DRAFT - DO NOT CITE OR QUOTE EPA/635/R-11/005C
www.epa.gov/iris
vvEPA
TOXICOLOGICAL REVIEW
OF
BIPHENYL
(CAS No. 92-52-4)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
July 2011
NOTICE
This document is an Interagency Science Consultation draft. This information is distributed
solely for the purpose of pre-dissemination peer review under applicable information quality
guidelines. It has not been formally disseminated by EPA. It does not represent and should not
be construed to represent any Agency determination or policy. It is being circulated for review
of its technical accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
DRAFT - DO NOT CITE OR QUOTE
-------
CONTENTS—TOXICOLOGICAL REVIEW OF BIPHENYL (CAS No. 92-52-4)
LIST OF TABLES vi
LIST OF ABBREVIATIONS AND ACRONYMS x
FOREWORD xii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiii
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION 3
3. TOXICOKINETICS 6
3.1. ABSORPTION 6
3.2. DISTRIBUTION 7
3.3. METABOLISM 8
3.3.1. Identification of Metabolites 8
3.3.1.1. Results from in vivo Animal Studies 8
3.3.1.2. Results from in vitro Studies with Animal and Human Cells or Tissues 10
3.3.2. Metabolic Pathways 11
3.3.2.1. Description of Metabolic Scheme and Enzymes Involved 11
3.3.3. Regulation of Metabolism, Sites of Metabolism, and Relationships to Toxic
Effects 14
3.3.3.1. Evidence for Induction of Phase I and II Enzymes 14
3.3.3.2. Demonstrated Tissue Sites of Metabolism 16
3.3.3.3. Possible Relationships Between Metabolites and Toxic Effects 16
3.4. ELIMINATION 17
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS 18
4. HAZARD IDENTIFICATION 19
4.1. STUDIES IN HUMANS 19
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION 23
4.2.1. Oral Exposure 24
4.2.1.1. Subchronic Toxicity 24
4.2.1.2. Chronic Toxicity and Carcinogenicity 25
4.2.2. Inhalation Studies 42
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND
INHALATION 44
4.3.1. Oral Exposure 44
4.3.2. Inhalation Exposure 47
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES 47
4.4.1. Acute and Short-term Toxicity Data 47
4.4.2. Kidney/Urinary Tract Endpoint Studies 49
4.4.3. Biphenyl as a Tumor Promoter 53
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE
OF ACTION 54
4.5.1. Effects on the Urinary Tract of Rats 54
4.5.2. Effects on the Liver of Mice 56
iii DRAFT - DO NOT CITE OR QUOTE
-------
4.5.3. Estrogenic Effects 56
4.5.4. Effects on Apoptosis 57
4.5.5. Mitochondrial Effects 58
4.5.6. Genotoxicity 58
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS 68
4.6.1. Oral 74
4.6.2. Inhalation 74
4.6.3. Mode-of-Action Information 75
4.7. EVALUATION OF CARCINOGENICITY 76
4.7.1. Summary of Overall Weight of Evidence 76
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence 79
4.7.3. Mode-of-Action Information 81
4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats 81
4.7.3.2. Mode-of-Action Information for Liver Tumors in Female Mice 87
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 90
4.8.1. Possible Childhood Susceptibility 90
4.8.2. Possible Gender Differences 91
4.8.3. Other 91
5. DOSE-RESPONSE ASSESSMENTS 93
5.1. ORAL REFERENCE DOSE (RfD) 93
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification 93
5.1.2. Methods of Analysis—Including Models 96
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs) 104
5.1.4. Previous RfD Assessment 105
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 106
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification... 106
5.2.2. Previous RfC Assessment 107
5.3. UNCERTAINTIES IN THE RfD and RfC 107
5.4. CANCER ASSESSMENT 108
5.4.1. Choice of Study/Data - with Rational and Justification 109
5.4.2. Dose-Response Data 109
5.4.3. Dose Adjustments and Extrapolation Method(s) 110
5.4.4. Oral Slope Factor and Inhalation Unit Risk 113
5.4.5. Uncertainties in Cancer Risk Values 113
5.4.5.1. Oral Slope Factor 113
5.4.5.2. Inhalation Unit Risk 114
5.4.6. Previous Cancer Assessment 115
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE RESPONSE 116
6.1. HUMAN HAZARD POTENTIAL 116
6.1.1. Noncancer 116
6.1.2. Cancer 117
6.2. DOSE RESPONSE 118
6.2.1. Noncancer/Oral 118
6.2.2. Noncancer/Inhalation 118
6.2.3. Cancer/Oral 118
6.2.4. Cancer/Inhalation 119
iv DRAFT - DO NOT CITE OR QUOTE
-------
7. REFERENCES 120
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION A-l
APPENDIX B. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE
DOSE B-l
APPENDIX C. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR C-l
v DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF TABLES
Table 2-1. Physicochemical properties of biphenyl 4
Table 3-1. Metabolites of biphenyl identified in urine, feces, and bile of male albino rats 9
Table 4-1. Biphenyl concentrations in the air of a Finnish paper mill producing biphenyl-
impregnated fruit wrapping paper 20
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl: comparison
with 60 unexposed males 21
Table 4-3. Exposure data and clinical features for five Parkinson's Disease patients with
occupational exposure to biphenyl 23
Table 4-4. Incidences of urinary bladder lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 27
Table 4-5. Incidences of ureter and kidney lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 29
Table 4-6. Body and organ weight data for male and female rats administered biphenyl in
the diet for 2 years 33
Table 4-7. Survival rate, body weight, food consumption, and daily biphenyl intake in
mice fed diets containing biphenyl for 2 years 36
Table 4-8. Dose-related changes in selected clinical chemistry values from male and
female BDFi mice exposed to biphenyl via the diet for 2 years 37
Table 4-9. Incidences of gross and histopathological findings in male and female BDFi
mice fed diets containing biphenyl for 2 years 38
Table 4-10. Incidences of selected tumor types among controls and mice administered
biphenyl orally for 18 months 41
Table 4-11. Incidences of selected histopathologic lesions in tissues of CD-I mice exposed
to biphenyl vapors 7 hours/day, 5 days/week for 13 weeks 44
Table 4-12. Prenatal effects following oral administration of biphenyl to pregnant Wistar
rats on GDs 6-15 45
Table 4-13. Summary of reproductive data in albino rats exposed to dietary biphenyl 47
Table 4-14. Number of Wistar rats exposed to biphenyl and the degree of change in kidney
weight and cellular architecture 51
Table 4-15. Content of biphenyl sulphate conjugates in urine and urinary crystals from
F344 rats treated with biphenyl and potassium bicarbonate (to elevate the pH
andK+ concentration of the urine) 55
Table 4-16. Genotoxicity test results for biphenyl 59
Table 4-17. Genotoxicity test results for biphenyl metabolites 63
Table 4-18. Summary of major studies evaluating effects of biphenyl after oral
administration in rats and mice 69
vi DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-19. Summary of major studies evaluating effects of biphenyl after inhalation
exposure in rats and mice 73
Table 5-1. BMD modeling datasets for incidences of nonneoplastic effects in the urinary
tract of male and female F344 rats exposed to biphenyl in the diet for 2 years 97
Table 5-2. BMD modeling datasets for body weight, selected clinical chemistry results,
and histopathological kidney effects in male and female BDFi mice exposed
to biphenyl in the diet for 2 years 98
Table 5-3. BMD modeling dataset for incidence of litters with fetal skeletal anomalies
from Wistar rat dams administered biphenyl by gavage on GDs 6-15 99
Table 5-4. Summary of BMDs/BMDLs for selected nonneoplastic effects following oral
exposure of rats and mice to biphenyl 102
Table 5-5. Incidence data for tumors in the urinary bladder of male and female F344 rats
exposed to biphenyl in the diet for 2 years 109
Table 5-6. Incidence data for liver tumors in male and female BDFi mice fed diets
containing biphenyl for 2 years 110
Table 5-7. Scaling factors for determining HEDs to use for BMD modeling of female
BDFi mouse liver tumor incidence data from Umeda et al. (2005) Ill
Table 5-8. Incidence of liver adenomas or carcinomas (combined) in female BDFi mice
fed diets containing biphenyl for2years 112
Table 5-9. POD and oral slope factor derived from liver tumor incidence data from BDFi
female mice exposed to biphenyl in the diet for 2 years 113
Table B-l. BMD modeling datasets for incidences of nonneoplastic effects in the urinary
tract of male and female F344 rats exposed to biphenyl in the diet for 2 years .... B-l
Table B-2. BMD modeling datasets for body weight, selected clinical chemistry results,
and histopathological kidney effects in male and female BDFi mice exposed
to biphenyl in the diet for 2 years B-2
Table B-3. BMD modeling dataset for incidence of litters with fetal skeletal anomalies
from Wistar rat dams administered biphenyl by gavage on GDs 6-15 B-3
Table B-4. Summary of BMD modeling results for incidence of renal nodular transitional
cell hyperplasia in male F344 rats exposed to biphenyl in the diet for 2 years B-3
Table B-5. Summary of BMD modeling results for incidence of renal nodular transitional
cell hyperplasia in female F344 rats exposed to biphenyl in the diet for 2 years.. B-5
Table B-6. Summary of BMD modeling results for incidence of renal simple transitional
cell hyperplasia in male F344 rats exposed to biphenyl in the diet for 2 years B-7
Table B-7. Summary of BMD modeling results for incidence of renal simple transitional
cell hyperplasia in female F344 rats exposed to biphenyl in the diet for 2 years.. B-9
Table B-8. Summary of BMD modeling results for incidence of mineralization in renal
pelvis of maleF344 rats exposed to biphenyl in the diet for 2 years B-ll
Table B-9. Summary of BMD modeling results for incidence of mineralization in renal
pelvis of female F344 rats exposed to biphenyl in the diet for 2 years B-l3
vii DRAFT - DO NOT CITE OR QUOTE
-------
Table B-10. Summary of BMD modeling results for incidence of hemosiderin deposits in
the kidney of female F344 rats exposed to biphenyl in the diet for 2 years B-15
Table B-l 1. Summary of BMD modeling results for incidence of papillary mineralization
in the kidney of male F344 rats exposed to biphenyl in the diet for 2 years B-l7
Table B-12. Summary of BMD modeling results for incidence of papillary mineralization
in the kidney of female F344 rats exposed to biphenyl in the diet for 2 years.... B-l 9
Table B-13. Summary of BMD modeling results for incidence of combined transitional
cell hyperplasia in the bladder of male F344 rats exposed to biphenyl in the
diet for 2 years B-21
Table B-14. Summary of BMD modeling results for incidence of mineralization in the
kidney (inner stripe outer medulla) of male BDFi mice exposed to biphenyl in
the diet for 2 years B-23
Table B-15. Summary of BMD modeling results for incidence of mineralization in the
kidney (inner stripe outer medulla) of female BDFi mice exposed to biphenyl
in the diet for 2 years B-25
Table B-16. BMD model results for serum LDH activity in female BDFi mice exposed to
biphenyl in the diet for 2 years B-27
Table B-17. BMD modeling results for serum AST activity in female BDFi mice exposed
to biphenyl in the diet for 2 years B-28
Table B-l8. BMD modeling results for serum ALT activity in female BDFi mice exposed
to biphenyl in the diet for 2 years B-31
Table B-19. BMD modeling results for serum AP activity in female BDFi mice exposed to
biphenyl in the diet for 2 years B-32
Table B-20. BMD modeling results for changes in BUN levels (mg/dL) in male BDFi
mice exposed to biphenyl in the diet for 2 years B-33
Table B-21. BMD modeling results for changes in BUN levels (mg/dL) in female BDFi
mice exposed to biphenyl in the diet for 2 years B-36
Table B-22. BMD modeling results for changes in mean terminal body weight in male
BDFi mice exposed to biphenyl in the diet for 2 years B-37
Table B-23. BMD modeling results for changes in mean terminal body weight in female
BDFi mice exposed to biphenyl in the diet for 2 years B-38
Table B-24. Summary of BMD modeling results for incidence of litters with fetal skeletal
anomalies from Wistar rat dams administered biphenyl by gavage on GDs 6-
15 B-40
Table C-l. Incidences of liver adenomas or carcinomas (combined) in female BDFi mice
fed diets containing biphenyl for 2 years C-l
Table C-2. Model predictions for liver tumors (adenomas or carcinomas combined) in
female BDFi mice exposed to biphenyl in the diet for 2 years C-2
viii DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
3-1. Schematic presentation of the metabolic pathways of biphenyl 13
5-1. NOAELs and LOAELs for noncancer effects in rats and mice from repeated oral
exposure to biphenyl 94
5-2. BMDs and BMDLs for selected noncancer effects in rats and mice from repeated oral
exposure to biphenyl 103
ix DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS
ACGIH
AIC
ALT
ALP
AP
AST
BBN
BMD
BMR
BMDS
BrdU
BUN
CA
CASRN
CHL
CHO
CYP
CVSF
DF
DNA
EEC
EHEN
EMG
ENMG
GC
GC-MS
GD
GOT
GPT
HED
HGPRT
HPLC
i.p.
IRIS
KO/W
Km
LD50
LDH
LOAEL
MCV
NOAEL
PBPK
PD
POD
PPAR
RfC
American Conference of Governmental Industrial Hygienists
Akaike's Information Criterion
alanine aminotransferase
alkaline phosphatase
alkaline phosphatase
aspartate aminotransferase
N-butyl-N-(4-hydroxybutyl)nitrosamine
benchmark dose
benchmark response
Benchmark Dose Software
5-bromo-2-deoxyuridine
blood urea nitrogen
chromosomal aberration
Chemical Abstracts Service Registry Number
Chinese hamster lung
Chinese hamster ovary
cytochrome P-450
conduction velocity of the slowest motor fibers
degrees of freedom
deoxyribonucleic acid
el ectroencephal ography
N-ethyl-N-hydroxyethylnitrosamine
electromyographic
electroneuromyography
gas chromatography
gas chromatography-mass spectrometry
gestation day
glutamate oxaloacetate transaminase
glutamate pyruvate transaminase
human equivalent doses
hypoxanthine guanine phosphoribosyl transferase
high-performance liquid chromatography
intraperitoneal or intraperitoneally
Integrated Risk Information System
octanol/water partition coefficient
Michaelis constant
median lethal dose
lactate dehydrogenase
lowest-observed-adverse-effect level
motor conduction velocity
no-observed-adverse-effect level
physiologically based pharmacokinetic
Parkinson's disease
point of departure
peroxisome proliferator activated receptors
reference concentration
DRAFT - DO NOT CITE OR QUOTE
-------
RfD reference dose
ROS reactive oxygen species
RR relative risk
SCE sister chromatid exchange
SD standard deviation
TLV threshold limit value
TMS trimethylsilyl
TWA time-weighted average
UDS unscheduled DNA synthesis
UF uncertainty factors
UGT uridine diphosphate glucuronosyl transferase
U.S. EPA U.S. Environmental Protection Agency
xi DRAFT - DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to biphenyl.
It is not intended to be a comprehensive treatise on the chemical or toxicological nature of
biphenyl.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of data and related uncertainties. The discussion is intended to convey the limitations
of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
xii DRAFT - DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER/AUTHOR
Zheng (Jenny) Li, Ph.D, DABT
U.S. EPA, ORD/NCEA
Washington, DC
CONTRIBUTORS
Christine Yuyang Cai, MS, PCP
U.S. EPA, ORD/NCEA
Washington, DC
J. Connie Kang-Sickel, Ph.D
U.S. EPA, ORD/NCEA
Washington, DC
CONTRACTOR SUPPORT
George Holdsworth, Ph.D
Lutz W. Weber, Ph.D, DABT
Oak Ridge Institute for Science and Education
Oak Ridge, TN
David Wohlers, Ph.D
Joan Garey, Ph.D
Peter McClure, Ph.D, DABT
SRC, Inc.
Syracuse, NY
REVIEWERS
This document has been provided for review to EPA scientists.
INTERNAL EPA REVIEWERS
Maureen Gwinn, Ph.D, DABT
U.S. EPA, ORD/NCEA
Washington, DC
Margaret Pratt, Ph.D
U.S. EPA, ORD/NCEA
Washington, DC
xiii DRAFT - DO NOT CITE OR QUOTE
-------
1 1. INTRODUCTION
2
3
4 This document presents background information and justification for the Integrated Risk
5 Information System (IRIS) Summary of the hazard and dose-response assessment of biphenyl.
6 IRIS Summaries may include oral reference dose (RfD) and inhalation reference concentration
7 (RfC) values for chronic and other exposure durations, and a carcinogenicity assessment.
8 The RfD and RfC, if derived, provide quantitative information for use in risk assessments
9 for health effects known or assumed to be produced through a nonlinear (presumed threshold)
10 mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
11 uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
12 population (including sensitive subgroups) that is likely to be without an appreciable risk of
13 deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
14 analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
15 inhalation RfC considers toxic effects for both the respiratory system (portal-of-entry) and for
16 effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
17 values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
18 acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
19 lifetime) exposure durations, all of which are derived based on an assumption of continuous
20 exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
21 derived for chronic exposure duration.
22 The carcinogenicity assessment provides information on the carcinogenic hazard
23 potential of the substance in question and quantitative estimates of risk from oral and inhalation
24 exposure may be derived. The information includes a weight-of-evidence judgment of the
25 likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
26 effects may be expressed. Quantitative risk estimates may be derived from the application of a
27 low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
28 the estimate of risk per mg/kg-day of oral exposure. Similarly, a plausible inhalation unit risk is
29 an upper bound on the estimate of risk per ug/m3 air breathed.
30 Development of these hazard identification and dose-response assessments for biphenyl
31 has followed the general guidelines for risk assessment as set forth by the National Research
32 Council (NRC, 1983). U.S. Environmental Protection Agency (U.S. EPA) Guidelines and Risk
33 Assessment Forum Technical Panel Reports that may have been used in the development of this
34 assessment include the following: Guidelines for the Health Risk Assessment of Chemical
35 Mixtures (U.S. EPA, 1986a), Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1986b),
36 Recommendations for and Documentation of Biological Values for Use in Risk Assessment (U.S.
37 EPA, 1988), Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Interim
38 Policy for Particle Size and Limit Concentration Issues in Inhalation Toxicity (U. S. EPA,
DRAFT - DO NOT CITE OR QUOTE
-------
1 1994a), Methods for Derivation of Inhalation Reference Concentrations and Application of
1 Inhalation Dosimetry (U.S. EPA, 1994b), Use of the Benchmark Dose Approach in Health Risk
3 Assessment (U.S. EPA, 1995), Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA,
4 1996), Guidelines for Neurotoxicity Risk Assessment (U.S. EPA, 1998), Science Policy Council
5 Handbook: Risk Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance
6 Document (U.S. EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment
7 of Chemical Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference
8 Concentration Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S.
9 EPA, 2005a), Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
10 Carcinogens (U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA,
11 2006a), and A Framework for Assessing Health Risks of Environmental Exposures to Children
12 (U.S. EPA, 2006b).
13 The literature search strategy employed for this compound was based on the Chemical
14 Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
15 scientific information submitted by the public to the IRIS Submission Desk was also considered
16 in the development of this document. The relevant literature was reviewed through June 2011.
17
DRAFT - DO NOT CITE OR QUOTE
-------
1 2. CHEMICAL AND PHYSICAL INFORMATION
2
3
4 Pure biphenyl is a white or colorless crystalline solid that usually forms leaflets or scales;
5 commercial preparations may be yellowish or slightly tan (NLM, 2007). Biphenyl is said to
6 have a pleasant odor that is variably described as peculiar, butter-like, or resembling geraniums
7 (NLM, 2007; IPCS, 1999). Biphenyl melts at 69°C and has a vapor pressure of 8.93 x 10"3 mm
8 Hg at 25°C, making it likely to enter the environment in its vaporized form (NLM, 2007). If
9 particle-bound biphenyl is precipitated to the ground, it is likely to be reintroduced to the
10 atmosphere by volatilization. The water solubility of biphenyl is 7.48 mg/L at 25°C. The
11 logarithm of the octanol/water partition coefficient (K0/w) of biphenyl of 3.98 suggests a potential
12 for bioaccumulation (NLM, 2007). Because it is biodegraded with an estimated half-life of 2
13 and 3 days in air and water, respectively (NLM, 2007), and is metabolized rapidly by humans
14 and animals (see Section 3), bioaccumulation does not occur (IPCS, 1999). Biphenyl is
15 ubiquitous in the environment, with reported indoor air concentrations of 0.16-1 |ig/m3 and
16 outdoor levels of approximately 0.03 |ig/m3 (IPCS, 1999). The physicochemical properties of
17 biphenyl are summarized in Table 2-1.
18
19
DRAFT - DO NOT CITE OR QUOTE
-------
Table 2-1. Physicochemical properties of biphenyl
Synonyms
CASRN
Chemical structure
Chemical formula
Molecular weight
Melting point
Boiling point
Specific gravity
Vapor pressure
Log K0/w
Water solubility
Henry's law constant
Conversion factors
Diphenyl, 1,1 '-biphenyl, l,l'-diphenyl, bibenzene, phenylbenzene,
lemonene, Carolid AL, Phenador-X, Tetrosine LY
92-52-4
O^O
C^HIO
154.2
69°C
256°C
1.041g/cm3at20°C
8.93 x 10-3mmHgat25°C
4.01
4.1T
4.17or5.27-5.46b
7.48 mg/L at 25°C
3.08 x 10"4 atm-m3/mol at 25°C
1 ppm = 6.31 mg/m3; 1 mg/m3 = 0.159 ppm
"Monsanto (1979).
Estimated by different methods: Dow (1971).
Source: NLM(2007).
1
2 Biphenyl exists naturally as a component of crude oil or coal tar. It is primarily produced
3 by debromination/dimerization of bromobenzene, is isolated as a byproduct of the
4 hydrodealkylation of toluene (yield approximately 1%), or is synthesized by catalytic
5 dehydrocondensation of benzene. Biphenyl is currently not registered for use as a pesticide in
6 the United States, but is still used in other countries as a fungistat, most commonly to preserve
7 packaged citrus fruits or in plant disease control (NLM, 2007). The major uses of biphenyl today
8 are as chemical synthesis intermediates (among them, the sodium salt of 2-hydroxy-biphenyl, a
9 pesticide known as Dowicide 1), as dye carriers in polyester dyeing, and as components in heat
10 transfer fluids (in particular Dowtherm A or Therminol® VP-1, consisting of 26.5% biphenyl
11 and 73.5% diphenyl oxide). Historically, biphenyl was the primary byproduct in the
12 manufacture of poly chlorinated biphenyls (PCBs) until PCBs were banned in the 1970's (U.S.
13 EPA 1978). The purity of technical biphenyl ranges from 93-99.9%. The prevalent impurities
14 in technical preparations are terphenyls, a side product from the dehydrocondensation of
DRAFT - DO NOT CITE OR QUOTE
-------
1 benzene. Biphenyl is rated as a high-volume production chemical. Annual U.S. production in
2 1990 was approximately 1.6 x 104 metric tons (NLM, 2007).
DRAFT - DO NOT CITE OR QUOTE
-------
1 3. TOXICOKINETICS
2
3
4 3.1. ABSORPTION
5 No quantitative studies on the absorption of biphenyl have been conducted in humans.
6 However, evidence of hepatic toxicity produced by a probable combination of inhalation and
7 dermal exposures to biphenyl was identified as the likely cause of death of a worker in a
8 biphenyl-impregnated fruit wrapping paper production facility and provides prima facie
9 qualitative evidence of absorption in a human subject (Hakkinen et al., 1973). This worker had
10 11 years of exposure to biphenyl; at the time of his death, air measurements in the factory were
11 as high as 123 mg/m3. Evidence of hepatic and nervous system toxicity was also observed in
12 eight co-workers (Hakkinen et al., 1973).
13 Animal studies in rats, rabbits, guinea pigs, and pigs indicate that biphenyl is rapidly and
14 readily absorbed following oral exposure, as evidenced by the detection of metabolites in urine
15 and bile (Meyer, 1977; Meyer and Scheline, 1976; Meyer et al., 1976a, b). Results from a study
16 with rats administered radiolabeled biphenyl indicate extensive oral absorption (about 85% of
17 administered dose) (Meyer et al., 1976a, see below), whereas results from studies of rabbits,
18 guinea pigs, and pigs administered nonlabeled biphenyl indicate less extensive oral absorption in
19 the range of 28-49% of the administered dose (Meyer, 1977; Meyer et al., 1976b).
20 In the most quantitative assessment of absorption using radiolabeled biphenyl, male
21 albino rats (n = 3; body weight = 200-300 g) given an oral dose of 100 mg/kg (0.7-1.0 uCi) of
22 [14C]-biphenyl (in soy oil) excreted 75-80% of the radioactivity in their urine within the first
23 24 hours, with a total average urinary excretion of 84.8% and fecal excretion of 7.3% during the
24 96-hour post-dosing period (Meyer et al., 1976a). Only a trace of [14C]-CO2 was detected in
25 expired air and <1% of the radioactivity was recovered from tissues obtained at the 96-hour
26 sacrifice of the rats. These results indicate that at least 85% of the administered dose was
27 absorbed in rats.
28 Less quantitative estimates of oral absorption have been provided in analytical studies of
29 biphenyl and metabolites in urine and feces from rabbits (Meyer, 1977), guinea pigs (Meyer,
30 1977), and pigs (Meyer et al., 1976b) following oral administration of single 100-mg/kg doses of
31 unlabeled biphenyl.
32 Male White Land rabbits and SffPIR guinea pigs were given biphenyl (100 mg/kg) by
33 gavage in soy oil, and urine and feces were collected at 24-hour intervals, up to 96 hours after
34 administration (Meyer, 1977). The phenolic metabolites of biphenyl were analyzed as
35 trimethylsilyl (TMS) ethers by combined gas chromatography/mass spectrometry (GC/MS)
36 (guinea pigs) or GC (rabbits). The biphenyl was hydroxylated to monohydroxylated biphenyls
37 and minor amounts of dihydroxylated derivatives, with the main route of excretion being through
38 the urine in both species and the major metabolite being 4-hydroxybiphenyl. In guinea pigs
6 DRAFT - DO NOT CITE OR QUOTE
-------
1 (n = 3), the mass of identified metabolites in urine collected for 24 or 96 hours accounted for
2 29.5 or 32.9% of the administered dose, respectively. In the first 24 hours, biphenyl and
3 biphenyl metabolites in feces accounted for 20.3% of the dose; most of this (14.3%) was
4 biphenyl, presumably unabsorbed. Bile was collected for 24 hours from another group of two
5 bile-cannulated guinea pigs dosed with 100 mg/kg biphenyl. No unchanged biphenyl was
6 detected in the collected bile, but conjugated mono- and dihydroxy metabolites accounted for
7 about 3% of the administered dose. The results with guinea pigs indicate that at least 33% of the
8 administered dose was absorbed. In rabbits, urinary metabolites accounted for 49.1% of the
9 dose, with most of this (25.4% on the first day and 15.9% on the second day) eliminated as
10 conjugates. In the first 24 hours, biphenyl and metabolites in feces accounted for 1.6% of the
11 dose with 1.4% being biphenyl. These results indicate that at least 49% of the administered dose
12 was absorbed in rabbits.
13 Absorption of single oral 100 mg/kg doses of biphenyl (in soy oil or propylene glycol)
14 has also been demonstrated in male and female Danish Landrace pigs weighing 31-35 kg (Meyer
15 et al., 1976b). Metabolites identified in urine collected at four 24-hour intervals after dose
16 administration included mono-, di-, and trihydroxybiphenyls, detected as TMS ethers by GC/MS
17 after enzyme hydrolysis of the samples by p-glucuronidase and sulphatase. Metabolites
18 identified and quantified in 24-hour urine samples accounted for averages of 17.5 and 26.5% of
19 the dose administered in soy oil to two female pigs and in propylene glycol to two male pigs,
20 respectively. Unchanged biphenyl was not detected in the urine samples. Metabolites in urine
21 collected for 96 hours accounted for averages of 27.6 and 44.8% of the doses administered to
22 female and male pigs, respectively. No phenolic metabolites of biphenyl were detected in feces
23 collected for 96 hours. Unchanged biphenyl was not detected in the feces collected from male
24 pigs, but the amount of unchanged biphenyl in feces from the two female pigs accounted for
25 18.4 and 5% of the administered dose. These results indicate that at least about 28 and 45% of
26 oral 100 mg/kg doses of biphenyl were absorbed in female and male pigs, respectively. It is
27 uncertain if the gender difference was due to vehicle differences or actual gender differences in
28 absorption efficiency.
29 No animal studies were located examining quantitative aspects of absorption of biphenyl
30 by the respiratory tract or skin.
31
32 3.2. DISTRIBUTION
33 No information was located regarding distribution of absorbed biphenyl in humans and
34 limited animal data are available. Meyer et al. (1976a) orally administered 100 mg/kg
35 [14C]-biphenyl to male albino rats and measured radioactivity in the lung, heart, kidney, brain,
36 spleen, liver, skeletal muscles, peritoneal fat, genital tract, and gastrointestinal tract at 96 hours
37 after dosing. Most of the radioactivity was excreted in urine (84.8%) and feces (7.3%) over the
38 96-hour period, and only 0.6% of the administered radioactivity remained in the animals at
7 DRAFT - DO NOT CITE OR QUOTE
-------
1 96 hours: 0.1% was found in peritoneal fat, 0.3% in the gastrointestinal tract (including its
2 contents) 0.1% in skeletal muscles, and 0.1% in the genital tract. Levels of radioactivity in other
3 examined tissues were very low. The results indicate that absorbed biphenyl is not preferentially
4 stored in tissues and is rapidly excreted, principally through the urine.
5
6 3.3. METABOLISM
7 3.3.1. Identification of Metabolites
8 3.3.1.1. Results from in vivo Animal Studies
9 No human studies have been identified on the in vivo metabolism of biphenyl. However,
10 the in vivo metabolism of biphenyl has been studied extensively in laboratory animals. These
11 studies have determined that in rats, rabbits, pigs, dogs, mice, and guinea pigs, biphenyl is
12 converted into a range of hydroxylated metabolites (Halpaap-Wood et al., 1981a; Meyer, 1977;
13 Meyer and Scheline, 1976; Meyer et al., 1976a, b). These metabolites have been detected in
14 urine both as nonconjugated compounds and as acidic conjugates.
15 The derivation of urinary metabolites and their subsequent analysis with GC has resulted
16 in the identification of more than 10 mono-, di-, and trihydroxybiphenyl metabolites from the
17 urine of rats, pigs, guinea pigs, and rabbits (Meyer, 1977; Meyer and Scheline, 1976; Meyer et
18 al., 1976a, b). These metabolites have been found as mercapturic acid conjugates and
19 glucuronide conjugates (Millburn et al., 1967). Comparable metabolites have been identified
20 among mammalian species tested, although quantitative differences in metabolite formation are
21 evident among species. A major metabolite in the rat, mouse, guinea pig, rabbit, and pig was
22 reportedly 4-hydroxybiphenyl (Halpaap-Wood et al. 198la; Meyer, 1977; Meyer and Scheline,
23 1976). 4,4'-Dihydroxybiphenyl was identified as a major metabolite in the pig (Meyer et al.,
24 1976b) and the rat (Halpaap-Wood et al., 198la; Meyer and Scheline, 1976), while 3,4-di-
25 hydroxybiphenyl was a major urinary metabolite in two strains of mice (Halpaap-Wood et al.,
26 1981a). Table 3-1 reviews the metabolites that have been identified in the excreta and bile of
27 male albino rats given single doses of 100 mg biphenyl/kg, as reported by Meyer and Scheline
28 (1976).
29
DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-1. Metabolites of biphenyl identified in urine, feces, and bile of male
albino rats
Metabolite"
Biphenyl
2-Hydroxybiphenyl
3 -Hy droxybipheny 1
4-Hydroxybiphenyl
3 ,4-Dihydroxybiphenyl
3,4' -Dihy droxybipheny 1
4,4 ' -Dihydroxybiphenyl
2,5-Dihydroxybiphenyl
Methoxy-hydroxybiphenyls
Methoxy-dihydroxybiphenyls
3 ,4,4 ' -Trihydroxybiphenyl
Total
Urine
Dayl
0.1
0.4
0.9
6.8
0.6
1.5
9.6
Trace
0.1
0.5
1.8
22.3
Day 2
0.1
0.5
0.4
0.7
0.2
0.3
1.7
ND
ND
0.3
0.9
5.1
Days 3 + 4
NDb
0.1
0.3
0.2
ND
0.8
0.1
ND
ND
0.1
0.5
2.1
Days 1-4
0.2
1.0
1.6
7.7
0.8
2.6
11.4
Trace
0.1
0.9
3.2
29.5
Feces
Dayl
ND
0.3
0.5
1.0
ND
ND
1.8
ND
ND
ND
1.1
4.7
Bile
Dayl
ND
0.1
0.5
1.5
0.1
0.3
1.9
ND
0.1
ND
0.7
5.2
aValues are percent of administered dose.
bND = not detected.
Source: Meyer and Scheline (1976).
1
2 The hydroxylation of biphenyl to produce 2-hydroxybiphenyl is a minor pathway in rats
3 and mice, but is more easily detected in mice than rats (Halpaap-Wood et al., 1981a, b).
4 Following intraperitoneal (i.p.) injection of [14C]-labeled biphenyl (30 mg/kg), the pattern of
5 percentages of radioactivity detected in urinary metabolites showed a relatively greater ability to
6 produce 2-hydroxybiphenyl in mice than rats. In Sprague-Dawley rats, metabolites identified in
7 order of abundance were (with percentage of total urinary radioactivity noted in parentheses):
8 4,4'-dihydroxybiphenyl (44.5%); 4-hydroxybiphenyl (28.5%); 3,4,4'-trihydroxybiphenyl (8.8%);
9 3,4'-dihydroxybiphenyl (8.5%); 3,4-dihydroxybiphenyl (5.1%); 3-hydroxybiphenyl (1.8%); and
10 2-hydroxybiphenyl (1.5%). In DBA/2Tex mice, major identified metabolites were: 4-hydroxy-
11 biphenyl (39.5%); 3,4-dihydroxybiphenyl (30.3%); 4,4'-dihydroxybiphenyl (10.2%);
12 3,4,4'-trihydroxybiphenyl (6.2%); 3-hydroxybiphenyl (4.3%); and 2-hydroxybiphenyl (4.2%).
13 In rats, 2,3-, 2,4-, and 2,5-dihyroxybiphenyl were detected at trace levels (<0.1%), whereas, in
14 mice, these metabolites were detected at levels of 0.3%, 0.8%, and 0.7%, respectively (Halpaap-
15 Woodetal., 1981a).
16 No in vivo studies have been identified that directly investigate differential metabolism of
17 biphenyl between males and females of any species. However, studies on urinary crystals and
18 calculi formation and composition after chronic exposure to biphenyl in the diet indicate that
19 male F344 rats are more susceptible than females to the formation of urinary bladder calculi
DRAFT - DO NOT CITE OR QUOTE
-------
1 (Ohnishi et al., 2001, 2000a, b). Urinary bladder calculi in males were predominantly composed
2 of the insoluble potassium salt of 4-hydroxybiphenyl-O-sulphate, whereas the less frequently
3 occurring urinary bladder calculi in females were composed mainly of 4-hydroxybiphenyl and
4 potassium sulphate, hydrolysis products of 4-hydroxybiphenyl-O-sulphate (Ohnishi et al., 2001,
5 2000a, b). These observations are consistent with observations that male rats have relatively
6 higher urinary potassium concentrations and pH values than female rats, and with the hypothesis
7 that gender differences in these urinary conditions (rather than gender differences in metabolism
8 of biphenyl) may be responsible for the gender differences in urinary calculi formation and the
9 subsequent development of non-neoplastic (hyperplasia) and neoplastic (papillomas and
10 carcinomas) lesions in male, but not female, F344 rats (Umeda et al., 2002; Ohnishi et al., 2001,
11 2000a, b).
12
13 3.3.1.2. Results from in vitro Studies with Animal and Human Cells or Tissues
14 The metabolism of biphenyl in vitro has been investigated using tissues of human origin,
15 resulting in evidence that the human metabolism of biphenyl is qualitatively similar to, but may
16 be quantitatively different from, rat metabolism. Benford et al. (1981) measured 2-, 3-, and
17 4-hydroxylation of biphenyl in microsomes prepared from the livers of five rats (sex not
18 identified) and four humans (sex not identified). The reaction products, after solvent extraction
19 and high-performance liquid chromatography (HPLC) quantitation, revealed that 2-hydroxylase
20 in the rat was 35 times higher than in humans, while 3- and 4-hydroxylases in humans were
21 1.5 and 1.2 times higher than in rats.
22 The evidence from studies of human tissue samples exposed to biphenyl metabolites in
23 vitro suggests differential Phase II metabolism contingent upon tissue origin. Powis et al. (1988)
24 have shown that/>-hydroxybiphenyl is conjugated with glucuronic acid and sulphate in human
25 liver and kidney tissue slices. In the liver, glucuronidation was the favored conjugation pathway,
26 while sulphation was favored in the kidney. Powis et al. (1989) also compared Phase I biphenyl
27 metabolism in human (from surgery), dog (mongrel), and rat (male F344) liver slices and
28 primary hepatocytes. It was found that liver slices from all three species had a similar capacity
29 to metabolize biphenyl, -3.5 nmol biphenyl/minute per g tissue, while hepatocyte preparations
30 from rats had about 4 times the metabolic capacity of dog hepatocytes and about 20 times that of
31 human hepatocytes. Powis et al. (1989) speculated that hepatocytes from dog and human liver
32 slices may have experienced more damage during isolation than rat hepatocytes.
33 A study of the sulphation of biphenyl metabolites in human surgical tissue samples was
34 conducted by Pacific! et al. (1991). Tissue samples of various types (liver, intestinal mucosa,
35 lung, kidney, bladder, and brain) were obtained from surgeries of patients of both sexes between
36 the ages of 49 and 76 years of age (each patient contributed only one tissue type, so that within-
37 patient organ comparisons were not made) and prepared 12,000 and 105,000 g supernatants to
38 study sulphation of biphenyl metabolites, specifically 2-, 3-, and 4-hydroxybiphenyl.
10 DRAFT - DO NOT CITE OR QUOTE
-------
1 Sulphotransferase activity for each of these substrates was detected in all tissues studied,
2 although marked tissue dependence was observed, with the highest activity found in the liver and
3 the lowest in the brain. The Michaelis constant (Km) of Sulphotransferase was dependent on the
4 substrate, but not on tissue type, with Km varying over a 500-fold range. The highest values of
5 Km were found with 4-hydroxybiphenyl and the lowest were found with 3-hydroxybiphenyl.
6 Several studies of biphenyl metabolism with in vitro animal systems support the findings
7 from the in vivo urinary metabolite investigations that: (1) a range of hydroxylated biphenyl
8 metabolites are formed, (2) 4-hydroxybiphenyl is a major metabolite, and (3) hydroxylated
9 biphenyl metabolites are conjugated to glucuronic acid or sulphate. Wiebkin et al. (1984, 1976)
10 reported that isolated rat and hamster hepatocytes metabolized biphenyl primarily to
11 4-hydroxybiphenyl and also to 4,4'-hydroxybiphenyl, both of which were then conjugated. A
12 small amount of 2-hydroxybiphenyl was produced. When 4-hydroxybiphenyl was incubated
13 with the hepatocytes, it was hydroxylated to 4,4'-dihydroxybiphenyl. Pretreatment of the
14 animals with either 5,6-benzoflavone or phenobarbital had little effect on the conjugate
15 formation rate in the in vitro experiment. Bianco et al. (1979) reported that rat hepatic
16 microsomes metabolize biphenyl to 4-, 2-, and 3-hydroxybiphenyl, which are conjugated to form
17 glucuronides and sulphates. The 4-hydroxybiphenyl isomer was the major metabolite. The
18 formation of 4-hydroxybiphenyl as a major metabolite in the hamster, mouse, and rabbit was
19 confirmed by Billings and McMahon (1978). 2-Hydroxybiphenyl and 3-hydroxybiphenyl were
20 detected in a lower amount in a ratio of 2:1 by hamster and rabbit microsomes, and in a 1:1 ratio
21 by mouse microsomes. In contrast, almost all hydroxylation of biphenyl in rat microsomes gave
22 rise to 4-hydroxybiphenyl.
23
24 3.3.2. Metabolic Pathways
25 3.3.2.1. Description of Metabolic Scheme and Enzymes Involved
26 Burke and Bridges (1975) suggested that biphenyl metabolism is mediated by
27 cytochrome P-450 (CYP) monooxygenases. Evidence of an arene oxide intermediate, which
28 may participate in binding to cellular macromolecules, was reported by Billings and McMahon
29 (1978). Support for CYP metabolism of biphenyl was provided by Halpaap-Wood et al.
30 (1981a, b), who reported that greater amounts of hydroxybiphenyls were obtained in in vitro
31 assays using liver homogenates when rats were treated first with p-naphthoflavone, 3-methyl-
32 cholanthrene or Aroclor 1254, which are known CYP inducers. In C57BL/6Tex mice, CYP
33 induction with P-naphthoflavone led to relatively greater amounts of urinary excretion of
34 2-hydroxybiphenyl, compared with uninduced mice, whereas pretreatment with p-naphtho-
35 flavone led to increases in urinary excretion of 2-, 3-, and 4-hydroxybiphenyl in Sprague-Dawley
36 rats and was without influence on the pattern of hydroxybiphenyl metabolites in DBA/2Tex mice
37 (Halpaap-Wood et al., 1981a).
11 DRAFT - DO NOT CITE OR QUOTE
-------
1 Figure 3-1 details combined evidence from the Halpaap-Wood et al. (1981a, b) and
2 Meyer and Scheline (1976) studies on the metabolic pathways of biphenyl. While sulphates and
3 glucuronides are formed on all three metabolic levels illustrated, only monosulphates and
4 monoglucuronides are identified. Monomethyl ethers are formed from dihydroxy and trihydroxy
5 metabolites alone. Glucuronides at the dihydroxy and trihydroxy levels are additionally labeled
6 with a question mark to suggest that, while these metabolites are likely, they have not been
7 identified.
12 DRAFT - DO NOT CITE OR QUOTE
-------
4-Hydroxybiphenyl
2-Hydroxybiphenyl
3-Hydroxybiphenyl
HO.
GYP
.OH
OH
ar.ar'-Dihydroxybiphenyl
ar.ar-Dihydroxybiphenyl
GYP
OH
HO.
OH
ar.ar.ar'-Trihydroxybiphenyl
1
2
o
6
4
5
6
7
Sulphono-
transferase Monohydroxysulfate
*" -glucuronide
UGT
Dihydroxy-monosulfate
-monoglycuronide (?)
-monomethyl ether
COMT Trihydroxy-monosulfate
*- -monoglucuronide (?)
-monomethyl ether
ar = aryl group; COMT = catechol-O-methyltransferase; UGT = undine
diphosphate glucuronosyl transferase; question marks denote tentative metabolites
(see text).
Sources: Halpaap-Wood et al. (1981a, b); Meyer and Scheline (1976).
9
10
Figure 3-1. Schematic presentation of the metabolic pathways of biphenyl.
13
DRAFT - DO NOT CITE OR QUOTE
-------
1 The metabolic scheme in Figure 3-1 does not include the possible redox cycling of
2 2,5-dihydroxybiphenyl (also known as phenylhydroquinone), which involves CYP-mediated
3 cycling between phenylhydroquinone and phenylbenzoquinone leading to the generation of
4 reactive oxygen species (ROS) (Balakrishnan et al. 2002; Kwok et al., 1999). This pathway is
5 thought to play a role in the carcinogenic effect of 2-hydroxybiphenyl (also known as
6 ort/zo-phenylphenol), a broad spectrum fungicide that, like biphenyl, induces urinary bladder
7 tumors in chronically exposed male rats (Kwok et al., 1999). Free 2,5-dihydroxybiphenyl and its
8 glucuronide or sulphate conjugates are readily detected in the urine of rats exposed to
9 2-hydroxybiphenyl, and the formation of 2,5-dihydoxybiphenyl and phenylbenzoquinone is the
10 principal metabolic pathway for 2-hydroxybiphenyl in the rat, especially at high exposure levels
11 associated with urinary bladder tumor formation (Kwok et al., 1999; Morimoto et al., 1989;
12 Nakao et al., 1983; Reitz et al., 1983; Meyer and Scheline, 1976). In contrast, the formation of
13 4-hydroxybiphenyl and 4,4'-dihydroxybiphenyl is the principal metabolic pathway for biphenyl
14 in rats and mice, and 2,5-dihydroxybiphenyl was not detected, or only detected at trace levels, in
15 the urine of rats exposed to 100 mg biphenyl/kg (Meyer and Scheline, 1976; see Table 3-1). In
16 mice exposed to i.p. doses of [14C]-biphenyl (30 mg/kg), radioactivity in 2-hydroxybiphenyl and
17 2,5-dihydroxybiphenyl in the urine accounted for only about 5% of the total radioactivity
18 detected in urinary metabolites (Halpaap-Wood et al., 1981a).
19
20 3.3.3. Regulation of Metabolism, Sites of Metabolism, and Relationships to Toxic Effects
21 3.3.3.1. Evidence for Induction of Phase I and II Enzymes
22 No studies of Phase I or II enzyme induction using liver microsomes of human origin
23 were identified. However, a number of studies have been conducted in rodents to investigate the
24 induction of Phase I enzymes that catalyze biphenyl hydroxylation. For example, Creaven and
25 Parke (1966) reported that pretreatment of weanling Wistar rats or ICI mice with phenobarbital
26 (an inducer of CYP3 A4, 2B6, and 2C8 as reported by Parkinson and Ogilvie, 2008) or
27 3-methylcholanthrene (an inducer of CYP1A2 as reported by Parkinson and Ogilvie, 2008)
28 increased NADPH-dependent activities of liver microsomes to produce 2-hydroxybiphenyl and
29 4-hydroxybiphenyl from biphenyl to varying degrees depending on the inducer. Haugen (1981)
30 reported that pretreatment of male CD rats with phenobarbital or 3-methylcholanthrene increased
31 NADPH-dependent activities of liver microsomes to produce 2-, 3-, and 4-hydroxybiphenyl from
32 biphenyl, again to varying degrees depending on the inducer. Stuehmeier et al. (1982) reported
33 that phenobarbital pretreatment of male C57BL/6JHan mice induced liver microsomal activities
34 to produce 4-hydroxybiphenyl, but not 2-hydroxybiphenyl, from biphenyl, whereas
35 3-methylcholanthrene induced activities for both 4- and 2-hydroxylation of biphenyl. Halpaap-
36 Wood et al. (198 la) reported that pretreatment of male Sprague-Dawley rats with
37 p-naphthoflavone (an inducer of CYP1A2 as reported by Parkinson and Ogilvie, 2008; also
38 known as 5,6-benzoflavone) enhanced the urinary excretion of 2-, 3-, and 4-hydroxybiphenyl,
14 DRAFT - DO NOT CITE OR QUOTE
-------
1 3,4-dihydroxybiphenyl, and 3,4,4'-trihydroxybiphenyl following i.p. administration of 30 mg
2 biphenyl/kg body weight. In contrast, pretreatment of male C57BL/6Tex mice with
3 p-naphthoflavone did not increase the overall urinary excretion of biphenyl metabolites
4 following i.p. administration of 60 mg biphenyl/kg, but shifted the principal metabolite from
5 4-hydroxybiphenyl to 2-hydroxybiphenyl and 2,5-dihydroxybiphenyl (Halpaap-Wood et al.,
6 1981a). Wiebkin et al. (1984) reported that P-naphthoflavone pretreatment of male Lewis rats or
7 male Syrian golden hamsters induced biphenyl hydroxylation activities in freshly isolated
8 pancreatic acinar cells or hepatocytes. From these observations and examination of patterns of
9 inhibition of biphenyl hydroxylation activities by CYP inhibitors (e.g., a-naphthoflavone and
10 1-benzyl-imidazole) under non-induced and induced conditions (see Haugen, 1981), it is
11 apparent that multiple CYP enzymes (e.g., CYP1A2 and CYP3A4) are likely involved in
12 biphenyl hydroxylation. However, no studies were located that used more modern techniques
13 (such as CYP knockout mice) to identify the principal CYP enzymes involved in the initial
14 hydroxylation of biphenyl or the formation of the dihydroxy- or trihydroxybiphenyl metabolites.
15 Several animal studies were located examining the possible coordinated induction of
16 Phase I enzymes with Phase II enzymes catalyzing the conjugation of hydroxylated biphenyl
17 metabolites to sulphate or glucuronic acid. Hepatocytes from rats (strain and sex were not noted)
18 pretreated with the CYP inducers, phenobarbital or 3-methylcholanthrene, produced glucuronide
19 and sulphate conjugates of 4-hydroxybiphenyl when incubated with biphenyl (Wiebkin et al.,
20 1978). Glucuronide conjugates were predominant under these "CYP-induced" conditions,
21 whereas hepatocytes from non-induced control rats produced predominant sulphate conjugates of
22 4-hydroxybiphenyl. These results suggest that induction (or possibly activation) of
23 glucuronidation enzymes may be coordinated with the induction of CYP enzymes. In contrast,
24 pretreatment of male Lewis rats with P-naphthoflavone (an inducer of CYP1A2) did not enhance
25 activities of freshly isolated pancreatic acinar cells to conjugate 4-hydroxybiphenyl with sulphate
26 or glucuronic acid, but the influence of this pretreatment on the conjugation capacity of
27 hepatocytes was not examined in this study (Wiebkin et al., 1984). In another study, uridine
28 diphosphate glucuronosyl transferase (UGT) activities with 1-naphthol or 3-hydroxy-
29 benzo[a]pyrene as substrates were higher in liver microsomes from male Wistar rats pretreated
30 with Aroclor 1254 (an inducer of several CYP enzymes) or phenobarbital, respectively,
31 compared with microsomes from control rats without pretreatment with CYP inducers (Bock et
32 al., 1980). Although Bock et al. (1980) measured UGT activities in microsomes from several
33 tissues from non-induced rats with 4-hydroxybiphenyl as a substrate, no comparisons between
34 induced and non-induced conditions were made using 4-hydroxybiphenyl as substrate. Paterson
35 and Fry (1985) reported that hepatocytes or liver slices from male Wistar rats pretreated with
36 P-naphthoflavone showed decreased rates of glucuronidation of 4-hydroxybiphenyl, compared
37 with hepatocytes or liver slices from rats without P-naphthoflavone pretreatment. Results from
38 this database provide equivocal evidence that the induction of Phase I enzymes catalyzing the
15 DRAFT - DO NOT CITE OR QUOTE
-------
1 hydroxylation of biphenyl may be coordinated with induction of Phase II enzymes catalyzing
2 glucuronidation of hydroxylated biphenyl metabolites.
o
J
4 3.3.3.2. Demonstrated Tissue Sites of Metabolism
5 CYP enzymes catalyzing hydroxylation of biphenyl and other substrates are present in
6 most, if not all, mammalian tissues, but the highest levels of activities are normally found in liver
7 (Parkinson and Ogilvie, 2008). In a study of male Sprague-Dawley rats, cytochrome P450
8 content was 20- to 40-fold higher in the microsomes from liver than from lung, although
9 biphenyl-4-hydrolase activity was only 1.7-fold higher in the microsomes from liver than from
10 lung (Matsubara et al., 1974). Wiebkin et al. (1984) observed 200- and 1,000-fold higher rates
11 of biphenyl metabolism in 5,6-benzoflavone-pretreated hepatocytes compared to similarly
12 treated pancreatic acinar cells from male Lewis rats and Syrian golden hamsters, respectively.
