DRAFT - DO NOT CITE OR QUOTE EPA/635/R-11/005A
www.epa.gov/iris
f/EPA
TOXICOLOGICAL REVIEW
OF
BIPHENYL
(CAS No. 92-52-4)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
September 2011
NOTICE
This document is an External Review 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DRAFT - DO NOT CITE OR QUOTE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 FIGURES ix
LIST OF ABBREVIATIONS AND ACRONYMS x
FOREWORD xii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiii
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION 3
3. TOXICOKINETICS 5
3.1. ABSORPTION 5
3.2. DISTRIBUTION 6
3.3. METABOLISM 7
3.3.1. Identification of Metabolites 7
3.3.1.1. Results from In Vivo Animal Studies 7
3.3.1.2. Results from In Vitro Studies with Animal and Human Cells or Tissues 9
3.3.2. Metabolic Pathways 10
3.3.2.1. Description of Metabolic Scheme and Enzymes Involved 10
3.3.3. Regulation of Metabolism, Sites of Metabolism, and Relationships to Toxic
Effects 13
3.3.3.1. Evidence for Induction of Phase I and II Enzymes 13
3.3.3.2. Demonstrated Tissue Sites of Metabolism 15
3.4. ELIMINATION 15
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS 16
4. HAZARD IDENTIFICATION 17
4.1. STUDIES IN HUMANS 17
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION 21
4.2.1. Oral Exposure 22
4.2.1.1. Subchronic Toxicity 22
4.2.1.2. Chronic Toxicity and Carcinogenicity 23
4.2.2. Inhalation Studies 39
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION... 40
4.3.1. Oral Exposure 40
4.3.2. Inhalation Exposure 44
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES 44
4.4.1. Acute and Short-term Toxicity Data 44
4.4.2. Kidney/Urinary Tract Endpoint Studies 45
4.4.3. Biphenyl as a Tumor Promoter 49
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
ACTION 51
4.5.1. Effects on the Urinary Tract of Rats 51
4.5.2. Genotoxicity 51
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS 52
4.6.1. Oral 58
iii DRAFT - DO NOT CITE OR QUOTE
-------
4.6.2. Inhalation 58
4.6.3. Mode-of-Action Information 59
4.7. EVALUATION OF CARCINOGENICITY 60
4.7.1. Summary of Overall Weight of Evidence 60
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence 62
4.7.3. Mode-of-Action Information 64
4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats 64
4.7.3.2. Mode-of-Action Information for Liver Tumors in Female Mice 70
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 73
4.8.1. Possible Childhood Susceptibility 73
4.8.2. Possible Gender Differences 74
4.8.3. Other 74
5. DOSE-RESPONSE ASSESSMENTS 75
5.1. ORAL REFERENCE DOSE (RfD) 75
5.1.1. Choice of Candidate Principal Studies and Candidate Critical Effects—with
Rationale and Justification 75
5.1.2. Methods of Analysis—Including Models (e.g., PBPK, BMD) 78
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs) 86
5.1.4. Previous RfD Assessment 87
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 87
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification 87
5.2.2. Previous RfC Assessment 89
5.3. UNCERTAINTIES IN THE RfD and RfC 89
5.4. CANCER ASSESSMENT 90
5.4.1. Choice of Study/Data—with Rationale and Justification 90
5.4.2. Dose-Response Data 90
5.4.3. Dose Adjustments and Extrapolation Method(s) 92
5.4.3.1. Bladder Tumors in Male Rats 92
5.4.3.2. Liver Tumors in Female Mice 92
5.4.4. Oral Slope Factor and Inhalation Unit Risk 94
5.4.5. Uncertainties in Cancer Risk Values 95
5.4.5.1. Oral Slope Factor 95
5.4.5.2. Inhalation Unit Risk 97
5.4.6. Previous Cancer Assessment 97
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE 98
6.1. HUMAN HAZARD POTENTIAL 98
6.1.1. Noncancer 98
6.1.2. Cancer 99
6.2. DOSE RESPONSE 100
6.2.1. Noncancer/Oral 100
6.2.2. Noncancer/Inhalation 100
6.2.3. Cancer/Oral 100
6.2.4. Cancer/Inhalation 101
7. REFERENCES 102
iv DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS
AND DISPOSITION A-l
APPENDIX B. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
MODE OF ACTION B-l
B.I. EFFECTS ON THE URINARY TRACT OF RATS B-l
B.2. EFFECTS ON THE LIVER OF MICE B-2
B.3. ESTROGENIC EFFECTS B-3
B.4. EFFECTS ON APOPTOSIS B-3
B.5. MITOCHONDRIAL EFFECTS B-4
B.6. GENOTOXICITY B-5
APPENDIX C. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE DOSE..C-1
APPENDIX D. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR D-1
v DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF TABLES
Table 2-1. Physicochemical properties of biphenyl 4
Table 3-1. Metabolites of biphenyl identified in urine, feces, and bile of male albino rats 8
Table 4-1. Biphenyl concentrations in the air of a Finnish paper mill producing biphenyl-
impregnated fruit wrapping paper 18
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl: comparison with 60
unexposed males 19
Table 4-3. Exposure data and clinical features for five PD patients with occupational exposure to
biphenyl 21
Table 4-4. Incidences of urinary bladder lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 25
Table 4-5. Incidences of ureter and kidney lesions in male and female F344 rats exposed to
biphenyl in the diet for 2 years 27
Table 4-6. Body and organ weight data for male and female rats administered biphenyl in the
diet for 2 years 31
Table 4-7. Dose-related changes in selected clinical chemistry values from male and female
BDFi mice exposed to biphenyl via the diet for 2 years 34
Table 4-8. Incidences of gross and histopathological findings in male and female BDFi mice fed
diets containing biphenyl for 2 years 35
Table 4-9. Incidences of selected tumor types among controls and mice administered biphenyl
orally for 18 months 38
Table 4-10. Incidences of selected histopathologic lesions in tissues of CD-I mice exposed to
biphenyl vapors 7 hours/day, 5 days/week for 13 weeks 40
Table 4-11. Prenatal effects following oral administration of biphenyl to pregnant Wistar rats on
GDs6-15 42
Table 4-12. Summary of reproductive data in albino rats exposed to dietary biphenyl 44
Table 4-13. Number of Wistar rats exposed to biphenyl and the degree of change in kidney
weight and cellular architecture 48
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in
rats and mice 53
Table 4-15. Summary of major studies evaluating effects of biphenyl after inhalation exposure
in rats, mice and rabbits 57
Table 5-1. Datasets employed in the BMD modeling of nonneoplastic effects in the urinary tract
of male and female F344 rats exposed to biphenyl in the diet for 2 years 78
Table 5-2. Datasets employed in the BMD modeling of body weight, selected clinical chemistry
results, and histopathological kidney effects in male and female BDFi mice exposed to
biphenyl in the diet for 2 years 79
vi 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 GDs 6-15 80
Table 5-4. Summary of BMDs/BMDLs for selected nonneoplastic effects following oral
exposure of rats and mice to biphenyl 83
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 91
Table 5-6. Incidence data for liver tumors in male and female BDFi mice fed diets containing
biphenyl for 2 years 91
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) 93
Table 5-8. Incidence of liver adenomas or carcinomas (combined) in female BDFi mice fed
diets containing biphenyl for 2 years 93
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 95
Table 5-10. Summary of uncertainties in the biphenyl cancer slope factor 96
Table B-l. 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) B-2
Table B-2. Genotoxicity test results for biphenyl B-6
Table B-3. Genotoxicity test results for biphenyl metabolites B-10
Table C-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 C-l
Table C-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 C-2
Table C-3. BMD modeling dataset for incidence of litters with fetal skeletal anomalies from
Wistar rat dams administered biphenyl by gavage on GDs 6-15 C-3
Table C-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 C-3
Table C-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 C-5
Table C-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 C-7
Table C-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 C-9
Table C-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 C-l 1
Table C-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 C-13
vii DRAFT - DO NOT CITE OR QUOTE
-------
Table C-10. Summary of BMD modeling results for incidence of hemosiderin deposits in the
kidney of female F344 rats exposed tobiphenyl in the diet for 2 years C-15
Table C-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 C-17
Table C-12. Summary of BMD modeling results for incidence of papillary mineralization in the
kidney of female F344 rats exposed tobiphenyl in the diet for 2 years C-19
Table C-l 3. 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 C-21
Table C-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 C-23
Table C-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 C-25
Table C-16. BMD model results for serum LDH activity in female BDFi mice exposed to
biphenyl in the diet for 2 years C-27
Table C-17. BMD modeling results for serum AST activity in female BDFi mice exposed to
biphenyl in the diet for 2 years C-28
Table C-l8. BMD modeling results for serum ALT activity in female BDFi mice exposed to
biphenyl in the diet for 2 years C-31
Table C-19. BMD modeling results for serum AP activity in female BDFi mice exposed to
biphenyl in the diet for 2 years C-32
Table C-20. BMD modeling results for changes in BUN levels (mg/dL) in male BDFi mice
exposed to biphenyl in the diet for 2 years C-33
Table C-21. BMD modeling results for changes in BUN levels (mg/dL) in female BDFi mice
exposed to biphenyl in the diet for 2 years C-36
Table C-22. BMD modeling results for changes in mean terminal body weight in male BDFi
mice exposed to biphenyl in the diet for 2 years C-37
Table C-23. BMD modeling results for changes in mean terminal body weight in female BDFi
mice exposed tobiphenyl in the diet for 2 years C-38
Table C-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 C-40
Table D-l. Incidences of liver adenomas or carcinomas (combined) in female BDFi mice fed
diets containing biphenyl for 2 years D-l
Table D-2. Model predictions for liver tumors (adenomas or carcinomas combined) in female
BDFi mice exposed to biphenyl in the diet for 2 years D-2
viii DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
Figure 3-1. Schematic presentation of the metabolic pathways ofbiphenyl 12
Figure 5-1. NOAELs and LOAELs for noncancer effects in rats and mice from repeated oral
exposure tobiphenyl 76
Figure 5-2. BMDs and BMDLs for selected noncancer effects in rats and mice from repeated
oral exposure to biphenyl 84
ix DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS
ACGIH
AIC
ALT
AP
AST
BBN
BMD
BMDL
BMR
BMDS
BrdU
BUN
CA
CASRN
CHL
CHO
CVSF
CYP
DNA
EEC
EHEN
EMG
ENMG
GC
GD
GOT
GPT
HED
HGPRT
HPLC
IARC
i.p.
IRIS
Km
LD50
LDH
LOAEL
MCV
MS
NOAEL
NRC
PBPK
PCB
PD
POD
American Conference of Governmental Industrial Hygienists
Akaike's Information Criterion
alanine aminotransferase
alkaline phosphatase
aspartate aminotransferase
N-butyl-N-(4-hydroxybutyl)nitrosamine
benchmark dose
95% lower confidence limit on the BMD
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
conduction velocity of the slowest motor fibers
cytochrome P-450
deoxyribonucleic acid
el ectroencephal ography
N-ethyl-N-hydroxyethylnitrosamine
el ectromy ographi c
el ectroneuromy ography
gas chromatography
gestation day
glutamate oxaloacetate transaminase
glutamate pyruvate transaminase
human equivalent doses
hypoxanthine guanine phosphoribosyl transferase
high-performance liquid chromatography
International Agency for Research on Cancer
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
mass spectrometry
no-observed-adverse-effect level
National Research Council
physiologically based pharmacokinetic
polychlorinated biphenyl
Parkinson's disease
point of departure
x
DRAFT - DO NOT CITE OR QUOTE
-------
PPAR
RD
RfC
RfD
ROS
RR
SCE
SD
SULT
TLV
TMS
TWA
UDS
UF
UGT
U.S. EPA
peroxisome proliferator activated receptors
relative deviation
reference concentration
reference dose
reactive oxygen species
relative risk
sister chromatid exchange
standard deviation
sulphotransferase
threshold limit value
trimethylsilyl
time-weighted average
unscheduled DNA synthesis
uncertainty factors
uridine diphosphate glucuronosyl transferase
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 and interagency reviewers
from other federal agencies and White House offices.
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. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of biphenyl.
IRIS Summaries may include oral reference dose (RfD) and inhalation reference concentration
(RfC) values for chronic and other exposure durations, and a carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal-of-entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, a plausible inhalation unit risk is
an upper bound on the estimate of risk per ug/m3 air breathed.
Development of these hazard identification and dose-response assessments for biphenyl
has followed the general guidelines for risk assessment as set forth by the National Research
Council (NRC, 1983). EPA Guidelines and Risk Assessment Forum Technical Panel Reports
that may have been used in the development of this assessment include the following:
Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986b), Guidelines
for Mutagenicity Risk Assessment (U.S. EPA, 1986a), Recommendations for and Documentation
of Biological Values for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for
Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and
Limit Concentration Issues in Inhalation Toxicity Studies (U.S. EPA, 1994a), Methods for
1 DRAFT - DO NOT CITE OR QUOTE
-------
Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA, 1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA,
1995), Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk
Characterization (U.S. EPA, 2000b), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000a), Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006b), A
Framework for Assessing Health Risk of Environmental Exposures to Children (U.S. EPA,
2006a), and Recommended Use of Body Weight3'4 as the Default Method in Derivation of the
Oral Reference Dose (U.S. EPA, 2011).
The literature search strategy employed for biphenyl was based on the chemical name,
Chemical Abstracts Service Registry Number (CASRN), and multiple common synonyms. Any
pertinent scientific information submitted by the public to the IRIS Submission Desk was also
considered in the development of this document. Primary, peer-reviewed literature identified
through August 2011 was included where that literature was determined to be critical to the
assessment. The relevant literature included publications on biphenyl that were identified
through Toxicology Literature Online (TOXLINE), PubMed, the Toxic Substance Control Act
Test Submission Database (TSCATS), the Registry of Toxic Effects of Chemical Substances
(RTECS), the Chemical Carcinogenesis Research Information System (CCRIS), the
Developmental and Reproductive Toxicology/Environmental Teratology Information Center
(DART/ETIC), the Hazardous Substances Data Bank (HSDB), the Genetic Toxicology Data
Bank (GENE-TOX), Chemical abstracts, and Current Contents. Other peer-reviewed
information, including health assessments developed by other organizations, review articles, and
independent analyses of the health effects data were retrieved and may be included in the
assessment where appropriate.
