1936), and rats (Horiguchi et
al., 1984; DuPont Haskell Laboratory, 1979; Stasenkova and Kochetkova, 1963) (See Appendix
C for study descriptions). Additionally, as reported in Section 4.1, human CNS effects may
result from THF occupational exposure. Based on the above findings, the CNS toxicity was
further considered as a candidate critical effect in the derivation of the RfC.
Chronic exposure to THF resulted in liver necrosis in the 5,310 mg/m3 exposure group for
female mice (NTP, 1998). Subchronic exposure to THF (NTP, 1998) provided evidence of
increased liver weights (both absolute and relative) in the 14,750 mg/m3 female rats and this
finding was accompanied by increased serum bile acid concentration in the absence of
cholestasis or hepatocellular necrosis. The study authors indicated that these changes were
consistent with decreased or altered hepatic function. In male mice, absolute and relative liver
weights were statistically significantly increased following exposure to concentrations of >1,770
mg/m3. The increases in absolute and relative liver weights in male mice were corroborated by
increased incidence of centrilobular cytomegaly, statistically significant at 14,750 mg/m3 (7/10
compared to 0/10 in controls). Also, relative and absolute liver weights were statistically
significantly increased in female mice beginning at 5,310 mg/m3 and were accompanied by
centrilobular cytomegaly (10/10 animals compared to 0/10 in controls) at 14,750 mg/m3. The
hepatocytes were additionally described as having slight karyomegaly (enlarged nucleus),
80
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
increased cytoplasmic volume, and granular cytoplasm with less vacuolation than that of
midzonal and periportal hepatocytes (NTP, 1998). No clinical chemistry measurements were
performed in mice; however, the finding of increased bile acids in rats, in the absence of
increased serum liver enzymes, was interpreted as possibly signifying decreased or altered
hepatocellular function in the 14,750 mg/m3 exposure group.
Further support in the database exists for liver effects following THF exposure.
Specifically, fatty liver degeneration (or infiltration) which was observed following short-term
inhalation exposure in female mice (Gamer et al., 2002; BASF, 2001a) is a likely adverse effect
since certain drugs which evoke fatty liver changes may predispose the liver to oxidative stress,
lipid peroxidation, and possible mitochondrial and organ damage (Begriche et al., 2006; Letteron
et al., 1996). In another subchronic inhalation toxicity study, Horiguchi et al. (1984) reported
mild liver toxicity in male rats in the form of increased serum liver enzymes, bilirubin, and
cholesterol at THF exposure concentrations of 2,950 and 14,750 mg/m3 in addition to increased
relative liver weight at 14,750 mg/m3 but no liver histopathology findings were reported (Section
4.2.1.2). Some earlier studies also reported liver effects when THF was administered in animals
using exposure routes other than inhalation (Stasenkova and Kochetkova, 1963; Komsta et al.,
1988). As reported in Section 4.1, the human liver also may be a target organ for THF
occupational exposure settings. While the reported liver findings may be confounded by the
likelihood of coexposure to other chemicals, it is reasonable to conclude that repeated
occupational exposure to high concentrations of THF may have contributed to the large increases
in serum liver enzymes and the palpable liver findings in some of the human studies (Gamier et
al., 1989; Horiuchi et al., 1967).
Subchronic exposure also resulted in effects including altered organ weights (thymus, and
spleen), increased bile acids, and altered hematological parameters at 14,750 mg/m3 THF in male
and female rats; however, no histopathological lesions were identified (NTP, 1998). The
biological significance of the decrease in thymus weight was considered questionable (Section
4.6.2). Degeneration of the adrenal cortex and uterine atrophy in the 14,750 mg/m3 female mice
was also observed. According to the study authors, degeneration of the adrenal cortex and
uterine atrophy may have been a direct effect of THF on these tissues or may be the result of a
hormonal effect, possibly through perturbation of the pituitary-hypothalamic-end organ axis
(NTP, 1998). On the other hand, no histopathological effects on the uterus or adrenals were
noted in the companion chronic bioassay (NTP, 1998) or in a short-term inhalation study that
evaluated histopathology of the uterus (BASF, 2001a). The effects on the thymus, spleen,
adrenal cortex and uterus were not considered further in the derivation of the RfC.
In consideration of the available studies reporting effects of chronic and subchronic THF
exposure in animals, the NTP (1998) study was chosen as the principal study. The subchronic
81
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
phase, rather than the chronic phase, of this study was selected to serve as the principal study due
comprehensive reporting in the subchronic study which better characterized the low-dose effects
associated with THF. Sensitive endpoints identified in this study, the effects in the CNS and
liver, were selected as the co-critical effects. The CNS effects were observed in rats and mice (at
concentrations >5,310 mg/m3) and the liver effects were observed in rats (at concentrations of
14,750 mg/m3) and mice (at concentrations >590 mg/m3). The toxicological significance of the
observed liver weight changes was considered to be uncertain at the low concentrations (590-
1,770 mg/m3), where the changes were of minimal severity and were not accompanied by other
signs of liver toxicity. The increases in absolute and relative liver weights at 5,310 mg/m3 were
greater than 10% above controls (statistically significant) and were accompanied by minimal
increases in histopathology findings (1/10 incidence in centrilobular cytomegaly) that progressed
with increases in THF concentration. The liver and CNS effects observed at the exposure
concentration of >5,310 mg/m3 were considered biologically significant and representative of
adverse effects.
5.2.2. Methods of Analysis
The most relevant endpoints for deriving the POD for the quantitative assessment were
CNS effects, hepatic centrilobular cytomegaly and increased liver weights in male mice in the
NTP (1998) subchronic study. Data in mice, rather than rats, were modeled because mice were
more sensitive to the THF-induced liver and CNS effects. The selection of the male mouse data
was based on the fact that the liver weight increased more steadily from lower administered
exposure in males than in females. Suitable data were available to model the liver weight and
liver histopathology findings using benchmark dose methods (see Table 5-3). Note that because
there was very little effect on body weight until the highest exposure, the absolute and relative
liver weight changes were essentially the same, and only the absolute liver weights were
considered for modeling. For CNS effects, no incidence data were available from the NTP
(1998) study, therefore BMC modeling could not be conducted for this endpoint, and a NOAEL
was identified for the POD. See Table 5-3 for the data considered for POD derivation for the
liver effects.
Human equivalent concentrations (HECs) for the potential critical effects were derived
(Section 5.2.2.1), and the final selection of the POD was made after the evaluation of effect
levels among multiple endpoints from the principal study (Section 5.2.2.2).
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
5.2.2.1. Calculation of HECs
The Methods for Derivation of Inhalation Reference Concentrations and Application of
Inhalation Dosimetry (hereafter referred to as the RfC Methodology) recommends converting the
POD[adj] to a human equivalent concentration (HEC) (U.S. EPA, 1994b). For the purposes of
this assessment, the induction of extrarespiratory tract effects in the liver and in the CNS is
consistent with properties of a category 3 gas as described under the RfC methodology (U.S.
EPA, 1994b).
For category 3 gases, HECs are calculated by multiplying the duration-adjusted exposure
concentration by the RGDR for the extrarespiratory region. The RGDR for extrarespiratory
effects is calculated by finding the ratio of the animal-to-human blood:gas (air) partition
coefficients. In cases where there are either no data available or where the animal partition
coefficient is larger than the human coefficient, a default value of 1 is used for the RGDR. For
THF, a human blood:gas partition coefficient was available from Ong et al. (1991); however, no
value was available for animals. Therefore, the default of 1 was applied in estimating the HECs
for extrarespiratory effects. For example, for the concentration of 1,770 mg/m3, which
corresponds to the NOAEL for the CNS toxicity (narcosis) in male and female mice in the NTP
(1998) study, the HEC based on the equation for a category 3 gas was calculated by estimating
continuous equivalent exposure and applying the RGDR, as follows:
NOAELadj= 1,770 mg/m3 x 6/24 hours x 5/7 days = 316 mg/m3
NOAELrec = 316 mg/m3 x default RGDR of 1 =316 mg/m3
The HECs calculated for each study concentration were used directly in conducting the
benchmark concentration (BMC) modeling of the liver effects. See Table 5-3 for the estimated
HECs.
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
Table 5-3. Measures of liver toxicity in B6C3Fi male mice following
subchronic inhalation exposure to THFa
Endpoint
Administered Concentration in ppm (concentration in mg/m3)
0(0)
66 (195)
200 (590)
600 (1770)
1,800 (5,310)
5,000 (14,750)
Human Equivalent Continuous Concentration (mg/m3)
0
35
105
316
948
2,634
Absolute liver
weight (g)
1.613 ±0.037
1.667 ±0.022
1.695 ±0.037
1.722 ±0.031b
1.789 ±0.035°
1.964 ±0.060°
Centrilobular
cytomegaly
0/10
NE
NE
NE
1/10
7/10
aMean ± standard error. All group sizes are 10 animals/group except for male mice in the 5,000 ppm group where
N = 7.
hp< 0.05.
><0.01.
NE = Not examined.
Source: Adapted from NTP (1998).
5.2.2.2. BMC Modeling
The modeling was conducted following EPA draft BMD technical guidance (U.S. EPA,
2000b) and used BMDS version 2.0 (U.S. EPA, 2008), as for the RfD (See Section 5.1.2). For
liver weights, a BMR of a 10% change relative to control was used, by analogy to its use in
evaluating body weight changes. In addition, a BMR of 1 standard deviation (SD) was also
estimated for each endpoint for comparison purposes (see Appendix B). For centrilobular
cytomegaly, no biological criterion for defining adversity was available, and a 10% extra risk
was used under the assumption that is represents a minimally biologically significant effect level.
For the liver weight data set, all of the continuous models fit the data adequately (see
Table B-2). BMDLs ranged over fourfold, leading to the selection of the unrestricted power
model, with the lowest BMDL, for providing the POD (see Appendix B). The EPA's BMD
technical guidance has generally recommended restricting the power parameter in the power
model to be greater than 1, primarily to avoid low-dose extrapolation in regions where the
estimated dose-response relationship is so steep that it may appear biologically implausible. For
these data, however, the BMRs of 10% change relative to the control mean and 1 SD both fell
well within the data range, and BMDLs estimated with unrestricted parameters provide more
accurate confidence interval coverage. The candidate POD for increased absolute liver weight
was the BMCL of 246 mg/m3 (Table 5-4).
For the centrilobular cytomegaly data set, the full suite of quantal models in BMDS was
considered. All of the models provided an adequate fit overall to the data set based on a
goodness-of-fit p value greater than 0.1. Of the models exhibiting adequate fit, BMDLs fell
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
within a threefold range, and the model yielding the lowest AIC value for a data set was selected
as the best-fit model (U.S. EPA, 2000b). The multistage model demonstrated the lowest AIC
(Table B-2). The candidate POD for centrilobular cytomegaly was the BMCL of 256 mg/m3
(Table 5-4).
Table 5-4. BMCa modeling results for noncancer effects in male mice,
resulting from subchronic inhalation exposure to THF
Dataset
Selected Model
BMCo.io"
BMCL010b
Absolute liver weight
Power (unrestricted)
783
246