13 Activities for enzymes catalyzing the conjugation of hydroxybiphenyls and other
14 hydroxylated aromatic compounds with glucuronic acid or sulphate have been detected in a
15 number of mammalian tissues, and, similar to CYP, the highest levels are found in the liver
16 (Parkinson and Ogilvie, 2008). Available data for conjugation activities with hydroxybiphenyls
17 in various mammalian tissues are consistent with this concept. Sulphotransferase activities with
18 2-, 3-, or 4-hydroxybiphenyl as substrates in microsomes from several human tissues showed an
19 approximate 100- to 500-fold range with the following order: liver > ileum > lung > colon >
20 kidney > bladder > brain (Pacific! et al., 1991). UGT activities with 4-hydroxybiphenyl as
21 substrate in microsomes from several male Wistar rat tissues showed the following order: liver >
22 intestine > kidney > testes ~ lung; activities were below the limit of detection in microsomes
23 from skin and spleen (Bock et al., 1980).
24
25 3.3.3.3. Possible Relationships Between Metabolites and Toxic Effects
26 Increased formation of urinary tract crystals and calculi in F344 rats chronically exposure
27 to biphenyl in the diet has been well documented. This phenomenon occurs predominantly in
28 males and can ultimately lead to non-neoplastic and neoplastic changes in the urinary bladder
29 (Umeda et al., 2002). Ohnishi et al. (2001, 2000a, b, 1998) have proposed mechanistic roles for
30 the potassium salt of the 4-hydroxybiphenyl sulphate conjugate, high urine potassium
31 concentrations, and relatively high urine pH in producing urinary calculi, which are found in
32 86% of male F344 rats and only 16% of female rats exposed to high biphenyl concentrations in
33 the diet (4,500 ppm) for 2 years (Umeda et al., 2002). Gender differences in calculi composition
34 were also observed, with calculi in male F344 rats being mainly composed of potassium 4-
35 hydroxybiphenyl-O-sulphate and calculi in female rats composed mainly of 4-hydroxybiphenyl
36 and potassium sulphate, presumably produced by the hydrolysis of 4-hydroxybiphenyl-O-
37 sulphate in the urine (Ohnishi et al., 2000a, b). As discussed earlier, these observations are
38 consistent with the hypothesis that gender differences in urinary conditions (higher urine
16 DRAFT - DO NOT CITE OR QUOTE
-------
1 potassium concentrations and pH) may be responsible for the gender differences in urinary
2 calculi formation and the subsequent development of nonneoplastic and neoplastic lesions in
3 male, but not female, F344 rats (Umeda et al., 2002; Ohnishi et al., 2001, 2000a, b).
4
5 3.4. ELIMINATION
6 No studies were located on the route or rate of elimination of biphenyl in humans, but
7 results from studies of orally exposed animals indicate that absorbed biphenyl is rapidly
8 eliminated from the body, principally as conjugated hydroxylated metabolites in the urine.
9 The most quantitative data on the routes and rates of elimination come from a study of
10 rats following administration of radiolabeled biphenyl (Meyer et al., 1976a). Urine collected for
11 24 hours after the oral administration of 100 mg/kg [14C]-labeled biphenyl in soy oil to male
12 albino rats contained 75.8% of the administered radioactivity, compared with 5.8% detected in
13 feces collected in the same period. Ninety-six hours after dose administration, <1% of the
14 administered radioactivity remained in tissues, 84.8% was in collected urine, 7.3% was in feces,
15 and 0.1% was in collected expired air (Meyer et al., 1976a). Although chemical identity analysis
16 of fecal radioactivity was not conducted by Meyer et al. (1976a), results from GC/MS analyses
17 of bile collected from bile-cannulated rats given single 100 mg/kg doses of unlabeled biphenyl
18 indicate that biliary excretion of metabolites represents a minor pathway of elimination (Meyer
19 and Scheline, 1976). In bile collected for 24 hours, unchanged biphenyl was not detected and
20 conjugated metabolites accounted for 5.2% of the administered dose; in contrast, conjugated
21 metabolites of biphenyl in 24-hour urine accounted for 22.3% of the dose (Meyer and Scheline,
22 1976).
23 Supporting evidence for the importance of urinary elimination of conjugated metabolites
24 is provided by the results of other studies, which analyzed biphenyl and biphenyl metabolites by
25 GC/MS or GC in urine and feces collected from rabbits (Meyer, 1977), guinea pigs (Meyer,
26 1977), and pigs (Meyer et al., 1976b) following oral administration of 100 mg/kg doses of
27 unlabeled biphenyl. In 24-hour urine samples, unchanged biphenyl was not detected, and total
28 metabolites accounted for averages of 25.4% of the administered dose in rabbits, 31.3% in
29 guinea pigs, 17.5% in female pigs, and 26.4% in male pigs. As in rats, biliary excretion
30 represents a minor elimination pathway in guinea pigs and rabbits; metabolites detected in bile
31 'collected for 24 hours from bile-cannulated guinea pigs accounted for 3.3% of the administered
32 dose, but for only 0.3% of the dose in bile collected for 7 hours from a rabbit given 100 mg/kg
33 biphenyl (Meyer, 1977). Neither unchanged biphenyl nor hydroxylated biphenyl metabolites
34 were detected in bile collected from a bile-cannulated pig for 24 hours after administration of
35 100 mg/kg biphenyl (Meyer et al., 1976b).
36 No studies were located examining quantitative aspects of elimination in animals
37 following inhalation or dermal exposure to biphenyl.
38
17 DRAFT - DO NOT CITE OR QUOTE
-------
1 3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS
2 No studies were located on the development of PBPK models for biphenyl in animals or
3 humans.
4
18 DRAFT - DO NOT CITE OR QUOTE
-------
1 4. HAZARD IDENTIFICATION
2
3
4 4.1. STUDIES IN HUMANS
5 Limited human data include assessments of workers exposed to biphenyl during
6 production of biphenyl-impregnated fruit wrapping paper at one mill in Finland (Seppalainen and
7 Hakkinen, 1975; Hakkinen et al., 1973, 1971) and another mill in Sweden (Wastensson et al.,
8 2006) and a single case report of reversible hepatotoxic effects attributed to biphenyl exposure
9 (Carella and Bettolo, 1994).
10
11 Seppalainen and Hakkinen, 1975; Hakkinen et al, 1973, 1971
12 Hakkinen and colleagues assessed the health of paper mill workers exposed to biphenyl
13 during the production of biphenyl-impregnated paper used to wrap citrus fruits. In 1959,
14 workers complained about a strong odor and irritation to the throat and eyes. Air measurements
15 made at various locations within the facility in June of 1959 resulted in estimated average
16 biphenyl concentrations of 4.4-128 mg/m3 (Table 4-1). In 1969, a 32-year-old worker at the
17 facility, who had worked for 11 years in the oil room where biphenyl levels were particularly
18 high, became ill. Despite aggressive medical intervention, the patient grew worse and died. Key
19 features at autopsy included necrosis of most liver cells, severe, but unspecified changes in the
20 kidneys, degeneration of the heart muscles, hyperactive bone marrow, and edematous changes in
21 the brain (Hakkinen et al., 1973, 1971). Subsequent measurements of biphenyl in the workplace
22 air (January 1970) resulted in estimated average concentrations ranging from 0.6 to 123 mg/m3
23 (Table 4-1). Measurements taken in both 1959 and 1971 indicated that biphenyl air
24 concentrations at multiple work areas greatly exceeded the current ACGIH (2008) threshold limit
25 value (TLV) of 0.2 ppm (1.3 mg/m3). In the location where biphenyl was mixed with paraffin oil
26 (the oil room), biphenyl occurred both as a vapor and as a dust, suggesting the possibility of both
27 dermal and inhalation exposures.
28
19 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-1. Biphenyl concentrations in the air of a Finnish paper mill
producing biphenyl-impregnated fruit wrapping paper
Sampling center locations
Paper mill hall
In front of paper reel
Behind impregnating roller
Near paper machine
Near rolling machine
Oil-room
Near measuring container
Above measuring container (lid open)
Near mixing container
During addition of biphenyl to mixing container
Average concentrations (mg/m3)
June 1959
17.9
128.0
7.2
4.4
19.5
No data
No data
No data
January 1970
7.2
64.0
1.5
0.6
3.5
123.0
15.5
74.5
Source: Hakkinen etal. (1973).
1
2 Thirty-one male workers at the Finnish facility were engaged in the biphenyl-
3 impregnation process; two other workers (one male stock keeper and one female paper cutter)
4 were thought to have been exposed to biphenyl and were therefore included in the study.
5 Common complaints among these workers included fatigue, headache, gastrointestinal
6 discomfort, numbness and aching of the limbs, and general fatigue; laboratory tests revealed
7 elevated serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (which
8 can indicate inflammation or damage to liver cells) in 10 of the 33 workers (Hakkinen et al.,
9 1973). Eight of the 33 workers were admitted to the hospital for further examination, including
10 liver biopsy. The majority of the 33 workers were subjected to neurophysiological examinations,
11 including electroencephalography (EEG) and electroneuromyography (ENMG, consisting of
12 nerve conduction velocity and electromyographic [EMG] tests). Seppalainen and Hakkinen
13 (1975) published the most comprehensive results of the neurophysiological examinations. In all,
14 24 subjects (including the 8 hospitalized workers) underwent neurophysiological examinations.
15 Exposure to biphenyl was terminated immediately following the initial neurophysiological
16 examinations, and 11 and 7 of these subjects were retested 1 and 2 years later, respectively.
17 EEG results. At initial examination, 10 of the 24 workers had abnormal EEGs, which
18 included diffuse slow wave abnormalities (6 cases), lateral spike and slow wave discharges
19 (2 cases), posterior slowing only (1 case), and mild slow wave abnormality in the right temporal
20 area (1 case). Six subjects exhibited unusual distribution of alpha rhythm with alpha activity also
21 prominent in the frontal areas. Four of the subjects exhibited no EEG abnormalities. In general,
22 the EEG results observed at initial examination were qualitatively similar in the 11 subjects
23 reexamined 1 year later. Exceptions included additional diffuse slow wave abnormalities in the
24 two subjects initially exhibiting only spike and wave discharges and the disappearance of the one
20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
case of mild temporal local abnormality. There was no discernable improvement in the EEGs of
the seven subjects reexamined after 2 years.
ENMG results. As shown in Table 4-2, the 24 biphenyl-exposed workers exhibited no
significant differences in mean maximal motor conduction velocity (MCV) relative to those of a
control group consisting of 60 healthy Finnish males, but significantly (p < 0.001) slower mean
conduction velocity of the slowest motor fibers (CVSF) of the ulnar nerves. Results at the 1-year
follow up of 11 of the biphenyl-exposed workers revealed no significant changes in initial
conduction velocity measures, but at the 2-year reexamination of 7 of the 11 subjects, the MCVs
of the median and deep peroneal nerves were significantly slower (p < 0.02 and/? < 0.01,
respectively) compared to the initial measurements. Abnormal EMGs among the biphenyl-
exposed workers included diminished numbers of motor units on maximal muscle contraction
(10 subjects) and fibrillations in some muscles (7 subjects). Workers exhibiting abnormal EMGs
typically displayed slowing of some nerve conduction velocities as well. Of those 11 subjects
undergoing repeat ENMG examination after 1 year, 5 subjects showed an increased level of
ENMG abnormality, while 4 remained unchanged and 2 had diminished abnormalities. At the
end of 2 years, three of seven subjects displayed diminished ENMG abnormalities, three of seven
were unchanged, and one of seven had the abnormality increased.
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl:
comparison with 60 unexposed males
Nerve
Median
MCV
Ulnar
MCV
CVSF
Deep peroneal
MCV
CVSF
Posterior tibial
MCV
Biphenyl group
(mean ± SD)
57.7 ±6.3
56.3 ±4.6
41.4 ±5.2
50.2 ±5.4
37.7 ±3. 9
43.4 ±3. 9
Control group
(mean ± SD)
58.0 ±3.8
56.6 ±4.0
45.5 ±3.2
50.3 ±3.5
38.2 ±5.6
42.4 ±4.7
f-test
Not significant
Not significant
p< 0.001
Not significant
Not significant
Not significant
19
20
21
22
23
SD = standard deviation
Source: Seppalainen and Hakkinen (1975).
Seppalainen and Hakkinen (1975) noted that subjects often exhibited signs of dysfunction
in both the peripheral nervous system, as evidenced by abnormal ENMGs, and the central
nervous system, as evidenced by abnormal EEGs and abnormal distribution of alpha activity.
Only five subjects (four men and the only woman in the biphenyl-exposed group) were found to
21
DRAFT - DO NOT CITE OR QUOTE
-------
1 have completely normal neurophysiological records. The authors interpreted their data to
2 indicate that biphenyl can attack the nervous system at different levels, the sites of greatest
3 vulnerability being the brain and peripheral nerves. Compound-related anomalies in nerve
4 conduction, EEG, and ENMG signals, while small, were consistent with the persistence of
5 incapacity and the incidence of subjective symptoms.
6
7 Carella andBettolo, 1994
8 Carella and Bettolo (1994) published a case report of a 46-year-old female who had
9 suffered from periodic asthenia while working over a 25-year period at a fruit-packing facility
10 where biphenyl-impregnated paper was used. The patient presented with hepatomegaly,
11 neutrophilic leukocytosis, and clinical chemistry findings indicative of hepatic perturbation. For
12 example, the activities of liver-specific enzymes in serum were 62 mU/mL for AST, 90 mU/mL
13 for ALT, 320 mU/mL for alkaline phosphatase (AP), and 970 IU/L for gamma glutamyl
14 transferase. Examination of a liver biopsy taken from the subject showed a polymorphic
15 inflammatory infiltrate with eosinophils in the portal and lobular regions. These findings are
16 indicative of chronic hepatitis.
17 Following cessation of work in citrus packing, the patient's asthenia gradually
18 disappeared and the serum enzyme abnormalities returned to normal. This permitted the
19 speculation that, in the absence of any other obvious causes of the liver abnormality,
20 occupational exposure to biphenyl may have been the principal etiological factor. It is possible
21 that, for this patient, exposure was via all of the major exposure pathways, inhalation, oral, and
22 dermal, with the latter route predominating.
23
24 Wastensson etal, 2006
25 At a facility manufacturing biphenyl-impregnated paper in Sweden, a cluster of five cases
26 of Parkinson's disease (PD) among the employees was investigated. Since, according to the
27 national average, only 0.9 cases would be expected from the 255 employees at the facility
28 (relative risk [RR] 5.6 [95% confidence interval 1.9-13]), it was suspected that the elevated PD
29 at the facility may have been related to biphenyl exposure. Four of the subjects worked in the
30 vicinity of a rewinder/dryer, while the fifth attended to another rewinder. Although no ambient
31 biphenyl levels were available for the subjects' work space, it was thought likely that the level of
32 biphenyl in air would be greater than the existing TLV of 1.3 mg/m3 (0.2 ppm) based on
33 measurements at a Finnish paper mill with similar production practices (Hakkinen et al., 1973).
34 Two subjects may have been exposed to higher levels of biphenyl than the others when they
35 created the paraffin oil/biphenyl mixture.
36 In addition to comparing existing PD cases to national trends, Wastensson et al. (2006)
37 examined the medical records of 222 former employees who had died. Nine cases of PD were
38 found among the decedents, compared with 4.3 cases of PD expected from data on the lifetime
22 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
risk of developing PD in the general population. This comparison yielded an RR of 2.1, with a
95% confidence interval of 0.96-4.0.
Clinical features and exposure data for the five living subjects, all of whom were
diagnosed with PD by a neurologist at a local hospital, are summarized in Table 4-3. With one
exception, the patients were in comparatively good health on initial diagnosis. The exception
was a 53-year-old male who had diabetes mellitus and withdrew from the study before his
neurological condition could be confirmed. Assuming that the diagnoses of PD were valid, the
central issue is whether these data indicate a causal relationship between PD and exposure to
biphenyl. Wastensson et al. (2006) discussed this issue in the context of other studies that have
pointed to a possible cause-and-effect relationship between pesticide exposure and PD, but were
unable to draw any firm conclusions from their limited data.
Table 4-3. Exposure data and clinical features for five Parkinson's Disease
patients with occupational exposure to biphenyl
Case
1
2
3
4
5
Exposure data
Age
Workplace
Years of exposure3
Age at onset of exposure
Age at onset of symptoms
63
PM3
12
19
52
63
PM3
4
26
55
58
PM4
9
17
44
54
PM3
4
18
51
63
PM3
2
21
55
Clinical features
Resting tremor
Cogwheel rigidity
Brady kinesia
Positive response to levodopab
+
+
+
+
+
+
+
+
+
+
+
+
+
-
+
+
+
+
-
+
13
14
15
16
17
18
19
20
""Exposure to biphenyl about one-third of each year.
bAll five patients improved with levodopa, which is a medication for Parkinson's Disease.
PM = paper mill
Source: Wastensson et al. (2006).
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
Overview. Available oral data for biphenyl include two well-designed chronic toxicity
and carcinogenicity studies, one in F344 rats (Umeda et al., 2002) and one in BDFi mice (Umeda
et al., 2005). Increased incidence of urinary bladder transitional cell papillomas and carcinomas,
associated with the formation of urinary bladder calculi, occurred in male, but not female,
F344 rats at the highest tested dietary concentration, 4,500 ppm, but were not found at lower
23
DRAFT - DO NOT CITE OR QUOTE
-------
1 exposure levels of 1,500 or 500 ppm. Non-neoplastic kidney lesions (simple transitional cell
2 hyperplasia in the renal pelvis and hemosiderin deposits) were found in female F344 rats at
3 biphenyl dietary concentrations >1,500 ppm (Umeda et al., 2002). Several other rat studies
4 provide supporting evidence that the kidney and other urinary tract regions are critical targets for
5 biphenyl in rats (Shiraiwa et al., 1989; Ambrose et al., 1960; Pecchiai and Saffiotti, 1957; Dow
6 Chemical Co., 1953). In BDFi mice, increased incidence of liver tumors (hepatocellular
7 adenomas and carcinomas) and non-neoplastic effects on the kidney (mineralization) and liver
8 (increased activities of plasma ALT and AST) were found in females exposed to biphenyl dietary
9 concentrations of 2,000 or 6,000 ppm (Umeda et al., 2005). In contrast, no carcinogenic
10 responses or noncancer adverse effects were found in female ddY mice exposed to 5,000 ppm
11 biphenyl in the diet for 2 years (Imai et al., 1983) or in B6C3Fi and B6AKFi mice exposed to
12 517 ppm biphenyl in the diet for 18 months (Innes et al., 1969; NCI, 1968).
13 No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
14 toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
15 biphenyl (7 hours/day, 5 days/week for up to about 90 days) were reported in mice exposed to
16 5 mg/m3 and in rats exposed to 300 mg/m3, but not in rabbits exposed to 300 mg/m3 (Deichmann
17 et al., 1947; Monsanto, 1946). Congestion or edema of the lung, kidney, and liver, accompanied
18 by hyperplasia with inflammation of the trachea, was found in CD-I mice exposed to biphenyl
19 vapors at 25 or 50 ppm (158 or 315 mg/m3) for 13 weeks (Sun Company Inc., 1977b).
20 Detailed study descriptions for all available subchronic and chronic toxicity and
21 carcinogenicity studies follow.
22
23 4.2.1. Oral Exposure
24 4.2.1.1. Subchronic Toxicity
25 Dow Chemical Co., 1953
26 Twenty-one-day-old female Long-Evans rats (8/group) were exposed to 0, 0.01, 0.03, or
27 0.1% biphenyl in the diet for 90 days. Body weights were monitored 3 times/week, and the
28 weights of the liver, kidneys, adrenals, and spleen were recorded at necropsy. Sections of heart,
29 liver, kidney, spleen, adrenals, pancreas, ovary, uterus, stomach, small and large intestine,
30 voluntary muscle, lung, thyroid, and pituitary from each rat were preserved in formalin.
31 Hematoxylin and eosin stained sections of the preserved sections from two rats of each group
32 were examined pathologically.
33 Based on U.S. EPA (1988) subchronic reference values for body weight and food
34 consumption in female Long-Evans rats, doses of biphenyl estimated for the dietary levels of
35 0.01, 0.03, and 0.1% are estimated to have been 10, 30, and 100 mg/kg-day, respectively. There
36 were no significant treatment-related effects on body weight, food consumption, or organ
37 weights. Results of histopathologic examinations were unremarkable. Biphenyl-exposed groups
38 exhibited lower average plasma blood urea nitrogen (BUN) levels than controls (28.2, 25.7, and
24 DRAFT - DO NOT CITE OR QUOTE
-------
1 26.3 mg percent for low-, mid-, and high-dose groups, respectively, compared to 35.3 mg percent
2 for controls), although the statistical significance of these apparent treatment-related differences
3 was not reported and the biological significance is uncertain.
4
5 Umeda et al, 2004
6 Six-week-old BDFi mice (10/sex/group) were exposed to biphenyl at dietary
7 concentrations of 0, 500, 2,000, 4,000, 8,000, 10,000, or 16,000 ppm for 13 weeks. To overcome
8 possible problems with taste aversion, mice assigned to the 8,000 and 10,000 ppm groups were
9 fed 4,000 ppm dietary biphenyl for the first week and 8,000 or 10,000 ppm for the remaining
10 12 weeks. Mice designated to receive 16,000 ppm were fed 4,000 ppm dietary biphenyl for the
11 first week, 8,000 ppm for the second week, and 16,000 ppm for the remaining 11 weeks.
12 Animals were checked daily for clinical signs; body weight and food consumption were recorded
13 weekly; organ weights were noted at term; and liver sections were processed for light
14 microscopic examination. Electron microscopy was carried out on liver tissue from one control
15 and one 16,000 ppm female.
16 Based on U.S. EPA (1988) subchronic default reference values for body weight and food
17 consumption (average values for combined sexes), doses of biphenyl for the dietary
18 concentrations of 500, 2,000, 4,000, 8,000, 10,000, and 16,000 ppm are estimated to have been
19 93, 374, 747, 1,495, 1,868, and 2,989 mg/kg-day, respectively. A single 16,000 ppm female
20 mouse died during the study; all other mice survived until terminal sacrifice. Final body weights
21 of mice of both sexes in the 8,000, 10,000, and 16,000 ppm groups were significantly lower than
22 gender-matched controls (for males: 83.3, 84.9, and 75.1% of controls; for females: 93.7, 91.6,
23 and 85.8% of controls, respectively). Umeda et al. (2004) noted that absolute liver weights were
24 significantly higher in 8,000 and 16,000 ppm female mice, but did not include the extent of these
25 increases in the study report. Light microscopic examination of liver specimens from all
26 16,000 ppm female mice revealed enlarged centrilobular hepatocytes, the cytoplasm of which
27 was filled with numerous eosinophilic fine granules. Upon electron microscopic examination,
28 these eosinophilic granules were identified as peroxisomes, indicative of a peroxisome
29 proliferative effect in the liver of the 16,000 ppm female mice. Evidence of histopathologic liver
30 lesions was not found in females of the 8,000 or 10,000 ppm groups. There were no signs of
31 treatment-related increased liver weight or histopathologic evidence of clearly enlarged
32 hepatocytes in any of the biphenyl-treated groups of male mice.
33
34 4.2.1.2. Chronic Toxicity and Carcinogenicity
35 4.2.1.2.1. Chronic rat studies
36 Umeda et al., 2002
37 In a chronic toxicity and carcinogenicity study of F344 rats (50/sex/group), biphenyl was
38 administered in the diet for 2 years at concentrations of 0, 500, 1,500, or 4,500 ppm. All animals
25 DRAFT - DO NOT CITE OR QUOTE
-------
1 were examined daily for clinical signs; body weights and food intake were determined once a
2 week for the first 14 weeks and every 4 weeks thereafter. Urinalysis was performed on all
3 surviving rats during week 105. Upon necropsy, weights of all major organs were recorded; all
4 major organs and tissues were subjected to histopathologic examination.
5 The study report included a plot of mean body weights during the 2-year study, but did
6 not include food consumption data. Estimated doses, therefore, were calculated using time-
7 weighted average (TWA) body weights from the graphically-depicted data (Figure 1 of Umeda
8 et al., 2002) and U.S. EPA (1988) chronic reference values for food consumption in F344 rats.
9 The resulting estimated doses for the 500, 1,500, and 4,500 ppm exposure groups were 36.4, 110,
10 and 378 mg/kg-day, respectively, for males and 42.7, 128, and 438 mg/kg-day, respectively, for
11 females. The study authors reported significantly lower mean body weights among 4,500 ppm
12 rats of both sexes compared to their respective controls. Mean body weights of 4,500 ppm male
13 and female rats were lower than those of controls throughout most of the study period and were
14 approximately 20% lower than respective controls at terminal sacrifice. There was no significant
15 effect on mean body weights of 500 or 1,500 ppm males or females. Survival of low- and mid-
16 dose male and female rats was not significantly different from controls. The study authors
17 reported that 3/50 of the 4,500 ppm female rats died after 13-26 weeks of biphenyl exposure and
18 attributed the deaths to marked mineralization of the kidneys and heart. However, they also
19 indicated that survival of this group was not adversely affected thereafter. Significantly
20 decreased survival was noted only for the group of 4,500 ppm male rats, 19/50 of which died
21 prior to terminal sacrifice. The first death occurred around treatment week 36; this rat exhibited
22 urinary bladder calculi. Survival data for the other groups were not provided. Evidence of
23 hematuria was first noted in 4,500 ppm male rats around week 40 and was observed in a total of
24 32/50 of the 4,500 ppm males during the remainder of the treatment period; 14 of these rats
25 appeared anemic. Hematuria and bladder tumors were primarily considered as causes of death
26 among the 4,500 ppm males (n = 19) that died prior to terminal sacrifice. Urinalysis performed
27 during the final treatment week revealed significantly increased urinary pH in the 31 remaining
28 4,500 ppm male rats (pH of 7.97 vs. 7.66 for controls; p < 0.05); occult blood was noted in the
29 urine of 23 of these males. Urine samples in 10/37 surviving 4,500 ppm females tested positive
30 for occult blood. Significant increases in relative kidney weights of 4,500 ppm males and
31 females and absolute kidney weights of 4,500 ppm males were reported, but actual data were not
32 presented.
33 Gross pathologic examinations at premature death or terminal sacrifice revealed the
34 presence of calculi in the bladder of 43/50 of the 4,500 ppm males and 8/50 of the 4,500 ppm
35 females (Table 4-4); these lesions were not seen in 500 or 1,500 ppm male or female rats. The
36 bladder calculi in the male rats were white, yellow, brown, gray, and black in color, ranged from
37 0.3 to 1.0 cm in size, and exhibited triangular, pyramidal, cuboidal, and spherical shapes. The
38 bladder calculi in the female rats were white and yellow in color, of uniform spheroidal shape,
26 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
and similar in size to those of the male rats. Forty-one of the 4,500 ppm male rats exhibited
polyp-like or papillary nodules protruding into the lumen from the bladder wall; bladder calculi
were noted in 38 of these males. Four of the eight calculi-bearing 4,500 ppm female rats also
exhibited thickening of the bladder wall. It was noted that 30/32 of the 4,500 ppm male rats with
hematuria also exhibited kidney or urinary bladder calculi.
Table 4-4. Incidences of urinary bladder lesions in male and female F344
rats exposed to biphenyl in the diet for 2 years
Dietary concentration (ppm)
Calculated dose (mg/kg-d)
Lesion
Transitional cell
Simple hyperplasia3
Nodular hyperplasia3
Papillary hyperplasia3
Combined
Papilloma
Carcinoma
Papilloma or carcinoma
(combined)
Squamous cell
Metaplasia3
Hyperplasia3
Papilloma or carcinoma
(combined)
Inflammatory polyp3
Calculi
Males (n = 50)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
500
36.4
0
0
0
0
0
0
0
0
0
0
0
0
1,500
110
0
0
0
0
0
0
0
0
0
0
0
0
4,500
378
12b
40b
17b
45
10b
24b
31b
19b
13b
1
10b
43b
Females (n = 50)
0
0
0
1
0
1
0
0
0
0
0
0
0
0
500
42.7
0
0
0
0
0
0
0
0
0
0
0
0
1,500
128
1
0
0
1
0
0
0
0
0
0
0
0
4,500
438
1
5C
4
10b
0
0
0
4
1
0
0
8b
9
10
11
12
13
14
15
aThe number is the sum of animals with severity grades of slight, moderate, marked, or severe.
bSignificantly different from control group (p < 0.01) according to Fisher's exact test.
Significantly different from control group (p < 0.05) according to Fisher's exact test.
Source: Umeda et al. (2002).
Histopathologic examinations at death or terminal sacrifice revealed no indications of
biphenyl-induced tumors or tumor-related lesions in organs or tissues other than those associated
with the urinary tract. As shown in Table 4-4, neoplastic and nonneoplastic lesions of the
urinary bladder were essentially limited to the 4,500 ppm rats and predominantly the males.
Only 4,500 ppm male rats exhibited papilloma (10/50) or carcinoma (24/50) of transitional cell
epithelium, three of which exhibited both papilloma and carcinoma. Most of the transitional cell
carcinomas (20/24) projected into the lumen, and the tumor cells invaded the entire body wall.
Bladder calculi were found in all 24 males with transitional cell carcinoma and 8/10 of the males
27
DRAFT - DO NOT CITE OR QUOTE
-------
1 with transitional cell papilloma. Among noncancerous responses in the bladder, simple, nodular,
2 and papillary hyperplasias were evident in 4,500 ppm animals. These hyperplasias developed in
3 the focal area of the bladder epithelium. Simple hyperplasia occurred less frequently than
4 nodular and papillary hyperplasias; furthermore, simple hyperplasia was almost always
5 accompanied by either nodular or papillary hyperplasia in the 4,500 ppm males. Ten of the
6 4,500 ppm males had polyps in the bladder epithelium, which were composed of spindle fibers
7 proliferated around transitional epithelial cells accompanied by inflammatory infiltration of
8 submucosal bladder epithelium. Squamous metaplasia was noted on the surface of the polyps,
9 which were found at different loci than the bladder tumors.
10 Table 4-5 summarizes the incidences of lesions of the ureter and kidney in the male and
11 female rats. The incidence of simple transitional cell hyperplasia in the ureter was greater in the
12 4,500 ppm males than the 4,500 ppm females. Other responses, such as mineralization of the
13 corticomedullary junction, were increased over controls to a greater extent in males compared to
14 females. In the renal pelvis, simple and nodular hyperplasia was frequently observed in
15 4,500 ppm males and 500 and 1,500 ppm females. Responses such as papillary necrosis, infarct,
16 and hemosiderin deposition occurred predominantly in exposed females.
17
28 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-5. Incidences of ureter and kidney lesions in male and female
F344 rats exposed to biphenyl in the diet for 2 years
Dietary concentration (ppm)
Calculated dose (mg/kg-d)
Response
Ureter
Transitional cell hyperplasia
Simple hyperplasia
Nodular hyperplasia
Dilatation
Kidney
Renal pelvis
Transitional cell hyperplasia
Simple hyperplasia
Nodular hyperplasia
Squamous metaplasia
Mineralization
Desquamation
Calculi
Other
Mineralization of
corticomedullary junction
Mineralization of papilla
Papillary necrosis
Infarct
Hemosiderin deposits
Chronic nephropathy
Males (n = 50)
0
0
1
0
0
6
0
0
9
1
0
0
9
0
0
0
45
500
36.4
0
0
0
8
1
0
6
0
0
0
9
0
0
0
45
1,500
110
0
0
0
5
1
0
10
0
0
0
14
0
0
0
43
4,500
378
8a
1
14a
19C
21a
2
18b
lla
13a
10a
23C
?d
0
0
34
Females (n = 50)
0
0
0
0
0
3
0
0
12
0
0
21
2
0
1
4
33
500
42.7
0
0
0
5
0
0
12
0
0
2
6
0
0
8
35
1,500
128
0
0
0
12C
1
0
18
0
0
26
3
0
0
22a
30
4,500
438
2
0
6b
25a
12a
0
27a
2
3
18
12a
23a
8C
25a
26
1
2
3
4
5
6
7
"Significantly different from control group (p < 0.01) according to % test.
bSignificantly different from control group (p < 0.05) according to Fisher's exact test.
Significantly different from control group (p < 0.05) according to %2 test.
dSignificantly different from control group (p < 0.01) according to Fisher's exact test.
Source: Umeda et al. (2002).
In summary, the chronic toxicity and carcinogen!city study of male and female F344 rats
administered biphenyl in the diet for 2 years (Umeda et al., 2002) provides evidence for
biphenyl-induced bladder tumors in males, but not females, based on the development of
transitional cell papillomas and carcinomas in the 4,500 ppm (438 mg/kg-day) males (Table 4-4).
This study identified a no-observed-adverse-effect level (NOAEL) of 500 ppm (42.7 mg/kg-day)
and a lowest-observed-adverse-effect level (LOAEL) of 1,500 ppm (128 mg/kg-day) for
nonneoplastic kidney lesions (simple transitional cell hyperplasia in the renal pelvis and
hemosiderin deposits) in female F344 rats exposed to biphenyl in the diet for 2 years.
29
DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Shiraiwa etal, 1989
3 The chronic toxicity of biphenyl was assessed in Wistar rats (50/sex/group) administered
4 the chemical at 0, 0.25, or 0.5% (0, 2,500, or 5,000 ppm) in the diet for up to 75 weeks. The rats
5 were observed daily for clinical signs. Body weight and food consumption were measured
6 weekly. At death or scheduled sacrifice, gross pathologic examinations were performed and all
7 organs were removed and preserved. Other than body weight and compound consumption data,
8 the published results of this study were limited to kidney weight data and urolithiasis findings.
9 Based on reported values for mean daily biphenyl intake (mg biphenyl/rat) and mean initial and
10 final body weights for each study group, doses of biphenyl at the 0.25 and 0.5% dietary levels
11 are estimated to have been 165 and 353 mg/kg-day for males, respectively, and 178 and
12 370 mg/kg-day for females, respectively. Mean final body weights in both 2,500 and 5,000 ppm
13 groups of biphenyl-exposed male and female rats were significantly lower (approximately 15 and
14 25% lower; p < 0.01) than their respective controls. Absolute and relative kidney weights of
15 control and biphenyl-exposed rats were similar, with the exception of significantly increased
16 (p < 0.001) mean relative kidney weight in 2,500 ppm female rats. The study authors reported
17 the occurrence of hematuria (in both the 2,500 and 5,000 ppm groups) as early as week 16 and
18 stated that it was more recognizable at 60 weeks. Kidney stone formation was reported in
19 6/46 and 1/43 of the 2,500 ppm males and females, respectively, and in 19/47 and 20/39 of the
20 5,000 ppm males and females, respectively. Detection of stones in other regions of the urinary
21 tract was essentially limited to the 5,000 ppm groups and included the ureter (2/47 males and
22 2/39 females) and urinary bladder (13/47 males and 6/39 females). Kidney stones were hard,
23 black, and located from the pelvic area to the medullary region. Stones in the ureter were hard,
24 black, and composed of protein. Stones in the urinary bladder were hard, yellowish-white, round
25 to oval in shape, and composed of ammonium magnesium phosphate. Histologically, kidneys
26 with stones exhibited obstructive pyelonephritis accompanied by hemorrhage, lymphocytic
27 infiltration, tubular atrophy, cystic changes of tubules, and fibrosis. Urinary bladders with stones
28 exhibited simple or diffuse hyperplasia and papillomatosis of the mucosa; however, neoplastic
29 lesions were not seen. No control rats (44 males and 43 females) showed stones in the kidney,
30 ureter, or urinary bladder. The lowest exposure level in this study, 2,500 ppm in the diet for
31 75 weeks, was a LOAEL for formation of kidney stones associated with pyelonephritis in Wistar
32 rats (dose levels of 165 and 178 mg/kg-day for males and females, respectively). Urinary
33 bladder stones associated with simple or diffuse hyperplasia and papillomatosis of the mucosa of
34 the urinary bladder was observed at the highest exposure level, 5,000 ppm biphenyl in the diet
35 (dose levels of 353 and 370 mg/kg-day for males and females, respectively).
36 Shiraiwa et al. (1989) also reported the results of an initiation-promotion study in male
37 Wistar rats (25/group) that included three groups administered a basal diet for 2 weeks followed
38 by diets containing 0, 0.125, or 0.5% biphenyl (0, 1,250, or 5,000 ppm) for 34 weeks. Three
30 DRAFT - DO NOT CITE OR QUOTE
-------
1 other groups received diets containing 0.1% N-ethyl-N-hydroxyethylnitrosamine (EHEN, an
2 initiator of kidney tumors in rats) for 2 weeks followed by diets containing 0, 0.125, or 0.5%
3 biphenyl (0, 1,250, or 5,000 ppm) for 34 weeks. Initial and final body weights were recorded.
4 At terminal sacrifice, gross pathologic examinations were performed. The study report included
5 information regarding kidney weights, but did not indicate whether weights of other organs were
6 measured. Kidney and urinary bladder were fixed; kidneys were sectioned transversely (10-
7 12 serial slices) and urinary bladders were cut into 4-6 serial slices. The authors used a
8 computer-linked image analyzer to determine the incidence of kidney lesions and dysplastic foci.
9 The presence of stones in the kidney and urinary bladder was assessed qualitatively using an
10 infrared spectrophotometer. Based on reported values for mean daily biphenyl intake (mg
11 biphenyl/rat) and average body weight (mean initial body weight + one-half the difference
12 between mean initial and mean final body weight) for each study group, doses of biphenyl at the
13 0.125 and 0.5% dietary levels are estimated to have been 59.28 and 248.3 mg/kg-day,
14 respectively, for rats on basal diet alone for the first 2 weeks and 62.0 and 248.2 mg/kg-day,
15 respectively, for rats receiving EHEN in the diet for the first 2 weeks. The mean final body
16 weight of the rats receiving basal diet followed by diet containing 0.5% biphenyl was
17 significantly lower (p < 0.001) than that of controls (0.389 ± 22 vs. 0.432 ± 30 kg). It was stated
18 that relative kidney weights were increased this group of biphenyl-exposed rats compared to the
19 basal diet control group, but the actual data were not presented. Stones were detected only in the
20 rats receiving 0.5% biphenyl in the diet; incidences were 4/25 (kidney), 1/25 (ureter), and
21 3/25 (urinary bladder) in rats that had received that basal diet for the first 2 weeks. Similar
22 results regarding final body weight and the detection of stones in the urinary tract were reported
23 for the rats that had received EHEN in the diet prior to the administration of biphenyl.
24 Incidences of dysplastic foci and renal cell tumors were determined in the kidneys of all groups
25 of rats. Only rats that had received EHEN during the initial 2 weeks exhibited neoplastic kidney
26 lesions (dysplastic foci, renal cell tumors). For the EHEN + 0% biphenyl, EHEN + 0.125%
27 biphenyl, and EHEN + 0.5% biphenyl groups, incidences of rats with dysplastic foci were 25/25,
28 21/25, and 25/25, respectively, and incidences of rats with renal cell tumors were 13/25, 12/25,
29 and 7/25, respectively. Under the conditions of this study, biphenyl did not exhibit tumor
30 promoting characteristics for the kidney tumor initiator, EHEN.
31
32 Ambrose et al, 1960
33 Weanling albino rats (15/sex/group) were administered biphenyl in the diet at
34 concentrations of 0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, or 1% for 2 years (10, 50, 100, 500, 1,000,
35 5,000, or 10,000 ppm). Body weights were monitored every week during the period of active
36 growth and then at 50-day intervals. Hemoglobin was monitored every 100 days in control and
37 high-dose rats; at 500, 600, and 700 days in rats receiving 0.5% biphenyl, and at 500 and
38 600 days in rats receiving 0.1% dietary biphenyl. A 98-day paired-feeding experiment was
31 DRAFT - DO NOT CITE OR QUOTE
-------
1 conducted in which control rats were provided the same amount of food that rats of the 0.5 and
2 1.0% dietary biphenyl groups consumed to assess whether possible differences in growth would
3 indicate a biphenyl exposure-related toxicological response or decreased palatability. At
4 necropsy, the weights of liver, kidneys, heart, and testes were determined for all groups except
5 those receiving 1.0% biphenyl in the diet. Stained sections of heart, lung, liver, kidney, adrenal,
6 spleen, pancreas, stomach, intestine, bladder, thyroid, brain, pituitary, and gonads were prepared
7 for histopathologic examinations. In some cases, bone marrow smears were prepared.
8 The study report of Ambrose et al. (1960) did not include sufficient information from
9 which daily biphenyl doses could be calculated. Biphenyl doses are estimated at 1, 4, 8, 42, 84,
10 420, and 840 mg/kg-day for the dietary levels of 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, and 1.0%,
11 respectively, based on U.S. EPA (1988) reference values for body weight and food consumption
12 in F344 rats (averages of values for males and females). There is greater uncertainty in the dose
13 estimates at the two highest exposure levels because the magnitude of reported decreased food
14 consumption in these groups was not specified in the study report. Decreased longevity was
15 apparent in male and female rats of the 0.5 and 1.0% biphenyl exposure groups, but was not
16 evident at lower exposure levels. Growth rates appeared similar among controls and groups
17 exposed to biphenyl levels <0.1%. At the two highest exposure levels, markedly decreased
18 growth was evident, but was attributable to decreased food consumption and indicative of
19 decreased palatability based on results of the paired-feeding experiment. Decreased hemoglobin
20 levels were reported in male and female rats of the two highest exposure levels after 300-
21 400 and 500-600 days, respectively, but were considered at least partially related to lower food
22 consumption in these groups relative to controls. Selected organ weights are summarized in
23 Table 4-6. There were no statistically significant treatment-related effects on organ weights at
24 dietary levels <0.1%, which were below those associated with decreases in food consumption,
25 body weight, and survival (i.e., 0.5 and 1.0%). Relative liver and kidney weights of female rats
26 of the 0.5% biphenyl exposure group were significantly (p < 0.05) increased, approximately
27 45 and 215% higher than those of respective controls. The only significant compound-related
28 histopathological change occurred in the kidneys, which, in all members of the two highest
29 exposure groups, showed irregular scarring, lymphocytic infiltration, tubular atrophy, and tubular
30 dilation associated with cyst formation. Some evidence of hemorrhage was present, and calculi
31 were frequently noted in the renal pelvis. Evidence of metaplasia in the epithelium of the renal
32 pelvis did not implicate neoplastic activity, and, taking the histopathological results as a whole,
33 there appeared to be no clear-cut, compound-related tumor development. However, the small
34 number of animals in each group and the decreased survival in the two highest dose groups may
35 have impaired the ability to detect late-developing tumors. The study identified 1,000 ppm
36 biphenyl in the diet (84 mg/kg-day) as a NOAEL and 5,000 ppm (420 mg/kg-day) as the LOAEL
37 for kidney effects including tubular atrophy and dilation associated with cyst formation and
38 calculi formation in the renal pelvis of albino rats of both sexes.
32 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-6. Body and organ weight data for male and female rats
administered biphenyl in the diet for 2 years
Percent biphenyl
in diet
Days on
diets
Number
of rats
Mean body weight
(g)±SE
Mean relative organ weight (g) ± SE
Liver
Kidneys
Heart
Testes
Males
0.0
0.001
0.005
0.01
0.05
0.1
0.5
745
744
747
752
730
746
746
9
8
10
11
13
10
2
396 ±24.6
424 ±5.1
383 ±19.8
394 ±14.2
371 ±15.8
366 ±23.7
345
2.89 ±0.16
2.66 ± 0.06
2.84 ±0.15
2.47 ± 0.07
3.03 ±0.12
2.98 ±0.19
3.12
0.75 ±0.02
0.70 ±0.03
0.73 ±0.02
0.72 ±0.01
0.74 ±0.02
0.83 ±0.05
1.17
0.32 ±0.015
0.28 ± 0.008
0.30 ±0.01
0.31 ±0.008
0.31 ±0.007
0.34 ±0.012
0.36
0.72 ±0.03
0.62 ± 0.07
0.56 ±0.06
0.67 ± 0.07
0.65 ±0.06
0.60 ±0.08
0.36
Females
0.0
0.001
0.005
0.01
0.05
0.1
0.5
745
744
747
752
730
746
746
9
6
5
11
5
5
5
333 ±9.4
369 ±13.4
335 ±16.6
341 ±9.1
306 ±12.5
327 ±6.8
226 ±25.8
3.11±0.15
3.21 ±0.17
2.81 ±0.28
3.46 ±0.74
3.51±0.12
3.18±0.10
4.52±0.20a
0.65 ±0.01
0.62 ± 0.02
0.64 ± 0.02
0.62 ± 0.02
0.68 ±0.02
0.65 ±0.01
1.39±0.14a
0.33 ±0.01
0.28 ±0.07
0.31 ±0.03
0.30 ±0.01
0.31 ±0.01
0.32 ±0.01
0.46 ± 0.04
NA
NA
NA
NA
NA
NA
NA
"Significantly different from controls (p < 0.05) according to two-tailed Student's t-test.
NA = not applicable; SE = standard error of the mean
Source: Ambrose et al. (1960)
2
3 Pecchiai and Saffiotti, 1957
4 Male albino rats (8/group; strain not stated) were given biphenyl in the diet for up to
5 13 months at concentrations resulting in estimated doses of 250 or 450 mg/kg-day. Upon
6 sacrifice, liver, kidney, spleen, heart, lung, thyroid, parathyroid, adrenal, pancreas, testis,
7 stomach, and intestine were processed for histopathological examination. At 2-month interim
8 sacrifices, moderate degenerative changes in liver and kidney were observed at both dose levels.
9 Liver effects consisted of moderate degeneration and hypertrophy of the Kupffer cells with a
10 generally well-preserved structure. Renal glomeruli were undamaged, but tubuli showed mild
11 signs of degeneration. The liver and kidney effects did not appear to increase in severity in rats
12 treated for up to 13 months. Other histopathologic effects noted in the biphenyl-treated rats
13 included hypertrophied splenic reticular cells, small follicles with sparse colloid and
14 desquamation of follicular epithelium in the thyroid, and hyperplastic and hyperkeratinized
15 forestomach epithelium with occasional desquamation. Although the study report did not
16 include tumor incidence data for the two dose groups, the study authors reported neoplastic
17 lesions in the forestomach of three biphenyl-treated rats. Two of the rats exhibited papillomas of
33
DRAFT - DO NOT CITE OR QUOTE
-------
1 the forestomach epithelium (one after 7 weeks and one after 7 months of treatment); a squamous
2 cell carcinoma was diagnosed in the other rat after 1 year of treatment. The study authors noted
3 two sequential responses to chronic biphenyl exposure: degenerative changes of nuclei and
4 cytoplasm in the parenchyma of liver and kidney, spleen, thyroid, and adrenals within 2 months
5 followed within 1 month or more by functional-regenerative changes that resulted in hyperplasia
6 and nuclear hypertrophy of liver and kidney parenchyma as well as functional hyperactivity of
7 the thyroid and parathyroid. Signs of cirrhosis were not seen, but irritation and hyperplasia were
8 evident in the lower urinary tract. The lowest dose, 250 mg/kg-day biphenyl, was an apparent
9 LOAEL for nonneoplastic degenerative changes in the liver, kidney, thyroid, and parathyroid of
10 male albino rats resulting in hyperplasia of liver, kidney, and thyroid.
11
12 Dow Chemical Co., 1953
13 Sprague-Dawley rats (12/sex/group) were exposed to biphenyl in the diet for 2 years at
14 exposure levels of 0, 0.01, 0.1, or 1% (0, 100, 1,000, or 10,000 ppm). Body weights were
15 monitored twice weekly for 3 months, then weekly. Blood samples were taken from all animals
16 at the start of the experiment, approximately every 3 months thereafter, and at term. Hemoglobin
17 levels, red and white blood cell counts and differential cell counts, and BUN concentrations were
18 recorded. At death or scheduled necropsy, organ weights were recorded for liver, lung, kidneys,
19 heart, and spleen. Sections from heart, liver, kidney, spleen, adrenals, pancreas, gonads,
20 stomach, small and large intestine, voluntary muscle, lung, bladder, and brain were fixed and
21 stained for histopathologic examination.
22 Based on U.S. EPA (1988) chronic reference values for body weight and food
23 consumption in Sprague-Dawley rats (average values for combined sexes), doses of biphenyl for
24 the dietary levels of 0.01, 0.1, and 1% are estimated to have been 7, 73, and 732 mg/kg-day,
25 respectively. It is unclear to what extent the data in the study were compromised by an outbreak
26 of pneumonia that affected the colony during the course of the experiment. Survival was poor in
27 control males, all of which had died by 18 months. Only two of the females receiving 0.1%
28 biphenyl in the diet survived to the end of the 21st month, and none had survived by the end of
29 the 23rd month. However, the authors considered the decreased survival in this group of females
30 to have been compound-related. Striking biphenyl concentration-related reductions in body
31 weight gain were observed among the groups, although, in monitoring food efficiency, the
32 authors indicated that the reduced growth was likely due to a lower daily consumption of food
33 rather than to the toxicological consequences of biphenyl. There were no clear indications of
34 exposure-related changes in hematological parameters, but the authors reported significant
35 (p < 0.05) increases in average (combined sexes) relative liver and kidney weights at the highest
36 exposure level, compared with control values (4.71 vs. 3.05 g/100 g and 1.68 vs. 1.00 g/100 g,
37 respectively). Histopathologic examinations revealed dilatation of the kidney tubules, an effect
38 that appeared to be associated with secondary inflammation, uremia, disruption of the filtration
34 DRAFT - DO NOT CITE OR QUOTE
-------
1 system, and an increase in BUN in affected animals. Since tubular dilatation was evident in
2 controls as well as treated animals, the authors presented their data on a semi quantitative severity
3 scale (0-4) in which 0 = no observed changes, 1 = tissue changes in occasional isolated areas,
4 2 = tissue changes in multiple areas, 3 = tissue changes in numerous areas, and 4 = extensive
5 tissue changes involving all or almost all areas. Among the controls, low-, mid-, and high-dose
6 rats, respective incidences for tubular dilatation with severity scores >2 were 1/12, 6/12, 7/12,
7 and 11/12 for males and 1/12, 3/12, 4/12, and 11/12 for females. Respectively, incidences for
8 tubular dilatation with severity scores >3 were 0/12, 1/12, 2/12, and 9/12 for males and 1/12,
9 2/12, 2/12, and 11/12 for females. Severity scores >3 for tubular dilatation are considered to
10 represent adverse renal effects. Calcification and intratubular inflammation were frequently
11 observed at the highest biphenyl exposure level. The incidence data for renal tubular dilatation
12 with a severity score >3 indicate that 100 ppm biphenyl in the diet (73 mg/kg-day) was a
13 NOAEL and that 1,000 ppm (732 mg/kg-day) was a LOAEL for renal effects in Sprague-Dawley
14 rats. The small number of rats in the exposure groups and the decreased survival at the highest
15 exposure level may have impaired the ability to detect late-developing tumors in this study.
16
17 4.2.1.2.2. Chronic mouse studies
18 Umeda et al, 2005
19 In a chronic toxicity and carcinogenicity study of BDFi mice (50/sex/group), biphenyl
20 was administered in the diet for 2 years at concentrations of 0, 667, 2,000 or 6,000 ppm. All
21 animals were observed daily for clinical signs and mortality. Body weights and food
22 consumption were recorded weekly for the first 14 weeks and every 4 weeks thereafter.
23 Hematological and clinical chemistry parameters were measured in blood samples drawn from
24 all 2-year survivors just prior to terminal sacrifice. At death or terminal sacrifice, gross
25 pathological examinations were performed and organs were removed and weighed. Specific
26 tissues prepared for microscopic examination were not listed in the study report, but included
27 liver and kidney.