DRAFT - DO NOT CITE OR QUOTE
-------
2. CHEMICAL AND PHYSICAL INFORMATION
Pure biphenyl is a white or colorless crystalline solid that usually forms leaflets or scales;
commercial preparations may be yellowish or slightly tan (HSDB, 2005). Biphenyl is said to
have a pleasant odor that is variably described as peculiar, butter-like, or resembling geraniums
(HSDB. 2005: Boehncke et al.. 1999). Biphenyl melts at 69°C and has a vapor pressure of
8.93 x 10"3 mm Hg at 25°C, making it likely to enter the environment in its vaporized form
(HSDB, 2005). If particle-bound biphenyl is precipitated to the ground, it is likely to be
reintroduced to the atmosphere by volatilization. The water solubility of biphenyl is 7.48 mg/L
at 25°C. The logarithm of the octanol/water partition coefficient (Kow) of biphenyl of 3.98
suggests a potential for bioaccumulation (HSDB, 2005). Because it is biodegraded with an
estimated half-life of 2 and 3 days in air and water, respectively (HSDB, 2005), and is
metabolized rapidly by humans and animals (see Section 3), bioaccumulation does not occur
(Boehncke et al., 1999). Biphenyl is ubiquitous in the environment, with reported indoor air
concentrations of 0.16-1 |ig/m3 and outdoor levels of approximately 0.03 |ig/m3 (Boehncke et
al., 1999). The physicochemical properties of biphenyl are summarized in Table 2-1.
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 Kow
Water solubility
Henry's law constant
Conversion factors
Diphenyl,
lemonene,
l,l'-biphenyl, l,l'-diphenyl, bibenzene, phenylbenzene,
Carolid AL, Phenador-X, Tetrosine LY
92-52-4
oo
C12H10
154.2
69°C
256°C
1.041g/cm3at20°C
8.93 x 10-3mmHgat25°C
4.01;4.11a
7.48 mg/L
3.08 x 10'4
;4.17or5.27-5.46b
at 25°C
atm-m3/mol at 25°C
1 ppm= 6.31 mg/m3; 1 mg/m3 = 0.159 ppm
aMonsanto (1946).
Estimated by different methods: Dow Chemical Co. (1983).
Source: HSDB (2005).
Biphenyl exists naturally as a component of crude oil or coal tar. It is primarily produced
by debromination/dimerization of bromobenzene, is isolated as a byproduct of the
hydrodealkylation of toluene (yield approximately 1%), or is synthesized by catalytic
dehydrocondensation of benzene. Biphenyl is currently not registered for use as a pesticide in
the United States, but is still used in other countries as a fungistat, most commonly to preserve
packaged citrus fruits or in plant disease control (HSDB, 2005). The current major uses of
biphenyl are as chemical synthesis intermediates (among them, the sodium salt of
2-hydroxy-biphenyl, a pesticide known as Dowicide 1), as dye carriers in polyester dyeing, and
as components in heat transfer fluids (in particular Dowtherm A or Therminol® VP-1, consisting
of 26.5% biphenyl and 73.5% diphenyl oxide). Historically, biphenyl was the primary byproduct
in the manufacture of polychlorinated biphenyls (PCBs) until PCBs were banned in the 1970s
(U.S. EPA, 1978). The purity of technical biphenyl ranges from 93 to 99.9%. The prevalent
impurities in technical preparations are terphenyls, a side product from the dehydrocondensation
of benzene. Biphenyl is rated as a high-volume production chemical. Annual U.S. production in
1990 was approximately 1.6 x 104 metric tons (HSDB. 2005).
DRAFT - DO NOT CITE OR QUOTE
-------
3. TOXICOKINETICS
3.1. ABSORPTION
No quantitative studies on the absorption of biphenyl have been conducted in humans.
Animal studies in rats, rabbits, guinea pigs, and pigs indicate that biphenyl is rapidly and readily
absorbed following oral exposure, as evidenced by the detection of metabolites in urine and bile
(Meyer, 1977; Meyer and Scheline, 1976; Meyer et al., 1976a: Meyer etal., 1976b). Results
from a study with rats administered radiolabeled biphenyl indicate extensive oral absorption
(Meyer et al., 1976b) (see below), whereas results from studies of rabbits, guinea pigs, and pigs
administered nonlabeled biphenyl indicate less extensive oral absorption in the range of 28-49%
of the administered dose (Meyer, 1977; Meyer et al., 1976a).
In the most quantitative assessment of absorption using radiolabeled biphenyl, male
albino rats (n = 3; body weight = 200-300 g) given an oral dose of 100 mg/kg (0.7-1.0 uCi) of
[14C]-biphenyl (in soy oil) excreted 75-80% of the radioactivity in their urine within the first
24 hours, with a total average urinary excretion of 84.8% and fecal excretion of 7.3% during the
96-hour postdosing period (Meyer et al., 1976b). Only a trace of [14C]-CO2 was detected in
expired air and <1% of the radioactivity was recovered from tissues obtained at the 96-hour
sacrifice of the rats. These results indicate that at least 85% of the administered dose was
absorbed and excreted from rats through urine or feces.
Less quantitative estimates of oral absorption have been provided in analytical studies of
biphenyl and metabolites in urine and feces from rabbits (Meyer, 1977), guinea pigs (Meyer,
1977), and pigs (Meyer et al., 1976a) following oral administration of single 100-mg/kg doses of
unlabeled biphenyl.
Male White Land rabbits and Sff:PIR guinea pigs were given biphenyl (100 mg/kg) by
gavage in soy oil, and urine and feces were collected at 24-hour intervals, up to 96 hours after
administration (Meyer, 1977). The phenolic metabolites of biphenyl were analyzed as
trimethylsilyl (TMS) ethers by combined gas chromatography (GC)/mass spectrometry (MS)
(guinea pigs) or GC (rabbits). The biphenyl was hydroxylated to monohydroxylated biphenyls
and minor amounts of dihydroxylated derivatives, with the main route of excretion being through
the urine in both species and the major metabolite being 4-hydroxybiphenyl. In guinea pigs
(n = 3), the mass of identified metabolites in urine collected at 24 or 96 hours post-exposure
accounted for 29.5 or 32.9% of the administered dose, respectively. In the first 24 hours,
biphenyl and biphenyl metabolites in feces accounted for 20.3% of the dose; most of this
(14.3%) was biphenyl, presumably unabsorbed. Bile was collected for 24 hours from another
group of two bile-cannulated guinea pigs dosed with 100 mg/kg biphenyl. No unchanged
biphenyl was detected in the collected bile, but conjugated mono- and dihydroxy metabolites
accounted for about 3% of the administered dose. The results with guinea pigs indicate that at
5 DRAFT - DO NOT CITE OR QUOTE
-------
least 33% of the administered dose was absorbed. In rabbits, urinary metabolites accounted for
49.1% of the dose, with most of this (25.4% on the first day and 15.9% on the second day)
eliminated as conjugates. In the first 24 hours, biphenyl and metabolites in feces accounted for
1.6% of the dose with 1.4% being biphenyl. These results indicate that at least 49% of the
administered dose was absorbed in rabbits.
Absorption of single oral 100 mg/kg doses of biphenyl (in soy oil or propylene glycol)
has also been demonstrated in male and female Danish Landrace pigs weighing 31-35 kg (Meyer
et al., 1976b). Metabolites identified in urine collected at four 24-hour intervals after dose
administration included mono-, di-, and trihydroxybiphenyls, detected as IMS ethers by GC/MS
after enzyme hydrolysis of the samples by p-glucuronidase and sulphatase. Metabolites
identified and quantified in 24-hour urine samples accounted for averages of 17.5 and 26.5% of
the dose administered in soy oil to two female pigs and in propylene glycol to two male pigs,
respectively. Unchanged biphenyl was not detected in the urine samples. Metabolites in urine
collected for 96 hours accounted for averages of 27.6 and 44.8% of the doses administered to
female and male pigs, respectively. No phenolic metabolites of biphenyl were detected in feces
collected for 96 hours. Unchanged biphenyl was not detected in the feces collected from male
pigs, but the amount of unchanged biphenyl in feces from the two female pigs accounted for
18.4 and 5% of the administered dose. These results indicate that at least about 28 and 45% of
oral 100 mg/kg doses of biphenyl were absorbed in female and male pigs, respectively. It is
uncertain if the gender difference was due to vehicle differences or actual gender differences in
absorption efficiency.
No animal studies were located examining quantitative aspects of absorption of biphenyl
by the respiratory tract or skin.
3.2. DISTRIBUTION
No information was located regarding distribution of absorbed biphenyl in humans and
limited animal data are available. Meyer et al. (1976a) orally administered 100 mg/kg
[14C]-biphenyl to male albino rats and measured radioactivity in the lung, heart, kidney, brain,
spleen, liver, skeletal muscles, peritoneal fat, genital tract, and gastrointestinal tract at 96 hours
after dosing. Most of the radioactivity was excreted in urine (84.8%) and feces (7.3%) over the
96-hour period, and only 0.6% of the administered radioactivity remained in the animals at
96 hours: 0.1% was found in peritoneal fat, 0.3% in the gastrointestinal tract (including its
contents), 0.1% in skeletal muscles, and 0.1% in the genital tract. Levels of radioactivity in other
examined tissues were very low. The results indicate that absorbed biphenyl is not preferentially
stored in tissues and is rapidly excreted, principally through the urine.
DRAFT - DO NOT CITE OR QUOTE
-------
3.3. METABOLISM
3.3.1. Identification of Metabolites
3.3.1.1. Results from In Vivo Animal Studies
No human studies on the in vivo metabolism of biphenyl have been identified. However,
the in vivo metabolism of biphenyl has been studied extensively in laboratory animals. These
studies have determined that in rats, rabbits, pigs, dogs, mice, and guinea pigs, biphenyl is
converted into a range of hydroxylated metabolites (Halpaap-Wood et al., 1981b: Meyer, 1977;
Meyer and Scheline, 1976; Meyer et al., 1976a: Meyer et al., 1976b). These metabolites have
been detected in urine as both nonconjugated compounds and acidic conjugates.
The derivation of urinary metabolites and their subsequent analysis with GC has resulted
in the identification of >10 mono-, di-, and trihydroxybiphenyl metabolites from the urine of rats,
pigs, guinea pigs, and rabbits (Meyer, 1977; Meyer and Scheline, 1976; Meyer et al., 1976a:
Meyer et al., 1976b). These metabolites have been found as mercapturic acid conjugates and
glucuronide conjugates (Millburn et al., 1967). Comparable metabolites have been identified
among mammalian species tested, although quantitative differences in metabolite formation are
evident among species. A major metabolite in the rat, mouse, guinea pig, rabbit, and pig was
reportedly 4-hydroxybiphenyl (Halpaap-Wood et al., 1981b: Meyer, 1977; Meyer and Scheline,
1976). 4,4'-Dihydroxybiphenyl was identified as a major metabolite in the pig (Meyer et al.,
1976b) and the rat (Halpaap-Wood et al., 1981b: Meyer and Scheline, 1976), while 3,4-di-
hydroxybiphenyl was a major urinary metabolite in two strains of mice (Halpaap-Wood et al.,
1981b). Table 3-1 reviews the metabolites that have been identified in the excreta and bile of
male albino rats given single doses of 100 mg biphenyl/kg, as reported by Meyer and Scheline
(1976).
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 -Hydroxybiphenyl
4-Hydroxybiphenyl
3 ,4 -Dihy droxybiphenyl
3 ,4 ' -Dihy droxybiphenyl
4,4 ' -Dihy droxybiphenyl
2,5 -Dihy droxybiphenyl
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).
The hydroxylation of biphenyl to produce 2-hydroxybiphenyl is a minor pathway in rats
and mice, but is more easily detected in mice than rats (Halpaap-Wood et al., 1981a, b).
Following intraperitoneal (i.p.) injection of [14C]-labeled biphenyl (30 mg/kg), the pattern of
percentages of radioactivity detected in urinary metabolites showed a relatively greater ability to
produce 2-hydroxybiphenyl in mice than rats. In Sprague-Dawley rats, metabolites identified in
order of abundance were (with percentage of total urinary radioactivity noted in parentheses):
4,4'-dihydroxybiphenyl (44.5%); 4-hydroxybiphenyl (28.5%); 3,4,4'-trihydroxybiphenyl (8.8%);
3,4'-dihydroxybiphenyl (8.5%); 3,4-dihydroxybiphenyl (5.1%); 3-hydroxybiphenyl (1.8%); and
2-hydroxybiphenyl (1.5%). In DBA/2Tex mice, major identified metabolites were: 4-hydroxy-
biphenyl (39.5%); 3,4-dihydroxybiphenyl (30.3%); 4,4'-dihydroxybiphenyl (10.2%);
3,4,4'-trihydroxybiphenyl (6.2%); 3-hydroxybiphenyl (4.3%); and 2-hydroxybiphenyl (4.2%).
In rats, 2,3-, 2,4-, and 2,5-dihyroxybiphenyl were detected at trace levels (<0.1%), whereas in
mice, these metabolites were detected at levels of 0.3, 0.8, and 0.7%, respectively (Halpaap-
Wood etal.. 1981b).
No in vivo studies have been identified that directly investigate differential metabolism of
biphenyl between males and females of any species. However, studies on urinary crystals and
calculi formation and composition after chronic exposure to biphenyl in the diet indicate that
male F344 rats are more susceptible than females to the formation of urinary bladder calculi
(Ohnishi etal., 2001; Ohnishi et al., 2000a; Ohnishi et al., 2000b). Urinary bladder calculi in
DRAFT - DO NOT CITE OR QUOTE
-------
males were predominantly composed of the insoluble potassium salt of 4-hydroxybiphenyl-O-
sulphate, whereas the less frequently occurring urinary bladder calculi in females were composed
mainly of 4-hydroxybiphenyl and potassium sulphate, hydrolysis products of 4-
hydroxybiphenyl-O-sulphate (Ohnishi etal., 2001; Ohnishi et al., 2000a: Ohnishi et al., 2000b).
These observations are consistent with observations that male rats have relatively higher urinary
potassium concentrations and pH values than female rats, and with the hypothesis that gender
differences in these urinary conditions (rather than gender differences in metabolism of biphenyl)
may be responsible for the gender differences in urinary calculi formation and the subsequent
development of non-neoplastic (hyperplasia) and neoplastic (papillomas and carcinomas) lesions
in male, but not female, F344 rats (Umeda et al.. 2002: Ohnishi etal.. 2001: Ohnishi et al..
2000a: Ohnishi et al.. 2000b).
3.3.1.2. Results from In Vitro Studies with Animal and Human Cells or Tissues
The metabolism of biphenyl in vitro has been investigated using tissues of human origin,
resulting in evidence that the human metabolism of biphenyl is qualitatively similar to, but may
be quantitatively different from, rat metabolism. Benford et al. (1981) measured 2-, 3-, and
4-hydroxylation of biphenyl in microsomes prepared from the livers of five rats (sex not
identified) and four humans (sex not identified). The reaction products, after solvent extraction
and high-performance liquid chromatography (HPLC) quantitation, revealed that 2-hydroxylase
in the rat was 35 times higher than in humans, while 3- and 4-hydroxylases in humans were
1.5 and 1.2 times higher than in rats.
The evidence from studies of human tissue samples exposed to biphenyl metabolites in
vitro suggests differential Phase II metabolism contingent upon tissue origin. Powis et al. (1988)
have shown that/>-hydroxybiphenyl is conjugated with glucuronic acid and sulphate in human
liver and kidney tissue slices. In the liver, glucuronidation was the favored conjugation pathway,
while sulphation was favored in the kidney. Powis et al. (1989) also compared Phase I biphenyl
metabolism in human (from surgery), dog (mongrel), and rat (male F344) liver slices and
primary hepatocytes. It was found that liver slices from all three species had a similar capacity
to metabolize biphenyl, -3.5 nmol biphenyl/minute per g tissue, while hepatocyte preparations
from rats had about 4 times the metabolic capacity of dog hepatocytes and about 20 times that of
human hepatocytes. Powis et al. (1989) speculated that hepatocytes from dog and human liver
slices may have experienced more damage during isolation than rat hepatocytes.