BMC10
BMCL10
Centrilobular cytomegaly
Multistage, degree 2 (coefficients
>0)
805
256
Concentrations used in the modeling were the HECs in mg/m3 (see Table 5-3).
bFor liver weights, BMC 0 io and BMCL0.i0 refer to a BMR of 10% increase in the control mean, while for
centrilobular cytomegaly, BMCio and BMCL,(J refer to 10% extra risk.
Data Source: NTP(1998).
For CNS effects in male and female mice, no incidence data were available, and a
NOAEL of 1,770 mg/m3 was identified as the POD. The adjustment for human equivalent
continuous concentration corresponds to a candidate POD of 316 mg/m3.
Of the three candidate PODs, the BMCLio of 246 mg/m3 based on findings of increased
absolute liver weight in male mice, was selected as the POD for deriving the RfC because it was
the most sensitive endpoint. However, a derivation of a potential RfC based on the NOAELrec
of 316 mg/m3 for CNS toxicity is presented for comparison purposes in Section 5.2.3.
5.2.3. RfC Derivation—Including Application of Uncertainty Factors (UFs)
The BMCL io of 246 mg/m3 for increased absolute liver weight in male B6C3Fi mice
exposed to THF for 6 hours/day, 5 days/week for 90 days (NTP, 1998) was selected as the POD
in the derivation of the RfC. A composite UF of 100 was applied to the POD.
A default UF of 10 was applied for inter-individual variability (UFH) to account for
human-to-human variability in susceptibility in the absence of quantitative information to assess
the pharmacokinetics and pharmacodynamics of THF in humans. Although a human PBPK
model based on inhalation exposure of volunteers (Droz et al., 1999) is available, information on
human variability relating to toxicodynamics and toxicokinetics in response to exposure to THF
is not available.
A default UF of 3 was applied for interspecies extrapolation (UFA) to account for the
uncertainty in extrapolating from laboratory animals to humans. This value is adopted by
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
convention where an adjustment from an animal-specific PODadj to a PODrec has been
incorporated. Application of an UF of 10 would depend on two areas of uncertainty (i.e.,
toxicokinetic and toxicodynamic uncertainties). In this assessment, the toxicokinetic component
associated with exposure to THF is mostly addressed by the determination of an HEC as
described in the RfC methodology (U.S. EPA, 1994b). The toxicodynamic uncertainty is also
accounted for to a certain degree by the use of the applied dosimetry method and an UF of 3 is
retained to account for residual uncertainty regarding the toxicodynamic differences between
mice and humans.
An UF of 1 was applied to account for extrapolation from subchronic-to-chronic
exposure (UFS), due to the lack of evidence that increased duration of exposure to THF may not
increase the incidence or severity of these effects. The 14-week study for THF (NTP, 1998),
selected as the principal study, reported critical findings of CNS effects and increased liver
weight which was supported by hepatic centrilobular cytomegaly. In the chronic exposure phase
of the study, while no organ weights were taken, no hepatic cytomegaly was identified at any
exposure level including the high exposure group of 5,310 mg/m3. However, the incidence of
liver necrosis in the female mice of the 5,310 mg/m3 exposure group was increased (although not
statistically significant) from 3/50 in the control to 7/48. The available chronic information
suggests that liver damage observed in rodents following subchronic exposure to THF (NTP,
1998) may not progress to more severe effects following chronic exposures near the POD,
considering that cytomegaly was not reported at chronic exposures < 5,310 mg/m3 and that
necrosis was only observed at 5,310 mg/m3 (the highest concentration), the same concentration
as the LOAEL for the CNS and liver effects in the subchronic study. Additionally, the CNS
effects were observed following exposure to 5,310 mg/m3 in both the subchronic and chronic
studies but with no evidence of effects at lower concentrations in the chronic study. A full
comparison of the studies is not possible given the incidence data were not reported for these
effects in either study. However, the available evidence suggests that increased duration of
exposure to THF may not increase the incidence or severity of these effects; thus, a 1-fold UF
was applied.
An UF of 1 was applied for LOAEL-to-NOAEL extrapolation (UFL) because the current
approach is to address this factor as one of the considerations in selecting a BMR for benchmark
dose modeling. In this case, a BMR of 10% change in absolute liver weight in male mice was
selected under an assumption that it represents a minimal biologically significant change.
An UF of 3 was applied to account for deficiencies in the database (UFD) for THF.
Chronic and subchronic inhalation bioassays and developmental toxicity studies are available in
rats and mice (NTP, 1998; Mast et al., 1992; DuPont Haskell Laboratory, 1980). No two-
generation reproductive toxicity study by the inhalation route is available. The inhalation data
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
for THF (see Section 4.2) suggest that fetuses and weanling animals may not be more sensitive
than adult animals given that the observed LOAELs for developmental effects were greater than
the LOAELs for systemic toxicity (CNS and liver weight changes) in adult animals (see Table 4-
10). However, the inhalation developmental studies are limited, since they did not provide an
evaluation of postnatal development. In the oral two-generation reproductive toxicity study for
THF, postnatal development (decreased pup body weight gain, in addition to delayed eye
opening and increased incidence of sloped incisors) was affected at drinking water
concentrations that had minimal effects on the dams. Therefore, a database UF of 3 was applied
to account for the lack of a two-generational reproductive study.
The RfC based on the BMCL io for increased absolute liver weight, and supported by the
co-critical effects, comprising CNS effects and increased incidence of centrilobular cytomegaly,
in male B6C3Fi mice (NTP, 1998), was derived as follows:
RfC = BMCLio - (UFh x UFa x UFd)
= 246 mg/m3 - 100
= 2.46 mg/m3
= 2 mg/m3 (rounded to 1 significant figure)
For comparison, a potential RfC can be derived from the PODHec based on the NOAEL
for CNS effects as follows:
RfC = NOAELrec - (UFh x UFa x UFd)
= 316 mg/m3 100
= 3.16 mg/m3
= 3 mg/m3 (rounded to 1 significant figure)
5.2.4. Previous RfC Assessment
This is the first IRIS assessment for THF; thus, no inhalation RfC was previously
available on IRIS.
5.3. CANCER ASSESSMENT
5.3.1. Choice of Study/Data—with Rationale and Justification
No studies evaluating the carcinogenicity of THF by the oral or inhalation route were
identified in humans (see Section 4.1.). A 2-year NTP (1998) inhalation cancer bioassay
reported a statistically significant positive trend in renal tubule adenomas or carcinomas in male
F344/N rats and a statistically significant positive trend in hepatocellular adenomas or
87
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
carcinomas in female B6C3Fi mice following inhalation exposure to 200, 600, and 1,800 ppm
(NTP, 1998) (see Section 4.7.2). Adenoma and carcinoma incidences within each site were
combined by counting animals with either of these responses. This practice was performed
under the assumption that adenomas and carcinomas originating from the same cell type
represent stages along a continuum of carcinogenic effects resulting from the same mechanism,
as recommended by the EPA cancer guidelines (U.S. EPA, 2005a). Table 5-4 summarizes the
incidences of mouse hepatocellular and rat renal neoplasms.
Table 5-5. Incidences of neoplastic lesions of the livers of female B6C3Fi
mice and kidneys of male F344/N rats exposed to THF 6 hours/day,
5 days/week for 105 weeks
Lesion
Concentration (ppm)
0
200
600
1,800
Female B6C3F, mice
Hepatocellular adenoma or carcinoma
Overall incidence3
17/50
24/50
26/50
41/48
Adjusted rateb
46.3%
61.3%
69.1%
93.0%
Adjusted incidence0
17/37
24/39
26/38
41/44
Trend test p-valuesd
p < 0.001
Male F344/N rats
Renal adenoma or carcinoma
Overall incidence3
1/50
1/50
4/50
5/50
Adjusted rateb
8.3%
16.7%
18.8%
38.3%
Adjusted incidence0
1/12
1/6
4/21
5/13
Trend test p-valuesd
p < 0.037
3Number of animals with tumors per number of animals examined.
bKaplan-Meier estimated tumor incidence at the end of the study, incorporating an adjustment for intercurrent
mortality.
0 Adjusted denominator estimated by dividing numerator (tumors) by the adjusted rate expressed as a proportion
(e.g., 0.083 rather than 8.3%).
dTrend tests: logistic regression
Source: NTP (1998).
Although no human studies were available, a chronic study in two rodent species
provides suggestive evidence of THF-induced carcinogenicity. The data from these studies are
adequate to support a quantitative cancer dose-response assessment. The NTP (1998) cancer
bioassay for THF is a well-conducted study showing evidence of increased incidence of tumors
in differing sexes of two species at all exposure levels. Both the overall and adjusted rates of
hepatocellular adenoma or carcinoma were increased in female mice, starting at an
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
32
33
34
35
36
approximately 15% increase over control at the lowest exposure, while the adjusted rate of renal
adenoma or carcinoma was increased in male rats, starting at an approximately 8% increase over
control at the lowest exposure. Considering that a tumor response was noted and that the data
are amenable to modeling, EPA concluded that quantitative analyses may be useful for providing
a sense of the magnitude of potential carcinogenic risk. As discussed below, BMC modeling
was performed on both male rat kidney tumors and female mouse liver tumors.
5.3.2. Exposure Adjustments and Extrapolation Method
THF is water soluble, and pharmacokinetics information suggests that it is systemically
absorbed and widely distributed following inhalation exposure in both humans and animals
(Droz et al., 1999; Ong et al., 1991; Kageyama, 1988; Elovaara et al., 1984; Kawata and Ito,
1984; Wagner, 1974). Accordingly, the liver and kidney tumors observed following inhalation
exposure to THF are considered extrarespiratory effects of a category 3 gas as defined by EPA's
RfC Methodology (U.S. EPA, 1994b). Experimental exposure concentrations were converted to
mg/m3 (0, 590, 1,770, and 5,310 mg/m3), and adjusted to a continuous exposure basis (mg/m3 x
6 hours/24 hours x 5 days/7 days = mg/m3 x 0.1786: 0, 105, 316, and 948 mg/m3). For the
category 3 equations, HECs for gases are calculated by multiplying the duration-adjusted
exposure concentration by the RGDR for the extrarespiratory region. The RGDR for
extrarespiratory effects is calculated by finding the ratio of the animal-to-human blood:gas (air)
partition coefficients. In cases where there are either no data available or where the animal
partition coefficient is larger than the human coefficient, a default value of 1 is used for the
RGDR. For THF, a human blood:gas partition coefficient was available from Ong et al. (1991);
however, no value was available for animals. Therefore, the default of 1 was applied in
estimating the HECs for extrarespiratory effects.
The U.S. EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a)
recommend that the method used to characterize and quantify cancer risk from a chemical is
determined by what is known about the mode of action of the carcinogen and the shape of the
cancer dose-response curve. The linear approach is recommended if the mode of action of
carcinogenicity is not understood (U.S. EPA, 2005a). In the case of THF, although there is some
information available, the data are inadequate to establish the mode of carcinogenic action for
kidney and liver tumors. Therefore, a linear low-dose extrapolation approach was used to
estimate human carcinogenic risk associated with THF exposure.
Several of the external peer review panel members (see Appendix A: Summary of
External Peer Review and Public Comments and Disposition) recommended that a non-linear
extrapolation approach to estimate the human carcinogenic risk associated with exposure to THF
should be presented in the Toxicological Review. The reviewers agreed with EPA's conclusion
89
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
that based on the available data the modes of action for both kidney and liver tumors induced by
THF are unknown. However, some of the reviewers suggested that THF is a weak carcinogen
and not highly toxic and that the biological effects identified for THF are those that commonly
exhibit thresholds. Specifically, they stated that THF does not appear to be genotoxic, does not
produce irreversible damage and/or proliferative lesions that are preneoplastic, and is not
bioaccumulative. The reviewers that recommended a nonlinear approach suggested that a
nongenotoxic carcinogen would have a nonlinear cancer response at low dose.
Very little data are available to inform the mode of action and no data are available to
indicate the shape of the dose-response curve at low doses. If data were available to better
inform the mode of action, and the data were indicative of a threshold response, then a reference
value could be derived based on a precursor endpoint (i.e., key event in the mode of action) and
considered for the RfC. In such cases, the reference value would be considered protective
against tumor development following inhalation exposures. For THF, there were no noncancer
effects reported that could serve as a precursor endpoint upon which to base a nonlinear analysis.
EPA considered whether the cell proliferation reported in the livers of mice following short-term
exposure to THF was a potential key event in the development of female mice liver tumors;
however, given the absence of proliferation data in any of the subchronic or chronic studies, the
use of this endpoint is not supported. In the absence of such information and under the U.S. EPA
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), EPA concluded that the data are
insufficient to provide significant biological support for either a linear or a nonlinear approach.
Therefore, extrapolation from the POD to lower doses is conducted by using a default linear
approach.
Because there are no biologically based dose-response models suitable for the tumor data
identified above, the data sets for incidence of hepatocellular adenoma or carcinoma observed in
female B6C3Fi mice and for incidence of renal tubule adenoma or carcinoma in male F344/N
rats, both adjusted for intercurrent mortality as estimated by NTP (1998) (see Table 5-5), were
modeled using the multistage model in the BMDS version 2.0 (U.S. EPA, 2008). A 10% BMR
was used with each tumor type (U.S. EPA, 2005a). BMC modeling results are shown in
Appendix B. The results of this analysis are summarized in Table 5-6.
90
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 5-6. Cancer Multistage modeling results for THF
Endpoint
/7-Value
BMCio/hec a
BIVIC L i o/i 11. ( a
Hepatocellular adenoma or
carcinoma (female mice)
0.47
52
35
Renal tubule adenoma or
carcinoma (male rats)
0.59
260
127
Concentrations used in the modeling were the HECs reported in mg/m3 and assuming the ratio of animal to
human ainblood partition coefficients is 1.
Data Source: Modeling based on data from NTP (1998).
In both cases, this model provided adequate data fits with goodness-of-fit/^-values higher
than 0.10; consequently, these results were used since there was no compelling biological reason
to use another empirical model. For the hepatocellular adenoma or carcinoma data set, the
BMCio/hec and BMCLi0/Hec are 52 and 35 mg/m3, respectively. For the renal tubule adenoma or
carcinoma data in male F344/N rats, the model gives the BMCio/hec of 260 mg/m3 and
corresponding BMCLio/mx of 127 mg/m3. The data for female mouse liver tumors were selected
for the derivation of the POD for the quantitative assessment since the data provided the
strongest carcinogenic response to inhalation exposure in animals. Therefore, the BMCLi0/hec of
35 mg/m3 was selected as the POD for the cancer assessment.
5.3.3. Inhalation Unit Risk
The inhalation unit risk (IUR) is derived from the BMCLio/mx (the lower bound on the
exposure associated with a 10% extra cancer risk) by dividing the risk (as a fraction) by the
BMCLio/hec and represents an upper bound estimate on human extra cancer risk from continuous
lifetime inhalation exposure to THF. The HEC BMCLio for extra risk of hepatocellular
adenomas or carcinomas in female B6C3Fi mice exposed to THF results in an IUR of
0.1/(35 mg/m3) = 0.0029 (mg/m3)"1 or 3 x 10~6 (|a,g/m3)-1 (rounded to one significant figure).
This value was derived by linear extrapolation to the origin from the POD of 35 mg/m3 and
represents an upper bound estimate. This unit risk should not be used with exposures >35
mg/m3, because above this level, the modeled dose-response relationship better characterizes
what is known about the carcinogenicity of THF than the inhalation unit risk. The slope of the
linear extrapolation from the BMCio is calculated as 0.1/(52 mg/m3) = 0.0019 (mg/m3)"1 or 2 x
10 6 (|ag/m3)
91
DRAFT - DO NOT CITE OR QUOTE
-------
1	5.3.4. Previous Cancer Assessment
2	This is the first IRIS assessment for THF; thus, no cancer assessment was previously
3	available on IRIS.
4
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
34
35
36
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
6.1.1.	Oral Noncancer
The database for oral toxicity of THF is limited. No human data following oral exposure
to THF are available. A number of human occupational exposure and case report studies
suggesting CNS and liver effects following exposure to THF are available; however, these
studies do not report levels of exposure to THF and all studies included co-exposures to other
chemicals known to cause similar toxicity. CNS toxicity appears to be the primary health effect
following acute exposure in animals, although no CNS effects were reported in the rodent
drinking water reproductive toxicity studies of longer exposure duration. Short-term and
subchronic oral exposure studies (4 weeks to approximately 90 days) suggest that general
toxicity (characterized by altered food and water consumption and decreased body weight) and
liver and kidney toxicity are potential health effects of oral exposure to THF. The available
reproductive toxicity studies suggest that THF induces effects in the offspring of exposed dams.
The two-generation reproduction toxicity in rats (BASF, 1996) was selected as the
principal study for the derivation of the RfD. The RfD of 0.3 mg/kg-day is based on a BMDLqs
of 303 mg/kg-day for decreased pup body weight gain (BASF, 1996). A composite UF of 1,000
was used. This factor is based on selection of an uncertainty factor of 10 to account for
intrahuman variability, 10 for interspecies extrapolation, and 10 for uncertainties in the database.
There is medium confidence in the principal study (BASF, 1996), however, the overall
confidence in the oral THF database is low, with several key data gaps identified, including lack
of a full systemic toxicity study and developmental toxicity studies. Therefore, the confidence in
the RfD is characterized as low-to-medium.
6.1.2.	Inhalation Noncancer
Although no epidemiological studies of THF have been conducted, several occupational
exposure case studies in humans suggest that target organs in humans are the CNS, respiratory
tract, liver, and kidney (Gamier et al., 1989; Albrecht et al., 1987; Juntunen et al., 1984; Edling,
1982; Emmett, 1976). The major uncertainties associated with all of the reported human case
studies are that none reported exposure levels for THF and the workers were exposed to other
solvents and chemicals in addition to THF; therefore, it is not possible to attribute the observed
effects to THF exposure alone. Nevertheless, inhalation studies in animals generally identified a
similar array of target organs (see Table 4-10) including clinical signs of CNS toxicity and liver
toxicity.
93
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
Respiratory tract irritation was reported in multiple human and animal studies. One
consideration in evaluating the potential health consequences due to THF-induced respiratory
tract irritation is the role of the exposure duration on the severity of the effect. Several acute or
short-term exposure studies (Ikeoka et al., 1988; Horiguchi et al., 1984; Ohashi et al., 1983)
identified concentrations inducing irritant responses that were lower than the concentrations that
induced toxicity in subchronic and chronic studies (NTP, 1998). There is direct evidence that
respiratory tract responses are transient in nature, waning with increasing exposure duration
(Horiguchi et al., 1984). These data suggest that irritant responses not observed with subchronic
or chronic exposure could occur in individuals who were not previously exposed.
Several systemic effects have been observed following subchronic or chronic inhalation
exposure to THF. Decreased body weight has been observed in rats (Horiguchi et al., 1984;
Kawata and Ito, 1984). Decreased blood pressure was observed in dogs (BASF, 1938) and rats
(Stasenkova and Kochetkova, 1963). Altered hematological parameters were observed in rats
(NTP, 1998; Horiguchi et al., 1984), mice (NTP, 1998; Stasenkova and Kochetkova, 1963), and
dogs (BASF, 1938). Following 14 weeks of inhalation exposure, rats of both sexes had
significantly increased relative liver weight and significantly relative weights for thymus and
spleen; male rats also had significantly increased relative kidney and lung weights (NTP, 1998).
In the same study, mice of both sexes showed increased relative liver weight and decreased
relative spleen weight, while male mice only had decreased relative thymus weight and female
mice had a slightly reduced relative lung weight (NTP, 1998). In addition, Horiguchi et al.
(1984) observed increased relative weights of brain, lung, liver, pancreas, spleen, and kidney.
Developmental studies by the inhalation route have been conducted in both rats (Mast et
al., 1992; DuPont Haskell Laboratory, 1980) and mice (Mast et al., 1992). In both studies and
both species, maternal toxicity included symptoms of CNS effects and significant decreases in
body weight accompanied by decreases in gravid uterine weight (Mast et al., 1992) or food
consumption (DuPont Haskell Laboratory, 1980). Decreased fetal weight was observed at the
same concentration that resulted in maternal toxicity in rats (Mast et al., 1992). In both mice
(Mast et al., 1992) and rats (DuPont Haskell Laboratory, 1980), decreased fetal survival also
occurred at the same concentrations that resulted in maternal toxicity. With regard to potential
teratogenic effects, Mast et al. (1992) noted that in mice that survived the exposure period, no
increase was observed in the incidence of fetal abnormalities. However, an increased incidence
of incomplete sternal ossification in rat fetuses was observed (DuPont Haskell Laboratory,
1980).
After consideration of all endpoints, the CNS effects and liver toxicity were determined
to be the most sensitive effects detected in the subchronic NTP (1998) study. Furthermore, the
THF database contains additional support for these endpoints from both human and animal
94
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
studies (Gamier et al., 1989; Horiuchi et al., 1967; Stasenkova and Kochetkova, 1963; Komsta et
al., 1988; Horiguchi et al., 1984; Gamer et al., 2002; BASF, 2001a; DuPont Haskell Laboratory,
1979; 1980; 1996a; 1996b).
The RfC of 2 mg/m3 is based on findings of CNS and liver toxicity in male mice in a
subchronic NTP (1998) study, with a POD of 246 mg/m3 derived from the BMCLio value for
increased absolute liver weight. A composite UF of 100 was used. This factor is based on a
default factor of 10 to account for intrahuman variability, 3 for extrapolation from an animal
study for which effect levels were adjusted by appropriate animal-to-human dosimetry, and 3 to
account for uncertainties in the overall toxicity database.
No sensitive subpopulations have been identified. The existing data do not provide
convincing evidence for age- or gender-related differences in sensitivity to noncancer effects of
THF, although there is uncertainty regarding the ability of THF to affect postnatal development.
A number of findings raise questions about the potential for increased susceptibility based on
gender, including potential effects in the uterus of mice, apparent sex-specific tumor formation,
and pharmacokinetic differences between male and female rodents.
The principal study used to derive the RfC (NTP, 1998) was a well-conducted and
documented study reflecting high confidence. The study included subchronic and chronic
exposure duration components in two species by the relevant route of exposure, evaluated a
comprehensive array of tissues, and covered a well-spaced concentration range. Confidence in
the supporting database is medium to high. Although chronic toxicity studies (NTP, 1998) and
developmental toxicity studies (Mast et al., 1992; DuPont Haskell Laboratory, 1980) were
available for the inhalation route, no multigeneration reproduction toxicity study by the
inhalation route is available. Both the inhalation developmental toxicity studies (Mast et al.,
1992; DuPont Haskell Laboratory, 1980) and the oral two-generation reproduction toxicity study
(BASF, 1996) show that effects in fetuses and pups occur at doses that cause at least minimal
maternal effects and that these doses are higher than the NOAEL for organ weight changes in
mice (NTP, 1998). Based on high confidence in the well-conducted critical study and medium-
to-high confidence in the database, the overall confidence in the RfC can be characterized as
medium to high.
6.1.3. Cancer
No epidemiological studies were identified that evaluated the carcinogenic potential of
THF via the oral, inhalation, or dermal routes of exposure.
A two-year NTP (1998) inhalation cancer bioassay reported a statistically significant
increasing trend for renal tubule adenomas and carcinomas in male F344/N rats and of
hepatocellular adenomas and carcinomas in female B6C3Fi mice following inhalation exposure
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
20
21
22
23
24
25
26
27
28
29
30
to 200, 600, and 1,800 ppm of THF. Data for female mouse liver tumors were selected as the
basis for the derivation of the inhalation unit risk because this was the strongest carcinogenic
response to inhalation THF exposures observed in animals.
The available mechanistic information and possible modes of action were evaluated for
the male rat kidney tumors and female mice liver tumors. For the rat kidney tumors, there are
some data suggesting that following the inhalation exposure in the NTP (1998) bioassay, tumors
developed due to the accumulation of a,2U-globulin. However, data were insufficient to support
this mode of action. For mouse liver tumors, although increased cell proliferation was noted in
short-term studies, the data are not adequate to support a mode of action. The absence of a
significant increase in cell proliferation in tissues obtained from the subchronic NTP (1998)
study suggests that cell proliferation might not be a sustained response even with continued
dosing. Furthermore, key precursor events linked to observed cell proliferation have not been
identified. The data on other potential modes of action are too limited to establish the mode of
action for the liver tumor induction of THF.
Exposure concentrations were adjusted to HECs prior to BMD modeling according to
EPA (U.S. EPA, 1994b) default dosimetric equations for a category 3 gas. The tumors observed
in the kidney and liver following inhalation exposure to THF are consistent with the expected
site of action for a category 3 gas. The incidence of hepatocellular adenoma or carcinoma
observed in female B6C3Fi mice in the NTP (1998) study were modeled using the multistage
model. Concentrations associated with a 10% extra risk for tumors at the lower 95% confidence
bounds for the animal curves were determined. The BMCLio of 35 mg/m3 for hepatocellular
adenomas and carcinomas was selected as the POD for the quantitative cancer assessment. A
linear extrapolation from the origin to the POD resulted in the derivation of an IUR of 3 x 10-6
(Hg/m3)"1, which represents an upper bound risk estimate for human exposures not exceeding 35
mg/m3. The slope of the linear extrapolation from the BMCio is 0.1/(52 mg/m3) or 2 x 10 6
(Hg/mY1.
96
DRAFT - DO NOT CITE OR QUOTE
-------
7. REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). (2001) Tetrahydrofuran. In: Documentation
of the threshold limit values and biological exposure indices. Cincinnati, OH: American Conference of
Governmental Industrial Hygienists.
Adams, TB; Greer, DB; Doull, J; et al. (1998) The FEMA GRAS assessment of lactones used as flavour ingredients.
Food Chem Toxicol 36:249-278.
Addolorato, G; Cibin, M; Caprista, E; et al. (1998) Maintaining abstinence from alcohol with y-hydroxybutyric acid.
Lancet 351:38.
Albrecht, WN; Boiano, JM; Smith, RD. (1987) IgA glomerulonephritis in a plumber working with solvent-based
pipe cement. Ind Health 25:157-158.
Andriamampandry, C; Taleb, O; Kemmel, V; et al. (2007) Cloning and functional characterization of a gamma-
hydroxybutyrate receptor identified in the human brain. FASEB J 21:885-895.
Arimoto, S; Nakano, N; Ohara, Y; et al. (1982) A solvent effect on the mutagenicity of tryptophan-pyrolysate
mutagens in the Salmonella/mammalian microsome assay. Mutat Res 102(2): 105-112.
BASF. (1938) Toxicity of tetrahydrofuran, with cover letter dated 05/10/94 (sanitized). Submitted under TSCA
Section 8D; EPA Document No." 86940000738S; NTIS No. OTS0557148.
BASF. (1993) Safety data sheet of pure distilled tetrahydrofuran.
BASF. (1994) Brief report: One-generation reproduction toxicity study of tetrahydrofuran in rats; administration in
the drinking water; range-finding study. Project No. 16R0144/93020.
BASF. (1996) Tetrahydrofuran: two-generation reproduction toxicity study in Wistar rats, continuous administration
in the drinking water, with cover letter dated 8/30/96. Study No. 71R0144/93038. Submitted under TSCA Section
8D. EPA Document No. 86960000573. NTIS No. OTS558774.
BASF. (1998) Tetrahydrofuran: study on cell proliferation in F344/N rats and B6C3F1 mice, with cover letter dated
10/14/1998. Study No. 97055. Submitted under TSCA Section 8D. EPA Document No. 86990000001. NTIS No.
OTS0573851.
BASF. (2001a) Tetrahydrofurane: subacute inhalation study in F344 rats and B6C3F1 mice 20 exposures to vapors
including interim sacrifices of satellite groups after 5 exposures, with a cover letter from the Tetrahydrofuran Task
Force dated 04/10/2001. Study No. 9910151/99007.
BASF. (2001b) Tetrahydrofurane: 5-day inhalation study in female B6C3F1 mice vapor exposure. Study No.
9910151/99129.
Begriche, K; Igoudjil, A; Pessayre, D; et al. (2006) Mitochondrial dysfunction in NASH: causes, consequences and
possible means to prevent it. Mitochondrion 6(1): 1—28.
Billecke, SD; Draganov, R; Counsell, P; et al. (2000) Human serum paraoxonase (PON1) isozymes Q and R
hydrolyze lactones and cyclic carbonate esters. Drug Metab Dispos 28(11): 1335-1342.
Boorman, G; Dixon, D; Elwell, M; et al.. (2003) Assessment of hyperplastic lesions in rodent carcinogenicity
studies. Toxicol Pathol 31(6):709-10.
Brooke, I; Cocker, J; Delic, JL; et al. (1998) Dermal uptake of solvents from the vapor phase: an experimental study
in humans. Ann Occup Hyg 42:531-540.
97
DRAFT - DO NOT CITE OR QUOTE
-------
Cascorbi, I. (2006) Genetic basis of toxic reactions to drags and chemicals. Toxicol Lett 162(l):16-28.
Chen, TH; Kavanagh, TJ; Chang, CC; et al. (1984) Inhibition of metabolic cooperation in Chinese hamster V79 cells
by various organic solvents and simple compounds. Cell Biol Toxicol 1(1):155—171.
Chhabra, RS; Elwell, MR; Chou, B; et al. (1990) Subchronic toxicity of tetrahydrofuran vapors in rats and mice.
Fundam Appl Toxicol 14(2):338-345.
Chhabra, RS; Herbert, RA; Roycroft, JH; et al. (1998) Carcinogenesis studies of tetrahydrofuran vapors in rats and
mice. Toxicol Sci 41(2): 183-188.
Collins, JL; Patek, PQ; Colin, M. (1982) In vivo surveillance of tumorigenic cells transformed in vitro. Nature
299(5879): 169-171.
Couper, FJ; Marinetti, LJ. (2002) v-Hydroxybutyrate (GHB)—Effects on human performance and behavior.
Forensic Sci Rev 14(1): 101-121.
Cramp, KS. (1995) Calculation of benchmark doses from continuous data. Risk Anal 15(1):79—89.
Curvall, M; Enzell, CR; Pettersson, B. (1984) An evaluation of the utility of four in vitro short term tests for
predicting the cytotoxicity of individual compounds derived from tobacco smoke. Cell Biol Toxicol 1(1): 173-193.
Dammann, M. (2005) Statistical analysis of the THF (tetrahydrofuran) kidney carcinomas and adenomas of the male
rat based on the expert report by Dr. Gordon C. Hard dated March 14, 2005. BASF, Germany, May 9, 2005.
(unpublished report available through the IRIS Submission Desk)
Debeljuk, L; del Carmen Diaz, M; Maines, VM; et al. (1983) Prolonged treatment with g-aminobutyric acid (gaba)-
mimetic substances in prepubertal male rats. Arch Androl 10:239-243.
DeFeudis, FV; Collier, B. (1970). Conversion of y-hydroxybutyrate to y-aminobutyrate by mouse brain in vivo.
Experientia 26:1072-1073.
Dierickx, PJ. (1989) Cytotoxicity testing of 114 compounds by the determination of the protein content in HepG2
cell cultures. Toxicol In Vitro 3(3): 189—193.
Dietert, RR; Etzel, RA; Chen, D; et al. (2000) Workshop to identify critical windows of exposure for children's
health: Immune and respiratory systems work group summary. Environ Health Perspect 108(Suppl 3):483-490.
Doherty, JD; Hattox, SE; Snead, OC; et al. (1978) Identification of endogenous y-hydroxybutyrate in human and
bovine brain and its regional distribution in human, guinea pigs and rhesus monkey brain. J Pharmacol Exp Ther
207:130-139.
Doi, AM; Hill, G; Seely, J; et al. (2007). Alpha 2|i-globulin nephropathy and renal tumors in national toxicology
program studies. Toxicol Pathol 35(2):533-540.
Draganov, DI; Teiber, JF; Speelman, A; et al. (2005) Human paraoxonases (PON1, PON2, and PON3) are
lactonases with overlapping and distinct substrate specificities. J Lipid Res 46(6): 1239-1247.
Droz, PO; Wu, MM; Cumberlan, WG; et al. (1989) Variability in biological monitoring of solvent exposure. I.
Development of a population physiological model. Br J Ind Med 46:447-460.
Droz, PO; Berode, M; Jang, JY. (1999) Biological monitoring of tetrahydrofuran: contribution of a physiologically
based pharmacokinetic model. Am Ind Hyg Assoc J 60(2):243-248.
98
DRAFT - DO NOT CITE OR QUOTE
-------
DuPont Haskell Laboratory. (1979) Initial submission: acute inhalation toxicity with tetrahydrofuran in rats with
cover letter dated 061592 and attachments. E.I. DuPont de Nemours and Company, Newark, DE; HLR-848-79.
Submitted under TSCA Section 8ECP; EPA Document No. 88-920004255; NTIS No. OTS0540603.
DuPont Haskell Laboratory. (1980) Tetrahydrofuran (THF) inhalation: effect on the rat conceptus. E.I. DuPont de
Nemours and Company, Newark, DE; HLR-750-82. Submitted under TSCA Section 8ECP; EPA Document No. 88-
920001524; NTIS No. OTS0535908.
DuPont Haskell Laboratory. (1996a) Acute inhalation neurotoxicity study of tetrahydrofuran in rats. E.I. DuPont de
Nemours and Company, Newark, DE; Haskel Laboratory Report No. 548-94.
DuPont Haskell Laboratory. (1996b) 90-Day inhalation neurotoxicity study with tetrahydrofuran in rats, with cover
letter dated 11/26/96. E.I. DuPont de Nemours and Company, Newark, DE; HLR-97-96. Submitted under TSCA
Section 4; EPA Document No. 44635; NTIS No. OTS0558874.
DuPont Haskell Laboratory. (1998) 14C-Tetrahydrofuran: disposition and pharmacokinetics in rats and mice, with
cover letter dated 10/14/1998. E.I. DuPont de Nemours and Company, Newark, DE; HLR-1998-01377. Submitted
under TSCA Section 8D; EPA Document No. 86990000002; NTIS No. OTS0573852.
DuPont Haskell Laboratory. (2000) Tetrahydrofuran: comparative in vitro microsomal metabolism. E.I. DuPont de
Nemours and Company, Newark, DE; DuPont-1103.
Dusdiecker, LB; Booth, BM; Stumbo, PJ; et al. (1985) Effect of supplemental fluids on human milk production. J
Pediatr 106(2):207-211.
Edling, C. (1982) Interaction between drugs and solvents as a cause of fatty change in the liver? Br JIndMed
39(2):198-199.
Eldefors, S; Ravn-Jonsen, A. (1992) Effect of organic solvents on nervous cell membrane as measured by changes
9+ ?+
inthe(Ca /Mg ) ATPase activity and fluidity of synaptosomal membrane. Pharmacol Toxicol 70:181-187.
Elovaara, E; Pfaffli, P; Savolainen, H. (1984) Burden and biochemical effects of extended tetrahydrofuran vapour
inhalation of three concentration levels. Acta Pharmacol Toxicol (Copenli) 54(3):221-226.
El Sayed, YM; Sadee, W. (1983) Metabolic activation of i?,5,-l-(tetrahydro-2-furanyl)-5-fluorouracil (ftorafur) to
5-fluorouracil by soluble enzymes. Cancer Res 43:4039-4044.
Emmett, E A. (1976) Parosmia and hyposmia induced by solvent exposure. Br J Ind Med 33(3): 196-198.
Eustis, SL; Hailey, JR; Boonnan, GA; et al. (1994) The utility of multiple sampling in the histopathological
evaluation of the kidney for carcinogenicity studies. Toxicol Pathol 22(5), 457-472.
Fenner-Crisp, P. (2007). Tetrahydrofuran tumor mode of action ~ human relevance analysis (Sept. 2007) (submitted
for publication) (Task Force Report). A final draft will be submitted to EPA by the Task Force when the report has
been accepted for publication.
Ferrara, SD; Giorgetti, R; Zancaner, S; et al. (1999) Effects of single dose of gamma-hydroxybutyric acid and
lorazepam on psychomotor performance and subjective feelings in healthy volunteers. Eur J Clin Pharmacol
54:821-827.
Florin, I; Rutberg, L; Curvall, M; et al. (1980) Screening of tobacco smoke constituents for mutagenicity using the
Ames' test. Toxicology 18(3):219-232.
Fujita, T; Suzuoki, Z. (1973). Enzymatic studies on the metabolism of the tetrahydrofuran mercaptan moiety of
thiamine tetrahydrofurfuryl disulfide. III. Oxidative cleavage of the tetrahydrofuran moiety. J Biochem 74:733-738.
99
DRAFT - DO NOT CITE OR QUOTE
-------
Funes-Cravioto, F; Zapata-Gayon, C; Kolmodin-Hedman, B; et al. (1977) Chromosome aberrations and sister-
chromatid exchange in workers in chemical laboratories and a rotoprinting factory and in children of women
laboratory workers. Lancet 2(8033):322-325.
Gallimberti, L; Ferri, M; Ferrara, SD; et al. (1992) Gamma-hydroxybutyric acid in the treatment of alcohol
dependence: a double-blind study. Alcohol Clin Exp Res 16(4):673-676.
Gallimberti, L; Cibin, M; Pagnin, P; et al. (1993) Gamma-hydroxybutyric acid for treatment of opiate withdrawal
syndrome. Neuropsychophannacology 9(1):77—81.
Galloway, SM; Armstrong, MJ; Reuben, C; et al. (1987) Chromosome aberrations and sister chromatid exchanges in
Chinese hamster ovary cells: evaluations of 108 chemicals. Environ Mol Mutagen 10(Suppl. 10): 1-175.
Gamer, AO; Jaeckh, R; Leibold, E; et al. (2002) Investigations on cell proliferation and enzyme induction in male
rat kidney and female mouse liver caused by tetrahydrofuran. Toxicol Sci 70:140-149.
Gamier, R; Rosenberg, N; Puissant, JM; et al. (1989) Tetrahydrofuran poisoning after occupational exposure. Br J
Ind Med 46(9):677-678.
Gibson, DP; Brauninger, R; Shaffi, HS; et al. (1997) Induction of micronuclei in Syrian hamster embryo cells:
comparison to results in the SHE cell transformation assay for National Toxicology Program test chemicals. Mutat
Res 392(l-2):61-70.
Gibson, KM; Sweetman, L; Nyhan, WL; et al. (1983) Succinic semialdehyde dehydrogenase deficiency: an inborn
error of gamma-aminobutyric acid metabolism. Clin Cliim Acta 133:33—42.
Ginsberg, G; Smolenski, S; Neafsey, P; et al. (2009). The influence of genetic polymorphisms on population
variability in six xenobiotic-metabolizing enzymes. J Toxicol Environ Health B 12(5-6):307-333.
Hageman, G; Kikken R; TenHoor, F; et al. (1988) Assessment of mutagenic activity of repeatedly used deep-frying
fats. Mutat Res 204(4):593-604.
Hara, K; Nagata, T; Kimura, K. (1987) Forensic toxicological analysis of tetrahydrofuran in body materials. Z
Rechtsmed 98:49-55.
Hard, GC; Alden, CL; Stula, EF; et al. (1995) Proliferative lesions of the kidney in rats. In: Guides for Toxicologic
Pathology. STP/ARP/AFIP, Washigton, DC, pp. 1-19.
Hard, GC. (2002). Significance of the renal effects of ethyl benzene in rodents for assessing human carcinogenic
risk. Toxicol Sci 69(1):30-41.
Hard, GC. (2005) Expert report on renal histopathology induced in F344 rats in subchronic toxicity and
carcinogenicity studies with tetrahydrofuran. Prepared for SOCMA Tetrahydrofuran Task Force, Washington DC,
March 14, 2005 (unpublished report available through the IRIS Submission Desk).
Hard, GC; Khan, KN. (2004). A contemporary overview of chronic progressive nephropathy in the laboratory rat,
and its significance for human risk assessment. Toxicol Pathol 32(2): 171-180.
Hard, GC; Seely, JC (2005) Recommendations for the interpretation of renal tubule proliferative lesions occurring in