28 There were no overt clinical signs or effects on food consumption or survival among
29 biphenyl-exposed mice of either sex compared to respective controls. However, mean terminal
30 body weights of 2,000 and 6,000 ppm mice of both sexes were significantly less than those of
31 respective controls (Table 4-7). Based on body weight and food consumption data, the study
32 authors estimated that the 667, 2,000, and 6,000 ppm dietary levels resulted in average daily
33 biphenyl doses of 97, 291, and 1,050 mg/kg-day in the males and 134, 414, and 1,420 mg/kg-day
34 in the females (Table 4-5).
35
3 5 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-7. Survival rate, body weight, food consumption, and daily biphenyl
intake in mice fed diets containing biphenyl for 2 years
Biphenyl in diet (ppm)
Survival at term
Average (± SD) body
weight at term (g)
Average food
consumption (g/d)
Average dose
(mg/kg-d)
Males
0
667
2,000
6,000
35/50
41/50
41/50
39/50
46.9 ±4.9
43.1 ±7.9
42.9±6.0a
32.4±3.6b
5.6
5.5
5.5
5.4
0
97
291
1,050
Females
0
667
2,000
6,000
31/50
22/50
25/50
32/49
34.0 ±4.0
32.5 ±3.3
30.5 ± 3. lb
25.5±3.0b
5.9
5.8
5.9
5.9
0
134
414
1,420
1
2
o
5
4
5
6
7
8
9
10
11
12
13
"Significantly different from controls (p < 0.05) according to Dunnett's test.
bSignificantly different from controls (p < 0.01) according to Dunnett's test.
Source: Umeda et al. (2005).
Although there were no compound-related changes in hematological parameters, some
clinical chemistry parameters showed marked changes in relation to dose, including a biphenyl
dose-related increase in BUN that achieved statistical significance in 6,000 ppm males and
females and 2,000 ppm males. Particularly striking were dose-related increases in activities of
the plasma enzymes AP, lactate dehydrogenase (LDH), glutamate oxaloacetate transaminase
(GOT; also referred to as AST), and glutamate pyruvate transaminase (GPT; also referred to as
ALT) in the female mice. These data are shown in Table 4-8 and are suggestive of biphenyl-
related hepatocellular disruption. Umeda et al. (2005) noted that females with malignant liver
tumors exhibited extremely high AST, ALT, and LDH activities. Biphenyl effects on these
parameters in males were less obvious, although AP activity was significantly greater than
controls in 6,000 ppm males (261 ± 102 vs. 178 ± 111 IU/L) (Table 4-8).
36
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-8. Dose-related changes in selected clinical chemistry values from
male and female BDFi mice exposed to biphenyl via the diet for 2 years
Males
Biphenyl dietary
concentration (ppm)
Dose (mg/kg-d)
Endpoint (mean ± SD)
AST (IU/L)
ALT (IU/L)
AP(IU/L)
LDH (IU/L)
BUN (mg/dL)
0
0
n = 34
85 ±92
73 ±113
178±111
321 ±230
20.2 ±3.6
667
97
n = 39
58 ±38
34±31
155 ±30
252 ± 126
22.0 ±4.0
2,000
291
n = 37
69 ±60
36 ±49
169 ±36
432 ± 868
23.2±4.4b
6,000
1,050
n = 37
88 ±151
43 ±80
261 ± 102a
283 ± 200
22.9±2.7a
Females
Biphenyl dietary
concentration (ppm)
Dose (mg/kg-d)
Endpoint (mean ± SD)
AST (IU/L)
ALT (IU/L)
AP(IU/L)
LDH (IU/L)
BUN (mg/dL)
0
0
n = 28
75 ±27
32 ±18
242 ± 90
268 ± 98
14.9 ±2.0
667
134
n = 20
120 ±110
56 ±46
256 ±121
461 ± 452
14.8 ±3.4
2,000
414
n = 22
211±373a
134±231a
428 ± 499
838 ± 2,000
21.0 ±20.5
6,000
1,420
n = 31
325 ± 448a
206 ± 280a
556 ± 228a
1,416 ±4,161b
23.8±11.7a
aSignificantly different from controls (p < 0.01) according to Dunnett's test.
bSignificantly different from controls (p < 0.05) according to Dunnett's test.
ALT (GPT) = alanine aminotransferase (glutamic pyruvic transaminase); AP (ALP) = alkaline phosphatase;
AST (GOT) = aspartate aminotransferase (glutamic oxaloacetic transaminase)
Source: Umeda et al. (2005).
1
2 The only apparent exposure-related effect on organ weights was 1.3-, 1.4-, and 1.6-fold
3 increases in relative liver weights of 667, 2,000, and 6,000 ppm female mice, respectively (the
4 data for liver weight group means and standard deviations [SDs] were not presented in Umeda et
5 al. [2005]). Incidences of gross and histopathological findings are presented in Table 4-9. Gross
6 pathologic examinations revealed biphenyl dose-related increased incidences of liver nodules in
7 females, but not males. The nodules were round- or oval-shaped cystic or solid masses
8 approximately 3-23 mm in diameter of the largest axis. Histopathological examinations
9 revealed that 5, 16, and 19 of the nodule-bearing 667, 2,000, and 6,000 ppm female mice also
10 exhibited proliferative lesions of hepatocellular origin. Significantly increased incidences of
11 basophilic cell foci were observed in 2,000 and 6,000 ppm female mice. Although incidences of
12 basophilic cell foci were significantly increased in 667 ppm male mice as well, a dose-related
13 effect was not evident because incidences of this lesion were not significantly increased in
14 2,000 or 6,000 ppm males compared to controls. Incidences of hepatocellular adenomas and
37
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
14
combined incidences of hepatocellular adenomas or carcinomas were significantly increased in
the 2,000 and 6,000 ppm females and Peto's trend tests confirmed significant positive trends for
dose-related increased incidences of hepatocellular adenomas (p < 0.05) and combined
incidences of hepatocellular adenomas or carcinomas (p < 0.01). Incidences of hepatocellular
carcinomas were significantly increased in 2,000 ppm females, but not 667 or 6,000 ppm
females. However, Umeda et al. (2005) noted that the incidences of hepatocellular carcinomas
(5/50 or 10%) in each of the 667 and 6,000 ppm groups of females exceeded a range of historical
control data (26 hepatocellular carcinomas in 1,048 female mice [2.5% incidence]) at the
laboratory where the study was conducted. No significant biphenyl exposure-related effects on
liver tumor incidences were seen in male mice. Incidences of desquamation of the urothelium in
the renal pelvis were increased in 6,000 ppm male and female mice. Incidences of
mineralization in the inner stripe of the outer medulla of the kidney were significantly increased
in the 2,000 and 6,000 ppm female mice.
Table 4-9. Incidences of gross and histopathological findings in male and
female BDFi mice fed diets containing biphenyl for 2 years
Parameter
Dietary concentration of biphenyl (ppm)
Males
0
667
2,000
6,000
Females
0
667
2,000
6,000
Average dose (mg/kg-d)
0
97
291
1,050
0
134
414
1,420
Necropsy
Liver nodules
20/50
16/49
14/50
11/50
7/50
13/50
24/50
26/49
Histopathology
Liver
Adenoma
Carcinoma
Adenoma or carcinoma
(combined)
Basophilic cell foci
Clear cell foci
Eosinophilic cell foci
Kidney
Desquamation: pelvis
Mineralization inner stripe-
outer medulla
8/50
8/50
16/50
0/50
0/50
0/50
0/50
9/50
6/49
8/49
12/49
6/49b
6/49b
0/49
0/49
8/49
7/50
5/50
9/50
1/50
2/50
0/50
0/50
14/50
3/50
4/50
7/50
2/50
0/50
0/50
10/50b
14/50
2/50
1/50
3/50
1/50
2/50
0/50
4/50
3/50
3/50
5/50
8/50
1/50
1/50
1/50
0/50
5/50
12/503
7/50a
16/50b
12/50b
3/50
0/50
0/50
12/503
10/493
5/49
14/493
6/49a
2/49
0/49
15/49b
26/49b
15
16
17
"Significantly different from controls (p < 0.05) according to Fisher's exact test.
bSignificantly different from controls (p < 0.01) according to Fisher's exact test.
Source: Umeda et al. (2005).
In summary, the chronic toxicity and carcinogenicity study of male and female
mice administered biphenyl in the diet for 2 years (Umeda et al., 2005) provides evidence for
38
DRAFT - DO NOT CITE OR QUOTE
-------
1 biphenyl-induced liver tumors in females, but not males, based on significantly increased
2 incidences of hepatocellular adenomas and combined carcinomas or adenomas in the female
3 mice receiving biphenyl from the diet at 414 and 1,420 mg/kg-day (Table 4-9). This study
4 identified a NOAEL of 134 mg/kg-day and a LOAEL of 414 mg/kg-day for nonneoplastic
5 effects (mineralization in the kidney and significantly increased plasma ALT and AST activities)
6 in female BDFi mice exposed to biphenyl in the diet for 2 years.
7
8 Imaietal, 1983
9 Groups of female ddY mice were fed diets containing 0 (n = 37 mice) or 0.5%
10 (n = 34 mice) biphenyl (5,000 ppm) in the diet for 2 years. This study also included groups
11 exposed to dietary concentrations of 0.2% thiabendazole or a mixture of 0.25% biphenyl and
12 0.1% thiabendazole (results from this part of the study are not further described herein). Food
13 consumption, body weights, and survival were assessed at intervals throughout exposure. At
14 terminal sacrifice, several organs were weighed and prepared for microscopic histology (brain,
15 pituitary, thymus, liver, spleen, pancreas, lung, heart, adrenal, kidney, ovaries, and uterus); in
16 addition, the thyroid, stomach, small intestine, and large intestine were prepared for histology
17 only. Urine samples collected from 10 control and 9 treated mice at terminal sacrifice were
18 analyzed for protein glucose, ketones, bilirubin, urobilogen, and pH. Blood samples collected at
19 the terminal sacrifice from 12 control and 9 treated mice were assessed for hematological
20 endpoints, and serum samples (n = 6 for control and treated groups) were also assessed for
21 biochemical endpoints including GOT, GPT, AP, cholinesterase, glucose, albumin, and total
22 protein. Based on U.S. EPA (1988) methodology for estimating food consumption rates from
23 body weight data and the reported average terminal body weight for the 5,000 ppm mice
24 (0.037 kg), an oral dose of 855 mg/kg-day is estimated from the dietary exposure. Exposure to
25 biphenyl did not influence survival, food consumption, or growth compared with controls (as
26 shown in Figures 1, 2, and 3 in Imai et al. [1983]). No marked exposure-related effects were
27 found on terminal organ and body weights (Tables 5 and 6 in Imai et al. [1983]) or on the
28 urinalytic, hematologic, or serum biochemical endpoints (Tables 2, 3, and 4 in Imai et al.
29 [1983]). Histological examination revealed no increased incidence of non-neoplastic lesions in
30 examined tissues in the 5,000 ppm biphenyl group, compared with the control group (Table 7 in
31 Imai et al. [1983]). The only tissues showing tumors at elevated incidence in the 5,000 ppm
32 mice, compared with the control group, were the lung (11/34 [32.4%] vs. 9/37 [24.3%] in
33 controls) and lymphatic tissues (lymphomas: 5/34 [14.7%] vs. 4/37 [10.8%]; leukemia: 3/34
34 [8.8%] vs. 2/37 [5.4%]), but these increases are not statistically significant (p > 0.05 by the
35 Fisher's exact test). In summary, exposure of female ddY mice to 5,000 ppm biphenyl in the diet
36 for 2 years was a NOAEL for non-neoplastic lesions, survival, body and organ weight changes,
37 and changes in urinalytic, hematologic, and serum chemistry endpoints. In contrast to the 2-year
38 bioassay with BDFi mice that found increased liver tumors in female mice exposed to dietary
39 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations >2,000 ppm (Umeda et al., 2005), no carcinogenic response occurred in female
2 ddY mice exposed to 5,000 ppm biphenyl in the diet (estimated dose of 855 mg/kg-day) for
3 2 years (Imai et al., 1983).
4
5 Innes et al, 1969; NCI, 1968
6 The carcinogenic potentials of 130 chemicals, including biphenyl, were assessed in a
7 protocol that exposed groups of two strains of Fl hybrid mice (18/sex/strain/group), produced by
8 mating female C57BL/6 mice to either male C3H/Anf mice (Fl designated as strain A) or male
9 AKR mice (Fl designated as strain B) to individual chemicals by the oral route for 18 months.
10 Four groups of untreated controls and a group of gelatin vehicle controls (18/sex/strain/group)
11 were included in the study. In the case of biphenyl, the chemical was administered via gavage to
12 mice for 3 weeks, starting at the age of 7 days at 215 mg biphenyl/kg body weight in 0.5%
13 gelatin (the report of Innes et al. [1969] appears to have erroneously reported the gavage dose as
14 2.5 mg/kg). Thereafter, and for the rest of the experimental period, biphenyl was mixed with
15 chow to a final concentration of 517 ppm. The gavage dose level and food concentration of
16 biphenyl were selected to reflect the maximum tolerated dose identified in preliminary range-
17 finding single-dose subcutaneous injection and single- and repeated-dose oral administration
18 studies. Initial gavage dose and dietary levels of biphenyl were not adjusted for weight gain
19 during the 18-month study. Based on U.S. EPA (1988) chronic reference values for body weight
20 and food consumption in strain A mice (average values for combined sexes), an average oral
21 dose of 91 mg/kg-day is estimated from the dietary exposure. Blood smears were prepared from
22 mice that showed splenomegaly, liver enlargement, or lymph adenopathy at necropsy. At term,
23 mice were examined for any gross pathological features. Major organs were processed for
24 histopathologic examination (including total chest contents, liver, spleen, kidneys with adrenals,
25 stomach, and genital organs). Innes et al. (1969) reported incidences for hepatomas, pulmonary
26 tumors, and lymphomas in control mice (Table 5 of Innes et al., 1969) and for tested chemicals
27 that were judged to give "high tumor yield" (Table 7 of Innes et al., 1969); biphenyl was
28 reported to be noncarcinogenic, but tumor incidence data for biphenyl were not reported. The
29 NCI (1968) report included tabulated incidences of hepatomas, pulmonary tumors, and
30 lymphomas in control mice and biphenyl-treated mice, which are summarized in Table 4-10. In
31 summary, the results provide no evidence of a carcinogenic response to 18 months of oral
32 exposure to biphenyl (215 mg/kg by gavage for 3 weeks, followed by dietary exposure to
33 517 ppm biphenyl).
34
40 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-10. Incidences of selected tumor types among controls and mice
administered biphenyl orally for 18 months
Group
Incidences of selected tumor types"
Hepatoma
Pulmonary tumors
Reticular cell sarcoma
Strain A male mice
Controls
Biphenyl-treated
8/79
2/17
5/79
3/17
5/79
1/17
Strain A female mice
Controls
Biphenyl-treated
0/87
0/18
3/87
1/18
4/87
0/18
Strain B male mice
Controls
Biphenyl-treated
5/90
3/17
10/90
1/17
1/90
0/17
Strain B female mice
Controls
Biphenyl-treated
1/82
0/17
3/82
0/17
4/82
4/17
aTumor incidences were tallied from those mice for which histopathologic examinations were performed.
Source: NCI (1968).
1
2 4.2.1.2.3. Chronic studies in other animal species
3 Monsanto, 1956
4 Mongrel dogs (two males and one female/group) were administered 0, 2.5, or 25 mg/kg
5 biphenyl in corn oil by capsule 5 days/week for 1 year. Dogs were examined daily for clinical
6 signs and weighed weekly. Blood samples were withdrawn at 3-month intervals to measure such
7 hematological parameters as hemoglobin, hematocrit, blood cell count, sedimentation rate,
8 icterus index, bromosulphalein retention, and, among clinical chemistry parameters, BUN. Urine
9 samples were obtained at similar intervals to measure specific gravity, sugar, protein, bile
10 pigments, occult blood, and microscopic sediment. Samples of urine from the high-dose dogs
11 were collected during week 18, pooled, and analyzed for the presence of biphenyl and
12 metabolites. At termination, gross necropsies were performed, and sections of large and small
13 intestine, pancreas, ovary or testis, adrenal, urinary bladder, stomach, lung, thyroid, brain, heart,
14 spleen, and liver were prepared for histopathologic examination. Although slight fluctuations
15 were seen in body weight during the study, the dogs generally exhibited a net weight gain.
16 Fluctuations in hematological parameters and urine analysis were inconsistent and not
17 considered compound-related. Gross pathological examination of the dogs showed no obviously
18 compound-related effects. Histopathologic examinations revealed lung congestion consistent
19 with bronchial pneumonia in one high-dose dog; histopathology was unremarkable for each of
20 the other dogs in the study.
21
41
DRAFT - DO NOT CITE OR QUOTE
-------
1 Dow Chemical Co., 1953
1 Dow Chemical Co. (1953) described a biphenyl feeding experiment in which four groups
3 of Rhesus monkeys (two males and one female/group) were exposed to 0, 0.01,0.1, or 1%
4 biphenyl in chow for 1 year, during which time most of the animals experienced ill health not
5 related to biphenyl exposure. Despite this caveat, hematological parameters were normal. The
6 authors considered an increase in relative liver weight in high-dose monkeys (4.65 g/100 g body
7 weight vs. 3.90 g/100 g body weight in controls) to possibly be compound-related.
8
9 4.2.2. Inhalation Studies
10 Deichmann et al, 1947; Monsanto, 1946
11 In three separate experiments, albino rabbits (sex and strain not stated), Sprague-Dawley
12 rats (sex not stated), and mice (sex and strain not stated) were repeatedly exposed to dusts
13 composed of 50% biphenyl attached to celite for 7 hours/day, 5 days/week. In the first
14 experiment, 3 rabbits and 10 rats were exposed to an average concentration of 300 mg/m3 on
15 each of 64 days over a period of 94 days. The rats exhibited irritation of the nasal mucosa
16 accompanied by serosanguineous discharge. Five of the rats died prior to term, and the survivors
17 lost weight. The rabbits exhibited no exposure-related adverse signs. In the second experiment,
18 three rabbits and six rats were exposed to an average concentration of 40 mg/m3 on each of
19 46 days over a total period of 68 days. One rat died prior to term. The surviving rats showed
20 signs of mucous membrane irritation, but appeared to gain weight at a normal rate. The rabbits
21 exhibited no exposure-related adverse signs. In the third experiment, 12 mice and 4 rats were
22 exposed to an average concentration of 5 mg/m3 on each of 62 days over a total period of
23 92 days. While the rats were unaffected at this concentration, all of the mice showed signs of
24 irritation of the upper respiratory tract and two died prior to term. Bronchopulmonary lesions
25 (including acute emphysema, congestion, edema, bronchitis, widespread lobular pneumonia, and
26 multiple pulmonary abscesses) were reported in rats from experiments 1 and 2 and in mice of
27 experiments. Some unspecified minor liver and kidney lesions were also noted. Based on the
28 results of these three experiments, a LOAEL of 5 mg/m3 for upper respiratory tract irritation in
29 the mice was identified.
30
31 Sun Company Inc., 1977b
32 Groups of CD-I mice (50/sex/group) were exposed to airborne biphenyl at vapor
33 concentrations of 0, 25, or 50 ppm (0, 157.7, and 315.3 mg/m3, respectively) for 7 hours/day,
34 5 days/week for 13 weeks. Mice were maintained and exposed to biphenyl in groups of 5 (for a
35 total of 10 groups/sex/exposure group). All animals were checked daily for clinical signs and
36 mortality, and body weight data were collected. Upon completion of the 13-week exposure
37 period, surviving mice were placed in metabolic cages for 12-hour collection of urine for
38 urinalysis. Blood samples were collected for blood chemistry and hematology assessments.
42 DRAFT - DO NOT CITE OR QUOTE
-------
1 Gross and histopathologic examinations were performed on all mice. Ten surviving
2 mice/sex/group were held for a 30-day recovery period prior to terminal sacrifice.
3 During the first few days of biphenyl exposure, some of the test material crystallized in
4 the delivery system; analysis of biphenyl exposure levels was not performed on these days.
5 Daily measured biphenyl exposure concentrations were highly variable during the first half of
6 the 13-week exposure period, whereas subsequently measured concentrations were closer to
7 target concentrations. For example, during the first 45 exposure sessions, measured daily
8 biphenyl concentrations in the 50 ppm target groups ranged from as low as 5 ppm to as high as
9 102 ppm and subsequent measurements ranged from 48 to 55 ppm. Mean biphenyl
10 concentrations (± 1 SD) calculated for the entire 13 weeks of exposure were 25 ± 7 and
11 50 ± 16 ppm for the 25 and 50 ppm target groups, respectively. The authors reported the loss of
12 46/100 of the 25 ppm mice due to overheating and cannibalization. Since the overheating event
13 occurred after 46 exposures, the overall study duration ran for 117 days to ensure that
14 replacement mice received a total of 65 exposures as called for in the protocol. The study report
15 did not mention results of clinical observations, and mortality data were not specifically
16 summarized. There were no clear indications of exposure-related effects on body weights.
17 Results of urinalysis, hematology, and clinical chemistry did not indicate any clear exposure-
18 related changes that could be attributed to biphenyl toxicity. Gross and histopathological
19 examinations revealed congested and hemorrhagic lungs, hyperplasia of the trachea with
20 inflammation accompanied by a high incidence of pneumonia, and congestion and edema in liver
21 and kidney of biphenyl-exposed mice (see Table 4-11). The pathologist considered the
22 congestion in the lung, liver, and kidney a likely effect of the anesthetic used for killing the mice,
23 although control mice did not exhibit these effects at 13-week sacrifice. The hemorrhagic lungs
24 and tracheal hyperplasia were considered effects of biphenyl exposure. Results from the 30-day
25 recovery groups suggest that the biphenyl exposure-related pulmonary effects were reversible.
26 This study identified a LOAEL of 25 ppm for histopathologic lung, liver, and kidney lesions in
27 male and female CD-I mice exposed to biphenyl by inhalation for 7 hours/day, 5 days/week for
28 13 weeks.
43 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-11. Incidences of selected histopathologic lesions in tissues of CD-I
mice exposed to biphenyl vapors 7 hours/day, 5 days/week for 13 weeks
Effect
Pulmonary congestion, edema
Pneumonia
Trachea! hyperplasia
Hepatic congestion, edema
Renal congestion, edema
13-Week exposure groups"
0 ppm
0/80
0/80
0/80
0/80
0/80
25 ppm
95/98
15/98
80/98
87/98
87/98
50 ppm
71/71
20/71
70/71
71/71
71/71
aThe study report presented incidences of histopathologic lesions for combined male and female mice only; no
statistical analyses were conducted.
Source: Sun Company Inc. (1977b).
2
3 4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
4 4.3.1. Oral Exposure
5 Khera etal, 1979
6 Pregnant female Wistar rats (18-20 group) were gavaged with 0, 125, 250, 500, or
7 1,000 mg/kg-day biphenyl in corn oil on gestation days (GDs) 6-15. Body weights of dams
8 were recorded on GDs 1, 6-15, and 22, at which point all dams were sacrificed. Parameters
9 evaluated at autopsy included the number of corpora lutea, fetal weights and viability, and early
10 resorptions. Two-thirds of the live fetuses/litter were examined for skeletal development and the
11 rest were examined for the presence of visceral abnormalities. Five of the 20 high-dose dams
12 died prior to sacrifice. Doses <500 mg/kg-day produced no clinical signs of maternal toxicity or
13 evidence of treatment-related effects on maternal weight gain. As shown in Table 4-12, a
14 significantly increased number of dams without live fetuses was observed in the high-dose
15 group, compared with controls. Mean numbers of corpora lutea and live fetuses in the high-dose
16 dams were similar to those of controls and dams of all other dose levels. However, the percent
17 of dead fetuses and resorption sites was clearly higher in the high-dose group, and the numbers
18 of anomalous fetuses and litters bearing anomalous fetuses appeared to increase with increasing
19 dose. Khera et al. (1979) noted that the slight increases in the number of fetuses with anomalies,
20 such as missing and unossified sternebrae or delayed calvarial ossification, were not statistically
21 significant, but, as shown in Table 4-12, the incidence of litters with any type of fetal anomalies
22 ("anomalous litters/number examined") was elevated (p < 0.05 by Fisher's exact test) at
23 500 mg/kg-day, but not at lower doses, compared with control incidences. This study identified
24 a NOAEL of 500 mg/kg-day and a LOAEL of 1,000 mg/kg-day for frank maternal toxicity
25 (increased mortality and decreased dams with live fetuses) and lethal fetal effects. For less
44
DRAFT - DO NOT CITE OR QUOTE
-------
1 severe developmentally toxic effects (increased incidence of anomalous litters), 500 mg/kg-day
2 was a LOAEL and 250 mg/kg-day was a NOAEL.
3
Table 4-12. Prenatal effects following oral administration of biphenyl to
pregnant Wistar rats on GDs 6-15
Effect
Rats without live fetuses at term/number
mated
Corpora lutea/pregnancy (mean ± SE)
Live fetuses/pregnancy (mean ± SE)
Dead or resorbed fetuses (%)
Fetal weight (g mean ± SE)
Anomalous fetuses/number examined
Anomalous litters/number examined
Anomalies (number of fetuses affected)
Wavy ribs, uni- and bilateral
Extra ribs, uni- and bilateral
13th rib, small sized
Sternebrae, missing or unossified
Calvarium, delayed ossification
Miscellaneous
Dose (mg/kg-d)
0
2/18
12.6 ±0.4
11.3 ±0.7
4.8
5.1±0.1
17/176
8/16
3
9
1
4
0
1
125
0/20
12.9 ±0.4
11.8 ±0.6
3.3
5.3 ±0.1
22/236
11/20
7
12
1
3
2
1
250
1/19
13.7 ±0.5
11.9 ±0.6
6.1
5.2 ±0.1
22/213
13/18
9
9
2
4
0
1
500
2/20
13. 3 ±0.4
11.2±0.5
7.8
5.2 ±0.1
35/199c
15/18C
8
15
1
16
0
0
1,000
11/203
12.5 ±0.7
10.7 ±1.3
13.7b
4.5 ±0.3
25/107c
6/9
5
6
0
17
8
0
4
5
6
7
8
9
10
11
12
13
14
"Significantly (p < 0.05) different from control incidence according to Fisher's exact test. Five dams died prior to
scheduled sacrifice, five other dams were not pregnant at term, and one dam had seven resorption sites and no live
fetuses.
bDerived from nine pregnant dams with live fetuses and one dam with seven resorptions and no live fetuses. The
study author stated that the percentage of dead or resorbed fetuses in the 1,000 mg/kg dose group was not
statistically significantly different from controls.
Significantly (p < 0.05) different from controls according to Fisher's exact test.
Source: Kheraetal. (1979).
Dow Chemical Co., 1953
Dow Chemical Co. (1953) reported the results of a multigenerational study in which
groups of 4-month-old male and female Long Evans rats (three males and nine females/group)
were fed diets containing 0, 0.01, 0.1, or 1.0% biphenyl. Based on U.S. EPA (1988) subchronic
reference values for body weight and food consumption in male and female Long Evans rats,
doses of biphenyl for the dietary levels of 0.01, 0.1, and 1.0% are estimated to have been 9, 89,
and 887 mg/kg-day, respectively, for the males and 10, 101, and 1,006 mg/kg-day, respectively,
for the females. Average cross-gender doses for males and females were 10, 95, and 947 mg/kg-
day. For breeding, three females were placed together with one male. Following the breeding
phase, females were separated and number of litters cast, number of days between mating and
45
DRAFT - DO NOT CITE OR QUOTE
-------
1 delivery, and average number of pups/litter at delivery were recorded. Fl pups were weighed
2 and culled to seven/litter at 2 days of age and weaned at 3 weeks of age, and weights were
3 recorded weekly for postnatal weeks 3-6. The Fl rats were continued on the same diets as their
4 parents, and, at 10 weeks of age, nine Fl females and three Fl males were mated to produce an
5 F2 generation of pups. F2 pups were selected (by the same procedure) for mating and
6 production of an F3 generation that were sacrificed at 3 weeks of age; twelve F3 pups from each
7 diet group were subjected to gross pathologic examinations.
8 There were no significant differences between controls and 0.01 and 0.1% biphenyl-fed
9 groups regarding litters cast; gestation length; or average number or weight of pups/litter at birth
10 or at 3 or 6 weeks of age. Decreased fertility in the 1% biphenyl-fed group of females was
11 observed (6/9, 7/9, and 8/9 confirmed pregnancies for the three successive generations of 1.0%
12 biphenyl-fed groups vs. 8/9, 9/9, and 8/9 confirmed pregnancies for controls). Averaged for Fl,
13 F2, and F3 pups combined, the 1.0% biphenyl-fed group exhibited significantly (p < 0.05)
14 decreased number of pups/litter at birth (6.2/litter vs. 8.6/litter for controls) and lower average
15 body weight at 3 weeks of age (36 vs. 48 g for controls) and 6 weeks of age (78 vs. 113 g for
16 controls). Gross pathologic evaluations of F3 weanlings revealed no signs of biphenyl treatment-
17 related effects. There was no evidence of a cumulative effect over the three generations. The
18 study authors indicated that the decreased fertility, smaller litter size, and reduced rate of growth
19 in the 1.0% biphenyl-fed group may have been associated with unpalatability and resultant
20 decreased food intake.
21
22 Ambrose et al, 1960
23 The research report of Ambrose et al. (1960) contains a subsection in which the
24 reproductive toxicity of biphenyl was examined in two experimental series. In the first
25 experiment, weanling albino rats were administered 0 or 0.1% biphenyl (5 males and
26 10 females/group) or 0.5% biphenyl (3 males and 9 females) in the diet for 60 days prior to
27 mating. In the second experiment, groups of 90-day-old albino rats were administered 0 or 0.1%
28 biphenyl (4 males and 8 females/group) or 0.5% biphenyl (3 males and 9 females) in the diet for
29 11 days prior to mating. Based on U.S. EPA (1988) subchronic reference values for body weight
30 and food consumption in rats of unspecified strain (average values for combined sexes), doses of
31 biphenyl for the dietary levels of 0.1 and 0.5% are estimated to have been 105 and 525 mg/kg-
32 day, respectively. All rats were maintained on their respective diets throughout mating and until
33 the progeny of all litters were weaned. Insufficient information is provided in the report to
34 permit a judgment as to whether dietary exposure to biphenyl was associated with reproductive
35 deficits. However, the authors presented tabular data for number of rats casting litters, total
36 born, and range of litter size (Table 4-13) and concluded that the compound had no significant
37 effect on reproduction.
38
46 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-13. Summary of reproductive data in albino rats exposed to dietary
biphenyl
Experimental series
First3
Secondb
Diet
Control
0.1% biphenyl
0.5% biphenyl
Control
0.1% biphenyl
0.5% biphenyl
Dams with litters
9/10
10/10
8/9
8/8
6/8
8/9
Total offspring
59
67
53
64
63
48
Litter size (range)
3-9
2-10
3-9
5-13
3-10
3-9
"Weanling rats on diets for 60 days before mating.
b90-Day-old rats on diets for 11 days before mating.
Source: Ambrose et al. (1960).
1
2 4.3.2. Inhalation Exposure
3 No studies were identified that examined the reproductive/developmental toxicity of
4 biphenyl via the inhalation route.
5
6 4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
7 4.4.1. Acute and Short-term Toxicity Data
8 Acute oral toxicity studies of biphenyl provide median lethal dose (LD50) values ranging
9 from 2,180 to 5,040 mg/kg for rats (Monsanto, 1976; Pecchiai and Saffiotti, 1957; Union
10 Carbide, 1949; Deichmann et al., 1947) and an LD50 value of 2,410 mg/kg for rabbits
11 (Deichmann et al., 1947). Dow Chemical Co., (1939) reported 100% survival and 100% lethal
12 doses of 1,600 and 3,000 mg/kg, respectively, in rats. Clinical signs commonly observed
13 following single oral dosing in these studies included increased respiration, lacrimation, loss of
14 appetite and body weight, and muscular weakness. Deaths occurred in the first few days
15 following dosing. Typical targets of histopathologic lesions were lungs, liver, and upper
16 gastrointestinal tract.
17 In another acute study, Pecchiai and Saffiotti (1957) administered single gavage doses of
18 biphenyl at 1-2.5, 3-6, 7, 9-11, or 10-13 mg/kg to groups of rats (n = 2-10) and observed them
19 for up to 7 months following dosing. Histopathological changes to the liver, kidney, thyroid,
20 parathyroid, and gastrointestinal mucosa were reported in biphenyl-treated rats; however, the
21 study report did not provide information regarding numbers of treatment-related deaths or
22 incidences of lesions in the various treatment groups. Among surviving rats, signs of
23 regeneration were evident within 1-4 months after treatment. By 7 months after treatment, most
24 of the changes had disappeared, but hepatocytes displayed modest vacuolization of the
47
DRAFT - DO NOT CITE OR QUOTE
-------
1 cytoplasm and numerous binucleate cells in the periphery of the lobules. In renal tubuli, a
2 moderate number of cytoplasmic granules were observed.
o
J
4 Sun Company Inc., 1977a
5 Groups of mice (10/sex of unspecified strain) were exposed to biphenyl by inhalation for
6 4 hours at average analytical concentrations of 14.11, 38.40, or 42.80 ppm (89.0, 242.2, and
7 270.0 mg/m3, respectively) and observed for up to 14 days following exposure. Clinical signs of
8 hyperactivity and mild respiratory discomfort were noted during exposure, but resolved during
9 postexposure observation. A solitary male mouse of the 42.80 ppm group died after 2 hours of
10 exposure, but this death was not attributed to biphenyl exposure. All other mice survived
11 throughout the 14-day postexposure observation period. Slight lung congestion was noted in
12 most mice upon gross pathological examination.
13 Sun Company Inc. (1977a) also provided details of a study in which groups of mice
14 (10/sex of unspecified strain) were exposed to biphenyl for 7 hours/day, 5 days/week for 2 weeks
15 at average analytical concentrations of 0, 24.8, or 54.75 ppm (0, 156.4, and 345.5 mg/m3,
16 respectively). Five animals/group were sacrificed immediately after exposure; the remaining
17 animals were sacrificed following a 14-day recovery period. Clinical signs were monitored
18 daily. Gross pathologic examinations at necropsy included assessment of lungs, trachea, heart,
19 spleen, liver, kidneys, stomach, and intestines. Histopathologic examinations included tissues
20 from lung, trachea, kidney, spleen, and liver. The study authors reported signs of hyperactivity
21 in some mice during the first few exposure periods. One female mouse of the 24.8 ppm
22 exposure group died prior to the third exposure session and one control female mouse died prior
23 the final exposure session. No abnormal clinical signs were seen during the 14-day recovery
24 period. Gross and histopathologic examinations revealed no signs of exposure-related adverse
25 effects.
26
27 Deichmann et al, 1947; Monsanto, 1946
28 Four rabbits (sex and strain unspecified) received up to 20 daily doses of 500 mg/kg
29 "purified" biphenyl to the skin; the compound was applied as a 25% preparation in olive oil.
30 Three rabbits received the same concentration of technical biphenyl. The compound was left on
31 the skin for 2 hours and then washed off with soap and water. Some biphenyl derivatives were
32 similarly assessed. One rabbit receiving purified biphenyl died after eight applications, and the
33 rest of the animals survived to term. The only reported sublethal effect clearly associated with
34 biphenyl exposure was that of weight loss, averaging 45 and 172 g for the rabbits receiving
35 purified and technical biphenyl, respectively.
36
48 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.4.2. Kidney/Urinary Tract Endpoint Studies
2 Several endpoint-specific studies assessed biphenyl-induced urinary tract effects in rats
3 (Shibata et al., 1989a, b; Kluwe, 1982; S0ndergaard and Blom, 1979; Booth et al., 1961) and
4 provide support to findings of the chronic oral rat studies described in Section 4.2.1.2 (Chronic
5 Toxicity and Carcinogenicity). Detailed descriptions of these endpoint-specific studies are
6 presented below.
7
8 Booth et al., 1961
9 In a preliminary study, five adult rats (sex and strain unspecified) were administered
10 biphenyl in the diet at 1% (w/w) for 26 days followed by a 29-day postexposure recovery period
11 for a total study period of 55 days. Total urine volume and the volume of sulfosalicylic acid-
12 precipitable sediment were recorded from urine collected from all five rats on study days 4, 8,
13 18, 20, and 26 (exposure days), and study days 28, 32, 35, and 54 (recovery period). Volumes of
14 both urine and sulfosalicylic acid-precipitable sediment increased from 7 and 0.56 mL,
15 respectively, on exposure day 4 to 32 and 2.24 mL, respectively, on exposure day 20. Both
16 values remained relatively high (approximately 27 and 2.2 mL, respectively) on exposure day 26
17 and decreased to approximately 14 and 0.8 mL, respectively, by the end of the recovery period.
18 Fractionation and analysis of the precipitate suggested the presence of p-hydroxybiphenyl and its
19 glucuronide. The study authors indicated that similar effects were noted in male and female rats
20 receiving biphenyl at a level of 0.5% in the diet, but not at the 0.1% dietary level.
21 A follow-up study employed 42 rats of each sex and exposure group and biphenyl dietary
22 levels of 0, 0.1, 0.25, or 0.5% (w/w). Biphenyl doses are estimated at 83.7, 209, and 419 mg/kg-
23 day for the dietary levels of 0.1, 0.25, and 0.5%, respectively, based on U.S. EPA (1988) chronic
24 reference values for body weight and food consumption in F344 rats (averages of values for
25 males and females). Rats were exposed for up to 165 days and followed for 0, 30, or 60 days of
26 recovery. Urine samples were collected periodically from five rats/sex/exposure group. Interim
27 sacrifices of five rats/sex/exposure group were performed after 30, 60, and 120 days on the diet
28 in order to assess the progression of biphenyl-induced histopathological effects on the kidney.
29 As noted in the preliminary study, the rats of the 0.5% exposure group in the follow-up study
30 exhibited gradual increases in the urine volume and sulfosalicylic acid-precipitable sediment and
31 decreased in both parameters during postexposure recovery. The study authors indicated that
32 these effects were much less pronounced in the 0.25% exposure group and absent in the 0.1%
33 exposure group. At the 0.5% exposure level, kidney lesions were noted in 1/5 of the males
34 (several small cysts and dilated tubules in the medulla and inner cortex) and 2/5 of the females
35 (mild local tubular dilation with some epithelial flattening) following 30 days of exposure.
36 Similar, but more extensive, kidney lesions were noted in 3/5 males and 5/5 females following
37 60 days of exposure. The kidney lesions were even more prominent following 120 days of
38 exposure. Reported histopathologic findings in the kidneys of rats from the 0.25% exposure
49 DRAFT - DO NOT CITE OR QUOTE
-------
1 group were limited to a single instance of an unspecified "prominent kidney lesion" at 60 days,
2 and one small calculus in the pelvis of one rat and a small calcareous deposit in the renal
3 pyramid of another rat following 120 days of exposure. Based on available information in the
4 study report, there were no apparent assessments of urinary and histopathologic renal effects at
5 the end of the 165-day treatment period. However, during the 60-day postexposure recovery
6 period, rats of the 0.5% biphenyl exposure group exhibited a regression of kidney lesions and
7 improvement in urine quality.
8
9 Kluwe, 1982
10 Kluwe (1982) examined changes in urine composition and kidney morphology in F344
11 rats exposed to biphenyl. Groups of male F344 rats were administered biphenyl (in corn oil) by
12 single gavage dosing at 0, 250, 500, or 1,000 mg/kg and observed for 15 days following
13 treatment. Body weights were recorded, and urine was collected on days 1, 2, 3, 4, 8, and
14 15 following treatment for urinalysis. Interim sacrifices were performed on eight control and
15 eight high-dose rats on posttreatment days 1, 2, 3, 8, and 15 for assessment of weight and
16 histopathology of the kidney. The study authors presented body weight data as mean percent
17 (n = 6) of preexposure body weight; results of urinalyses were presented as mean values (n = 6)
18 for each group. There were no significant effects on body weight in the low-dose group. Mean
19 body weight gains of mid- and high-dose groups were consistently 6-10% lower than control
20 values (p < 0.05), beginning as early as day 2 following the initiation of dosing and continuing
21 through day 15. Dose-related increases in polyuria, proteinuria, and glucosuria were observed on
22 day 1; polyuria and glucosuria were no longer apparent by day 4 and proteinuria resolved
23 between days 8 and 15. The study authors presented no data to indicate that single oral dosing
24 caused changes in kidney weight. Histopathologic examinations of kidneys revealed renal
25 papillary necrosis in 8/32 high-dose rats; this effect was observed as early as day 1 and persisted
26 during the 15-day posttreatment period.
27 Kluwe et al. (1982) conducted a similar experiment in which groups of male F344 rats
28 received biphenyl at doses of 0, 250, or 500 mg/kg-day by gavage for 14 days. In this
29 experiment, polyuria persisted throughout the treatment period; glucosuria was no longer
30 apparent by day 4 and proteinuria resolved between treatment days 8 and 15. Relative kidney
31 weight of high-dose rats was significantly increased during the second half of the treatment
32 period, but the magnitude of this effect was small and considered by the study authors to be of
33 little biological significance. There was some indication of tubular dilatation in focal areas of
34 kidneys from the high-dose rats.
35
36 S0ndergaard andBlom, 1979
37 Groups of male and female SPF-Wistar rats were administered diets consisting of
38 semisynthetic chow and biphenyl at concentrations resulting in biphenyl doses of 0, 50, 150, 300,
50 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
or 450 mg/kg-day. Other groups were administered diets consisting of commercial chow and
biphenyl at concentrations resulting in biphenyl doses of 0, 50, 150, 300, 500, or 1,000 mg/kg-
day. The treatment period lasted for up to 21 days. The numbers of male and female rats in each
treatment group are specified in Table 4-14. Urine was collected on days 4, 10, and 17 for
urinalysis. At terminal sacrifice, absolute and relative kidney weights were determined and
kidney tissues were prepared for light and electron microscopic assessment. Apparently, interim
sacrifices (days 1, 2, 4, and 10) were performed in order to assess the activity of AP in proximal
tubules. Table 4-14 presents semiquantitative study results, which include increases in urine
volume/specific gravity and relative kidney weight, as well as polycystic kidney changes. No
changes in AP levels were seen as a result of biphenyl exposure. The kidney effects of biphenyl
appeared to be more pronounced when added to the semisynthetic diet vs. the commercial diet,
with 50 mg/kg-day as a LOAEL for the onset of kidney changes.
Table 4-14. Number of Wistar rats exposed to biphenyl and the degree of
change in kidney weight and cellular architecture
Exposure
(mg/kg-d)
Number of animals
(male/female)
Relative kidney weight
increases
Cystic change
Increases of urine
volume/specific gravity
Semisynthetic diet
0
50
150
300
450
3/14
4/3
0/10
14/14
4/4
-
+
+
+++
+++
-
-
*
***
***
-/-
•/•
Commercial chow
0
50
150
300
500a
l,000a
10/20
10/10
10/10
10/10
0/10
0/10
-
-
-
-
+b
+++b
-
-
-
-
-
**
-/-
•/•
•/•
14
15
16
"Dose for 14 days.
bAbsolute organ weight.
+ = statistically significant compared with controls (p < 0.05), as calculated by the authors (Student's t-test);
+++ = statistically significant compared with controls (p < 0.001), as calculated by the authors (Student's t-test);
* = less than one-third of the area; ** = less than two-thirds of the area; *** = greater than two-thirds of the area;
• = effect; - = no effect.
Source: Sendergaard and Blom (1979).
Shibataetal, 1989a, b
Male F344 rats (20/group) were exposed to 0 or 0.5% (w/w) biphenyl in the diet for
51
DRAFT - DO NOT CITE OR QUOTE
-------
1 24 weeks (Shibata et al., 1989a). After 4 weeks, 5 rats/group were injected with 100 mg/kg
2 5-bromo-2-deoxyuridine (BrdU) and sacrificed 1 hour later. One kidney from each rat was
3 processed for immune-histopathologic identification of BrdU as an index of cell proliferation,
4 while the second kidney was processed for light and scanning electron microscopic examination.
5 The remaining rats were sacrificed after 8, 16, and 24 weeks to monitor further development of
6 morphological alterations in the renal papilla and pelvis. Survival was unaffected by treatment
7 and biphenyl-treated animals showed no adverse clinical signs. The study authors reported that
8 treatment resulted in significantly lower mean body weight compared to controls; food
9 consumption was unaffected and water consumption was slightly higher than that of controls.
10 There were no significant treatment-related effects on labeling indices of cell proliferation (BrdU
11 incorporation) in renal papilla or pelvic epithelia and no histopathologic lesions of the renal
12 papilla and pelvis were evident. Focal calcification of the renal medulla was observed in the
13 majority of the biphenyl-treated rats. The study authors stated that urinalysis demonstrated an
14 association between biphenyl exposure and microcalculi formation, but provided no additional
15 information regarding urinalysis results.
16 In a similar study (Shibata et al., 1989b), a group of 10 male F344 rats received 0.5%
17 (w/w) biphenyl in the diet for up to 8 weeks. Based on U.S. EPA (1988) subchronic reference
18 values for body weight and food consumption in male F344 rats, the dose was estimated at
19 500 mg/kg-day. At 4 weeks, five rats/group were processed as described by Shibata et al.
20 (1989a) for assessment of BrdU incorporation, but in the urinary bladder rather than in the
21 kidney. During week 4, urine samples were taken for urinalysis. At terminal sacrifice, urinary
22 bladder tissues were processed for scanning electron microscopic examinations. There were no
23 treatment-related deaths or adverse clinical signs. Although food and water consumption were
24 similar to controls, biphenyl-treated rats showed a consistent reduction in average body weight
25 (229 vs. 247 g after 4 weeks and 300 vs. 327 g after 8 weeks, for treated vs. controls,
26 respectively \p < 0.01]). A greater than fourfold increase in the BrdU labeling index was
27 observed in urinary bladder epithelium of the biphenyl-fed rats (mean percent labeling index of
28 0.58 ± 0.31 compared to 0.13 ± 0.09 in controls;/* < 0.05). Urinalysis revealed numerous
29 microcalculi in the urinary sediment of the biphenyl-treated rats. This condition, designated as
30 "severe" by the authors, was associated with histopathological lesions of the epithelium of the
31 urinary bladder that included simple hyperplasia with moderate severity in 5/5 rats, moderate
32 pleomorphic microvilli (5/5), moderate uniform microvilli (5/5), and the occurrence of ropey or
33 leafy microridges (5/5), the latter condition designated as severe. Scanning electron microscope
34 images of the luminal surface of bladder epithelial cells showed pleomorphic microvilli that
35 varied in size and shape and the formation of microridges.
36
52 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.4.3. Biphenyl as a Tumor Promoter
2 Tamano etal, 1993
3 Male B6C3Fi mice (10-20/group) received the bladder carcinogen BBN at 0 or 0.05% in
4 the drinking water for 4 weeks followed by 0 or 1% biphenyl in the feed for 32 weeks. The mice
5 were observed for clinical signs and body weight and food consumption were monitored. At
6 37-week terminal sacrifice, kidneys and urinary bladders were prepared for histopathological
7 examination. No treatment-related clinical signs were observed. Mean body weight of the BBN
8 + 1% biphenyl-treated mice was significantly (p < 0.01) lower than that of mice receiving BBN
9 treatment only (32.2 ±1.8 vs. 38.4 ± 2.6 g). Biphenyl treatment did not result in increased
10 incidences of simple hyperplasia or papillary or nodular dysplasia in the BBN-initiated mice.
11 Administration of 1% biphenyl in the feed to eight mice for 8 weeks did not significantly affect
12 indices of cell proliferation (BrdU incorporation) in urinary bladder epithelium.
13
14 Shiraiwa etal, 1989
15 In the initiation-promotion portion of a chronic toxicity study designed to assess the
16 ability of biphenyl to promote carcinogenesis by EHEN in the kidney (see Section 4.2.1.2 for a
17 detailed study description), male Wistar rats (25/group) received basal diet with either 0 or 0.1%
18 dietary EHEN for 2 weeks, followed by a basal diet containing either 0, 0.125, or 0.5% biphenyl
19 for 34 weeks (Shiraiwa et al., 1989). At terminal sacrifice, gross pathologic examinations were
20 performed. Kidney and urinary bladder were fixed; kidneys were sectioned transversely (10-
21 12 serial slices) and urinary bladders were cut into 4-6 serial slices. The authors used a
22 computer-linked image analyzer to determine the incidence of kidney lesions and dysplastic foci.
23 The presence of stones in the kidney and urinary bladder was assessed qualitatively using an
24 infrared spectrophotometer.
25 Based on reported values for mean daily biphenyl intake (mg biphenyl/rat) and average
26 body weight (mean initial body weight + one-half the difference between mean initial and mean
27 final body weight) for each study group, doses of biphenyl at the 0.125 and 0.5% dietary levels
28 are estimated to have been 59.28 and 248.3 mg/kg-day, respectively, for rats on basal diet alone
29 for the first 2 weeks and 62.0 and 248.2 mg/kg-day, respectively, for rats on basal diet and
30 EHEN for the first 2 weeks. Stones were present in the kidney, ureter, and urinary bladder of
31 high-dose rats irrespective of whether animals were initially exposed to the basal or
32 EHEN-containing diet (combined incidences of 6/25 and 8/25, respectively). The incidence of
33 rats with renal cell tumors after EHEN and subsequent biphenyl administration was lower than
34 that of rats receiving EHEN followed by basal diet) (7/25 and 13/25, respectively). This finding
35 indicates that biphenyl was not a promoter of renal cell tumors in male Wistar rats under the
36 conditions of the study.
37
53 DRAFT - DO NOT CITE OR QUOTE
-------
1 Kurataetal, 1986
2 Male F344 rats (25/group) were exposed to 0.05% N-butyl-N-(4-hydroxybutyl)
3 nitrosamine (BBN, a bladder carcinogen) in the drinking water for 4 weeks followed by diets
4 containing either 0 or 0.5% biphenyl for 32 weeks. One group of five rats received biphenyl
5 without pretreatment with BBN. The rats receiving biphenyl either with or without pretreatment
6 with BBN gained less weight than control rats or those receiving only BBN. Incidences of
7 urinary bladder hyperplasia, papilloma, and carcinoma were 17/18 (94%), 15/18 (83%), and
8 11/18 (61%), respectively, in the group of rats that survived treatment of BBN followed by
9 biphenyl, compared to 6/24 (25%), 3/24 (12%), and 0/24 (0%), respectively, in the rats receiving
10 BBN only. These urinary bladder lesions were not seen in any of the five rats receiving biphenyl
11 without BBN pretreatment. Urinary bladder calculi were found in 25% of the rats receiving
12 BBN followed by biphenyl and in 12% of the rats receiving BBN only. Biphenyl was
13 considered a urinary bladder tumor promoter in male F344 rats under the conditions of the study.
14
15 Boutwell and Bosch, 1959
16 Biphenyl was negative for tumor promotion in a skin-painting experiment in which the
17 initiator was 0.3% 9,10-dimethyl-l,2-benzanthracene in benzene (Boutwell and Bosch, 1959). In
18 the 16/20 mice that survived the topical application of 20% biphenyl for 16 weeks, none had
19 developed papillomas or carcinomas as a result of treatment.
20
21 Ito et al, 1984
22 Six-week-old male F344 rats (20-30/group) were exposed to BBN in drinking water at
23 0.01 or 0.05% for 4 weeks, followed by 0.5% biphenyl in the feed for 32 weeks. Controls
24 receiving only BBN and controls receiving only biphenyl were included. After sacrifice, urinary
25 bladders were prepared for light microscopic assessment of neoplastic and cancerous lesions.
26 The study authors reported that biphenyl exhibited moderate bladder cancer-promoting activity,
27 but data to support this finding were not included in the study report.