A study of the sulphation of biphenyl metabolites in human surgical tissue samples was
conducted by Pacific! et al. (1991). Tissue samples of various types (liver, intestinal mucosa,
lung, kidney, bladder, and brain) were obtained from surgeries of patients of both sexes between
the ages of 49 and 76 years of age (each patient contributed only one tissue type, so that within-
patient organ comparisons were not made). The tissues were homogenized, filtered, and
centrifuged at 12,000 and 105,000 g to obtain supernatants to study sulphation of biphenyl
9 DRAFT - DO NOT CITE OR QUOTE
-------
metabolites, specifically 2-, 3-, and 4-hydroxybiphenyl. Sulphotransferase activity for each of
these substrates was detected in all tissues studied, although marked tissue dependence was
observed, with the highest activity found in the liver and the lowest in the brain. The Michaelis
constant (Km) of Sulphotransferase was dependent on the substrate, but not on tissue type, with
Km varying over a 500-fold range. The highest values of Km were found with 4-hydroxybiphenyl
and the lowest were found with 3-hydroxybiphenyl.
Several studies of biphenyl metabolism with in vitro animal systems support the findings
from the in vivo urinary metabolite investigations that: (1) a range of hydroxylated biphenyl
metabolites are formed, (2) 4-hydroxybiphenyl is a major metabolite, and (3) hydroxylated
biphenyl metabolites are conjugated to glucuronic acid or sulphate. Wiebkin et al. (1984; 1976)
reported that isolated rat and hamster hepatocytes metabolized biphenyl primarily to
4-hydroxybiphenyl and also to 4,4'-hydroxybiphenyl, both of which were then conjugated. A
small amount of 2-hydroxybiphenyl was produced. When 4-hydroxybiphenyl was incubated
with the hepatocytes, it was hydroxylated to 4,4'-dihydroxybiphenyl. Pretreatment of the
animals with either 5,6-benzoflavone or phenobarbital had little effect on the conjugate
formation rate in the in vitro experiment. Bianco et al. (1979) reported that rat hepatic
microsomes metabolize biphenyl to 4-, 2-, and 3-hydroxybiphenyl, which are conjugated to form
glucuronides and sulphates. The 4-hydroxybiphenyl isomer was the major metabolite. The
formation of 4-hydroxybiphenyl as a major metabolite in the hamster, mouse, and rabbit was
confirmed by Billings and McMahon (1978) 2-Hydroxybiphenyl and 3-hydroxybiphenyl were
detected in a lower amount in a ratio of 2:1 by hamster and rabbit microsomes, and in a 1:1 ratio
by mouse microsomes. In contrast, almost all hydroxylation of biphenyl in rat microsomes gave
rise to 4-hydroxybiphenyl.
3.3.2. Metabolic Pathways
3.3.2.1. Description of Metabolic Scheme and Enzymes Involved
Burke and Bridges (1975) suggested that biphenyl metabolism is mediated by
cytochrome P-450 (CYP) monooxygenases. Evidence of an arene oxide intermediate, which
may participate in binding to cellular macromolecules, was reported by Billings and McMahon
(1978). Support for CYP metabolism of biphenyl was provided by Halpaap-Wood et al. (1981a,
b), who reported that greater amounts of hydroxybiphenyls were obtained in in vitro assays using
liver homogenates when rats were treated first with p-naphthoflavone, 3-methylcholanthrene, or
Aroclor 1254, which are known CYP inducers. In C57BL/6Tex mice, CYP induction with
P-naphthoflavone led to relatively greater amounts of urinary excretion of 2-hydroxybiphenyl,
compared with uninduced mice, whereas pretreatment with P-naphthoflavone led to increases in
urinary excretion of 2-, 3-, and 4-hydroxybiphenyl in Sprague-Dawley rats and was without
influence on the pattern of hydroxybiphenyl metabolites in DBA/2Tex mice (Halpaap-Wood et
al.. 1981b).
10 DRAFT - DO NOT CITE OR QUOTE
-------
Figure 3-1 details combined evidence from the Halpaap-Wood et al. (1981a, b) and
Meyer and Scheline (1976b) studies on the metabolic pathways of biphenyl. While sulphates
and glucuronides are formed on all three metabolic levels illustrated, only monosulphates and
monoglucuronides are identified. Monomethyl ethers are formed from dihydroxy and trihydroxy
metabolites alone. Glucuronides at the dihydroxy and trihydroxy levels are additionally labeled
with a question mark to suggest that, while these metabolites are likely, they have not been
identified.
11 DRAFT - DO NOT CITE OR QUOTE
-------
HO
OH
4-Hydroxybiphenyl
2-Hydroxybiphenyl
CYP
3-Hydroxybiphenyl
ar.ar'-Dihydroxybiphenyl
OH
ar.ar-Dihydroxybiphenyl
CYP
OH
HO.
'OH
ar.ar.ar'-Trihydroxybiphenyl
Sulphono-
transferase^ Monohydroxysulfate
*" -glucuronide
UGT
COMT
Dihydroxy-monosulfate
-monoglycuronide (?)
-monomethyl ether
Trihydroxy-monosulfate
-monoglucuronide (?)
-monomethyl ether
ar = aryl group; COMT = catechol-O-methyltransferase; UGT = uridine
diphosphate glucuronosyl transferase; question marks denote tentative metabolites
(see text).
Sources: Halpaap-Wood et al. (1981a, b); Meyer and Scheline (1976b).
Figure 3-1. Schematic presentation of the metabolic pathways of biphenyl.
12
DRAFT - DO NOT CITE OR QUOTE
-------
The metabolic scheme in Figure 3-1 does not include the possible redox cycling of
2,5-dihydroxybiphenyl (also known as phenylhydroquinone), which involves CYP-mediated
cycling between phenylhydroquinone and phenylbenzoquinone leading to the generation of
reactive oxygen species (ROS) (Balakrishnan et al., 2002; Kwoketal., 1999). This pathway is
thought to play a role in the carcinogenic effect of 2-hydroxybiphenyl (also known as
ort/70-phenylphenol), a broad spectrum fungicide that, like biphenyl, induces urinary bladder
tumors in chronically exposed male rats (Kwok et al., 1999). Free 2,5-dihydroxybiphenyl and its
glucuronide or sulphate conjugates are readily detected in the urine of rats exposed to
2-hydroxybiphenyl, and the formation of 2,5-dihydoxybiphenyl and phenylbenzoquinone is the
principal metabolic pathway for 2-hydroxybiphenyl in the rat, especially at high exposure levels
associated with urinary bladder tumor formation (Kwoketal., 1999; Morimoto et al., 1989;
Nakao et al., 1983; Reitz et al., 1983; Meyer and Scheline, 1976). In contrast, the formation of
4-hydroxybiphenyl and 4,4'-dihydroxybiphenyl is the principal metabolic pathway for biphenyl
in rats and mice, and 2,5-dihydroxybiphenyl was not detected, or only detected at trace levels, in
the urine of rats exposed to 100 mg biphenyl/kg (Meyer et al., 1976b) (see Table 3-1). In mice
exposed to i.p. doses of [14C]-biphenyl (30 mg/kg), radioactivity in 2-hydroxybiphenyl and 2,5-
dihydroxybiphenyl in the urine accounted for only about 5% of the total radioactivity detected in
urinary metabolites (Halpaap-Wood et al., 1981b).
3.3.3. Regulation of Metabolism, Sites of Metabolism, and Relationships to Toxic Effects
3.3.3.1. Evidence for Induction of Phase I and II Enzymes
No studies of Phase I or II enzyme induction using liver microsomes of human origin
were identified. However, a number of studies have been conducted in rodents to investigate the
induction of Phase I enzymes that catalyze biphenyl hydroxylation. For example, Creaven and
Parke (1966) reported that pretreatment of weanling Wistar rats or ICI mice with phenobarbital
[an inducer of CYP3A4, 2B6, and 2C8 as reported by Parkinson and Ogilvie (2008)1 or
3-methylcholanthrene [an inducer of CYP1A2 as reported by Parkinson and Ogilvie (2008)1
increased NADPH-dependent activities of liver microsomes to produce 2-hydroxybiphenyl and
4-hydroxybiphenyl from biphenyl to varying degrees depending on the inducer. Haugen (1981)
reported that pretreatment of male CD rats with phenobarbital or 3-methylcholanthrene increased
NADPH-dependent activities of liver microsomes to produce 2-, 3-, and 4-hydroxybiphenyl from
biphenyl, again to varying degrees depending on the inducer. Stuehmeier et al. (1982) reported
that phenobarbital pretreatment of male C57BL/6JHan mice induced liver microsomal activities
to produce 4-hydroxybiphenyl, but not 2-hydroxybiphenyl, from biphenyl, whereas
3-methylcholanthrene induced activities for both 4- and 2-hydroxylation of biphenyl. Halpaap-
Wood et al. (1981b) reported that pretreatment of male Sprague-Dawley rats with
p-naphthoflavone [an inducer of CYP1A2 as reported by Parkinson and Ogilvie (2008): also
known as 5,6-benzoflavone] enhanced the urinary excretion of 2-, 3-, and 4-hydroxybiphenyl,
13 DRAFT - DO NOT CITE OR QUOTE
-------
3,4-dihydroxybiphenyl, and 3,4,4'-trihydroxybiphenyl following i.p. administration of 30 mg
biphenyl/kg body weight. In contrast, pretreatment of male C57BL/6Tex mice with
p-naphthoflavone did not increase the overall urinary excretion of biphenyl metabolites
following i.p. administration of 60 mg biphenyl/kg, but shifted the principal metabolite from
4-hydroxybiphenyl to 2-hydroxybiphenyl and 2,5-dihydroxybiphenyl (Halpaap-Wood et al.,
1981b). Wiebkin et al. (1984) reported that P-naphthoflavone pretreatment of male Lewis rats or
male Syrian golden hamsters induced biphenyl hydroxylation activities in freshly isolated
pancreatic acinar cells or hepatocytes. From these observations and examination of patterns of
inhibition of biphenyl hydroxylation activities by CYP inhibitors (e.g., a-naphthoflavone and
1-benzyl-imidazole) under non-induced and induced conditions (Haugen, 1981), it is apparent
that multiple CYP enzymes (e.g., CYP1A2 and CYP3A4) are likely involved in biphenyl
hydroxylation. However, no studies were located that used more modern techniques (such as
CYP knockout mice) to identify the principal CYP enzymes involved in the initial hydroxylation
of biphenyl or the formation of the dihydroxy- or trihydroxybiphenyl metabolites.
Several animal studies were located examining the possible coordinated induction of
Phase I enzymes with Phase II enzymes catalyzing the conjugation of hydroxylated biphenyl
metabolites to sulphate or glucuronic acid. Hepatocytes from rats (strain and sex were not noted)
pretreated with the CYP inducers, phenobarbital or 3-methylcholanthrene, produced glucuronide
and sulphate conjugates of 4-hydroxybiphenyl when incubated with biphenyl (Wiebkin et al.,
1978). Glucuronide conjugates were predominant under these "CYP-induced" conditions,
whereas hepatocytes from non-induced control rats produced predominant sulphate conjugates of
4-hydroxybiphenyl. These results suggest that induction (or possibly activation) of
glucuronidation enzymes may be coordinated with the induction of CYP enzymes. In contrast,
pretreatment of male Lewis rats with P-naphthoflavone (an inducer of CYP1A2) did not enhance
activities of freshly isolated pancreatic acinar cells to conjugate 4-hydroxybiphenyl with sulphate
or glucuronic acid, but the influence of this pretreatment on the conjugation capacity of
hepatocytes was not examined in this study (Wiebkin et al., 1984). In another study, uridine
diphosphate glucuronosyl transferase (UGT) activities with 1-naphthol or 3-hydroxy-
benzo[a]pyrene as substrates were higher in liver microsomes from male Wistar rats pretreated
with Aroclor 1254 (an inducer of several CYP enzymes) or phenobarbital, respectively,
compared with microsomes from control rats without pretreatment with CYP inducers (Bock et
al., 1980). Although Bock et al. (1980) measured UGT activities in microsomes from several
tissues from non-induced rats with 4-hydroxybiphenyl as a substrate, no comparisons between
induced and non-induced conditions were made using 4-hydroxybiphenyl as substrate. Paterson
and Fry (1985) reported that hepatocytes or liver slices from male Wistar rats pretreated with
P-naphthoflavone showed decreased rates of glucuronidation of 4-hydroxybiphenyl, compared
with hepatocytes or liver slices from rats without P-naphthoflavone pretreatment. Results from
this database provide equivocal evidence that the induction of Phase I enzymes catalyzing the
14 DRAFT - DO NOT CITE OR QUOTE
-------
hydroxylation of biphenyl may be coordinated with induction of Phase II enzymes catalyzing
glucuronidation of hydroxylated biphenyl metabolites.
3.3.3.2. Demonstrated Tissue Sites of Metabolism
CYP enzymes catalyzing hydroxylation of biphenyl and other substrates are present in
most, if not all, mammalian tissues, but the highest levels of activities are normally found in liver
(Parkinson and Ogilvie, 2008). In a study of male Sprague-Dawley rats, CYP content was 20-
40-fold higher in the microsomes from liver than from lung, although biphenyl-4-hydrolase
activity was only 1.7-fold higher in the microsomes from liver than from lung (Matsubara et al.,
1974). Wiebkin et al. (1984) observed 200- and 1,000-fold higher rates of biphenyl metabolism
in 5,6-benzoflavone-pretreated hepatocytes compared to similarly treated pancreatic acinar cells
from male Lewis rats and Syrian golden hamsters, respectively.
Activities for enzymes catalyzing the conjugation of hydroxybiphenyls and other
hydroxylated aromatic compounds with glucuronic acid or sulphate have been detected in a
number of mammalian tissues, and, similar to CYP, the highest levels are found in the liver
(Parkinson and Ogilvie, 2008). Available data for conjugation activities with hydroxybiphenyls
in various mammalian tissues are consistent with this concept. Sulphotransferase activities with
2-, 3-, or 4-hydroxybiphenyl as substrates in microsomes from several human tissues showed an
approximate 100- to 500-fold range with the following order: liver > ileum > lung > colon >
kidney > bladder > brain (Pacifici etal., 1991). UGT activities with 4-hydroxybiphenyl as
substrate in microsomes from several male Wistar rat tissues showed the following order: liver >
intestine > kidney > testes ~ lung; activities were below the limit of detection in microsomes
from skin and spleen (Bock et al., 1980).
3.4. ELIMINATION
No studies were located on the route or rate of elimination of biphenyl in humans, but
results from studies of orally exposed animals indicate that absorbed biphenyl is rapidly
eliminated from the body, principally as conjugated hydroxylated metabolites in the urine.