rat kidneys with advanced chronic progressive nephropathy (CPN). Toxicol Pathol 33(6):641-649.
Hellwig, J; Gembardt, C; Jasti, S. (2002) Tetrahydrofuran: two-generation reproduction toxicity in Wistar rats by
continuous administration in the drinking water. Food Chem Toxicol 40(10): 1515-1523.
Henderson, VE; Smith, AHR. (1936) Anaesthetic effects of some furan derivatives. J Pharmacol Exp Ther 57:394-
398.
100
DRAFT - DO NOT CITE OR QUOTE
-------
Hofmann, HT; Oettel, H. (1954) Concerning the toxicity of tetrahydrofuran. Phannakologie 222:233-235.
Horiguclii, S; Teramoto, K; Katahira, T. (1984) Acute and repeated inhalation toxicity of tetrahydrofuran in
laboratory animals. Sumitomo Sangyo Eisei 20:141-157.
Horiuchi, K; Horiguclii, S; Utsunomiya, T; et al. (1967) Toxicity of an organic solvent, tetrahydrofuran, on the basis
of industrial health studies at a certain factory. Sumitomo Bull Ind Health 3:49-56.
Hossaini-Hilali, J; Benlamlih, S; Dahlborn, K. (1994) Effects of dehydration, rehydration, and hyperhydration in the
lactating and non-lactating black Moroccan goat. Comp Biochem Physiol A Physiol 109(4): 1017-1026.
Ikeoka, H; Nakai, Y; Ohaslii, Y; et al. (1988) Experimental studies on the respiratory toxicity of tetrahydrofuran in a
short term exposure. Sumitomo Sangyo Eisei 19:113-119.
Juntunen, J; Kaste, M; Harkonen, H. (1984) Cerebral convulsion after enfluran anaesthesia and occupational
exposure to tetrahydrofuran. J Neurol Neurosurg Psychiatry 47(11): 1258.
Kageyama, M. (1988) Exposure of humans to inhalation of tetrahydrofuran: elimination through expiration and
decay in alveolar air and blood. J Osaka-shi Igakkai Zasslii 37(1): 19-33.
Katahira, T; Teramoto, K; Horiguclii, S. (1982) Experimental studies on the toxicity of tetrahydrofuran administered
to animals by repeated inhalation. Jpn J Ind Health 24:379-387.
Kaufman, EE; Nelson, T. (1987) Evidence for the participation of a cytosolic NADP+-dependent oxidoreductase in
the catabolism of y-hydroxybutyrate in vivo. J Neurochem 48:1935-1941.
Kaufman, EE; Nelson, T; Goochee, C; et al. (1979) Purification and characterization of an NADP+-linked alcohol
oxido-reductase which catalyzes the interconversion of y-Hydroxyl-butyrate and succinic semialdehyde. J
Neurochem 32:699-712.
Kavlock, RJ; Allen, BC; Faustman, EM; et al. (1995) Dose-response assessment for development toxicity IV.
Benchmark dose for fetal weight changes. Fundam Appl Toxicol 26:211-222.
Kawata, F; Ito, A. (1984) Experimental studies on the effects of organic solvents in living bodies: Changes of
tetrahydrofuran concentration in rats' organs and histological observations after inhalation. Nihon Hoigaku Zasslii
8(3):367-375.
Kawata, F; Sliimizu, T; Ozono, S. (1986) Determination and fluorescent-liistochemical approach to catecholamines
in the rat brain after inhalation of tetrahydrofuran. Nihon Hoigaku Zasslii 40(6):811-820.
Kerckaert, GA; Brauninger, R; LeBouef, RA; et al. (1996) Use of the Syrian hamster embryo cell transformation
assay for carcinogenicity prediction of chemicals currently being tested by the National Toxicology Program in
rodent bioassays. Environ Health Perspect 104(Suppl 5): 1075-1084.
Kimura, ET; Ebert, DM; Dodge, PW. (1971) Acute toxicity and limits of solvent residue for sixteen organic
solvents. Toxicol Appl Pharmacol 19(4):699-704.
Klaunig, JE; Ruch, RJ; DeAngelo, AB; et al. (1998) Inhibition of mouse hepatocyte intercellular communication by
phthalate monoesters. Cancer Lett 43(l-2):65-71.
Kobayaslii, K; Urasliima, K; Sliimada, N; et al. (2002) Substrate specificity for rat cytochrome P450 (CYP)
isofonns: Screening with cDNA-expressed systems of the rat. Biochem Pharmacol 63(5):889-896.
Komsta, E; Chu, I; Secours, VE; et al. (1988) Results of a short-term toxicity study for three organic chemicals
found in Niagara River drinking water. Bull Environ Contain Toxicol 41(4):515-522.
101
DRAFT - DO NOT CITE OR QUOTE
-------
Kronevi, T; Holmberg, B; Arvidsson, S. (1988) Teratogenicity test of y-butyrolactone in the Sprague-Dawley rat.
Pharmacol Toxicol 62:57-58.
LaBelle, CW; Brieger, H. (1955) The vapor toxicity of a compound solvent and its principal components. Arch Ind
Health 12:623-627.
Letteron, P; Fromenty, B; Terris, B; et al. (1996) Acute and chronic hepatic steatosis lead to in vivo lipid
peroxidation in mice. J Hepatol 24(2):200-208.
Little, W; Collis, KA; Gleed, PT; et al. (1980) Effect of reduced water intake by lactating dairy cows on behavior,
milk yield and blood composition. VetRec 106(26):547-51.
Lock, EA; Hard, GC (2004) Chemically induced renal tubule tumors in the laboratory rat and mouse: Review of the
NCI/NTP database and categorization of renal carcinogens based on mechanistic information. Crit Rev Toxicol
34(3):211-299.
Lopez, V; Falco, C; Mori, G; et al. (1999) Apoptosis is regulated by polyamines in the cell cycle of Chinese hamster
ovary cells. Biocell 23(3):223-228.
Luster, MI; Portier, C; Pait, DG; et al. (1992) Risk assessment in immunotoxicity: I. Sensitivity and predictability of
immune tests. Fundam Appl Toxicol 18:200-210.
Luster, MI; Portier, C; Pait, DG; et al. (1993) Risk assessment in immunotoxicology: II. Relationships between
immune and host resistance tests. Fundam Appl Toxicol 21:71-82.
Malley, LA; Christoph, GR; Stadler, JC; et al. (2001) Acute and subchronic neurotoxicology evaluation of
tetrahydrofuran by inhalation in rats. Drug Chem Toxicol 24(3):201-219.
Marcus, RJ; Winters, WD; Hultin, E. (1976) Neurophannacological effects induced by butanol, 4-hydroxybutyrate,
4-mercaptobutyric acid thiolactone, tetrahydrofuran, pyrrolidine, 2-deoxy-d-glucose and related substances in the
rat. Neuropharmacology 15(1):29—38.
Maron, D; Katzenellenbogen, J; Ames, BN. (1981) Compatibility of organic solvents with the salmonella/
microsome test. Mutat Res 88(4):343-350.
Mast, TJ; Weigel, RJ; Westerberg, RB; et al. (1992) Evaluation of the potential for developmental toxicity in rats
and mice following inhalation exposure to tetrahydrofuran. Fundam Appl Toxicol 18(2):255-265.
Matthews, EJ; Spalding, JW; Tennant, RW. (1993) Transformation of BALB/c-3T3 cells: V. Transformation
responses of 168 chemicals compared with mutagenicity in salmonella and carcinogenicity in rodent bioassays.
Environ Health Perspect 101(Suppl. 2):347-482.
McMahon, RE; Cline, JC; Thompson. CZ. (1979) Assay of 855 test chemicals in ten tester strains using a new
modification of the Ames test for bacterial mutagens. Cancer Res 39:682-693.
Metcalf, DR; Emde, RN; Stripe, JT. (1966) An EEG-behavioral study of sodium hydroxybutyrate in humans.
Electroencephalogr Clin Neurophysiol 20:506-512.
Miotto, K; Darakjian, J; Basch, J; et al. (2001) Gamma-hydroxybutyric acid: patterns of use, effects and withdrawal.
Am J Addict 10:232-241.
Mirsalis, J; Tyson, K; Beck, J; et al. (1983) Induction of unscheduled DNA synthesis (UDS) in hepatocytes
following in vitro and in vivo treatment. Environ Mutagen 5:482.
102
DRAFT - DO NOT CITE OR QUOTE
-------
Morse, JM; Ewing, G; Gamble, D; et al. (1992) The effect of maternal fluid intake on breast milk supply: a pilot
study. Can J Public Health 83(3):213-216.
Mortelmans, K; Haworth, S; Lawlor, T; et al. (1986) Salmonella mutagenicity tests: II. Results from the testing of
270 chemicals. Environ Mutagen 8(Suppl. 7): 1-119.
Nelson, T; Kaufman, E; Mine, J; et al. (1981) The extraneural distribution of y-hydroxybutyrate. J Neurochem
37(5): 1345—1348.
Nimmerrichter, AA; Walter, H; Gutierrez-Lobos, KE; et al. (2002) Double-blind controlled trial of
y-hydroxybutyrate and clomethiazole in the treatment of alcohol withdrawal. Alcohol Alcohol 37(l):67-73.
NIOSH (National Institute for Occupational Safety and Health). (1991) Health hazard evaluation report, Flexlab,
Inc., Hastings, Michigan. Hazard Evaluations and Technical Assistance Branch, NIOSH, U.S. Department of
Health and Human Services, Cincinnati, OH; Report No. HETA 89-267-2139.
NIOSH. (1997) NIOSH pocket guide to chemical hazards. Washington, DC: U.S. Government Printing Office.
Available online at http://www.cdc.gov/niosli/npg/ (accessed July 29, 2009).
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press.
NSF (National Sanitation Foundation). (2003) Gamma-butyrolactone CASRN 96-48-0 oral risk assessment
document [draft]. Available online at http://aec.ilis.com/collections/nsf/index.htm (accessed July 29, 2009).
NTP (National Toxicology Program). (1992) Toxicology and carcinogenesis studies of gamma-butyrolactone (CAS
No. 96-48-0) in F344/N rats and B6C3F1 mice. Public Health Service, U.S. Department of Health and Human
Services; NTP TR- 406. Available from the National Institute of Enviromnental Health Services, Research Triangle
Park, NC and online at http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr406.pdf (accessed July 28, 2009).
NTP. (1998) Toxicology and carcinogenesis studies of tetrahydrofuran (CAS No. 109-99-9) inF344/N rats and
B6C3F1 mice. Public Health Service, U.S. Department of Health and Human Services; NTP TR- 475. Available
from the National Institute of Enviromnental Health Services, Research Triangle Park, NC.
Ohaslii, Y; Nakai, Y; Nakata, J; et al. (1983) Effects on the ciliary activity and morphology of rabbit's nasal
epithelium exposed to tetrahydrofuran. Osaka City Med J 29(1): 1—14.
Ong, CN; Cilia, SE; Phoon, WH; et al. (1991) Biological monitoring of occupational exposure to tetrahydrofuran.
Br J Ind Med 48(9):616-621.
Pellizzari, ED; Hartwell, TD; Harris, BS; et al. (1982) Purgeable organic compounds in mother's milk. Bull
Environ Contain Toxicol 28(3):322-328.
Pettersson, B; Curvall, M; Enzell, CR. (1982) Effects of tobacco smoke compounds on the ciliary activity of the
embryo chicken trachea in vitro. Toxicology 23(1):41—55.
Popjak, G. (1945) Lipids of the human kidney cortex and medulla in fatty change. J Pathol 57:87-100.
Pozdnyakova, AG. (1965) TrLeningr Sanit Gig Med 81:91-96.
RIVM (Dutch National Institute for Public Health and the Enviromnent). (2001) Re-evaluation of human-
toxicological maximum permissible risk levels. RIVM, National Institute of Public Health and the Environment
Bilthoven, The Netherlands; RIVM Report No. 711701 025; p. 276..
Root, B. (1965) Oral premedication of children with 4-hydroxybutyrate. Anesthesiology 26:259-260.
103
DRAFT - DO NOT CITE OR QUOTE
-------
Roth, RH; Giarman, NJ. (1966) Gamma-butyrolactone and gamma-hydroxybutyric acid -1. Distribution and
metabolism. BiochemPharmacol 15:1333-1348.
Roth, RH; Giarman, NJ. (1968) Evidence that central nervous system depression by 1,4-butanediol is mediated
through a metabolite, gamma-hydroxybutyrate. Biochem Pharmacol 17:735-739.
Sawyer, TW; Baer-Dubowska, W; Chang, K; et al. (1988) Tumor-initiating activity of the bay-region dihydrodiols
and diol-epoxides of dibenz[a,j]anthracene and cholanthrene on mouse skin. Carcinogenesis 9(12):2203-2207.
Scharf, MB; Hauck, M; Stover, R; et al. (1998) Effect of gamma-hydroxybutyrate on pain, fatigue, and the alpha
sleep anomaly in patients with fibromyalgia. Preliminary report. J Rheumatol 25:1986-1990.
SRC (Syracuse Research Corporation). (2001) Enviromnental Fate Data Base. SRC, North Syracuse, New York.
Available online at http://www.syrres.com/esc/efdb.htm (accessed July 29, 2009).
Stasenkova, KP; Kochetkova, TA. (1963) The toxicity of tetrahydrofuran. Toksikol NovuknProm Kliim 5:21-34.
Stoughton, RW; Robbins, BH. (1936) The anesthetic properties of tetrahydrofurane. J Pharmacol Exp Ther 58:171—
173.
Tabib, A; Bachrach, U. (1999) Role of polyamines in mediating malignant transformation and oncogene expression.
Int J Biochem Cell Biol 31:1289-1295.
Takizawa, N; Tanaka, M; Liu, Z; et al. (2003) A dissociation of gamma-butyrolactone-induced absence seizure and
CRE- and AP-1 DNA-binding activities in the developing rat brain. Neurosci Res 45 (4):483-490.
Tassaneeyakul, W; Veronese, ME; Birkett, DJ; et al. (1993) Validation of 4-nitrophenol as an in vitro substrate
probe for human liver CYP2E1 using cDNA expression and microsomal kinetic techniques. Biochem Pharmacol
46(11): 1975—1981.
Teiber, JF; Draganov, DI; La Du, BN. (2003) Lactonase and lactonizing activities of human serum paraoxonase
(PON1) and rabbit serum PON3. Biochem Pharmacol 66:887-896.
Teramoto, K; Wakitani, F; Tanaka, H; et al. (1989) Elimination of acetone, 2-propanol, styrene and tetrahydrofuran
via exhaled air in rats. Toxicol Sci 14:325.
U.S. EPA (Enviromnental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
mixtures. FedRegist 51 (185):34014-34025.
U.S. EPA. (1986b) Guidelines for mutagenicity risk assessment. Fed Regist 51(185):34006-34012.
U.S. EPA. (1988) Recommendations for and documentation of biological values for use in risk assessment.
Enviromnental Criteria and Assessment Office, Office of Health and Enviromnental Assessment, Cincinnati, OH;
EPA/600/6-87/008. Available from the National Technical Information Service, Springfield, VA; PB88-179874/AS
and online at http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=34855#Download.
U.S. EPA. (1991a) Guidelines for developmental toxicity risk assessment. Fed Regist 56(234):63798-63826.
U.S. EPA. (1991b) Alpha 2u-globulin: association with chemically induced renal toxicity and neoplasia in the male
rat. Risk Assessment Forum, Washington, DC; EPA/625/3-91/019F.
U.S. EPA. (1994a) Interim policy for particle size and limit concentration issues in inhalation toxicity: notice of
availability. FedRegist 59(206):53799.
U.S. EPA. (1994b) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry. Enviromnental Criteria and Assessment Office, Office of Health and Enviromnental Assessment,
104
DRAFT - DO NOT CITE OR QUOTE
-------
Cincinnati, OH; EPA/600/8-90/066F. Available from the National Technical Information Service, Springfield, VA,
PB2000-500023, and online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=71993 (accessed July 28,
2009).
U.S. EPA. (1995) Use of the benchmark dose approach in health risk assessment. Risk Assessment Forum,
Washington, DC; EPA/630/R-94/007. Available from the National Technical Information Service, Springfield, VA,
PB95-213765, and online at http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive
(accessed July 28, 2009).
U.S. EPA. (1996) Guidelines for reproductive toxicity risk assessment. Fed Regist 61(212):56274-56322.
U.S. EPA. (1998) Guidelines for neurotoxicity risk assessment. Fed Regist 63(93):26926-26954.
U.S. EPA. (2000a) Science policy council handbook: risk characterization. Office of Science Policy, Office of
Research and Development, Washington, DC. EPA/100-B-00-002.
U.S. EPA. (2000b) Benchmark dose technical guidance document [external review draft]. Risk Assessment Forum,
Washington, DC; EPA/630/R-00/001.
U.S. EPA. (2000c) Supplementary guidance for conducting health risk assessment of chemical mixtures. Risk
Assessment Forum, Washington DC; EPA/630/R-00/002.
U.S. EPA. (2001) Help manual for benclunark dose software version 1.3. Office of Research and Development,
Washington, DC; EPA 600/R-00/014F.
U.S. EPA. (2002) A review of the reference dose concentration and reference concentration processess. Risk
Assessment Forum, Washington DC; EPA/630/P-02/002F. Available online at
http://c!pub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive (accessed July 29, 2009).
U.S. EPA. (2005a) Guidelines for carcinogen risk assessment. Fed Regist 70(66):17765-18717.
U.S. EPA. (2005b) Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens. Risk
Assessment Forum, Washington DC; EPA/630/R-03/003F.
U.S. EPA. (2006a) Science policy council handbook: peer review. 3rd edition. Office of Science Policy, Office of
Research and Development, Washington, DC; EPA/100/B-06/002. Available online at
http://www.epa.gov/OSA/spc/2peerrev.htm (accessed July 29, 2009).
U.S. EPA. (2006b) A framework for assessing health risk of enviromnental exposures to children. National Center
for Enviromnental Assessment, Washington DC; EPA/600/R-05/093F. Available online at
http://c!pub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363 (accessed July 29, 2009).
U.S. EPA. (2008) Benclunark dose software (BMDS) version 2.0. Available through
http://www.epa.gov/ncea/bmds/.
Valencia, R; Mason, JM; Woodruff, RC; et al. (1985) Chemical mutagenesis testing in drosophila. III. Results of 48
coded compounds tested for the National Toxicology Program. Environ Mutagen 7(3):325-348.
van Himbergen, TM; van Tits, LJH; Roest, M; et al. (2006) The story of PON1: how an organophosphate-
hydrolysing enzyme is becoming a player in cardiovascular medicine. Netherlands J Med 64(2):34-38.
van Ravenzwaay, B; Gamer, AO; Leibold, E; et al. (2003) Effect of cytochrome P-450 inhibition on
tetrahydrofuran-induced hepatocellular proliferation in female mice. Arch Toxicol 77(8):459-464.
Vayer, P; Mandel, P; Maitre, M. (1985) Conversion of y-hydroxybutyrate to y-aminobutyrate in vitro. J Neurochem
45:810-814.
105
DRAFT - DO NOT CITE OR QUOTE
-------
Verschueren, K. (2001) Handbook of environmental data on organic chemicals. 4th edition. Vol. 2. New York, NY:
Wiley-Interscience; pp. 1971-1974.
Vickers, MD. (1969) Gammahydroxybutyric acid. Int Anesthesia Clinics 7:75-89.
Wagner, HM. (1974) Retention einiger Kohlenwasserstoffe bei der Inhalation. Ver Wasser Boden Lufthyg J Water,
Soil Air Hyg:225-229.
Weaver, RJ; Thompson, S; Smith, G; et al. (1994) A comparative study of constitutive and induced alkoxyresorufin
O-dealkylation and individual cytochrome P450 forms in cynomolgus monkey (Macaca fascicularis), human,
mouse, rat, and hamster liver microsomes. Biochem Pharmacol 47(5):763-773.
Woo, Y-T; Arcos, JC; Argus, MF; et al. (1977) Structural identification of /?-dio\anc-2-onc as the major urinary
metabolite of /?-dio\anc. Naunyn-Schmiedebergs Arch Pharmacol 299:283-287.
REFERENCES ADDEDD AFTER EXTERNAL PEER REVIEW
Bruner, RH; Greaves, P; Hard, GC; et al. (2010) Histopathologic changes in the kidneys of male F344 rats from a 2-
year inhalation carcinogenicity study of tetrahydrofuran: A pathology working group review and re-evaluation.
Regul Toxicol Pharmacol 58:100-105.
Hennida, SAS; Possari, EPM; Souza, DB; et al. (2006) 2'-Deoxyguanosine, 2'-deoxycytidine, and 2'-deoxy-
adenosine adducts resulting from the reaction of tetrahydrofuran with DNA bases. Chem Res Toxicol 19:927-936.
Loureiro, APM; Di Mascio, P; Gomes, O F; et al. (2000) ;ra«.v.;ra«.v-2.4-Dccadicnal-induced l,7V2-etheno-2'-
deoxyguanosine adduct formation. Chem Res Toxicol 13:601-609.
Loureiro, APM; Campos, IPdA; Gomes, OF; et al. (2004) Structural characterization of diastereoisomeric ethano
adducts derived from the reaction of 2'-deoxyguanosine with trans.irans-2.4-dccadicna 1. Chem Res Toxicol 17:
641-649.
Loureiro, APM; Campos, IPdA; Gomes, OF; et al. (2005) Structural characterization of an etheno-
2'-deoxyguanosine adduct modified by tetrahydrofuran. Chem Res Toxicol 18:290-299.
1
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
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
The Toxicological Review of Tetrahydrofuran has undergone a formal external peer
review performed by scientists in accordance with EPA guidance on peer review (U.S. EPA,
2006a; 2000a). 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 arranged by charge question 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.
EXTERNAL PEER REVIEWER 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.
(A) GENERAL CHARGE QUESTIONS
QUESTION Al. Is the Toxicological Review logical, clear, and concise? Has EPA
accurately, clearly, and objectively represented and synthesized the scientific evidence for
noncancer and cancer hazard?
Comments: Generally, the reviewers regarded the Toxicological Review for Tetrahydrofuran to
be comprehensive, clear, and concise. One of the reviewers commented that the presentation of
mode of action information was repetitious. An additional reviewer suggested subheadings be
added to better orient presentation by topic rather than study in Section 4.4.3 (Mode of Action
Studies). One reviewer also cited some errors and recommended clarification or changes to
Section 3.3 (Metabolism) and Figure 3-1. Several of the reviewers raised issues specifically
related to the derivation of the oral RfD, inhalation RfC, and the cancer assessment that were
repeated in their responses to the Chemical-Specific charge questions. These comments are
addressed under the relevant charge questions below.
EPA Response to Comments: Section 4.4.3 (now moved to Section 4.5.2 and Appendix C.2.2)
has been revised by adding subheadings to orient presentation by topic rather than study as
suggested by the reviewers. Some toxicity data unrelated to mode of action (e.g., uterine effects
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
in mice) have been moved to Section 4.2. Section 3.3 and Figure 3.1 have been revised for
clarification and additional information on THF metabolism has been included. In addition,
redundant text has been removed in several sections.
QUESTION A2. Please identify any additional studies that should be considered in the
assessment of the noncancer and cancer health effects of THF.
Comments: None of the reviewers provided any additional studies for consideration.
EPA Response to Comments: No response needed.
CHEMICAL-SPECIFIC CHARGE QUESTIONS
(B) ORAL REFERENCE DOSE (RfD) FOR TETRAHYDROFURAN
QUESTION Bl. A chronic RfD for THF has been derived from the oral drinking water
2-generation reproductive toxicity study (BASF, 1996; Hellwig et al., 2002) in rats. Please
comment on whether the selection of this study as the principal study has been scientifically
justified and transparently and objectively described in the document. Please identify and
provide the rationale for any other studies that should be selected as the principal study.
Comments: The reviewers agreed with the selection of the Hellwig et al. (2002)/BASF (1996)
study as the principal study for derivation of the chronic oral RfD. One reviewer did not provide
any response.
EPA Response to Comments: EPA agrees with the reviewers and retained the principal study
as selected.
QUESTION B2. Decreased F2 male pup body weight was selected as the most appropriate
critical effect. Please comment on whether the selection of this critical effect has been
scientifically justified and transparently and objectively described in the document. Please
provide detailed explanation. Please identify and provide the rationale for any other
endpoints that should be considered in the selection of the critical effect.
Comments: Two of the reviewers agreed with the selection of decreased F2 male pup body
weight. Three of the reviewers commented that decreased pup body weight gain represented a
minimally adverse or non-adverse effect and questioned the effect of maternal water
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
consumption on these effects. They also noted that the supporting data including delayed eye
opening were weak. However, these reviewers collectively agreed that given the database as a
whole and the available oral toxicity studies, decreased F2 male pup body weight was the most
appropriate critical effect that could be used to derive the RfD. One reviewer did not provide
any response to this question.
EPA Response to Comments: EPA agrees with the reviewers recommendations. Decreased
pup body weight gain may be related to alterations in normal neonatal development as
demonstrated by the associated developmental findings of delayed eye opening and increased
incidence of sloped incisors observed following oral exposure to THF. Data on the possible
relationship between decreased water intake in dams and decreased production of milk was not
provided in this study. As detailed in Section 4.3.1, the decreased gain in pup body weight is
supported by the statistically significant correlation between F2 pup body weight gain and
maternal THF intake after multivariable regression analyses were conducted to control for the
other possible confounding factors, namely average water intake and number of pups in each
litter. Thus, the observed responses in the pups appear to be related to THF exposure.
QUESTION B3. The chronic RfD has been derived utilizing benchmark dose (BMD)
modeling to define the point of departure (POD). All available models were fit to the
individual male and female and combined incidence data (F1 and F2 pup body weight
gain). Please comment on the appropriateness and scientific justification presented for
individual and combined body weights to obtain a data set for BMD modeling. Please
provide comments with regards to whether BMD modeling is the best approach for
determining the point of departure. Has the BMD modeling been appropriately conducted
and objectively and transparently described? Has the benchmark response selected for use
in deriving the POD been scientifically justified and transparently and objectively
described? Please identify and provide rationale for any alternative approaches (including
the selection of BMR, model, etc.) for the determination of the point of departure, and if
such approaches are preferred to EPA's approach.
Comments: All reviewers agreed that BMD modeling was the most appropriate approach to
derive the RfD and that the F1 and F2 pup weight gain data were suitable endpoints for deriving
BMD estimates. However, four of the reviewers recommended using 1 SD below the mean for
body weight gain instead of a 5% reduction in body weight gain as the BMR to establish the
POD, on the basis that a percentage reduction in body weight gain is an arbitrary choice
compared with a measure of effect that considers the variation among animals.
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
EPA Response to Comments: EPA agrees that use of a BMR of 1 SD for decreased pup weight
gain can provide a useful characterization of this continuous variable by defining the exposure at
which 10% of exposed animals would be expected to have body weights lower than -98% of the
control group [draft U.S. EPA Benchmark Dose Technical (iuidance Document (U.S. EPA,
2000b)]. However, a 1 SD reduction does not necessarily consider biological significance,
because in this case the adversity of a weight reduction of that size is not considered. EPA
agrees with reviewers that extreme percentiles of the control group, such as the 98th percentile,
may not be an adverse level of response. Some scientific consensus on adversity of body weight
decreases is available, as a 10% decrease in adult body weight has been a long-standing
convention for identifying maximum tolerated doses (e.g., NTP bioassay protocols). For
younger animals, a 5% change in fetal or pup weight has been considered to convey similar
biological significance (Kavlock et al., 1995). However, in response to the commenters,
modeling was conducted using 1 SD below the mean for pup body weight gain as the BMR. The
results can be found in Appendix B. Note that in this instance, the BMDisd is nearly identical to
the BMDio for the data considered because the SD was essentially 10% of the control mean.
QUESTION B4. Please comment on the selection of the uncertainty factors applied to the
POD for the derivation of the RfD. For instance, are they scientifically justified and
transparently and objectively described in the document?
Comments: There were differences in opinion among the reviewers on the selection of UFs.
Three of the reviewers agreed with the selection of the UFs applied to the POD for the derivation
of the RfD. One of the reviewers did not provide comments on the selection of UFs. Two
reviewers questioned the total UF citing relatively low toxicity observed following oral exposure
to THF. Specifically, one of these reviewers noted that the interspecies UF could be reduced
based on the fact that the water solubility of THF would make it unlikely that THF would be
absorbed and distributed differently in rodents compared to humans. The same reviewer also
suggested a reduction in the database UF by using the inhalation toxicity database to inform the
oral toxicity database. Another reviewer commented that both the inter- and intraspecies UFs
could be reduced based on the available biotransformation data. This reviewer indicated that the
Toxicological Review presents data suggesting that metabolism of THF does not have a role in
THF-induced toxicity (i.e., metabolism is not a rate-limiting step) and that there may not be a 10-
fold variability among individuals or among species. The reviewer, therefore, was of the opinion
that each of the intra- and interspecies UFs should be no more than 3. The reviewer also thought
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
that the oral database deficiency didn't warrant a UFD of 10 but rather a factor of 3 would suffice
since the RfD was based on a well conducted study and a very sensitive endpoint.
EPA Response to Comments: In response to the comments from the reviewers who questioned
the selection of UFs for the RfD, EPA re-evaluated the rationale for each of the UFs. Regarding
the UF for possible human variability, there is no information on the toxicokinetics of THF
following exposure by the oral route or on differential sensitivity of human populations to THF.
However, blood kinetics data were highly variable among volunteers exposed to THF by the
inhalation route (Kageyama, 1988 covered in Section 3.1.2). Additionally, the metabolism
section (Section 3.3) has been revised to include literature on THF metabolism and on the role
that lactonase (also known as PON1) may play in hydrolyzing GBL (a lactone intermediate) to
GHB. There is a wide inter-individual variation in PON1 concentration and activity (up to 13-
fold) and possibly in some CYP450 isoenzymes, which may be involved in the early steps of
oxidative metabolism of THF to GBL. It is not clear if and to what extent genetic variability in
expression and activity of PON1 and CYP450 may influence the kinetics of THF
biotransformation, and how, in turn, such variability might influence human risk to THF
exposure (see Section 4.8.3). Furthermore, no information is available on life-stage
susceptibility to THF exposure. Therefore, the default value of 10 for UFH was retained.
With respect to interspecies variability, there are some limited data by the inhalation
route of exposure suggesting qualitative toxicokinetic similarities between humans (Droz et al.,
1999; Ong et al., 1991; Kageyama, 1988) and rats (Elovaara et al., 1984; Kawata and Ito, 1984).
For instance, THF was rapidly excreted following repeated inhalation exposure in both species
with limited bioaccumulation (see Section 3). However, these data are not adequate to provide a
quantitative estimate of toxicokinetic differences. Also, the human inhalation exposure PBPK
model for estimating THF concentration in blood, breath, and urine (Droz et al., 1999) does not
account for the toxicokinetic and toxicodynamic variability in humans, and no similar PBPK
model has been identified in animals (see Section 3.6). Furthermore, there are no comparative
toxicokinetic or toxicodynamic studies following exposure to THF by the oral route in humans
and animals. Thus, a UFA of 10 was retained to account for interspecies differences.
The comments on the database UF are addressed below in response to comments to
Charge Question B6.
QUESTION B5. A two-generation reproductive toxicity study was used for the selection of
the POD for the derivation of the RfD. Please comment on whether the rationale and
justification for not applying a subchronic to chronic uncertainty factor has been
scientifically justified and transparently described in the document.
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
Comments: Five of the reviewers agreed with the rationale and justification for not applying a
subchronic to chronic UF. One reviewer did not provide comments.
EPA Response to Comments: No response needed.
QUESTION B6. Please comment on whether the rationale and justification for the
selection of the database uncertainty factor has been scientifically justified and
transparently described in the document.
Comments: Two reviewers agreed with the selection of the database UF and stated that the
rationale and scientific justification for this selection was transparently described. Three
reviewers commented that the data suggest that the overall oral toxicity of THF is low and that
both the oral and inhalation toxicity data for THF should be utilized in the selection of the UFD,
thus reducing the UFD. One reviewer did not provide any response.
EPA Response to Comments: In response to the comments from the reviewers who suggested
utilizing both the oral and inhalation databases to inform the selection of the oral database UF,
the rationale for the UF was re-examined. The oral database for THF contains a two-generation
reproductive toxicity study and a range-finding one-generation reproductive study (Hellwig et
al., 2002; BASF, 1996, 1994). There are no available human occupational or epidemiological
studies or standard toxicity studies, including developmental toxicity studies, in animals. Based
on the limitations in the oral database for THF, alternative approaches for deriving the RfD were
considered (described in Section 5.1.1). These alternatives included the use of the inhalation
data and application of a route-to-route extrapolation approach or use of the oral data for
metabolites of THF. However, EPA concluded that both of these approaches were precluded by
deficiencies in the database. Thus, a database UF of 10 was retained in the derivation of the
RfD. The text in Section 5.1.3 has been augmented to include a more clear description of the
available database.
(C) INHALATION REFERENCE CONCENTRATION (RfC) FOR
TETRAHYDROFURAN
QUESTION CI. A chronic RfC for THF has been derived from data from a 105 week
chronic inhalation study (NTP, 1998) in mice and rats. Please comment on whether the
selection of this study as the principal study has been scientifically justified and
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
transparently and objectively described in the document. Please identify and provide the
rationale for any other studies that should be selected as the principal study.
Comments: All reviewers stated that they were supportive of EPA's selection of the 105-week
chronic inhalation study as the principal study to derive the RfC for THF.
EPA Response to Comments: Charge Question CI inaccurately states that the RfC for THF
was derived from data from a 105 week chronic inhalation study (NTP, 1998) in mice and rats.
The RfC derived in the external peer review draft was based on the 14-week subchronic NTP
(1998) study which identified both CNS and liver effects in mice. Both the 14-week and 105-
week studies are reported as NTP (1998).
The external peer review draft stated that based on clinical signs of CNS toxicity and
liver effects, a NOAEL of 1770 mg/m3 and a LOAEL of 5310 mg/m3 were identified. CNS
effects were observed at 5310 mg/m3and 14,750 mg/m3 in the subchronic study, and at 5310
mg/m3 in the chronic study. The draft also noted that THF induced a concentration-dependent
increase in liver weight in male and female mice and rats and centrilobular cytomegaly in male
and female mice in the subchronic study. The chronic study evaluated body weight and clinical
signs of toxicity and organs were subjected to histopathological examination at necropsy.
However, no measurements were taken for organ weights, hematology, or clinical chemistry. In
addition, the chronic study did not identify liver cytomegaly in any of the exposure groups (a
slight increase in necrosis was observed in the livers of the 5310 mg/m3 female mice). Thus, the
liver weights and histopathology (cytomegaly) cited in the discussion of the selection of the
principal study and critical effect in the external peer review draft were those observed in the
subchronic component of the NTP (1998) study. The document has been revised to further
clarify the effects reported by each study component (subchronic versus chronic exposure
duration) from NTP (1998) and the rationale for the selection of the principal study.
QUESTION C2. Liver toxicity and CNS effects were selected as the co-critical
toxicological effects. Please comment on whether the selection of this critical effect has
been scientifically justified and transparently and objectively described in the document.
Specifically, please address whether the selection of liver effects and CNS toxicity as the co-
critical effects instead of increased thymus weight has been adequately and transparently
described. Please identify and provide the rationale for any other endpoints that should be
considered in the selection of the critical effect.
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
Comments: Four of the reviewers agreed with the selection of liver toxicity and CNS effects as
co-critical effects. Two of these reviewers stated that the liver effects were minimally adverse,
but appropriate to select nonetheless. One of these two reviewers expressed preference for using
cytomegaly over increased liver weight as an endpoint while the second reviewer stated that
often cytomegaly may be a reversible effect and that, in the absence of other key effects, an
argument can be made against using the liver changes as a critical effect. Another reviewer
disagreed with the designation of liver toxicity as a co-critical effect stating that observations of
hepatomegaly, without well characterized events such as sustained cell proliferation or decreased
apoptosis, is a questionable critical effect. One reviewer did not provide comment.
Five of the reviewers agreed that thymus weight should not be used as a critical effect
since it was not accompanied by either histopathological changes or measured alterations in
immune competence. No other endpoints were identified by the reviewers as effects that should
be considered in the selection of the critical effect.
EPA Response to Comments: EPA agrees with the reviewers who indicated that the CNS and
liver effects were appropriate for use as the co-critical effects. In addition, Section 5.2.1 has
been augmented to include additional discussion of liver and CNS findings. In the subchronic
NTP (1998) study, liver weights (both absolute and relative) were increased in the 14,750 mg/m3
female rats and this finding was accompanied by increased serum bile acid concentration in the
absence of cholestatsis or hepatocellular necrosis. The study authors indicated that these changes
were consistent with decreased or altered hepatic function. In male mice, absolute and relative
liver weights were statistically significantly increased following exposure to concentrations of
>1,770 mg/m3. The increases in absolute and relative liver weights in male mice were
corroborated by increased incidence of centrilobular cytomegaly, statistically significant at
14,750 mg/m3 (7/10 compared to 0/10 in the control group). Also, relative and absolute liver
weights were statistically significantly increased in female mice beginning at 5,310 mg/m3 and
were accompanied by centrilobular cytomegaly (10/10 animals compared to 0/10 in controls) at
14,750 mg/m3. The hepatocytes were additionally described as having slight karyomegaly
(enlarged nucleus), increased cytoplasmic volume, and granular cytoplasm with less vacuolation
than that of midzonal and periportal hepatocytes (NTP, 1998). No clinical chemistry
measurements were performed in mice. The study authors concluded that the histopathological
changes observed in the high exposure male and female mice group suggested that the liver is the
target organ for toxicity. They also stated that the liver weight increase and mild
histopathological changes observed at the lower THF exposure concentration (5,310 mg/m3)
were consistent with a treatment related effect (NTP, 1998). Furthermore, in the chronic study,
liver necrosis was noted in female mice treated with 5,310 mg/m3 THF. Considering the
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
information described above as well as the supporting data in acute and short-term studies
(described in Appendix C), EPA concluded that liver effects and CNS effects are appropriate as
co-critical effects for derivation of the RfC. Section 5.2.3 includes the candidate PODs
associated with these effects as well as the potential RfCs (which are similar) for the liver and
CNS effects.
EPA agrees with the reviewers regarding thymus weight as inappropriate for use as a
critical effect for the derivation of the RfC.
QUESTION C3. The chronic RfC has been derived utilizing benchmark dose modeling to
define the point of departure (based on liver cytomegaly). BMD modeling was conducted
on liver weight and cytomegaly data in both males and females. Has the BMD modeling
been appropriately conducted and objectively and transparently described? Has the
benchmark response selected for use in deriving the POD been scientifically justified and
transparently and objectively described? Please provide comments on whether the
selection of a POD based on liver cytomegaly instead of liver weight is scientifically
justified and transparently described. Please identify and provide rationale for any
alternative approaches (including the selection of BMR, model, etc.) for the determination
of the point of departure, and if such approaches are preferred to EPA's approach.
Comments: All of the reviewers commented that the BMD modeling was appropriate. One