28
29 4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
30 ACTION
31 4.5.1. Effects on the Urinary Tract of Rats
32 Urinary tract effects in male rats chronically exposed to biphenyl in the diet are
33 associated with the formation of urinary bladder calculi. Mechanistic studies performed by
34 Ohnishi and coworkers (Ohnishi et al., 2001, 2000a, b) were designed to identify urinary
35 metabolites of biphenyl, to assess conditions leading to calculi formation, and to determine the
36 composition of urinary crystals and calculi. Ohnishi et al. (2000a) identified sulphate conjugates
37 of mono- and dihydroxy biphenyl metabolites in urine and urinary crystals from F344 rats treated
54 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
with biphenyl and KHCOs (to elevate the pH and K+ concentration of the urine). Male F344 rats
(five per group) were administered a diet containing 1.6% biphenyl and 5% potassium
bicarbonate for 7 days (Ohnishi et al., 2000a). Urine was collected on days 6 and 7 and pooled.
Urinary crystals (i.e., precipitates) were collected and dissolved in acetonitrile and were analyzed
by HPLC to identify metabolites or by inductively coupled plasma spectroscopy to identify
inorganic elements. As shown in Table 4-15, biphenyl sulphate conjugates in the urine consisted
primarily of 3,4-dihydroxybiphenyl-3-O-sulphate (40.9% of the total biphenyl sulphate
conjugates) and 3-hydroxybiphenyl (23.4%). No bisulphates were observed (Ohnishi et al.,
2000a). In contrast; about 90% of sulphate conjugates in urinary crystals were 4-hydroxy-
biphenyl-O-sulphate, and only 3.9 and 1.06% were 3,4-dihydroxybiphenyl-3-O-sulphate and
3-hydroxybiphenyl, respectively. In a follow-up study, Ohnishi et al. (2000b) evaluated the
composition of urinary calculi in male and female rats exposed to 4,500 ppm biphenyl in the diet
for 104 weeks. Urinary calculi in chronically exposed male rats were composed mainly of
4-hydroxybiphenyl-O-sulphate, whereas calculi in female rats were composed primarily of
4-hydroxybiphenyl and potassium sulphate, the hydrolysis products of 4-hydroxybiphenyl-
O-sulphate (Ohnishi et al., 2000b).
Table 4-15. Content of biphenyl sulphate conjugates in urine and urinary
crystals from F344 rats treated with biphenyl and potassium bicarbonate (to
elevate the pH and K^ concentration of the urine)
Biphenyl sulphate conjugates
2-Hydroxybiphenyl-O-sulphate
3 -Hy droxybipheny 1-O-sulphate
4-Hydroxybiphenyl-O-sulphate
4,4 ' -Dihydroxybipheny 1-O-sulphate
2,5-Dihydroxybiphenyl-O-sulphate
3 ,4-Dihydroxybiphenyl-3 -O-sulphate
3 ,4- Dihydroxybiphenyl-4-O-sulphate
2,3 - Dihydroxybiphenyl-3 -O-sulphate
Urine (%)
3.32a
23.37
11.94
7.17
5.62
40.88
2.27
5.43
Urine crystals (%)
0.06
1.06
89.45
3.11
0.02
3.90
2.28
0.12
18
19
20
21
22
23
24
aThe component fraction (%) for each of the sulphate conjugates was estimated from the ratio of the liquid
chromatography tandem mass spectrometry peak area of the sulfate to the total area.
Source: Ohnishi et al. (2000a).
Using the same experimental protocol as that described in Ohnishi et al. (2000a), but
adding potassium bicarbonate (5%), potassium chloride (5%), or sodium bicarbonate (8%) to the
diet for 13 weeks, Ohnishi et al. (2001) reported hydronephrosis and blood in the urine only in
those animals receiving biphenyl plus potassium bicarbonate. Feed consumption was not
affected by the dietary additions, while water intake was greatly increased in all groups of
animals that received biphenyl and/or salts. Neither high urinary potassium levels alone, as
55
DRAFT - DO NOT CITE OR QUOTE
-------
1 induced by cofeeding of potassium chloride, nor high urinary pH alone, as induced by cofeeding
2 of sodium bicarbonate, were sufficient to cause kidney damage. It was concluded that a
3 combination of high urinary pH and high potassium levels was necessary to cause precipitation
4 of biphenyl sulphate. It was proposed that the crystalline precipitate caused obstruction that led
5 to hydronephrosis or damaged the transitional epithelium in the bladder causing hyperplasia.
6
7 4.5.2. Effects on the Liver of Mice
8 Based on findings of biphenyl-induced liver tumors in female BDFi mice administered
9 high dietary concentrations of biphenyl for 2 years (Umeda et al., 2005), a 13-week oral study
10 was performed to assess whether peroxisome proliferation might be induced (Umeda et al.,
11 2004). Groups of male and female BDFi mice (10/sex/group) were administered biphenyl in the
12 diet at six different concentrations ranging from 500 to 16,000 ppm. Biphenyl concentrations
13 >8,000 ppm resulted in significantly decreased final body weights of males and females.
14 Significantly increased liver weights were noted in the 8,000 and 16,000 ppm groups of female
15 mice. Evidence of peroxisome proliferation was restricted to the 16,000 ppm group of female
16 mice and included light microscopy findings of clearly enlarged hepatocytes filled with
17 eosinophilic fine granules and electron microscopy confirmation that the granules corresponded
18 to increased numbers of peroxisomes. Light microscopy of livers from rats exposed to
19 concentrations <8,000 ppm showed no indications of proliferation of peroxisomes. There were
20 no indications of other biphenyl-induced liver effects in any of the groups of male mice.
21
22 4.5.3. Estrogenic Effects
23 Several biphenyl derivatives display estrogenic activity. Schultz et al. (2002) used the
24 Saccharomyces cerevisiae/LacZ reporter assay to study the estrogenic activity of 120 chemicals
25 to identify chemical structures that impart estrogenic activity to a molecule. Chemicals without a
26 hydroxy group, among them biphenyl, were inactive in this assay. The estrogenic activities of
27 biphenyl metabolites in this assay were 4,4'-dihydroxybiphenyl (median effective concentration
28 = (2.6 x 10"7 M) > 4-hydroxybiphenyl (1.2 x 10"6 M) > 3-hydroxybiphenyl (9.2 x 10"6 M)
29 > 2-hydroxybiphenyl (1.8 x 10"5 M). Estrogenic activities of the corresponding hydroxylated di-,
30 tri-, or tetrachlorobiphenyl metabolites were approximately two orders of magnitude higher,
31 provided there were no chlorines and hydroxy groups on the same ring.
32 Kitamura et al. (2003) used MCF-7 cells transfected with an estrogen receptor-luciferase
33 reporter construct to test biphenyl and its metabolites for estrogenic activity. The starting point
34 for this investigation was the structural similarity between hydroxylated metabolites of biphenyl
35 and of 2,2-diphenyl propane, the 4,4'-dihydroxy metabolite of which is bisphenol A, a known
36 endocrine disrupter. Biphenyl per se displayed no estrogenic activity in this assay. Metabolites
37 of biphenyl formed by liver microsome preparations were identified after solvent extraction from
56 DRAFT - DO NOT CITE OR QUOTE
-------
1 reaction media by HPLC-MS. The compounds were also tested in an in vitro competitive
2 estrogen receptor binding assay. The biphenyl metabolites, 2-, 3-, 4-hydroxybiphenyl, and
3 4,4'-dihydroxybiphenyl, all exhibited estrogenic activity when the cell culture contained
4 microsomes from 3-methylcholanthrene-induced rat livers and to a lesser extent, phenobarbital-
5 induced rat livers, in the presence of NADPH. In the competitive estrogen receptor binding
6 assay, 4,4'-dihydroxybiphenyl displayed weak binding affinity, while biphenyl and its
7 monohydroxy metabolites did not show any activity. 4,4'-Dihydroxybiphenyl is one of two
8 major biphenyl metabolites in rats and mice (Halpaap-Wood et al., 1981a, b; Meyer and
9 Scheline, 1976), suggesting that high doses of biphenyl, in the form of this metabolite, might
10 induce some minor estrogenic effect.
11
12 4.5.4. Effects on Apoptosis
13 Kokel and Xue (2006) tested a series of benzenoid chemicals (including mesitylene,
14 cyclohexane, benzene, toluene, and biphenyl) for their ability to suppress apoptosis in the
15 nematode, Caenorhabditis elegans, a model suitable for the characterization of carcinogens that
16 act by way of apoptosis inhibition. The study included wild type and three strains of C. elegans
17 mutants; the ced-3(n2438) mutant (which carries a partial loss-of-function mutation in the ced-
18 3 gene), the ced-3(n2273) mutant (also partly defective in cell death), and the ced-(n2433)
19 mutant (a strong loss-of-function ced-3 mutant). Effects on apoptosis were assessed by counting
20 the numbers of cells that should have died during embryogenesis, but inappropriately survived.
21 The results indicated that these chemicals did not significantly affect apoptosis in wild type
22 C. elegans. However, inhibition of apoptosis was apparent in mutant strains ced-3(n2438) and
23 ced-3(n2273) exposed to benzene, toluene, or biphenyl. The study authors interpreted these
24 results as indicative of apoptosis-inhibitory activity that does not depend on mutations in a
25 specific cell-death gene. A lack of apparent apoptosis-inhibitory activity in the strong loss-of-
26 function ced-3(n2433) mutant was interpreted as indicative that inhibition of apoptosis, rather
27 than transformation of cell fates, caused the increase in extra cell observed in the other two
28 mutant strains. All three chemicals also displayed embryotoxicity. Biphenyl and naphthalene
29 were both shown to suppress apoptosis in C. elegans mutant strain ced-3(n2438) by causing
30 overexpression of the CED-3 caspase. The authors proposed that benzenoid chemicals that can
31 form quinones suppress apoptosis in C. elegans via this reactive intermediate, although this was
32 proven only for benzene, toluene, and naphthalene.
33 Regulation of apoptosis during embryogenesis is critical, and a recent study by Tan et al.
34 (2011) showed that inhibition of apoptosis during this stage of development may have
35 detrimental effects on the nervous system. No literature was identified, however, that
36 specifically supports an association between inhibition of apoptosis by biphenyl and effects on
37 embryogenesis.
38
57 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.5.5. Mitochondrial Effects
2 Nishihara (1985) assessed the effects of biphenyl on the respiratory and energy linked
3 activities of rat liver mitochondria that had been isolated from male Wistar rats. Biphenyl (5-
4 60 ug/mL in acetone solvent) was added to liver mitochondria and effects on rates of succinate
5 oxidation and a-ketoglutarate/malate oxidation were assessed by measuring oxygen
6 consumption. Solvent controls were included in the study. Biphenyl significantly inhibited state
7 3 respiration at concentrations >20 ug/mL. The inhibition was greater for a-ketoglutarate/malate
8 oxidation than for succinate oxidation. State 4 respiration was significantly stimulated by
9 biphenyl; the effect was greater in magnitude for succinate than for a-ketoglutarate/malate
10 oxidation. Biphenyl also altered mitochondrial membrane permeability, as evidenced by the
11 instantaneous release of endogenous K+, leading to instantaneous dissipation of the
12 mitochondrial membrane potential. Inhibition of state 3 respiration is generally considered to
13 reflect an interference with electron transport. The study author suggested that the biphenyl-
14 induced stimulation of state 4 respiration may be explained by an uncoupling action on
15 respiration.
16
17 4.5.6. Genotoxicity
18 Biphenyl. The results of genotoxicity studies of biphenyl are summarized in Table 4-16.
19 Reverse mutation assays using Salmonella typhimurium and Escherichia coli provide
20 consistently negative results both with and without the addition of a mammalian metabolic
21 activation system (rat S9 mix). Biphenyl was not genotoxic in a host-mediated deoxyribonucleic
22 acid (DNA) repair assay of E. coli in the presence of S9 (Hellmer and Bolcsfoldi, 1992). In rec
23 assays of Bacillus subtilis, two studies reported negative results both with and without S9
24 (Garrett et al., 1986; Kojima and Hiraga, 1978), one study reported negative results without S9
25 (Kawachi et al., 1980) and one study reported equivocal results with S9 (Hanada, 1977).
26 Biphenyl was reported to induce mitotic recombination both with and without S9 in
27 Saccharomyces cerevisiae strain D3 (Pagano et al., 1988), but not in S. cerevisiae strain Diploid
28 D7 (Garrett et al., 1986).
29
58 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-16. Genotoxicity test results for biphenyl
Organism
Strain or test
system
Endpoint
Test substance
concentrations
Metabolic
activation"
+S9
-S9
Reference
Bacterial and prokary otic assays
S. typhimurium
E. coli
TA98, 100
TA98, 100, 1535,
1538
TA98, 100
TA97, 98, 100
TA98, 1535
TA98, 100,
YG1041
TA98, 100,
1535,1537, 1538,
C3076, D3052, G46
TA98, 100, 1537
TA98, 100
TA1535, 1536,
1537-1, 1538-1
TA98, 100
TA98, 100, 1535,
1537
TA98, 100
TA98, 100, 1535,
1537, 2637
TA98, 100, 1532,
1535, 1537, 1538,
2636
C3076, D3052,
G46, TA98, 1000,
1535, 1537, 1538
TA98, 100, 1535,
1537, 1538, 1978
Chromotest
WP2, WP2 uvrA~
WP2, WP2 uvrA~
WP uvrA~, polA~
B/y WP2try~,
B/y WP2try~hcr~
B/y WP2trylicr
K-12 uvrB/recA+
K-12 uvrB/recA"
Mutation
Mutation
Host-mediated
DNA repair
NS
NS
NS
1-100 ug/plate
5-1,000 ug/plateb
5-250 ug/plateb
0.1-1,000 ug/mL
1-105 ug/mLb
25-800 ug/plate
Units provided in
Japanese
1-1,000 ug/plate
1-100 ug/plate
0.15-2 ug/plate
Up to 5 mg/plate
0. 1-500 ug/plateb
104-fold range
77 ug/plate
2.4-154 ug/mL
1-1,000 ug/mL
104-fold range
1-105 ug/mL
Units provided in
Japanese
<1,000 ug/mLb
Up to 161 mM
-
-
-
-
-
-
"
-
-
-
-
-
-
-
"
"
-
-
-
-
-
-
-
-
NT
NT
-
-
NT
-
"
-
-
-
-
-
NT
"
"
-
-
-
-
-
-
NT
Bosetal., 1988
Purchase et al., 1978
Kawachi et al., 1980
Bramsetal., 1987
Narbonne et al., 1987
Chung and Adris, 2003,
2002
Cline and McMahon,
1977
Garrettetal., 1986;
Waters etal., 1982
Glattetal., 1992
Hanada, 1977
Kojima and Hiraga,
1978
Haworthetal., 1983
Houketal., 1989
Ishidateetal., 1984
Paganoetal., 1988;
1983)
Probst etal., 1981
Westinghouse, 1977
Bramsetal., 1987
Cline and McMahon,
1977
Probst etal., 1981
Garrettetal., 1986
Hanada, 1977
Kojima and Hiraga,
1978
Hellmer and
Bolcsfoldi, 1992
59
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-16. Genotoxicity test results for biphenyl
Organism
B. subtilis
S. cerevisiae
Strain or test
system
Not given
recA~
H17 (rec+)
M45 (reel
H17 (rec+)
M45 (reel
D3
Diploid D7
Endpoint
Rec assay
Mitotic
recombination
Test substance
concentrations
NS
1-105 ug/mL
Units provided in
Japanese
1 or 10 mg
1-105 ug/mL
10"5 or 10"3 M
10'5 Ma
Metabolic
activation"
+S9
NT
-
+/-
+/-
-
-
+
-S9
-
-
-
-
+
Reference
Kawachi et al., 1980
Garrettetal., 1986
Hanada, 1977
Kojima and Hiraga,
1978
Garrettetal., 1986
Paganoetal., 1988
Tests with cultured mammalian cells
Hamster
Human
Rat
V79
DON
CHL
DON
Kidney
V79
CHO
Peripheral blood
lymphocytes
Diploid lung
fibroblast
Liver-derived cells
HSBP diploid lung
fibroblast
WI-38 lung
fibroblasts
Primary hepatocyte
Immortalized liver
epithelial cells
Mutation
SCE
CA
Cell
transformation
CA
SCE
CA
Micronuclei
Cell
transformation
DNA repair
UDS
Excision
repair
DNA repair
HGPRT
mutation
5-100 ug/mL
100 ug/mLb
0.1-1 mM
NS
NS
Up to 25 ug/mL
Up to 60 ug/mL
75-125 ug/mL
0.1-1 mM
0.025-250 ug/mL
<100 ug/mL
3. 1-200 ug/mL
100 ug/mLa
10-70 uL/mL
10-70 uL/mL
10-70 uL/mL
0.025-250 ug/mL
0.025-250 ug/mL
100 uM
1-105 ug/mL
0.01-100 uM
100 uM
0.01-1,000 uM
100 uMc
100 uM
+
NT
NT
-
-
+
-
+
-
NT
NT
NT
-
-
-
-
-
-
-
-
NT
NT
-
-
NT
-
+/-
+
+
NT
NT
-
-
-
-
-
-
Glattetal. (1992)
Abe and Sasaki, 1977
Kawachi et al., 1980
Kawachi et al., 1980
Ishidateetal., 1984
Ishidate and Odashima,
1977
Sofunietal., 1985
Abe and Sasaki, 1977
Purchase et al., 1978
Glattetal., 1992
Yoshidaetal., 1978
Rencuzogullari et al.,
2008
Purchase et al., 1978
Purchase et al., 1978
Snyder and Matheson,
1985
Garrettetal., 1986;
Waters etal., 1982
Hsiaetal., 1983a,b
Probst etal., 1981
Brounsetal., 1979
Williams et al., 1989
Williams, 1980
60
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-16. Genotoxicity test results for biphenyl
Organism
Mouse
Strain or test
system
L5178Y/TK+/-
Endpoint
Mutation
Test substance
concentrations
50-500 uM
150 uMa
50-1,500 uM
500 uMa
98.7-395 uM
98.7 uMa
5-60 uM
10uMa
Metabolic
activation"
+S9
+
+d
-S9
-
+d
Reference
Garberg et al., 1988
Wangenheim and
Bolcsfoldi, 1988, 1986
In vivo tests
Rat
Mouse
Mouse
Silkworm
Bone marrow
CD-I/stomach,
colon, liver, kidney,
bladder, lung, brain,
bone marrow
CD-I/stomach,
liver, kidney,
bladder, lung, brain,
bone marrow
SCE
CA
DNA damage,
Comet assay
DNA damage,
Comet assay
Mutation
NS
NS
10-2,000 mg/kg
2,000 mg/kg
NS
-
-
+
+
-
Kawachi et al., 1980
Sasaki etal., 2002
Sasaki etal., 1997
Kawachi et al., 1980
"Lowest concentration resulting in cytotoxicity.
bLowest concentration resulting in precipitation.
'Highest concentration not causing cytotoxicity.
dPositive result only at cytotoxic concentrations.
CA = chromosomal aberrations; CHL = Chinese hamster lung; CHO = Chinese hamster ovary;
HGPRT = hypoxanthine guanine phosphoribosyl transferase; NS = not specified; NT = not tested; +/- = weakly
positive or equivocal result; empty cell = no information available; SCE = sister chromatid exchanges;
UDS = unscheduled DNA synthesis
1
2 Assays of biphenyl-exposed cultured mammalian cells provide mixed results. In the
3 absence of exogenous metabolic activation, biphenyl produced negative results for sister
4 chromatid exchanges (SCE) and/or chromosomal aberrations (CA) in the DON Chinese hamster
5 cell line (Abe and Sasaki, 1977) or Chinese hamster lung (CHL) fibroblasts (Sofuni et al., 1985;
6 Kawachi et al., 1980); cell transformations in Chinese hamster kidney cells (Purchase et al.,
7 1978) and human diploid lung fibroblasts (Purchase et al., 1978); unscheduled DNA synthesis,
8 excision repair, and DNA repair in rat hepatocytes (Brouns et al., 1979); and hypoxanthine
9 guanine phosphoribosyl transferase (HGPRT) mutation in rat immortalized liver epithelial cells
10 (Williams, 1980). In the presence of S9 mix, biphenyl produced negative results for CAs in
11 CHL fibroblasts (Ishidate et al., 1984; Ishidate and Odashima, 1977) or Chinese hamster ovary
12 (CHO) cells (Yoshida et al., 1978); DNA repair in human HSBP diploid lung fibroblasts (Snyder
61
DRAFT - DO NOT CITE OR QUOTE
-------
1 and Matheson, 1985); and unscheduled DNA synthesis in human lung WI-38 lung fibroblasts
2 (with or without S9; Garrett et al., 1986).
3 Positive results were obtained for CA in CHL fibroblasts (Sofuni et al., 1985) and
4 mutations in Chinese hamster V79 cells (Glatt et al., 1992) in the presence, but not absence, of
5 S9. Biphenyl induced forward mutations in mouse L5178Y/TK+" lymphoma cells with and
6 without S9 (Wangenheim and Bolcsfoldi, 1988, 1986); another study provided similar results in
7 the presence, but not the absence, of S9 (Garberg et al., 1988). Significant increases in SCE
8 (< twofold higher than solvent controls), CA (two- to fourfold higher than solvent controls), and
9 micronuclei (approximately 2.5-fold higher than solvent controls) were reported in human
10 peripheral blood lymphocytes exposed to biphenyl for 24-48 hours at concentrations >50 |iL/mL
11 (Rencuzogullari et al., 2008).
12 Evaluations of the potential genotoxicity of biphenyl in vivo have been performed in rats,
13 mice, and silkworms. Biphenyl did not induce SCE or CA in bone marrow cells of rats or
14 mutations in silkworms, but limited information is available for these studies (Kawachi et al.,
15 1980). In a Comet assay, positive results were reported for DNA damage in stomach, blood,
16 liver, bone marrow, kidney, bladder, lung, and brain cells of CD-I mice administered single
17 doses of 2,000 mg biphenyl/kg (Sasaki et al., 2002, 1997). It is unknown if the DNA damage
18 was caused by direct reaction with biphenyl or its metabolites, or by indirect damage from
19 cytotoxicity or ROS generated from redox cycling of hydroquinone metabolites.
20 Biphenyl metabolites. Table 4-17 summarizes results from genotoxicity tests of several
21 biphenyl metabolites, 2-hydroxybiphenyl (also known as o-phenylphenol), 4-hydroxybiphenyl
22 (the principal metabolite of biphenyl), and 2,5-dihydroxybiphenyl. 2-Hydroxybiphenyl and its
23 sodium salt have received the most research attention because they are used as fungicides and
24 anti-bacterial agents and have been found to cause urinary bladder tumors in male F344 rats with
25 chronic exposure to high concentrations in the diet (see Balakrishna et al., 2002; Kwok et al.,
26 1999; Smith et al., 1998 for review).
27
62 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-17. Genotoxicity test results for biphenyl metabolites
Organism
Strain or test
system
Endpoint
Test substance
concentrations
Metabolic
activation"
+S9
-S9
Reference
2-Hydroxybiphenyl in vitro tests
S. typhimurium
E. coli
B. subtilis
Hamster
Rat
TA98, TA100
TA98, 100, 1535,
1537
TA98, 100
TA97a, 102
TA98, 100, 1535,
1537, 2637
TA98, 100
TA1535, 1537-1,
1538-1
TA1536
B/y WP2trylicr
B/yWP2try-
WP2 lacking
catalase and
superoxide
dismutase
WP2, WP2 uvrA~,
CM571, WP100
Not given
H17 (rec+)
M45 (recT)
CHL
CHO
Liver DNA
Mutation
Streptomycin
resistance
mutation
DNA repair
Rec assay
CA
DNA adducts,
[32P]-post
labeling
method
NS
3.3-250 ug/plate
1-1,000 ug/plate
1-100 ug/plate
Up to 0.5 mg/plate
NS
Units provided in
Japanese
1-1,000 ug/mL
1,000 ug/mLa
0-10 uM
NS
10-10,000 mg/plate
Units provided in
Japanese
NS
Up to 0.05 mg/mL
3. 1-200 ug/mL
94 ug/mLa
1 mM, in presence
of rat skin
homogenate, CYP,
or prostaglandin
synthase activation
systems
-
-
-
-
-
+/-
+/-
+
+/-
NT
+
-
+
NT
-
-
+b
-
-
-
-
NT
+/-
+/-
+
+
-
+
-
NT
Kawachietal., 1980
Haworth et al., 1983
Kojima and Hiraga, 1978
Fujitaetal., 1985
Ishidateetal., 1984
Nishioka and Ogasawara,
1978
Hanada, 1977
Kojima and Hiraga, 1978
Tani et al., 2007
Nishioka and Ogasawara,
1978
Kawachietal., 1980
Kojima and Hiraga, 1978;
Hanada, 1977
Kawachietal., 1980
Ishidateetal., 1984
Yoshidaetal., 1978
Pathak and Roy, 1993
2-Hydroxybiphenyl in vivo tests
Rat
Bone marrow
F344/bladder
epithelium
SCE
Micronuclei
Hyperdiploidy/
hypodiploidy
Cell
proliferation
NS
2,000 ppm in diet,
14 days
-
+
+
Kawachietal., 1980
Balakrishnan et al., 2002
63
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-17. Genotoxicity test results for biphenyl metabolites
Organism
Rat
Mouse
Mouse
Mouse
Mouse
Mouse
Rat
Rat
Silkworm
Strain or test
system
F344/bladder
epithelium
CD-1/stomach,
colon, liver,
kidney, bladder,
lung
CD-1/brain, bone
marrow
CD-1/stomach,
liver, kidney,
bladder, lung
CD-1/brain, bone
marrow
CD-1/skin
F344/bladder
epithelium
F344/bladder
epithelium
Endpoint
DNA damage,
alkaline elution
assay
DNA damage,
Comet assay
DNA damage,
Comet assay
DNA damage,
Comet assay
DNA damage,
Comet assay
DNA adduct,
[32P]-post
labeling
method
DNA adduct,
[32P]-post
labeling
method
Cell
proliferation
DNA binding
Mutation
Test substance
concentrations
1,000 or 2,000 ppm,
sodium salt in diet
for 3 months; no
damage at 250 or
500 ppm
10-2,000 mg/kg
10-2,000 mg/kg
2,000 mg/kg
2,000 mg/kg
10 or 20 mg applied
to skin
800-12,500 ppm in
diet
15-1,000 mg/kg by
gavage, labeled with
14C]-2-hydroxy-
biphenyl, uniformly
labeled in phenol
ring
NS
Metabolic
activation"
+S9 -S9
+
+
-
+
-
+
+
-
Reference
Morimoto etal., 1989
Sasaki et al., 2002
Sasaki et al., 2002
Sasaki etal., 1997
Sasaki etal., 1997
Pathak and Roy, 1993
Smith etal., 1998
Kwoketal., 1999
Kawachietal., 1980
4-Hydroxybiphenyl in vitro tests
S. typhimurium
B. subtilis
TA98
TA1535
TA1535, 1536,
1537-1, 1538-1
H17 (rec+)
M45 (rec~)
Mutation
Rec assay
5-1,000 ug/plate
1,000 ug/platec
Units provided in
Japanese
Units provided in
Japanese
+ NT
NT
-
-
Narbonne et al., 1987
Hanada, 1977
Hanada, 1977
2,5-Dihydroxybiphenyl in vitro or in vivo tests
Human
DNA fragments
from plasmid
pbcNI
DNA damage,
Comet assay
0.1 mM
+d
Inoueetal., 1990
64
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-17. Genotoxicity test results for biphenyl metabolites
Organism
Rat
Mouse
Strain or test
system
F344/bladder
epithelium
CD-1/skin
Endpoint
DNA damage,
alkaline elution
assay
DNA adduct,
[32P]-post
labeling
method
Test substance
concentrations
0.05% injected
intravesically into
bladder wall
10 or 20 mg applied
to skin
Metabolic
activation"
+S9 -S9
e
+
Reference
Morimoto etal., 1989
Pathak and Roy, 1993
aLowest concentration resulting in cytotoxicity.
bMetabolic activation system derived from rat skin homogenate.
°Lowest concentration resulting in precipitation.
dPositive response only in the presence of Cu(II)
Injection with 0.05% or 0.1% phenylbenzoquinone, a metabolite of 2,5-dihydroxybiphenyl, produced DNA
damage at concentrations of 0.05 or 0.1%, but not at 0.005 or 0.0005%.
NS = not specified; NT = not tested; +/- = weakly positive or equivocal result; empty cell = no information
available
1
2 In bacterial mutagenicity tests or in vitro mammalian tests of 2-hydroxybiphenyl, results
3 were mostly negative or equivocal, but other tests with bacterial systems suggest that oxidative
4 DNA damage following metabolism of 2-hydroxybiphenyl to 2,5-dihydroxybiphenyl is possible
5 (see Table 4-17 for references). 2-Hydroxybiphenyl induced DNA repair in E. coli strains both
6 with and without S9 (Nishioka and Ogasawara, 1978). Tani et al. (2007) provided evidence that
7 redox cycling of a semiquinone/quinone pair causes oxidative DNA damage following exposure
8 of a mutant E. coli strain (WP2, lacking catalase and superoxide dismutase) to 2-hydroxy-
9 biphenyl: 2-hydroxybiphenyl induced streptomycin resistance mutations in the mutant, but not
10 in the wild type. Exposure of B. subtilis to 2-hydroxybiphenyl both with and without S9 in the
11 rec assay yielded positive (Kojima and Hiraga, 1978; Hanada, 1977) and negative (Kawachi et
12 al., 1980) results. 2-Hydroxybiphenyl did not induce CAs in CHL fibroblasts without S9 in one
13 study (Kawachi et al., 1980), or with S9 in other studies of CHL fibroblasts (Ishidate et al., 1984)
14 and CHO cells (Yoshida et al., 1978).
15 Results from in vivo mammalian genotoxicity test systems provide limited evidence for
16 possible genotoxic actions (DNA damage and micronuclei formation) from 2-hydroxybiphenyl
17 through its metabolites, 2,5-dihydroxybiphenyl and phenylbenzoquinone (Table 4-17).
18 DNA damage was detected by the Comet assay in the urinary bladder of CD-I mice
19 administered single oral doses of 2,000 mg 2-hydroxybiphenyl/kg, but it is unknown if the
20 damage was due to cytotoxicity, direct reaction of DNA with 2-hydroxybiphenyl or its
21 metabolites, or possible oxidative DNA damage from redox cycling of 2,5-dihydroxybiphenyl
22 (Sasaki et al., 2002, 1997). DNA damage was also detected in the urinary bladder of male or
65
DRAFT - DO NOT CITE OR QUOTE
-------
1 female rats intravesically injected with 0.05 or 0.1% phenylbenzoquinone, but not with injections
2 of 0.05% 2-hydroxybiphenyl or 2,5-dihydroxybiphenyl, although DNA damage was found in
3 urinary bladders from male F344 rats fed the sodium salt of 2-hydroxybiphenyl in the diet for
4 3 months at 1,000 or 2,000 ppm, but not at 500 or 250 ppm (Morimoto et al., 1989). Topical
5 application of 10 or 20 mg of the sodium salt of 2-hydroxybiphenyl or 5 mg of 2,5-dihydroxy-
6 biphenyl to the skin of female CD-I mice produced several DNA adducts in the skin that were
7 detected by the [32P]-post labeling technique (Pathak and Roy, 1993). Similar adducts were
8 formed in vitro when DNA was incubated with 2-hydroxybiphenyl (1 mM) in the presence
9 metabolic activation from rat skin homogenates, a CYP system, or a prostaglandin synthase
10 system (Pathak and Roy, 1993). In contrast, Smith et al. (1998), using a similar technique to that
11 used by Pathak and Roy (1993), were unable to detect exposure-related DNA adducts in bladder
12 epithelial tissue from male F344 rats fed 800, 4,000, 8,000, or 12,500 ppm 2-hydroxybiphenyl in
13 the diet for 13 weeks. In this experiment, increased bladder cell epithelium proliferation (i.e.,
14 increased BrdU incorporation) was observed at 8,000 and 12,500 ppm, dietary concentrations
15 associated with the development of urinary bladder tumors in chronically exposed rats (Smith et
16 al., 1998). Kwok et al. (1999) found no evidence of binding of radioactivity to DNA extracted
17 from the bladder epithelium of male F344 rats given single gavage doses of [14C]-labeled
18 2-hydroxybiphenyl at 15, 50, 250, 500, or 1,000 mg/kg, but increased protein binding occurred
19 with increasing doses of 250, 500, and 1,000 mg/kg. Kwok et al. (1999) noted that the increase
20 in protein binding increased with increasing dose levels of 250, 500, and 1,000 mg/kg, in parallel
21 with increasing incidence of bladder epithelial lesions (hyperplasia, papillomas, and carcinomas)
22 in rats chronically exposed to 2-hydroxybiphenyl in the diet at 0, 269, and 531 mg/kg.
23 Increased micronuclei (about threefold increase over controls) and increased cell
24 proliferation (>200-fold increased incorporation of BrdU in DNA) were found in the bladder
25 epithelium of male F344 rats exposed to 2% (2,000 ppm) 2-hydroxybiphenyl in the diet for
26 2 weeks, without evidence for hypo- or hyperploidy as assayed by fluorescence in situ
27 hybridization with a DNA probe for rat chromosome 4 (Balakrishnan et al., 2002). Similar
28 exposure to 2% NaCl or 2% 2-hydroxybiphenyl + 2% NaCl, produced about two- or six-fold
29 increases of micronuclei in the bladder epithelium, respectively, but neither treatment stimulated
30 bladder epithelium cell proliferation to the same degree as 2% 2-hydroxybiphenyl in the diet
31 (Balakrishan et al., 2002). 2-Hydroxybiphenyl reportedly did not induce SCE in the bone
32 marrow of rats, but exposure parameters were not specified in the report by Kawachi et al.
33 (1980). The mechanism of 2-hydroxybiphenyl-induced micronuclei is not understood, but, as
34 discussed by Balakrishan et al. (2002), possible mechanisms include: (1) DNA damage from
35 ROS from redox cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone,
36 (2) interference of the mitotic spindle through covalent modification of proteins, (3) inhibition of
37 enzymes regulating DNA replication, or (4) micronuclei generation as a secondary response to
38 cytotoxicity or regenerative hyperplasia.
66 DRAFT - DO NOT CITE OR QUOTE
-------
1 Bacterial mutation assays of the major biphenyl metabolite, 4-hydroxybiphenyl, yielded
2 negative results in all but one case that was accompanied by overt cytotoxicity (Narbonne et al.,
3 1987). 2,5-Dihydroxybiphenyl (i.e., phenylhydroquinone) caused in vitro damage to human
4 DNA from plasmidpbcNI in the presence of Cu(II) (Inoue et al., 1990), DNA adducts when
5 applied to mouse skin (Pathak and Roy, 1993), but did not cause DNA damage when injected
6 intravesically into the urinary bladder of F344 rats at a concentration of 0.05% (Morimoto et al.,
7 1989).
8 In summary, the overall weight of evidence for biphenyl genotoxicity from short-term
9 tests is negative or equivocal (Table 4-16). Biphenyl did not induce mutations in a variety of
10 bacterial test systems (in the absence or presence of exogenous metabolic activation), but in vitro
11 assays of genotoxicity in mammalian test systems yielded a mix of negative and positive results,
12 with positive results mostly in the presence of metabolic activation. In tests of clastogenic
13 effects in mammalian systems, biphenyl induced SCE, CAs, and micronuclei in cultured human
14 peripheral blood lymphocytes (Rencuzogullari et al., 2008) and CAs in one assay of CHL
15 fibroblasts in the presence, but not the absence, of rat liver metabolic activation (Sofuni et al.,
16 1985). However, biphenyl did not induce clastogenic effects (in the presence of metabolic
17 activation) in other assays with Chinese hamster fibroblasts (Ishidate et al., 1984; Ishidate and
18 Odashima, 1977) or CHO cells (Yoshida et al., 1978). In the only adequately reported in vivo
19 genotoxicity studies with biphenyl, single oral doses of 2,000 mg/kg of biphenyl or
20 2-hydroxybiphenyl induced DNA damage in several organs of CD-I mice (including liver and
21 bladder), but it is uncertain if the damage was due to a direct effect on DNA by biphenyl or its
22 metabolites or indirectly due to cytotoxicity or ROS generated by redox cycling of a
23 hydroquinone metabolite of 2-hydroxybiphenyl (Sasaki et al., 2002, 1997).
24 The overall weight of evidence for 2-hydroxybiphenyl genotoxicity suggests that
25 oxidative DNA damage from redox cycling between 2,5-dihydroxybiphenyl and phenylbenzo-
26 quinone is possible (Sasaki et al., 2002, 1997; Pathak and Roy, 1993; Morimoto et al., 1989), but
27 no evidence for DNA adducts or DNA binding in urinary bladder epithelium tissue was found in
28 rats following short-term (Kwok et al., 1999) or subchronic (Smith et al., 1998) oral exposure to
29 2-hydroxybiphenyl at high doses associated with the formation of urinary bladder tumors.
30 Increased micronuclei in urinary bladder epithelium were detected in rats exposed to 2%
31 2-hydroxybiphenyl or its sodium salt in the diet for 14 days (Balakrishnan et al., 2002). The
32 mechanism of this clastogenic effect is uncertain, but could involve micronuclei formation in
33 secondary response to cytotoxicity or regenerative cell proliferation, DNA damage from ROS
34 generated from redox cycling of a hydroquinone metabolite, or protein modifications leading to
35 mitotic spindle interference or inhibition of enzymes important in DNA replication.
36 4-Hydroxybiphenyl, the predominant metabolite of biphenyl, was not mutagenic in
37 bacterial testing at noncytotoxic concentrations (Narbonne et al., 1987; Hanada, 1977).
38 2,5-Dihydroxybiphenyl (i.e., phenylhydroquinone) caused in vitro damage to human DNA from
67 DRAFT - DO NOT CITE OR QUOTE
-------
1 plasmidpbcNI in the presence of Cu(II) (Inoue et al., 1990) and DNA adducts when applied to
2 mouse skin (Pathak and Roy, 1993), but did not cause DNA damage when injected intravesically
3 into the urinary bladder of F344 rats at a concentration of 0.05% (Morimoto et al., 1989).
4
5 4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS
6 Tables 4-18 and 4-19 include the major studies and the observed effects for oral and
7 inhalation exposure to biphenyl, respectively.
68 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-18. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the
LOAEL
Comments
Reference
Subchronic studies
Rat, Long-Evans
(female, 8/group)
Mice, BDFj
(10/sex/group)
Diet
Diet
0, 10, 30, or 100
90 d
0, 93, 347, 747,
1495, 1868, or 2989
13wks
ND
M:747
F: 1868
ND
M: 1495
F: 2989
Lower average plasma
BUN levels in all
exposed groups
(statistical significance
not reported and
biological significance
is uncertain).
M: Decreased body
weight.
F: Decreased body
weight >10% and
histopathological
changes within the liver
(enlarged centrilobular
hepatocytes with
numerous eosinophilic
fine granules in the
cytoplasm).
To overcome possible
problems with taste
aversion, animals in the
3 highest dose groups
received lower doses
for the first 1-2 wks of
exposure followed by
the final dose for the
remaining time.
Dow Chemical
Co., 1953a
Umedaetal.,
2004
Chronic studies
Rats, F344
(50/sex/group)
Diet
M: 0,36.4, 110, or
378
F: 0, 42.7, 128, or
438
2yrs
M: 110
F: 42.7
M:378
F: 128
M: Bladder tumors and
transitional cell
hyperplasia.
F: Nonneoplastic
kidney lesions (simple
transitional cell
hyperplasia in the renal
pelvis and hemosiderin
deposits).
Umedaetal.,
2002
69
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-18. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Rats, Wistar
(50/sex/group)
Rats, Wistar
(male, 25/group)
Rats, albino
(weanling,
15/sex/group)
Rats, albino
(male, 8/group)
Rats, Sprague-
Dawley
(12/sex/group)
Exposure route
Diet
Diet
Diet
Diet
Diet
Dose (mg/kg-d),
duration
M:0, 165, or 353
F: 0,178, or 370
75wks
Control groups:
basal diet for 2 wks
followed by
exposure at 0, 59.28,
or 248.3 for 34 wks
Exposure groups:
diet containing 0.1%
EHEN for 2 wks
followed by 0, 62,
or 248.2 for 34 wks
0, 1, 4, 8, 42, 84,
420, and 840
2yrs
0,250, or 450
13 mo
0,7, 73, or 732
2yrs
NOAEL
(mg/kg-d)
M:ND
F:ND
Control:
59.28
Exposure:
62
84
ND
73
LOAEL
(mg/kg-d)
M: 165
F: 178
Control:
248.3
Exposure:
248.2
420
250
732
Effect(s) at the
LOAEL
Formation of kidney
stones associated with
pyelonephritis in both
sexes.
Formation of kidney
stones associated with
pyelonephritis in both
sexes.
Kidney effects
including tubular
atrophy and dilation
associated with cyst
formation and calculi
formation in the renal
pelvis of both sexes.
Nonneoplastic
degenerative changes
in the liver, kidney,
thyroid, and
parathyroid resulting in
hyperplasia of liver,
kidney, and thyroid.
Renal effects (tubular
dilatation, calcification,
and intratubular
inflammation).
Comments
Biphenyl did not
exhibit tumor
promoting
characteristics for the
kidney tumor initiator,
EHEN, under the
conditions of this study.
Reference
Shiraiwaetal.,
1989
Ambrose et al.,
1960
Pecchiai and
Saffiotti, 1957
Dow Chemical
Co., 1953a
70
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-18. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Mice, BDFj
(50/sex/group)
Mice, ddY
(female, 34-
37/group)
Mice, hybrid
(2 strains,
1 8/sex/strain/group)
Dogs, Mongrel
Monkey, Rhesus
(2 M/dose, IF/dose)
Exposure route
Diet
Diet
Gavage (215
mg/kg body weight
in 0.5% gelatin) for
the first 3 wks,
followed by dietary
exposure for the
remaining time
Capsule in corn oil
Diet
Dose (mg/kg-d),
duration
M: 0,97, 291, or
1050
F:0, 134,414, or
1420
2yrs
0 or 855
2yrs
Oor91
18 mo
0,2.5 or 25
5 d/wk for 1 yr
0,0.01, 0.1, or 1%
for 1 yr
NOAEL
(mg/kg-d)
M:97
F: 134
855
91
ND
ND
LOAEL
(mg/kg-d)
M:291
F:414
ND
ND
ND
ND
Effect(s) at the
LOAEL
M: Decreased body
weight.
F: Nonneoplastic
effects (mineralization
in the kidney and
significantly increased
plasma ALT and AST
activities) in female
mice.
No adverse effects
observed at the highest
dose tested.
No evidence of a
carcinogenic response.
ND
ND
Comments
Two strains of Fl
hybrid mice were
produced by mating
female C57BL/6 mice
with either male
C3H/Anf mice or male
AKR mice.
Reference
Umedaetal.,
2005
Imaietal., 1983
Innesetal., 1969
NCI, 1968
Monsanto, 1956a
Dow Chemical
Co. 1953a
Reproductive and developmental studies
Rats, Wistar
(18-20/dose),
pregnant
Gavage in corn oil
0, 125, 250, 500 or
1,000 on CDs 6-15.
Dam: 500
Offspring:
250
Dam:
1000
Offspring:
500
Dam: maternal toxicity
(increased mortality),
increased in dead
fetuses and resorption.
Offspring: missing and
unossified sternebrae,
delayed calvarial
ossification.
Kheraetal., 1979
71
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-18. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Rats, Long Evans
(9 F/dose; 3 M/dose)
Rats, Albino (F/M)
Exposure route
Diet
Diet
Dose (mg/kg-d),
duration
M: 9, 89, or 887
F: 10, 101, or 1006
continuous breeding
0, 105, or 525
continuous breeding
NOAEL
(mg/kg-d)
M:ND
F: 101
ND
LOAEL
(mg/kg-d)
M:ND
F: 1006
ND
Effect(s) at the
LOAEL
M:ND
F: decreased fertility,
litter size, reduced fetal
growth rate.
ND
Comments
The effects seen in the
high dose group may
be associated with
unpalatability and
resultant decreased
food intake.
Reference
Dow Chemical
Co. 1953a
Ambrose et al.,
1960
a Report was not peer reviewed.
F = female; M = male; ND = not determined
72
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-19. Summary of major studies evaluating effects of biphenyl after inhalation exposure in rats and mice
Species, strain
Rabbits, albino
(3/dose)
Rats, Sprague-
Dawley (10/dose)
Rabbits, albino
(3/dose)
Rats, Sprague-
Dawley (6/dose)
Mice (12/dose)
Rats, Sprague-
Dawley (4/dose)
Mice, GDI
(50/sex/dose)
Dose (mg/m3), duration
300 mg/m3 (7 hours/day, 5 days/wk)
64 days over 94 days period
40 mg/m3 (7 hours/day, 5 days/wk)
46 days over 68 days period
5 mg/m3 (7 hours/day, 5 days/wk)
62 days over 92 days period
0, 157.7, or 315.3 mg/m3 (7 hours/day,
5 days/week), 13 weeks
NOAEL
(mg/m3)
ND
ND
ND
ND
LOAEL
(mg/m3)
ND
ND
5
157.7
Effect(s) at the LOAEL
ND
ND
Mice: upper respiratory tract irritation (acute
emphysema, congestion, edema, bronchitis, lobular
pneumonia, and multiple pulmonary abscesses)
Histopathologic lung, liver and kidney lesions
(congested and hemorrhagic lungs, tracheal
hyperplasia, and congestion and edema in the liver
and kidney) in both sexes.
References
Deichmannetal., 1947
Sun Company Inc.,
1977a
aReport was not published.
ND = not determined
73
DRAFT - DO NOT CITE OR QUOTE
-------
1 4.6.1. Oral
2 Biphenyl displays a relatively low acute oral toxicity, with LDso values in laboratory
3 animals in the 2-3 g/kg range (see Section 4.4.1). The major symptoms of biphenyl intoxication
4 typically associated with short-term, high-dose oral exposure of animals are labored breathing,
5 loss of body weight, and weakness. Following medium- or long-term oral exposure, reduced
6 body weight gain has been reported frequently (Umeda et al., 2005, 2004, 2002; Ambrose et al.,
7 1960; Dow Chemical Co., 1953) and attributed to low palatability of the feed (Ambrose et al.,
8 1960; Dow Chemical Co., 1953); however, the feed intake data of Umeda et al. (2005) in mice
9 did not support this notion. Increased liver and kidney weights were observed frequently
10 (Umeda et al., 2004, 2002; S0ndergaard and Blom, 1979; Ambrose et al., 1960; Monsanto, 1956;
11 Dow Chemical Co., 1953). A reduction in hemoglobin levels of rats receiving biphenyl for
12 700 days was reported (Ambrose et al., 1960). Signs of liver damage (increased serum activities
13 of ALT, AST, AP, and LDH) were observed in mice (Umeda et al., 2005). Pathological effects
14 on the urinary system dominated the spectrum of symptoms in dogs (Monsanto, 1956), rats
15 (Umeda et al., 2002; Dow Chemical Co., 1953), and mice (Umeda et al., 2005).
16 Urinary system effects, such as increased urine volume with increased specific gravity,
17 poly cystic changes, nephritis, and precipitation of free 4-OH-biphenyl and its glucuronide in
18 urine are commonly reported following oral exposure to biphenyl (Kluwe, 1982; S0ndergaard
19 and Blom, 1979; Monsanto, 1976; Booth et al., 1961). Calculi appeared in the urine of male rats
20 only (Umeda et al., 2002; Ohnishi et al., 2001, 2000a, b; Shibata et al., 1989b; Ambrose et al.,
21 1960). Urothelial hyperplasia with increased indices of cell proliferation have been described in
22 rats but not in mice and were attributed to irritation by calculi (Umeda et al., 2005, 2002; Shibata
23 et al., 1989b). Tubular dilatation and morphological changes in papillae and pelvis, kidney
24 stones, obstructive pyelonephritis, tubular atrophy, fibrosis, and pelvic hyperplasia were
25 observed (Shibata et al., 1989a, b; Shiraiwa et al., 1989; Takita, 1983; Kluwe, 1982; Booth et al.,
26 1961).
27 Increased incidences of fetuses with skeletal anomalies were reported following gavage
28 administration of biphenyl to Wistar rats during gestation (Kheraetal., 1979). Athree-
29 generation study in rats (Dow Chemical Co., 1953) found general reproductive toxicity at high
30 doses (about 947 mg/kg-day).
31
32 4.6.2. Inhalation
33 In a case study of workers engaged in the production of biphenyl-impregnated paper,
34 Hakkinen et al. (1973, 1971) observed liver damage (elevated levels of serum AST and ALT;
35 incipient cirrhosis and fatty changes in biopsy specimens) and effects on the central and
36 peripheral nervous systems (polyneuritic symptoms [abnormal EEGs and ENMGs], giddiness,
37 fatigue) that were attributed to long-term exposure to high concentrations of biphenyl. In one
38 fatal case, autopsy revealed kidney and bone marrow damage and heart muscle degeneration, as
74 DRAFT - DO NOT CITE OR QUOTE
-------
1 well as brain edema (Hakkinen et al., 1973, 1971). More recently the possibility has been
2 discussed that long-term exposure to biphenyl might contribute to the onset of PD (Wastensson
3 et al., 2006). The workplace conditions reported for these studies (Wastensson et al., 2006;
4 Hakkinen et al., 1973, 1971) suggested that inhalation represented the predominant route of
5 exposure, but dermal absorption as well as oral uptake (hand to mouth) might have occurred at a
6 significant level.
7 In mice, short-term biphenyl inhalation at concentrations as high as 54.75 ppm
8 (345.5 mg/m3) appeared to cause no symptoms (Sun Company Inc., 1977a). In another study,
9 3 rabbits, 4-6 rats, or 12 mice/group were exposed to biphenyl by inhalation for 7-13 weeks at
10 concentrations ranging from 5 to 300 mg/m3 (Deichmann et al., 1947). No adverse effects were
11 observed in rabbits, while rats and mice showed irritation of mucous membranes and succumbed
12 to high concentrations. Mice were far more sensitive than rats in these experiments, additionally
13 showing congestion and hemorrhage of the lungs (Deichmann et al., 1947). Repeated exposure
14 of mice to biphenyl at vapor concentrations of 25 or 50 ppm (157.75 or 315.5 mg/m3) for
15 13 weeks resulted in high incidences of pneumonia and tracheal hyperplasia, and high incidences
16 of congestion and edema in the lungs, liver, and kidney (Sun Company Inc., 1977b).
17 Reproductive or developmental studies using the inhalation route of exposure were not
18 identified.
19
20 4.6.3. Mode-of-Action Information
21 The studies described above have demonstrated that exposure to biphenyl may lead to a
22 variety of noncancer health effects (i.e., weight loss, liver toxicity, urinary tract toxicity).
23 However, there is not sufficient information to determine the mode of action for noncancer
24 health effects following exposure to biphenyl.
25 Weight loss or lack of weight gain has been consistently associated with oral exposure to
26 biphenyl (Umeda et al., 2005, 2002; Ambrose et al., 1960; Dow Chemical Co., 1953). The work
27 of Nishihara (1985) provides a possible explanation for this toxic effect. This author found that,
28 in vitro, biphenyl can act as an uncoupler of respiration. It may be speculated that long-term,
29 high-dose exposure to biphenyl uncouples mitochondrial respiration to a certain extent, resulting
30 in a futile cycle that diverts the use of nutrients from building body mass into maintaining
31 necessary energy stores. It is not clear at what level of in vivo exposure this effect might become
32 operative.