The most quantitative data on the routes and rates of elimination come from a study of
rats following administration of radiolabeled biphenyl (Meyer et al., 1976b). Urine collected for
24 hours after the oral administration of 100 mg/kg [14C]-labeled biphenyl in soy oil to male
albino rats contained 75.8% of the administered radioactivity, compared with 5.8% detected in
feces collected in the same period. Ninety-six hours after dose administration, <1% of the
administered radioactivity remained in tissues, 84.8% was in collected urine, 7.3% was in feces,
and 0.1% was in collected expired air (Meyer et al., 1976a). Although chemical identity analysis
of fecal radioactivity was not conducted by Meyer et al. (1976a) (results from GC/MS analyses
of bile collected from bile-cannulated rats given single 100 mg/kg doses of unlabeled biphenyl
indicate that biliary excretion of metabolites represents a minor pathway of elimination (Meyer
15 DRAFT - DO NOT CITE OR QUOTE
-------
and Scheline, 1976). In bile collected for 24 hours, unchanged biphenyl was not detected and
conjugated metabolites accounted for 5.2% of the administered dose; in contrast, conjugated
metabolites of biphenyl in 24-hour urine accounted for 22.3% of the dose (Meyer and Scheline,
1976).
Supporting evidence for the importance of urinary elimination of conjugated metabolites
is provided by the results of other studies, which analyzed biphenyl and biphenyl metabolites by
GC/MS or GC in urine and feces collected from rabbits (Meyer, 1977), guinea pigs (Meyer,
1977), and pigs (Meyer et al., 1976b) following oral administration of 100 mg/kg doses of
unlabeled biphenyl. In 24-hour urine samples, unchanged biphenyl was not detected, and total
metabolites accounted for averages of 25.4% of the administered dose in rabbits, 31.3% in
guinea pigs, 17.5% in female pigs, and 26.4% in male pigs. As in rats, biliary excretion
represents a minor elimination pathway in guinea pigs and rabbits; metabolites detected in bile
collected for 24 hours from bile-cannulated guinea pigs accounted for 3.3% of the administered
dose, but for only 0.3% of the dose in bile collected for 7 hours from a rabbit given 100 mg/kg
biphenyl (Meyer, 1977). Neither unchanged biphenyl nor hydroxylated biphenyl metabolites
were detected in bile collected from a bile-cannulated pig for 24 hours after administration of
100 mg/kg biphenyl (Meyer et al., 1976b).
No studies were located examining quantitative aspects of elimination in animals
following inhalation or dermal exposure to biphenyl.
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS
No studies were located on the development of PBPK models for biphenyl in animals or
humans.
16 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; Hakkinen et al., 1971) and another mill in Sweden
8 (Wastensson et al.. 2006).
9 A case report of a 46-year-old female who worked at a fruit-packing facility over a 25-
10 year period where biphenyl-impregnated paper was used presented with hepatomegaly,
11 neutrophilic leukocytosis, clinical chemistry findings indicative of hepatic perturbation, and liver
12 biopsy indicative of chronic hepatitis (Carella and Bettolo, 1994). Following cessation of work
13 in citrus packing, serum enzymes returned to normal, suggesting that occupational exposure to
14 biphenyl may have been the principal etiological factor.
15 Hakkinen and colleagues assessed the health of paper mill workers exposed to biphenyl
16 during the production of biphenyl-impregnated paper used to wrap citrus fruits. In 1959,
17 workers complained about a strong odor and irritation to the throat and eyes. Air measurements
18 made at various locations within the facility in June of 1959 resulted in estimated average
19 biphenyl concentrations of 4.4-128 mg/m3 (Table 4-1). In 1969, a 32-year-old worker at the
20 facility, who had worked for 11 years in the oil room where biphenyl levels were particularly
21 high, became ill. Despite aggressive medical intervention, the patient grew worse and died. Key
22 features at autopsy included necrosis of most liver cells, severe, but unspecified changes in the
23 kidneys, degeneration of the heart muscles, hyperactive bone marrow, and edematous changes in
24 the brain (Hakkinen et al., 1973; 1971). Subsequent measurements of biphenyl in the workplace
25 air (January 1970) resulted in estimated average concentrations ranging from 0.6 to 123 mg/m3
26 (Table 4-1). Measurements taken in both 1959 and 1971 indicated that biphenyl air
27 concentrations at multiple work areas greatly exceeded the current American Conference of
28 Governmental Industrial Hygienists (ACGUI 2001) threshold limit value (TLV) of 0.2 ppm
29 (1.3 mg/m3). In the location where biphenyl was mixed with paraffin oil (the oil room), biphenyl
30 occurred both as a vapor and as a dust, suggesting the possibility of both dermal and inhalation
31 exposures.
32
17 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
Average concentrations (mg/m3)
June 1959
January 1970
Paper mill hall
In front of paper reel
Behind impregnating roller
Near paper machine
Near rolling machine
17.9
128.0
7.2
4.4
7.2
64.0
1.5
0.6
Oil-room
Near measuring container
Above measuring container (lid open)
Near mixing container
During addition of biphenyl to mixing container
19.5
No data
No data
No data
3.5
123.0
15.5
74.5
Source: Hakkinen et al. (1973).
1
2 Thirty-one male workers at the Finnish facility engaged in the biphenyl-impregnation
3 process and two other workers exposed to biphenyl were included in the study. Common
4 complaints among these workers included fatigue, headache, gastrointestinal discomfort,
5 numbness and aching of the limbs, and general fatigue; laboratory tests revealed elevated serum
6 aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (which can indicate
7 inflammation or damage to liver cells) in 10 of the 33 workers (Hakkinen et al., 1973). Eight of
8 the 33 workers were admitted to the hospital for further examination, including liver biopsy. The
9 majority of the 33 workers were subjected to neurophysiological examinations, including
10 electroencephalography (EEG) and electroneuromyography (ENMG, consisting of nerve
11 conduction velocity and electromyographic [EMG] tests). Seppalainen and Hakkinen (1975)
12 published the most comprehensive results of the neurophysiological examinations. In all, 24
13 subjects (including the 8 hospitalized workers) underwent neurophysiological examinations.
14 Exposure to biphenyl was terminated immediately following the initial neurophysiological
15 examinations, and 11 and 7 of these subjects were retested 1 and 2 years later, respectively.
16 EEG results. At initial examination, 10 of the 24 workers had abnormal EEGs, which
17 included diffuse slow wave abnormalities (6 cases), lateral spike and slow wave discharges
18 (2 cases), posterior slowing only (1 case), and mild slow wave abnormality in the right temporal
19 area (1 case). Six subjects exhibited unusual distribution of alpha rhythm with alpha activity also
20 prominent in the frontal areas. Four of the subjects exhibited no EEG abnormalities. In general,
21 the EEG results observed at initial examination were qualitatively similar in the 11 subjects
22 reexamined 1 year later. Exceptions included additional diffuse slow wave abnormalities in the
23 two subjects initially exhibiting only spike and wave discharges and the disappearance of the one
18
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
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 andp < 0.01,
respectively) compared to the initial measurements. Abnormal EMGs among the biphenyl-
exposed workers included diminished numbers of motor units on maximal muscle contraction
(10 subjects) and fibrillations in some muscles (7 subjects). Workers exhibiting abnormal EMGs
typically displayed slowing of some nerve conduction velocities as well. Of those 11 subjects
undergoing repeat ENMG examination after 1 year, 5 subjects showed an increased level of
ENMG abnormality, while 4 remained unchanged and 2 had diminished abnormalities. At the
end of 2 years, three of seven subjects displayed diminished ENMG abnormalities, three of seven
were unchanged, and one of seven had the abnormality increased.
Table 4-2. Nerve conduction velocities of 24 persons exposed to biphenyl:
comparison with 60 unexposed males
Nerve
Biphenyl group
(mean ± SD)
Control group
(mean ± SD)
f-test
Median
MCV
57.7 ±6.3
58.0 ±3.8
Not significant
Ulnar
MCV
CVSF
56.3 ±4.6
41.4 ±5.2
56.6 ±4.0
45.5 ±3.2
Not significant
p< 0.001
Deep peroneal
MCV
CVSF
50.2 ±5.4
37.7 ±3.9
50.3 ±3.5
38.2 ±5.6
Not significant
Not significant
Posterior tibial
MCV
43.4 ±3.9
42.4 ±4.7
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
19
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. At a facility manufacturing biphenyl-
6 impregnated paper in Sweden, a cluster of five cases of Parkinson's disease (PD) among the
7 employees was investigated (Wastensson et al. 2006). Since, according to the report, the
8 expected prevalence of PD would be less than 0.9 cases from the 255 employees at the facility
9 (relative risk [RR] 5.6 [95% confidence interval 1.9-13]), it was suspected that the elevated PD
10 at the facility may have been related to biphenyl exposure. Four of the subjects worked in the
11 vicinity of a rewinder/dryer, while the fifth attended to another rewinder. Although no ambient
12 biphenyl levels were available for the subjects' work space, it was thought likely that the level of
13 biphenyl in air would be greater than the existing TLV of 1.3 mg/m3 (0.2 ppm) based on
14 measurements at a Finnish paper mill with similar production practices (Hakkinen et al., 1973).
15 Two subjects may have been exposed to higher levels of biphenyl than the others when they
16 created the paraffin oil/biphenyl mixture.
17 In addition to comparing existing PD cases to national trends, Wastensson et al. (2006)
18 examined the medical records of 222 former employees who had died. Nine cases of PD were
19 found among the decedents, compared with 4.3 cases of PD expected from data on the lifetime
20 risk of developing PD in the general population. This comparison yielded an RR of 2.1, with a
21 95% confidence interval of 0.96-4.0.
22 Clinical features and exposure data for the five living subjects, all of whom were
23 diagnosed with PD by a neurologist at a local hospital, are summarized in Table 4-3. With one
24 exception, the patients were in comparatively good health on initial diagnosis. The exception
25 was a 53-year-old male who had diabetes mellitus and withdrew from the study before his
26 neurological condition could be confirmed. Assuming that the diagnoses of PD were valid, the
27 central issue is whether these data indicate a causal relationship between PD and exposure to
28 biphenyl. Wastensson et al. (2006) discussed this issue in the context of other studies that have
29 pointed to a possible cause-and-effect relationship between pesticide exposure and PD, but were
30 unable to draw any firm conclusions from their limited data.
31
20 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-3. Exposure data and clinical features for five PD 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
Bradykinesia
Positive response to levodopab
+
+
+
+
+
+
+
+
+
+
+
+
+
-
+
+
+
+
-
+
""Exposure to biphenyl about one-third of each year.
bAll five patients improved with levodopa, which is a medication for PD.
PM = paper mill
Source: Wastensson et al. (2006)
1
2 4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
3 ANIMALS—ORAL AND INHALATION
4 Available oral data for biphenyl include two well-designed two-year chronic toxicity and
5 carcinogenicity studies, one in F344 rats (Umeda et al., 2002) and one in BDFi mice (Umeda et
6 al., 2005). Increased incidence of urinary bladder transitional cell papillomas and carcinomas,
7 associated with the formation of urinary bladder calculi, occurred in male, but not female,
8 F344 rats at the highest tested dietary concentration, 4,500 ppm, but were not found at lower
9 exposure levels of 1,500 or 500 ppm. Non-neoplastic kidney lesions (simple transitional cell
10 hyperplasia in the renal pelvis and hemosiderin deposits) were found in female F344 rats at
11 biphenyl dietary concentrations^ 1,500 ppm (Umeda et al., 2002). Several other rat studies
12 provide supporting evidence that the kidney and other urinary tract regions are critical targets for
13 biphenyl in rats (Shiraiwa et al., 1989; Ambrose etal., I960; Pecchiai and Saffiotti, 1957; Dow
14 Chemical Co, 1953). In BDFi mice, increased incidence of liver tumors (hepatocellular
15 adenomas and carcinomas) and non-neoplastic effects on the kidney (mineralization) and liver
16 (increased activities of plasma ALT and AST) were found in females exposed to biphenyl dietary
17 concentrations of 2,000 or 6,000 ppm (Umeda et al., 2005). In contrast, no carcinogenic
18 responses or noncancer adverse effects were found in female ddY mice exposed to 5,000 ppm
19 biphenyl in the diet for 2 years (Imai et al., 1983) or in B6C3Fi and B6AKFi mice exposed to
20 517 ppm biphenyl in the diet for 18 months (Innesetal.. 1969: NCL 1968).
21
DRAFT - DO NOT CITE OR QUOTE
-------
1 No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
2 toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
3 biphenyl (7 hours/day, 5 days/week for up to about 90 days) were reported in mice exposed to
4 5 mg/m3 and in rats exposed to 300 mg/m3, but not in rabbits exposed to 300 mg/m3 (Deichmann
5 et al., 1947; Monsanto, 1946). Congestion or edema of the lung, kidney, and liver, accompanied
6 by hyperplasia with inflammation of the trachea, was found in CD-I mice exposed to biphenyl
7 vapors at 25 or 50 ppm (158 or 315 mg/m3) for 13 weeks (Sun, 1977a).
8 Study descriptions for all available subchronic and chronic toxicity and carcinogenicity
9 studies follow.
10
11 4.2.1. Oral Exposure
12 4.2.1.1. Subchronic Toxicity
13 Twenty-one-day-old female Long-Evans rats (8/group) were exposed to 0, 0.01, 0.03, or
14 0.1% biphenyl in the diet for 90 days (Dow Chemical Co.). Body weights were monitored 3
15 times/week, and the weights of the liver, kidneys, adrenals, and spleen were recorded at
16 necropsy. Heart, liver, kidney, spleen, adrenals, pancreas, ovary, uterus, stomach, small and
17 large intestine, voluntary muscle, lung, thyroid, and pituitary from each rat were examined
18 histopathologically (2 rats/group).
19 Based on U.S. EPA (1988) subchronic reference values for body weight and food
20 consumption in female Long-Evans rats, doses of biphenyl resulting from the dietary levels of
21 0.01, 0.03, and 0.1% are estimated to have been 10, 30, and 100 mg/kg-day, respectively. There
22 were no significant treatment-related effects on body weight, food consumption, or organ
23 weights. Results of histopathologic examinations were unremarkable. Biphenyl-exposed groups
24 exhibited lower average plasma blood urea nitrogen (BUN) levels than controls (28.2, 25.7, and
25 26.3 mg percent for low-, mid-, and high-dose groups, respectively, compared to 35.3 mg percent
26 for controls), although the statistical significance of these apparent treatment-related differences
27 was not reported and the biological significance is uncertain.
28 Six-week-old BDFi mice (10/sex/group) were exposed to biphenyl at dietary
29 concentrations of 0, 500, 2,000, 4,000, 8,000, 10,000, or 16,000 ppm for 13 weeks (Umeda et al.