reviewer questioned the selection of liver cytomegaly rather than the liver weight modeling
results to define the POD for deriving the RfC. This reviewer stated that given the choice
between cytomegaly and increased liver weight data, the liver weight data may be a more
appropriate endpoint to model, but stated that liver weight changes not accompanied by cell
proliferation and/or apoptosis may not be representative of toxicity. This reviewer suggested that
the POD should be based on the CNS effects and a NOAEL/LOAEL approach using the CNS
effects was the preferred method for derivation of the RfC. In addition, one reviewer thought it
was unclear why only the male mouse data was modeled instead of the female mouse data or
both sexes combined. Two reviewers suggested expanding the explanation in Appendix B of the
AIC.
EPA Response to Comments: Further consideration of the BMD modeling and
NOAEL/LOAEL approaches described in Section 5.2.2, provides evidence of similar PODs for
both hepatocytomegaly and increased liver weight in the male mice as well as the POD (as a
NOAEL) for CNS effects. Sections 5.2.1 and 5.2.2 were revised to further discuss the
toxicological significance of the liver and CNS endpoints and to better characterize the modeling
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
and candidate PODs. The discussion of the selection of the POD for derivation of the RfC was
expanded in Section 5.2.2. Section 5.2.3 includes RfCs for both liver and CNS effects. The
selection of the male mouse data was based on the fact that the males were slightly more
sensitive (i.e., by about one dose-spacing unit) than females and text was added for clarification
to Section 5.2.2. Additional text has been added to Section 5.2.2 and Appendix B describing the
AIC.
QUESTION C4. No incidence data were presented for CNS effects. Thus, these data could
not be evaluated by BMD modeling. However, a NOAEL-LOAEL approach (based on the
CNS data) for the derivation of the RfC has been presented for comparison purposes.
Please provide comments as to whether the NOAEL-LOAEL approach based on the POD
for CNS effects is more appropriate for the derivation of the RfC. Please provide
comments with regards to whether BMD modeling is the best approach for determining the
point of departure.
Comments: Two reviewers considered the CNS effects as having greater toxicological
significance than the hepatic effects and therefore supported the use of a NOAEL/LOAEL
approach to derive the RfC. One of the reviewers commented that the NOAEL/LOAEL and
BMD modeling approaches yielded the same results and had a preference for the use of BMD
modeling. One reviewer agreed with the approach, analysis, and discussion and conclusions
presented in the Toxicological Review. One reviewer felt that both approaches were appropriate
and that confidence was increased by the fact that the approaches provided the same value.
Finally, one reviewer preferred the BMD modeling approach but agreed with the other reviewers
that the confidence was increased by the fact that the approaches provided the same value.
EPA Response to Comments: See responses to comments under Questions C2 and C3.
QUESTION C5. Please comment on whether the selection of the uncertainty factors
applied to the POD for the derivation of the RfCs. For instance, are they scientifically
justified and transparently and objectively described in the document.
Comments: A reviewer commented that the UF for interspecies differences should not be
reduced (UFA = 3). Specifically, the reviewer disagreed with the calculation of a human
equivalent concentration (HEC) to account for toxicokinetic differences between animals and
humans. The reviewer recommended that the UF not be reduced until it can be replaced with a
data-driven UF based on a physiologically-based pharmacokinetic model. Conversely, another
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
reviewer suggested that both the inter- and intraspecies UFs could be reduced based on the
available biotransformation data. This reviewer indicated that the Toxicological Review presents
data suggesting that metabolism of THF does not have a role in THF-induced toxicity (i.e.,
metabolism is not a rate-limiting step). Thus, there may not be a 10-fold variability among
individuals or among species. This reviewer also stated that the inhalation database was adequate
and that the available data were better documented than the oral database. In addition, some of
the reviewers commented that the total UF may be overly conservative and that additional
discussion should be added to the document to support a reduction of the overall uncertainty
factor. One reviewer provided no response to this question.
EPA Response to Comments: Regarding the intra- and interspecies UFs see response to
comment under Charge Question B4. In addition, the toxicokinetic component of the
interspecies UF is addressed by the HEC calculation according to EPA guidance (U.S. EPA,
1994b). No data are available to determine toxicodynamic differences between animals and
humans. Thus, an UF of 3 was retained to account for interspecies differences. The available
toxicity data following inhalation exposure to THF includes chronic inhalation bioassays in rats
and mice and an inhalation developmental toxicity study, but lacks multigeneration reproductive
toxicity studies. Based on the consideration of these areas of toxicity data gaps as discussed in
Section 5.2.3 and below, a UFD of 3 was retained for the derivation of the RfC.
QUESTION C6. Please comment on the transparency and scientific rationale and
justification for the selection of the database uncertainty factor. Please comment on
whether the application of the database uncertainty factor adequately represents the gap in
inhalation reproductive and developmental toxicity and immunotoxicity data for THF.
Please comment on whether the rationale for use of the oral data to inform this decision
scientifically justifiable and transparently described in the document.
Comments: Four of the reviewers agreed with the selection of the database UF of 3. One
reviewer stated that the explanation was transparent. Two reviewers commented that the lack of
immunotoxicity data may not be cause for concern. One reviewer specifically commented that
there was no evidence to indicate that lymphocyte cell populations would be selectively sensitive
to THF. In addition, this reviewer noted that cytotoxicity was not demonstrated in the available
mode of action studies for THF. This reviewer further suggested that due to the rapid
metabolism of THF, there was less concern for immunotoxicity at chronic low exposures to
THF. This reviewer suggested that secondary effects that may result from inflammatory
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
responses produced at high exposures would not be relevant to low exposures. One reviewer did
not provide any response to this question.
EPA Response to Comments: EPA agrees with the reviewers and has revised the text to
indicate that thymus effects observed following exposure to THF are not likely to represent a
specific uncertainly in the database. An uncertainty factor of 3 was retained to account for
deficiencies in the database (UFD) for THF.
QUESTION C7. THF induces a spectrum of effects consistent with both Category 1 and
Category 3 gases. Therefore, for the purposes of calculating HECs, respiratory tract effect
levels were calculated using the default equations for Category 1 gases and
extrarespiratory tract effect levels were calculated using default equations for Category 3
gases. Please comment on the explanation for the dosimetry choice in the derivation of the
RfC. Has the rationale been scientifically justified and transparently described?
Comments: Five reviewers agreed with EPA's dosimetry choices. One reviewer did not
respond. Two reviewers commented that this section could be improved by additional discussion
of the gas categories.
EPA Response to Comments: EPA agrees with the reviewers regarding the dosimetry choices.
Section 5.2.2.1 was revised to better characterize the approach used to calculate HECs for the
endpoints considered as the basis for the RfC as described under the RfC methodology (U.S.
EPA, 1994b). Detailed classification information for Category 1, 2, and 3 gases is also provided
in the EPA report cited (U.S. EPA 1994b).
(D) CARCINOGENICITY OF TETRAHYDROFURAN
QUESTION Dl. Under the EPA's 2005 Guidelines for Carcinogen Risk Assessment
(U.S.EPA, 2005), there is suggestive evidence for the human carcinogenic potential of THF.
Please comment on the scientific justification for the cancer weight of the evidence
characterization. A quantitative cancer assessment has been derived for THF. Do the data
support estimation of a cancer slope factor for THF? Please comment on the scientific
justification for deriving a quantitative cancer assessment considering the uncertainty in
the data and the suggestive nature of the weight of the evidence of carcinogenic potential.
Has the rationale and scientific justification for quantitation been transparently and
objectively described?
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
Comments: Five reviewers agreed with the "suggestive evidence of carcinogenic potential"
cancer descriptor. One reviewer did not comment. None of the reviewers disagreed with the
choice to derive a quantitative cancer assessment. Two reviewers commented that, while the
evidence for female mouse liver tumors can support quantitative estimation of cancer potency,
the dose-response of the male rat kidney tumor data was weak. One reviewer agreed with the
choice to quantify cancer risk but felt that a quantitative assessment would overestimate the risk
due to his opinion that THF is a very weak possible human carcinogen. Another reviewer noted
that the quantitative cancer assessment may provide a measure of the magnitude of the
carcinogenic concern. Several reviewers commented on the extrapolation approach utilized,
these comments are address under Charge Question D4.
EPA Response to Comments: The kidney and liver tumors in male rats and female mice,
respectively, support the qualitative characterization that there is suggestive evidence of
carcinogenic potential for THF. The utility of the quantitative cancer risk estimate is that it
characterizes the chemical's relative potency. For THF, the estimated PODs for kidney and liver
tumors (shown in Table 5-6) demonstrate the relative sensitivity of the two responses. The
response in female mice was more sensitive, and the response in the male rats is considered
supportive. The derivation of the inhalation cancer estimate is based on the female mouse liver
tumor data.
QUESTION D2. The available data suggest that a plausible mode of action for THF-
induced male rat kidney tumors may involve the accumulation of alpha-2u globulin. EPA
concluded that the available data do not provide significant biological support to establish
a mode of action for male rat kidney tumors and that these tumors are relevant to humans.
Please comment on the transparency and scientific rationale and justification for the
evaluation of these data and the conclusions regarding the possible mode(s) of action and
human relevance for the male rat kidney tumors.
Comments: Four reviewers agreed with the conclusions that the available data do not provide
significant biological support to establish a mode of action for male rat kidney tumors and that
these tumors are relevant to humans. Two reviewers did not comment.
EPA Response to Comments: No response needed.
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
QUESTION D3. The available data suggest that increased proliferation and promotion in
the liver may be a plausible mode of action for THF-induced female mouse liver tumors.
EPA concluded that the data do not provide significant biological support to establish a
mode of action for female mouse liver tumors and that these tumors are relevant to
humans. Please comment on the transparency and scientific rationale and justification for
the evaluation of these data and the conclusions regarding the possible mode(s) of action
and human relevance for the female mouse liver tumors.
Comments: Three of the reviewers agreed with the conclusion that the available data do not
provide significant biological support to establish a mode of action for female mouse liver
tumors and that the liver tumors are relevant to humans. One reviewer agreed that the mode of
action for THF-induced liver tumors is unknown, but suggested that chronic irritation (which this
reviewer considered as the most plausible mechanism of action for very low potency cancer-
causing chemicals) may be the mode of action of THF-induced liver tumors. The reviewer also
commented that the female mouse liver tumors were not relevant to humans because of the lack
of sufficiently high exposures and the very low incidence of liver cancer in humans compared to
B6C3F1 mice. Another reviewer's comments on the extrapolation approach are summarized and
addressed under Charge Question D4. One reviewer provided no response to this question.
EPA Response to Comments: As described in Sections 4.7.1 and 5.3.6, though it is possible
that THF may act as a tumor promoter, there is no information on potential precursor or key
events, and the possible role of chronic inflammation or specific mediators of tumorigenesis for
THF has not been examined. Thus, EPA maintains that in the absence of mode of action
information the mouse liver tumors are considered relevant.
QUESTION D4. An inhalation unit risk has been derived utilizing benchmark dose
modeling to define the point of departure of 10% extra risk followed by linear low-dose
extrapolation below the point of departure (i.e., the default assumption). Please comment
on the scientific justification and rationale supporting the estimation of an inhalation unit
risk from the available data for THF. Specifically, please comment on whether the
rationale for the quantitative analysis is objectively and transparently described,
considering the uncertainty in the data and the suggestive nature of the weight of evidence.
Please comment on the selection of linear low dose extrapolation. Has the justification of
linear low dose extrapolation been objectively and transparently presented? Please
identify and provide rationale for any alternative approaches for low dose extrapolation
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
that the data for THF would support and if such approaches are preferred to EPA's
approach.
Comments: One reviewer agreed with the selection of the linear extrapolation approach. Two
reviewers commented that the discussion of the decisions leading to the derivation of the
inhalation unit risk and the decision to select a linear low dose extrapolation was objectively and
transparently described. Another reviewer proposed survival adjustments to the tumor incidence
rates that should be considered in the dose-response modeling and derivation of the inhalation
unit risk. This reviewer presented several choices to consider for selecting dose-response
models. Four reviewers disagreed with the selection of a linear low dose extrapolation based on
the following reasons: THF is not genotoxic/DNA-reactive, its metabolism is rapid and doesn't
form a reactive metabolite, it doesn't cause irreversible damage, it induced a weak tumor
response at high doses, and it doesn't induce proliferative lesions considered to be pre-
neoplastic; concluding that the application of a nonthreshold model will overestimate cancer risk.
One reviewer also noted that all of the biological effects identified for THF are those which are
commonly thought to exhibit thresholds; this reviewer and another reviewer recommended using
a reference value approach to estimate a non-carcinogenic dose. One reviewer provided no
direct response to this question, although this reviewer provided relevant comments under
previous Charge Questions; these comments are incorporated above.
EPA Response to Comments: The reviewers that recommended a nonlinear approach
suggested that a nongenotoxic carcinogen would automatically have a nonlinear cancer response
at low dose. Very little data are available to inform the mode of action and no data are available
to indicate the shape of the dose-response curve at low exposures. If data were available to
better inform the mode of action, and the data were indicative of a threshold response, then a
reference value could be derived based on a precursor endpoint (i.e., key event in the mode of
action) and considered for the RfC. In the absence of such information and under the U.S. EPA
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), EPA concluded that the
available information does not establish a possible mode of action for THF and data are
insufficient to establish significant biological support for either a linear or nonlinear approach.
As such, a default linear extrapolation approach was applied and statements regarding a
nonlinear extrapolation approach were added to Section 5.3.2. In order to address the reviewers'
comments regarding the inclusion of a nonlinear approach, text was also been added to Section
5.3.2 to expand the discussion of the available biological support and the rationale for the
selection of the extrapolation approach. For THF, there were no noncancer effects reported that
could serve as a precursor endpoint upon which to base a nonlinear analysis. Thus, EPA
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
continues to recommend a default linear low-dose extrapolation approach in estimating
carcinogenic risk of THF to humans.
Addressing the differential survival among the female mice and male rats, as noted by
one reviewer, did affect the incidence rates (See Table 5-5). The PODs and associated unit risk
estimated from each data set were revised accordingly. EPA considered the model selection
options proposed by this reviewer, which included the multistage approach used in the external
peer review draft. As noted by the reviewer, the multistage model is supported by biological
plausibility through its parallelism to the multistage carcinogenic process, and has been widely
used for cancer risk assessments. In cases such as THF, where there is no biologically based
model available and insufficient support for considering nonlinear low-dose extrapolation, EPA
prefers to use the multistage model, which also maintains consistency across cancer assessments,
unless the multistage model does not fit the observed dose-response. In this case, a simpler
multistage model (one stage) fit the survival-adjusted data better than the fit offered by the
reviewer. EPA notes that the suggestions for simple or weighted BMDL averaging both produce
PODs that are not well-defined confidence limits; that is, for component BMDLs that are 95%
confidence limits, neither type of average leads to a 95% confidence limit for the combined
result, and the level of confidence is not easily determined. More comprehensive analysis
involving model averaging is a significant area of research currently, but scientific consensus
regarding how to implement model averaging has not yet been reached.
QUESTION D5. THF induces a spectrum of effects consistent with both Category 1 and
Category 3 gases. Therefore, for the purposes of calculating human equivalent
concentrations, respiratory tract effect levels were calculated using the default equations
for Category 1 gases and extrarespiratory tract effect levels were calculated using default
equations for Category 3 gases. Please comment on the explanation for the dosimetry
choice in the derivation of the inhalation unit risk. Has the rationale been scientifically
justified and transparently described?
Comments: Three reviewers agreed with the HEC calculations. Three reviewers did not
comment in response to this question.
EPA Response to Comments: No response needed.
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
PUBLIC COMMENTS
Comments: A commenter stated that the critical endpoint (decreased weight gain in F2 pups)
for deriving the RfD is weak and equivocal because the findings may have been due to reduced
maternal milk production associated with decreased water intake and/or food consumption. The
commenter also questioned the rationale behind the gender-specific effect on weight gain in
male, but not female, pups and expressed doubts about a direct association between this reduced
weight gain in male pups and THF exposures.
EPA Response to Comments: See response to charge question B2 in Appendix A. There are
no data evaluating the possible relationship between decreased water intake in dams and
decreased production of milk (i.e., milk output was not measured). However, after multivariable
regression analyses were conducted to control for possible confounding factors, including water
intake and number of pups in each litter, it was found that the decreased gain in F2 pup body
weight was statistically significantly correlated with maternal THF intake. Therefore, decreased
pup body weight gain is considered an appropriate endpoint for deriving the RfD. The study
authors concluded that the decreased pup weight gain could be a high concentration effect
reflecting general toxicity due to direct exposure to THF during lactation (Hellwig et al., 2002).
Specifically, the study authors suggested that, given that THF is slightly more soluble in lipid
than water, THF may have been more concentrated in the dam's milk fat than in the maternal
water compartment. Based on the developmental effects observed (decreased pup weight gain,
delayed eye opening, and increased incidence of sloped incisors) the study authors designated
3,000 ppm as the NOAEL. Maternal food consumption was marginally decreased in F0 and F1
lactating dams and the decrease was not associated with statistically significant reduction in
maternal body weight gain. Finally, there is no apparent gender-specific effect on weight gain in
both the F1 and F2 pups. While for F1 pups, the responses at each dose were different between
males and females, for F2 pups, the responses were comparable between males and females at all
doses. It is not clear if there is sex dependence for effects on F1 but not F2 pups or if there is a
biological basis or this difference reflects only statistical considerations. However, the BMD and
BMDL estimates for the combined F2 data were similar to the BMD and BMDL estimates
derived for either sex individually. Therefore, EPA concluded that data corresponding to the F2
females, described by the linear model, provided the best fit and the corresponding BMDLos
value of 303 mg/kg-day was used to derive the RfD.
Comments: A commenter stated that the composite UF (1,000) used to derive the RfD for THF
is excessive given the apparent toxicity of THF. Specifically, the commenter stated that an UF
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
of 10 to account for deficiencies in the oral database exaggerates the potential significance of
limitations in the oral data.
EPA Response to Comments: See response to Charge Question B4. As noted in Section 5.1.3,
an UFd of 10 was selected to account for deficiencies in the toxicity database for oral exposure to
THF. The oral database for THF contains a two-generation reproductive toxicity study and a
range-finding one-generation reproductive study (Hellwig et al., 2002; BASF, 1996, 1994).
There are no available human occupational or epidemiological studies or standard toxicity
studies, including developmental toxicity studies, in animals.
Comments: A commenter questioned the use of CNS depression as one of the critical endpoints
to derive the RfC because transient sedation from exposure to this volatile organic chemical is
reversible and does not by itself provide any evidence of sustained neurotoxicity.
EPA Response to Comments: EPA agreed with the peer reviewers' comments supporting the
use of CNS effects as a co-critical effect for the derivation of the RfC. Text was added to
Section 5.2.1 to further discuss significance of these effects and the rationale for the selection.
Comments: A commenter recommended using a combined UF of 30 rather than 100 to derive
the RfC for THF. The commenter stated that the inhalation database for THF is relatively robust
obviating the need for a database UF of 3. Among the reasons that were cited in support of this
view were that adult animals were more sensitive than fetuses or weanling animals, the offspring
findings in the oral two-generation study were unremarkable, and the absence of other
immunotoxicity findings (such as histopathology) that may lend support to the effect on thymus
weight.
EPA Response to Comments: Based on comments from the external peer reviewers, EPA has
revised the text to indicate that thymus effects observed following exposure to THF may not
represent an uncertainty in the database. An uncertainty factor of 3 was selected to account for
deficiencies in the database for THF. Chronic and subchronic inhalation bioassays and
developmental toxicity studies are available in rats and mice (NTP, 1998; Mast et al., 1992;
DuPont Haskell Laboratory, 1980). No two-generation reproductive toxicity study by the
inhalation route is available.
Comments: A commenter stated that the carcinogenicity data for THF, particularly the liver
tumor response in female mice, support at most a classification of suggestive evidence of
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
carcinogenic potential in humans. The commenter disagreed with EPA's determinations
regarding the mode of action for carcinogenicity. The commenter agreed with EPA's conclusion
that the currently available data do not clearly establish a,2u-globulin accumulation as mode of
action for kidney effects. The commenter suggested, however, that renal tumors may have been
related to THF-induced exacerbation of or interaction with CPN (Hard, 2005). The commenter
asserted that evidence suggests that CPN may not be relevant to humans (Hard and Khan, 2004)
and that there is a causal link between CPN, atypical tubule hyperplasia (ATH), and adenomas in
rats. The commenter stated that there is sufficient evidence to clearly establish CPN as the mode
of action for increased incidence of renal tumors in male rats based on two evaluations by the
Tetrahydrofuran Task Force (Fenner-Crisp, 2007; Hard, 2005). The commenter also cited a
recent NTP publication that evaluated a,2u-globulin-associated nephropathy and renal tumors in
rats (Doi et al., 2007) as support for the conclusion that there is a causal link between CPN and
proliferative lesions in the kidney.
EPA Response to Comments: EPA agreed with the peer reviewers' comments that the mode of
carcinogenic action for THF has not been established. Hard (2005) concluded that in the
chronically exposed control and high exposure male rat groups there were comparable incidences
of ATH (5/50 and 6/50, respectively). In addition, Hard (2005) reported that the treated male
and female group incidences and severity of CPN were almost identical to the respective male
and female control groups (Table 3).
There was no difference in the incidence or severity of CPN in male rats of the NTP 2-
year carcinogenicity study on THF (both the control and high-exposure groups have 13 males
with end-stage kidneys). Although THF did not exacerbate development of CPN, it was
postulated that it may have exacerbated the development of proliferative lesions within CPN-
affected tissue. No data in the peer-reviewed literature are available that support a role of CPN
in the induction of THF-induced kidney tumors in male rats.
Doi et al. (2007) concluded that it is possible that a,2u-globulin-associated nephropathy
may simply contribute to a weak background tumorigenic stimulus provided by age-related
chronic progressive nephropathy. However, the study authors stated that there is no direct
evidence for the histological alterations, including CPN and ATH, thought to be included with
a,2U-globulin nephropathy. The overall conclusions of Doi et al. (2007) were that the critical
component(s) of the nephropathy most closely associated with the development of tumors cannot
clearly be identified. As noted in Section 5.3.6, the mode of action of THF for the male rat
kidney tumors has not been determined.
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
Comments: In an unpublished report submitted in October 19, 2007 to EPA as part of the Public
Record, histology slides of kidneys from male and female F344/N rats of the 2-year
carcinogenicity and 14-week NTP studies (NTP, 1998) on THF were reexamined (Dammann,
2005; Hard, 2005). The authors of these unpublished reports suggested that the overall incidence
of kidney tumors in the male rats was 2/50 (4%), 1/50 (2%), 3/50 (6%), and 5/50 (10%), with all
tumors being adenomas (Hard, 2005). Use of the Cochran-Armitage trend test on the data
presented in the unpublished report showed no significant concentration-response trend in tumor
incidence (Dammann, 2005). The author also concluded that THF does not appear to act via the
a,2U-globulin mode of action. Instead, the author proposed that advanced CPN may play a role in
the development of ATH, and perhaps the kidney tumors from THF exposure.
Additional public comments were submitted to the IRIS Program on July 16, 2009.
Included in these comments was a report entitled: "Pathology Working Group Review of
Selected Histologic Changes in the Kidneys of Male Rats Assigned to a 2-Year Inhalation
Carcinogenicity Study of Tetrahydrofuran (NTP Study No. 05181-03)." The Pathology Working
Group (PWG) reevaluation was conducted during March 3-4, 2009 and included five voting
members including Dr. Gordon Hard who had conducted a previous evaluation of the same data.
For this discussion, the report will be referred to as PWG (2009). The specified objectives of the
new reevaluation were to establish the most appropriate diagnoses of proliferative kidney
changes; to provide comment on likely potential pathogenic mechanisms for male rat kidney
tumors; and to provide perspective on risk from potential human exposure to THF. In addition to
evaluating kidney sections from all the control and high concentration male rat groups, the PWG
(2009) examined kidneys that had proliferative lesions in the low and mid exposure male rat
groups (5 and 10, respectively). The criteria for proliferative changes were based on Hard et al.
(1995). The report by the PWG (2009) stated that the NTP pathologists consolidated all variants
of tubular hyperplasia under the diagnostic term "Renal Tubule Hyperplasia." In contrast, the
PWG (2009) differentiated between "simple" and "atypical" hyperplasia where, and according to
PWG (2009), the first was not recorded because it was regarded as a reactive tubular alteration
directly associated with CPN. Atypical tubular hyperplasia (ATH) was recorded by PWG (2009)
because it was considered to represent a potential pre-neoplastic lesion with strong relevance to
carcinogenicity, but severity grades were not assigned.
Both NTP (1998) and the PWG (2009) concluded that renal cell adenomas were
increased in the high exposure male rats compared to controls. The PWG (2009) considered
both preneoplastic and neoplastic lesions together and reported that when these effects were
combined, the incidence values were similar between treated and control rats. In this
determination, the PWG (2009) applied different criteria that distinguished between reactive
tubular hyperplasia (associated with CPN) and atypical tubule hyperplasia (deemed as
A-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
preneoplastic). The PWG (2009) concluded that adenomas and ATH were present in kidneys
that showed advanced CPN. Furthermore, they suggested that accelerated tubular cell
degeneration and regeneration associated with CPN was likely responsible for the development
of most proliferative lesions. The PWG (2009) indicated that THF-induced exacerbation of CPN
was not considered to be contributory because severity of CPN was similar between treated and
control rats. Additionally, the PWG (2009) report stated that there was no evidence of early
tumor occurrence or of tumor progression to carcinoma.
The report indicated that the PWG did not observe histological changes associated with
a,2U-globulin nephropathy in the chronic NTP (1998) cancer bioassay slides. However, the PWG
(2009) report did note that hyaline droplets were detected in the tubular epithelium of high
exposure male rats in the subchronic 14-week study, but that similar results were observed in
control males. The PWG (2009) also concurred with the results of the BASF 4-week inhalation
study. Specifically, hyaline droplets were increased in the proximal tubules and hot spots of
accelerated cell proliferation were identified in the cortex of male rats exposed to 1800 ppm THF
for 20 days They also noted that immunohistochemistry confirmed that the hyaline droplets
contained a,2u-globulin following the 4-week inhalation exposure. The PWG (2009) concluded
that these slight increases in cell proliferation associated with a2u-globulin may have contributed
to the development of adenomas in male rats exposed to the high THF concentration in the
chronic cancer bioassay.
The PWG (2009) report concluded that given the absence of data demonstrating
statistically significant differences in tumors and preneoplastic lesions, and the assertion that two
mechanisms (CPN and a,2u-globulin) likely resulted in the proliferative changes observed in the
kidney, which have no known counterpart in humans, the formation of renal tubule adenomas in
the 2-year carcinogencity THF study (NTP, 1998) have no relevance to humans.
EPA Response to Comments: EPA agreed with the peer reviewers' comments and continues to
conclude that the mode of carcinogenic action for hepatocellular and renal tumors is largely
unknown. Additional discussion of the role of CPN and ATH in the development of kidney
tumors in male rats observed following exposure to THF has been included in Section 4.7.3.1.
EPA disagrees with the characterization in the PWG (2009) report that renal tubule hyperplasia
in the NTP reports is a non-specific term for all variants of tubular hyperplasia and with the
approach by the PWG (2009) of combining ATH with neoplastic kidney findings for statistical
analyses. There was no difference in the incidence or severity of CPN in male rats of the NTP 2-
year carcinogenicity study on THF (both the control and high-dose groups have 13 males with
end-stage kidneys). Although THF did not exacerbate development of CPN, it was postulated
that it may have exacerbated the development of proliferative lesions within CPN-affected tissue.
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
33
34
Specifically, against a background of nephropathy that was uniform across all groups, there were
more renal tubular tumors in treated rats than in the controls, and those in the higher doses were
larger in size. Consideration should be given to the robustness and the gender specificity of the
renal tumor response. Thus, EPA concluded that the male rat kidney tumors were relevant to
humans and that the mode of action for these tumors has not been established.
Comments: A commenter asked why, after concluding in Section 5.3.1 of the external peer
review draft Toxicological Review that "quantitative analyses may be useful for providing a
sense of the magnitude of potential carcinogenic risk", EPA neglected to revisit this issue after
the quantitative cancer assessment was completed. The commenter also questioned why EPA
didn't provide greater specificity by discussing what the quantitative results may mean in terms
of magnitude of potential carcinogenic risk and whether the results appear sensible for the data.
The commenter added that quantitative risk assessment should be reserved for substances where
the evidence provides a greater scientific basis for concern, such as human evidence or a
confirmed genotoxic mechanism, or at least a clearly defined and established carcinogenic
response in multiple test species. The commenter also noted that THF has not been shown to be
carcinogenic in humans and has not been confirmed as genotoxic. The commenter added that the
male rat renal tumors should not be considered relevant to humans and the hepatic tumors are
significantly increased only in the high dose female mice. Based on these considerations, in
addition to the absence of any indication of age-dependent susceptibility, the commenter
concluded that the application of linear dose-response extrapolation results in an unduly
conservative and implausible cancer potency estimate for THF that is comparable in value to two
known human carcinogens, namely benzene and vinyl chloride.
EPA Response to Comments: In accordance with peer reviewer comments, EPA continued to
present a quantitative cancer assessment. See responses to comments under Charge Questions
D1 for discussion of the cancer descriptor and choice to perform a quantitative analysis, D2 and
D3 regarding modes of action and relevance to humans, and D4 for issues regarding the IUR
calculation and extrapolation approach.
Comments: A commenter listed three studies (shown below) that were not considered in the
draft THF Toxicological Review. Though not considered to likely materially affect the
conclusions, the commenter noted that the references should be cited and discussed in the THF
Toxicological Review.
A-22
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
•	Lehman (2005). Determination of the percutaneous absorption of THF, in vitro, using
human cadaver skin model, PRACS Inst., Ltd. (unpublished report provided as #5
supporting document with the comments).
•	Loureiro, AP; de Arruda Campos, IP; Gomes, OF; et al. (2005) Structural
characterization of an etheno-2'-deoxyguanosine adduct modified by tetrahydrofuran.
Chem Res Toxicol 18(2):290-299.
•	Hermida, SA; Possari, EP; Souza, DB; et al. (2006) 2'-Deoxyguanosine,
2'-deoxycytidine, and 2'-deoxyadenosine adducts resulting from the reaction of
tetrahydrofuran with DNA bases. Chem Res Toxicol 19(7):927-936.
EPA Response to Comments: Conclusions and summaries of the studies by Luoriero et al.
(2005) and Hermida et al. (2006) have been added to Section 4.5 (Genotoxicity Studies) and
Appendix C.2, respectively.
A-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
33
34
35
APPENDIX B. BMD MODELING
The THF data sets considered for dose-response modeling include both quantal and
continuous endpoints. EPA's BMDS version 2.0 (U.S. EPA, 2008 was used for the model fitting
and benchmark estimation.
Definition of the BMR and corresponding BMD and BMDL
Rationales for BMRs are provided in Section 5 (Section 5.1.2 for the RfD, Section 5.2.2
for the RfC, and Section 5.3.2 for the inhalation unit risk). For all of the quantal endpoints
analyzed here, cytomegaly and cancer incidence, the BMD and BMDL values were defined
based on BMR values of 10% extra risk. For the continuous endpoints, BMD and BMDL values
were defined using a BMR of 5% of the control mean for decreased pup body weight gain, and
10% of the control mean for increased liver weight. For the selected models, additional analyses
were carried out for a 1 SD change in the mean for comparison purposes.
For all of the BMD values estimated as described above, BMDL values were defined as
the 95% lower bound on the corresponding BMD. Confidence intervals were calculated using a
profile likelihood method.
Model Selection
For each noncancer endpoint, EPA guidance (US EPA, 2000b) was followed with regard
to the choice of model and BMDL to use as a POD:
1.	Models with an unacceptable fit (including consideration of local fit in the
low-dose region) are excluded.
2.	If the BMDL values for the remaining models for a given endpoint are within
a factor of 3, no model dependence is assumed, and the models are considered
indistinguishable in the context of the precision of the methods. The models
are then ranked according to the AIC, and the model with the lowest AIC is
chosen as the basis for the BMDL.
3.	If the BMDL values are not within a factor of 3, some model dependence is
assumed, and the lowest BMDL is generally selected as a reasonable
conservative estimate..
B-l DRAFT - DO NOT CITE OR QUOTE
-------
1	For each cancer endpoint, the multistage model was considered first. Models with stages
2	up to n-1, where n is the total number of groups, were applied. Among those with goodness-of-
3	fit p-values >0.05, the model with the most parsimonious fit was selected, based on whether
4	there was a statistically significant improvement in the overall fit when each additional stage was
5	added to the model. If no adequate fits for a cancer endpoint had been obtained with the
6	multistage model, the suite of dichotomous models in BMDS woud have been considered, but
7	this step was not necessary.
8
Table B-l. BMD modeling results for pup body weight gain in the Wistar
rat two-generation reproductive toxicity study
Endpoint and model
AICa
P-
Value/degre
e of freedom
BMDsd
(mg/kg-day)
BMDLsd
(mg/kg-day)
BMDo.os
(mg/kg-day)
BMDLo.os
(mg/kg-day)
Pup body weight gain F1 males, days 7-14
Linear
159.3
0.90/3
728
549
457
355
Polynomial (2-degree)
161.1
0.80/2