33 Several of the oral animal studies (Umeda et al., 2005; Sun Company Inc., 1977b;
34 Pecchiai and Saffiotti, 1957; Dow Chemical Co., 1953; Deichmann et al., 1947) and the
35 epidemiological study by Hakkinen et al. (1973) provide evidence that the liver is a target for
36 biphenyl toxicity by any route of exposure. This evidence consists of changes in blood
37 parameters that are indicative of liver toxicity; however, in animal studies, liver histopathology
38 does not support or explain this finding. Evidence for damage to the nervous system, as
75 DRAFT - DO NOT CITE OR QUOTE
-------
1 suggested by Hakkinen et al. (1973) and Seppalainen and Hakkinen (1975), has not been
2 reproduced in animal studies. The limited evidence for an estrogenic activity of
3 4,4'-dihydroxybiphenyl (Kitamura et al., 2003; Schultz et al., 2002) is insufficient to assign a
4 clear endocrine-disrupting effect to this important metabolite of biphenyl.
5 Damage to the urinary tract has been observed consistently in animals but not in humans.
6 The work of Ohnishi et al. (2001, 2000a, b) provides tenable evidence that, in the rat, this is due
7 to the precipitation in the urinary tract of crystals consisting mostly of 4-hydroxybiphenyl.
8 These crystals irritate the epithelia of ureters and bladder, leading to chronic inflammation and
9 possibly cancer as well as obstruction of the urinary tract with subsequent hydronephrosis. The
10 work of Ohnishi et al. (2001, 2000b) has made it clear that, at least in their animal model, two
11 conditions are required for this event to occur: (1) the pH in the urine of the animals needs to be
12 higher than normal and (2) elevated potassium levels need to accompany the elevated pH
13 because it is the potassium salt of 4-hydroxybiphenyl sulphate that has the lowest solubility in
14 high-pH urine. No damage to the urinary tract was observed in rabbits exposed via inhalation to
15 biphenyl for up to 13 weeks (Deichmann et al., 1947). Although this mode of action is likely to
16 explain the effects of biphenyl in the urinary tract of rats, it is unclear whether or not it has any
17 bearing on humans that are likely exposed by inhalation.
18 Gombar et al. (1991) developed structure activity relationship computer models for four
19 types of chemical compounds (carboaromatic, heteroaromatic, alicyclic, acyclic) to estimate the
20 teratogenic potential of 171 compounds (for which teratogenic data exist in >900 publications) in
21 an overall procedure (dosage, maternal toxicity, and affected organ systems were not factored
22 into these preliminary models). The models considered species, route of administration, and
23 duration and timing of exposure. Experimental endpoints entered into the model were number of
24 dams; maternal toxicity; teratogenic endpoints; numbers of viable implants, resorptions, and
25 abnormal fetuses; and dead/live fetus ratio. Fetal deaths per se, runting, delayed ossification, and
26 minor skeletal abnormalities such as extra or missing ribs were not rated as teratogenic effects.
27 The computerized modeling uses a coding system that represents only "heavy" atoms (i.e., no
28 hydrogens). The models included molecule fragments and their electronic descriptors to
29 represent functional groups, molecular shape descriptors, and connectivity descriptors. The
30 results of the calculations were presented as 24 different structural descriptor values. After
31 eliminating two types of results (outliers and "statistically influential"), the models returned a
32 96% correct classification of the teratogenic potential of chemicals. Biphenyl and
33 2-hydroxybiphenyl were negative in this computerized evaluation.
34
35 4.7. EVALUATION OF CARCINOGENICITY
36 4.7.1. Summary of Overall Weight of Evidence
37 Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the
38 database for biphenyl provides "suggestive evidence of carcinogenic potential." This cancer
76 DRAFT - DO NOT CITE OR QUOTE
-------
1 weight-of-evidence descriptor is based on urinary bladder tumors (transitional cell papillomas
2 and carcinomas) in male F344 rats (Umeda et al., 2002) and liver tumors (hepatocellular
3 adenomas and carcinomas) in female BDFi mice (Umeda et al., 2005) exposed to biphenyl in the
4 diet for 104 weeks. Earlier chronic toxicity and carcinogenicity assessments in orally exposed
5 animals found no clear evidence of biphenyl-induced carcinogenicity in rats (Shiraiwa et al.,
6 1989; Ambrose et al., 1960; Pecchiai and Saffiotti, 1957; Dow Chemical Co., 1953), mice (Imai
7 et al., 1983; Innes et al., 1969; NCI, 1968), dogs (Monsanto, 1956), or Rhesus monkeys (Dow
8 Chemical Co., 1953). The earlier studies had limitations including small numbers of animals in
9 exposure groups and shorter-than-lifetime durations of exposure due to design or decreased
10 survival unrelated to tumor development, with the exception of a mouse study that found no
11 evidence of carcinogenic responses in female ddY mice exposed to 5,000 ppm biphenyl in the
12 diet for 2 years (Imai et al., 1983).
13 Considerable evidence suggests that the development of urinary bladder tumors in male
14 rats exposed to biphenyl depends on the sustained occurrence of urinary bladder calculi
15 composed of precipitated 4-hydroxybiphenyl-O-sulphate, based on: (1) close correlation
16 between urinary bladder calculi formation and development of urinary bladder tumors in male
17 rats exposed to biphenyl, (2) dose-response and temporal concordance between biphenyl-induced
18 urinary calculi formation, regenerative hyperplasia, and urinary bladder tumor development,
19 (3) an overall negative or equivocal weight of evidence for the genotoxicity of biphenyl and
20 metabolites, and (4) the wide body of evidence that other nongenotoxic or weakly genotoxic
21 chemicals produce urinary bladder tumors in rodents at high exposure levels by a mode of action
22 involving calculi formation, followed by ulceration or inflammation and regenerative cell
23 proliferation (IARC, 1999b). Mode-of-action information is sufficient to conclude that these
24 tumors are high-dose phenomena; without the development of calculi, urinary bladder tumors are
25 not expected. The proposed mode of action is expected to be relevant to humans at exposure
26 levels sufficient to cause urinary bladder calculi in humans, because calculi in humans have been
27 associated with urinary bladder irritation, regeneration, and cancer (IARC, 1999b; Cohen, 1998,
28 1995) and the metabolism of biphenyl to sulphate conjugates of hydroxylated biphenyl
29 metabolites has been demonstrated in human tissues.
30 For liver tumors, a proposed mode of action (Umeda et al., 2004) includes activation of
31 peroxisome proliferator activated receptors (PPARs) by biphenyl or its metabolites in liver cells
32 or direct or indirect (through ROS) reactions with DNA in liver cells to produce mutations
33 leading to tumor initiation. However, available data are insufficient to establish a mode of
34 action for liver tumors in female mice (See Section 4.7.3.2.2.1 for more information). In the
35 absence of information to indicate otherwise, the development of liver tumors in female mice
36 with chronic exposure to biphenyl is assumed to be relevant to humans. EPA acknowledges that
37 some mouse strains (e.g., B6C3Fi) are relatively susceptible to liver tumors and the background
38 incidence of this tumor can be high. For these reasons, use of mouse liver tumor data in risk
77 DRAFT - DO NOT CITE OR QUOTE
-------
1 assessment has been a subject of controversy (King-Herbert and Thayer, 2006). The BDFi
2 mouse used in the Umeda et al. (2005) bioassay is a cross between female C57BL/6 and male
3 DBA/2 mice (Charles River Laboratories International, Inc., 1999), both of which are considered
4 to be relatively resistant to liver tumor induction (Maronpot, 2009). In the Umeda et al. (2005)
5 bioassay, the incidences of tumors in male and female concurrent control mice were 32 and 6%,
6 respectively. The relatively low background incidence of liver tumors in female control mice
7 from Umeda et al. (205) minimizes the possible confounding of compound-related liver tumors
8 in this sex.
9 The descriptor of "suggestive evidence of carcinogenic potential" is appropriate when the
10 weight of evidence is suggestive of carcinogenicity, i.e., a concern for potential carcinogenic
11 effects in humans is raised, but the data are judged not sufficient for a stronger conclusion (U.S.
12 EPA, 2005a). As discussed in Section 4.2.1.2, biphenyl exposure produced an increased
13 incidence of urinary bladder tumors in male F344 rats (Umeda et al., 2002) and liver tumors in
14 female BDFi mice (Umeda et al., 2005). Such data could be considered consistent with the
15 descriptor of "likely to be carcinogenic to humans." As stated in the Guidelines for Carcinogen
16 Risk Assessment (U.S. EPA, 2005a), a "likely" descriptor may include "an agent that has tested
17 positive in animal experiments in more than one species, sex, strain, site, or exposure route, with
18 or without evidence of carcinogenicity in humans." Biphenyl did induce tumors in two species
19 (rat and mouse) and at two sites (liver and urinary bladder); however, tumor findings across the
20 biphenyl database and the interpretation of some of these findings indicate some uncertainties
21 regarding the potential human carcinogenicity of biphenyl.
22 Both the liver tumors and urinary bladder tumors induced by dietary exposure to biphenyl
23 each occurred in only one sex and only one species. Liver tumors were induced in female BDFi
24 mice only, and urinary bladder tumors occurred in male F344 rats only. The incidence of liver
25 adenomas and carcinomas (separate and combined) in Umeda et al. (2002) was increased over
26 control in all groups of exposed female mice; however, the liver tumor incidence plateaued at the
27 mid- and high-dose groups (incidence of adenoma and carcinoma combined in the control and
28 low-, mid-, and high-dose groups were 3/48, 8/50, 16/49, and 14/48, respectively). Further,
29 female ddY mice exposed to 5000 ppm biphenyl in the diet for 2 years showed no increased
30 incidence of liver tumors (Imai et al., 1983). Urinary bladder tumors in F344 male rats induced
31 by dietary biphenyl exposure appear to be a high-dose phenomenon closely related to the
32 formation of calculi. A mode of action analysis (see Section 4.7.3.1) supports the conclusion
33 that exposures that do not lead to urinary bladder calculi will not produce tumors. While the
34 proposed mode of action for urinary bladder tumors in male rats is considered relevant to
35 humans, there is evidence that humans are likely to be less susceptible to these tumors than rats.
36 As discussed in Section 4.7.3.1.4.2, the rodent horizontal quadruped stature is expected to
37 promote calculi residency time in the bladder without causing obstruction, whereas the anatomy
38 of the urinary tract in humans and their upright bipedal stature result in more ready excretion of
78 DRAFT - DO NOT CITE OR QUOTE
-------
1 calculi in the urine or therapeutic removal of calculi that form obstructions (Cohen and Ellwein,
2 1992; Matanowki, 1981). Overall, the mode of action analysis suggests that biphenyl is not
3 likely to induce urinary bladder tumors in humans at environmental exposure levels. In light of
4 the above considerations related to biphenyl-induced female mouse liver tumors and male rat
5 bladder tumors, EPA concluded that the currently available information is most consistent with a
6 determination that the database for biphenyl provides "suggestive evidence of carcinogenic
7 potential."
8 U.S. EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) indicate that
9 for tumors occurring at a site other than the initial point of contact, the cancer descriptor may
10 apply to all routes of exposure that have not been adequately tested at sufficient doses. An
11 exception occurs when there is convincing toxicokinetic data that absorption does not occur by
12 other routes. Information available on the carcinogenic effects of biphenyl demonstrates that
13 tumors occur in tissues remote from the site of absorption following chronic oral exposure
14 (urinary bladder in male rats and liver in female mice). No information on the carcinogenic
15 effects of biphenyl via the inhalation or dermal routes in humans and animals is available.
16 Quantitative data demonstrating rapid and extensive absorption of biphenyl are restricted to the
17 oral route of exposure; a case report of hepatic toxicity produced by a probable combination of
18 inhalation and dermal exposures in a worker in a biphenyl-impregnated fruit wrapping paper
19 production facility provides qualitative evidence of absorption by these routes (Hakkinen et al.,
20 1973). Therefore, based on the observance of systemic tumors following oral exposure and
21 assumed absorption by all routes of exposure, it is assumed that an internal dose will be achieved
22 regardless of the route of exposure. Therefore, EPA considers the biphenyl database to provide
23 "suggestive evidence of carcinogenic potential" by all routes of exposure.
24
25 4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence
26 Available human studies were not designed to evaluate associations between exposure to
27 biphenyl and occurrence of cancer (see Section 4.1).
28 As discussed in Section 4.2, carcinogenicity studies in animals are limited to the oral
29 exposure route. In well-designed cancer bioassays of F344 rats (Umeda et al., 2002) and BDFi
30 mice (Umeda et al., 2005), dietary exposure to biphenyl resulted in the occurrence of urinary
31 bladder tumors in male rats and significantly increased incidences in liver tumors in female mice.
32 Earlier chronic toxicity and carcinogenicity assessments found no clear evidence of
33 biphenyl-induced carcinogenicity in orally exposed rats, mice, dogs, or Rhesus monkeys.
34 However, these studies were generally limited in design, with the exception of a study reporting
35 no evidence of carcinogenic responses in female ddY mice (n = 34 mice vs. 37 control mice)
36 exposed to 5,000 ppm biphenyl in the diet for 2 years (Imai et al., 1983). In a study of Wistar
37 rats, sufficient numbers of animals (50/sex/group) were exposed to biphenyl in the diet at
38 concentrations up to 5,000 ppm, but only for 75 weeks (Shiraiwa et al., 1989). Some of the male
79 DRAFT - DO NOT CITE OR QUOTE
-------
1 rats exhibited urinary bladder calculi and simple or diffuse hyperplasia and papillomatosis of the
2 urinary bladder mucosa in the absence of neoplastic lesions, but the study may have been
3 terminated prior to eventual urinary bladder tumor development. Ambrose et al. (1960) exposed
4 albino rats (15/sex/exposure level) to biphenyl in the diet at concentrations up to 10,000 ppm for
5 2 years (10, 50, 100, 500, 1,000, 5,000, or 10,000 ppm); however, decreased survival in rats
6 exposed to 5,000 or 10,000 ppm, presumably from decreased food consumption, and the
7 relatively small numbers of animals in each exposure group may have impaired the ability to
8 detect late-developing tumors. In another study, groups of Sprague-Dawley rats (12/sex/group)
9 received biphenyl in the diet at concentrations up to 10,000 ppm for up to 2 years (Dow
10 Chemical Co., 1953). However, this study suffered from a pneumonia outbreak, particularly
11 among control males, and the relatively small numbers of animals and the decreased survival
12 may have impaired the ability to detect late-developing tumors. A study of male albino rats
13 included small numbers of rats (8/group) and a short (13 months) exposure period (Pecchiai and
14 Saffiotti, 1957). A study of B6C3Fi or B6AkFi mice exposed to biphenyl in the diet for only
15 18 months (Innes et al., 1969; NCI, 1968) included relatively small numbers of mice
16 (18/sex/group) and only one exposure level (517 ppm) that was similar to the concentration
17 (667 ppm) without carcinogenic effect in the Umeda et al. (2005) 24-month BDFi mouse
18 bioassay. The dog study included two males and one female, a high dose of 25 mg/kg-day, and
19 an exposure period of only 1 year (Monsanto, 1956). Rhesus monkeys (two males and one
20 female) were exposed to biphenyl in the diet at a concentration of 10,000 ppm, but for only
21 1 year (Dow Chemical Co., 1953).
22 The overall weight of evidence for biphenyl genotoxicity from short-term tests is
23 negative or equivocal. Biphenyl did not induce mutations in a variety of bacterial test systems,
24 but both negative and positive results were obtained in mammalian in vitro test systems (see
25 section 4.5.6. for references). Single oral doses of 2,000 mg biphenyl/kg induced DNA damage
26 (detected by the Comet assay) in several organs of CD-I mice (including the liver and bladder),
27 but it is uncertain if the damage was due to a direct effect on DNA or was an indirect effect due
28 to cytotoxicity or ROS generated by redox cycling of phenylhydroquinone, a major urinary
29 metabolite of 2-hydroxybiphenyl and a minor metabolite of biphenyl in rats.
30 The overall weight of evidence for 2-hydroxybiphenyl genotoxicity suggests that
31 oxidative DNA damage from ROS from redox cycling between 2,5-dihydroxybiphenyl and
32 phenylbenzoquinone is possible. DNA damage was detected in liver and bladder of CD-I mice
33 exposed to 2,000 mg/kg of 2-hydroxybiphenyl (Sasaki et al., 2002, 1997) and in the urinary
34 bladder of male F344 rats fed the sodium salt of 2-hydroxybiphenyl at 1 or 2% in the diet for 3-
35 5 months (Morimoto et al., 1989). DNA adducts were detected by [32P]-post labeling in skin of
36 CD-I mice after topical application of the sodium salt of 2-hydroxybiphenyl or phenylhydro-
37 quinone (Pathak and Roy, 1993), and increased micronuclei were detected in urinary bladder
38 epithelium of male F344 rats exposed to 2,000 ppm 2-hydroxybiphenyl or 2,000 ppm NaCl plus
80 DRAFT - DO NOT CITE OR QUOTE
-------
1 2,000 ppm 2-hydroxybiphenyl in the diet for 2 weeks (Balakrishnan et al., 2002). However,
2 increased binding of radioactivity to DNA was not detected in DNA extracted from urinary
3 bladder epithelium of male F344 rats exposed to single gavage doses of 2-hydroxybiphenyl as
4 high as 1,000 mg/kg (Kwok et al., 1999), and DNA adducts were not detected in urinary bladder
5 epithelium of male F344 rats exposed for 13 weeks to biphenyl dietary concentrations as high as
6 12,500 ppm (Smith et al., 1998). The mechanism by which 2-hydroxybiphenyl may induce
7 micronuclei in the urinary bladder epithelium is uncertain, but could involve micronuclei
8 generation as a secondary response to cytotoxicity or regenerative cell proliferation, DNA
9 damage from ROS from redox cycling of 2,5-dihydroxybiphenyl, or protein modifications
10 leading to mitotic spindle interference or inhibition of enzymes important in DNA replication
11 (Balakrishnan et al., 2002). The hydroxylation of biphenyl to produce 2-hydroxybiphenyl is a
12 minor pathway in rats and mice (Halpaap-Wood et al., 1981a, b; Meyer and Scheline, 1976).
13 2-Hydroxybiphenyl and 2,5-dihydroxybiphenyl collectively accounted for less than 2% of
14 metabolites in urine of rats administered single oral doses of 100 mg biphenyl/kg (Meyer and
15 Scheline, 1976) or single i.p. doses of 30 mg biphenyl/kg (Halpaap-Wood et al., 1981a). In mice
16 given i.p. doses of 30 mg biphenyl/kg, these metabolites accounted for less than 5% of urinary
17 metabolites (Halpaap-Wood et al., 1981a).
18
19 4.7.3. Mode-of-Action Information
20 4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats
21 4.7.3.1.1. Hypothesized mode of action. The best-supported hypothesis proposes a mode of
22 action whereby the formation of urinary bladder calculi (from the precipitation of 4-
23 hydroxybiphenyl-O-sulphate) is a key event in the development of urinary bladder tumors in
24 male rats fed high levels of biphenyl in the diet for 2 years. According to this hypothesis, the
25 calculi (occurring in association with increased urinary pH and potassium, and predominantly
26 composed of 4-hydroxybiphenyl-O-sulphate) cause irritation to transitional epithelial cells of the
27 urinary bladder leading to sustained cell proliferation, which promotes the development of
28 initiated cells in the urinary bladder with progression to papillomas and carcinomas.
29
30 4.7.3.1.2. Experimental support for the hypothesized mode of action
31 4.7.3.1.2.1. Strength, consistency, and specificity of association, including support for the
32 hypothesized mode of action in male rats. The formation of urinary bladder calculi,
33 predominantly composed of potassium 4-hydroxybiphenyl-O-sulphate, is strongly, consistently,
34 and specifically associated with the formation of urinary bladder tumors in male rats chronically
35 exposed to high dietary concentrations of biphenyl. Several findings support this association.
36 Urinary bladder calculi were formed at a high prevalence (43/50; 86%) in a group of male rats
37 exposed to biphenyl in the diet at a concentration of 4,500 ppm, but were absent in male rats
38 receiving diets containing 0, 500, or 1,500 ppm biphenyl (Umeda et al., 2002). These
81 DRAFT - DO NOT CITE OR QUOTE
-------
1 observations were consistent with the detection of urinary bladder transitional cell papilloma
2 (10/50; 20%), carcinoma (24/50; 48%), and papilloma or carcinoma (31/50; 62%) in the
3 4,500 ppm group of male rats and total absence of urinary bladder papilloma or carcinoma in the
4 control, 500, or 1,500 ppm groups of male rats. Bladder calculi were found in all 24 of the male
5 rats with urinary bladder transitional cell carcinoma and in 8/10 of the male rats with transitional
6 cell papilloma.
7 The association between urinary bladder calculus formation and development of urinary
8 bladder tumors is both gender and species specific. Urinary bladder calculi, of similar size to
9 those observed in males, were observed at much lower incidence (8/50; 16%) in the 4,500 ppm
10 female rats, but they were of more uniform color (white and yellow versus white, yellow, brown,
11 gray, and black in males) and shape (spheroidal vs. triangular, pyramidal, cubical, and spheroidal
12 in males) and primarily composed of 4-hydroxybiphenyl and potassium bisulphate (which are
13 hydrolysis products of potassium 4-hydroxybiphenyl-O-sulphate) (Umeda et al., 2002; Ohnishi
14 et al., 2000b). No urinary bladder calculi were found in the 500 and 1,500 ppm groups of female
15 rats. Transitional cell hyperplasia was found in 10/50 4,500-ppm female rats, but no urinary
16 bladder transitional cell papillomas or carcinomas were seen in any of the biphenyl-exposed
17 groups of female rats. Furthermore, there was no evidence of biphenyl-induced urinary bladder
18 calculi or bladder tumors in male or female BDFi mice receiving dietary biphenyl at
19 concentrations as high as 6,000 ppm for 2 years (Umeda et al., 2005).
20 Urinary bladder calculi in male rats were associated with significantly increased urinary
21 pH (average pH of 7.97 in the 4,500 ppm group at the final week of exposure compared to
22 7.66 in controls) (Umeda et al., 2002) and were composed primarily of potassium
23 4-hydroxybiphenyl-O-sulphate (Ohnishi et al., 2000b). The urine pH of female rats exposed to
24 4,500 ppm for 104 weeks (pH = 7.26) was not elevated compared with controls (pH = 7.29)
25 (Umeda et al., 2002), and urinary calculi of a different composition than male rats (i.e.,
26 4-hydroxybiphenyl and potassium bisulphate, compared with potassium 4-hydroxybiphenyl-
27 O-sulphate in males) were found in only 8/50 4,500-ppm females (Ohnishi et al., 2000b). From
28 these observations, it appears that the formation of the calculi results from the precipitation of the
29 potassium salt of the sulphate conjugate of 4-hydroxybiphenyl under the elevated pH conditions
30 of the male rat urine. The mechanism responsible for increased urinary pH is unknown, although
31 Ohnishi et al. (2001, 2000a, b) proposed that gender differences in urinary conditions, such as
32 pH and potassium concentrations, and sulphatase activities in kidneys, may be responsible for
33 the gender differences in urinary calculi composition and formation and the subsequent
34 development of urinary bladder tumors in male, but not female, F344 rats.
35 Relatively strong, consistent, and specific associations between calculi formation and
36 transitional cell hyperplasia and between transitional cell hyperplasia and the development of
37 transitional cell tumors in the urinary bladder have been shown in male F344 rats chronically
38 exposed to high concentrations of biphenyl in the diet. Urinary bladder transitional cell
82 DRAFT - DO NOT CITE OR QUOTE
-------
1 hyperplasia (simple, nodular, papillary) occurred in 45/50 (90%) male rats receiving biphenyl in
2 the diet for 2 years at the same dietary concentration (4,500 ppm) at which high prevalences of
3 both urinary bladder calculi formation (43/50; 86%) and transitional cell tumors (31/50 62%)
4 were observed (Umeda et al., 2002). Forty-two of the 45 male rats with urinary bladder
5 transitional cell hyperplasia also exhibited urinary bladder calculi. In another study, evidence of
6 biphenyl-induced calculi formation (microcalculi in the urine) and increased indices of urinary
7 bladder transitional cell proliferation (greater than fourfold increase in BrdU incorporation) in
8 male F344 rats has been reported following as little as 4-8 weeks of dietary exposure to
9 5,000 ppm biphenyl (Shibata et al., 1989b).
10 The most convincing evidence that degenerative changes in the urinary bladder
11 epithelium lead to tumor formation is the site-concordance of associations between calculi
12 formation in the urinary bladder, transitional cell proliferation, transitional cell hyperplasia, and
13 transitional cell tumors (Umeda et al., 2002). In addition, the strong associations between
14 urinary tract calculi formation, ulcerations or inflammation, and subsequent hyperplasia
15 combined with repeated, high-level exposure to other chemicals that cause urinary bladder
16 tumors in rodents, including melamine, uracil, and the sodium salt of 2-hydroxybiphenyl (IARC
17 1999a, b, c; Cohen, 1998; 1995) provide further evidence that degenerative changes are involved
18 in the etiology of rodent urinary bladder tumors. It is not unusual to see extensive proliferation
19 or hyperplasia in bladder epithelium in response to urinary calculi from other rodent bladder
20 tumorigens without an associated ulceration or intense inflammatory response. In male rats
21 exposed to 4,500 ppm biphenyl, increasing numbers of rats with clinical hematuria were
22 observed beginning at about the 40th week of exposure, and histologic examinations at study
23 termination revealed focal hyperplasia in 45/50 rats, providing some evidence of calculi-induced
24 bladder epithelial damage followed by cell proliferation (Umeda et al., 2002). Over the course of
25 the study, 94% of male rats with hematuria had bladder or kidney calculi, but hematuria was not
26 found in any biphenyl-exposed females. In addition, with 8 weeks, but not 4 weeks, of exposure
27 to 5,000 ppm biphenyl in the diet, moderate urinary bladder epithelial hyperplasia and
28 microcalculi in urine were observed in 5/5 male F344 rats, but no descriptions of degenerative
29 changes were provided; these observations are consistent with a rapid repair response to
30 epithelial damage from biphenyl-induced urinary tract calculi (Shibata et al., 1989b).
31 The ability of repeated biphenyl exposure to promote previously initiated urinary bladder
32 cells to bladder tumors is supported by results of a bladder tumor initiation-promotion study
33 (Kurata et al., 1986). Incidences of urinary bladder hyperplasia, papilloma, and carcinoma were
34 significantly increased in male F344 rats initiated with dietary BBN for 4 weeks followed by
35 5,000 ppm biphenyl in the diet for 32 weeks, compared with rats receiving BBN only for
36 4 weeks. For example, 94 and 83% of rats treated with BBN followed by biphenyl developed
37 urinary bladder hyperplasia and papillomas, respectively, compared with 25 and 12% of rats
38 exposed to BBN alone.
83 DRAFT - DO NOT CITE OR QUOTE
-------
1 The hypothesis that the mode of action involves the development of urinary bladder
2 tumors in biphenyl-exposed male rats is further supported by an overall negative or equivocal
3 weight of evidence for the genotoxicity of biphenyl. As discussed earlier, there are consistently
4 negative results for biphenyl in bacterial mutation assays and inconsistent positive results for
5 biphenyl in in vitro mammalian assays mostly in the presence of metabolic activation. There is
6 evidence that 2,5-dihydroxybiphenyl (i.e., phenylhydroquinone), the principal urinary metabolite
7 in rats exposed to high doses of 2-hydroxybiphenyl, can undergo redox cycling to produce ROS
8 that may damage DNA and lead to tumor-initiating mutations; however, 2-hydroxybiphenyl is a
9 minor urinary metabolite of biphenyl in rats and 2,5-dihydroxybiphenyl was not detected in urine
10 of rats exposed to oral doses of 100 mg biphenyl/kg (Meyer and Scheline, 1976).
11
12 4.7.3.1.2.2. Dose-response concordance. Dose-response relationships for urinary bladder
13 calculi formation, transitional cell hyperplasia, and transitional cell tumor development show
14 concordance in the 2-year oral study of rats (Umeda et al., 2002). In male rats, urinary calculi,
15 nonneoplastic lesions (epithelial hyperplasia), and neoplastic lesions (papillomas and
16 carcinomas) of the urinary bladder were observed only at the highest exposure level
17 (4,500 ppm); no urinary bladder calculi, transitional cell hyperplasia, or transitional cell tumors
18 were found in control, 500, or 1,500 ppm male rats. Furthermore, urinary bladder calculi were
19 found in 43/45 high-dose male rats, in all 24 male rats with transitional cell carcinoma, and in
20 8/10 of the male rats with transitional cell papilloma.
21
22 4.7.3.1.2.3. Temporal relationship. Results from the 2-year oral study in rats (Umeda et al.,
23 2002) provide some evidence of a progression from urinary bladder calculi formation to the
24 development of bladder tumors. Urinary bladder calculi were observed in the first 4,500 ppm
25 male rat that died (week 36), evidence of blood in the urine was observed in 4,500 ppm male rats
26 by week 40, and incidences of bladder calculi and bloody urine that paralleled increases in
27 mortality and tumor formation were observed throughout the remainder of the study. In addition,
28 results of a short-term oral study demonstrate that microcalculi can be detected in the urine of
29 male rats after as little as 4 weeks of dietary exposure to 5,000 ppm biphenyl and that
30 hyperplasia of urinary bladder epithelium can be detected at least by week 8 (Shibata et al.,
31 1989b). Presumably, the development of biphenyl-induced urinary bladder tumors requires a
32 longer exposure period to urinary calculi of sufficient size, shape, and composition to induce
33 urinary bladder epithelial damage and a sustained proliferative response.
34
35 4.7.3.1.2.4. Biological plausibility and coherence. The proposed mode of action is consistent
36 with the current understanding of cancer biology and is supported by the wide body of evidence
37 that other chemicals with primarily nongenotoxic profiles produce urinary bladder tumors in
38 rodents at high exposure levels by a mode of action involving calculi formation, ulceration or
84 DRAFT - DO NOT CITE OR QUOTE
-------
1 inflammation, and regenerative cell proliferation ( IARC, 1999a, b, c; Cohen, 1998, 1995).
2 Additional information could strengthen the plausibility and coherence of the proposed mode of
3 action to explain the occurrence of biphenyl-induced urinary bladder tumors in male rats. These
4 additional data include results from investigations of earlier time points in the proposed temporal
5 progression from calculi formation to epithelial damage, regenerative cell proliferation, and
6 tumor development and further investigations into the factors underlying gender-specific
7 differences in precipitation of 4-hydroxybiphenyl-O-sulphate to form bladder calculi in rats.
8
9 4.7.3.1.3. Other possible modes of action for bladder tumors in male rats. Although the
10 weight of evidence from short-term standard genotoxicity tests with biphenyl and
11 4-hydroxybiphenyl is predominantly negative, evidence is available that suggests that oral
12 exposure to high doses of 2-hydroxybiphenyl is associated with the development of urinary
13 bladder tumors in male rats. The induction of genotoxic effects in the urinary bladder epithelium
14 leading to tumor initiation is proposed to occur via redox cycling between 2,5-
15 dihydroxybiphenyl and phenylbenzoquinone (Balakrishnan et al., 2002; Kwok et al.,1999;
16 Pathak and Roy, 1993; Morimoto et al., 1989). However, the strong, consistent, and specific
17 association between the occurrence of urinary bladder calculi composed of 4-hydroxybiphenyl-
18 O-sulphate and development of urinary bladder tumors in male but not female rats, the evidence
19 that 2-hydroxybiphenyl is a minor urinary metabolite of biphenyl and, finally, that 2,5-
20 dihydroxybiphenyl was not detected in the urine of biphenyl-exposed rats, demonstrate that the
21 support for a genotoxic mode of action involving key mutational events from biphenyl or its
22 metabolites in the urinary bladder leading to initiation of tumor cells is not compelling.
23 Additional support for a proposed genotoxic mode of action would come from studies showing
24 formation of 2,5-dihydroxylbiphenyl and phenylbenzoquinone in the urinary bladder epithelium
25 of rats exposed to low doses of biphenyl.
26
27 4.7.3.1.4. Conclusions about the hypothesized mode of action for bladder tumors in male rats.
28 4.7.3.1.4.1. Support for the hypothesized mode of action in rats. There is strong evidence that
29 urinary bladder tumors in male rats chronically exposed to biphenyl in the diet is a high-dose
30 phenomenon involving sustained occurrence of calculi in the urinary bladder leading to
31 transitional cell damage, sustained regenerative cell proliferation, and eventual promotion of
32 spontaneously initiated tumor cells in the urinary bladder epithelium.
33 To summarize, chronic exposure of male rats to a high dietary concentration of biphenyl
34 (4,500 ppm) caused increased urinary pH and high prevalence of urinary bladder calculi (from
35 the precipitation of 4-hydroxybiphenyl-O-sulphate in the urine), transitional cell hyperplasia, and
36 transitional cell tumors. Incidences of male rats with calculi and those with bladder tumors were
37 strongly correlated, and chronic exposure of male rats to lower dietary concentrations of
38 biphenyl (500 and 1,500 ppm) did not increase urinary pH and did not cause calculi formation,
85 DRAFT - DO NOT CITE OR QUOTE
-------
1 transitional cell hyperplasia, or bladder tumor development. There were relatively strong
2 associations between incidences of rats with calculi and those with transitional cell hyperplasia
3 and between incidences of rats with transitional cell hyperplasia and bladder tumors. In contrast,
4 high concentrations of biphenyl in the diet of female rats had no effect on urinary pH, caused a
5 much lower prevalence of urinary bladder calculi of a different composition, and resulted in no
6 urinary bladder tumors. The urinary bladder calculi in the male rats were mainly composed of
7 the conjugated biphenyl metabolite, potassium 4-hydroxybiphenyl-O-sulphate, whereas those of
8 the female rats were predominantly composed of 4-hydroxybiphenyl and potassium bisulphate
9 (which are hydrolysis products of potassium 4-hydroxybiphenyl-O-sulphate). There was no
10 evidence of urinary bladder calculi formation or tumor development in male and female mice
11 exposed to similar dietary concentrations of biphenyl. Results of a tumor initiation-promotion
12 study in male rats support the proposal that biphenyl-induced sustained cell proliferation
13 promotes initiated tumor cells in the urinary bladder. Finally, results of genotoxicity tests with
14 biphenyl are predominantly negative or equivocal at best. The preponderance of evidence
15 supports a mode of action for biphenyl in male rats only involving urinary tract calculi
16 formation, urinary epithelium damage, sustained regenerative cell proliferation and hyperplasia,
17 and subsequent bladder tumor formation. There is evidence that 2,5-dihydroxybiphenyl can
18 undergo redox cycling to produce ROS that may damage DNA leading to tumor-initiating
19 mutations, but it was not detected in urine of rats exposed to oral doses of 100 mg biphenyl/kg
20 and its metabolic precursor, 2-hydroxybiphenyl, is a minor urinary metabolite of biphenyl in rats
21 (Meyer and Scheline, 1976).
22
23 4.7.3.1.4.2. Relevance of the hypothesized mode of action to humans. Although there are no
24 studies in humans examining possible associations of biphenyl exposure with urinary bladder
25 calculi formation or cancer, urinary bladder calculi have been reported in humans following
26 exposure to other chemicals (IARC, 1999b; Cohen 1998, 1995). Urinary bladder calculi are, in
27 general, expected to be irritating and lead to reparative cell proliferation regardless of
28 composition or species; however, based on the anatomy of the urinary tract in humans and their
29 upright, bipedal stature, calculi are either quickly excreted in urine or cause obstruction leading
30 to pain and subsequent therapeutic removal of the calculi (Cohen, 1998, 1995). In contrast, the
31 rodent horizontal quadruped stature is expected to promote calculi residency time in the bladder
32 without causing obstruction (Cohen, 1998, 1995). In white populations, 95% of bladder tumors
33 are transitional cell carcinomas such as those found in male rats exposed to high concentrations
34 of biphenyl. IARC (1999b) noted that several case-control studies of urinary bladder cancer in
35 white human populations found relative risks for an association between a history of urinary tract
36 stones and bladder carcinomas ranging from about 1.0 to 2.5, suggesting a causative link. Thus,
37 the proposed mode of action is expected to be relevant to humans at exposure levels sufficient to
38 cause urinary bladder calculi in humans, because: (1) calculi resulting from human exposure to
86 DRAFT - DO NOT CITE OR QUOTE
-------
1 other substances have been associated with urinary bladder irritation, regeneration, and cancer
2 (IARC, 1999b; Cohen 1998, 1995) and (2) sulphate conjugation of hydroxylated biphenyl
3 metabolites has been demonstrated in human tissues (as briefly reviewed in Section 3.3).
4 The underlying physiological factors determining the precipitation of 4-hydroxybiphenyl-
5 O-sulphate in urine to form calculi in male rats, but not female rats, exposed to high dietary
6 biphenyl concentrations are unknown. Given this lack of understanding for rats and the absence
7 of specific human data on biphenyl-induced calculi or urinary stones, there is uncertainty in
8 extrapolation of the dose-response relationship for biphenyl-induced calculi formation in male
9 rats to humans.
10
11 4.7.3.1.4.3. Populations or lifestases particularly susceptible to the hypothesized mode of
12 action. IARC (1999b) noted that increased risks for bladder carcinoma in humans have been
13 associated with cigarette smoking, exposure to infectious agents, such as Shistosoma
14 haematobium, causing urinary tract inflammation, and a history for urinary tract infections in
15 general. As such, people with these types of exposure or history may be particularly susceptible
16 to the formation of urinary calculi and urinary bladder cancer, but evidence supporting this
17 inference is lacking. In addition, there are conditions (bladder diverticuli, neurogenic bladder,
18 and staghorn renal pelvic calculi) that can increase the residency time of calculi in humans; thus,
19 individuals with these conditions may also be particularly susceptible to biphenyl-induced
20 bladder tumors under the hypothesized mode of action.
21
22 4.7.3.2. Mode-of-Action Information for Liver Tumors in Female Mice
23 Evidence that chronic oral exposure to biphenyl can cause liver tumors comes from the 2-
24 year BDFi mouse bioassay by Umeda et al. (2005). Exposure to 2,000 or 6,000 ppm biphenyl in
25 the diet, but not to 667 ppm, produced increased incidences of hepatocellular adenomas or
26 carcinomas in female mice, but no carcinogenic response in male BDFi mice. Earlier studies
27 found no carcinogenic response in B6C3Fi or B6AkFi mice exposed to 517 ppm biphenyl in the
28 diet for 18 months (Innes et al., 1969; NCI, 1968) or in ddY female mice exposed to 5,000 ppm
29 biphenyl in the diet for 2 years (Imai et al., 1983). The only investigations into the mode of
30 action for biphenyl-induced liver tumors in mice involve examinations of indicators of
31 peroxisome proliferation following biphenyl exposure (Umeda et al., 2004; Sunouchi et al.,
32 1999). Thus, a mode of action involving PPARs is proposed and an evaluation of the supporting
33 data follows.
34
35 4.7.3.2.1. Hypothesized mode of action for liver tumors in female mice. Proliferation of
36 peroxisomes is regulated by a class of ligand-activated transcription factors known as PPARs.
37 PPARa regulates induction of the peroxisome proliferation response in rodents and is thought to
38 mediate at least some of the responses for hepatocarcinogens, including initiation of cellular
87 DRAFT - DO NOT CITE OR QUOTE
-------
1 events leading to transformation. Peroxisome proliferators (PPARa agonists) are a structurally
2 diverse group of non- or weakly mutagenic chemicals that induce a suite of responses including
3 the induction of tumors in rats and mice (Klaunig et al., 2003).
4 Klaunig et al. (2003) have proposed a mode of action for PPARa agonists involving the
5 following key events. PPARa agonists activate PPARa to transcribe genes involved in
6 peroxisome proliferation, cell cycling/apoptosis, and lipid metabolism. The changes in gene
7 expression lead to changes in cell proliferation and apoptosis, and to peroxisome proliferation.
8 Suppression of apoptosis coupled with increased cell proliferation allows transformed cells to
9 persist and proliferate, resulting in preneoplastic hepatic foci and ultimately promotion of tumor
10 growth via selective clonal expansion. Peroxisome proliferation may lead to oxidative stress,
11 which potentially contributes to the proposed mode of action by causing indirect DNA damage
12 and/or by causing cytotoxicity leading to reparative cell proliferation. PPARa agonists also
13 inhibit gap junction intercellular communication and stimulate non-parenchymal hepatic Kupffer
14 cells; these events are also thought to stimulate cell proliferation. Increases in the size and
15 number of peroxisomes and induction of peroxisome-related gene expression (e.g., palmitoyl-
16 CoA oxidase and acyl-CoA oxidase) are regarded as indicators that the PPARa agonism mode of
17 action is operative.
18
19 4.7.3.2.2. Experimental support for the hypothesized mode of action for liver tumors in female
20 mice.
21 4.7.3.2.2.1. Strength, consistency, specificity of association, including support for the
22 hypothesized mode of action in mice. There is limited support for a possible association
23 between biphenyl-induced proliferation of peroxisomes and liver tumors, because the following
24 findings were reported in female BDFi mice (which developed liver tumors following dietary
25 exposure to 2,000 or 6,000 ppm) but not in male BDFi mice (which did not develop liver tumors
26 following exposure to concentrations as high as 6,000 ppm biphenyl). Dietary exposure of
27 female BDFi mice to 16,000 ppm biphenyl for 13 weeks induced hepatocellular peroxisomes as
28 evidenced by light microscopy detection of enlarged hepatocytes filled with eosinophilic fine
29 granules and electron microscopy confirmation that the granules corresponded to increased
30 numbers of peroxisomes (Umeda et al., 2004). Significantly increased activities were measured
31 for potassium cyanide-insensitive palmitoyl CoA oxidation in liver homogenate (up to 1.9-fold)
32 and lauric acid 12-hydroxylation in liver microsomes (up to 3.8-fold) from female BDFi mice
33 given oral doses up to 5.2 mmol/kg-day (800 mg/kg-day) for 3 days (Sunouchi et al., 1999).
34 The available data do not demonstrate strong, consistent, or specific associations between
35 key events in the proposed mode of action and the development of liver tumors in female mice
36 exposed to biphenyl. Klaunig et al. (2003) proposed that an adequate data set to support a
37 PPARa agonism mode of action should meet the following demonstration criteria, most of which
88 DRAFT - DO NOT CITE OR QUOTE
-------
1 as noted in parentheses have not been investigated for biphenyl or its metabolites: (1) activation
2 of PPARa (no data), (2) expression of peroxisomal genes including PPARa-mediated expression
3 of cell cycle, growth, and apoptosis, and nonperoxisomal lipid gene expression (no data),
4 (3) peroxisomal proliferation (limited data for biphenyl in mice as summarized in previous
5 paragraph) and perturbation of cell proliferation and apoptosis (no data for mouse liver),
6 (4) inhibition of gap junction intercellular communication (no data), (5) hepatocyte oxidative
7 stress (no data), (6) Kupffer cell-mediated events (no data), and (7) selective clonal expansion
8 (no data).
9
10 4.7.3.2.2.2. Dose-response concordance. The available data do not show concordance between
11 the dose-response relationships for liver tumors in female BDFi mice exposed for 2 years to
12 biphenyl in the diet (liver tumors at 2,000 or 6,000 ppm, but not 667 ppm; Umeda et al., 2005)
13 and liver peroxisome proliferation, the only key event in the proposed mode of action that has
14 been investigated. Umeda (2004) reported that, compared with controls, increased liver
15 peroxisomes were detected in female BDFi mice exposed to 16,000 ppm biphenyl in the diet for
16 13 weeks, but not in mice exposed to 500, 2,000, 4,000, 8,000, or 10,000 ppm.
17
18 4.7.3.2.2.3. Temporal relationship. Indicators of liver peroxisome proliferation were elevated
19 in female mice, but not male mice, with oral exposure durations of 3 days following exposure to
20 800 mg/kg-day (increased activities of potassium cyanide-insensitive palmitoyl CoA oxidation
21 and lauric acid 12-hydroxylation; Sunouchi et al. 1999) and 13 weeks following exposure to
22 16,000 ppm in the diet, but not at lower dietary concentrations (increased numbers of liver
23 peroxisomes; Umeda et al. 2004).
24
25 4.7.3.2.2.4. Biological plausibility and coherence. The data are inadequate to evaluate the
26 biological plausibility and coherence of the proposed mode of action as it relates to liver tumors
27 in female mice exposed to biphenyl.
28
29 4.7.3.2.3. Other possible modes of action for liver tumors in mice. As discussed in
30 Section 4.5.5, the overall weight of evidence from short-term standard genotoxicity tests with
31 biphenyl and 4-hydroxybiphenyl is predominantly negative. A genotoxic mode of action for
32 biphenyl-induced liver tumors in mice could be proposed based on the large metabolic capacity
33 of the mouse liver to convert biphenyl to hydroxylated metabolites and evidence that metabolites
34 of 2-hydroxybiphenyl (2,5-dihydroxybiphenyl and 2,5'-benzoquinone) can produce DNA
35 damage (Tani et al., 2007; Balakrishnan et al., 2002; Sasaki et al. 2002, 1997; Pathak and Roy,
36 1993; Morimoto et al., 1989). However, hydroxylation of biphenyl to produce 2-hydroxy-
37 biphenyl appears to be a minor metabolic pathway in mice administered single intraperitoneal
38 doses of 30 mg biphenyl/kg (Halpaap-Wood et al., 1981a), and the available data are inadequate
89 DRAFT - DO NOT CITE OR QUOTE
-------
1 to establish that this genotoxic mode of action operates in the biphenyl induction of liver tumors
2 in mice. There have been no in vitro or in vivo investigations of biphenyl-induced DNA adducts
3 or ROS generation in mouse liver cells or of possible gender differences in the production of
4 biphenyl-induced DNA adducts or other genotoxic events. Current mode-of-action information
5 is inadequate to provide plausible explanations for why female BDFi mice exposed to high
6 dietary concentrations of biphenyl develop liver tumors, but male BDFi mice exposed to
7 6,000 ppm and female ddY mice exposed to 5,000 ppm do not (Umeda et al., 2005; Imai et al.,
8 1983).
9
10 4.7.3.2.4. Conclusions about the hypothesized mode of action for liver tumors in mice.
11 A PPARa agonism mode of action for liver tumors in female mice exposed to 2,000 or 4,000
12 ppm biphenyl in the diet for 2 years is not adequately supported by the experimental data. This
13 is based on the lack of concordance between dose-response relationships for biphenyl-induced
14 liver tumors and proliferation of hepatocellular peroxisomes in female mice. Evidence for
15 increased hepatocellular peroxisomes in female mice was only found with 13-week exposure to
16 16,000 ppm biphenyl and not at several concentrations <10,000 ppm (Umeda et al., 2004).
17 Furthermore, a series key events demonstrating PPARa agonism mode of action have not been
18 identified.
19 Available data are inadequate to support alternative modes of action that propose direct
20 or indirect genotoxic events from reactive biphenyl metabolites or ROS, respectively, as key
21 events. Results from standard short-term genotoxicity tests are mostly negative or equivocal for
22 biphenyl and 4-hydroxybiphenyl. Although there is some evidence for DNA damage from ROS
23 generated from redox cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone, there
24 are no investigations into the metabolic formation of 2-hydroxybiphenyl, 2,5-dihydroxybiphenyl,
25 and phenylbenzoquinone in livers of biphenyl-exposed mice exposed to a range of biphenyl
26 doses, no in vitro or in vivo investigations of biphenyl-induced DNA adducts or ROS generation
27 in mouse liver cells, and no investigations of possible gender differences in capability to produce
28 biphenyl-induced DNA adducts or other genotoxic events.
29
30 4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
31 4.8.1. Possible Childhood Susceptibility
32 No specific information was identified that would point specifically towards an early
33 childhood susceptibility for biphenyl toxicity. However, the developmental profiles of
34 superoxide dismutase and catalase in humans that were reported by McElroy et al. (1992)
35 indicate that the activities of both enzymes may be comparatively low before and at birth,
36 placing humans in the perinatal period at an increased risk of adverse effects elicited by quinoid
37 metabolites of biphenyl. Specifically, Buonocore et al. (2001) drew attention to the fact that the
90 DRAFT - DO NOT CITE OR QUOTE
-------
1 human brain has relatively low superoxide dismutase activity at birth. Given the limited data on
2 age-specific ROS scavenging enzymes, any suggestions of childhood susceptibility to biphenyl is
3 speculative.
4 Studies in animals provide evidence that biphenyl metabolism is mediated by CYP1A2
5 and CYP3A4 (Haugen, 1981). Phase II enzymes, such as sulphotransferases (SULTs) and
6 uridine diphosphate glucuronosyl transferases (UGTs), may be involved in conjugation activities
7 with hydroxybiphenyls in mammalian tissues (Pacifici et al., 1991; Bock et al., 1980). CYP1A2
8 expression is negligible in the early neonatal period, but is significantly increased to 50% of
9 adult levels by one year of age (Sonnier and Cresteil, 1998). In general, SULTs and UGTs,
10 depending on the isoforms, also exhibit differential expression during human development
11 (Duanmu et al., 2006; Strassburg et al., 2002). To the extent that metabolism increases or
12 reduces the toxicity of biphenyl, changes in the expression of Phase I and II enzymes during
13 development can influence susceptibility to biphenyl toxicity. Specific isoforms of cytochrome
14 P450s and Phase II enzymes have not been identified as the principal catalyzers involved in
15 biphenyl metabolism and the effect of differences in enzyme expression on childhood
16 susceptibility to biphenyl has not been established.
17
18 4.8.2. Possible Gender Differences
19 Benford and Bridges (1983) evaluated the sex- and tissue-specific induction of biphenyl
20 2-, 3-, and 4-hydroxylase activities in microsomal preparations or primary hepatocyte cultures
21 from male and female Wistar rats. No differences in biphenyl hydroxylase activities were
22 observed between the sexes. However, there were some sex differences in the way tissues
23 responded to the action of enzyme inducers. For example, the CYP1A inducer a-naphthoflavone
24 strongly induced 2-hydroxylase in male liver but had no effect on female liver. Betamethasone
25 induced 2-hydroxylase activity in female liver but inhibited it in male liver. The available
26 limited human data do not suggest that gender differences exist in the response to biphenyl
27 exposure. However, available animal data suggest gender-related differences in susceptibility to
28 tumors (i.e., bladder tumors in male but not female F344 rats and increased incidences of liver
29 tumors in female but not male BDFi mice administered biphenyl in the diet for a lifetime).
30
31 4.8.3. Other
32 The limited information on the specifics of biphenyl metabolism and toxic effects in
33 humans does not allow a meaningful assessment of populations that might be highly susceptible
34 to the adverse effects of biphenyl. For example, there is as yet no clear attribution of CYP
35 isozymes to the various biphenyl hydroxylases and no information on which sulphotransferases
36 and glucuronidases conjugate hydroxylated biphenyl metabolites. It is known that many CYP
37 isozymes, as well as glucuronidases, exist in polymorphic forms with catalytic activities that
91 DRAFT - DO NOT CITE OR QUOTE
-------
1 differ from the wild type. In addition, such enzyme polymorphisms display specific distributions
2 across populations and ethnicities that might put some at increased risk and others at decreased
3 risk of adversity from biphenyl exposure. This lack of information represents a data gap.
92 DRAFT - DO NOT CITE OR QUOTE
-------
1 5. DOSE-RESPONSE ASSESSMENTS
2
3
4 5.1. ORAL REFERENCE DOSE (RfD)
5 5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
6 No information was located regarding possible associations between oral exposure to
7 biphenyl and health outcomes in humans.
8 As discussed in Section 4.6.1, the major and most sensitive targets of toxicity following
9 oral exposure to biphenyl are the liver, urinary system, body weight, and developing organism
10 (see Figure 5-1). In the rat, chronic oral studies identified the kidney and urinary bladder as
11 critical noncancer targets (see Figure 5-1 for LOAELs and NOAELs found in these studies).