30 2004b). To overcome possible problems with taste aversion, mice assigned to the 8,000 and
31 10,000 ppm groups were fed 4,000 ppm dietary biphenyl for the first week and 8,000 or 10,000
32 ppm for the remaining 12 weeks. Mice designated to receive 16,000 ppm were fed 4,000 ppm
33 dietary biphenyl for the first week, 8,000 ppm for the second week, and 16,000 ppm for the
34 remaining 11 weeks. Animals were checked daily for clinical signs; body weight and food
35 consumption were recorded weekly; organ weights were noted at term; and liver sections were
36 processed for light microscopic examination. Electron microscopy was carried out on liver
37 tissue from one control and one 16,000 ppm female.
22 DRAFT - DO NOT CITE OR QUOTE
-------
1 Based on U.S. EPA (1988) subchronic default reference values for body weight and food
2 consumption (average values for combined sexes), doses of biphenyl for the dietary
3 concentrations of 500, 2,000, 4,000, 8,000, 10,000, and 16,000 ppm are estimated to have been
4 93, 374, 747, 1,495, 1,868, and 2,989 mg/kg-day, respectively. A single 16,000 ppm female
5 mouse died during the study; all other mice survived until terminal sacrifice. Final body weights
6 of mice of both sexes in the 8,000, 10,000, and 16,000 ppm groups were significantly lower than
7 gender-matched controls (for males: 83.3, 84.9, and 75.1% of controls; for females: 93.7, 91.6,
8 and 85.8% of controls, respectively). Umeda et al. (2004a) noted that absolute liver weights
9 were significantly higher in 8,000 and 16,000 ppm female mice, but did not include the extent of
10 these increases in the study report. Light microscopic examination of liver specimens from all
11 16,000 ppm female mice revealed enlarged centrilobular hepatocytes, the cytoplasm of which
12 was filled with numerous eosinophilic fine granules. Upon electron microscopic examination,
13 these eosinophilic granules were identified as peroxisomes, indicative of a peroxisome
14 proliferative effect in the liver of the 16,000 ppm female mice. Evidence of histopathologic liver
15 lesions was not found in females of the 8,000 or 10,000 ppm groups. There were no signs of
16 treatment-related increased liver weight or histopathologic evidence of clearly enlarged
17 hepatocytes in any of the biphenyl-treated groups of male mice. Based on the significant
18 decrease in body weight in both genders, EPA identified 1,495 mg/kg-day as the LOAEL and
19 747 mg/kg-day as the NOAEL.
20
21 4.2.1.2. Chronic Toxicity and Carcinogenicity
22 4.2.1.2.1. Chronic rat studies
23 In a chronic toxicity and carcinogenicity study of F344 rats (50/sex/group), biphenyl was
24 administered in the diet for 2 years at concentrations of 0, 500, 1,500, or 4,500 ppm (Umeda et
25 al., 2002). All animals were examined daily for clinical signs; body weights and food intake were
26 determined once a week for the first 14 weeks and every 4 weeks thereafter. Urinalysis was
27 performed on all surviving rats during week 105. Upon necropsy, weights of all major organs
28 were recorded; all major organs and tissues were subjected to histopathologic examination.
29 The study report included a plot of mean body weights during the 2-year study, but did
30 not include food consumption data. Estimated doses, therefore, were calculated using time-
31 weighted average (TWA) body weights from the graphically-depicted data (Umeda et al., 2002
32 Figure 1) and U.S. EPA (1988) chronic reference values for food consumption in F344 rats. The
33 resulting estimated doses for the 500, 1,500, and 4,500 ppm exposure groups were 36.4, 110, and
34 378 mg/kg-day, respectively, for males and 42.7, 128, and 438 mg/kg-day, respectively, for
35 females. The study authors reported significantly lower mean body weights among 4,500 ppm
36 rats of both sexes compared to their respective controls. Mean body weights of 4,500 ppm male
37 and female rats were lower than those of controls throughout most of the study period and were
38 approximately 20% lower than respective controls at terminal sacrifice. There was no significant
23 DRAFT - DO NOT CITE OR QUOTE
-------
1 effect on mean body weights of 500 or 1,500 ppm males or females. Survival of low- and mid-
2 dose male and female rats was not significantly different from controls. The study authors
3 reported that 3/50 of the 4,500 ppm female rats died after 13-26 weeks of biphenyl exposure and
4 attributed the deaths to marked mineralization of the kidneys and heart. However, they also
5 indicated that survival of this group was not adversely affected thereafter. Significantly
6 decreased survival was noted only for the group of 4,500 ppm male rats, 19/50 of which died
7 prior to terminal sacrifice. The first death occurred around treatment week 36; this rat exhibited
8 urinary bladder calculi. Survival data for the other groups were not provided. Evidence of
9 hematuria was first noted in 4,500 ppm male rats around week 40 and was observed in a total of
10 32/50 of the 4,500 ppm males during the remainder of the treatment period; 14 of these rats
11 appeared anemic. Hematuria and bladder tumors were primarily considered as causes of death
12 among the 4,500 ppm males (n = 19) that died prior to terminal sacrifice. Urinalysis performed
13 during the final treatment week revealed significantly increased urinary pH in the 31 remaining
14 4,500 ppm male rats (pH of 7.97 versus 7.66 for controls; p < 0.05); occult blood was noted in
15 the urine of 23 of these males. Urine samples in 10/37 surviving 4,500 ppm females tested
16 positive for occult blood. Significant increases in relative kidney weights of 1,500 and
17 4,500 ppm males and females and absolute kidney weights of 4,500 ppm males were reported,
18 but actual data were not presented.
19 Gross pathologic examinations at premature death or terminal sacrifice revealed the
20 presence of calculi in the bladder of 43/50 of the 4,500 ppm males and 8/50 of the 4,500 ppm
21 females (Table 4-4); these lesions were not seen in 500 or 1,500 ppm male or female rats. The
22 bladder calculi in the male rats were white, yellow, brown, gray, and black in color, ranged from
23 0.3 to 1.0 cm in size, and exhibited triangular, pyramidal, cuboidal, and spherical shapes. The
24 bladder calculi in the female rats were white and yellow in color, of uniform spheroidal shape,
25 and similar in size to those of the male rats. Forty-one of the 4,500 ppm male rats exhibited
26 polyp-like or papillary nodules protruding into the lumen from the bladder wall; bladder calculi
27 were noted in 38 of these males. Four of the eight calculi-bearing 4,500 ppm female rats also
28 exhibited thickening of the bladder wall. It was noted that 30/32 of the 4,500 ppm male rats with
29 hematuria also exhibited kidney or urinary bladder calculi.
30
24 DRAFT - DO NOT CITE OR QUOTE
-------
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)
Males (n = 50)
0
0
500
36.4
1,500
110
4,500
378
Females (n = 50)
0
0
500
42.7
1,500
128
4,500
438
Lesion
Transitional cell
Simple hyperplasia3
Nodular hyperplasia3
Papillary hyperplasia3
Combined
Papilloma
Carcinoma
Papilloma or carcinoma
(combined)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
12b
40b
1?b
45
10b
24b
31b
0
1
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
1
5C
4
10b
0
0
0
Squamous cell
Metaplasia3
Hyperplasia3
Papilloma or carcinoma
(combined)
Inflammatory polyp3
Calculi
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19b
13b
1
10b
43b
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
1
0
0
8b
3The 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)
1
2 Histopathologic examinations at death or terminal sacrifice revealed no indications of
3 biphenyl-induced tumors or tumor-related lesions in organs or tissues other than those associated
4 with the urinary tract. As shown in Table 4-4, neoplastic and nonneoplastic lesions of the
5 urinary bladder were essentially limited to the 4,500 ppm rats and predominantly the males.
6 Only 4,500 ppm male rats exhibited papilloma (10/50) or carcinoma (24/50) of transitional cell
7 epithelium, three of which exhibited both papilloma and carcinoma. Most of the transitional cell
8 carcinomas (20/24) projected into the lumen, and the tumor cells invaded the entire body wall.
9 Bladder calculi were found in all 24 males with transitional cell carcinoma and 8/10 of the males
10 with transitional cell papilloma. Among noncancerous responses in the bladder, simple, nodular,
11 and papillary hyperplasias were evident in 4,500 ppm animals. These hyperplasias developed in
12 the focal area of the bladder epithelium. Simple hyperplasia occurred less frequently than
13 nodular and papillary hyperplasias; furthermore, simple hyperplasia was almost always
14 accompanied by either nodular or papillary hyperplasia in the 4,500 ppm males. Ten of the
15 4,500 ppm males had polyps in the bladder epithelium, which were composed of spindle fibers
25
DRAFT - DO NOT CITE OR QUOTE
-------
1 proliferated around transitional epithelial cells accompanied by inflammatory infiltration of
2 submucosal bladder epithelium. Squamous metaplasia was noted on the surface of the polyps,
3 which were found at different loci than the bladder tumors.
4 Table 4-5 summarizes the incidences of lesions of the ureter and kidney in the male and
5 female rats. The incidence of simple transitional cell hyperplasia in the ureter was greater in the
6 4,500 ppm males than the 4,500 ppm females. Other responses, such as mineralization of the
7 corticomedullary junction, were increased over controls to a greater extent in males compared to
8 females. In the renal pelvis, simple and nodular hyperplasia was frequently observed in
9 4,500 ppm males and 500 and 1,500 ppm females. Responses such as papillary necrosis, infarct,
10 and hemosiderin deposition occurred predominantly in exposed females.
11
26 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)
Males (n = 50)
0
0
500
36.4
1,500
110
4,500
378
Females (n = 50)
0
0
500
42.7
1,500
128
4,500
438
Response
Ureter
Transitional cell simple hyperplasia
Transitional cell nodular hyperplasia
Dilatation
1
0
0
0
0
0
0
0
0
8a
1
14a
0
0
0
0
0
0
0
0
0
2
0
6b
Kidney
Renal pelvis
Transitional cell simple hyperplasia
Transitional cell nodular hyperplasia
Squamous metaplasia
Mineralization
Desquamation
Calculi
6
0
0
9
1
0
8
1
0
6
0
0
5
1
0
10
0
0
19C
21a
2
18b
lla
13a
3
0
0
12
0
0
5
0
0
12
0
0
12C
1
0
18
0
0
25a
12a
0
27a
2
3
Other
Mineralization of corticomedullary
junction
Mineralization of papilla
Papillary necrosis
Infarct
Hemosiderin deposits
Chronic nephropathy
0
9
0
0
0
45
0
9
0
0
0
45
0
14
0
0
0
43
10a
23C
?d
0
0
34
21
2
0
1
4
33
2
6
0
0
8
35
26
3
0
0
22a
30
18
12a
23a
8C
25a
26
""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).
2
3 In summary, the chronic toxicity and carcinogenicity study of male and female F344 rats
4 administered biphenyl in the diet for 2 years (Umeda et al., 2002) provides evidence for
5 biphenyl-induced bladder tumors in males, but not females, based on the development of
6 transitional cell papillomas and carcinomas in the 4,500 ppm (378 mg/kg-day) males (Table 4-4).
7 This study identified a no-observed-adverse-effect level (NOAEL) of 500 ppm (42.7 mg/kg-day)
8 and a lowest-observed-adverse-effect level (LOAEL) of 1,500 ppm (128 mg/kg-day) for
9 nonneoplastic kidney lesions (simple transitional cell hyperplasia in the renal pelvis and
10 hemosiderin deposits) in female F344 rats exposed to biphenyl in the diet for 2 years. The
11 chronic toxicity of biphenyl was assessed in Wistar rats (50/sex/group) administered the
27
DRAFT - DO NOT CITE OR QUOTE
-------
1 chemical at 0, 0.25, or 0.5% (0, 2,500, or 5,000 ppm) in the diet for up to 75 weeks (Shiraiwa et
2 al.,1989). The rats were observed daily for clinical signs. Body weight and food consumption
3 were measured weekly. At death or scheduled sacrifice, gross pathologic examinations were
4 performed and all organs were removed and preserved. Other than body weight and compound
5 consumption data, the published results of this study were limited to kidney weight data and
6 urolithiasis findings. Based on reported values for mean daily biphenyl intake (mg biphenyl/rat)
7 and mean initial and final body weights for each study group, doses of biphenyl at the 0.25 and
8 0.5% dietary levels are estimated to have been 165 and 353 mg/kg-day for males, respectively,
9 and 178 and 370 mg/kg-day for females, respectively.
10 Mean final body weights in both 2,500 and 5,000 ppm groups of biphenyl-exposed male
11 and female rats were significantly lower (approximately 15 and 25% lower; p < 0.01) than their
12 respective controls. Absolute and relative kidney weights of control and biphenyl-exposed rats
13 were similar; with the exception of significantly increased (p < 0.001) mean relative kidney
14 weight in 2,500 ppm female rats. The study authors reported the occurrence of hematuria (in
15 both the 2,500 and 5,000 ppm groups) as early as week 16 and stated that it was more
16 recognizable at 60 weeks (Shiraiwa et al., 1989). Kidney stone formation was reported in
17 6/46 and 1/43 of the 2,500 ppm males and females, respectively, and in 19/47 and 20/39 of the
18 5,000 ppm males and females, respectively. Detection of stones in other regions of the urinary
19 tract was essentially limited to the 5,000 ppm groups and included the ureter (2/47 males and
20 2/39 females) and urinary bladder (13/47 males and 6/39 females). Kidney stones were hard,
21 black, and located from the pelvic area to the medullary region. Stones in the ureter were hard,
22 black, and composed of protein. Stones in the urinary bladder were hard, yellowish-white, round
23 to oval in shape, and composed of ammonium magnesium phosphate. Histologically, kidneys
24 with stones exhibited obstructive pyelonephritis accompanied by hemorrhage, lymphocytic
25 infiltration, tubular atrophy, cystic changes of tubules, and fibrosis. Urinary bladders with stones
26 exhibited simple or diffuse hyperplasia and papillomatosis of the mucosa; however, neoplastic
27 lesions were not seen. No control rats (44 males and 43 females) showed stones in the kidney,
28 ureter, or urinary bladder. The lowest exposure level in this study, 2,500 ppm in the diet for
29 75 weeks, was a LOAEL for formation of kidney stones associated with pyelonephritis in Wistar
30 rats (dose levels of 165 and 178 mg/kg-day for males and females, respectively).
31 Shiraiwa et al. (1989) also reported the results of an initiation-promotion study in male
32 Wistar rats (25/group) that included three groups administered a basal diet for 2 weeks followed
33 by diets containing 0, 0.125, or 0.5% biphenyl (0, 1,250, or 5,000 ppm) for 34 weeks. Three
34 other groups received diets containing 0.1% N-ethyl-N-hydroxyethylnitrosamine (EHEN, an
35 initiator of kidney tumors in rats) for 2 weeks followed by diets containing 0, 0.125, or 0.5%
36 biphenyl (0, 1,250, or 5,000 ppm) for 34 weeks. Initial and final body weights were recorded.
37 At terminal sacrifice, gross pathologic examinations were performed. The study report included
38 information regarding kidney weights, but did not indicate whether weights of other organs were
28 DRAFT - DO NOT CITE OR QUOTE
-------
1 measured. Kidney and urinary bladder were fixed; kidneys were sectioned transversely (10-
2 12 serial slices) and urinary bladders were cut into 4-6 serial slices. The authors used a
3 computer-linked image analyzer to determine the incidence of kidney lesions and dysplastic foci.