552
357
Polynomial (3-degree)
161.1
0.78/1


539
357
Power (power >1)
161.1
0.73/2


524
356
Pup body weight gain F1 females, days 7-14
Linear
179.1
0.75/3
923
658
513
376
Power (power >1)
180.8
0.65/2


646
383
Polynomial (2-degree)
180.8
0.61/1


662
382
Pup body weight gain F2 males, days 7-14
Linear
198.6
0.33/3
831
593
417
306
Power (power >1)
199.8
0.24/2


571
320
Polynomial (2-degree)
199.9
0.22/1


602
318
Pup body weight gain F2 females, days 7-14
Linear (and higher
order polynomial
models), Power
(power >1)
206.3
0.27/2
974
665
440
303
aAIC = Akaike Information Criterion (see Appendix B).
''BIVIDL = 95% lower bound of the BMD. Subscript denotes the specified benchmark response (BMR) level,
either 1SD from the control mean or 0.05 x (control mean).
Sources: Hellwig et al. (2002); BASF (1996).
9
10
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
Pup Body Weight Gain, F1 Male Rats (Hellwig et al., 2002; BASF, 1996)
Linear Model with 0.95 Confidence Level
Linear
BMDL
BMD
0	200	400	600	800	1000
dose
Polynomial Model.
BMDS Model Run
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose^2 + ...
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Signs of the polynomial coefficients are not restricted
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 =	2.1082
rho =	0 Specified
beta_0 =	17.8231
beta 1 = -0.00194978
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 )
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
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
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
alpha
beta_0
beta 1
alpha
1
1.7e-008
-7 e-010
beta_0
1.7e-008
1
-0.69
beta_l
-7 e-010
-0.69
1
Parameter Estimates
95.0% Wald Confidence Interval
Variable
alpha
beta_0
beta 1
Estimate
2.01936
17.8234
-0.00195114
Std. Err.
0.301028
0.206968
0.000358764
Lower Conf. Limit
1.42935
17 . 4178
-0.0026543
Upper Conf. Limit
2.60936
18.2291
-0.00124798
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 24
134	21
381	22
1071	23
17 . 8
17 . 5
17 . 2
15. 7
17 . 8
17 . 6
17 .1
15. 7
1.15
1. 55
1. 43
1. 65
1. 42
1. 42
1. 42
1. 42
-0.0807
-0.2
0. 396
-0.114
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A2:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model
A1
A2
A3
fitted
R
Log(likelihood)
-76.516795
-74 . 898382
-76.516795
-76.625032
-89.411989
Param's
5
8
5
3
2
AIC
163.033591
165.796764
163.033591
159.250064
182.823978
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: 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
B-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
39
40
41
42
43
44
45
46
47
48
49
Test	-2*log(Likelihood Ratio)	Test df	p-value
Test 1	29.0272	6	<.0001
Test 2	3.23683	3	0.3565
Test 3	3.23683	3	0.3565
Test 4	0.216473	2	0.8974
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
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 adeguately describe the data
Benchmark Dose Computation
Specified effect =	0.05
Risk Type	=	Relative risk
Confidence level =	0.95
BMD =	456.743
BMDL =	354.564
Specified effect =	1
Risk Type =	Estimated standard deviations from the control mean
Confidence level =	0.95
BMD =	728.313
BMDL =	548.965
B-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
39
40
41
Pup Body Weight Gain, F1 Female Rats (Hellwig et al., 2002; BASF, 1996)
Linear Model with 0.95 Confidence Level
5
Linear
18
17.5
17
16.5
16
15.5
15
BMDL
BMD
0	200	400	600	800	1000
dose
Polynomial Model
BMDS Model Run
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose^2 + ...
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Signs of the polynomial coefficients are not restricted
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 =	2.56358
rho =	0 Specified
beta_0 =	17.479
beta 1 = -0.00171244
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
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
alpha
beta_0
beta 1
1
3e-00 9
-5.8e-010
3e-00 9
1
-0.69
-5.8e-010
-0.69
1
Parameter Estimates
95.0% Wald Confidence Interval
Variable
alpha
beta_0
beta 1
Estimate
2.4664
17.4699
-0.00170284
Std. Err.
0.365643
0. 228021
0.000396487
Lower Conf. Limit
1.74975
17.023
-0.00247994
Upper Conf. Limit
3.18304
17.9168
-0.000925736
Table of Data and Estimated Values of Interest
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 24
134	21
381	23
1071	23
17 . 3
17 . 4
16. 9
15. 6
17 . 5
17 . 2
16.8
15. 6
1. 47
1.72
1. 66
1. 56
1. 57
1. 57
1. 57
1. 57
-0. 53
0. 462
0.241
-0.141
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A2:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi
Var{e(i)}
Mu + e(i)
Sigma^2
Likelihoods of Interest
Model
A1
A2
A3
fitted
R
Log(likelihood)
-86.288553
-85.974284
-86.288553
-86.575503
-94.973257
Param's
5
8
5
3
2
AIC
182.577106
187.948568
182.577106
179.151006
193.946514
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: 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
B-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
39
40
41
42
43
44
45
46
47
48
49
Test 1
Test 2
Test 3
Test 4
17.9979
0.628539
0.628539
0.57 3 9
0.006237
0.8899
0.8899
0.7505
The p-value for Test 1 is less than .05. There appears to be a
difference between resp'Orise andAor variances among the dase levels
It seems appropriate to' model the 'data
The p-value tor Test 2 is 'greater than .1. A h'OmogerieO'US variance
model appears t'O be appropriate here
The p-value tor Test 3 is greater than .1. The modeled variance appears
t'O be appropriate here
The p-value tor Test 4 is greater than .1. The model chO'Sen seems
t'O adeguately describe the data
E'enchmark DO'Se C'Oirputati'Ori
Specified effect =	0.05
Risk Typ'e	=	Relative risk
C'Onfidence level =	0.95
E'MD =	512.964
E'MDL =	375.515
Specified effect =	1
Risk Typ'e =	Estimated standard deviations frcm the coritro'l mean
C'Onf idence level =	0.95
E'MD =	922.271
E'MDL =	657.7 53
B-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
39
40
Pup Body Weight Gain, F2 Male Rats (Hellwig et al., 2002; BASF, 1996)
Linear Model with 0.95 Confidence Level
Linear
19
5
18
17.5
17
16.5
16
15.5
15
BMDL
BMD
14.5
0
200
400
600
800
1000
dose
Polynomial Model.
BMDS Model Run
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose^2 + ...
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Signs of the polynomial coefficients are not restricted
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 =	3.1932
rho =	0 Specified
beta_0 =	17.779
beta 1 = -0.00216123
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter (s) -rho
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
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 1
alpha
1
9. le-010
1.5e-010
beta_0
9. le-010
1
-0.71
beta_l
1.5e-010
-0.71
1
Parameter Estimates
Variable
alpha
beta_0
beta 1
Estimate
3.12675
17 .7486
-0.00212706
Std. Err.
0. 466108
0. 262989
0. 000493474
95.0% Wald Confidence Interval
Lower Conf. Limit
2 . 21319
17.2331
-0.00309425
Upper Conf. Limit
4.0403
18.264
-0.00115987
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 24
129 20
385 23
974	23
17 . 4
17 . 9
17
15. 6
17 . 7
17 . 5
16. 9
15. 7
1. 56
1. 98
1. 94
1. 67
1.77
1.77
1.77
1.77
-0.966
1. 08
0.191
-0.208
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A2:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model
A1
A2
A3
fitted
R
Log(likelihood)
-95.200286
-94.325125
-95.200286
-96.299723
-104.744931
Param's
5
8
5
3
2
AIC
200.400571
204.650249
200.400571
198.599446
213.489861
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: 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.
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
Tests of Interest
Test -2*log(Likelihood Ratio) Test df	p-value
Test 1 20.8396 6	0.00196
Test 2 1.75032 3	0.6258
Test 3 1.75032 3	0.6258
Test 4 2.19888 2	0.3331
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
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 adeguately describe the data
Benchmark Dose Computation
Specified effect =	0.05
Risk Type	=	Relative risk
Confidence level =	0.95
BMD =	417 . 21
BMDL =	306.394
Specified effect =	1
Risk Type =	Estimated standard deviations from the control mean
Confidence level =	0.95
BMD =	831.318
BMDL =	593.354
B-l 1
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
Pup Body Weight Gain, F2 Female Rats (Hellwig et al., 2002; BASF, 1996)
Linear Model with 0.95 Confidence Level
Linear
Polynomial Model
BMDS Model Run
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose^2 + ...
Dependent variable = Mean
Independent variable = Dose
Signs of the polynomial coefficients are not restricted
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
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
lalpha =	1.26529
rho =	0
beta_0 =	17.139
beta 1 = -0.00191944
Asymptotic Correlation Matrix of Parameter Estimates
lalpha	rho	beta_0	beta_l
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
lalpha
rho
beta_0
beta 1
1
-1
0. 018
-0.024
-1
1
-0.018
0. 024
0. 018
-0.018
1
-0. 67
-0.024
0. 024
-0. 67
1
Parameter Estimates
95.0% Wald Confidence Interval
Variable
lalpha
rho
beta_0
beta 1
Estimate
9.98254
-3.13077
17 .14
.00194859
Std. Err.
10.5226
3.76257
0.262826
. 000537889
Lower Conf. Limit
-10.6413
-10.5053
16.6249
-0.00300284
Upper Conf. Limit
30.6064
4.24372
17.6552
-0.000894352
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 24
129 19
385 23
974	23
17 . 2
17 .1
16
15. 4
17 .1
16. 9
16.4
15. 2
1. 5
1. 62
2 . 41
1. 84
1.72
1.76
1. 85
2 . 07
0.171
0. 523
-1. 01
0. 366
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A2:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*In(Mu(i)))
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model	Log(likelihood)	# Param's	AIC
A1 -98.759122	5	207.518244
A2 -95.606538	8	207.213077
A3 -96.919491	6	205.838983
fitted -99.134279	4	206.268557
R -105.763005	2	215.526009
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: 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
B-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
39
40
41
42
43
44
45
46
47
Test -2*log(Likelihood Ratio) Test df
p-value
Test 1
Test 2
Test 3
Test 4
20.3129
6.30517
2.62591
4.42957
0.002436
0.09767
0.269
0.1092
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 adeguately describe the data
Benchmark Dose Computation
Specified effect =	0.05
Risk Type	=	Relative risk
Confidence level =	0.95
BMD =	4 3 9.8 05
BMDL =	303.273
Specified effect =	1
Risk Type =	Estimated standard deviations from the control mean
Confidence level =	0.95
BMD =	973.7 47
BMDL =	664.723
B-14
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. BMCa modeling results for noncancer effects resulting from
subchronic inhalation exposure to THF
Male mice: liver weight
Model
AIC
/j-valuc
bmc1sd
bmcl1sd
BMC01b
BMCL01b
Power (unrestricted)
-190.1
0.81
374
80
783
246
Hill
-189.0
0.55
607
275
1030
502
Linear (and higher order polynomials)
-189.9
0.53
912
710
1390
1110
Male mice: centrilobular cytomegaly
Model
AIC
/j-valuc