12 Kidney effects observed include: renal pelvis transitional cell hyperplasia and hemosiderin
13 deposits in female F344 rats at doses >128 mg/kg-day (Umeda et al., 2002); kidney stone
14 formation and obstructive pyelonephritis with tubular atrophy, tubular cysts, and fibrosis in male
15 and female Wistar rats at 165 and 370 mg/kg-day, respectively (Shiraiwa et al., 1989); renal
16 lymphocytic infiltration, tubular atrophy, and tubular cysts in male and female albino rats at
17 doses >420 mg/kg-day (Ambrose et al., 1960); mild renal tubular degeneration in male albino
18 rats at 250 or 450 mg/kg-day (Pecchiai and Saffioti, 1957; not plotted in Figure 5-1 because
19 quantitative data were not included in the study report); and renal tubular dilatation in male and
20 female Sprague-Dawley rats at 732 mg/kg-day (Dow Chemical Co., 1953). An increased
21 incidence of urinary bladder hyperplasia associated with calculi or "stones" was observed in
22 male and female F344 rats at 378 and 438 mg/kg-day, but not at 110 and 128 mg/kg-day,
23 respectively (Umeda et al., 2002). Elevated incidences of the same lesion were observed in male
24 and female Wistar rats at 353 and 370 mg/kg-day, respectively (Shiraiwa et al., 1989). In
25 contrast, urinary bladder hyperplasia and calculi were not observed in male or female albino rats
26 at doses as high as 840 mg/kg-day (Ambrose et al., 1960) or in male or female Sprague-Dawley
27 rats exposed to doses as high as 732 mg/kg-day (Dow Chemical Co., 1953).
93 DRAFT - DO NOT CITE OR QUOTE
-------
1600 4
1400 -j
1200 -j
1000 -
800 -
600 -
400 » i
A
i
200 -j 1
n 1________
A
I
_™™™™™™™»™~™J
Drop-lines span highest
I
A
i
1 ,
I™™™™™™™™™™™™™™™™™™™™™™™™™™
:o lowest doses tested
I '
I
I
i
m
' j
i
1
i
ANOAEL BLOAEL
.
'
1
k
I
i k
I
|
i
|
k
Female
BDFi
mouse
(1)
Female
ddY
mouse
(2)
Liver3
Male
BDFi
mouse
(1)
Female
ddY
mouse
(2)
Body weightb
Female
BDFi
mouse
(1)
Female
ddY
mouse
(2)
Female
F344 rat
(3)
Male
Wistar rat
(4)
Male,
female
albino rat
(5)
Male,
female
Sprague-
Dawley rat
(6)
Kidney0
Male
F344 rat
(3)
Male,
female
albino rat
(5)
Male
Wistar rat
(4)
Male,
female
Sprague-
Dawley rat
(6)
Urinary bladderd
Wistar rat
(7)
Develop-
mental6
Increased plasma liver enzymes in BDF1 mice.
bDecreased body weight (>10% lower than controls) in BDF1 mice.
Increased incidences of kidney lesions including: mineralization in outer medulla in BDF1 mice; renal pelvis transitional cell hyperplasia
and hemosiderin deposits in F344 rats; kidney stone formation in Wistar rats; renal tubular atrophy in albino rats; renal tubular dilatation in
Sprague-Dawley rats.
Increased incidences of urinary bladder calculi or stones and hyperplasia in F344 rats and Wistar rats.
Increased number of litters with fetal skeletal anomalies in Wistar rats.
(1) = Umeda et al., 2005; (2) = Imai et al., 1983; (3) = Umeda et al., 2002; (4) = Shiraiwa et al., 1989; (5) = Ambrose et al., 1960; (6) = Dow
Chemical Co., 1953; (7) = Khera et al., 1979
Figure 5-1. NOAELs and LOAELs for noncancer effects in rats and mice from repeated oral exposure to biphenyl.
94
DRAFT - DO NOT CITE OR QUOTE
-------
1
2 In mice, chronic oral toxicity studies identified the liver, kidney, and body weight as
3 critical noncancer targets (see Figure 5-1 for NOAELs and LOAELs for these effects). In
4 mice, significantly (p < 0.05) increased plasma levels of enzymes indicative of liver damage
5 were observed at dose levels of 1,050 mg/kg-day in males and >414 mg/kg-day in females
6 (Umeda et al., 2005), but no exposure-related changes in liver enzymes were observed in female
7 ddY mice at 885 mg/kg-day (Imai et al., 1983). Significantly increased incidence of
8 mineralization of the renal outer medulla was observed in female BDFi mice at 414 and
9 1,420 mg/kg-day (Umeda et al, 2005), but exposure-related histological changes in the kidney
10 were not found in female ddY mice at 885 mg/kg-day (Imai et al., 1983). Following the same
11 pattern of apparent strain difference in susceptibility to biphenyl toxicity, body weights were
12 decreased by >10% in male BDFi mice at >291 mg/kg-day and females at >414 mg/kg-day
13 (Umeda et al., 2005), but body weights in female ddY mice exposed to 885 mg/kg-day were
14 similar to control values (Imai et al., 1983). Shorter duration oral exposure (13 weeks) of mice
15 to biphenyl at higher dietary concentrations (estimated doses >1,500 mg/kg-day) has also been
16 shown to affect body and/or liver weights in mice (Umeda et al., 2004).
17 In the only available oral developmental toxicity study (Khera et al., 1979), frank
18 maternal toxicity (increased mortality [5/20 vs. 0/18 in controls] and decreased number of dams
19 with live fetuses [9/20 vs. 16/18 in controls]) occurred at the highest dose (1,000 mg/kg-day).
20 Significantly increased incidences of fetuses with skeletal anomalies were noted at doses
21 >500 mg/kg-day. The NOAEL and LOAEL of 250 and 500 mg/kg-day for delayed skeletal
22 development are noted in Figure 5-1.
23 The 2-year dietary studies in F344 rats (Umeda et al., 2002) and BDFi mice (Umeda et
24 al., 2005) and the developmental study in Wistar rats (Khera et al., 1979) were selected as
25 candidate principal studies for deriving the RfD because they provide the best available data
26 (adequate number of dose groups and dose spacing, sufficient group sizes, comprehensive
27 endpoint assessment and quantitation of results) to describe dose-response relationships for the
28 critical effects in rats and mice associated with chronic or gestational oral exposure to biphenyl.
29 In the 2-year dietary study of male and female F344 rats, biphenyl was administered in
30 the diet at 0, 500, 1,500, or 4,500 ppm (respective estimated doses were 36.4, 110, and
31 378 mg/kg-day for males and 42.7, 128, and 438 mg/kg-day for females) (Umeda et al., 2002).
32 At the highest dose, noncancer effects included significantly increased incidence of rats with
33 transitional cell hyperplasia in the renal pelvis, renal mineralization and hemosiderin deposits,
34 and urinary bladder transitional cell hyperplasia. Noncancer effects at the mid-dose level were
35 restricted to significantly increased incidences of females with renal transitional cell hyperplasia
36 and hemosiderin deposits. There were no significant biphenyl-related effects in low-dose males
37 or females.
95 DRAFT - DO NOT CITE OR QUOTE
-------
1 In the 2-year dietary study of male and female BDFi mice, biphenyl was administered in
2 the diet at 0, 667, 2,000, or 6,000 ppm (respective estimated doses were 0, 97, 291, and
3 1,050 mg/kg-day for males, and 0, 134, 414, and 1,420 mg/kg-day for females) (Umeda et al.,
4 2005). At the two highest dose levels, noncancer effects included increased incidence of mice
5 with renal mineralization, increased levels of BUN, increased levels of serum enzymes indicative
6 of liver damage, and decreased terminal body weights. No exposure-related effects were
7 observed at the lowest exposure level.
8 In the oral developmental toxicity study, pregnant Wistar rats were exposed by gavage to
9 0, 125, 250, 500, or 1,000 mg biphenyl/kg-day on GDs 6-15 (Khera et al., 1979). Significantly
10 increased numbers of fetuses with skeletal anomalies (wavy ribs, extra ribs, small 13th rib,
11 missing or unossified sternebrae, delayed ossification of the calvarium) were noted at doses
12 >500 mg/kg-day, and the number of litters exhibiting any of these anomalies was significantly
13 higher at the 500 mg/kg-day dose level relative to controls.
14 Candidate critical effects from the chronic study in F344 rats (Umeda et al., 2002) were:
15 (1) nodular or simple transitional cell hyperplasia in the renal pelvis of males and females,
16 (2) mineralization in the renal pelvis or renal papillary mineralization in males and females,
17 (3) renal hemosiderin deposits in females, and (4) transitional cell hyperplasia in the urinary
18 bladder of males. Candidate critical effects from the chronic study in BDFi mice (Umeda et al.,
19 2005) were: (1) decreased body weight in males and females, (2) mineralization of the renal
20 inner stripe-outer medulla in males and females, (3) BUN in males and females, and (4) serum
21 liver enzyme activities (AST [GOT], ALT [GPT], AP [ALP], and LDH) in females. The
22 candidate critical effect from the rat oral developmental toxicity study (Khera et al., 1979) was
23 litters with fetal skeletal anomalies from Wistar rat dams exposed during gestation.
24
25 5.1.2. Methods of Analysis—Including Models
26 Dichotomous datasets modeled include selected nonneoplastic lesions in the urinary
27 system of male and female F344 rats (Table 5-1) exposed to biphenyl in the diet for 2 years
28 (Umeda et al., 2002), mineralization in the kidney of male and female BDFi mice (Table 5-2)
29 exposed to biphenyl in the diet for 2 years (Umeda et al., 2005), and litters with skeletal
30 anomalies from Wistar rat dams (Table 5-3) administered biphenyl by gavage on GDs 6-15
31 (Khera etal., 1979).
32
96 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-1. BMD modeling datasets for incidences of nonneoplastic effects in
the urinary tract of male and female F344 rats exposed to biphenyl in the
diet for 2 years
Biphenyl dietary concentration (ppm)
TWA body weight (kg)a
Calculated dose (mg/kg-d)b
Effect
Renal pelvis
Nodular transitional cell hyperplasia
Simple transitional cell hyperplasia
Mineralization
Other kidney effects
Hemosiderin depositf
Papillary mineralization
Males (n = 50)
0
0.411
0
500
0.412
36.4
1,500
0.408
110
4,500
0.357
378
Females (n = 50)
0
0.251
0
500
0.246
42.7
1,500
0.246
128
4,500
0.216
438
0
6
9
1
8
6
1
5
10
21C
19d
18e
0
3
12
0
5
12
1
12d
18
12C
25C
27d
0
9
0
9
0
14
0
23d
4
2
8
6
22C
3
25C
12C
Bladder
Combined transitional cell hyperplasia8
0
0
0
45
1
0
1
10
aTWA body weight calculated using graphically-presented body weight data in the study report of Umeda et al.
(2002).
bCalculated doses based on calculated TWA body weights and chronic reference food consumption values for
F344 rats (0.030 and 0.021 kg/day for males and females, respectively; taken from Table 1-6 of U.S. EPA, 1988).
Significantly different from control group (p < 0.01) according to %2 test.
dSignificantly different from control group (p < 0.05) according to %2 test.
Significantly different from controls (p < 0.05) according to Fisher's exact test.
fMale data for incidences of hemosiderin deposits not selected for quantitative analysis.
8Female data for incidences of combined transitional cell hyperplasia not selected for quantitative analysis.
Source: Umeda et al. (2002).
97
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-2. BMD modeling datasets for body weight, selected clinical
chemistry results, and histopathological kidney effects in male and female
BDFi mice exposed to biphenyl in the diet for 2 years
Endpoint
Biphenyl concentration in the diet (ppm)
0
667
2,000
6,000
Males
Dose (mg/kg-d)
Histopathological kidney effect
Mineralization inner stripe-outer medulla
Clinical chemistry parameter
BUN (mg/dL)
Body weight
Mean terminal body weight (g)
0
n = 50
9
n = 34
20.2 ±3.6
n = 35
46.9 ±4.9
97
n=49
8
n=39
22.0 ±4.0
n=41
43.1 ±7.9
291
n=50
14
n=37
23.2±4.4a
n=41
42.9±6.0a
1,050
n=50
14
n=37
22.9±2.7b
n=39
32.4±3.6b
Females
Dose (mg/kg-d)
Histopathological kidney effect
Mineralization inner stripe-outer medulla
Clinical chemistry parameter
AST (IU/L)
ALT (IU/L)
AP (IU/L)
LDH (IU/L)
BUN (mg/dL)
Body weight
Mean terminal body weight (g)
0
n = 50
o
J
n = 28
75 ±27
32 ±18
242 ± 90
268 ± 98
14.9 ±2.0
n = 31
34.0 ±4.0
134
n=50
5
n=20
120 ±110
56 ±46
256 ±121
461 ±452
14.8 ±3.4
n=22
32.5 ±3. 3
414
n=50
12C
n=22
211±373b
134±231b
428 ± 499
838 ± 2,000
21.0 ±20.5
n=25
30.5 ± 3. lb
1,420
n=49
26d
n=31
325 ± 448b
206 ± 280b
556 ± 228b
1,416 ±4,161a
23.8±11.7b
n=32
25.5±3.0b
"Significantly different from controls (p < 0.05) according to Dunnett's test.
bSignificantly different from controls (p < 0.01) according to Dunnett's test.
Significantly different from controls (p < 0.05) according to Fisher's exact test.
dSignificantly different from controls (p < 0.01) according to Fisher's exact test.
ALT (GPT) = alanine aminotransferase (glutamic pyruvic transaminase); AP (ALP) = alkaline phosphatase;
AST (GOT) = aspartate aminotransferase (glutamic oxaloacetic transaminase)
Source: Umeda et al. (2005).
98
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-3. BMD modeling dataset for incidence of litters with fetal skeletal
anomalies from Wistar rat dams administered biphenyl by gavage on
CDs 6-15
Effect
Litters with fetal skeletal anomaliesVlitters
examined
Dose (mg/kg-d)
0
8/16
125
11/20
250
13/18
500
15b/18
1,000
6/9
aThe study authors reported one runted fetus in the control group and one fetus with kinky tail in the 250 mg/kg-day
dose group, which may have influenced the reported incidence data for anomalous litters/litters examined.
bSignificantly different from controls (p < 0.05) according to Fisher's exact test conducted for this review.
Source: Kheraetal. (1979).
1
2 All available core dichotomous models in the EPA Benchmark Dose Software (BMDS)
3 (version 2.1.2) were fit to the incidence data for each dataset. The multistage model was run for
4 all polynomial degrees up to n-1 (where n is the number of dose groups including control).
5 Adequate model fit was judged by three criteria: goodness-of-fit^-value (p > 0.1), visual
6 inspection of the dose-response curve, and a value of <2 for the largest scaled residual for any
7 data point in the dataset (including the control). Among all of the models providing adequate fit
8 to the data, the lowest BMDL was selected as the potential point of departure (POD) when the
9 difference between the BMDLs estimated from these models was more than threefold; otherwise,
10 the BMDL from the model with the lowest Akaike's Information Criterion (AIC) was chosen as
11 the candidate POD. In accordance with U.S. EPA (2000b) guidance, BMDs and BMDLs
12 associated with an extra risk of 10% were calculated for all models. In the absence of
13 information to identify the biologically significant level of response for an endpoint, a
14 (benchmark response) BMR of 10% extra risk is typically chosen as a response level for
15 dichotomous data and is recommended for the BMR when using dichotomous models to
16 facilitate a consistent basis of comparison across assessments and endpoints.
17 A BMR of 10% extra risk was selected to derive the POD for development effects from
18 the Khera et al. (1979) study because the endpoints were characterized as affected litters. A
19 BMR of 5% extra risk has typically been used for quantal data in reproductive and
20 developmental studies when data are available to characterize individual pups within litters (U.S.
21 EPA, 2000b). Since this level of reporting was not available, nested models could not be used.
22 Thus, a BMR of 10% extra risk among affected litters was employed in order to better
23 approximate a 5% extra risk in affected offspring and to recognize the litter as the experimental
24 unit. BMDs and BMDLs associated with extra risk of 5% for all endpoints were also calculated
25 for comparison.
26 When core models failed to provide adequate fit to the data, optimizations of the models
27 (model restriction adjustments, specification of initial parameters, and use of alternative models)
99
DRAFT - DO NOT CITE OR QUOTE
-------
1 were attempted in an effort to achieve adequate fit. If these optimizations failed to achieve better
2 fit, the highest dose was dropped and the entire modeling procedure was repeated. If an adequate
3 fit could not be achieved after dropping the highest dose, then the dataset was determined to be
4 unsuitable for BMD modeling.
5 For continuous data, all core continuous models available in the EPA BMDS
6 (version 2.1.2) were first applied to the data while assuming constant variance. If the data were
7 consistent with the assumption of constant variance (p>0.1), then the fit of all the continuous
8 models to the mean were evaluated while assuming constant variance. In the absence of
9 information to indicate the biologically significant level of response, BMDs and BMDLs were
10 calculated based on a benchmark response (BMR) representing a change of 1 SD from the
11 control. BMDs and BMDLs for decreased body weight were also calculated for a BMR of 10%
12 decrease from the control (i.e., 10% relative deviation [RD]) because a 10% decrease in body
13 weight is generally considered to represent a minimally biologically significant effect. For
14 serum enzyme activities (AST, ALT, AP, LDH), BMDs and BMDLs were also calculated for a
15 BMR of 100% increase from the control (i.e., twofold or 1 RD; BMDmD and BMDLmD).
16 Several expert organizations, particularly those concerned with early signs of drug-induced
17 hepatotoxicity, have identified an increase in liver enzymes (AST, ALT, AP) compared with
18 concurrent controls of two- to fivefold as an indicator of concern for hepatic injury (EMEA,
19 2006; Boone et al., 2005). Because LDH, like liver enzymes, is one of the more specific
20 indicators of hepatocellular damage in most animal species and generally parallels changes in
21 liver enzymes in toxicity studies where liver injury occurs, a similar twofold increase in LDH is
22 considered indicative of liver injury in experimental animals. A similar approach was taken for
23 BUN.
24 Adequate model fit was judged by three criteria: goodness-of-fit^-value (p > 0.1), visual
25 inspection of the dose-response curve, and a value of <2 for the largest scaled residual for any
26 data point in the data set (including the control). Among all of the models providing adequate fit
27 to the data, the lowest BMDL was selected as the potential POD when the BMDLs estimated
28 from these models varied by more than threefold; otherwise, the BMDL from the model with the
29 lowest AIC was chosen as the candidate POD. When the test for constant variance was negative,
30 all models were run again while applying the power model integrated into the BMDS to account
31 for nonhomogeneous variance. When the nonhomogeneous variance model provided an
32 adequate fit (p > 0.1) to the variance data, the models were evaluated using the nonhomogeneous
33 variance model. Model fit and POD selection proceeded as described earlier. When both tests
34 for variance (constant and nonhomogeneous) provided inadequate fit to the variance data, model
35 restriction adjustments were attempted in an effort to achieve adequate fit. If these
36 manipulations failed to achieve better fit, the highest dose was dropped and the entire modeling
37 procedure was repeated. If an adequate fit could not be achieved after dropping the highest dose,
38 then the dataset was determined to be unsuitable for BMD modeling.
100 DRAFT - DO NOT CITE OR QUOTE
-------
1 Summary modeling results are presented in Table 5-4 and Figure 5-2; more detailed
2 modeling results are presented in Appendix B (Tables B-4 through B-24 and respective model
3 output files). The BMDs and BMDLs shown in Table 5-4 and Figure 5-2 are those from the
4 best-fitting models for each endpoint. For datasets to which no model could be fit, NOAELs and
5 LOAELs were considered for the candidate POD.
6
101 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-4. Summary of BMDs/BMDLs for selected nonneoplastic effects
following oral exposure of rats and mice to biphenyl
Males
Best fitting
model
Benchmark result
(mg/kg-d)
Females
Best fitting
model
Benchmark result
(mg/kg-d)
F344 rats (Umeda et aL, 2002); biphenyl in the diet for 2 yrs
Kidney
BMD10
BMDL10
BMD10
BMDL10
Renal pelvis
Transitional cell nodular
hyperplasia
Transitional cell simple
hyperplasia
Mineralization
Multistage
3 -degree
Gamma
Log-probit
193
314
208
127
113
138
Multistage
2-degree
Gamma
Multistage
1 -degree
274
71
88
212
52
56
Kidney - other
Hemosiderin deposit
Papillary mineralization
Bladder
Transitional cell hyperplasia
Not selected13
Multistage
1 -degree
Gamma
-
92
BMD10
205
-
58
BMDL10
147
Dichotomous-Hill
Logistic
Not selected13
45
292
BMD10
-
23
219
BMDL10
-
BDFi mice (Umeda et al., 2005); biphenyl in the diet for 2 yrs
Kidney
Mineralization
Clinical chemistry
AST
ALT
LDH
AP
BUN
Body weight
Terminal body weight
Log-logistic
Not selected13
Not selected13
Not selected13
Not selected13
Linear
No adequate fit0
BMD10
721
BMD1RD
-
-
-
-
BMD1SD
415a
BMD01RD
-
BMDL10
276
BMDL1RD
-
-
-
-
BMDL1SD
267a
BMDLoiRD
-
Log-logistic
Power
No adequate fit0
No adequate fit0
No adequate fit0
No adequate fit0
Linear
Wistar rats (Khera et al., 1979); biphenyl by gavage to dams on GDs 6-15
Litters with fetal skeletal anomalies
Log-logistic
BMD10
233
BMD1RD
190a
-
-
-
BMD1SD
-
BMD01RD
583
BMD10
57
BMDL10
122
BMDL1RD
122a
-
-
-
BMDL1SD
-
BMDL01RD
511
BMDL10
20
"Adequate fit obtained only after excluding results from the highest dose group.
b"Not selected" indicates that the data set was not selected for dose-response analysis because either a treatment-
related effect was not observed or because the response observed in the other sex in the same study was more
robust.
°"No adequate fit" indicates that none of the models in BMDS provided an adequate fit to the data.
BMD = maximum likelihood estimate of the dose associated with the selected BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10% extra risk;
IRD = 100% relative deviation from control mean value; 0 IRD = 10% relative deviation from control mean value;
ISD = 1 SD from control mean value)
102
DRAFT - DO NOT CITE OR QUOTE
-------
400 -
350 -
£• 250 -
•D
1&
E 200 -
150 -
100 -
4 BMDL = 5hl mg/kg-d
J^ ! ! ! X
^f ! ! ! x
i i i
i i i
i i i
i i i
! ! ! '
i i i
i i
; ; I
! 1
i
i
Female
mouse (1)
Terminal
Body
weight
•
1 !
i
i
> i
i
i
i
i
i
i 1
Female
mouse (1)
AST
Liver
1
1 BIVIU - /z i mg/Kg-a
1
>
| I
I
I
I
i
Male
rat (2)
TC
hyperplasia
Bladder
<
1
<
• BMD »BMDL !
1 !
i
i
i
1
>
1
>
Male Female
mouse (1) | mouse (1)
Mineralization
I
Male
mouse (1)
BUN
<
1
<
1
>
1
>
1 i
Male
rat (2)
Female
rat (2)
Nodular hyperplasia
•
•
I
I
»
I
I
!
T T T !
I * * ' f
i i!
i i i i i i i
Male Female Male Female Female Male Female Anomalous
rat (2) | rat (2) rat (2) rat (2) rat (2) rat (2) rat (2) litters (3)
, . , . Renal pelvis Memo- _ ... . .. ..
Simple hyperplasia minerairzation siderin Papillary mineralization
Kidney Deve'°P-
mental
TC = transitional cell
(1) = Umeda et al. (2005); (2) = Umeda et al. (2002); (3) = Khera et al. (1979)
Figure 5-2. BMDs and BMDLs for selected noncancer effects in rats and mice from repeated oral exposure to biphenyl.
103
DRAFT - DO NOT CITE OR QUOTE
-------
1 Examination of the BMD and BMDL values in Table 5-4 and Figure 5-2 reveals
2 BMD/BMDL pairs for four kidney effects and for the developmental effect that are clustered
3 below BMD/BMDL pairs for the other effects. The BMDL values in this cluster range from
4 20 to 58 mg/kg-day and identify the following as the most sensitive nonneoplastic effects
5 associated with repeated oral exposure to biphenyl in animals: (1) renal transitional cell
6 hyperplasia (simple) in female F344 rats (52 mg/kg-day), (2) renal mineralization in female F344
7 rats (56 mg/kg-day), (3) renal hemosiderin deposition in female F344 rats (23 mg/kg-day),
8 (4) renal papillary mineralization in male F344 rats (58 mg/kg-day), and (5) increased litters with
9 fetal skeletal anomalies in Wistar rats (20 mg/kg-day).
10 NOAEL values for endpoints with datasets for which adequate model fits could not be
11 obtained using BMDS were higher than the BMDL values for these five kidney and
12 developmental endpoints. These include selected clinical chemistry parameters in female BDFi
13 mice (NOAELs for LDH, AP, and BUN: 414 mg/kg-day; NOAEL for ALT: 134 mg/kg-day) and
14 terminal body weight in male BDFi mice (NOAEL: 97 mg/kg-day).
15 The increased fetal skeletal anomalies in Wistar rats was selected as the critical effect for
16 deriving an oral RfD because it was considered to be an adverse effect and resulted in the most
17 sensitive POD (BMDLio of 20 mg/kg-day) observed compared with other PODs for biphenyl-
18 induced kidney effects.
19
20 5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs)
21 The RfD for biphenyl was derived by dividing the POD of 20 mg/kg-day (i.e., the
22 BMDLio based on fetal skeletal anomalies in litters from biphenyl-treated pregnant Wistar rats)
23 by a total UF of 100, comprised of 10 for interindividual variability and 10 for interspecies
24 extrapolation, as described below.
25
26 • An UF of 10 was applied to account for interspecies variability in extrapolation from
27 laboratory animals (rats) to humans because information is not available to quantitatively
28 assess toxicokinetic or toxicodynamic differences between animals and humans.
29
30 • An UF of 10 was applied to account for intraspecies variability in susceptibility to
31 biphenyl, as quantitative information for evaluating toxicokinetic and toxicodynamic
32 differences among humans are not available.
33
34 • An UF of 1 was applied for use of data from a subchronic study to assess potential effects
35 from chronic exposure because developmental toxicity resulting from a narrow period of
36 exposure was used as the critical effect. The developmental period is recognized as a
37 susceptible life stage when exposure during a time window of development is more
38 relevant to the induction of developmental effects than lifetime exposure.
39
40 • An UF of 1 was applied for extrapolation from a LOAEL to a NOAEL because the
41 current approach is to address this factor as one of the considerations in selecting a BMR
104 DRAFT - DO NOT CITE OR QUOTE
-------
1 for BMD modeling. In this case, a BMR of 10% increase in incidence of litters with
2 skeletal anomalies was selected under an assumption that it represents a minimal
3 biologically significant change.
4
5 • An UF of 1 to account for database deficiencies was applied. The biphenyl database
6 includes chronic toxicity studies in rats (Umeda et al., 2002; Shiraiwa et al., 1989;
7 Ambrose et al., 1960; Pecchiai and Saffioti, 1957; Dow Chemical Co., 1953) and mice
8 (Umeda et al., 2005; Imai et al., 1983); subchronic toxicity studies in rats (Shibata et al.,
9 1989a, b; Kluwe et al., 1982; S0ndergaard and Blom, 1979; Booth et al., 1961) and mice
10 (Umeda et al., 2004); a developmental toxicity study in rats (Khera et al., 1979); and one-
11 and three-generation reproductive toxicity studies in rats (Ambrose et al., 1960; Dow
12 Chemical Co., 1953). Epidemiological studies provide some evidence that biphenyl may
13 induce functional changes in the nervous system at concentrations in excess of
14 occupational exposure limits. Seppalainen and Hakkinen (1975) reported small increases
15 in anomalies in nerve conduction, EEG, and ENMG signals in workers exposed to
16 biphenyl during the production of biphenyl-impregnated paper at concentrations that
17 exceeded the occupational limit by up to 100-fold, and Wastensson et al. (2006) reported
18 a cluster of Parkinson's disease in a Swedish factory manufacturing biphenyl-
19 impregnated paper. No other clusters of Parkinson's disease have been reported in
20 biphenyl exposed populations, and Wastensson et al. (2006) acknowledged that chance is
21 an alternative explanation for this cluster. Studies in experimental animal models have
22 not identified effects on the nervous system following biphenyl exposure. Accordingly,
23 these epidemiologic studies do not suggest that the nervous system is a sensitive target of
24 biphenyl toxicity and therefore the lack of nervous system-specific studies is not
25 considered a gap in the biphenyl toxicity database.
26
27 The RfD for biphenyl was calculated as follows:
28 RfD = BMDLio - UF
29 = 20 mg/kg-day - 100
30 = 0.2 mg/kg-day
31
32 5.1.4. Previous RfD Assessment
33 The previous IRIS assessment for biphenyl (U.S. EPA, 1989) derived an oral RfD of
34 0.05 mg/kg-day based on kidney damage in albino rats administered biphenyl for 2 years at
35 dietary levels >0.5% (Ambrose et al., 1960). U.S. EPA considered the dietary level of 0.1% (50
36 mg/kg-day using a food factor of 0.05/day) to represent a NOAEL due to the following: (1)
37 uncertainty in the significance of effects observed at lower doses as compared to the more certain
38 adverse effect level of 0.5% in the diet and (2) supporting findings of 0.1% biphenyl as a
39 NOAEL in an unpublished report of a subchronic rat feeding study and a three-generation rat
40 reproduction study performed by Stanford Research Institute (Dow Chemical Co., 1953). The
41 NOAEL of 50 mg/kg-day was divided by a total UF of 1,000 (10 for extrapolation from animals
42 to humans, 10 for protection of sensitive human subpopulations, and a modifying factor of 10 to
43 account for intraspecies variability demonstrated in the threshold suggested by the data in the
44 chronic animal study).
45
105 DRAFT - DO NOT CITE OR QUOTE
-------
1 5.2. INHALATION REFERENCE CONCENTRATION (RfC)
2 5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
3 Human data are limited to assessments of possible associations between occupational
4 exposure to biphenyl and health outcomes where inhalation is presumed to have been the major
5 exposure route. Clinical signs and abnormal electrophysiological test results among workers
6 exposed to biphenyl during the production of biphenyl-impregnated fruit wrapping paper provide
7 evidence of biphenyl-induced neurological effects (Seppalainen and Hakkinen, 1975; Hakkinen
8 et al., 1973, 1971). Case reports include an account of periodic loss of strength and eventual
9 signs of chronic hepatitis in a woman during a 25-year period of employment at a fruit-packing
10 facility where biphenyl-impregnated paper was used (Carella and Bettolo, 1994) and a cluster of
11 five cases of Parkinson's Disease (0.9 cases expected) at a facility manufacturing biphenyl-
12 impregnated paper (Wastensson et al., 2006). None of these studies provided air monitoring data
13 adequate to characterize workplace exposures to biphenyl. Therefore, data from the available
14 human studies could not be used for dose-response analysis and derivation of an RfC.
15 Limited information is available regarding the effects of inhaled biphenyl in laboratory
16 animals. In mice, repeated airborne exposure to biphenyl (7 hours/day, 5 days/week for 2 weeks)
17 at concentrations as high as 54.75 ppm (345.5 mg/m3) appeared to cause no symptoms (Sun
18 Company Inc., 1977a). In a series of studies that included repeated inhalation exposure of
19 rabbits, rats, and mice to atmospheres containing biphenyl for periods of 68-94 days
20 (Deichmann et al., 1947; Monsanto, 1946), rabbits exhibited no signs of exposure-related
21 adverse effects at concentrations as high as 300 mg/m3. Irritation of mucous membranes was
22 observed in rats at concentrations of 40 and 300 mg/m3. Mice were the most sensitive to inhaled
23 biphenyl; irritation of the upper respiratory tract was noted at a concentration of 5 mg/m3
24 (Deichmann et al., 1947; Monsanto, 1946), but other biphenyl concentrations were not tested in
25 this experiment. The limitations of a single exposure level and poorly-reported study details
26 preclude the use of this study for RfC derivation.
27 Repeated exposure of mice to biphenyl at vapor concentrations of 25 or 50 ppm
28 (157.75 or 315.5 mg/m3) for 13 weeks resulted in high incidences of pneumonia and tracheal
29 hyperplasia, and high incidences of congestion and edema in the lungs, liver, and kidney (Sun
30 Company Inc., 1977b). The following study limitations and lack of supporting data preclude the
31 usefulness of this study for deriving an RfC for biphenyl. Measured biphenyl exposure
32 concentrations varied greatly during the first half of the 13-week exposure period; for example,
33 in the high concentration group (target concentration of 50 ppm), the measured concentrations
34 ranged from 5 ppm to 102 ppm during the first 45 exposure sessions. High mortality in 25 ppm
35 male mice (40/50) after 46 exposures necessitated the use of replacement animals; these
36 replacement animals received the same total number of exposure sessions as the surviving
37 animals from the original 25 ppm group but exposures were not concurrent. Histopathological
38 findings were reported only for males and females combined. Reports of lung congestion and
106 DRAFT - DO NOT CITE OR QUOTE
-------
1 hemorrhagic lungs in some control mice were not confirmed histopathologically, and congestion
2 in the lung, liver, and kidney were considered by the study pathologist a likely effect of the
3 anesthetic used for killing the mice. The severity of reported histopathologic lesions was not
4 specified.
5 The 13-week inhalation mouse study of Sun Company Inc. (1977b) is the only available
6 study that employed at least subchronic-duration exposure and included multiple biphenyl
7 exposure levels. This study is considered inadequate for RfC derivation because: (1) exposure
8 levels were highly variable during the first half of the 13-week exposure period, (2) one of the
9 exposure groups experienced high losses (46/100) due to an overheating event and
10 cannibalization after 46 exposures, although replacement mice were subsequently added and
11 received a total of 65 exposures, and (3) limitations in the reporting of histopathological findings.
12 An RfC was not derived due to the significant uncertainty associated with the inhalation
13 database for biphenyl, and route-to-route extrapolation was not supported in the absence of a
14 PBPK model. Although an RfC cannot be derived, it should be noted that the available
15 inhalation data provides some evidence that inhalation exposure to biphenyl could induce
16 respiratory or systemic lesions.
17
18 5.2.2. Previous RfC Assessment
19 No RfC was derived in the previous (1985) IRIS assessment.
20
21 5.3. UNCERTAINTIES IN THE RfD and RfC
22 Risk assessments should include a discussion of uncertainties associated with the derived
23 toxicity values. To derive the oral RfD, the UF approach (U.S. EPA, 2002, 1994b) was applied
24 to a POD of 10 mg/kg-day (see Section 5.1). Factors were applied to the POD to account for
25 extrapolating from responses observed in an animal bioassay to humans or a diverse human
26 population of varying susceptibilities. Uncertainties associated with the data set used to derive
27 the biphenyl RfD are more fully described below.
28 The available database was determined to be inadequate for deriving a chronic inhalation
29 RfC for biphenyl (see Section 5.2).
30
31 Selection of the critical effect for reference value determination. The critical effect
32 selected for derivation of the RfD was skeletal anomalies in fetuses from rat dams administered
33 biphenyl by gavage during GDs 6-15. An increased incidence of these anomalies was reported
34 at doses >500 mg/kg-day; frank maternal toxicity, including death, was observed at the highest
35 dose level (1,000 mg/kg-day). There is some degree of uncertainty regarding the toxicological
36 significance of the reported skeletal anomalies (wavy or extra ribs and delayed ossification most
37 commonly observed) and the relevance of gavage dosing used in the developmental toxicity
38 study to human exposures. Supporting developmental toxicity studies are not available.
107 DRAFT - DO NOT CITE OR QUOTE
-------
1 Dose-response modeling. BMD modeling was used to estimate the POD for the biphenyl
2 RfD. BMD modeling has advantages over a POD based on a NOAEL or LOAEL because, in
3 part, the latter are a reflection of the particular exposure concentration or dose at which a study
4 was conducted. A NOAEL or LOAEL lacks characterization of the dose-response curve, and for
5 this reason, is less informative than a POD obtained from BMD modeling. The selected model,
6 i.e., the log-logistic model, provided the best mathematical fit to the experimental data set (as
7 determined by the lowest AIC), but does not necessarily have greater biological support over the
8 various models included in BMDS. Other models in BMDS yield estimates of the POD higher
9 than the POD derived using the log-logistic model (by up to 5.8-fold).
10 Interspecies extrapolation ofdosimetry and toxicodynamics. Limited information is
11 available regarding species-specific toxicokinetic and toxicodynamic differences in biphenyl
12 metabolism. Results of available in vitro assays of human and rat liver preparations suggest
13 qualitative similarities and quantitative differences in biphenyl metabolism (Powis et al., 1989,
14 1988; Benford et al., 1981). Available in vivo animal data demonstrate qualitative and
15 quantitative differences between rats and mice (Halpaap-Wood et al., 1981a; Meyer and Scheline
16 1976; Meyer et al., 1976a). However, in vivo human data are lacking and it is uncertain which
17 animal species, the rat or the mouse, would be more comparable to humans. Other areas of
18 biphenyl toxicokinetics (absorption, distribution, elimination), have received some attention in
19 animal studies, but comparative human data are not available. PBPK models for biphenyl to
20 address differences in toxicokinetics between animal and human are lacking. An UF of 10 was
21 used to account for animal to human extrapolation in the absence of adequate comparative
22 animal and human toxicokinetic and toxicodynamic data for biphenyl.
23 Sensitive human populations. Heterogeneity among humans is another uncertainty
24 associated with extrapolating doses from animals to humans. Identification of populations that
25 might be relatively more susceptible to the toxic effects of biphenyl is not feasible because of the
26 limited information on biphenyl metabolism and mode of action of biphenyl toxicity. It is
27 known, however, that many CYP isozymes and glucuronidase exist in polymorphic forms. Such
28 enzyme polymorphism may put some populations at increased risk from biphenyl exposure. In
29 the absence of biphenyl-specific data on human variation, a factor of 10 was used to account for
30 uncertainty associated with human variation. Human variation may be larger or smaller;
31 however, biphenyl-specific data to examine the potential magnitude of over- or under-estimation
32 are absent.
33
34 5.4. CANCER ASSESSMENT
35 5.4.1. Choice of Study/Data—with Rationale and Justification
36 No information was located regarding possible associations between oral exposure to
37 biphenyl and cancer in humans. Two animal bioassays found statistically significant
108 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
14
15
associations between lifetime oral exposure to biphenyl and tumor development. Biphenyl was
associated with urinary bladder tumors in male, but not female, F344 rats (Umeda et al., 2002)
and liver tumors in female, but not male, BDFi mice (Umeda et al., 2005). Tumor data for these
two sites were selected for dose-response analysis.
No studies were identified that examined the association between inhalation exposure to
biphenyl and cancer in humans or animals.
5.4.2. Dose-Response Data
The dose-response data for urinary bladder tumor formation resulting from lifetime oral
exposure of male and female F344 rats (Umeda et al., 2002) are shown in Table 5-5. The dose-
response data for liver tumor formation resulting from lifetime oral exposure of male and female
BDFi mice (Umeda et al., 2005) are shown in Table 5-6. The datasets selected for BMD
analysis include urinary bladder transitional cell papilloma or carcinoma (combined) in the male
F344 rats and liver adenoma or carcinoma (combined) in the female BDFi mice.
Table 5-5. Incidence data for tumors in the urinary bladder of male and
female F344 rats exposed to biphenyl in the diet for 2 years
Biphenyl dietary concentration (ppm)
Calculated dose (mg/kg-d)a
Tumor incidence0
Transitional cell
Papilloma
Carcinoma
Papilloma or carcinoma
Males
0
0
0/50
0/50
050
500
36.4
0/50
0/50
0/50
1,500
110
0/50
0/50
0/50
4,500
378
10/49b'
24/49b'
31/49b'
Females
0
0
0/50
0/50
0/50
500
42.7
0/50
0/50
0/50
1,500
128
0/50
0/50
0/50
4,500
438
0/50
0/50
0/50
"Calculated doses based on TWA body weights (calculated from body weight data presented graphically in Figure 1
of Umeda et al., 2002) and chronic reference food consumption values for F344 rats listed in Table 1-6 of U.S. EPA
(1988).
bSignificantly different from control group (p < 0.01) according to Fisher's exact test.
°One high-dose male rat was excluded from the denominator because it died prior to week 52. It is assumed that
this rat did not have a tumor and was not exposed for a sufficient time to be at risk for developing a tumor. Umeda
et al. (2002) did not specify the time of appearance of the first tumor.
Source: Umeda et al. (2002).
16
109
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-6. Incidence data for liver tumors in male and female BDFi mice
fed diets containing biphenyl for 2 years
Biphenyl dietary
concentration (ppm)
Reported dose (mg/kg-d)
Tumor incidence0
Adenoma
Carcinoma
Adenoma or carcinoma
Dietary concentration of biphenyl (ppm)
Males
0
0
8/50
8/50
16/50
667
97
6/49
8/49
12/49
2,000
291
7/49
5/49
9/49
6,000
1,050
3/50
4/50
7/50
Females
0
0
2/48
1/48
3/48
667
134
3/50
5/50
8/50
2,000
414
12/493
7/49a
16/49b
6,000
1,420
10/483
5/48
14/483
"Significantly different from controls (p < 0.05) according to Fisher's exact test as reported by Umeda et al. (2005).
bSignificantly different from controls (p < 0.01) according to Fisher's exact test as reported by Umeda et al. (2005).
°One low-dose, one mid-dose male, two control, one mid-dose, and two high-dose female mice were excluded from
denominators because they died prior to week 52. It is assumed that they did not have tumors and were not exposed
for a sufficient time to be at risk for developing a tumor. Umeda et al. (2005) did not specify the time of appearance
of the first tumor.
Source: Umeda et al. (2005).
1
2 5.4.3. Dose Adjustments and Extrapolation Method(s)
3 5.4.3.1. Bladder Tumors in Male Rats
4 There is strong evidence that the occurrence of urinary bladder tumors in male rats
5 chronically exposed to biphenyl in the diet is a high-dose nongenotoxic phenomenon involving
6 occurrence of calculi in the urinary bladder leading to transitional cell damage, sustained
7 regenerative cell proliferation, and eventual promotion of spontaneously initiated tumor cells in
8 the urinary bladder epithelium (see Section 4.7.3.1 for a detailed discussion of the hypothetized
9 mode of action for urinary bladder tumors in biphenyl-exposed male rats). No increased risk of
10 bladder tumors is expected as long as the exposure to biphenyl is below the dose needed to form
11 calculi (Cohen and Ellwein, 1992). As noted in the EPA Guidelines for Carcinogen Risk
12 Assessment (U.S. EPA, 2005a), a nonlinear approach to dose-response analysis is used when
13 there are sufficient data to ascertain the mode of action and conclude that it is not linear at low
14 doses and the agent does not demonstrate mutagenic or other activity consistent with linearity at
15 low doses. Therefore, consistent with the cancer guidelines, a nonlinear extrapolation approach
16 for biphenyl-induced urinary bladder tumors was selected.
17 Based on the proposed mode of action, the available evidence indicates that doses below
18 the oral RfD would not result in the sequence of events that includes calculus formation,
19 consequent epithelial cell damage, and sustained regenerative cellular proliferation.
20 Accordingly, the RfD of 0.2 mg/kg-day derived for noncancer effects of biphenyl was judged to
21 be protective against an increased risk of biphenyl-induced urinary bladder cancer.
22
110
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
11
12
13
5.4.3.2. Liver Tumors in Female Mice
In the study report of their 2-year bioassay in BDFi mice, Umeda et al. (2005) provided
averaged food consumption and biphenyl dose estimates for each exposure group (Table 1 of
Umeda et al., 2005). The study report did not include average body weights for the exposure
groups. Therefore, the biphenyl concentration in the food was multiplied by the corresponding
average daily food consumption value to determine the average daily biphenyl intake. Dividing
this average daily biphenyl intake by the author-calculated daily dose yielded the average body
weight that would have been used by the study authors to calculate the average daily biphenyl
dose. Scaling factors were calculated using U.S. EPA (1988) reference body weight for humans
(70 kg) and the average body weight for each dose group of female mice: (average body
weight/70)0 25 = scaling factor. The human equivalent dose (HED) was calculated as: HED =
scaling factor x reported dose (Table 5-7).
Table 5-7. Scaling factors for determining HEDs to use for BMD modeling
of female BDFi mouse liver tumor incidence data from Umeda et al. (2005)
Biphenyl dietary concentration (mg/kg food)
Reported dose (mg/kg-d)
Reported average food consumption (kg/d)
Average mouse body weight (kg)a
Scaling factorb
HED (mg/kg-d)c
667
134
0.0058
0.0289
0.143
19
2,000
414
0.0059
0.0285
0.142
59
6,000
1,420
0.0059
0.0249
0.137
195
14
15
16
17
18
19
20
21
22
23
24
25
26
"(Biphenyl concentration in food [mg/kg food] x reported average food consumption [kg/day]) -^ reported average
daily dose of biphenyl (mg/kg-day) = calculated average mouse body weight (kg).
bCalculated using reference body weight for humans (70 kg; U.S. EPA, 1988), and the average body weights for
each dose group: mouse-to-human scaling factor = (average mouse body weight/70)0 25.
°HED = reported dose x scaling factor.
The EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) recommend
that when the weight of evidence evaluation of all available data are insufficient to establish the
mode of action for a tumor site and when scientifically plausible based on the available data,
linear extrapolation is used as a default approach. Accordingly, a linear approach to low-dose
extrapolation for biphenyl-induced liver tumors in female mice was selected because the mode of
action for this tumor site has not been established (see Section 4.7.3.2).
Incidence data for liver adenoma or carcinoma (combined) in the female mouse used to
derive the oral slope factor are presented in Table 5-8. Tumor incidence data were adjusted to
account for mortalities before 52 weeks; it was assumed that animals dying before 52 weeks
were not exposed for sufficient time to be at risk for developing tumors (see footnote a in Table
5-8).
Ill
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-8. Incidence of liver adenomas or carcinomas (combined) in female
BDFi mice fed diets containing biphenyl for 2 years
Biphenyl dietary concentration (ppm)
RED (mg/kg-d)
Tumor incidence
Adenoma or carcinoma (combined)
0
0
3/48a
667
19
8/50
2,000
59
16149s*
6,000
195
14/48a'c
aTwo control, one mid-dose, and two high-dose female mice were excluded from denominators because they died
prior to week 52. It is assumed that they did not have tumors and were not exposed for a sufficient time to be at
risk for developing a tumor. Umeda et al. (2005) did not specify the time of appearance of the first tumor.
bSignificantly different from controls (p < 0.05) according to Fisher's exact test.
Significantly different from controls (p < 0.01) according to Fisher's exact test.
Source: Umeda et al. (2005).
1
2 The multistage-cancer model in the EPA BMDS (version 2.1.2), using the extra risk
3 option, was fit to the female mouse liver tumor incidence data. The multistage model has been
4 used by EPA in the vast majority of quantitative cancer assessments because it is thought to
5 reflect the multistage carcinogenic process and it fits a broad array of dose-response patterns.
6 The multistage-cancer model was run for all polynomial degrees up to n-1 (where n is the
7 number of dose groups including control). An extra risk of 10% tumor incidence was selected as
8 the benchmark response. Adequate model fit was judged by three criteria: goodness-of-fit/>-
9 value (p > 0.05), visual inspection of the dose-response curve, and a value of <2 for the largest
10 scaled residual for any data-point in the dataset (including the control). If an adequate fit to the
11 data was not achieved using the protocol above, the other dichotomous models were fit to the
12 data. If none of the models achieved an adequate fit for the full dataset, the highest dose was
13 dropped and the entire modeling procedure was repeated.
14 When liver tumor incidence data for all dose groups were modeled, none of the models in
15 BMDS, including the multistage model, provided an adequate fit of the data (see Appendix C,
16 Table C-2). The animals in the highest dose group, while exhibiting a statistically significantly
17 increased incidence in liver tumors compared with controls, did not show a monotonic increase
18 in tumor response compared with the responses at the lower doses. To better estimate responses
19 in the low dose region, the high dose group was excluded as a means of improving the fit of the
20 model in the region of interest. When the high-dose group was dropped, the multistage model
21 provided an adequate fit to the data (see Appendix C, Table C-2). The BMDHEDio and
22 BMDLHEDio using this latter dataset were 18.7 and 12.2 mg/kg-day, respectively. See Appendix
23 C for more information.
24
112
DRAFT - DO NOT CITE OR QUOTE
-------
1 5.4.4. Oral Slope Factor and Inhalation Unit Risk
2 A low-dose linear extrapolation approach results in calculation of an oral slope factor that
3 describes the cancer risk per unit dose of the chemical at low doses. The oral slope factor was
4 calculated by dividing the risk (i.e., BMR of 10% extra risk) at the POD by the corresponding
5 BMDL (0. l/BMDLHEDio). Using linear extrapolation from the BMDLHEDio, the human
6 equivalent oral slope factor of 8.2 x 10"3 (mg/kg-d)"1 (round to one significant figure, 8 x 10"3
7 (mg/kg-d)"1) was derived for liver tumors in female BDFi mice (Table 5-9).
8
Table 5-9. POD and oral slope factor derived from liver tumor incidence
data from BDFi female mice exposed to biphenyl in the diet for 2 years
Species/tissue site
Female mouse liver tumors
BMDaEDio
(mg/kg-d)
18.7
BMDLaEDio
(mg/kg-d)
12.2
Slope factor" (risk per
[mg/kg-d])
8.2 x 10'3
aHuman equivalent slope factor = 0.1/BMDL10HED; see Appendix C for details of modeling results.
9
10 This slope factor should not be used with exposures >12.2 mg/kg-day (the POD for this
11 dataset), because above the POD, the fitted dose-response model better characterizes what is
12 known about the carcinogenicity of biphenyl (i.e., the slope factor may not approximate the
13 observed dose-response relationship adequately at exposure exceeding 12.2 mg/kg-day).
14 An inhalation unit risk for biphenyl was not derived in this assessment. The potential
15 carcinogenicity of inhaled biphenyl has not been evaluated in human or animal studies, and
16 route-to-route extrapolation was not possible in the absence of a PBPK model.
17
18 5.4.5. Uncertainties in Cancer Risk Values
19 5.4.5.1. Oral Slope Factor
20 A number of uncertainties underlie the cancer unit risk for biphenyl. Table 5-10
21 summarizes the impact on the assessment of issues such as the use of models and extrapolation
22 approaches (particularly those underlying the Guidelines for Carcinogen Risk Assessment (U. S.
23 EPA, 2005a), the effect of reasonable alternatives, the decision concerning the preferred
24 approach, and its justification.
25 The uncertainties presented in Table 5-10 have a varied impact on risk estimates. Some
26 suggest risks could be higher than was estimated, while others would decrease risk estimates or
27 have an impact of an uncertain direction. Several uncertainties are quantitatively characterized
28 for the significantly increased rodent tumors. These include the statistical uncertainty in the
29 multistage modeling estimate. Due to limitations in the data, particularly regarding the MOA
30 and relative human sensitivity and variability, the quantitative impact of other uncertainties of
31 potentially equal or greater impact has not been explored. As a result, an integrated quantitative
32 analysis that considers all of these factors was not undertaken.
113 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-10. Summary of uncertainties in the biphenyl cancer slope factor
Consideration/
approach
Selection of data
set
Cross-species
scaling
Extrapolation
procedure for rat
urinary bladder
tumors
Extrapolation
procedure for
mouse liver tumors
Human relevance
of female mouse
liver tumor data
Model uncertainty
Statistical
uncertainty at POD
Human population
variability /
sensitive
subpopulations
Impact on slope
factor
No other studies or
data sets could be
used to derive a slope
factors
Alternatives (i.e.
scaling by [body
weight] or [body
weight]273) could t or
I slope factor
No impact on the
slope factor because
the MOA for male rat
bladder tumors does
not support low-dose
linear extrapolation.