4 The presence of stones in the kidney and urinary bladder was assessed qualitatively using an
5 infrared spectrophotometer. Based on reported values for mean daily biphenyl intake (mg
6 biphenyl/rat) and average body weight (mean initial body weight + one-half the difference
7 between mean initial and mean final body weight) for each study group, doses of biphenyl at the
8 0.125 and 0.5% dietary levels are estimated to have been 59.3 and 248.3 mg/kg-day,
9 respectively, for rats on basal diet alone for the first 2 weeks and 62.0 and 248.2 mg/kg-day,
10 respectively, for rats receiving EHEN in the diet for the first 2 weeks.
11 The mean final body weight of the rats receiving basal diet followed by diet containing
12 0.5% biphenyl was significantly lower (p < 0.001) than that of controls (0.389 ± 22 versus 0.432
13 ± 30 kg). It was stated that relative kidney weights were increased in this group of biphenyl-
14 exposed rats compared to the basal diet control group, but the actual data were not presented.
15 Stones were detected only in the rats receiving 0.5% biphenyl in the diet; incidences were 4/25
16 (kidney), 1/25 (ureter), and 3/25 (urinary bladder) in rats that had received that basal diet for the
17 first 2 weeks. Similar results regarding final body weight and the detection of stones in the
18 urinary tract were reported for the rats that had received EHEN in the diet prior to the
19 administration of biphenyl. Incidences of dysplastic foci and renal cell tumors were determined
20 in the kidneys of all groups of rats. Only rats that had received EHEN during the initial 2 weeks
21 exhibited neoplastic kidney lesions (dysplastic foci, renal cell tumors). For the EHEN + 0%
22 biphenyl, EHEN + 0.125% biphenyl, and EHEN + 0.5% biphenyl groups, incidences of rats with
23 dysplastic foci were 25/25, 21/25, and 25/25, respectively, and incidences of rats with renal cell
24 tumors were 13/25, 12/25, and 7/25, respectively. Under the conditions of this study, biphenyl
25 did not exhibit tumor promoting characteristics for the kidney tumor initiator, EHEN.
26 Weanling albino rats (15/sex/group) were administered biphenyl in the diet at
27 concentrations of 0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, or 1% for 2 years (0, 10, 50, 100, 500,
28 1,000, 5,000, or 10,000 ppm) (Ambrose et al., 1960). Body weights were monitored every week
29 during the period of active growth and then at 50-day intervals. Hemoglobin was monitored
30 every 100 days in control and high-dose rats; at 500, 600, and 700 days in rats receiving 0.5%
31 biphenyl, and at 500 and 600 days in rats receiving 0.1% dietary biphenyl. A 98-day paired-
32 feeding experiment was conducted in which control rats were provided the same amount of food
33 that rats of the 0.5 and 1.0% dietary biphenyl groups consumed to assess whether possible
34 differences in growth would indicate a biphenyl exposure-related toxicological response or
35 decreased palatability. At necropsy, the weights of liver, kidneys, heart, and testes were
36 determined for all groups except those receiving 1.0% biphenyl in the diet. Stained sections of
37 heart, lung, liver, kidney, adrenal, spleen, pancreas, stomach, intestine, bladder, thyroid, brain,
29 DRAFT - DO NOT CITE OR QUOTE
-------
1 pituitary, and gonads were prepared for histopathologic examinations. In some cases, bone
2 marrow smears were prepared.
3 The study report of Ambrose et al. (1960) did not include sufficient information from
4 which daily biphenyl doses could be calculated. Biphenyl doses are estimated at 1, 4, 8, 42, 84,
5 420, 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%,
6 respectively, based on U.S. EPA (1988) reference values for body weight and food consumption
7 in F344 rats (averages of values for males and females). There is greater uncertainty in the dose
8 estimates at the two highest exposure levels because the magnitude of reported decreased food
9 consumption in these groups was not specified in the study report. Decreased longevity was
10 apparent in male and female rats of the 0.5 and 1.0% biphenyl exposure groups, but was not
11 evident at lower exposure levels. Growth rates appeared similar among controls and groups
12 exposed to biphenyl levels <0.1%. At the two highest exposure levels, markedly decreased
13 growth was evident, but was attributable to decreased food consumption and indicative of
14 decreased palatability based on results of the paired-feeding experiment. Decreased hemoglobin
15 levels were reported in male and female rats of the two highest exposure levels after 300-
16 400 and 500-600 days, respectively, but were considered at least partially related to lower food
17 consumption in these groups relative to controls. Selected organ weights are summarized in
18 Table 4-6. There were no statistically significant treatment-related effects on organ weights at
19 dietary levels <0.1%, which were below those associated with decreases in food consumption,
20 body weight, and survival (i.e., 0.5 and 1.0%). Relative liver and kidney weights of female rats
21 of the 0.5% biphenyl exposure group were significantly (p < 0.05) increased, approximately
22 45 and 215% higher than those of respective controls. The only significant compound-related
23 histopathological change occurred in the kidneys, which, in all members of the two highest
24 exposure groups, showed irregular scarring, lymphocytic infiltration, tubular atrophy, and tubular
25 dilation associated with cyst formation. Some evidence of hemorrhage was present, and calculi
26 were frequently noted in the renal pelvis. Evidence of metaplasia in the epithelium of the renal
27 pelvis did not implicate neoplastic activity, and, taking the histopathological results as a whole,
28 there appeared to be no clear-cut, compound-related tumor development. However, the small
29 number of animals in each group and the decreased survival in the two highest dose groups may
30 have impaired the ability to detect late-developing tumors. The study identified 1,000 ppm
31 biphenyl in the diet (84 mg/kg-day) as a NOAEL and 5,000 ppm (420 mg/kg-day) as the LOAEL
32 for kidney effects including tubular atrophy and dilation associated with cyst formation and
33 calculi formation in the renal pelvis of albino rats of both sexes.
34
30 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. (I960).
1
2 Male albino rats (8/group; strain not stated) were given biphenyl in the diet for up to
3 13 months at concentrations resulting in estimated doses of 250 or 450 mg/kg-day (Pecchiai and
4 Saffiotti, 1957). Upon sacrifice, liver, kidney, spleen, heart, lung, thyroid, parathyroid, adrenal,
5 pancreas, testis, stomach, and intestine were processed for histopathological examination. At 2-
6 month interim sacrifices, moderate degenerative changes in liver and kidney were observed at
7 both dose levels. Liver effects consisted of moderate degeneration and hypertrophy of the
8 Kupffer cells with a generally well-preserved structure. Renal glomeruli were undamaged, but
9 tubuli showed mild signs of degeneration. The liver and kidney effects did not appear to
10 increase in severity in rats treated for up to 13 months. Other histopathologic effects noted in the
11 biphenyl-treated rats included hypertrophied splenic reticular cells, small follicles with sparse
12 colloid and desquamation of follicular epithelium in the thyroid, and hyperplastic and
13 hyperkeratinized forestomach epithelium with occasional desquamation. Although the study
14 report did not include tumor incidence data for the two dose groups, the study authors reported
15 neoplastic lesions in the forestomach of three biphenyl-treated rats. Two of the rats exhibited
16 papillomas of the forestomach epithelium (one after 7 weeks and one after 7 months of
17 treatment); a squamous cell carcinoma was diagnosed in the other rat after 1 year of treatment.
31
DRAFT - DO NOT CITE OR QUOTE
-------
1 The study authors noted two sequential responses to chronic biphenyl exposure: degenerative
2 changes of nuclei and cytoplasm in the parenchyma of liver and kidney, spleen, thyroid, and
3 adrenals within 2 months followed within 1 month or more by functional-regenerative changes
4 that resulted in hyperplasia and nuclear hypertrophy of liver and kidney parenchyma as well as
5 functional hyperactivity of the thyroid and parathyroid. Signs of cirrhosis were not seen, but
6 irritation and hyperplasia were evident in the lower urinary tract. The lowest dose, 250 mg/kg-
7 day biphenyl, was an apparent LOAEL for nonneoplastic degenerative changes in the liver,
8 kidney, thyroid, and parathyroid of male albino rats resulting in hyperplasia of liver, kidney, and
9 thyroid.
10 Sprague-Dawley rats (12/sex/group) were exposed to biphenyl in the diet for 2 years at
11 exposure levels of 0, 0.01, 0.1, or 1% (0, 100, 1,000, or 10,000 ppm) (Dow Chemical Co., 1953).
12 Body weights were monitored twice weekly for 3 months, then weekly. Blood samples were
13 taken from all animals at the start of the experiment, approximately every 3 months thereafter,
14 and at term. Hemoglobin levels, red and white blood cell counts and differential cell counts, and
15 BUN concentrations were recorded. At death or scheduled necropsy, organ weights were
16 recorded for liver, lung, kidneys, heart, and spleen. Sections from heart, liver, kidney, spleen,
17 adrenals, pancreas, gonads, stomach, small and large intestine, voluntary muscle, lung, bladder,
18 and brain were fixed and stained for histopathologic examination.
19 Based on U.S. EPA (1988) chronic reference values for body weight and food
20 consumption in Sprague-Dawley rats (average values for combined sexes), doses of biphenyl for
21 the dietary levels of 0.01, 0.1, and 1% are estimated to have been 7, 73, and 732 mg/kg-day,
22 respectively. It is unclear to what extent the data in the study were compromised by an outbreak
23 of pneumonia that affected the colony during the course of the experiment. Survival was poor in
24 control males, all of which had died by 18 months. Only two of the females receiving 0.1%
25 biphenyl in the diet survived to the end of the 21st month, and none had survived by the end of
26 the 23r month. However, the authors considered the decreased survival in this group of females
27 to have been compound-related. Striking biphenyl concentration-related reductions in body
28 weight gain were observed among the groups, although, in monitoring food efficiency, the
29 authors indicated that the reduced growth was likely due to a lower daily consumption of food
30 rather than to biphenyl toxicity. There were no clear indications of exposure-related changes in
31 hematological parameters. The authors reported significant (p < 0.05) increases in average
32 (combined sexes) relative liver and kidney weights at the highest exposure level, compared with
33 control values (4.71 versus 3.05 g/100 g and 1.68 versus 1.00 g/100 g, respectively).
34 Histopathologic examinations revealed dilatation of the kidney tubules, an effect that appeared to
35 be associated with secondary inflammation, uremia, disruption of the filtration system, and an
36 increase in BUN in affected animals. Tubular dilatation was evident in controls as well as
37 treated animals, but increased in severity with dose (measured on a scale of 0-4). Among the
38 controls, low-, mid-, and high-dose rats, respective incidences for tubular dilatation with severity
32 DRAFT - DO NOT CITE OR QUOTE
-------
1 scores >2 were 1/12, 6/12, 7/12, and 11/12 for males and 1/12, 3/12, 4/12, and 11/12 for females.
2 Incidences of tubular dilatation with severity scores >3 were 0/12, 1/12, 2/12, and 9/12 for males
3 and 1/12, 2/12, 2/12, and 11/12 for females, respectively. Calcification and intratubular
4 inflammation were frequently observed at the highest biphenyl exposure level. The incidence
5 data for renal tubular dilatation with a severity score >3 (considered adverse) indicate that
6 100 ppm biphenyl in the diet (73 mg/kg-day) was a NOAEL and that 1,000 ppm (732 mg/kg-
7 day) was a LOAEL for renal effects in Sprague-Dawley rats. The small number of rats in the
8 exposure groups and the decreased survival at the highest exposure level may have impaired the
9 ability to detect late-developing tumors in this study.
10
11 4.2.1.2.2. Chronic mouse studies
12 In a chronic toxicity and carcinogenicity study of BDFi mice (50/sex/group), biphenyl
13 was administered in the diet for 2 years at concentrations of 0, 667, 2,000 or 6,000 ppm (Umeda
14 etal., 2005). All animals were observed daily for clinical signs and mortality. Body weights and
15 food consumption were recorded weekly for the first 14 weeks and every 4 weeks thereafter.
16 Hematological and clinical chemistry parameters were measured in blood samples drawn from
17 all 2-year survivors just prior to terminal sacrifice. At death or terminal sacrifice, gross
18 pathological examinations were performed and organs were removed and weighed. Specific
19 tissues prepared for microscopic examination were not listed in the study report, but included
20 liver and kidney.
21 There were no overt clinical signs or effects on food consumption or survival among
22 biphenyl-exposed mice of either sex compared to respective controls. However, mean terminal
23 body weights showed a dose-related decrease; body weights were significantly less than those of
24 respective controls at 2,000 and 6,000 ppm (males: 46.9, 43.1, 42.9, and 32.4 kg; females: 34.0,
25 32.5, 30.5, and 25.5 kg, at 0, 667, 2,000, and 6,000 ppm, respectively). Based on body weight
26 and food consumption data, the study authors estimated that the 667, 2,000, and 6,000 ppm
27 dietary levels resulted in average daily biphenyl doses of 97, 291, and 1,050 mg/kg-day in the
28 males and 134, 414, and 1,420 mg/kg-day in the females.
29 Although there were no compound-related changes in hematological parameters, some
30 clinical chemistry parameters showed marked changes in relation to dose, including a biphenyl
31 dose-related increase in BUN that achieved statistical significance in 6,000 ppm males and
32 females and 2,000 ppm males. In the female mice, dose-related increases in activities of the
33 plasma enzymes AP, lactate dehydrogenase (LDH), glutamate oxaloacetate transaminase (GOT;
34 also referred to as AST), and glutamate pyruvate transaminase (GPT; also referred to as ALT)
35 (see Table 4-7) suggested effects of biphenyl on the liver. Umeda et al. (2005) noted that
36 females with malignant liver tumors exhibited extremely high AST, ALT, and LDH activities.
37 In general, biphenyl did not induce dose-related changes in liver enzymes in male mice, although
38 AP activity was significantly greater than controls in 6,000 ppm males (Table 4-7).
33 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-7. Dose-related changes in selected clinical chemistry values from
male and female BDFi mice exposed to biphenyl via the diet for 2 years
Males
Biphenyl dietary
concentration (ppm)
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
"Significantly different from controls (p < 0.01) according to Dunnett's test.
bSignificantly different from controls (p < 0.05) according to Dunnett's test.
Source: Umeda et al. (2005).
2
3 The only apparent exposure-related effect on organ weights was 1.3-, 1.4-, and 1.6-fold
4 increases in relative liver weights of 667, 2,000, and 6,000 ppm female mice, respectively (the
5 data for liver weight group means and standard deviations [SDs] were not presented in Umeda et
6 al. [2005]) Gross pathologic examinations revealed biphenyl dose-related increased incidences
7 of liver nodules in females, but not males (Table 4-8). The nodules were round- or oval-shaped
8 cystic or solid masses approximately 3-23 mm in diameter of the largest axis. Histopathological
9 examinations revealed that 5, 16, and 19 of the nodule-bearing 667, 2,000, and 6,000 ppm female
10 mice also exhibited proliferative lesions of hepatocellular origin (Table 4-8). Significantly
11 increased incidences of basophilic cell foci were observed in 2,000 and 6,000 ppm female mice.