BMC10
BMCL10
Gamma, Weibull (power >1)
22.72
1.0


948
266
Log-logistic (slope >1)
22.72
1.0


948
322
Logistic
23.04
0.66


1138
645
Multistage, degree 2 (coefficients_>0)
20.86
0.93


805
256
Probit
22.89
0.75


1061
602
Log-probit
22.72
1.0


948
358
Concentrations used in the modeling were the HECs in mg/m3.
bFor the liver weight endpoints, BMC0.i/BMCL0.i refers to 10% relative increase in control value. For liver
pathology, BMCh/BMCLio refers to 10% extra risk in incidence of centrilobular cytomegaly.
Source: Based on data from NTP (1998).
B-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
Liver Centrilobular Cytomegaly, Male Mice (NTP, 1998)
Multistage Model with 0.95 Confidence Level
0.6
0.4
0.2
Multistage
BMDL
500
1000
1500
dose
2000
2500
Polynomial Model
BMDS MODEL RUN
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*dose/sl-beta2*dose/N2 ) ]
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 = 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
Beta(2) = 1.76579e-007
B-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
39
40
41
42
43
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s)
have been estimated at
do not appear in the correlation matrix
-Background -Beta(l)
boundary point, or have been specified by the user, and
Beta(2)
Beta(2)
1
Variable
Background
Beta(1)
Beta(2)
Parameter Estimates
Estimate	Std. Err.
0	NA
0	NA
1.627 7 6e-007	7.75991e-008
MA - Indicates that this parameter has hit a bound
implied by some ineguality constraint and thus
has no standard error.
Analysis of Deviance Table
Model	Log(likelihood) Deviance Test DF
Full model	-9.35947
Fitted model	-9.43218	0.145417	2
Reduced model	-17.3975	16.076	2
AIC:	20.8644
P-value
0.9299
0.000323
Goodness of Fit
Dose	Est._Prob. Expected Observed
Chi/N2 Res.
0.0000
948.0000
2634.0000
Chi-sguare
0.0000
0.1361
0.6768
0.14
0.	000
1.	361
6. 768
DF = 2
0	10	0.000
1	10	-0.307
7	10	0.106
P-value	= 0.9345
Benchmark Dose Computation
Confidence level
Specified effect
Risk Type
BMD
BMDL
0. 95
0.1
Extra risk
804.532
255. 8
B-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
39
40
41
Absolute liver weight, male mice, NTP (1998)
Power Model with 0.95 Confidence Level
Power
BMDL
BMD
1500
2000
2500
0
500
1000
dose
Power Model.
BMDS MODEL RUN
The form of the response function is:
Y[dose] = control + slope * doseApower
Dependent variable = MEAN
Independent variable = mg_cum_hec
rho is set to 0
The power is not restricted
A constant variance model is fit
Total number of dose groups = 6
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 = 0.0125082
rho =	0 Specified
control =	1.613
slope = 0.0133873
power =	0.414725
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,
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
and do not appear in the correlation matrix )
alpha
control
slope
power
alpha
1
-1. 8e-009
-4.8e-00 9
6. 2e-009
control
-1.8e-00 9
1
-0. 82
0.79
slope
-4 . 8e-009
-0. 82
1
-1
power
5.2e-00 9
0.79
-1
1
Parameter Estimates
95.0% Wald Confidence Interval
Variable
alpha
control
slope
power
Estimate
0.0113839
1.62729
0.00362732
0.5708
Std. Err.
0.00213239
0.032356
0.00585053
0. 202297
Lower Conf. Limit
0.00720445
1.56387
-0.0078395
0.174305
Upper Conf. Limit
0.0155633
1.69071
0.0150941
0.967294
Table of Data and Estimated Values of Interest
Dose
N
Obs Mean
Est
Mean
Obs Std Dev
Est
Std Dev
Scaled Res
0
10
1. 61
1.
63
0.117
0
107
-0.424
35
10
1. 67
1.
65
0.06 96
0
107
0. 359
105
10
1. 7
1.
68
0.117
0
107
0.475
316
10
1.72
1.
72
0. 098
0
107
-0.0655
948
10
1.79
1.
81
0. Ill
0
107
-0.585
2634
7
1. 96
1.
95
0.159
0
107
0. 287
Model Descriptions for likelihoods calculated
Model A1:	Yij
V a r { e ( ij ) }
Model A2:	Yij
Var{ e ( ij ) }
Mu ( i) + e(ij;
Sigma^2
Mu(i) + e(ij;
Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi
Var{e(i)}
Mu + e(i)
Sigma^2
Likelihoods of Interest
Model
A1
A2
A3
fitted
R
Log(likelihood)
99.538919
102.357731
99.538919
99.053425
80.470340
Param's
7
12
7
4
2
AIC
-185.077839
-180.715462
-185.077839
-190.106851
-156.940680
Test 1:
Test 2
Test 3
Test 4
(Note:
Explanation of Tests
Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Are Variances Homogeneous? (A1 vs A2)
Are variances adeguately modeled? (A2 vs. A3)
Does the Model for the Mean Fit? (A3 vs. fitted)
When rho=0 the results of Test 3 and Test 2 will be the same.
B-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
39
40
41
42
43
44
45
46
47
48
49
Tests of Interest
Test -2*log(Likelihood Ratio) Test df
p-value
Test 1
Test 2
Test 3
Test 4
43.7748
5.63762
5.63762
0.970988
10
5
5
3
<.0001
0.3431
0.3431
0.8083
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
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 adeguately describe the data
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Relative risk
Confidence level =	0.95
BMD
783.381
BMDL = 2 4 6.114
Specified effect
1
Risk Type
Estimated standard deviations from the control mean
Confidence level
0. 95
BMD = 373.946
BMDL = 7 9.97 32
B-20
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-3. Summary of model selection and modeling results for best-
fitting multistage models for cancer effects resulting from chronic
inhalation exposure to THF



Goodness





Data Set
Degree

-





See Section
of

of-fit


BMD10
BMDL10
Model selection
5.3.1.,
Model
df
/j-valuc
LLb
x"
(mg/m3)
(mg/m3)
rationale"
Female mouse
1
1
0.95
-86.4707
NR
51.7
35.2
Most parsimonious fit
hepatocellular
2
2
0.75
-86.4345
0.07
61.4
35.4

tumors
3
2
0.76
-86.4118
0.05
61.2
35.5

Male rat kidney
1
1
0.50
-25.0786
NR
260
127
Most parsimonious fit
tumors
2
2
0.38
-25.0783
<0.1
268
127


3
2
0.43
-25.0775
<0.1
273
127

3 Adequate fit: goodness-of-fit p>0.05, scaled residuals <2.0, good fit near BMR, lack of extreme curvature not reflected in the observed
data.
b LL=Log-likelihood.
0 %2 = 2 x |(LLj - LLj)|, where i and j are consecutive numbers of stages. The test was evaluated for 1 degree of freedom (df). %2 for 1 df at
a = 0.05 is 3.84.
B-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
35
36
37
38
39
40
41
Hepatocellular adenomas or carcinomas, female mice (NTP, 1998)
Multistage Cancer Model with 0.95 Confidence Level
0.9
0.8
0.7
0.6
0.5
0.4
0.3
200	400	600	800
dose
Multistage Cancer Model
BMDS Model Run
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*dose^l)]
The parameter betas are restricted to be positive
Dependent variable = Effect
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-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background =	0.461466
Beta (1) = 0.00214268
Asymptotic Correlation Matrix of Parameter Estimates
t—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r
Multistage Cancer 	
Linear extrapolation		
iWIDL ^ [ BMP
B-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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
Background
Beta(1)
Background
1
-0. 56
Beta(1)
-0. 56
1
Variable
Background
Beta(1)
Parameter Estimates
Estimate
0.473223
0.00203941
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
Full model
Fitted model
Reduced model
Log(likelihood)	# Param'
-86.1605	4
-86.4707	2
-98.6187	1
Deviance Test d.f.
0.6204
24.9164
P-value
0.7333
C.0001
AIC:
176.941
Dose
Goodness of Fit
Est. Prob.
Expected
Observed
Scaled
Residual
0.0000
105.0000
316.0000
948.0000
0.4732
0.5748
0.7235
0.9238
17.509
22.416
27.492
40.647
17.000
24.000
26.000
41.000
37
39
38
44
-0.168
0. 513
-0.541
0. 201
Chi ^2
0. 62
d.f.
P-value
0.7319
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	51.6621
BMDL =	35.2535
BMDU =	84.437 6
Taken together, (35.2535, 84.4376) is a 90	% two-sided confidence
interval for the BMD
Multistage Cancer Slope Factor =	0.0028366
B-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
33
34
35
36
37
38
39
40
41
Renal adenomas or carcinomas, male rats (NTP, 1998)
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
BMDL
BMD
0
200
400
600
800
dose
Multistage Cancer Model
BMDS Model Run
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*dose^l)]
The parameter betas are restricted to be positive
Dependent variable = Effect
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-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.0998638
Beta (1) = 0.000398329
Asymptotic Correlation Matrix of Parameter Estimates
B-24
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
50
51
52
53
54
55
56
57
58
59
60
61
62
Background
Beta(1)
Background
1
-0.7
Beta(1)
-0.7
1
Variable
Background
Beta(1)
Parameter Estimates
Estimate
0.0903959
0.000405707
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
Full model
Fitted model
Reduced model
Log(likelihood)	# Param'
-25.0322	4
-25.0786	2
-26.8314	1
Deviance Test d.f.
0.0929231
3.59837
P-value
0.9546
0.3082
AIC:
54.1573
Dose
Goodness of Fit
Est. Prob.
Expected
Observed
Scaled
Residual
0.0000
105.0000
316.0000
948.0000
0.0904
0.1283
0.1998
0.3808
1. 085
0.770
4 .197
4 . 951
000
000
000
5. 000
12
21
13
-0.085
0. 281
-0.107
0. 028
Chi ^2
0.10
d.f.
P-value
0.9520
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	259.696
BMDL =	126.522
BMDU =	2285.4
Taken together, (126.522, 2285.4 ) is a 90	% two-sided confidence
interval for the BMD
Multistage Cancer Slope Factor = 0.000790379
B-25 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
APPENDIX C. SUPPLEMENTAL INFORMATION
C.l. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
C.l.l. Acute Toxicity Studies
Oral
Hofmann and Oettel (1954) examined the effects of THF following oral exposure. Cats
(13), rabbits (12), and rats (62, strain and sex not specified) received oral doses (route not
specified) ranging from a single administration of 3 cm3/kg (2,670 mg/kg) to 25 administrations
of 1 cm3/kg (890 mg/kg). The authors reported that no functional or histopathological damage to
the liver was observed. Also, no changes were observed in urine analysis, serum urea content, or
histopathology of the kidney.
Stasenkova and Kochetkova (1963) evaluated the acute toxicity of THF administered by
gavage. White rats (10/group, sex and strain not specified) received THF doses of 1, 1.5, 2, 3, 4,
or 5 g/kg by gavage as a solution in 2 mL of distilled water. The rats received a total of six
doses. The rats were observed for clinical signs and mortality. Necropsy and histopathology of
major organs was conducted in animals that died during the study exposure period. It does not
appear that histopathology was performed on the animals that survived exposure. No mortality
was observed at a dose of 2 g/kg. However, a dose of 3 g/kg resulted in 20% mortality, and
doses of 4-5 g/kg resulted in 90-100% mortality, respectively. Clinical signs of sedation,
including immobility, drowsiness, reduced response to external stimuli, and reduced respiratory
rate, were observed after 3-9 minutes of exposure. Mucous membranes appeared to have a
cyanotic discoloration. Histopathological lesions were observed in the stomach, brain, liver,
heart, spleen, and kidneys and included necrosis, edema, hemorrhage, and excess of blood or
fluid in the blood vessels or tissues.
Kimura et al. (1971) investigated the acute oral toxicity of THF in male Sprague Dawley
rats (6-12/group). The median lethal dose (LD50) values were estimated for four ages of rats:
newborns (24-48 hours old), 14 days old, young adult (80-160 g), and older adult (300-470 g).
Single doses of THF (doses unspecified) were administered by gavage; a microsyringe was used
for the newborn animals. The oral LD50 values for THF were estimated as 2.3 mL/kg for
14-day-old rats, 3.6 mL/kg for young adult rats, and 3.2 mL/kg for older adult rats. The LD50
values for the young animals were not statistically different from the values for the older adult
rats.
Inhalation
Stoughton and Robbins (1936) tested the effects of acute inhalation exposure to THF in
both mice and dogs. Mice (10/group, strain and sex not specified) were exposed to THF
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
35
36
concentrations of 0,0.5, 1.0, 1.5,2.2, or3.0mmol/L (0,36,050,72,100, 108,150, 158,620, or
216,300 mg/m3) for a single 2-hour exposure. The parameters evaluated included the time
required for onset of anesthesia and the time to respiratory failure or death. At the end of the
2-hour exposure, the animals still alive were observed until recovery or death. THF
concentrations of 2.2 mmol/L were 100% fatal; at these concentrations, time to onset of
anesthesia was 5-8 minutes and time to death was 30-51 minutes. The 1.0 mmol/L dose of THF
resulted in 50% mortality, with time to anesthesia of 50 minutes and time to death of
109 minutes. No mortality was observed at a THF concentration of 0.5 mmol/L. Animals
surviving at the end of the exposure period regained the ability to walk in 6-8 hours following
exposure to THF. One dog (strain and sex not specified) was anesthetized with THF and
maintained for 1.5 hours at a THF atmospheric concentration of 5-6%. During this exposure,
electroencephalogram (EEG), respiration, and blood pressure were measured. Two days
following exposure, the dog was sacrificed and autopsied. Symptoms observed in the dog
included increased saliva and mucus flow, decrease in blood pressure, stimulation of respiration,
and prolonged sleep up to 6-8 hours after exposure stopped. No gross abnormalities were
observed on autopsy.
Henderson and Smith (1936) exposed six rats (strain and sex not specified) to increasing
concentrations of THF vapor for 1 hour. The exact concentrations of THF vapor used were not
reported, but the authors noted that anesthesia occurred at 6.47% THF. Two animals exposed to
just the anesthetic concentration for 30 minutes recovered within 2 minutes after exposure. Two
rats that died within 24 hours of exposure had congested, mottled lungs. One rat that initially
recovered but appeared ill 4 days later showed fatty changes in the liver.
Hofmann and Oettel (1954) examined the effects of acute inhalation exposure to THF in
18 cats, 20 rabbits, 52 rats, and 150 mice. The sex and strain of the animals were not specified.
Animals were exposed to THF vapors at concentrations ranging from 3,400-60,000 cm3/m3
(equivalent concentrations reported by the authors were 10,000-193,000 mg/m3). Exposure
regimens ranged from one 2-hour exposure to 30 6-hour exposures. No additional information
was provided on exposure durations and concentrations. Therefore, it is not possible to estimate
adjusted exposure concentrations. Liver function was assessed by using a bromosulfalein test
(decreased clearance of bromosulfalein from the blood is indicative of liver dysfunction).
Kidney function was also assessed by urinalysis and serum urea content. Blood cell count was
evaluated. Both the liver and kidney were evaluated histopathologically. The authors reported a
slight, transient retention of bromosulfalein immediately following exposure to narcotic
concentrations of THF.
LaBelle and Brieger (1955) evaluated the effects of acute THF inhalation exposure in rats
and mice. Groups of eight male albino rats were exposed to a fixed concentration of THF for a
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
single 4-hour exposure period. Those animals surviving were observed for 14 days. The range
of concentrations tested was not specified. This procedure was repeated until the median lethal
concentration (LC50) could be determined. In addition, groups of white mice (6/group, sex not
specified) were exposed continuously to saturated THF vapor (approximately 47,000 ppm or
138,650 mg/m3), and survival time was recorded. For mice, the mean survival time following
exposure to saturated vapor was 41 minutes. In rats, the LC50 reported by the authors was
18,000 ppm (53,100 mg/m3). Narcosis was reported in rats prior to death.
Stasenkova and Kochetkova (1963) evaluated the effects of a single 2-hour inhalation
exposure to THF in white mice and rats (10/group, sex and strain not specified). THF vapor was
generated by allowing it to evaporate from a filter paper, so constant air concentrations were not
maintained for the duration of the exposure period. For example, at the highest target
concentration of 180 mg/L, air concentrations in the test chamber were reported as 140 mg/L
after 15 minutes and 65 mg/L after 2 hours. Based on the average of the measurements at
15 minutes and 2 hours, actual mean exposure concentrations were 0, 7, 13, 19, 27, 42, 73, 80,
and 103 mg/L (0, 7,000, 13,000, 19,000, 27,000, 42,000, 73,000, 80,000, and 103,000 mg/m3).
Animals were evaluated for clinical signs and mortality. Histopathological examination was
conducted on animals that died. The authors did not indicate whether histopathological
examinations were conducted on the animals that survived exposure. In mice, the average
concentration of 19 mg/L resulted in 80% mortality, and 27 mg/L resulted in 100% mortality.
Rats were less sensitive to THF. The average concentration of 42 mg/L resulted in 20%
mortality, and 80 mg/L resulted in 100% mortality. The animals displayed symptoms of sedation
and narcosis, including depressed activity, interrupted breathing, and reduced coordination of
movement. In addition, mucus membranes were pale and bluish in color. Lesions observed in
lungs and bronchi included excess blood or fluid, edema, perivascular hemorrhage, and catarrhal
condition of the mucus membrane. Histopathological lesions were also observed in brain, liver,
kidney, and spleen, including excess blood or fluid, edema, and dystrophic changes.
DuPont Haskell Laboratory (1979) conducted an acute inhalation study of THF in order
to determine the highest concentration of THF that would not produce narcosis in rats. ChR-CD
rats (6/sex/group) were exposed to THF concentrations, ranging from 3,010-20,500 ppm (8,880-
60,475 mg/m3) for a single 6-hour exposure period. Following exposure, all rats were weighed
daily and clinical signs were observed for 14 days. The authors determined that the nonnarcotic
concentration in male rats was 5,380 ppm (15,871 mg/m3) and in female rats was 5,700 ppm
(16,815 mg/m3). During the exposure period, both male and female rats demonstrated clinical
signs of pawing and scratching and decreased or no response to sound at all concentrations.
Male rats also exhibited signs of rapid respiration, and females showed signs of paralysis. Based
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
on clinical signs of CNS toxicity, the lowest exposure concentration of 8,880 mg/m3 is the study
LOAEL.
Ohashi et al. (1983) evaluated the effects of acute inhalation exposure to THF on the
upper respiratory tract (nasal mucosa) of rabbits. Adult rabbits (sex and number not specified)
were exposed to THF concentrations of 100, 250, 1,000, 2,000, 6,000, or 12,000 ppm (295, 738,
2,950, 5,900, 17,770, or 35,400 mg/m3) for a single 4-hour exposure period. The rabbits were
sacrificed by air embolization, and their nasal mucus membranes were obtained at 0, 20, 40, 60,
120, or 180 minutes following exposure. The membranes were evaluated for ciliary beating
frequency and examined by scanning electron microscopy. No other organs or systems were
evaluated. THF caused a dose-related decrease in ciliary beating frequency. Concentrations of
250 ppm caused about a 50% decrease in beat frequency that returned to normal within 3 hours
following exposure. Concentrations of 1,000 ppm almost completely eliminated ciliary beating,
and at these concentrations beat activity did not return to normal. THF concentrations of
250 ppm resulted in the appearance of sporadic compound cilia, but no other morphological
changes. Concentrations of 1,000, 2,000, and 6,000 ppm resulted in the increased incidence of
compound cilia and the vacuolation of epithelial cells, indicating moderate degeneration. At
12,000 ppm THF, observations included many large compound cilia, vacuolation, cytoplasmic
protuberances, and sloughing of the epithelial cells, indicating severe degeneration. Based on
significant morphological changes to nasal epithelial cells, 1,000 ppm (2,950 mg/m3) is the
LOAEL and 250 ppm (738 mg/m3) is the NOAEL.
Horiguchi et al. (1984) evaluated the acute toxicity of THF following inhalation exposure
in rats. Sprague-Dawley rats (6 males/group) received a single 3-hour exposure to THF at
concentrations of 200, 1,000, 5,000, 10,000, 15,000, 25,000, or 30,000 ppm (590, 2,950, 14,750,
29,500, 44,250, 73,750, or 88,500 mg/m3). The animals were observed for clinical signs of
toxicity, abnormal behavior, and mortality for 72 hours following exposure. The LC50 value was
estimated to be 21,000 ppm (61,950 mg/m3) by using a probit method. Animals in the 200 ppm
group displayed signs of head shaking and face washing, as well as patches of mild irritation on
nose, ears, and eyelids, and sleep. Symptoms of irritation increased with the exposure
concentration. At 5,000 ppm, animals displayed intense salivation, tearing, and bleeding from
the nose. In addition, animals developed clonic muscle spasms, had altered respiratory patterns,
and became comatose about 1 hour following the start of exposure. All animals in the
25,000 ppm group died within 72 hours following exposure. No information was provided
regarding the observations in other dose groups. Based on clinical signs of irritation and
neurotoxicity, the concentration of 5,000 ppm (14,750 mg/m3) is the LOAEL in this study.
Ikeoka et al. (1988) investigated the effects of acute inhalation exposure to THF on the
lower respiratory tract (tracheal mucosa) of rabbits as a follow-up to the earlier study by Ohashi
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
et al. (1983). Adult rabbits (sex and number not specified) were exposed to THF at
concentrations of 100, 250, 1,000, 2,000, 6,000, or 12,000 ppm (295, 738, 2,950, 5,900, 17,770,
or 35,400 mg/m3) for a single 4-hour exposure period. The authors did not state if a control
group was also included. The rabbits were sacrificed by air embolization, and their tracheal
mucosa membranes were obtained at 0, 20, 40, 60, 120, or 180 minutes following exposure. The
membranes were evaluated for ciliary beating frequency and examined by scanning electron
microscopy. No other organs or systems were evaluated. THF caused a dose-related decrease in
ciliary beating frequency. Concentrations of 250 ppm caused about a 50% decrease in beat
frequency that returned to normal within 3 hours following exposure. Concentrations of
1,000 ppm almost completely eliminated ciliary beating, and at these concentrations beat activity
did not return to normal within 3 hours. Compound cilia, ballooning, and vacuolation of tracheal
epithelial cells were observed in the high-concentration group. However, the areas of severe
degeneration observed in the nasal epithelium following the same exposure protocol were not
observed in the trachea in the current study. The effects on the tracheal morphology were mild
compared with those observed in nasal epithelium by Ohashi et al. (1983). Based on tracheal
histopathology, 12,000 ppm (35,400 mg/m3) is the LOAEL and 6,000 ppm (17,770 mg/m3) is the
NOAEL.
Dermal
Stasenkova and Kochetkova (1963) evaluated the effects of THF application to the skin
of white mice (20, strain and sex not specified) and rabbits (number, sex, and strain not
specified). Pure THF (1 mL) was applied to the skin of rabbits. THF caused reddening of the
skin, which subsequently thickened and sloughed off. Pure THF applied to the eyes of rabbits
caused edema of the eyelid, vasodilation, and corneal opacity. The tails of mice were immersed
in pure THF for 2 hours. This treatment resulted in mortality, symptoms typical of THF
poisoning, as well as excess blood or fluid and hemorrhage of internal organs.
C.1.2. Short-term Studies
Oral
Komsta et al. (1988) reported the results of a short-term oral toxicity study of THF in
rats. Sprague-Dawley rats (10/sex/group) were administered THF in drinking water at
concentrations of 0, 1, 10, 100, or 1,000 mg/L for 4 weeks. The equivalent doses estimated by
the study authors based on measured water consumption and body weights were 0, 0.1, 0.8, 10.2,
and 95.5 mg/kg-day. Clinical signs, body weight gain, and food and water consumption were
evaluated weekly. Following the exposure period, the animals were sacrificed and examined at
gross necropsy. Organ weights were obtained for brain, heart, liver, spleen, and kidney. Blood
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
28
29
30
31
32
33
34
was collected for hematology and serum chemistry evaluation. A selection of tissues from the
control and high-dose group was evaluated histopathologically.
There was no increase in mortality in any of the dose groups, and no clinical signs were
observed in any of the treated animals. In addition, body weight gain and food and water
consumption were not significantly different between treated and control animals. No changes in
hematology or serum chemistry were observed in treated animals. Some sporadic observations
of histopathological changes were observed in the thyroid, liver, and kidney; however, the
incidence for these findings was comparable in treated and control animals. Male rats in the
high-dose group demonstrated a higher incidence of increased cytoplasmic homogeneity in liver
compared with controls (3/10 and 7/10 for control and high-dose animals, respectively). No
changes in any of the biochemical parameters evaluated were observed. Female rats showed an
increased incidence of anisokaryosis (unequal size of cell nuclei) in the liver (0/10 and 7/10 for
control and high-dose animals, respectively) and tubular cytoplasmic inclusions in the kidney
(0/10 and 3/10 for control and high-dose animals, respectively). The authors did not conduct a
statistical analysis of the incidence data. In addition, histopathology was not performed on the
lower dose groups, so it is not possible to evaluate the dose-response relationship for these
endpoints. The study authors concluded that THF in drinking water at doses up to 1,000 mg/L
did not produce overt toxicity. Komsta et al. (1988) also indicated that the effects observed at
the high dose of THF were considered mild and adaptive and could not be related to any
functional changes (i.e., altered biochemical parameters).
Pozdnyakova (1965) evaluated the effects of short-term exposure to THF in drinking
water. White mice (number, sex, and strain not specified) received THF in the drinking water at
concentrations of 40 and 100 mg/L for 45 days. Mice in the high-dose group exhibited
decreased body weight, paralysis of hind legs, leukocytosis, and decreased hemoglobin. No
significant changes were observed in the low-dose group. No additional information was
provided about the study.
In the same study report, Pozdnyakova (1965) exposed 20 rabbits (sex and strain not
specified) and 50 white rats (sex and strain not specified) to THF in drinking water at doses of
10 and 20 mg/kg. The study was classified as being chronic in duration by the study authors;
however, the actual duration of exposure was not specified. Rabbits in the high-dose group
exhibited a change in cholinesterase activity, an increase in prothrombin time, and a low serum
antibody titer compared with controls. Rats in the high-dose group showed a reduction in BW
and a change in serum albumin content. No additional information was provided in the study.
C-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
Inhalation
Horiguchi et al. (1984) evaluated the ability of THF to irritate the respiratory tract
following short-term inhalation exposure to THF. Male Sprague-Dawley rats (3-6/group) were
exposed to 0, 100, or 5,000 ppm (0, 295, or 14,750 mg/m3) THF vapor for up to 3 weeks. No
information was provided on the duration of each exposure period or the number of days/week
the animals were exposed, and therefore duration-adjusted exposure concentrations could not be
calculated. A single animal was randomly selected from each exposure group 1 day, 1 week, and
3 weeks following the start of exposure. The animals were sacrificed the next day (24 hours
later) and the respiratory tract mucous membrane was extracted and prepared for histological
examination. No differences were observed between the tracheal mucosa of the treated groups
and the controls following 1 day or 1 week of exposure. By 3 weeks of exposure, the tracheal
mucosa of animals in the high-concentration group exhibited disordered cilia and epithelial cells
and darkening of cell bodies compared with control animals. Also, by 3 weeks of exposure, the
nasal mucosa of animals in the low-concentration group (100 ppm) exhibited the same type of
changes described above for the tracheal mucosa (e.g., disordered cilia and epithelial cells and
darkening of cell bodies) without significant histopathological effects. The nasal mucosa of
animals exposed to 5,000 ppm for either 1 week or 3 weeks, however, demonstrated disruption
of the epithelial architecture, congestion, and sloughing of ciliary and goblet cells, in addition to
vacuolation and darkening of cell bodies. Based on these effects at the nasal mucosa, the
LOAEL is determined to be 5,000 ppm and the NOAEL is 100 ppm.
Stasenkova and Kochetkova (1963) evaluated the short-term effects of THF inhalation in
male rats and mice (20/group; strain not specified). The animals were exposed for 2-hour
periods, twice a day, every day for 2 months to air concentrations of THF ranging from 6 to
8 mg/L (6,000-8,000 mg/m3). However, THF vapor was generated by allowing it to evaporate
from a filter paper, so constant air concentrations were not maintained for the duration of each
exposure period. Animals were evaluated for clinical signs, mortality, and body weight.
Endpoints evaluated included the threshold of neuromuscular irritability (method of
measurement was not specified), arterial blood pressure, blood cell counts, liver function
(measured by synthetic capacity), and kidney function (measured by albumin in urine). After
2 months, the animals were sacrificed and histopathological examination of major organs was
conducted.
All animals developed symptoms of narcosis during the exposure; however, this effect
was not observed during the periods between exposures. By day 40 of exposure, treated rats had
reduced body weight compared with controls. At the end of the 2-month study period, mean
body weights of treated rats was 30% less than controls. In addition, treated rats had a lower
threshold of neuromuscular irritability than controls. No effects in rats were observed on blood
C-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
pressure, blood cell count, or liver or kidney function. Histopathological lesions in the
respiratory tract included catarrhal rhinitis, bronchitis, proliferative reaction in lungs,
emphysema, and hypertrophy of muscle fibers in the walls of the bronchi. Histopathological
lesions, including hypertrophy of muscle fibers and perivascular sclerosis, were observed in the
heart, liver, and kidneys. Incidence data were not provided for any of these histopathological
findings.
Treated mice initially developed symptoms of eye and respiratory tract irritation and had
an increase in the threshold of neuromuscular irritability compared with controls. After 1 month
of treatment, mortality in mice increased. The authors indicated that mice died of bronchial
pneumonia. It was not clear if mortality in controls was increased and if the bronchial
pneumonia was a cause of THF treatment or a bacterial infection in the mice. The mice still
living at the end of the 2-month treatment period had a 15-20% decrease in body weight
compared with controls. No information was provided on the results of other endpoints
evaluated in mice. Because of poor reporting of this study, no NOAEL-LOAEL can be
determined.
C.1.3. Neurotoxicity Studies
DuPont Haskell Laboratory (1996a), published in the peer-reviewed literature as Malley
et al. (2001) investigated the neurotoxicity of acute inhalation exposure to THF in rats. Crl:CD
BR rats (12/sex/group) were exposed to THF vapor at concentrations of 0, 500, 2,500, or
5,000 ppm (0, 1,475, 7,375, or 14,750 mg/m3) for a single 6-hour exposure (designated as test
day 1). The animals were then observed for 2 weeks following exposure. Clinical signs, body
weight, and food consumption were evaluated weekly. The response to an alerting stimulus was
determined as a group for each exposure concentration, prior to the start of exposure and
approximately 2 and 4 hours after initiation of exposure. All rats were evaluated for
neurobehavioral effects. Motor activity assessments and functional observational battery (FOB)
assessments were conducted before exposure and on test days 2, 8, and 15. For the motor
activity assessments, animals were individually tested in an automated activity monitor that
measured both duration of continuous movements and number of movements. The FOB
assessment consisted of a series of quantified behavioral evaluations conducted in a sequence
that proceeded from the least interactive to the most interactive. During the FOB assessment,
each rat was evaluated in three environments: inside the home cage, on removal from the home
cage while being handled, and in a standard open field arena.
Exposure to 2,500 ppm THF appeared to have an effect on response to alerting stimulus
in rats. Six of 24 rats in the 2,500 ppm group had a diminished response after 2 hours of
exposure, and all 24 rats in this group had diminished response after 4 hours of exposure. Half
C-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
the rats in the 5,000 ppm group had diminished response after 2 hours of exposure, and all of the
rats had either no response or diminished response to stimulus after 4 hours of exposure. Other
signs of sedation in the high concentration group included a significant increase in the incidence
of lethargy and abnormal gait in both male and female rats at 5,000 ppm. Male rats in the
5,000 ppm group had significantly decreased body weight gain and food consumption in the
interval between test day 1 and 2, although these values were comparable to controls for the
remainder of the observation period. Several parameters in the FOB were affected in the 5,000
ppm groups immediately following the exposure period only, including the righting reflex in
males and females, palpebral closure in females, and ease of handling in females. The effects on
FOB parameters were not observed during test days 2, 8, or 15, suggesting that the sedative
effects of THF were short-lived. The LOAEL for this study is 2,500 ppm (7,375 mg/m3), based
on observations of sedative effects, and the NOAEL for this study is 500 ppm (1,475 mg/m3).
DuPont Haskell Laboratory (1996b; Malley et al., 2001) investigated neurotoxicity
following subchronic inhalation exposure to THF in rats. Crl:CD BR rats (12-18/sex/group)
were exposed to THF vapor at concentrations of 0, 500, 1,500, or 3,000 ppm (0, 1,475, 4,425, or
8,850 mg/m3) 6 hours/day, 5 days/week over a 13- to 14-week exposure period. Clinical signs,
body weight, and food consumption were evaluated weekly. Prior to the start of exposure and
approximately 2, 4, and 6 hours after initiation of exposure, the response to an alerting stimulus
was determined for the rats as a group for each exposure concentration. All rats were evaluated
for neurobehavioral effects. Motor activity assessments and FOB assessments were conducted
before the first exposure and at 4, 8, and 13 weeks. For the motor activity assessments, animals
were individually tested in an automated activity monitor that measured both duration of
continuous movements and number of movements. The FOB assessment consisted of a series of
quantified behavioral evaluations conducted in a sequence that proceeded from the least
interactive to the most interactive. During the FOB assessment, each rat was evaluated in three
environments: inside the home cage, after removal from the home cage while being handled,
and in a standard open field arena. Rats (6/sex/group) were sacrificed after 13 weeks of
exposure, and tissue from the nervous system and muscle was assessed histopathologically.
The only effects observed in this study appeared to be related to the acute sedative effects
of THF characterized by the study authors as acute behavioral sedation, which dissipates rapidly
upon termination of exposure (Malley et al., 2001). A diminished response to alerting stimulus
during exposure was observed in male and female rats in the 1,500 and 3,000 ppm exposure
groups. In the 3,000 ppm group, diminished response was observed consistently, beginning on
the second day of exposure. In the 1,500 ppm group, diminished response was observed
sporadically from days 16 to 49 of exposure and observed consistently on the remaining
exposure days. Diminished response was observed sporadically from days 16 to 49 of exposure
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
and observed consistently on the remaining exposure days. Compound-related clinical signs,
including stained nose and stained/wet perineum, were also observed in male and female rats in
the 1,500 and 3,000 ppm groups. These signs were not observed on Mondays prior to the start of
exposure for the week or on the days of the motor activity and FOB assessment. Therefore,
these signs were considered to be transient. No effects were observed on body weight, body
weight gain, food consumption, motor activity, any of the parameters in the FOB, or
neuropathology in either male or female rats at any concentration. Based on clinical signs of
sedation during exposure to THF, 1,500 ppm (4,425 mg/m3) is the study LOAEL, and the
NOAEL for this study is 500 ppm (1,475 mg/m3). However, the authors suggested that these
effects were transient.
Marcus et al. (1976) evaluated the neuropharmacological effects of THF administered by
i.p. injection. Male Sprague-Dawley rats (number/group not specified) were implanted with
electrodes to facilitate continuous EEG recordings. THF was administered by i.p. injection at
doses of 15 and 21 mmol/kg (1,156 and 1,619 mg/kg). After a 2-minute latency period,
21 mmol/kg THF induced high amplitude slow wave activity in the EEG, which lasted
2 minutes. The EEG pattern then changed to spiking and electrical silence, which lasted for
20 minutes. The altered EEG pattern was accompanied by loss of the righting reflex. A dose of
15 mmol/kg induced a desynchronization of the EEG activity without loss of the righting reflex.
In an in vitro study, THF caused a decrease in adenosine triphosphatase (ATPase) activity
and membrane fluidity in a dose-dependent manner in an assay using rat brain synaptosomes
(Edelfors and Ravn-Jonsen, 1992).
C.2. METABOLITE AND MECHANISTIC DATA AND OTHER STUDIES
C.2.1. Metabolite Studies
The nervous system is one of the primary targets of THF toxicity. As discussed under
Metabolism (Section 3.3), the effects of THF on the nervous system may be due to its
metabolites, GBL and GHB. Major study findings of these compounds are briefly summarized
(Table C-l) to facilitate an evaluation of THF toxicity data, but a more detailed review is
available (NSF, 2003).
C-10
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-l. Comparison of target organ toxicity for THF and its metabolites
Target organ
THF
GBL
GHB
CNS
No effect in rat drinking water study
at 882 mg/kg-day. Narcosis observed
in inhalation studies at estimated
systemic doses of 2,260 mg/kg-day in
mice3 and 5,822 mg/kg-day in ratsb.
Lethargy in rat and mice
subchronic gavage at 225 mg/kg-
day (NTP, 1992); EEG changes
beginning at 50 mg/kg i.p. in
young rats in mode of action
studies (Takizawa et al., 2003)
Dizziness in human
clinical studies at
12.5 mg/kg LOAEL
(Ferrara et al., 1999)
Liver
No effect in rat drinking water study
at 788 mg/kg-day. Increased absolute
and relative liver weight in mice in the
inhalation study at estimated systemic
dose of 753 mg/kg-day.
No effect in subchronic gavage
study at 900 mg/kg-day (rats) and
1,050 mg/kg-day (mice) (NTP,
1992)
No data
Kidney
Increased kidney weight in rat
drinking water study at 714 mg/kg-
day.
No effect in subchronic gavage
study at 900 mg/kg-day (rats) and
1,050 mg/kg-day (mice) (NTP,
1992)
No data
Thymus
No oral data. Decreased thymus
weight at 753 mg/kg-day and thymus
atrophy at 2,260 mg/kg-day in mouse
inhalation study.
Thymus depletion at 262 mg/kg-
day in mouse 2-year gavage
study0 (NTP, 1992)
No data
BW
Minimally decreased body weight in
rat drinking water study at 714 mg/kg-
day.
Decreased body weight in rat 2-
year gavage study at 450 mg/kg-
day and in mice at 262 mg/kg-
day (NTP, 1992)
No data
Development
Decreased pup body weight gain,
delayed eye opening, and increased
incidence of sloped incisors at 782
mg/kg-day in rat drinking water study.
Fetal weight, skeletal alterations in rat
inhalation studies.
No effects in rat gavage at
500 mg/kg-day (Kronevi et al.,
1988)
No data
Reproductive
No effect in rat drinking water study
on reproductive function or testes
weight at 788 mg/kg-day.
Decreased testes weight in rat
gavage study at LOAEL of
667 mg/kg-day (Debeljuk et al.,
1983)
No data
aFor this cursory analysis, estimated systemic doses were calculated from the inhalation studies assuming 100%
absorption and EPA default parameter values for mice as follows: LOEL exposure concentration (mg/m3) x
default EPA ventilation rate (0.063 m3/day) x study exposure duration (6 hours/24 hours)/default EPA BW
(0.037 kg) = mg/kg-day.
bFor this cursory analysis, estimated systemic doses were calculated from the inhalation studies assuming 100%
absorption and EPA default parameter values for rats as follows: LOEL exposure concentration (mg/m3) x default
EPA ventilation rate (0.36 m3/day) x study exposure duration (6 hours/24 hours)/default EPA BW (0.38 kg) =
mg/kg-day.
°No effects on thymus weight were observed in the 13-week study (NTP, 1992). Thymus histopathology in the
chronic study (NTP, 1992) was attributed by the authors to injuries secondary to fighting.
1
2	There is no specific organ toxicity information following repeated human exposure to
3	GBL; however, chronic use of GBL as a drug of abuse can lead to neurotoxicity, including
4	addiction, anxiety, depression, insomnia, and tremor (Herold and Sneed, 2002). The systemic
C-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
toxicity of GBL has been investigated in a full 2-year bioassay in rats and mice that employed
gavage dosing (NTP, 1992). The most sensitive effect observed in these studies was clinical
signs of CNS toxicity (lethargy) with a NOAEL of 112 mg/kg-day in rats. The only other
treatment-related effect observed in rats and mice was for decreased body weight. NTP (1992)
also reported a statistically significantly increased incidence of thymic depletion and epithelial
hyperplasia of the thymus in the mid- and high-dose male mice (0/42, 5/39, and 6/38 and 0/42,
4/39, and 4/38, respectively). The study authors concluded that the observed dose-related
increase in these nonneoplastic lesions was related to fighting in the male mice. Specifically, the
depletion of lymphocytes in the thymus (often seen with debilitation and stress in animals) was
most often observed in mice dying early as a result of wounds received from fighting. The
relevance of the observed effects on the thymus remains uncertain.
In other studies on GBL, no prenatal developmental effects were observed in rats at doses
up to 500 mg/kg-day (Kronevi et al., 1988), while decreased testicular weight was reported in a
short-term reproductive study (Debeljuk et al., 1983) with a LOAEL of 667 mg/kg-day.
The oral toxicity data for GHB are primarily from clinical studies in human subjects or
from case reports of oral poisonings. Transient dizziness and a sense of dullness in 50% of
human subjects following a single oral dose of 12.5 mg/kg were observed by Ferrara et al.
(1999). Standardized measure of psychomotor performance was not affected at this dose
(Ferrara et al., 1999). Metcalf et al. (1966) reported the effects of single oral doses of 35-
63 mg/kg GHB in volunteers. All participants reported drowsiness during the experiment and
some participants receiving doses >50 mg/kg were rendered unconscious. Medical anesthetic
doses of GHB are typically in the range of 60 mg/kg (Miotto et al., 2001; Vickers, 1969; Root,
1965).
In the case of GHB, the dosing regimen seems to play an important role on the induction
of CNS effects. The human clinical studies make it clear that for the CNS effects of GHB, bolus
dosing regimens have an important effect. For example, as shown in Table C-2, large
differences in total daily dose did not show a significant change in overall response rate and
severity when the individual doses were similar (Gallimberti et al., 1993, 1992). Furthermore,
the incidence of effects and their severity generally corresponds to the individual doses rather
than the total daily dose (Nimmerrichter et al., 2002; Gallimberti et al., 1993).
C-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
Table C-2. Comparative effects of single and multiple daily dosing of GHB
Reference
Single dose
(mg/kg)
Maximum total daily
dose (mg/kg-day)
Effect
Gallimberti et al. (1993)
25
300
Dizziness (5/41)
Gallimberti et al. (1992)
17
50
Dizziness (3/41)
Addolorato et al. (1998)
50
150
Vertigo and lethargy (30% of 109 patients)
Nimmerrichter et al. (2002)
10-20
50
Vertigo (9/31); majority after the double dose