Departure from EPA's
Guidelines for
Carcinogen Risk
Assessment POD
paradigm, if justified,
could | or t slope
factor by an unknown
extent
Human risk could for
|, depending on
relative sensitivity
For poorly fitting liver
tumors dataset,
alternatives could J, or
t slope factor
I slope factor 1.5 -fold
if BMD10 used rather
thanBMDL10
Low-dose risk f to an
unknown extent
Decision
Umeda et al. (2005)
studies were selected.
Administered dose was
scaled to humans on
the basis of
equivalence of
mg/kg3/4-day (default
approach)
Nonlinear
extrapolation. The
RfDof0.2mg/kg-day
is considered to protect
against the risk of
urinary bladder tumors.
Multistage model to
determine the POD,
linear low-dose
extrapolation from
POD (default
approach)
Liver tumors in female
mice are relevant to
human exposure
Drop highest dose of
the liver tumors
dataset.
BMDL (default
approach for
calculating plausible
upper bound)
Considered
qualitatively
Justification
The bioassay by Umeda et al. (2005) was a
well conducted experiment with sufficient
dose groups (four dose groups, including
control) and animal numbers (50 animals/sex)
per group.
There are no data to support alternatives. Use
of [body weight]374 for cross-species scaling is
consistent with data that allow comparison of
potencies in humans and animals, and it is
supported by analysis of the allometric
variation of key physiological parameters
across mammalian species. No PBPK model is
available to derive internal doses.
Available MOA data for urinary bladder
tumors support nonlinearity (i.e., that bladder
tumor is a high-dose phenomena, and is
closely related to calculi formation in the
urinary bladder of male rats).
Available MOA data do not inform selection
of dose-response model; linear approach in
absence of clear support for an alternative is
generally consistent with scientific
deliberations supporting EPA's Guide lines for
Carcinogen Risk Assessment.
It was assumed that humans are as sensitive as
the most sensitive rodent gender/species
tested; true correspondence is unknown.
Model options explored with full liver tumor
datasets did not generate a p > 0.05, which is
one of the indications of dropping the highest
dose according to the draft Benchmark Dose
Technical Guidance (U.S. EPA, 2000b).
Limited size of bioassay results in sampling
variability; lower bound is 95% confidence
interval on dose.
No data to support range of human
variability /sensitivity in metabolism or
response, including whether children are more
sensitive.
BMDL10 = 95% lower confidence limits on the doses associated with a 10% extra risk of cancer incidence.
1 5.4.5.2. Inhalation Unit Risk
1 The potential carcinogen!city of inhaled biphenyl has not been assessed. Therefore, a
3 quantitative cancer assessment for biphenyl by the inhalation pathway was not performed.
114
DRAFT - DO NOT CITE OR QUOTE
-------
1
2 5.4.6. Previous Cancer Assessment
3 In the previous IRIS cancer assessment (U.S. EPA, 1991), biphenyl was listed in Group
4 D; not classifiable as to human carcinogenicity based on no human data and inadequate studies
5 in mice and rats. Neither an oral slope factor nor inhalation unit risk was derived in the previous
6 cancer assessment.
7
115 DRAFT - DO NOT CITE OR QUOTE
-------
1 6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
2 RESPONSE
o
3
4
5 6.1. HUMAN HAZARD POTENTIAL
6 6.1.1. Noncancer
7 Toxicokinetic studies of animals indicate that orally administered biphenyl is rapidly and
8 readily absorbed, distributed widely to tissues following absorption, and rapidly eliminated from
9 the body, principally as conjugated hydroxylated metabolites in the urine (Meyer, 1977; Meyer
10 and Scheline, 1976; Meyer et al., 1976a, b). Data for absorption, distribution, and elimination
11 are not available for inhaled or dermally applied biphenyl. Metabolism to a range of
12 hydroxylated metabolites has been demonstrated in in vitro systems with rat and human cells and
13 tissues. Human metabolism of biphenyl appears to be qualitatively similar to metabolism in the
14 rat, although some reports of quantitative differences are available (Powis et al., 1989, 1988;
15 Benford et al., 1981).
16 Available human health hazard data consist of limited assessments of workers exposed to
17 biphenyl during the production or use of biphenyl-impregnated fruit wrapping paper in which
18 signs of hepatic and nervous system toxicity were observed.
19 Chronic oral studies in rats and mice identify the liver and urinary system as principal
20 targets of biphenyl toxicity, the rat kidney being the most sensitive. Results of a developmental
21 toxicity study in rats indicate that skeletal development is a sensitive indicator of biphenyl
22 toxicity. In chronically exposed rats, non-neoplastic kidney lesions (simple transitional cell
23 hyperplasia in the renal pelvis and hemosiderin deposits) were found in females at >1,500 ppm
24 biphenyl in the diet (128 mg/kg-day), and urinary bladder tumors, associated with urinary
25 bladder calculi and transitional cell hyperplasia, were found in males, but not females, at the
26 highest tested concentration, 4,500 ppm (378 mg/kg-day) (Umeda et al., 2002). Several other rat
27 studies provide supporting evidence that the kidney and other urinary tract regions are sensitive
28 targets for biphenyl in rats (Shiraiwa et al., 1989; Ambrose et al., 1960; Pecchiai and Saffiotti,
29 1957; Dow Chemical Co., 1953). In chronically exposed BDFi mice, increased incidence of
30 nonneoplastic effects on the kidney (mineralization) and liver (increased activities of plasma
31 ALT and AST) were found in females exposed to >2,000 ppm biphenyl in the diet (414 mg/kg-
32 day) (Umeda et al., 2005). In contrast, no exposure-related nonneoplastic or neoplastic effects
33 on the liver or kidney were found in female ddY mice exposed to 5,000 ppm biphenyl in the diet
34 for 2 years (Imai et al., 1983) or in B6C3Fi and B6AKFi mice exposed to 517 ppm biphenyl in
35 the diet for 18 months (Innes et al., 1969; NCI, 1968). In the only available developmental
36 toxicity study for biphenyl, increased incidences of litters with fetuses showing skeletal
37 anomalies were reported following exposure of pregnant rats to gavage doses >500 mg/kg-day
38 on GDs 6-15 (Khera et al., 1979).
116 DRAFT - DO NOT CITE OR QUOTE
-------
1 Biphenyl effects on reproductive function in rats have been reported at a higher exposure
2 level than the lowest exposure levels associated with urinary tract, liver, or developmental
3 toxicity. No exposure-related effect on the number of dams with litters was found following
4 exposure of male and female albino rats to up to 5,000 ppm biphenyl in the diet (525 mg/kg-day)
5 for 11 or 60 days prior to mating (Ambrose et al., 1960). In a three-generation rat study,
6 decreased fertility, decreased number of pups/litter, and decreased pup body weight were
7 observed at 10,000 ppm biphenyl in the diet; (947 mg/kg-day), but not at <1,000 ppm (Dow
8 Chemical Co., 1953).
9 No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
10 toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
11 biphenyl (7 hours/day, 5 days/week for up to about 90 days) were reported in mice exposed to
12 5 mg/m3 and in rats exposed to 300 mg/m3, but not in rabbits exposed to 300 mg/m3 (Deichmann
13 et al., 1947; Monsanto, 1946). Congestion or edema of the lung, kidney, and liver, accompanied
14 by hyperplasia with inflammation of the trachea, was reported in CD-I mice exposed to biphenyl
15 vapors at 25 or 50 ppm (158 or 315 mg/m3) for 13 weeks (Sun Company Inc., 1977b).
16
17 6.1.2. Cancer
18 No assessments are available regarding possible associations between exposure to
19 biphenyl and increased risk of cancer in humans.
20 In a 2-year study of F344 rats administered biphenyl in the diet, significantly increased
21 incidences of urinary bladder tumors in males were observed at the highest dose level
22 (378 mg/kg-day). There is strong evidence that the occurrence of urinary bladder tumors in the
23 male rats is a high-dose nongenotoxic phenomenon involving occurrence of calculi in the urinary
24 bladder leading to transitional cell damage, sustained regenerative cell proliferation, and eventual
25 promotion of spontaneously initiated tumor cells in the urinary bladder epithelium. Urinary
26 bladder calculi in the high-dose (438 mg/kg-day) female rats were observed at much lower
27 incidence and were different in physical appearance and chemical composition; furthermore,
28 there were no urinary bladder tumors in any of the biphenyl-exposed female rats.
29 In a 2-year study of BDFi mice administered biphenyl in the diet, the incidence of liver
30 tumors in female mice was significantly increased at doses >414 mg/kg-day, but not in males at
31 doses up to and including 1,050 mg/kg-day. Available data are insufficient to establish a mode
32 of action for liver tumors in female mice.
33 Under EPA Guidelines for Carcinogenic Risk Assessment (U.S. EPA, 2005a), the
34 database for biphenyl provides "suggestive evidence of carcinogenic potential" based on
35 evidence of female mouse liver tumors and male rat bladder tumors.
36
117 DRAFT - DO NOT CITE OR QUOTE
-------
1 6.2. DOSE RESPONSE
2 6.2.1. Noncancer/Oral
3 The RfD of 0.2 mg/kg-day was based on an increased incidence of litters with fetal
4 skeletal anomalies from Wistar rat dams administered biphenyl by gavage on GDs 6-15 (Khera
5 et al., 1979). The BMDLio of 20 mg/kg-day was selected as the POD. To derive the RfD, the
6 POD was divided by a total UF of 100 (10 for animal-to-human extrapolation and 10 for human
7 interindividual variability in susceptibility). The interspecies uncertainty factor was applied to
8 account for the lack of quantitative information to assess toxicokinetic and toxicodynamic
9 differences between animals and humans. The intraspecies uncertainty factor was applied to
10 account for the lack of information regarding the range of responses to biphenyl in the human
11 population.
12 The overall confidence in the RfD assessment is medium to high. Confidence in the
13 principal study (Khera et al., 1979) is medium to high. The design, conduct and reporting of this
14 developmental toxicity study in Wistar rats were adequate; however, only litter average data
15 were available that did not permit a nested analysis based on individual fetal data. Confidence in
16 the database is high. The database is robust in that it includes chronic-duration oral exposure
17 studies in several rat and mouse strains, a developmental toxicity study in Wistar rats, and one-
18 and three-generation reproductive toxicity studies in rats.
19
20 6.2.2. Noncancer/Inhalation
21 No inhalation RfC was derived due to the lack of studies of biphenyl toxicity following
22 chronic exposure and studies involving subchronic exposure that were inadequate for RfC
23 derivation. Repeated exposure of mice to biphenyl vapors for 13 weeks resulted in high
24 incidences of pneumonia and tracheal hyperplasia, and high incidences of congestion and edema
25 in the lungs, liver, and kidney (Sun Company Inc., 1977b); however, study limitations and lack
26 of supporting data preclude the use of this study for deriving an RfC for biphenyl. Study
27 limitations include highly variable biphenyl exposure concentrations during the first half of the
28 study, high mortality after 46 exposures in one group of biphenyl-exposed mice due to an
29 overheating event and cannibalization that necessitated the use of replacement animals, and
30 limitations in the reporting of histopathological findings.
31
32 6.2.3. Cancer/Oral
33 The oral slope factor of 0.008 per mg/kg-day is based on the tumor response in the liver
34 of female BDFi mice exposed to biphenyl in the diet for 2 years (Umeda et al., 2005). The slope
35 factor was derived by linear extrapolation from a human equivalent BMDLio of 12.2 mg/kg-day
36 for liver adenomas or carcinomas.
118 DRAFT - DO NOT CITE OR QUOTE
-------
1 A nonlinear extrapolation approach for biphenyl-induced urinary bladder tumors in male
2 rats was used because evidence show that the occurrence of urinary bladder tumors is a high-
3 dose nongenotoxic phenomenon involving occurrence of calculi in the urinary bladder leading to
4 transitional cell damage, sustained regenerative cell proliferation, and eventual promotion of
5 spontaneously initiated tumor cells in the urinary bladder epithelium. As long as the dose is
6 below that which is needed to form calculi, no increased risk of bladder tumors is expected.
7 Therefore, the RfD of 0.2 mg/kg-day derived for noncancer effects of biphenyl was judged to be
8 protective against increased risk of biphenyl-induced urinary bladder cancer.
9
10 6.2.4. Cancer/Inhalation
11 No human or animal data on the potential carcinogenicity of inhaled biphenyl are
12 available. Therefore, a quantitative cancer assessment for biphenyl by the inhalation pathway
13 was not performed.
119 DRAFT - DO NOT CITE OR QUOTE
-------
7. REFERENCES
Abe, S; Sasaki, M. (1977) Chromosome aberrations and sister chromatic! exchanges in Chinese hamster cells
exposed to various chemicals. J Natl Cancer Inst 58(6):1635-1641.
ACGIH (American Conference of Governmental Industrial Hygienists). (2008) Biphenyl. In: Threshold limit values
for chemical substances and physical agents, and biological exposure indices. Cincinnati, OH: American
Conference of Governmental Industrial Hygienists, p. 14.
Ambrose, A; Booth, A; DeEds, F; et al. (1960) A lexicological study of biphenyl, a citrus fungistat. Food Res
25:328-336.
Balakrishnan, S; Uppala, PT; Rupa, DS; et al. (2002) Detection of micronuclei, cell proliferation and hyperdiploidy
in bladder epithelial cells of rats treated with o-phenylphenol. Mutagenesis 17:89-93.
Benford, D; Bridges, J. (1983) Tissue and sex differences in the activation of aromatic hydrocarbon hydroxylases in
rats. BiochemPharmacol 32:309-313.
Benford, D; Bridges, J; Boobis, A; et al. (1981) The selective activation of cytochrome P-450 dependent microsomal
hydroxylases in human and rat liver microsomes. Biochem Pharmacol 30(12): 1702-1703.
Bianco, PJ; Jones, RS; Parke, DV. (1979) Effects of carcinogens on biphenyl hydroxylation in isolated rat
hepatocytes. Biochem Soc Trans 7:639-641.
Billings, RE; McMahon, RE. (1978) Microsomal biphenyl hydroxylation, the formation of 3-hydroxybiphenyl and
biphenyl catechol. Mol Pharmacol 14:145-154.
Bock, KW; von Clausbruch, UC; Kaufmann, R; et al. (1980) Functional heterogeneity of UDP-
glucuronyltransferase in rat tissues. Biochem Pharmacol 29(4):495-500.
Boone, L; Meyer, D; Cusick, P; et al. (2005) Selection and interpretation of clinical pathology indicators of hepatic
injury in preclinical studies. Vet ClinPathol 34(3): 182-188.
Booth, A; Ambrose, A; Deeds, F; et al. (1961) The reversible nephrotoxic effects of biphenyl. Toxicol Appl
Pharmacol 3:560-567.
Bos, RP; Theuws, JL; Jongeneelen, FJ; et al. (1988) Mutagenicity of bi-, tri- and tetra-cyclic aromatic hydrocarbons
in the "taped-plate assay" and in the conventional Salmonella mutagenicity assay. Mutat Res 204(2):203-206.
Boutwell, R; Bosch, D. (1959) The tumor-promoting action of phenol and related compounds for mouse skin.
Cancer Res 19(4):413^24.
Brams, A; Buchet, J; Crutzen-Fayt, M; et al. (1987) A comparative study, with 40 chemicals, of the efficiency of the
Salmonella assay and the SOS chromotest (kit procedure). Toxicol Lett 38:123-133.
Brouns, R; Foot, M; De Vrind, R; et al. (1979) Measurement of DNA-excision repair in suspensions of freshly
isolated rat hepatocytes after exposure to some carcinogenic compounds: its possible use in carcinogenicity
screening. Mutat Res 64(6):425-432.
Buonocore, G; Perrone, S; Bracci, R. (2001) Free radicals and brain damage in the newborn. BiolNeonate 79:180-
186.
Burke, MD; Bridges, JW. (1975) Biphenyl hydroxylations and spectrally apparent interactions with liver
microsomes from hamsters pre-treated with phenobarbitone and 3-methylcholanthrene. Xenobiotica 5(6):357-376.
Carella, G; Bettolo, P. (1994) Reversible hepatotoxic effects of diphenyl: report of a case and a review of the
literature. J Occup Med 36(5):575-576.
120 DRAFT - DO NOT CITE OR QUOTE
-------
Charles River Laboratories International, Inc. (1999) B6D2F1 (BDF1) Mouse. Wilmington, MA. Available online at
http://www.criver.com/en-US/ProdServ/ByType/ResModOver/ResMod/Pages/B6D2FlMouse.aspx (accessed July
6,2011).
Chung, K; Adris, P. (2002) Growth inhibition of intestinal bacteria and mutagenicity of aminobiphenyls, biphenyl
and benzidine. Abstr Gen Meeting Am Soc Microbiol 102:10.
Chung, KT; Adris, P. (2003) Growth inhibition of intestinal bacteria and mutagenicity of 2-, 3-, 4-aminobiphenyls,
benzidine, and biphenyl. Toxicol In Vitro 17(2): 145-152.
Cline, J; McMahon, R. (1977) Detection of chemical mutagens: use of concentration gradient plates in a high
capacity screen. Res Commun Chem Pathol Pharmacol 16:523-533.
Cohen, SM. (1995) Cell proliferation in the bladder and implications for cancer risk assessment. Toxicology
102:149-159.
Cohen, SM. (1998) Cell proliferation and carcinogenesis. Drug Metab Rev 30(2):339-357.
Cohen, SM; Ellwein LB. (1992) Risk assessment based on high-dose animal exposure experiments. Chem. Res.
Toxicol. 5: 742-748
Creaven, PJ; Parke, DV. (1966) The stimulation of hydroxylation by carcinogenic and non-carcinogenic compounds.
Biochem Pharmacol 15:7-16.
Deichman, WB; Kitzmiller, K; Dierker, M; et al. (1947) Observations on the effects of diphenyl, o- and
p-aminodiphenyl, o- and p-nitrodiphenyl and dihydroxyoctachlorodiphenyl upon experimental animals. J Ind Hyg
Toxicol 29:1-13.
Dow Chemical Co. (1939) Toxicity of diphenyl and diphenyl oxide (sanitized). Submitted under TSCA Section 8D;
EPA Document No. 86-890001205S; NTIS No. OTS0520717.
Dow Chemical Co. (1953) Toxicological study of diphenyl in citrus wraps with cover letter. Prepared by Stanford
Research Institute. Submitted under TSCA Section 8D. EPA Document No. 878213721; NTIS No. OTS0206456.
Dow Chemical Co. (1971) Partition coefficients of biphenyl, diphenyl oxide and dowtherm a between 1-octanol and
water-another look (1983). Dow Chemical Company, Midland, MI. Submitted under TSCA Section 8D; EPA/OTS
Doc No. 878213735; NTIS No. OTS0206456.
Duanmu, Z; Weckle, A; Koukouritaki, SB; et al. (2006) Developmental expression of aryl, estrogen, and
hydroxysteroid sulfotransferases in pre- and postnatal human liver. J Pharmacol Exp Ther 316(3): 1310-7.
EMEA (European Medicines Agency). (2006) Draft guidelines on detection of early signals of drug-induced
hepatotoxicity in non-clinical studies. Committee for Medicinal Products for Human Use (CHMP), London.
Adoption by CHMP for release for consultation 28 June 2006. EMEA/CHMP/SWP/150115/2006.
Fujita, H; Kojima, A; Sasaki, M; et al. (1985) Mutagenicity test of antioxidants and fungicides with Salmonella
typhimurium TA97a, TA102. Kenkyu Nenpo-Tokyo-Toritsu Eisei Kenkyusho 36:413-417.
Garberg, P; Akerblom, E-L; Bolcsfoldi, G. (1988) Evaluation of a genotoxicity test measuring DNA-strand breaks in
mouse lymphoma cells by alkaline unwinding and hydroxyapatite elution. Mutat Res 203(3): 155-176.
Garrett, NE; Stack, HF; Waters, MD. (1986) Evaluation of the genetic activity profiles of 65 pesticides. Mutat Res
168:301-325.
Glatt, H; Anklam, E; Robertson, LW. (1992) Biphenyl and fluorinated derivatives: liver enzyme-mediated
mutagenicity detected in Salmonella typhimurium and Chinese hamster V79 cells. Mutat Res 281(3):151-156.
Gombar, V; Borgstedt, H; Enslein, K; et al. (1991) A QSAR model of teratogenesis. Quant Struct-Act Relat
10(4):306-332.
121 DRAFT - DO NOT CITE OR QUOTE
-------
Hakkinen, I; Vikkula, E; Hernberg, S. (1971) The clinical picture of diphenyl poisoning. Scand J Clin Lab Invest
Suppl27(116).
Hakkinen, I; Siltanen, E; Hernberg, S; et al. (1973) Diphenyl poisoning in fruit paper production: a new health
hazard. Arch Environ Health 26(2):70-74.
Halpaap-Wood, K; Horning, E; Horning, M. (1981a) The effect of phenobarbital and beta-naphthoflavone induction
on the metabolism of biphenyl in the rat and mouse. Drug Metab Dispos 9(2):97-102.
Halpaap-Wood, K; Horning, E; Horning, M. (1981b) The effect of 3-methylcholanthrene, Aroclor 1254, and
phenobarbital induction on the metabolism of biphenyl by rat and mouse 9,000 g supernatant liver fractions. Drug
Metab Dispos 9(2): 103-107.
Hanada, S. (1977) Studies on food additives, diphenyl (biphenyl) and O-phenyl phenol from the view point of public
health. Part 2. On the toxicities of diphenyl and O-phenyl phenol. J Nagoya City Univ Med Sch 28(3):983-995.
Haugen, D. (1981) Biphenyl metabolism by rat liver microsomes: regioselective effects of inducers, inhibitors, and
solvents. Drug Metab Dispos 9(3):212-218.
Haworth, S; Lawlor, T; Mortelmans, K; et al. (1983) Salmonella mutagenicity test results for 250 chemicals.
Environ Mutagen 5(Suppl. 1): 1-142.
Hellmer, L; Bolcsfoldi, G. (1992) An evaluation on the Escherichia coli K-12 uvrB/recA DNA repair host-mediated
assay: I. In vitro sensitivity of the bacteria to 61 compounds. Mutat Res 272(2): 145-160.
Houk, VS; Schlakowsky, S; Claxton, LD. (1989) Development and validation of the spiral Salmonella assay: an
automated approach to bacterial mutagenicity testing. Mutat Res 223(l):49-64.
Hsia, M; Kreamer, B; Dolara, P. (1983a) Quantitation of chemically induced DNA damage and repair in isolated rat
hepatocytes by a filter elution method. Dev Toxicol Environ Sci 11(375):378.
Hsia, M; Kreamer, B; Dolara, P. (1983b) A rapid and simple method to quantitate chemically induced unscheduled
DNA synthesis in freshly isolated rat hepatocytes facilitated by DNA retention of membrane filters. Mutat Res
122(2): 177-186.
IARC (International Agency for Research on Cancer). (1999a) Melamine. In: IARC Monographs on the evaluation
of the carcinogenic risk of chemicals to humans. Some chemicals that cause tumours of the kidney or urinary
bladder in rodents and some other substances. Lyon, France: International Agency for Research on Cancer; pp.
329-338.
IARC. (1999b) Consensus report. In: Capen, CC; Dybing, E: Rice, JM; et al., eds. Species differences in thyroid,
kidney and urinary bladder carcinogenesis. Lyon, France: IARC Sci Publ 147, pp. 1-14.
IARC. (1999c) ortho-Phenylphenol and its sodium salt. In: IARC Monographs on the evaluation of the
carcinogenic risk of chemicals to humans. Some chemicals that cause tumours of the kidney or urinary bladder in
rodents and some other substances. Lyon, France: International Agency for Research on Cancer; pp. 451-480.
Imai, S; Morimoto, J; Sekigawa, S; et al. (1983) Additive toxicity test of thiabenzadole and diphenyl in mice. J Nara
Med Assoc 34:512-522.
Innes, JR; Ulland, BM; Valerio, MG; et al. (1969) Bioassay of pesticides and industrial chemicals for tumorigenicity
in mice: apreliminary note. JNatl Cancer Inst42(6):1101-1114.
Inoue, S; Yamamoto, K; Kawanishi, S. (1990) DNA damage induced by metabolites of o-phenylphenol in the
presence of copper(II)ion. Chem Res Toxicol 3(2): 144-149.
122 DRAFT - DO NOT CITE OR QUOTE
-------
IPCS (International Programme on Chemical Safety). (1999) Biphenyl. Concise international chemical assessment
document (CICAD). Vol. 6. World Health Organization, Geneva, Switzerland. Available online at
http://www.inchem.org/documents/cicads/cicads/cicad06.htm (accessed January 15, 2009).
Ishidate, M, Jr.; Odashima, S. (1977) Chromosome tests with 134 compounds on Chinese hamster cells in vitro - a
screening for chemical carcinogens. Mutat Res 48(3-4):337-354.
Ishidate, M, Jr.; Sofuni, T; Yoshikawa, K; et al. (1984) Primary mutagenicity screening of food additives currently
used in Japan. Food Chem Toxicol 22(8):623-636.
Ito, N; Fukushima, S; Shirai, T; et al. (1984) Drugs food additives and natural products as promoters in rat urinary
bladder carcinogenesis. IARC Sci Publ 56:399-407.
Kawachi, T; Yahagi, T; Kada, T; et al. (1980) Cooperative programme on short-term assays for carcinogenicity in
Japan. IARC Sci Publ 27:323-330.
Khera, KS; Whalen, C; Angers, G; et al. (1979) Assessment of the teratogenic potential of piperonyl butoxide,
biphenyl, and phosalone in the rat. Toxicol Appl Pharmacol 47(2):353-358.
King-Herbert, A; Thayer, K. (2006) NTP workshop: animal models for the NTP rodent cancer bioassay: stocks and
strains-should we switch? Toxicol Pathol 34(6):802-5.
Kitamura, S; Sanoh, S; Kohta, R; et al. (2003) Metabolic activation of proestrogenic diphenyl and related
compounds by rat liver microsomes. J Health Sci 49(4):298-310.
Klaunig, JE; Babich, MA; Baetcke, KP; et al. (2003) PPARa agonist-induced rodent tumors: modes of action and
human relevance. Crit Rev Toxicol 33(6):655-780.
Kluwe, WM. (1982) Development of resistance to nephrotoxic insult: changes in urine composition and kidney
morphology on repeated exposures to mercuric chloride or biphenyl. J Toxicol Environ Health 9(4) :619-63 5.
Kojima, A; Hiraga, K. (1978) Mutagenicity of citrus fungicides in the microbial system. Tokyo Toritsu Eisei
Kenkyusho Nempo 29:83-85
Kokel, D; Xue, D. (2006) A class of benzenoid chemicals suppresses apoptosis in C. elegans. Chembiochem
7(12):2010-2015.
Kurata, Y; Asamoto, M; Hagiwara, A; et al. (1986) Promoting effects of various agents in rat urinary bladder
carcinogenesis initiated by A^-butyl-Ar.-(4-hydroxybutyl) nitrosamine. Cancer Lett 32:125-135.
Kwok, ESC; Bucholz, BA: Vogel, JS; et al. (1999) Dose-dependent binding of or//zo-phenylphenol to protein but
not DNA in the urinary bladder of male F344 rats. Toxicol Appl Pharmacol 159:18-24.
Maronpot, RR. (2009) Biological basis of differential susceptibility to hepatocarcinogenesis among mouse strains. J
Toxicol Pathol 22:11-33.
Matanowki, G; Elliott, CA. (1981) Bladder cancer epidemiology. Epidemiol. Rev. 3:203-228
Matsubara, T; Prough, RA; Burke, MD; et al. (1974) The preparation of microsomal fractions of rodent respiratory
tract and their characterization. Cancer Res 34(9):2196-2203.
McElroy, MC; Postle, AD; Kelly, FJ. (1992) Catalase, superoxide dismutase and glutathione peroxidase activities of
lung and liver during human development. Biochim Biophys Acta 1117:153-158.
Meyer, T. (1977) The metabolism of biphenyl: IV. Phenolic metabolites in the guinea pig and the rabbit. Acta
Pharmacol Toxicol Suppl 40(2): 193-200.
Meyer, T; Scheline, RR. (1976) The metabolism of biphenyl. II. Phenolic metabolites in the rat. Acta Pharmacol
Toxicol 39(4):419-432.
123 DRAFT - DO NOT CITE OR QUOTE
-------
Meyer, T; Larsen, J; Hansen, EV; et al. (1976a) The metabolism of biphenyl. III. Phenolic metabolites in the pig.
Acta Pharmacol Toxicol 3 9(4) :43 3-441.
Meyer, T; Aarbakke, J; Scheline, RR. (1976b) The metabolism of biphenyl. I. Metabolic disposition of 14C-
biphenyl in the rat. Acta Pharmacol Toxicol 39(4):412^U8.
Millburn, P; Smith, RL; Williams, RT. (1967) Biliary excretion of foreign compounds. Biochem J 105:1275-1281.
Mole, ML; Sanders, L; Oglesby, LA. (1988) High-performance liquid chromatographic assay of biphenyl
metabolism by hepatocytes cultured in an embryo/hepatocyte co-culture medium. Anal Biochem 175(l):74-84.
Monsanto (Monsanto Company). (1946) Final report on the physiological response of experimental animals to the
absorption of diphenyl, and several resins, elastomers and plastics with cover letter (sanitized). Submitted under
TSCA Section 8D; EPA Document No. 878213563;NTISNo. OTS0206411.
Monsanto (Monsanto Company). (1956) Chronic oral administration metabolic studies on dogs. Prepared by
Hazleton Laboratories. Submitted under TSCA Section 8D; EPA Document No. 878213568; NTIS No.
OTS0206411.
Monsanto (Monsanto Company). (1976) Toxicological investigation of biphenyl. Prepared by Younger
Laboratories Inc. Submitted under TSCA Section 8D; EPA Document No. 878213572; NTIS No. OTS0206411.
Monsanto (Monsanto Company). (1979) Aqueous solubility and octanol/water partition coefficient with
attachment—biphenyl. Submitted under TSCA Section 8D; EPA Document No. 878213573; NTIS No.
OTS0206411.
Morimoto,K.; Sato, M; Fukuoka, M; et al. (1989) Correlation between the DNA damage in urinary bladder
epithelium and the urinary 2-phenyl-l,4-benzoquinone levels from F344 rats fed sodium o-phenylphenate in the
diet. Carcinogenesis 10:1823-1827.
Nakao,T; Ushiyama, J.;Kabashima, J.; et al. (1983 ) The metabolic profile of sodium o-phenylphenate after
subchronic oral administration to rats. Food Chem Toxicol 21:325-329.
Narbonne, JF; Cassand, P; Alzieu, P; et al. (1987) Structure-activity relationships of the n-methylcarbamate series in
Salmonella typhimurium. MutatRes 191:21-27.
NCI (National Cancer Institute). (1968) Evaluation of carcinogenic, teratogenic and mutagenic activities of selected
pesticides and industrial chemicals. Vol. I. Carcinogenic study. Bethesda, MD: Available from the National
Technical Information Service, Springfield, VA; NTIS PB-223159.
Nishihara, Y. (1985) Comparative study of the effects of biphenyl and Kanechlor-400 on the respiratory and energy
linked activities of rat liver mitochondria. Br J Ind Med 42(2): 128-132.
Nishioka, H; Ogasawara, H. (1978) Mutagenicity testing for di phenyl derivatives in bacterial systems. Mutat Res
54:248-249.
NLM (National Library of Medine). (2007) Biphenyl. Hazardous Substances Data Bank (HSDB). National
Institutes of Health, U.S. Department of Health and Human Services, Bethesda, MD. Available online at
http://toxnet.nlm.nih.gov.
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press. Available online at http://books.nap.edu/books/POD115/html/index.html
(accessed November 2, 2009).
Ohnishi, M; Take, M; Sagawa, M; et al. (1998) Analysis of the components of biphenyl induced urinary bladder
calculus in male rats. Jpn J Toxicol Environ Health 44:256-263.
124 DRAFT - DO NOT CITE OR QUOTE
-------
Ohnishi, M; Yajima, H; Takemura, T; et al. (2000a) Characterization of hydroxy-biphenyl-O-sulfates in urine and
urine crystals induced by biphenyl and KHCO3 administration in rats. J Health Sci 46(4):299-303.
Ohnishi, M; Yajima, H; Yamamoto, S; et al. (2000b) Sex dependence of the components and structure of urinary
calculi induced by biphenyl administration in rats. Chem Res Toxicol 13(8):727-735.
Ohnishi, M; Yajima, H; Takeuchi, T; et al. (2001) Mechanism of urinary tract crystal formation following biphenyl
treatment. Toxicol Appl Pharmacol 174(2): 122-129.
Pacifici, GM; Vannucci, L; Bencini, C; etal. (1991) Sulphation of hydroxybiphenyls in human tissues. Xenobiotica
21:1113-1118.
Pagano, G; Esposito, A; Giordano, GG; et al. (1983) Genotoxicity and teratogenicity of diphenyl and diphenyl ether:
a study of sea urchins, yeast, and Salmonella typhimurium. Teratog Carcinog Mutagen 3(4):377-393.
Pagano, G; Cipollaro, M; Corsale, G; et al. (1988) Comparative toxicity of diphenyl, diphenyl ester, and some of
their hydroxy derivatives. Medicine Biologie Environment 16:291-297.
Parkinson, A; Ogilvie, BW. (2008) Biotransformation of xenobiotics. In: Klaassen, CD, ed. Casarett & Doull's
Toxicology. The basic science of poisons. New York, NY: McGraw-Hill Companies, Inc., pp. 161, 236-237.
Paterson, P; Fry, JR. (1985) Influence of cytochrome P-450 type on the pattern of conjugation of 4-hydroxybiphenyl
generated from biphenyl or 4-methoxybiphenyl. Xenobiotica 15:493-502.
Pathak, DN; Roy, D. (1993) In vivo genotoxicity of sodium ortho-phenylphenol: phenylbenzoquinone is one of the
DNA-binding metabolite(s) of sodium ortho-phenylphenol. Mutat Res 286(2):309-319.
Pecchiai, L; Saffiotti, U. (1957) Studio della tossicita' del difenile, dell' ossidifenile, e della loro miscela
("Dowtherm") [Study of the toxicity of biphenyl, oxydiphenyl and their mixture (Dowtherm)]. Med Lav 48(4) :247-
254.
Powis, G; Jardine, I; Van Dyke, R; et al. (1988) Foreign compound metabolism studies with human liver obtained as
surgical waste relation to donor characteristics and effects of tissue storage. Drug Metab Dispos 16(4):582-589.
Powis, G; Melder, DC; Wilke, TJ. (1989) Human and dog, but not rat, isolated hepatocytes have decreased foreign
compound-metabolizing activity compared to liver slices. Drug Metab Dispos 17:526-531.
Probst, G; McMahon, R; Hill, L; et al. (1981) Chemically-induced unscheduled DNA synthesis in primary rat
hepatocyte cultures: a comparison with bacterial mutagenicity using 218 compounds. Environ Mutagen 3:11-32.
Purchase, IF; Longstaff, E; Ashby, J; et al. (1978) An evaluation of 6 short-term tests for detecting organic chemical
carcinogens. Br J Cancer 37(6):873-903.
Reitz, RH; Fox, TR; Quast, JF; et al. (1983) Molecular mechanisms involved in the toxicity of orthophenylphenol
and its sodium salt. Chem Biol Interact 43:99-l 19.
Rencuzogullari, E; Parlak, S; Basrilla, H. (2008) The effects of food protector biphenyl on sister chromatid
exchange, chromosome aberrations, and micronucleus in human lymphocytes. Drug Chem Toxicol 31:263-274.
Sasaki, YF; Saga, A; Akasaka, M; et al. (1997) In vivo genotoxicity of ortho-phenylphenol, biphenyl, and
thiabendazole detected in multiple mouse organs by the alkaline single cell gel electrophoresis assay. Mutat Res
395(2-3): 189-198.
Sasaki, YF; Kawaguchi, S; Kamaya, A; et al. (2002) The comet assay with 8 mouse organs: results with 39 currently
used food additives. Mutat Res 519:103-119.
Seppalainen, AM; Hakkinen, I. (1975) Electrophysiological findings in diphenyl poisoning. J Neurol Neurosurg
Psychiat 38(3):248-252.
125 DRAFT - DO NOT CITE OR QUOTE
-------
Shibata, MA; Tanaka, H; Yamada, M; et al. (1989a) Proliferative response of renal pelvic epithelium in rats to oral
administration of ortho-phenylphenol, sodium ortho-phenyrphenate and diphenyl. Cancer Lett 48(1): 19-28.
Shibata, M-A; Yamada, M; Tanaka, H; et al. (1989b) Changes in urine composition, bladder epithelial morphology,
and DNA synthesis in male F344 rats in response to ingestion of bladder tumor promoters. Toxicol Appl Pharmacol
99:37-49.
Shiraiwa, K; Takita, M; Tsutsumi, M; et al. (1989) Diphenyl induces urolithiasis but does not possess the ability to
promote carcinogenesis by N-ethyl-N-hydroxyethylnitrosamine in kidneys of rats. J Toxicol Pathol 2:41-48.
Smith, RA; Christenson, WR; Bartels, MJ; et al. (1998) Urinary physiologic and chemical metabolic effects on the
urothelial cytotoxicity and potential DNA adducts of o-phenylphenol in male rats. Toxicol Appl Pharmacol
150:402^13.
Snyder, R; Matheson, D. (1985) Nick translation-a new assay for monitoring DNA damage and repair in cultured
human fibroblasts. Environ Mutagen 7:267-279.
Sofuni, T; Hayashi, M; Matsuoka, A; et al. (1985) Mutagenicity tests on organic chemical contaminants in city
water and related compounds II. Chromosome aberration tests in cultured mammalian cells. Bull Natl Inst Hyg Sci
(Tokyo) 103:64-75.
Sendergaard, D; Blom, L. (1979) Poly cystic changes in rat kidney induced by biphenyl fed in different diets. Arch
Toxicol Suppl (2):499-502.
Sonnier, M; Cresteil, T. (1998) Delayed ontogenesis of CYP1A2 in the human liver. Eur JBiochem 251:893-898.
Strassburg, CP; Strassburg, A; Kneip, S; et al. (2002) Developmental aspects of human hepatic drug glucuronidation
in young children and adults. Gut 50(2):259-65.
Stuehmeier, G; Legrum, W; Netter, KJ. (1982) Does cobalt pretreatment of mice induce a phenobarbitone-type
cytochrome P-450? Xenobiotica 12(5):273-282.
Sun Company Inc. (1977a) Acute inhalation toxicity of biphenyl with cover letter. Prepared by Cannon
Laboratories, Inc. Submitted under TSCA Section 8D; EPA Document No. 878213530; NTIS No. OTS0206401.
Sun Company Inc. (1977b) Initial submission: 90-day inhalation toxicity study of biphenyl (99 + % purity) in GDI
mice (final report) with cover letter dated 022892. Prepared by Cannon Laboratories, Inc. Submitted under TSCA
Section 8ECP; EPA Document No. 88-920001856; NTIS No. OTS0539116.
Sunouchi, M; Miyajima, A; Ozawa, S; et al. (1999). Effects of diphenyl on hepatic peroxysomal enzyme and drug-
metabolizing enzyme activities in BDF 1 mice. J Toxicol Sci 24:333.
Takita, M. (1983) Urolithiasis induced by oral administration of diphenyl in rats. J Nara Med Univ Med Assoc
34:565-584.
Tamano, S; Asakawa, E; Boomyaphiphat, P; et al. (1993) Lack of promotion of N-butyl-N-(4-hydroxybutyl)
nitrosamine-initiated urinary bladder carcinogenesis in mice by rat cancer promoters. Teratog Carcinog Mutagen
13(2):89-96.
Tan, Y; Yamada-Mabuchi, M; Arya, R; et al. (2011) Coordinated expression of cell death genes regulates neuroblast
apoptosis. Development 138(11):2197-206.
Tani, S; Yonezawa, Y; Morisawa, S; et al. (2007) Development of a new E. coli strain to detect oxidative mutation
and its application to the fungicide o-phenylphenol and its metabolites. Mutat Res 628(2): 123-128.
Terlecky, SR; Koepke, JI; Walton, PA. (2006) Peroxisomes and aging. Biochim Biophys Acta 1763:1749-1754.
Umeda, Y; Arito, H; Kano, H; et al. (2002) Two-year study of carcinogenicity and chronic toxicity of biphenyl in
rats. JOccup Health44(3):176-183.
126 DRAFT - DO NOT CITE OR QUOTE
-------
Umeda, Y; Also, S; Arito, H; et al. (2004) Induction of peroxisome proliferation in the liver of biphenyl-fed female
mice. JOccup Health 46(6):486-488.
Umeda, Y; Aiso, S; Yamazaki, K; et al. (2005) Carcinogenicity of biphenyl in mice by two years feeding. J Vet
MedSci67(4):417-424.
Union Carbide. (1949) Range finding tests on diphenyl. Tables of protocols attached. With cover letter. Prepared by
Mellon Institute. Submitted under TSCA Section 8D; EPA Document No. 878213680; NTIS No. OTS0206426.
U.S. EPA (Environmental Protection Agency). (1978) 40 CFR Part 761, Subpart B (Manufacturing, Processing,
Distribution in Commerce and Use of PCBs and PCB Items. Available online at
http://www.epa.gov/wastes/hazard/tsd/pcbs/pubs/laws.htm (accessed June 28, 2011).
U.S. EPA. (1986a) Guidelines for the health risk assessment of chemical mixtures. Federal Register 51(185):34014-
34025. Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1986b) Guidelines for mutagenicity risk assessment. Federal Register 51(185):34006-34012. Available
online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1988) Recommendations for and documentation of biological values for use in risk assessment.
Prepared by the Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH for the Office of Solid Waste and Emergency Response, Washington, DC; EPA 600/6-87/008.
Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1991) Guidelines for developmental toxicity risk assessment. Federal Register 56(234):63798-63826.
Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1994a) Interim policy for particle size and limit concentration issues in inhalation toxicity studies.
Federal Register 59(206):53799. Available online at
http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1994b) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry. Office of Research and Development, Washington, DC; EPA/600/8-90/066F. Available online at
http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 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 online at
http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=42601 (accessed January 15, 2009).
U.S. EPA. (1996) Guidelines for reproductive toxicity risk assessment. Federal Register 61(212):56274-56322.
Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (1998) Guidelines for neurotoxicity risk assessment. Federal Register 63(93):26926-26954. Available
online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
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. Available online at
http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (2000b) Benchmark dose technical guidance document [external review draft]. Risk Assessment Forum,
Washington, DC; EPA/630/R-00/001. Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January
15, 2009).
U.S. EPA. (2000c) Supplementary guidance for conducting for health risk assessment of chemical mixtures. Risk
Assessment Forum, Washington, DC; EPA/630/R-00/002. Available online at http://www.epa.gov/iris/backgr-d.htm
(accessed January 15, 2009).
U.S. EPA. (2002) A review of the reference dose and reference concentration processes. Risk Assessment Forum,
127 DRAFT - DO NOT CITE OR QUOTE
-------
Washington, DC; EPA/630/P-02/0002F. Available online at http://www.epa.gov/iris/backgr-d.htm (accessed
January 15, 2009).
U.S. EPA. (2005a) Guidelines for carcinogen risk assessment. Risk Assessment Forum, Washington, DC;
EPA/630/P-03/001B. Available online at http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
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. Available online at http://www.epa.gov/iris/backgr-
d.htm (accessed January 15, 2009).
U.S. EPA. (2006a) Science policy council handbook: peer review. Third edition. Office of Science Policy, Office of
Research and Development, Washington, DC; EPA/100/B-06/002. Available online at
http://www.epa.gov/iris/backgr-d.htm (accessed January 15, 2009).
U.S. EPA. (2006b) A framework for assessing health risk of environmental exposures to children. National Center
for Environmental Assessment, Washington, DC, EPA/600/R-05/093F. Available online at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363 (accessed January 15, 2009).
Wangenheim, J; Bolcsfoldi, G. (1986) Mouse lymphoma tk+/- assay of 30 compounds. Environ Mutagen 8(Suppl.
6):90.
Wangenheim, J; Bolcsfoldi, G. (1988) Mouse lymphoma L5178Y thymidine kinase locus assay of 50 compounds.
Mutagenesis 3(3): 193-205.
Wastensson, G; Hagberg, S; Andersson, E; et al. (2006) Parkinson's disease in diphenyl-exposed workers—a causal
association? Parkinsonism Relat Disord 12(l):29-34.
Waters, MD; Sandhu, SS; Simmon, V; et al. (1982) Study of pesticide genotoxicity. Basic Life Sci 21:275-326.
Westinghouse (Westinghouse Electric Corporation). (1977) Potential carcinogenicity testing of PCB replacements
using the Ames test with cover letter. Submitted under TSCA Section 8D; EPA Document No. 878214672; NTIS
No. OTS0206616.
Wiebkin, P; Fry, JR; Jones, CA; et al. (1976) The metabolism of biphenyl by isolated viable rat hepatocytes.
Xenobiotica 6(12):725-743.
Wiebkin, P; Fry, JR; Jones, C; et al. (1978) Biphenyl metabolism in isolated rat hepatocytes: effect of induction and
nature of the conjugates. Biochem Pharmacol 27:1899-1907.
Wiebkin, P; Schaeffer, B; Longnecker, D; et al. (1984) Oxidative and conjugative metabolism of xenobiotics by
isolated rat and hamster acinar cells. Drug Metab Dispos 12(4):427-431.
Williams, G. (1980) DNA repair and mutagenesis in liver cultures as indicators in chemical carcinogen screening,
in: mammalian cell transformation by chemical carcinogens. Adv Mod Environ Toxicol 1:273-296.
Williams, G; Mori, H; McQueen, C. (1989) Structure-activity relationships in the rat hepatocyte DNA-repair test for
300 chemicals. Mutat Res 221:263-286.
Yoshida, S; Masubuchi, M; Hiraga, K. (1978) Cytogenetic studies of antimicrobials on cultured cells. Tokyo
Toritsu Eisei Kenkyusho Kenkyo Nempo (Annu Rep Tokyo Metrop Res Lab Public Health) 29(2):86-88.
128 DRAFT - DO NOT CITE OR QUOTE
-------
1 APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
2 COMMENTS AND DISPOSITION
3
4
5 [Page intentionally left blank]
A-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 APPENDIX B. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE DOSE
2
3
4
5
6
7
8
9
10
11
Datasets used for modeling incidences of nonneoplastic effects in the urinary tract of
male and female F344 rats exposed to biphenyl in the diet for 2 years (Umeda et al., 2002) are
shown in Table B-l. Datasets used for modeling body weight data, selected clinical chemistry
results, and histopathological kidney effects in male and female BDFi mice exposed to biphenyl
in the diet for 2 years (Umeda et al., 2005) are shown in Table B-2. The dataset for incidence of
litters with fetal skeletal anomalies, tallied from evaluation of fetuses from Wistar rat dams
administered biphenyl by gavage on GDs 6-15 (Khera et al., 1979) is shown in Table B-3.
Table B-l. BMD modeling datasets for incidences of nonneoplastic effects in
the urinary tract of male and female F344 rats exposed to biphenyl in the
diet for 2 years
Biphenyl dietary concentration (ppm)
TWA body weight (kg)a
Calculated dose (mg/kg-d)b
Effect
Renal pelvis
Nodular transitional cell hyperplasia
Simple transitional cell hyperplasia
Mineralization
Other kidney effects
Hemosiderin depositf
Papillary mineralization
Males (n = 50)
0
0.411
0
500
0.412
36.4
1,500
0.408
110
4,500
0.357
378
Females (n = 50)
0
0.251
0
500
0.246
42.7
1,500
0.246
128
4,500
0.216
438
0
6
9
1
8
6
1
5
10
21C
19d
18e
0
3
12
0
5
12
1
12d
18
12C
25C
27d
0
9
0
9
0
14
0
23d
4
2
8
6
22C
3
25C
12C
Bladder
Combined transitional cell hyperplasia8
0
0
0
45
1
0
1
10
aTWA body weight calculated using graphically-presented body weight data from Umeda et al. (2002).
bCalculated doses based on TWA body weights and chronic reference food consumption values for F344 rats
(0.030 kg/day for males and 0.021 kg/day for females; taken from Table 1-6 of U.S. EPA, 1988).
Significantly different from control group (p < 0.01) according to %2 test.
dSignificantly different from control group (p < 0.05) according to %2 test.
Significantly different from controls (p < 0.05) according to Fisher's exact test.
fMale data for incidences of hemosiderin deposits not selected for quantitative analysis..
8Female data for incidences of combined transitional cell hyperplasia not selected for quantitative analysis.
Source: Umeda et al. (2002).
12
B-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. BMD modeling datasets for body weight, selected clinical
chemistry results, and histopathological kidney effects in male and female
BDFi mice exposed to biphenyl in the diet for 2 years
Endpoint
Biphenyl concentration in the diet (ppm)
0
667
2,000
6,000
Males
Dose (mg/kg-d)
Histopathological kidney effect
Mineralization inner stripe-outer medulla
Clinical chemistry parameter
BUN (mg/dL)
Body weight
Mean terminal body weight (g)
0
n = 50
9
n = 34
20.2 ±3.6
n=35
46.9 ±4.9
97
n=49
8
n=39
22.0 ±4.0
n=41
43.1 ±7.9
291
n=50
14
n=37
23.2±4.4a
n=41
42.9±6.0a
1,050
n=50
14
n=37
22.9±2.7b
n=39
32.4±3.6b
Females
Dose (mg/kg-d)
Histopathological kidney effect
Mineralization inner stripe-outer medulla
Clinical chemistry parameter
AST (IU/L)
ALT (IU/L)
AP (IU/L)
LDH (IU/L)
BUN (mg/dL)
Body weight
Mean terminal body weight (g)
0
n = 50
o
J
n = 28
75 ±27
32 ±18
242 ± 90
268 ± 98
14.9 ±2.0
n=31
34.0 ±4.0
134
n=50
5
n=20
120 ±110
56 ±46
256 ±121
461 ±452
14.8 ±3.4
n=22
32.5 ±3. 3
414
n=50
12C
n=22
211±373b
134±231b
428 ± 499
838 ± 2,000
21.0 ±20.5
n=25
30.5 ± 3. lb
1,420
n=49
26d
n=31
325 ± 448b
206 ± 280b
556 ± 228b
1,416 ±4,161a
23.8±11.7b
n=32
25.5±3.0b
"Significantly different from controls (p < 0.05) according to Dunnett's test.
bSignificantly different from controls (p < 0.01) according to Dunnett's test.
Significantly different from controls (p < 0.05) according to Fisher's exact test.
dSignificantly different from controls (p < 0.01) according to Fisher's exact test.
ALT (GPT) = alanine aminotransferase (glutamic pyruvic transaminase); AP (ALP) = alkaline phosphatase;
AST (GOT) = aspartate aminotransferase (glutamic oxaloacetic transaminase)
Source: Umeda et al. (2005).
B-2
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-3. BMD modeling dataset for incidence of litters with fetal skeletal
anomalies from Wistar rat dams administered biphenyl by gavage on
CDs 6-15
Effect
Litters with fetal skeletal anomaliesVlitters
examined
Dose (mg/kg-d)
0
8/16
125
11/20
250
13/18
500
15/18b
1,000
6/9
1
2
3
4
5
6
7
aThe study authors reported one runted fetus in the control group and one fetus with kinky tail in the 250 mg/kg-day
dose group, which may have influenced the reported incidence data for anomalous litters/litters examined.
bSignificantly different from controls (p < 0.05) according to Fisher's exact test conducted for this review.