12 Although incidences of basophilic cell foci were significantly increased in 667 ppm male mice as
13 well, a dose-related effect was not evident because incidences of this lesion were not
14 significantly increased in 2,000 or 6,000 ppm males compared to controls. Incidences of
15 hepatocellular adenomas and combined incidences of hepatocellular adenomas or carcinomas
16 were significantly increased in the 2,000 and 6,000 ppm females and Peto's trend tests confirmed
34
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
4
5
6
7
8
9
10
11
12
significant positive trends for dose-related increased incidences of hepatocellular adenomas (p <
0.05) and combined incidences of hepatocellular adenomas or carcinomas (p < 0.01). Incidences
of hepatocellular carcinomas were significantly increased in 2,000 ppm females, but not 667 or
6,000 ppm females. However, Umeda et al. (2005) noted that the incidences of hepatocellular
carcinomas (5/50 or 10%) in each of the 667 and 6,000 ppm groups of females exceeded the
range of historical control data for that laboratory (26 hepatocellular carcinomas in 1,048 female
mice [2.5% incidence in 21 bioassays, with a maximum incidence of 8%]). 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 4-8. Incidences of gross and histopathological findings in male and
female BDFi mice fed diets containing biphenyl for 2 years
Parameter
Dietary concentration of biphenyl (ppm)
Males
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
Liver0
Adenoma
Carcinoma
Adenoma or carcinoma
(combined)
Basophilic cell foci
Clear cell foci
Eosinophilic cell foci
8/50
8/50
16/50
0/50
0/50
0/50
6/49
8/49
12/49
6/49b
6/49b
0/49
7/50
5/50
9/50
1/50
2/50
0/50
3/50
4/50
7/50
2/50
0/50
0/50
2/50
1/50
3/50
1/50
2/50
0/50
3/50
5/50
8/50
1/50
1/50
1/50
12/503
7/50a
16/50b
12/50b
3/50
0/50
10/493
5/49
14/493
6/49a
2/49
0/49
Kidney
Desquamation: pelvis
Mineralization inner stripe-
outer medulla
0/50
9/50
0/49
8/49
0/50
14/50
10/50b
14/50
4/50
3/50
0/50
5/50
0/50
12/503
15/49b
26/49b
13
14
15
""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.
Historical control data for hepatocellular tumors: Male BDFi mouse: adenoma—17.2% (4-34%), carcinoma—
18.8% (2-42%), adenoma/carcinoma—32.2% (10-68%). Female BDFj mouse: adenoma-^.8% (0-10%),
carcinoma—2.3% (0-8%), adenoma/carcinoma—7.1% (2-14%). Source: email dated July 25, 2011, from Umi
Umeda, JBRC, to Connie Kang, NCEA, ORD, U.S. EPA.
Source: Umeda et al. (2005).
In summary, the chronic toxicity and carcinogenicity study of male and female BDFi
mice administered biphenyl in the diet for 2 years (Umeda et al., 2005) provides evidence for
35
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 4-8). 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 Groups of female ddY mice were fed diets containing 0 (n = 37 mice) or 0.5%
8 (n = 34 mice) biphenyl (5,000 ppm) in the diet for 2 years (Imai et al., 1983). This study also
9 included groups exposed to dietary concentrations of 0.2% thiabendazole or a mixture of 0.25%
10 biphenyl and 0.1% thiabendazole (results from this part of the study are not further described
11 herein). Food consumption, body weights, and survival were assessed at intervals throughout
12 exposure. At terminal sacrifice, several organs were weighed. The following organs were
13 examined for histopathological changes: brain, pituitary, thymus, liver, spleen, pancreas, lung,
14 heart, adrenal, kidney, ovaries, uterus, thyroid, stomach, small intestine, and large intestine.
15 Urine and blood samples were collected from mice (6-12/group) at terminal sacrifice and were
16 analyzed for urinalysis, hematological, and serum chemistry endpoints. Based on U.S. EPA
17 (1988) methodology for estimating food consumption rates from body weight data and the
18 reported average terminal body weight for the 5,000 ppm mice (0.037 kg), an oral dose of 855
19 mg/kg-day is estimated from the dietary exposure.
20 Exposure to biphenyl did not influence survival, food consumption, or growth compared
21 with controls. No marked exposure-related effects were found on terminal organ and body
22 weights or on the urinalytic, hematologic, or serum chemistry endpoints. Histological
23 examination revealed no increased incidence of non-neoplastic lesions in examined tissues in the
24 5,000 ppm biphenyl group, compared with the control group. The only tissues showing tumors
25 at elevated incidence in the 5,000 ppm mice, compared with the control group, were the lung
26 (11/34 [32.4%] versus 9/37 [24.3%] in controls) and lymphatic tissues (lymphomas: 5/34
27 [14.7%] versus 4/37 [10.8%]; leukemia: 3/34 [8.8%] versus 2/37 [5.4%]), but these increases
28 were not statistically significant (p > 0.05 by the Fisher's exact test). In summary, 5,000 ppm
29 biphenyl in the diet of female ddY mice for 2 years was a NOAEL for non-neoplastic lesions,
30 survival, body and organ weight changes, and changes in urinalytic, hematologic, and serum
31 chemistry endpoints. No carcinogenic response occurred in female ddY mice exposed to
32 5,000 ppm biphenyl in the diet (estimated dose of 855 mg/kg-day) for 2 years (Imai et al., 1983).
33 The carcinogenic potentials of 130 chemicals, including biphenyl, were assessed in a
34 protocol that exposed groups of two strains of Fl hybrid mice (18/sex/strain/group), produced by
35 mating female C57BL/6 mice to either male C3H/Anf mice (Fl designated as strain A) or male
36 AKR mice (Fl designated as strain B) to individual chemicals by the oral route for 18 months
37 (Innes et al., 1969; NCI, 1968). Four groups of untreated controls and a group of gelatin vehicle
38 controls (18/sex/strain/group) were included in the study. In the case of biphenyl, the chemical
36 DRAFT - DO NOT CITE OR QUOTE
-------
1 was administered via gavage to mice for 3 weeks, starting at the age of 7 days at 215 mg
2 biphenyl/kg body weight in 0.5% gelatin (the report of Innes et al. (1969) appears to have
3 erroneously reported the gavage dose as 2.5 mg/kg). Thereafter, and for the rest of the
4 experimental period, biphenyl was mixed with chow to a final concentration of 517 ppm. The
5 gavage dose level and food concentration of biphenyl were selected to reflect the maximum
6 tolerated dose identified in preliminary range-finding, single-dose subcutaneous injection and
7 single- and repeated-dose oral administration studies. Initial gavage dose and dietary levels of
8 biphenyl were not adjusted for weight gain during the 18-month study. Based on U.S. EPA
9 (1988) chronic reference values for body weight and food consumption in strain A mice (average
10 values for combined sexes), an average oral dose of 91 mg/kg-day is estimated from the dietary
11 exposure. Blood smears were prepared from mice that showed splenomegaly, liver enlargement,
12 or lymph adenopathy at necropsy. At term, mice were examined for any gross pathological
13 features. Major organs were processed for histopathologic examination (including total chest
14 contents, liver, spleen, kidneys with adrenals, stomach, and genital organs). Innes et al. (1969)
15 reported incidences for hepatomas, pulmonary tumors, and lymphomas in control mice and for
16 tested chemicals that were judged to give "high tumor yield"; biphenyl was reported to be
17 noncarcinogenic, but tumor incidence data for biphenyl were not reported. The NCI (1968)
18 report included tabulated incidences of hepatomas, pulmonary tumors, and lymphomas in control
19 mice and biphenyl-treated mice, which are summarized in Table 4-9. In summary, the results
20 provide no evidence of a carcinogenic response to 18 months of oral exposure to biphenyl in
21 mice (215 mg/kg by gavage for 3 weeks, followed by dietary exposure to 517 ppm biphenyl).
22
37 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-9. Incidences of selected tumor types among controls and mice
administered biphenyl orally for 18 months
Group
Incidences of selected tumor types"
Hepatoma
Pulmonary tumors
Reticular cell sarcoma
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 Mongrel dogs (two males and one female/group) were administered 0, 2.5, or 25 mg/kg
4 biphenyl in corn oil by capsule 5 days/week for 1 year (Monsanto, 1946). Dogs were examined
5 daily for clinical signs and weighed weekly. Blood samples were drawn at 3-month intervals to
6 measure hematological and clinical chemistry parameters. Urine samples were obtained at
7 similar intervals to measure specific gravity, sugar, protein, bile pigments, occult blood, and
8 microscopic sediment. Samples of urine from the high-dose dogs were collected during week
9 18, pooled, and analyzed for the presence of biphenyl and metabolites. At termination, gross
10 necropsies were performed, and sections of large and small intestine, pancreas, ovary or testis,
11 adrenal, urinary bladder, stomach, lung, thyroid, brain, heart, spleen, and liver were prepared for
12 histopathologic examination. Although slight fluctuations were seen in body weight during the
13 study, the dogs generally exhibited a net weight gain. Fluctuations in hematological parameters
14 and urine analysis were inconsistent and not considered compound-related. Gross pathological
15 examination of the dogs showed no obviously compound-related effects. Histopathologic
16 examinations revealed lung congestion consistent with bronchial pneumonia in one high-dose
17 dog; histopathology was unremarkable for each of the other dogs in the study.
18 Dow Chemical Co. (1953) described a biphenyl feeding experiment in which four groups
19 of Rhesus monkeys (two males and one female/group) were exposed to 0, 0.01, 0.1, or 1%
20 biphenyl in chow for 1 year, during which time most of the animals experienced ill health not
21 related to biphenyl exposure. Despite this caveat, hematological parameters were normal. The
38
DRAFT - DO NOT CITE OR QUOTE
-------
1 authors considered an increase in relative liver weight in high-dose monkeys (4.65 g/100 g body
2 weight versus 3.90 g/100 g body weight in controls) to possibly be compound-related.
o
J
4 4.2.2. Inhalation Studies
5 In three separate experiments, albino rabbits (sex and strain not stated), Sprague-Dawley
6 rats (sex not stated), and mice (sex and strain not stated) were repeatedly exposed to dusts
7 composed of 50% biphenyl attached to celite for 7 hours/day, 5 days/week (Deichmann et al.,
8 1947; Monsanto, 1946). In the first experiment, 3 rabbits and 10 rats were exposed to an average
9 concentration of 300 mg/m3 on each of 64 days over a period of 94 days. The rats exhibited
10 irritation of the nasal mucosa accompanied by serosanguineous discharge. Five of the rats died
11 prior to term, and the survivors lost weight. The rabbits exhibited no exposure-related adverse
12 signs. In the second experiment, three rabbits and six rats were exposed to an average
13 concentration of 40 mg/m3 on each of 46 days over a total period of 68 days. One rat died prior
14 to term. The surviving rats showed signs of mucous membrane irritation, but appeared to gain
15 weight at a normal rate. The rabbits exhibited no exposure-related adverse signs. In the third
16 experiment, 12 mice and 4 rats were exposed to an average concentration of 5 mg/m3 on each of
17 62 days over a total period of 92 days. While the rats were unaffected at this concentration, all
18 of the mice showed signs of irritation of the upper respiratory tract and two died prior to term.
19 Bronchopulmonary lesions (including acute emphysema, congestion, edema, bronchitis,
20 widespread lobular pneumonia, and multiple pulmonary abscesses) were reported in rats from
21 experiments 1 and 2 and in mice of experiment 3. Some unspecified minor liver and kidney
22 lesions were also noted. Based on the results of these three experiments, a LOAEL of 5 mg/m3
23 in mice and a LOAEL of 40 mg/m3 in rats for upper respiratory tract irritation was identified.
24 Groups of CD-I mice (50/sex/group) were exposed to airborne biphenyl at vapor
25 concentrations of 0, 25, or 50 ppm (0, 157.7, and 315.3 mg/m3, respectively) for 7 hours/day,
26 5 days/week for 13 weeks (Sun Company Inc., 1977a). Mice were maintained and exposed to
27 biphenyl in groups of 5 (for a total of 10 groups/sex/exposure group). All animals were checked
28 daily for clinical signs and mortality, and body weight data were collected. Upon completion of
29 the 13-week exposure period, surviving mice were placed in metabolic cages for 12-hour
30 collection of urine for urinalysis. Blood samples were collected for blood chemistry and
31 hematology assessments. Gross and histopathologic examinations were performed on all mice.
32 Ten surviving mice/sex/group were held for a 30-day recovery period prior to terminal sacrifice.
33 During the first few days of biphenyl exposure, some of the test material crystallized in
34 the delivery system; analysis of biphenyl exposure levels was not performed on these days.
35 Daily measured biphenyl exposure concentrations were highly variable during the first half of
36 the 13-week exposure period, whereas subsequently measured concentrations were closer to
37 target concentrations. For example, during the first 45 exposure sessions, measured daily
38 biphenyl concentrations in the 50 ppm target groups ranged from as low as 5 ppm to as high as
39 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
102 ppm and subsequent measurements ranged from 48 to 55 ppm. Mean biphenyl
concentrations (± 1 SD) calculated for the entire 13 weeks of exposure were 25 ± 7 and
50 ± 16 ppm for the 25 and 50 ppm target groups, respectively. The authors reported the loss of
46 mice (40 males and 1 female at 25 ppm and 5 males at 50 ppm) due to overheating and
cannibalization. Since the overheating event occurred after 46 exposures, the overall study
duration ran for 117 days to ensure that replacement mice received a total of 65 exposures as
called for in the protocol. The study report did not mention results of clinical observations, and
mortality data were not specifically summarized. There were no clear indications of exposure-
related effects on body weights. Results of urinalysis, hematology, and clinical chemistry did
not indicate any clear exposure-related changes that could be attributed to biphenyl toxicity.
Gross and histopathological examinations revealed congested and hemorrhagic lungs,
hyperplasia of the trachea with inflammation accompanied by a high incidence of pneumonia,
and congestion and edema in liver and kidney of biphenyl-exposed mice (Table 4-10). The
pathologist considered the congestion in the lung, liver, and kidney a likely effect of the
anesthetic used for killing the mice, although control mice did not exhibit these effects at 13-
week sacrifice. The hemorrhagic lungs and tracheal hyperplasia were considered effects of
biphenyl exposure. Results from the 30-day recovery groups suggest that the biphenyl exposure-
related pulmonary effects were reversible. This study identified a LOAEL of 25 ppm for
histopathologic lung, liver, and kidney lesions in male and female CD-I mice exposed to
biphenyl by inhalation for 7 hours/day, 5 days/week for 13 weeks.
Table 4-10. 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
Tracheal 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
22
23
24
25
26
aThe study report presented incidences of histopathologic lesions for combined male and female mice only; no
statistical analyses were conducted.
Source: Sun Company Inc. Q977a).