20-40
100
Vertigo (17/33); seizure (1/33); disorientation
(1/33)—majority after the double dose
Scharf et al. (1998)
30
60
Altered brain wave measurements during sleep
Peak doses rather than cumulative doses appear to drive the CNS response to
administration of GHB. The absence of observed CNS effects in the two-generation THF
drinking water study in rats (Hellwig et al., 2002; BASF, 1996) at higher daily doses than the
daily gavage doses for GBL, which also caused CNS effects (NTP, 1992), may reasonably be
explained by differences in exposure patterns. Continuous drinking water exposures might not
result in sufficient peak levels of exposure to induce the CNS effect. Other explanations may
exist for the absence of reported CNS effects in the two-generation study including, for instance,
lack of a more detailed neurobehavioral evaluation and other limitations in study design
including lack of sensitivity or suitability for analyzing neurotoxicity potential.
C.2.2. Mechanistic Studies
C.2.2.1. Cytotoxicity
THF was evaluated in a series of short-term in vitro assays to assess its potential for
cytotoxicity (Curvall et al., 1984): inhibition of cell growth of ascites sarcoma BP 8 cells grown
as stationary suspension cultures, inhibition of oxidative metabolism in isolated brown fat cells,
plasma membrane damage (leakage of a cytoplasmic nucleotide marker from prelabeled cells),
and ciliotoxicity as measured by time to ciliostasis in cultures of trachea from unborn chickens.
To facilitate comparison of multiple chemicals, the results from each individual assay were
expressed as a percentage of control responses and then these percentages were converted to a
10-point scale where 0 corresponded to 0-9%. The response observed in each of the individual
assays of THF was <10%. THF was scored 0 for each of the individual assays and for its mean
cytotoxicity activity. In contrast, several chemicals, mostly alkylphenols, were highly active in
the test systems, having activity of 7 in each of the test systems. In a related study, a 5 mM
concentration of THF took >60 minutes to cause ciliostasis in an in vitro assay in embryonic
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
24
25
26
27
28
29
30
31
32
33
34
35
36
chicken trachea, whereas highly cytotoxic compounds caused ciliostasis in <5 minutes
(Pettersson et al., 1982). Therefore, the results of these studies suggest that THF is not cytotoxic.
The cytotoxicity of THF was evaluated in an in vitro assay of protein content in cell
cultures (Dierickx, 1989). Human hepatoma, HepG2 cells were maintained in culture in 24 well
tissue culture test plates. THF and other test compounds were dissolved directly in culture
medium at five different concentrations (not specified) and incubated with test cells for 24 hours.
The cells were lysed and protein content measured. The relative toxicity of THF and the other
test compounds was determined by estimating the concentration in mM required to induce a 50%
reduction of cell protein content (PIso)- Very toxic compounds, such as heavy metals and
surfactants, consistently had PI50 values of less than 1 mM. In contrast, the PI50 for THF was
372. The results of this study suggest that THF is not cytotoxic.
The cytotoxicity of 168 chemicals, including THF, was characterized as part of a
validation of the BALB/c-3T3 cell transformation assay (Matthews et al., 1993). The LC50 for
THF was 90.3 mM. The authors noted that in the analysis of the entire data set of 168 chemicals,
in vitro cytotoxicity did not correlate to in vivo carcinogenic activity. THF was considered by
the authors as noncytotoxic (defined as having an LC50 ranging from 5 to 100 mM).
C.2.2.2. CYP450 Activity, Cell Proliferation, andApoptosis
BASF (1998) reevaluated kidney tissues from male rats and liver tissues from female
mice from the NTP (1998) study to examine the relationship between cell proliferation responses
and increase in tumors observed in these tissues following THF administration.
Histopathological examination and evaluation of cell proliferation as measured by proliferating
cell nuclear antigen (PCNA) staining were conducted using tissue samples from the 0, 200, 600,
and 1,800 ppm (0, 590, 1,770, and 5,310 mg/m3) exposure groups (10/group) from the NTP
(1998) subchronic (13 weeks) study. For the male rat kidneys, tissues from the cortex, outer
stripe of the outer medulla, inner stripe of the outer medulla, and inner medulla were evaluated
separately. For the female mouse liver, no zonal subdivision was made.
The histopathology examination revealed increased incidence of moderate grade hyaline
droplet accumulation in the male rat kidney tissues of the high-concentration group as compared
to controls, but these changes were not accompanied by evidence of cell degeneration. No other
differences between controls and exposure groups were noted. No increase in cell proliferation
was found in any of the kidney compartments or in evaluation of all compartments combined.
Cell proliferation index was statistically significantly decreased in individual kidney
compartments, although these changes did not show a concentration-dependent pattern. For the
female mouse liver tissues, no treatment-related histopathology was observed. The cell
proliferation index was increased by approximately 39% in tissues from the high-concentration
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
mice compared with controls. However, this result was not statistically significant and was
noted as being predominantly based on the results from 2/10 animals. Furthermore, a significant
decrease in proliferation index was observed in the mid-concentration group, but no clear
concentration-response pattern was observed. Based on these results, the study authors
concluded that the examination of the tissues from the 13-week NTP (1998) study revealed no
clear increase in cell proliferation that can be correlated to a tumorigenic mechanism.
BASF (Gamer et al., 2002; BASF, 2001a) evaluated a series of endpoints in male F344
rats (6/group plus 5/group at the control and high concentrations for enzyme assays) and female
B6C3Fi mice (10/group plus 5 in the control and high concentrations for enzyme assays) in
tissues for which THF-treated animals developed tumors. Animals were placed in one of three
groups that were exposed 6 hours/day for either 5 consecutive days, 5 consecutive days followed
by a 21-day observation period, or 20 consecutive days over a period of approximately 28 days.
Test animals were exposed nose only to 0, 199, 604, or 1,794 ppm THF (average THF
concentrations of 0, 598, 1,811, or 5,382 mg/m3), corresponding to the concentrations used in the
NTP (1998) cancer bioassay. Concentrations adjusted for continuous exposure were 0, 107, 323,
or 961 mg/m3. For the animals in each of the four concentration groups, a full necropsy was
done, including histopathological evaluation of the kidney (rat) and liver (mouse). Additional
evaluations in these same organs included measurements of cell proliferation (S-phase response
by 5-bromo-2-deoxyuridine [BrdU] staining) and terminal deoxynucleotidyl transferase
deoxyuridine triphosphate (dUTP) nick-end-labeling staining (TUNEL) apoptosis assay. For the
male rat kidneys, immunohistochemical detection of a2u-globulin was also performed. Five
animals from the control and high-concentration groups that were exposed for 5 consecutive
days were also harvested for measurement of CYP450 content and for CYP450 isozyme activity
as measured by ethoxyresorufin-O-deethylase (EROD) and pentoxyresorufin-O-depentylase
(PROD) activity.
The results of the BASF (2001a), evaluating cell proliferation, apoptosis, and
a,2U-globulin accumulation in the kidneys of male F344 rats, are shown in Table C-3. Although
no significant increase in labeling index in the renal cortex was determined by standard
assessment methods, focal areas of increased BrdU labeling were noted. Quantitation of these
areas revealed increased cell proliferation in subcapsular proximal tubules (cortex 1) in animals
exposed to THF at the mid and high concentration for 20 days and at the high concentration for 5
consecutive days. No increase in labeling was observed in the groups given a 21-day recovery
period. An increase in cell proliferation was also noted in the proximal tubules between the
outer stripe of the outer medulla and the subcapsular layer (cortex 2) at the highest concentration
following 20 exposures.
C-15
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-3. Mode of action study findings in male F344 rat kidneys following
exposure to THF by inhalation
Exposure protocol
Control
600 mg/m3
1,800 mg/m3
5,400 mg/m3
%
LC (M)a
%
LC (M)
%
LC (M)
%
LC (M)
5 Exposures








BrdU labeling: cortex 1
100
112
95
107
109
122
153b
171
BrdU labeling: cortex 2
100
132
102
134
99
131
125
165
TUNEL: whole cortex
100
13°
115
15
107
14
92
12
5 Exposures + 3 weeks recovery






BrdU labeling: cortex 1
100
138
78d
107
88
121
110
152
BrdU labeling: cortex 2
100
140
86
121
86
120
105
147
TUNEL: whole cortex
100
9
45
4
145
13
478b
43
20 Exposures








BrdU labeling: cortex 1
100
118
119
140
159b
188
298b
352
BrdU labeling: cortex 2
100
156
101
158
113
176
186b
290
TUNEL: whole cortex
100
35
74
26
157
55
234b
82