Source: Kheraetal. (1979).
Goodness of fit statistics and benchmark results for each of the modeled biphenyl-
induced nonneoplastic effects from the chronically-exposed rats (Umeda et al., 2002) and mice
(Umeda et al., 2005) and the gestationally-exposed rats (Khera et al., 1979) are summarized in
Tables B-4 through B-22. Each table of modeled results for a particular effect is followed by the
information from the output file of the best-fitting model for that effect.
Table B-4. Summary of BMD modeling results for incidence of renal
nodular transitional cell hyperplasia in male F344 rats exposed to biphenyl
in the diet for 2 years
Model
Gammab
Logistic
Log-Logistic13
Log-Probitb
Multistage (3-degree)c'd
Probit
Weibmf
Goodness of fit
%2/7-valuea
0.31
0.64
0.31
0.31
0.58
0.59
0.31
Largest
residual
0.73
0.74
0.74
0.71
0.84
0.84
0.75
AIC
95.02
92.72
95.01
95.03
92.60
92.76
95.00
Benchmark result (mg/kg-d)
BMD5
169.71
178.92
172.40
163.38
133.82
157.59
175.08
BMDL5
74.44
133.35
75.93
89.50
69.08
117.53
73.08
BMD10
212.00
233.81
216.08
202.25
193.30
212.09
221.75
BMDL10
120.62
192.35
120.70
128.71
126.95
173.76
121.01
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
B-3
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
6
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.3
0 50 100 150 200 250 300 350
0.2
0.1
10:3801/122011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:\USEPA\IRIS\biphenyl\rat\renalnodularhyper\male\mst_nodhypMrev_MS_3.(d)
Gnuplot Plotting File:
C:\USEPA\IRIS\biphenyl\rat\renalnodularhyper\male\mst_nodhypMrev_MS_3.plt
Wed Jan 12 10:38:57 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*dose^l-beta2*dose^2-beta3*dose^3) ]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 4
Total number of specified parameters = 0
Degree of polynomial = 3
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 = 0.00721859
Beta(l) = 3.68302e-005
Beta(2) = 0
Beta(3) = 9.69211e-009
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Beta(2) have been estimated at a boundary point, or
have been specified by the user, and do not appear in the correlation matrix )
Beta(l) Beta(3)
Beta(l) 1 -0.95
Beta(3) -0.95 1
Variable
Background 0 *
Beta(l) 0.000234424 *
Beta(2) 0 *
Beta(3) 8.31393e-009 *
- Indicates that this value is not calculated.
Analysis of Deviance Table
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
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -43.8185 4
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 193.298
BMDL = 126.946
BMDU = 248.35
Taken together, (126.946, 248.35 ) is a 90% two-sided confidence interval for the BMD
Table B-5. Summary of BMD modeling results for incidence of renal
nodular transitional cell hyperplasia in female F344 rats exposed to biphenyl
in the diet for 2 years
Model
Gammab
Logistic
Log-Logistic13
Log-Probitb
Multistage (2-degree)c'd
Probit
Weibullb
Goodness of fit
%2/7-valuea
0.96
0.69
0.96
0.99
0.99
0.76
0.95
Largest
residual
-0.24
0.63
-0.26
-0.15
-0.36
0.54
-0.27
AIC
69.04
69.93
69.07
68.96
67.19
69.69
69.08
Benchmark result (mg/kg-d)
BMDS
200.54
277.38
203.45
188.92
191.47
253.65
207.16
BMDLS
118.95
211.02
118.10
134.61
121.69
190.94
119.11
BMD10
276.46
343.52
279.78
261.35
274.42
324.08
285.37
BMDL10
198.73
289.03
196.91
193.58
211.52
268.17
201.63
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
28
B-5
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
0.4
0.35
0.3
T3 0.25
-ffi
8
< 0.2
£=
g
I 0.15
Ll_
0.1
0.05
0
Multistage
BMDL
BMD
11:4801/132011
0 50 100 150 200 250 300 350 400 450
dose
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalnodularhyper/female/mst_nodhypFrev_MS_2.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalnodularhyper/female/mst_nodhypFrev_MS_2.pit
Thu Jan 13 11:48:49 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*dose^l-beta2*dose^2)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Beta(l) have been estimated at a boundary point, or
have been specified by the user, and do not appear in the correlation matrix )
Beta(2)
Beta(2) 1
Variable
Background 0 *
Beta(l) 0 *
Beta(2) 1.39908e-006 *
- Indicates that this value is not calculated.
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
Fitted model -32
Reduced model -48
AIC: 67
Dose Est. Prob.
0.0000 0.0000
42.7000 0.0025
128.0000 0.0227
438.0000 0.2354
Chi~2 = 0.15 d.f. =
Benchmark Dose Computation
Specified effect =
.5947
.1018
.1895
Goodn
Expected
0.000
0.127
1.133
11.770
3 P-va
0.1
1 0.277585 3 0.9642
1 31.2917 3 <.0001
ess of Fit
Scaled
Observed Size Residual
0.000 50 0.000
0.000 50 -0.357
1.000 50 -0.126
12.000 50 0.077
lue = 0. 9853
Risk Type = Extra risk
Confidence level =
BMD =
BMDL =
BMDU =
Taken together, (211.518,
0. 95
274 . 422
211.518
351. 444
351.444) is a
90% two-sided confidence interval for the BMD
Table B-6. Summary of BMD modeling results for incidence of renal simple
transitional cell hyperplasia in male F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab'c
Logistic
Log-Logistic13
Log-Probitb
Multistage (3 -degree)*1
Probit
Weibullb
Goodness of fit
%2/7-valuea
0.66
0.35
0.36
0.36
0.60
0.33
0.36
Largest
residual
0.71
-1.18
0.71
0.71
0.74
-1.22
0.71
AIC
184.41
185.78
186.41
186.41
184.59
185.92
186.41
Benchmark result (mg/kg-d)
BMD5
284.70
96.07
320.26
284.12
201.02
90.26
324.89
BMDL5
55.27
73.33
58.80
100.23
52.30
68.00
55.27
BMD10
313.76
171.37
340.21
312.44
255.53
164.29
344.08
BMDL10
113.22
131.76
115.09
144.14
107.40
124.13
113.14
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit differed by less than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
27
B-7
DRAFT - DO NOT CITE OR QUOTE
-------
Gamma Multi-Hit Model with 0.95 Confidence Level
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.1
Gamma Multi-Hit
BMDL
BMD
0 50 100 150 200 250 300 350
dose
11:5501/132011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Gamma Model. (Version: 2.15; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalsimplehyper/male/gam_rensimphypMrev_gamma.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalsimplehyper/male/gam_rensimphypMrev_gamma.plt
Thu Jan 13 11:55:07 2011
BMDS_Model_Run
The form of the probability function is: P[response]= background+(1-
background)*CumGamma[slope*dose,power], where CumGamma(.) is the cummulative Gamma distribution
function
Dependent variable = incidence
Independent variable = dose
Power parameter is restricted as power >=1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.134615
Slope = 0.00398471
Power = 2.55235
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Power have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
Background Slope
Background
Slope
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.126666 0.0271566 0.0734404 0.179892
0.0408652 0.00241924 0.0361236 0.0456068
18 NA
NA - 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) # Param's Deviance Test d.f.
Full model -89.7871 4
Fitted model -90.2033 2 0.832451 2
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
Reduced model -97.2446 1
AIC: 184.407
Goodness of Fit
Dose Est._Prob. Expected Observed Size
Scaled
Residual
BMDL =
Table B-7. Summary of BMD modeling results for incidence of renal simple
transitional cell hyperplasia in female F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab, Weibullb,
Multistage (l-degree)c,d
Logistic
Log-Logistic13
Log-Probitb
Probit
Goodness of fit
%2/7-valuea
0.89
0.28
0.71
0.41
0.33
Largest
residual
0.34
1.29
-0.26
1.00
1.22
AIC
183.87
186.14
185.77
185.39
185.77
Benchmark result (mg/kg-d)
BMD5
34.63
83.08
37.52
84.12
75.68
BMDL5
25.35
66.43
18.90
62.52
60.94
BMD10
71.12
145.87
71.51
120.97
135.30
BMDL10
52.08
119.22
39.91
89.91
110.85
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the gamma and Weibull models took the form of a 1-degree polynomial multistage model and
produced identical goodness of fit statistics and BMD values; the model with the lowest AIC was selected because
BMDL values for models providing adequate fit differed by less than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
23
B-9
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage
BMDL BMD
0 50 100 150 200 250 300 350 400 450
dose
14:01 01/132011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalsimplehyper/female/mst_simplehypFrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalsimplehyper/female/mst_simplehypFrev_MS_l.pit
Thu Jan 13 14:01:13 2011
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 = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.0607741
Beta(l) = 0.00145231
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background
Beta(l)
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.057038 * * *
- Indicates that this value is not calculated.
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f. P-value
-89.8139 4
-89.9369 2 0.246113 2 0.8842
-106.633 1 33.6378 3 <.0001
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
Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 71.1248
BMDL = 52.0766
BMDU = 105.072
Taken together, (52.0766, 105.072) is a 90% two-sided confidence interval for the BMD
Table B-8. Summary of BMD modeling results for incidence of
mineralization in renal pelvis of male F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab
Logistic
Log-Logistic13
Log-Probitb'c
Multistage (l-degree)d
Probit
Weibullb
Goodness of fit
% /7-value3
0.35
0.58
0.34
0.64
0.51
0.57
0.34
Largest
residual
-0.75
-0.79
-0.75
-0.74
-0.84
-0.80
-0.75
AIC
206.13
204.33
206.14
204.13
204.60
204.35
206.15
Benchmark result (mg/kg-d)
BMD5
130.11
98.62
128.13
144.55
70.84
94.16
131.37
BMDLS
42.91
70.79
36.96
96.05
41.20
66.44
42.84
BMD10
201.71
181.36
199.42
207.88
145.51
175.86
205.20
BMDL10
88.15
130.04
78.03
138.13
84.62
123.70
88.00
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
21
B-ll
DRAFT - DO NOT CITE OR QUOTE
-------
LogProbit Model with 0.95 Confidence Leve
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
I
<
0.3
50 100 150 200 250 300 350
0.2
0.1
15:3801/132011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Probit Model. (Version: 3.2; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalmineral/male/lnp_minpelvMrev_logprobit.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalmineral/male/lnp_minpelvMrev_logprobit.pit
Thu Jan 13 15:38:28 2011
BMDS Model Run
The form of the probability function is: P[response] = Background + (1-Background) *
CumNorm(Intercept+Slope^Log(Dose)), where CumNorm(.) is the cumulative normal distribution
function
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0.18
intercept = -6.59931
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
background intercept
background 1 -0.46
intercept -0.46 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.157045 0.0325697 0.0932095 0.22088
-6.61851 0.281947 -7.17111 -6.0659
1 NA
NA - Indicates that this parameter has hit a bound implied by some ineguality constraint and thus
has no standard error.
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-99.607 4
-100.063 2 0.91202 2
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
Reduced model -104.101 1 8.98864
AIC: 204.126
Goodness of Fit
Dose Est._Prob. Expected Observed Size
Scaled
Residual
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 207.879
BMDL =
Table B-9. Summary of BMD modeling results for incidence of
mineralization in renal pelvis of female F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab
Logistic
Log-Logistic13
Log-Probitb
Multistage (l-degree)c'd
Probit
Weibullb
Goodness of fit
X2/>-valuea
0.57
0.76
O.001
<0.001
0.85
0.77
0.56
Largest
residual
-0.43
0.59
2.90
2.90
-0.44
0.57
-0.44
AIC
250.89
249.10
263.72
263.72
248.89
249.08
250.89
Benchmark result (mg/kg-d)
BMDS
44.66
64.48
1.33 x 1015
1.54 x 1014
42.68
62.20
43.32
BMDL5
27.40
48.11
NA
NA
27.40
46.34
27.40
BMD10
90.32
123.84
1.58 x 1015
2.21 x 1014
87.67
120.41
88.56
BMDL10
56.28
92.31
NA
NA
56.28
89.56
56.28
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
23
B-13
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
I
<
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage
BMDL BMD
0 50 100 150 200 250 300 350 400 450
dose
16:2401/132011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalmineral/female/mst_minpelvlFrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalmineral/female/mst_minpelvlFrev_MS_l.pit
Thu Jan 13 16:24:18 2011
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 = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.230737
Beta(l) = 0.00118679
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(1)
Background 1 -0.62
Beta(l) -0.62 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.228898 * * *
Variable
Background
Beta(l) 0.0012018 *
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log (likelihood) # Param's Deviance Test d.f. P-value
Full model -122.276 4
Fitted model -122.443 2 0.334544 2 0.846
Reduced model -128.859 1 13.1664 3 0.00429
AIC :
B-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
Prob.
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
Table B-10. Summary of BMD modeling results for incidence of
hemosiderin deposits in the kidney of female F344 rats exposed to biphenyl
in the diet for 2 years
Model
Gammab, Weibullb,
Multistage (1 -degree)0
Logistic
Log-Logistic13
Log-Probitb
Probit
Dichotomous-Hilld'e
Goodness of fit
%2/7-valuea
0.022
0.002
0.093
0.002
0.002
0.9997
Largest
residual
2.36
2.92
1.75
2.82
2.90
0.026
AIC
220.99
225.98
218.35
225.97
225.57
213.75
Benchmark result (mg/kg-d)
BMDS
29.64
66.06
19.21
74.77
61.90
34.28
BMDL5
21.20
52.04
12.74
52.43
49.07
12.76
BMD10
60.87
123.37
40.56
107.53
116.90
45.32
BMDL10
43.54
97.71
26.89
75.40
92.96
23.29
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the only model with an adequate fit (%2/>-value > 0.1).
ev = 0.5 (specified), g = 0.16 (specified), intercept = 0.08 (initialized), slope = 1 (initialized).
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
21
B-15
DRAFT - DO NOT CITE OR QUOTE
-------
Dichotomous-Hill Model with 0.95 Confidence Level
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Dichotomous-h
BMDL 3MD
0 100 200 300 400
dose
09:1401/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Dichotomous Hill Model. (Version: 1.2; Date: 12/11/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/hemosiderin/female/dhl_hemosidFrev_dichotomous
hill.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/hemosiderin/female/dhl_hemosidFrev_dichotomous
hill.pit
Fri Jan 14 09:14:35 2011
The form of the probability function is: P[response] = v*g +(v-v*g)/[1+EXP(-intercept-
slope^Log(dose))] where: 0 <= g < 1, 0 < v <= Iv is the maximum probability of response predicted
by the model, and v*g is the background estimate of that probability.
Dependent variable = incidence
Independent variable = dose
Parameter v is set to 0.5
Parameter g is set to 0.16
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
v = -9999 Specified
g = -9999 Specified
intercept = 0.08
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -v -g have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
intercept slope
intercept 1 -0.99
slope -0.99 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
-12.5334 5.83724 -23.9742 -1.09265
2.95297 1.43635 0.137773 5.76817
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
Log(likelihood) # Param's Deviance Test d.f. P-value
-104.876 4
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 45.3249
BMDL = 23.2881
Table B-ll. Summary of BMD modeling results for incidence of papillary
mineralization in the kidney of male F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab
Logistic
Log-Logistic13
Log-Probitb
Multistage (l-degree)c'd
Probit
Weibullb
Goodness of fit
%2/7-valuea
0.63
0.81
0.001
0.001
0.88
0.82
0.63
Largest
residual
-0.37
0.51
2.93
2.93
-0.40
0.48
-0.37
AIC
228.81
226.99
241.27
239.27
226.82
226.96
228.81
Benchmark result (mg/kg-d)
BMDS
51.08
70.07
5.64 x 1014
5.13 x 1013
44.66
66.59
49.89
BMDLS
28.48
52.70
NA
NA
28.45
49.79
28.47
BMD10
99.83
131.45
6.68 x 1014
7.38 x 1013
91.74
126.42
98.66
BMDL10
58.49
98.95
NA
NA
58.44
94.42
58.48
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
26
B-17
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
I
<
0.6
0.5
0.4
0.3
0.2
0.1
Multistage
BMDL
BMD
0
50 100 150 200 250 300 350
dose
11:2501/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Model. (Version: 3.2; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/male/mst_papminMrev_MS_l.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/male/mst_papminMrev_MS_l.pit
Fri Jan 14 11:25:01 2011
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 = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.168963
Beta(l) = 0.00114658
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.62
Beta(l) -0.62 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.168634 * * *
0.00114846 * * *
Variable
Background
Beta(l)
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log (likelihood) # Param's Deviance Test d.f. P-value
Full model -111.284 4
Fitted model -111.409 2 0.250221 2 0.8824
Reduced model -117.634 1 12.6991 3 0.005335
AIC :
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
Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 91.741
BMDL = 58.4361
BMDU = 182. 915
Taken together, (58.4361, 182.915) is a 90% two-sided confidence interval for the BMD
Table B-12. Summary of BMD modeling results for incidence of papillary
mineralization in the kidney of female F344 rats exposed to biphenyl in the
diet for 2 years
Model
Gammab
Logistic0
Log-Logistic13
Log-Probitb
Multistage (l-degree)d
Probit
Weibullb
Goodness of fit
%2/7-valuea
0.11
0.23
0.11
0.11
0.21
0.23
0.11
Largest
residual
1.27
1.37
1.27
1.27
1.28
1.36
1.27
AIC
139.76
138.04
139.76
139.76
138.38
138.08
139.76
Benchmark result (mg/kg-d)
BMD5
360.00
175.24
388.83
356.94
113.15
164.88
391.23
BMDL5
68.91
129.91
61.62
150.95
65.01
119.64
68.91
BMD10
397.57
292.33
413.84
395.27
232.43
282.98
415.47
BMDL10
141.55
219.17
130.08
217.08
133.53
206.34
141.55
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
21
B-19
DRAFT - DO NOT CITE OR QUOTE
-------
Logistic Model with 0.95 Confidence Level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Logistic -
BMDL
BMD
50
100 150 200 250 300 350 400 450
dose
13:0001/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/female/log_papmineralFrev_logistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/pappmineral/female/log_papmineralFrev_logistic.plt
Fri Jan 14 13:00:44 2011
The form of the probability function is: P[response] = I/[1+EXP(-intercept-slope*dose)]
Dependent variable = incidence
Independent variable = dose
Slope parameter is not restricted
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
background = 0 Specified
intercept = -2.67819
slope = 0.00343504
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
intercept slope
intercept 1 -0.78
slope -0.78 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
-2.72974 0.364791 -3.44472 -2.01477
0.00353956 0.00119641 0.00119464 0.00588449
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -65.6458 4
Fitted model -67.0198 2 2.74796 2 0.2531
Reduced model -71.3686 1 11.4455 3 0.009545
AIC:
B-20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Dose
Est. Prob.
Residual
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 292.331
BMDL = 219.166
Table B-13. Summary of BMD modeling results for incidence of combined
transitional cell hyperplasia in the bladder of male F344 rats exposed to
biphenyl in the diet for 2 years
Model
Gammab'c
Logistic
Log-Logistic13
Log-Probitb
Multistage (3-degree)d
Probit
Weibullb
Goodness of fit
%2/7-valuea
1.00
1.00
1.00
1.00
0.39
1.00
1.00
Largest
residual
-0.12
0.00
0.00
0.00
-1.63
0.00
0.00
AIC
34.54
36.51
36.51
36.51
40.12
36.51
36.51
Benchmark result (mg/kg-d)
BMDS
186.38
314.74
283.35
227.03
109.67
266.72
300.36
BMDLS
125.23
151.02
126.46
122.78
93.51
137.23
131.93
BMD10
205.40
323.93
295.47
241.87
139.41
280.54
313.72
BMDL10
146.73
182.76
147.96
140.96
123.14
166.54
160.88
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2002).
17
B-21
DRAFT - DO NOT CITE OR QUOTE
-------
Gamma Multi-Hit Model with 0.95 Confidence Level
I
<
0.8
0.6
0.4
0.2
Gamma Multi-Hit
BMDL
1
14:1501/142011
0 50 100 150 200 250 300 350
dose
2 BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
3
4 ====================================================================
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/bladdercombinedhyper/male/gam_bladcomhypMrev_gamma.pl
t
Fri Jan 14 14:15:19 2011
The form of the probability function is: P[response]= background+(1-
background)^CumGamma[slope^dose,power], where CumGamma(.) is the cummulative Gamma distribution
function
Dependent variable = incidence
Independent variable = dose
Power parameter is restricted as power >=1
Total number of observations = 4Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial (and Specified) Parameter Values
Background = 0.0192308
Slope = 0.0320399
Power = 8.56462
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Background -Power have been estimated at a boundary point, or
have been specified by the user, and do not appear in the correlation matrix )
Slope
Slope 1
Variable
Background
Slope 0.0624215 0.00323795
Power 18 NA
NA - Indicates that this parameter has hit a bound implied by some ineguality constraint and thus
has no standard error.
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
Dose
Est. Prob.
Scaled
Residual
50
50
50
50
Chi'
d. f.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 205.404
BMDL = 146.733
Table B-14. Summary of BMD modeling results for incidence of
mineralization in the kidney (inner stripe outer medulla) of male BDFi mice
exposed to biphenyl in the diet for 2 years
Model
Gammab, Weibullb,
Multistage (1 -degree)0
Logistic
Log-LogisticM
Log-Probitb
Probit
Goodness of fit
%2/7-valuea
0.46
0.43
0.48
0.33
0.44
Largest
residual
1.03
1.07
1.01
1.24
1.07
AIC
214.84
214.97
214.79
215.51
214.95
Benchmark result (mg/kg-d)
BMD5
369.24
454.16
341.66
710.74
442.78
BMDL5
155.65
238.75
130.84
377.36
227.50
BMD10
758.45
856.07
721.28
1,022.10
844.26
BMDL10
319.71
446.12
276.22
542.66
430.21
aValues O.10 fail to meet conventional goodness-of-fit criteria.
'Tower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
BMD = maximum likelihood estimate of the dose associated with the selected
benchmark response; BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark
response: i.e., 10 = dose associated with 10% extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2005).
24
B-23
DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
Log-Logistic
BMDL
BMD
200
400 600
dose
800
1000
12:5701/172011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/male/lnl_minmedullM_loglogistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/male/lnl_minmedullM_loglogistic.plt
Mon Jan 17 12:57:13 2011
The form of the probability function is: P[response] = background+(1-background)/[1+EXP(-
intercept-slope*Log(dose))]
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.18
intercept = -8.98323
slope = 1.06986
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
background intercept
background 1 -0.64
intercept -0.64 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.185925 * * *
-8.77824 * * *
Variable
background
intercept
slope 1 *
- Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -104.672 4
Fitted model -105.397 2 1.44976 2
Reduced model -106.377 1 3.40987 3
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
Goodness of Fit
Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 721.275
BMDL = 276.216
Table B-15. Summary of BMD modeling results for incidence of
mineralization in the kidney (inner stripe outer medulla) of female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Gammab
Logistic
Log-Logistic1"'0
Log-Probitb
Multistage (l-degree)d
Probit
Weibullb
Goodness of fit
%2/7-valuea
0.70
0.31
0.80
0.53
0.92
0.38
0.69
Largest
residual
-0.27
1.22
-0.18
0.80
-0.34
1.14
-0.28
AIC
184.21
184.34
184.12
183.33
182.23
183.96
184.22
Benchmark result (mg/kg-d)
BMD5
116.20
257.38
127.12
253.31
104.00
234.00
113.82
BMDL5
76.96
205.80
57.98
189.78
76.86
188.80
76.94
BMD10
229.86
451.19
233.39
364.28
213.63
417.63
227.40
BMDL10
158.09
369.40
122.40
272.92
157.88
343.46
158.04
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest BMDL10 was selected because BMDL values for models providing
adequate fit differed by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Umeda et al. (2005).
21
B-25
DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
0.7
0.6
0.5
0.4
0.3
0.2
0.1
200 400 600 800 1000 1200 1400
1
2
3
4.
*T
5
6
7
8
9
10
12
14
15
16
17
18
19
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
13:2701/172011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day .
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/20
Gnuplot Plotting File:
11/BMD/mice/minmedulla/ female/ Inl minmedullF loglogistic. (d)
C: /Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/female/lnl minmedullF loglogistic.plt
BMDS Model Run
Mon Jan 17 13:27:41 2011
The form of the probability function is: P[response] = background+ ( 1-background )/[ 1+EXP (-
intercept-slope*Log (dose ) ) ]
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as
Total number of observations = 4
Total number of records with miss
Maximum number of iterations = 25
Relative Function Convergence has
Parameter Convergence has been se
User has chosen the log transform
Default Initial
background =
intercept =
slope =
Asymptotic Correlation
background interc
background 1 -0
intercept -0.48
slope 0.44 -0
Variable Estimate
background 0.05773
intercept -8.90345
slope 1.22989
* - Indicates that this value is
slope >= 1
ing values = 0
0
been set to: le-008
t to: le-008
ed model
Parameter Values
0.06
-9.5037
1.31777
Matrix of Parameter Estimates
ept slope
.48 0.44
1 -0.99
. 99 1
Parameter Estimates
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
* * *
* * *
* * *
not calculated.
Analysis of Deviance Table
Model Log(likelihood)
Full model -89.0288
Fitted model -89.0609
Reduced model -107.593
# Param's Deviance Test d.f. P-value
4
3 0.0641982 1 0.8
1 37 1286 3 < 0001
B-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
Prob.
d.f. =1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 233.39
BMDL = 122.401
Table B-16. BMD model results for serum LDH activity in female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
All doses
Constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Polynomial (3 -degree)0
Powerd
0.0001
O.OOOl
0.0001
0.0001
O.OOOl
NA
0.38
0.30
0.93
0.93
0.00
0.34
0.34
0.31
0.31
1,687.59
1,685.52
1,686.01
1,683.73
1,683.73
CF
2,914.91
2,882.07
3,194.19
3,193.16
CF
1,491.53
1,450.54
1,595.47
1,449.38
182.66
465.81
465.80
465.86
465.81
0.0000
0.0026
0.0011
1.1 x 1Q-8
0.0036
Non constant variance
Hill
Linearb
Polynomial (2-degree)b
Polynomial (3-degree)b
Powerd
0.91
0.91
0.91
0.91
0.91
NA
O.OOOl
0.0001
0.0001
O.OOOl
-0.22
5.08
1.86
5.08
1.33
1,461.52
1,544.20
1,537.72
1,544.20
1,486.07
72.34
-9,999.00
554.86
-9,999.00
60.83
CF
720.55
25.81
1,947.93
41.31
161.83
53.40
42.35
53.40
107.91
107.12
19.49
6.96
0.88
81.24
21
22
23
aValues O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable (degrees of freedom for the test of mean
fit are <0, the %2 test for fit is not valid)
Source: Umeda et al. (2005).
None of the models provided an adequate fit to both the variance model and the means model.
B-27
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-17. BMD modeling results for serum AST activity in female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
All doses
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)', Powerd
O.OOOl
0.0001
NA
0.72
-5.69 x 107
0.68
1,264.30
1,260.96
6,722.40
1,826.88
566.24
1,205.47
213.62
595.87
0.00
135.74
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd
0.52
0.52
0.52
0.52
NA
0.0001
0.0001
O.OOOl
0.82
5.04
-2.55 x 109
-2.13
1,121.84
1,219.20
8.00
1,164.51
83.86
CF
0.00
106.70
CF
90.71
CF
69.43
154.69
21.60
185.08
150.64
114.05
2.76
CF
110.24
Highest dose dropped
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)0, Power
Not modeled; number of dose groups less than number of model parameters
O.OOOl
0.99
0.01
826.48
648.56
372.37
229.54
33.18
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd'e
Not modeled; number of dose groups less than number of model parameters
0.78
0.78
0.78
O.OOOl
0.0001
0.28
3.24 x 108
-2.20 x 109
-0.29
6
8
709.33
0
0
72.36
CF
CF
44.29
228.57
219.67
190.33
CF
CF
121.53
aValues O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
Selected model; only model providing adequate fit to modeled variance and means.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable (degrees of freedom for the test of mean
fit are <0, the %2 test for fit is not valid)
Source: Umeda et al. (2005).
B-28
DRAFT - DO NOT CITE OR QUOTE
-------
Power Model with 0.95 Confidence Level
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
400
350
300
250
200
150
100
50
Power
BMDL
BMD
0 50 100 150 200 250 300 350 400
dose
10:4701/182011
BMD and BMDL indicated are associated with a twofold increase from control, and are in units of mg/kg-day.
Power Model. (Version: 2.16; Date: 10/28/2009)
Input Data File: C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.pit
Tue Jan 18 10:47:11 2011
BMDS Model Run
The form of the response function is: Y[dose] = control + slope * dose^power
Dependent variable = mean
Independent variable = dose
The power is restricted to be greater than or egual to 1
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
lalpha = 10.765
rho = 0
control = 75
slope = 0.369536
power = 0.980467
lalpha
rho
control
slope
lalpha
1
-1
-0. 43
Variable
lalpha
rho
control
slope
power
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
4.06805 -20.8791 -4.93268
0.905641 2.7739 6.32395
5.21212 63.8097 84.2409
0.113823 0.165841 0.61202
NA
-12.9059
4.54893
74.0253
0.38893
1
NA - Indicates that this parameter has hit a bound implied by some ineguality constraint and thus
has no standard error.
B-29
DRAFT - DO NOT CITE OR QUOTE
-------
1 Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
2
3
4
5
6
7 Model Descriptions for likelihoods calculated
8 Model Al: Yi j = Mu ( i ) + e ( i j ) Var { e ( i j ) } = Sigma ^2
9 Model A2: Yij = Mu(i) + e(ij) Varfe(ij)} = Sigma(i)~2
10 Model A3: Yij = Mu(i) + e(ij) Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)
11 Model A3 uses any fixed variance parameters that were specified by the user
12 Model R: Yi = Mu + e(i) Var{e(i)} = Sigma"2
13
14 Likelihoods of Interest
15 Model Log(likelihood) # Param's
16 Al
17 A2
18 A3
19 fitted
20 R
21
22 Explanation of Tests
23 Test 1: Do responses and/or variances differ among Dose levels? (A2 vs. R)
24 Test 2: Are Variances Homogeneous? (Al vs A2)
25 Test 3: Are variances adequately modeled? (A2 vs. A3)
26 Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
27 (Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
28
29 Tests of Interest
30 Test -2*log(Likelihood Ratio) Test df
31 Test 1 125.335 4
32 Test 2 120.413 2
33 Test 3 0.0775771 1
34 Test 4 1.18681 1
35
36
37
38
39
40
41
42 The p-value for Test 3 is greater than .1. The modeled variance appears to be appropriate here
43
44
45
46
47 Benchmark Dose Computation
48 Specified effect = 1
49 Risk Type = Relative risk
50 Confidence level = 0.95
51 BMD = 190.33
52 BMDL = 121.534
B-30 DRAFT - DO NOT CITE OR QUOTE
-------
Table B-18. BMD modeling results for serum ALT activity in female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
All doses
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)', Powerd
O.0001
O.0001
NA
0.55
9.61 x 10'7
0.94
1,167.39
1,164.57
3,911.09
1,613.62
436.97
1,106.30
160.82
412.90
0.00
38.31
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd
0.78
0.78
0.78
0.78
NA
O.0001
O.0001
O.0001
-0.49
1.69 x 1010
-1.39 x 1011
-1.88
1,013.25
6
8
1,047.49
116.28
0
0
90.73
CF
CF
CF
62.72
148.75
419.08
87.64
108.55
121.30
CF
CF
77.76
Highest dose dropped
Constant variance
Hillb
Linear0,
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
0.0001
0.0001
0.0001
0.79
NA
NA
-0.22
4.25 x 1Q-7
-3.00 x 1Q-9
756.72
758.65
758.65
518.80
488.92
497.95
324.41
325.96
325.96
116.10
170.36
167.69
0.00
0.00
0.00
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
0.89
0.89
0.89
0.0001
0.0001
NA
-2.59 x 109
-5.85 x 107
0.10
6
8
631.43
0
0
110.52
CF
CF
67.61
111.13
169.57
172.25
CF
CF
117.98
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
None of the models provided an adequate fit to both the variance model and the means model.
B-31
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-19. BMD modeling results for serum AP activity in female BDFi
mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
All doses
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)', Powerd
O.0001
0.0001
NA
0.31
.4.74 x lO'8
1.32
1,240.81
1,239.14
642.90
1,253.51
320.63
919.17
540.57
1,208.38
180.68
720.75
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Polynomial (3 -degree)0
Powerd
0.006
0.006
0.006
0.006
0.006
NA
0.0001
0.0001
O.0001
0.0001
-0.93
5.04
-2.57 x 1011
1.89
1.41
1,180.07
1,334.76
8
1,242.58
1,236.21
147.47
-9,999.00
0
1,495.81
665.13
CF
244.46
CF
213.20
345.69
177.26
28.02
390.64
1,506.34
815.01
CF
0.05
CF
333.91
482.17
Highest dose dropped
Constant variance
Hillb
Linear0,
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
O.0001
O.0001
0.0001
0.55
0.95
NA
-0.51
-0.05
1.09E-8
868.21
867.85
869.84
617.91
510.80
499.45
361.78
393.46
372.60
487.67
467.69
464.35
201.11
315.45
213.97
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
0.77
0.77
0.77
0.0001
NA
NA
4.52 x 109
0.13
-0.21
6
794.19
794.19
0
287.55
285.46
CF
183.20
179.35
465.02
480.63
482.75
CF
334.12
333.04
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
None of the models provided an adequate fit to both the variance model and the means model.
B-32
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-20. BMD modeling results for changes in BUN levels (mg/dL) in
male BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
Males
All doses
Constant variance
Hillb
Linear°'d, Polynomial
(2 -degree)0, Power
0.03
0.03
NA
0.01
0.25
-2.00
540.50
545.04
CF
2,254.69
CF
1,288.77
CF
12,777.10
CF
7,154.72
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Polynomial (3 -degree)0
Powerd
0.01
0.01
0.01
0.01
0.01
NA
0.28
0.13
0.13
0.13
0.25
-1.95
-2.23
-2.25
-2.32
542.49
540.78
542.57
542.52
542.51
CF
3,134.77
2,029.81
1,688.06
1,170.31
CF
1,690.32
1,459.55
1,324.21
1,092.10
CF
15,745.20
4,649.85
2,974.25
1,334.64
CF
8,512.03
3,312.21
2,291.81
1,196.80
Highest dose dropped
Constant variance
Hillb
Linear0, Polynomial
(2-degree)c, Powerd
Not modeled; number of dose groups less than number of model parameters
0.49
0.32
0.77
420.23
414.78
266.77
2,140.93
1,335.54
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
B-33
DRAFT - DO NOT CITE OR QUOTE
-------
Linear Model with 0.95 Confidence Level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
25
24
23
22
20
19
Linear
BMDL
BMi)
50
100
150
11:0301/192011
200 250
dose
300
350
400
BMD and BMDL indicated are associated with a 1SD change from control, and are in units of mg/kg-day.
Polynomial Model. (Version: 2.16; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/BUN/male/lin_BUNMHDD_linear.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/BUN/male/lin_BUNMHDD_linear.plt
Wed Jan 19 11:03:37 2011
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
The polynomial coefficients are restricted to be positive
A constant variance model is fit
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 16.1929
rho = 0 Specified
beta_0 = 20.5429
beta_l = 0.00972018
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
alpha beta_0 beta_l
alpha 1 -3.8e-008 3.2e-008
beta_0 -3.8e-008 1 -0.74
beta 1 3.2e-008 -0.74 1
Variable
alpha
beta_0
beta 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
15.8907 2.14271 11.6911 20.0904
19.4657 21.6863
0.00338636 0.0158352
34
B-34
DRAFT - DO NOT CITE OR QUOTE
-------
1 97 39 22 21.5 4
2 291 37 23.2 23.4 4.4
3
4 Model Descriptions for likelihoods calculated
5 Model Al: Yij = Mu(i) + e(ij)
6 Var{e(ij)} = Sigma~2
7 Model A2: Yij = Mu(i) + e(ij)
8 Var{e(ij)} = Sigma(1)^2
9 Model A3: Yij = Mu(i) + e(ij)
10 Var{e(ij)} = Sigma^2
11 Model A3 uses any fixed variance parameters that were specified by the user
12 Model R: Yi = Mu + e(i)
13 Var{e(i)} = Sigma "'2
14
15 Likelihoods of Interest
16 Model Log(likelihood) # Param's
17 Al -206.630664 4
18 A2 -205.915695 6
19 A3 -206.630664 4
20 fitted -207.115525 3
21 R -211.514015 2
22
23 Explanation of Tests
24
25 Test 1: Do responses and/or variances differ among Dose levels? (A2 vs. R)
26 Test 2: Are Variances Homogeneous? (Al vs A2)
27 Test 3: Are variances adequately modeled? (A2 vs. A3)
28 Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
29 (Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
30
31 Tests of Interest
32 Test -2*log(Likelihood Ratio) Test df
33 Test 1 11.1966 4
34 Test 2 1.42994 2
35 Test 3 1.42994 2
36 Test 4 0.969721 1
37
38
39
40
41 The p-value for Test 2 is greater than .1. A homogeneous variance model appears to be
42 appropriate here
43
44
45
46
47
48
49 Benchmark Dose Computation
50 Specified effect = 1
51 Risk Type = Estimated standard deviations from the control mean
52 Confidence level = 0.95
53 BMD = 414.775
54 BMDL = 266.77
55
B-3 5 DRAFT - DO NOT CITE OR QUOTE
-------
Table B-21. BMD modeling results for changes in BUN levels (mg/dL) in
female BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD1RD
BMDL1RD
All doses
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)', Powerd
O.0001
O.0001
NA
0.38
-3.45 x 10'8
1.18
603.61
601.53
CF
1,869.01
CF
1,224.15
CF
2,507.85
CF
1,434.76
Non constant variance
Hillb
Linear0, Polynomial
(2 -degree)0, Powerd
0.08
0.08
NA
O.0001
-1.21
-1.63
493.48
590.70
141.72
519.60
CF
216.41
CF
1,191.69
CF
683.73
Highest dose dropped
Constant variance
Hillb
Linear0,
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
O.0001
O.0001
0.0001
0.50
0.82
NA
-0.57
-0.18
-2.11 x 1Q-10
417.59
417.19
419.13
744.99
555.48
430.03
403.07
413.38
414.77
921.79
627.58
436.97
410.67
432.73
417.75
Non constant variance
Hillb
Linear0
Polynomial (2 -degree)0
Powerd
Not modeled; number of dose groups less than number of model parameters
0.23
0.23
0.23
0.07
NA
0.0001
-1.38
-0.93
-0.93
300.36
299.05
297.05
180.70
263.22
256.90
114.17
152.60
151.17
1,416.07
842.06
925.84
916.09
495.16
490.39
"Values O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; IRD = dose associated with a 100% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
None of the models provided an adequate fit to both the variance model and the means model.
B-36
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-22. BMD modeling results for changes in mean terminal body
weight in male BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD01RD
BMDL01RD
All doses
Constant variance
Hillb
Linear0, Powerd
Polynomial (3 -degree)0
O.OOOl
0.0001
O.OOOl
0.03
0.10
0.03
-1.68
-1.68
-1.66
716.95
714.95
716.89
459.61
460.46
498.04
390.85
391.75
392.48
358.30
359.04
390.52
316.09
316.87
317.33
Non constant variance
Hillb
Linear0,
Polynomial (3 -degree)0
Powerd
0.002
0.002
0.002
0.002
NA
0.59
0.44
0.38
-1.52
-1.52
-1.42
-1.51
704.84
701.13
702.64
702.84
600.48
541.68
643.20
600.89
CF
460.24
467.09
464.26
421.46
357.54
450.96
421.53
325.00
326.02
328.74
327.62
Highest dose dropped
Constant variance
Hillb
Linear0, Polynomial
(2 -degree)0, Powerd
Not modeled; number of dose groups less than number of model parameters
0.01
0.05
-1.49
560.11
566.99
328.79
400.33
238.24
Non constant variance
Hillb
Linear0, Polynomial
(2 -degree)0, Powerd
Not modeled; number of dose groups less than number of model parameters
0.18
0.001
-1.5
562.10
561.56
308.43
398.66
235.32
"Values O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be negative.
dRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; 0 IRD = dose associated with a 10% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
None of the models provided an adequate fit to both the variance model and the means model.
B-37
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-23. BMD modeling results for changes in mean terminal body
weight in female BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
Variance
model
/7-valuea
Means
model
/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMDoiRD
BMDL01RD
All doses
Constant variance
Hillb
Linearc'd, Polynomial
(2-degree)c, Power"
0.36
0.36
0.80
0.42
-0.21
-0.93
382.59
382.26
387.90
584.12
230.17
489.94
397.06
583.33
243.57
510.85
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
Coefficients restricted to be negative.
dSelected model; the model with the lowest AIC was selected because BMDL values for models providing adequate
fit did not differ by more than threefold.
eRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote benchmark response: i.e., iSD = dose
associated with 1 standard deviation from control mean value; 0 IRD = dose associated with a 10% relative deviation
from control mean value); CF = computation failed; NA = not applicable
Source: Umeda et al. (2005).
Linear Model with 0.95 Confidence Leve
2
3
4
5
6
7
8
9
10
11
12
13
14
15
36
34
32
30
28
26
24
0
Linear
BMDL 3MD
200 400 600 800 1000 1200 1400
dose
09:2001/202011
BMD and BMDL indicated are associated with a 10% decrease from control, and are in units of mg/kg-day.
Polynomial Model. (Version: 2.16; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/termbdwt/female/lin_termbdwtF_linear.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/termbdwt/female/lin_termbdwtF_linear.pit
Thu Jan 20 09:20:01 2011
B-38
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
_> *J
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
72
73
74
75
Dependent variable = mean
Independent variable = dose
rho is set to 0
The polynomial coefficients are restricted to be negative
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 11.4937
rho = 0 Specified
beta 0 = 33.4391
beta 1 = -0.00571961
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho have been estimated at a boundary point, or
specified by the user, and do not appear in the correlation matrix )
alpha beta 0 beta 1
alpha 1 -9.6e-009 9.1e-009
beta 0 -9.6e-009 1 -0.67
beta 1 9.1e-009 -0.67 1
Parameter Estimates
95.0% Wald Confidence
Variable Estimate Std. Err. Lower Conf. Limit Upper
alpha 11.2518 1.5172 8.27818
beta 0 33.4983 0.432523 32.6505
beta 1 -0.00574262 0.000545303 -0.0068114
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
0 31 34 33.5 4 3.35 0.833
134 22 32.5 32.7 3.3 3.35 -0.32
414 25 30.5 31.1 3.1 3.35 -0.925
1420 32 25.5 25.3 3 3.35 0.264
Model Descriptions for likelihoods calculated
Model Al : Yij = Mu ( i ) + e(ij) Var{e(ij)} = Sigma^2
Model A2 : Yij = Mu ( i ) + e(ij) Varfe(ij)} = Sigma(i)^2
Model A3: Yij = Mu ( i ) + e(ij) Var{e(ij)} = Sigma"2
have been
Interval
Conf. Limit
14.2255
34 .346
0. 00467385
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
Al -187.261579 5 384.523158
A2 -185.643849 8 387.287698
A3 -187.261579 5 384.523158
fitted -188.129218 3 382.258435
R -226.477701 2 456.955401
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels? (A2 vs. R)
Test 2: Are Variances Homogeneous? (Al 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
Test 1 81.6677 6 <.0001
Test 2 3.23546 3 0.3567
Test 3 3.23546 3 0.3567
Test 4 1.73528 2 0.4199
The p-value for Test 1 is less than .05. There appears to be a difference betw
and/or variances among the dose levels. It seems appropriate to model the data
een response
The p-value for Test 2 is greater than .1. A homogeneous variance model appears to be
appropriate here
B-39
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Relative risk
Confidence level = 0.95
BMD = 583.327
BMDL = 510.848
Table B-24. Summary of BMD modeling results for incidence of litters with
fetal skeletal anomalies from Wistar rat dams administered biphenyl by
gavage on GDs 6-15
Model
Gammab, Weibullb,
Multistage (1 -degree)0
Logistic
Log-LogisticM
Log-Probitb
Probit
Goodness of fit
%2/7-valuea
0.31
0.28
0.41
0.23
0.28
Largest
residual
-1.25
1.17
-1.32
-1.59
1.20
AIC
106.11
106.42
105.33
106.55
106.50
Benchmark result (mg/kg-d)
BMDS
54.45
73.97
27.03
125.14
79.59
BMDLS
24.15
36.73
9.59
55.10
41.02
BMD10
111.84
149.18
57.06
179.97
160.27
BMDL10
49.61
73.79
20.24
79.23
82.37
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
°Betas restricted to >0.
dSelected model; the model with the lowest BMDL was selected because BMDL values for models providing
adequate fit differed by more than threefold; this model also had the lowest AIC.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., 10 = dose associated with 10%
extra risk; 5 = dose associated with 5% extra risk)
Source: Kheraetal. (1979).
B-40
DRAFT - DO NOT CITE OR QUOTE
-------
02 HMDL
16:0601/142011
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
400 600
dose
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
Logistic Model. (Version: 2.13; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/develop/anomlitt/lnl_anomlitt_loglogistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/develop/anomlitt/lnl_anomlitt_loglogistic.pit
Fri Jan 14 16:06:43 2011
BMDS Model Run
The form of the probability function is: P[response] = background+(1-background)/[1+EXP(-
intercept-slope^Log(dose))]
Dependent variable = incidence
Independent variable = dose
Slope parameter is restricted as slope >= 1
Total number of observations = 5
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.5
intercept = -6.54827
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
background intercept
background 1 -0.77
intercept -0.77 1
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.503241 * * *
Variable
background
intercept -6.24131 *
slope 1 *
- Indicates that this value is not calculated.
B-41
DRAFT - DO NOT CITE OR QUOTE
-------
1 Fitted model -50.6629 2
2 Reduced model -52.2232 1
3
4 AIC: 105.326
5
6 Goodness of Fit
7 Scaled
8 Dose Est. Prob. Expected Observed Size Residual
9 1
10
11
12
13
14
15
16 Chi~2 = 2.90 d.f. = 3 P-value = 0.406
17
18 Benchmark Dose Computation
19 Specified effect = 0.1
20 Risk Type = Extra risk
21 Confidence level = 0.95
22 BMD = 57.0591
23 BMDL = 20.2399
B-42 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX C. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR
3
4
5
6
7
The mouse liver tumor dataset from Umeda et al. (2005) for which dose-response
modeling was performed is shown in Table C-l.
Table C-l. Incidences of liver adenomas or carcinomas (combined) in
female BDFi mice fed diets containing biphenyl for 2 years
Biphenyl dietary concentration (ppm)
Reported dose (mg/kg-d)
RED (mg/kg-d)
Tumor incidence
Adenoma or carcinoma (combined)
0
0
0
3/48a
667
134
19
8/50
2,000
414
59
16/49^
6,000
1,420
195
14/48a'c
8
9
10
11
12
13
14
15
16
aTwo control, one mid-dose, and two high-dose female mice were excluded from denominators because they died
prior to week 52. It is assumed that they did not have tumors and were not exposed for a sufficient time to be at
risk for developing a tumor. Umeda et al. (2005) did not specify the time of appearance of the first tumor.
bSignificantly different from controls (p < 0.05) according to Fisher's exact test.
Significantly different from controls (p < 0.01) according to Fisher's exact test.
Source: Umeda et al. (2005).
Summaries of the BMDs, BMDLs, and the derived oral slope factors for the modeled
mouse data are presented in Table C-2, followed by the plot and model output file from the best-
fitting model. The animals in the highest dose group, while exhibiting a statistically
significantly increased incidence in liver tumors compared with controls, did not show a
monotonic increase in tumor response compared with the responses at the lower doses. To better
estimate responses in the low dose region, the high dose group was excluded as a means of
improving the fit of the model in the region of interest.
C-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-2. Model predictions for liver tumors (adenomas or carcinomas
combined) in female BDFi mice exposed to biphenyl in the diet for 2 years
Model
Goodness of fit
%2/7-valuea
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMDnEDio
BMDLnEDio
Cancer
slope
factor
(risk per
mg/kg-d)
All doses
Multistage (1-, 2-, 3-degree)b,
Gammac, Weibulf
Logistic
Log-Logistic0
Log-Probif
Probit
0.03
0.01
0.04
0.005
0.01
2.14
2.31
1.97
2.58
2.30
197.37
198.96
196.62
201.06
198.80
64.76
104.91
50.68
128.52
100.16
37.29
71.27
26.80
74.43
67.23
0.003
0.001
0.004
0.001
0.001
Highest dose dropped
Multistage (l-degree)M
Multistage (2-degree)b
0.96
0.96
0.04
0.04
132.32
132.32
18.72
18.72
12.15
12.15
0.008
0.008
aValues <0.05 fail to meet conventional goodness-of-fit criteria.
bBetas restricted to >0.
°Power restricted to >1.
dSelected model.
BMD = maximum likelihood estimate of the dose associated with the selected benchmark response; BMDL = 95%
lower confidence limit on the BMD (subscripts denote benchmark response: i.e., HEDIO = human equivalent dose
associated with 10% extra risk
Source: Umeda et al. (2005).
Multistage Cancer Model with 0.95 Confidence Level
2
3
4
0.5
0.4
0.3
0.2
0.1
0
Multistage Cancer
Linear extrapolation
BMDL BMD
0 10 20 30 40 50 60
dose
09:33 02/03 2011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
C-2
DRAFT - DO NOT CITE OR QUOTE
-------
1 ====================================================================
2 Multistage Cancer Model. (Version: 1.9; Date: 05/26/2010)
3 Input Data File:
4 C: /Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l .
5 (d)
6 Gnuplot Plotting File:
7 C: /Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l .
8 pit
9 Thu Feb 03 09:33:34 2011
10 ====================================================================
11
13 The form of the probability function is: P[response] = background + ( 1-background ) * [ 1-EXP ( -
14 betal*dose^l) ]
15 The parameter betas are restricted to be positive
16 Dependent variable = incidence
17 Independent variable = dose
18 Total number of observations = 3
19 Total number of records with missing values = 0
20 Total number of parameters in model = 2
21 Total number of specified parameters = 0
22 Degree of polynomial = 1
23 Maximum number of iterations = 250
24 Relative Function Convergence has been set to: 2.22045e-016
25 Parameter Convergence has been set to: 1.49012e-008
26 ±±±± we are sorry but Relative Function and Parameter Convergence are currently unavailable in
27 this model. Please keep checking the web sight for model updates which will eventually
28 incorporate these convergence criterion. Default values used. ****
29
30
31
32
33
34 Asymptotic Correlation Matrix of Parameter Estimates
35 Background Beta ( 1 )
36
37
38
39 Parameter Estimates
40 95.0% Wald Confidence Interval
41 Variable Estimate Std. Err. Lower Conf . Limit Upper Conf . Limit
42 Background
43 Beta(l)
44
45
46
47
48 Model Log (likelihood) # Param's Deviance Test d.f. P-value
49 Full model -64.1585 3
50 Fitted model -64.1595 2
51 Reduced model -70.107 1
52
53 AIC: 132.319
54
55 Goodness of Fit
56 Scaled
57 Dose Est._Prob. Expected Observed Size Residual
58 ---------------- - ------------------------------------------------------
59
60
61
62
63 Chi^2 = 0.00 d.f. = 1
64
65 Benchmark Dose Computation
66 Specified effect = 0.1
67 Risk Type = Extra risk
68 Confidence level = 0.95
69 BMD = 18.7158
70 BMDL = 12.1518
71 BMDU = 36.3895
72 Taken together, (12.1518, 36.3895) is a 90% two-sided confidence interval for the BMD
73 Multistage Cancer Slope Factor = 0.00822924
74
C-3 DRAFT - DO NOT CITE OR QUOTE
------- |