Exposure protocol
Control
600 mg/m3
1,800 mg/m3
5,400 mg/m3
%
LA (%)
%
LA (%)
%
LA (%)
%
LA (%)
5 Exposures








a2u-globulin: whole cortex
100
6.16
136d
8.37
17 lb
10.53
178b
10.95
a2u-globulin: cortex 1
100
7.30
125
9.14
167b
12.18
175d
12.75
a2u-globulin: cortex 2
100
5.01
131
6.57
176b
8.82
188b
9.42
5 Exposures + 3 weeks recovery






a2u-globulin: whole cortex
100
5.57
150
8.35
212b
11.80
299b
16.66
a2u-globulin: cortex 1
100
6.68
154
10.32
213b
14.22
280b
18.70
a2u-globulin: cortex 2
100
4.47
141
6.30
205d
9.18
324b
14.49
20 Exposures








a2u-globulin: whole cortex
100
5.34
149d
7.97
22 lb
11.79
259b
13.84
a2u-globulin: cortex 1
100
6.20
149b
9.21
212b
13.15
253b
15.70
a2u-globulin: cortex 2
100
4.47
149d
6.66
236b
10.53
265b
11.86
aLC (M) = positively labeled cells (LCs) mean value.
V<0.01.
°Number of apoptotic cells.
dp <0.05.
Source: Adapted from BASF (2001a).
1
2	To determine whether changes in cell proliferation might reflect altered apoptosis rates,
3	apoptotic cells were also quantitated (Table C-3). The number of cells undergoing apoptosis was
4	significantly increased in the high-concentration groups exposed for 5 days and observed for
5	21 days or after 20 exposure days. Marginal increases were observed in the mid-concentration
C-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
groups for these two exposure regimens, but the results were not statistically significant. The
authors suggested the increase in apoptosis observed in the group with a recovery period might
be greater than in the 20-day exposure group, because in the latter group competing cell
proliferation and apoptosis events might have reduced the degree of apoptosis.
THF exposure also induced a,2u-globulin accumulation in male rats treated under all three
of the separate exposure regimens (Table C-3). Increases were generally concentration related,
with increases at the high concentration ranging from 175 to 280% of control levels for cortex 1
and from 188 to 324% of control levels for cortex 2, among the three exposure regimens. When
the whole cortex was used as the labeled area (LA) for the analysis, accumulation was
significantly elevated beginning at the low concentration following 5 consecutive days or
20 days of exposure. Maximum effects observed at the high concentration ranged from 178 to
299% of controls among the three exposure regimens. The accumulation of a,2u-globulin as
measured by the immunohistochemical staining technique was supported by histopathological
evaluation of control and high-concentration animals exposed to THF for 20 days. The
incidence of proximal tubule cells with grade 2 (slightly increased) staining for hyaline droplets
was 1/6 and 5/6 for controls and high-concentration animals, respectively. THF exposure had no
effect on CYP450 content or CYP450 enzyme activities in the male rat kidneys.
BASF (2001a) and Gamer et al., (2002) also evaluated cell proliferation in female
B6C3Fi mice liver following inhalation exposure to THF (Table C-4). Since chemical exposures
can have varying effects in different regions of the liver lobule, cell proliferation was evaluated
separately for zone 1 (the region adjacent to the portal triad), zone 3 (the region adjacent to the
central vein), and zone 2 (the area of the lobule intermediate between zones 1 and 3). Increased
cell proliferation was observed in zones 2 and 3 of the liver following THF exposure for 5 days
and in zone 3 following 20 exposures. No concentration-dependent increase in BrdU labeling
was observed in the animals given a 21-day recovery period, suggesting that the increases in cell
proliferation may be reversible. Coincident with the increase in BrdU labeling, the mitotic index
(MI) was increased in zone 3 after 5 or 20 exposures in the high-concentration groups. No
treatment-related change in the number of liver cells undergoing apoptosis was observed. The
number of stained cells was small, suggesting that THF did not induce an apoptotic response
under the exposure conditions. Five consecutive days of exposure to THF at the high
concentration generated a statistically significant increase in CYP450 content in the liver (125%
of controls; p < 0.05), EROD activity (192% of controls; p < 0.01), and PROD activity (321% of
controls; p < 0.05). The authors concluded that THF-induced liver tumors in female mice may
be related to increased cell proliferation, based on the increased liver weight, BrdU labeling, and
MI observed in the liver. Some histological changes were noted, including fatty infiltration and
cell proliferation including altered texture of the cytoplasm in zones 3 and 2 (more homogeneous
C-17
DRAFT - DO NOT CITE OR QUOTE
-------
1	and eosinophilic); however, no morphological signs of cell degeneration, such as cloudy
2	swelling, vacuolar degeneration, or necrosis, were found.
3
Table C-4. BrdU labeling and MI as a measure of cell proliferation in
female B6C3Fi mouse livers following exposure to THF by inhalation
Exposure protocol
Control
600 mg/m3
1,800 mg/m3
5,400 mg/m3
5 Exposures
BrdU labeling
(% of control)

%
LP (%)
%
LI (%)
%
LI (%)
%
LI (%)
Zone 1
100
1.01
110
1.11
122
1.23
143
1.44
Zone 2
100
2.54
98
2.48
117
2.96
183d
4.66
Zone 3
100
0.85
147
1.25
188
1.60
401d
3.41
Zone 1, 2, 3
100
1.46
110
1.61
132
1.93
217d
3.17
Hematoxylin
and eosin: MI

MI (%)
MI (%)
MI (%)
MI (%)
Zone 1
0.01
0.01
0.03
0.04
Zone 2
0.14
0.14
0.17
0.48b
Zone 3
0.00
0.01
0.00
0.19°
Zone 1, 2, 3
0.05
0.05
0.07
0.23b
5 Exposures + 3-week recovery
BrdU labeling
(% of control)

%
LI (%)
%
LI (%)
%
LI (%)
%
LI (%)
Zone 1
100
0.88
120
1.06
100
0.88
109
0.96
Zone 2
100
2.75
107
2.95
85
2.35
76
2.08
Zone 3
100
1.09
170b
1.85
148
1.61
137
1.49
Zone 1, 2, 3
100
1.57
124
1.95
103
1.61
96
1.51
Hematoxylin
and eosin: MI

MI (%)
MI (%)
MI (%)
MI (%)
Zone 1
0.00
0.00
0.01
0.00
Zone 2
0.02
0.01
0.04
0.08
Zone 3
0.00
0.00
0.04
0.03
Zone 1, 2, 3
0.01
0.00
0.03
0.04
20 Exposures
BrdU labeling
(% of control)

%
LI (%)
%
LI (%)
%
LI (%)
%
LI (%)
Zone 1
100
1.39
106
1.48
91
1.27
104
1.45
Zone 2
100
3.53
86
3.02
95
3.35
118
4.16
Zone 3
100
1.52
133
2.02
134
2.04
230b
3.49
Zone 1, 2, 3
100
2.51
101
2.17
103
2.22
141
3.03
Hematoxylin
and eosin: MI

MI (%)
MI (%)
MI (%)
MI (%)
Zone 1
0.05
0.05
0.01
0.05
Zone 2
0.04
0.16
0.32°
0.24b
Zone 3
0.01
0.01
0.07
0.20b
Zone 1, 2, 3
0.03
0.07
0.13°
0.16b
C-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
aLI = labeling index.
V
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
CYP450 enzyme activity that has broad activity for many CYP450 isoforms. The parallel
exposure group did not receive this pretreatment with ABT and was used to test the effects of
THF without CYP450 inhibition. The livers of the mice were evaluated for total CYP450
content and some of the CYP450 activities including EROD, PROD, and nitrophenol
hydroxylase (NPH), as well as cell proliferation (as measured by PCNA staining), and
examination by electron microscopy.
Exposure of animals at the high concentration induced a narcotic effect. Three of
18 mice died in the high-concentration group without CYP450 inhibition, and 1 of 18 mice died
in the high-concentration group pretreated with ABT. The high-concentration mice also had
reduced body weight compared with controls. No clinical effects of THF were observed at the
low concentration. No THF-related histopathology changes were observed in any of the
treatment groups, although, in the livers of ABT-pretreated mice, centrilobular fatty changes
were noted. Measurements of CYP450 content and activity revealed that CYP450s were induced
in the high-concentration mice. Liver CYP450 content was increased by 98% in the high-
concentration group, and this increase was blocked by ABT pretreatment. THF treatment
induced PROD activity by about sixfold in the high-concentration group. In the mice pretreated
with ABT, PROD activity was induced by approximately twofold by THF. EROD activity was
increased by 160% in the high-concentration mice as compared to controls in the absence of
ABT, and no induction of EROD activity was observed in the mice pretreated with ABT. These
results show that THF induces both EROD and PROD activity and that the ABT pretreatment
was an effective inhibitor of CYP450 isoform activity. In contrast to the results for PROD and
EROD, NPH activity, known to be predominantly catalyzed by human and rat CYP2E1
(Kobayashi et al., 2002; Tassaneeyakul et al., 1993) was decreased in a concentration-dependent
manner by THF and was not affected by ABT pretreatment. CYP450 content or associated
enzyme activities were not induced above basal levels in the low-concentration group.
THF exposure induced cell proliferation at the high concentration, regardless of
pretreatment with ABT. In mice exposed to 14,739 mg/m3 THF without ABT pretreatment,
PCNA staining was increased 814% relative to controls in zone 3, although a decrease to 59% of
control levels that was not statistically significant was observed in zone 2, and no difference was
observed for zone 1. The overall increase in PCNA staining for the three zones (pooled data)
was 133% of controls (not statistically significant). In the high-concentration group pretreated
with ABT, cell proliferation was even greater than the parallel THF group without pretreatment.
PCNA staining was 150, 280, and 1,050% of control levels in liver zones 1, 2, and 3,
respectively. In ABT-pretreated mice, the overall PCNA labeling for the three zones (data
pooled) was 329% of controls. No change in PCNA staining was observed in the low-
concentration groups regardless of pretreatment with inhibitor.
C-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
The data indicated that THF is an inducer of CYP450s and that THF induces cell
proliferation in the livers of female mice, particularly in zone 3 hepatocytes. Pretreatment with
the CYP450 inhibitor ABT enhanced the degree of PCNA staining, suggesting that THF itself,
rather than a downstream oxidative metabolite, is responsible for the cell proliferative response.
In mice with enzyme inhibition, the cell proliferation response was enhanced only moderately. It
is possible that this effect would have been even more dramatic if the basal as well as inducible
CYP450 activity had been blocked by the ABT pretreatments. ABT did not provide a complete
inhibition of response, producing some uncertainty about the role that CYP450s play in THF-
induced cell proliferation. A second area of uncertainty is that there were qualitative differences
in the histopathology in the ABT-pretreated mice (i.e., centrilobular fatty changes) compared to
mice without ABT pretreatment. It is not clear whether these histopathological changes that
were unique to ABT-pretreated mice could have caused hepatocytes to be more susceptible to
THF-induced liver toxicity. Even though these areas of uncertainty remain, the most possible
interpretation of the data is that the cell proliferative response of the liver in female mice is not
dependent on CYP450 activity, since treatment with the CYP450 inhibitor did not decrease the
proliferative response. This interpretation suggests that THF itself, not a metabolite, is the active
moiety in inducing cell proliferation. However, in the absence of further in vitro (or in vivo)
metabolism data with and without ABT, it is not possible to determine if THF metabolism is
actually inhibited and to what extent.
C.2.2.3. Initiation
Other than the NTP (1998) study, no direct animal cancer bioassays have been
conducted. The use of THF as a solvent control in cancer studies for other compounds provides
some limited data on the potential cancer mode of action for THF. Sawyer et al. (1988)
evaluated the tumor-initiating properties of dibenz[a,j]anthracene, cholanthrene, and their diol
and epoxide metabolites on the skin of SENCAR mice. The test compounds were dissolved in
either acetone (30 mice/group) or THF (24 mice/group). The number of papillomas/mouse and
percent of mice with papillomas was lower for THF-treated controls (5%) than for acetone-
treated controls (16%) and was much lower than for the animals treated with the test compounds
(39-97% for various treatment groups), suggesting that THF is not a potent tumor initiator.
However, interpretation of this study is limited for a number of reasons. The study authors did
not provide data on the historical incidence of papillomas. A tumor-screening protocol was used,
which did not include a control group, an adequate number of dose levels, or adequate numbers
of animals/dose group. Another complication in evaluating this study is that the tumor
incidences for the test compounds dissolved in acetone or THF could reflect cocarcinogenic
interactions.
C-21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
C.2.2.4. Inhibition of Gap Junctional Intercellular Communication
Chen et al. (1984) investigated the ability of organic solvents to inhibit gap junctional
intercellular communication (GJIC). Cocultures of 6-thioguanine-sensitive and resistant Chinese
hamster V79 fibroblast cells were treated with the test compound and the degree of metabolic
cooperation was determined by the survival of the resistant cells. The killing of resistant cells
serves as an indicator of metabolic cooperation, because the toxic 6-thioguanine metabolite that
is formed only in the sensitive cells can be passed on to normally resistant cells when gap
junctions are intact. Therefore, robust growth of the resistant cells in this assay system would
suggest that GJIC is inhibited. THF was judged to be positive (as defined by a doubling in
recovery of resistant colonies) in the metabolic cooperation assays, suggesting that THF can
inhibit GJIC. The recovery rate of resistant cells increased with increasing concentration (up to
100 [iL of THF/5 mL of medium).
Table C-5. Summary of studies on the direct mutagenicity/genotoxicity of
THF
Endpoint
Assay system
Results
(without/
with
activation)
Comments
Reference
In vitro studies
Gene mutation
—bacteria
S. typhimurium TA1535,
TA1537, TA98, TA100
-/-
Used preincubation modification of
the standard assay (NTP [1998]
study)
Mortelmans et al.
(1986)

S. typhimurium G46,
TA1535, TA100, C3076,
TA1537, D3052, TA1538,
TA98,
Escherichia coli WP2, WP2
uvrA'
-/-
Gradient technique was used in
which the mutagenic concentration
range was identified as the lowest
and highest concentration at which
distinct colonies were observed;
results presented in a summary
table without data.
McMahon et al.
(1979)

S. typhimurium TA1535,
TA1537, TA98, TA100
-/-
Screening only using a spot test
was done in strains TA1535,
TA1537, TA98; results presented in
a summary table without data
Florin et al.
(1980)

S. typhimurium TA98
ntT-
Results presented in summary text
without data
Arimoto et al.
(1982)
Clastogenicity
Micronuclei, Syrian
hamster embryo cells
nt/-
None
Gibson et al.
(1997)
Chromosome
aberration
Chinese hamster ovary cells
-/±
Slight increase with S9 not
considered positive by study
authors.; NTP study (1998)
Galloway et al.
(1987)
C-22
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Table C-5. Summary of studies on the direct mutagenicity/genotoxicity of
THF
Endpoint
Assay system
Results
(without/
with
activation)
Comments
Reference
DNA damage
Sister chromatid exchange,
Chinese hamster ovary cells
-/-
NTP study (1998)
Galloway et al.
(1987)
Cell
transformation
BALB/C-3T3 cells
-/nt
Limited activity was noted in one
of two trials in the data tables, but
not in the text of the study
Matthews et al.
(1993)
Syrian hamster embryo
cells
-/nt
No cytotoxicity was observed at the
highest test concentration
Kerckaert et al.
(1996)
NIH/3T3 cells
-/nt
THF used as control; cells treated
in vitro were injected in mice to
assess tumorigenicity
Collins et al.
(1982)
In vivo studies
Gene mutation
Drosophila sex-linked
recessive lethal
-
NTP (1998) study
Valencia et al.
(1985)
Clastogenicity
Mouse erythrocyte
micronucleus
±
Positive response only in mid-
concentration males (NTP [1998]
study)
NTP (1998)
Chromosome
aberration
Mouse bone marrow
-
NTP (1998) study
NTP (1998)
DNA damage
Mouse bone marrow, sister
chromatid exchange
-
NTP (1998) study
NTP (1998)
Mouse hepatocyte
unscheduled DNA synthesis
-
NTP (1998) study
Mirsalis et al.
(1983)
ant = not tested.
Mortelmans et al. (1986) reported that THF did not induce reverse mutations with or
without metabolic activation in four tester strains of the S. typhimurium test system. THF was
also negative (with or without activation) when tested in a battery of eight strains of
S. typhimurium and two Escherichia coli strains by using a modification of the standard assay
(McMahon et al., 1979) or in four S. typhimurium strains (Florin et al., 1980). Several studies
used or specifically examined the effects of THF as a soluble solvent in the S. typhimurium
mutagenicity assays and generally support the conclusions of the above-mentioned more
definitive studies. Hageman et al. (1988), in a study of the mutagenicity of frying oils, reported
that THF solvent controls were nonmutagenic (with or without activation) in tester strains TA97,
TA100, and TA104 relative to mutagen-containing oil samples. Maron et al. (1981) screened a
series of solvents for compatibility with the S. typhimurium test system and reported that, while
high-dose THF was toxic to the four tester strains used, it did not affect the mutagenicity of
C-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
33
34
35
36
benzo(a)pyrene at lower levels (50 (iL/plate) in strain TA100 in the plate incorporation protocol.
THF was judged to be an unsatisfactory solvent for the preincubation assay due to higher
cytotoxicity observed in this protocol modification. Finally, THF was reported to enhance the
mutagenicity of tryptophan pyrolysate mutagens in S. typhimurium preincubation assay when
used as a solvent (Arimoto et al., 1982). No potential mode of action for this effect was given,
but the authors reported (no quantitative data provided) that the solvent was not itself mutagenic
in tester strain TA98 with activation. The studies by Hageman et al. (1988), Arimoto et al.
(1982), and Maron et al. (1981) are of limited value for assessing the mutagenic potential of THF
because THF served as the control solvent in these studies and it is not clear if the results for
THF were compared to untreated samples.
THF was also negative in a variety of in vitro assays evaluating chromosome and DNA
damage up to cytotoxic concentrations. Gibson et al. (1997) reported that THF did not increase
micronuclei formation when assayed in Syrian hamster embryo cells at concentrations that
significantly reduced cell number. Galloway et al. (1987) reported some increase in total
chromosome aberrations in the presence of S9 activation in Chinese hamster ovary cells. A
majority of the aberrations were classified as simple, including breaks and terminal deletions.
The study authors suggested that these increases were insufficient to be scored as a positive
result. As part of this same study, Galloway et al. (1987) reported that THF did not induce sister
chromatid exchanges in this cell system at cytotoxic doses.
THF was judged to be inactive when tested in the standard BALB/c-3T3 mouse cell
transformation assay (Matthews et al., 1993). A Syrian hamster embryo cell assay was also
negative for cell transformation when THF was tested at concentrations up to 5 mg/mL
(Kerckaert et al., 1996). Collins et al. (1982) evaluated the in vivo tumorigenicity of NIH/3T3
cells transformed in vitro by benzo[a]pyrene-trans-7,8-dihydrodiol-9,10-epoxide (BPE)
dissolved in THF. The ability of BPE-treated cells to induce tumors in normal mice (strain not
specified) and ATxFL mice having a compromised immune response (thymectomized, lethally
irradiated, and restored with syngeneic liver cells) was greater than the tumorigenicity of cells
treated with THF only. Cells from 46/57 BPE-treated plates were tumorigenic in vivo, whereas
cells from only 2/20 of the THF-treated plates were tumorigenic when injected in mice. The
background tumor rate for untreated mice was not reported, but the low incidence of tumors
induced by THF-treated cells as compared with positive controls suggested that THF did not
significantly increase the rate of cell transformation.
THF has also generated negative findings in in vivo genotoxicity assays. THF did not
induce sex-linked recessive lethal mutation in Drosophila melanogaster in a screening test for
48 chemicals for NTP (Valencia et al. 1985). NTP (1998) evaluated the formation of
micronuclei in peripheral blood erythrocytes in male and female mice at the end of their 13-week
C-24
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
inhalation study. There was only a statistically significant increased incidence of micronucleated
normochromatic erythrocytes at the mid concentration in males. The effect was not
concentration dependent, and no corresponding increase was seen for females. The results were
considered by NTP to be equivocal. In a bioassay for chromosomal aberrations, male B6C3Fi
mice received THF by i.p. injection at doses of up to 2,000 mg/kg. No significant increase in the
number of aberrations/cell or percent of bone marrow cells with aberrations was observed (NTP,
1998).
In vivo assays for DNA damage have also been conducted for THF. Male B6C3Fi mice
received THF doses of up to 2,000 mg/kg by i.p. injection. Bone marrow cells were harvested
after 23 or 42 hours of exposure. In the 23-hour treatment protocol, a significant increase in the
mean number of sister chromatid exchanges/cell was reported for the high-dose animals.
However, this effect was observed in only one of the two replicate trials. No increase in sister
chromatid exchanges was reported for the animals exposed for 42 hours. NTP (1998)
characterized these results as negative. In another assay for DNA damage, Mirsalis et al. (1983
[published abstract]) reported that in vivo treatment of male rats with THF did not induce
unscheduled DNA synthesis in hepatocytes.
Loureiro et al. (2004, 2000) reported formation of three DNA adducts from reaction of
2'-deoxyguanosine with trans, trans-2,4-decadienal occurring in the presence of oxidized THF.
Later on, the same investigators structurally characterized these novel stable adducts produced
by the reaction of THF oxidation products with l,N2-etheno-2'-deoxy-guanosine (Hermida et al.,
2006; Loureiro et al., 2005). They also claim that an interaction leading to DNA-THF adducts
may be a contributing factor to the observed toxicological effects associated with THF exposure.
However, the limited information available from in vitro and in vivo genotoxicity studies point
to THF as non-mutagenic (NTP, 1998). Further investigations are necessary to evaluate the
possible interaction of THF oxidation products with DNA and their role in mutagenic mode(s) of
action or THF-induced carcinogenic activity in rodents.
In summary, the genotoxic potential of THF has been evaluated in a variety of in vitro
and in vivo assays. Nearly all the results are conclusively negative, with equivocal findings
reported in a small number of assays that have been conducted. Taken together, these data
support the conclusion that THF is not likely genotoxic.
C.2.3. Noncancer Mode of Action Information
THF was evaluated in a series of short-term in vitro tests to assess its potential for
cytotoxicity (Matthews et al., 1993; Dierickx, 1989; Curvall et al., 1984; Pettersson et al., 1982).
The results of these studies suggest that THF is not cytotoxic.
C-25
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
The available data suggest that THF metabolism is extensive and that oxidative
metabolism may be due to CYP450 isozymes. However, the identity of the isozymes responsible
for THF metabolism has not been elucidated. In addition, whether THF or one of its metabolites
is responsible for the observed toxicological effects is unknown. Some mode of action data
(BASF, 2001b) suggest that the parent compound might be the active form for liver toxicity and
that metabolites might be responsible for neurological effects.
In the two-generation reproduction study (Hellwig et al., 2002; BASF, 1996) of THF in
rats by the oral route, increased kidney weights in F0/F1 adults were observed in the high-dose
groups. The mode of action for THF-induced kidney toxicity is unknown. Two possible modes
of action were considered. First, THF exposure by the inhalation route induces CYP450 activity
in the mouse liver (Gamer et al., 2002; BASF, 2001a, b), and therefore it is possible that a
similar response could occur in the rat kidney. However, data available showed that acute
inhalation exposures had no effect on kidney CYP450 activity in male F344 rats (Gamer et al.,
2002; BASF, 2001a). These results are not directly comparable to the oral two-generation study
since the exposure duration and rat strains differed between the two studies. Nevertheless, the
only directly available data do not support the idea that CYP450 induction is responsible for the
observed increase in kidney weight. Furthermore, since it is not known whether THF itself or a
metabolite is the active moiety with respect to the kidney effects, it is not clear whether an
induction of CYP450 activity is likely to increase or decrease THF toxicity in the affected organ.
Some data suggest that an a2U-globulin-associated mode of action could contribute to THF-
induced nephrotoxicity. However, there is insufficient evidence to conclude that the kidney
effects observed following THF exposure are related to the accumulation of a2u-globulin for the
following reasons (See Section 4.7.3.1 for analysis of the available data).
Decreased body weight gain in F1/F2 pups and delayed developmental stages (delayed
eye opening) in F2 pups were also observed in the high-concentration groups of the two-
generation reproduction study of THF in rats by the oral route (BASF, 1996). Two hypotheses
for the observed decrease in pup body weight gain were considered. First, decreased maternal
water intake during the lactation period could limit maternal milk production, resulting in
decreased nutrition for pups and corresponding decreases in their growth, assuming that the
composition of the milk did not change to maintain its nutritional value at times when water
intake is low. Published studies have showed an association between water restriction and
decreased volume of milk production in both humans and livestock (Hossaini-Hilali et al., 1994;
Morse et al., 1992; Dusdiecker et al., 1985; Little et al., 1980), and, therefore, the proposed
explanation of decreased pup weight due to decreased milk production is biologically possible.
The temporal pattern of decreased pup body weight gain (significant decrements only during
PNDs 4-14) correlates well to the postnatal lactation period where milk intake is greatest, and
C-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
thus demand on a limited maternal milk supply would be expected to be most dramatic. The
absence of an effect on pup body weight gain for PNDs 14-21 corresponds to the period where
pups begin direct food and water intake and therefore depend less on milk production as a source
of nutrition. Whether the observed decrease in water intake was due to a toxic effect of THF or
was secondary to poor palatability is not clear from the available data. No study was conducted
to test specifically whether THF, at the concentrations tested, reduced water intake solely
because of taste aversion. Also, the two-generation study (Hellwig et al., 2002; BASF, 1996) did
not include a water-restricted control group to separate the effects of decreased water intake from
those that are induced directly by THF. In some cases the temporal pattern of water intake can
provide evidence for decreased palatability, where decreased water consumption at initial
introduction of the treated water is greater than the decrease observed at later exposure periods.
However, for the two-generation study (Hellwig et al., 2002; BASF, 1996), the decrease in water
intake was not greater for week 1 versus other weeks during the premating period. This result by
itself is not sufficient to determine whether decreased water intake was secondary to palatability,
since water intake data for initial days of exposure were not reported (weekly summaries were
provided in the report), and this is only an indirect measure of potential taste aversion.
The second hypothesis is that THF itself induces a direct effect on pup development. Several
considerations provide indirect support for a role of THF in the observed decreased pup body
weight gain. In the two-generation study (Hellwig et al., 2002; BASF, 1996), THF induced
developmental effects in the F2 pups (delayed eye opening and increased incidence of sloped
incisors) in addition to decreased pup weight gain. While this observation that other
developmental indices are affected by THF treatment supports a role of THF exposure, it could
simply reflect additional developmental delays resulting from decreased milk availability. The
developmental effects of THF have also been tested in inhalation exposures in rodents, which
would not be subject to issues of water palatability. However, the available studies did not
assess postnatal development (sacrifice was at the end of gestation) and therefore do not provide
directly comparable responses to the oral two-generation study. In the inhalation studies,
maternally toxic concentrations of THF reduced fetal survival and weight and increased the
incidence of fetal skeletal alterations in rats and mice (Mast et al., 1992; DuPont Haskell
Laboratory, 1980). These inhalation data are consistent with the hypothesis that THF can induce
developmental effects. On the other hand, even though the two-generation study did not fully
evaluate fetal toxicity outcomes, the absence of a THF effect on litter size or pup weight during
the early postnatal period (days 1-4) suggests that fetal effects were not occurring in the oral
dosing study. One explanation for the absence of an indication of fetal effects in the two-
generation study, other than dose route, is that the degree of maternal toxicity in the inhalation
studies was more severe than in the drinking water study. However, a subtle effect on male rat
C-27
DRAFT - DO NOT CITE OR QUOTE
-------
1	fertility/fecundity may exist following exposure to a high concentration of THF in drinking water
2	based on a slight decrease (not statistically significant) in the mean number of delivered F2 pups
3	and a finding of one infertile F1 parental male rat in the high dose group (Section 4.3.1).
C-28
DRAFT - DO NOT CITE OR QUOTE
-------