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
4.3.1. Oral Exposure
Pregnant female Wistar rats (18-20 group) were gavaged with 0, 125, 250, 500, or
1,000 mg/kg-day biphenyl in corn oil on gestation days (GDs) 6-15 (Khera et al., 1979). Body
40
DRAFT - DO NOT CITE OR QUOTE
-------
1 weights of dams were recorded on GDs 1, 6-15, and 22, at which point all dams were sacrificed.
2 Parameters evaluated at autopsy included the number of corpora lutea, fetal weights and
3 viability, and early resorptions. Two-thirds of the live fetuses/litter were examined for skeletal
4 development and the rest were examined for the presence of visceral abnormalities.
5 Five of the 20 high-dose dams died prior to sacrifice. Doses <500 mg/kg-day produced
6 no clinical signs of maternal toxicity or evidence of treatment-related effects on maternal weight
7 gain. As shown in Table 4-11, a significantly increased number of dams without live fetuses was
8 observed in the high-dose group, compared with controls. Mean numbers of corpora lutea and
9 live fetuses in the high-dose dams were similar to those of controls and dams of all other dose
10 levels. However, the percent of dead fetuses and resorption sites was higher in the high-dose
11 group, and the numbers of anomalous fetuses and litters bearing anomalous fetuses appeared to
12 increase with increasing dose. The increases in the number of fetuses with anomalies, such as
13 missing and unossified sternebrae or delayed calvarial ossification, were not statistically
14 significant, but, as shown in Table 4-11, the incidence of litters with any type of fetal anomalies
15 ("anomalous litters/number examined") was statistically significantly elevated (p < 0.05 by
16 Fisher's exact test) at >500 mg/kg-day compared with control incidences. This study identified a
17 NOAEL of 500 mg/kg-day and a LOAEL of 1,000 mg/kg-day for frank maternal toxicity
18 (increased mortality and decreased dams with live fetuses) and lethal fetal effects. For less
19 severe developmentally toxic effects (increased incidence of anomalous litters), 500 mg/kg-day
20 was a LOAEL and 250 mg/kg-day was a NOAEL.
21
41 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-11. 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
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
125
0/20
12.9 ±0.4
11.8 ±0.6
3.3
5.3 ±0.1
22/236
11/20
250
1/19
13.7 ±0.5
11.9±0.6
6.1
5.2 ±0.1
22/213
13/18
500
2/20
13.3 ±0.4
11.2 ±0.5
7.8
5.2 ±0.1
35/199c
15/18C
1,000
11/203
12.5 ±0.7
10.7 ±1.3
13.7b
4.5 ±0.3
25/107c
6/9
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
o
J
9
1
4
0
1
7
12
1
3
2
1
9
9
2
4
0
1
8
15
1
16
0
0
5
6
0
17
8
0
"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).
1
2 Dow Chemical Co. (1953) reported the results of a multigenerational study in which
3 groups of 4-month-old male and female Long Evans rats (three males and nine females/group)
4 were fed diets containing 0, 0.01, 0.1, or 1.0% biphenyl. Based on U.S. EPA (1988) subchronic
5 reference values for body weight and food consumption in male and female Long Evans rats,
6 doses of biphenyl for the dietary levels of 0.01, 0.1, and 1.0% are estimated to have been 9, 89,
7 and 887 mg/kg-day, respectively, for the males and 10, 101, and 1,006 mg/kg-day, respectively,
8 for the females. Average cross-gender doses for males and females were 10, 95, and 947 mg/kg-
9 day. For breeding, three females were placed together with one male. Following the breeding
10 phase, females were separated and number of litters cast, number of days between mating and
11 delivery, and average number of pups/litter at delivery were recorded. Fl pups were weighed
12 and culled to seven/litter at 2 days of age and weaned at 3 weeks of age, and weights were
13 recorded weekly for postnatal weeks 3-6. The Fl rats were continued on the same diets as their
14 parents, and, at 10 weeks of age, nine Fl females and three Fl males were mated to produce an
15 F2 generation of pups. F2 pups were selected (by the same procedure) for mating and
42
DRAFT - DO NOT CITE OR QUOTE
-------
1 production of an F3 generation that were sacrificed at 3 weeks of age; 12 F3 pups from each diet
2 group were subjected to gross pathologic examinations.
3 There were no significant differences between controls and 0.01 and 0.1% biphenyl-fed
4 groups regarding litters cast, gestation length, or average number or weight of pups/litter at birth
5 or at 3 or 6 weeks of age. Decreased fertility in the 1% biphenyl-fed group of females was
6 observed (6/9, 7/9, and 8/9 confirmed pregnancies for the three successive generations of 1.0%
7 biphenyl-fed groups versus 8/9, 9/9, and 8/9 confirmed pregnancies for controls). Averaged for
8 Fl, F2, and F3 pups combined, the 1.0% biphenyl-fed group exhibited significantly (p < 0.05)
9 decreased number of pups/litter at birth (6.2/litter versus 8.6/litter for controls) and lower
10 average body weight at 3 weeks of age (36 versus 48 g for controls) and 6 weeks of age
11 (78 versus 113 g for controls). Gross pathologic evaluations of F3 weanlings revealed no signs
12 of biphenyl treatment-related effects. There was no evidence of a cumulative effect over the
13 three generations. The study authors indicated that the decreased fertility, smaller litter size, and
14 reduced rate of growth in the 1.0% biphenyl-fed group may have been associated with
15 unpalatability and resultant decreased food intake.
16 The research report of Ambrose et al. (1960) contains a subsection in which the
17 reproductive toxicity of biphenyl was examined in two experimental series. In the first
18 experiment, weanling albino rats were administered 0 or 0.1% biphenyl (5 males and
19 10 females/group) or 0.5% biphenyl (3 males and 9 females) in the diet for 60 days prior to
20 mating. In the second experiment, groups of 90-day-old albino rats were administered 0 or 0.1%
21 biphenyl (4 males and 8 females/group) or 0.5% biphenyl (3 males and 9 females) in the diet for
22 11 days prior to mating. Based on U.S. EPA (1988) subchronic reference values for body weight
23 and food consumption in rats of unspecified strain (average values for combined sexes), doses of
24 biphenyl for the dietary levels of 0.1 and 0.5% are estimated to have been 105 and 525 mg/kg-
25 day, respectively. All rats were maintained on their respective diets throughout mating and until
26 the progeny of all litters were weaned. Insufficient information is provided in the report to
27 permit a judgment as to whether dietary exposure to biphenyl was associated with reproductive
28 deficits. However, the authors presented tabular data for number of rats casting litters, total
29 born, and range of litter size (Table 4-12) and concluded that the compound had no significant
30 effect on reproduction.
31
43 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-12. 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.
C0.1% = 105 mg/kg and 0.5% = 525 mg/kg/day
Source: Ambrose et al. (I960).
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 (Pecchiai and Saffiottl 1957: Union Carbide. 1949:
10 Deichmann et al., 1947: Monsanto, 1946) and an LDso 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 Groups of mice (10/sex of unspecified strain) were exposed to biphenyl by inhalation for
18 4 hours at average analytical concentrations of 14.11, 38.40, or 42.80 ppm (89.0, 242.2, and
19 270.0 mg/m3, respectively) and observed for up to 14 days following exposure (Sun Company
20 Inc., 1977a, b). Clinical signs of hyperactivity and mild respiratory discomfort were noted
21 during exposure, but resolved during postexposure observation. One male mouse of the
22 42.80 ppm group died after 2 hours of exposure, but this death was not attributed to biphenyl
23 exposure. All other mice survived throughout the 14-day postexposure observation period.
24 Slight lung congestion was noted in most mice upon gross pathological examination.
44
DRAFT - DO NOT CITE OR QUOTE
-------
1 Sun Company Inc. (1977b) also provided details of a study in which groups of mice
2 (10/sex of unspecified strain) were exposed to biphenyl for 7 hours/day, 5 days/week for 2 weeks
3 at average analytical concentrations of 0, 24.8, or 54.75 ppm (0, 156.4, and 345.5 mg/m3,
4 respectively). Five animals/group were sacrificed immediately after exposure; the remaining
5 animals were sacrificed following a 14-day recovery period. Clinical signs were monitored
6 daily. Gross pathologic examinations at necropsy included assessment of lungs, trachea, heart,
7 spleen, liver, kidneys, stomach, and intestines. Histopathologic examinations included tissues
8 from lung, trachea, kidney, spleen, and liver. The study authors reported signs of hyperactivity
9 in some mice during the first few exposure periods. One female mouse of the 24.8 ppm
10 exposure group died prior to the third exposure session and one control female mouse died prior
11 the final exposure session. No abnormal clinical signs were seen during the 14-day recovery
12 period. Gross and histopathologic examinations revealed no signs of exposure-related adverse
13 effects.
14 Four rabbits (sex and strain unspecified) received up to 20 daily doses of 500 mg/kg
15 "purified" biphenyl to the skin; the compound was applied as a 25% preparation in olive oil.
16 Three rabbits received the same concentration of technical biphenyl (Deichmann et al., 1947;
17 Monsanto, 1946). The compound was left on the skin for 2 hours and then washed off with soap
18 and water. Some biphenyl derivatives were similarly assessed. One rabbit receiving purified
19 biphenyl died after eight applications, and the rest of the animals survived to term. The only
20 reported sublethal effect clearly associated with biphenyl exposure was that of weight loss,
21 averaging 45 and 172 g for the rabbits receiving purified and technical biphenyl, respectively.
22
23 4.4.2. Kidney/Urinary Tract Endpoint Studies
24 Endpoint-specific studies of biphenyl-induced urinary tract effects in rats (Shibata et al.,
25 1989b: Shibata et al.. 1989a: Kluwe. 1982: S0ndergaard and Blom. 1979: Booth etal.. 1961)
26 support findings of the chronic oral rat studies described in Section 4.2.1.2 (Chronic Toxicity and
27 Carcinogenicity). Detailed descriptions of these endpoint-specific studies are presented below.
28 In a preliminary study, five adult rats (sex and strain unspecified) were administered
29 biphenyl in the diet at 1% (w/w) for 26 days followed by a 29-day postexposure recovery period
30 for a total study period of 55 days (Booth et al., 1961). Total urine volume and the volume of
31 sulfosalicylic acid-precipitable sediment were recorded from urine collected from all five rats on
32 study days 4, 8, 18, 20, and 26 (exposure days), and study days 28, 32, 35, and 54 (recovery
33 period). Volumes of both urine and sulfosalicylic acid-precipitable sediment increased from 7
34 and 0.56 mL, respectively, on exposure day 4 to 32 and 2.24 mL, respectively, on exposure day
35 20. Both values remained relatively high (approximately 27 and 2.2 mL, respectively) on
36 exposure day 26 and decreased to approximately 14 and 0.8 mL, respectively, by the end of the
37 recovery period. Fractionation and analysis of the precipitate suggested the presence of p-
38 hydroxybiphenyl and its glucuronide. The study authors indicated that similar effects were noted
45 DRAFT - DO NOT CITE OR QUOTE
-------
1 in male and female rats receiving biphenyl at a level of 0.5% in the diet, but not at the 0.1%
2 dietary level.
3 A follow-up study employed 42 rats/sex/group and biphenyl dietary levels of 0, 0.1, 0.25,
4 or 0.5% (w/w). Biphenyl doses are estimated at 83.7, 209, and 419 mg/kg-day for the dietary
5 levels of 0.1, 0.25, and 0.5%, respectively, based on U.S. EPA (1988) chronic reference values
6 for body weight and food consumption in F344 rats (averages of values for males and females).
7 Rats were exposed for up to 165 days and followed for 0, 30, or 60 days of recovery. Urine
8 samples were collected periodically from five rats/sex/exposure group. Interim sacrifices of five
9 rats/sex/exposure group were performed after 30, 60, and 120 days on the diet in order to assess
10 the progression of biphenyl-induced histopathological effects on the kidney. As noted in the
11 preliminary study, the rats of the 0.5% exposure group in the follow-up study exhibited gradual
12 increases in the urine volume and sulfosalicylic acid-precipitable sediment and decreased in both
13 parameters during postexposure recovery. The study authors indicated that these effects were
14 much less pronounced in the 0.25% exposure group and absent in the 0.1% exposure group. At
15 the 0.5% exposure level, kidney lesions were noted in 1/5 of the males (several small cysts and
16 dilated tubules in the medulla and inner cortex) and 2/5 of the females (mild local tubular
17 dilation with some epithelial flattening) following 30 days of exposure. Similar, but more
18 extensive, kidney lesions were noted in 3/5 males and 5/5 females following 60 days of
19 exposure. The kidney lesions were even more prominent following 120 days of exposure.
20 Reported histopathologic findings in the kidneys of rats from the 0.25% exposure group were
21 limited to a single instance of an unspecified "prominent kidney lesion" at 60 days, and one
22 small calculus in the pelvis of one rat and a small calcareous deposit in the renal pyramid of
23 another rat following 120 days of exposure. Urinary and histopathologic renal effects were not
24 assessed at the end of the 165-day treatment period; however, during the 60-day postexposure
25 recovery period, rats of the 0.5% biphenyl exposure group exhibited a regression of kidney
26 lesions and improvement in urine quality.
27 Kluwe (1982) examined changes in urine composition and kidney morphology in F344
28 rats exposed to biphenyl. Groups of male F344 rats were administered biphenyl (in corn oil) by
29 single gavage dosing at 0, 250, 500, or 1,000 mg/kg and observed for 15 days following
30 treatment. Body weights were recorded, and urine was collected on days 1, 2, 3, 4, 8, and
31 15 following treatment for urinalysis. Interim sacrifices were performed on eight control and
32 eight high-dose rats on posttreatment days 1, 2, 3, 8, and 15 for assessment of weight and
33 histopathology of the kidney. There were no significant effects on body weight in the low-dose
34 group. Mean body weight gains of mid- and high-dose groups were consistently 6-10% lower
35 than control values (p < 0.05), beginning as early as day 2 following the initiation of dosing and
36 continuing through day 15. Dose-related increases in polyuria, proteinuria, and glucosuria were
37 observed on day 1; polyuria and glucosuria were no longer apparent by day 4 and proteinuria
38 resolved between days 8 and 15. Histopathologic examinations of kidneys revealed renal
46 DRAFT - DO NOT CITE OR QUOTE
-------
1 papillary necrosis in 8/32 high-dose rats; this effect was observed as early as day 1 and persisted
2 during the 15-day posttreatment period.
3 Kluwe et al. (1982) conducted a similar experiment in which groups of male F344 rats
4 received biphenyl at doses of 0, 250, or 500 mg/kg-day by gavage for 14 days. In this
5 experiment, polyuria persisted throughout the treatment period; glucosuria was no longer
6 apparent by day 4 and proteinuria resolved between treatment days 8 and 15. Relative kidney
7 weight of high-dose rats was significantly increased during the second half of the treatment
8 period, but the magnitude of this effect was small and considered by the study authors to be of
9 little biological significance. There was some indication of tubular dilatation in focal areas of
10 kidneys from the high-dose rats.
11 Groups of male and female SPF-Wistar rats were administered diets consisting of
12 semisynthetic chow and biphenyl at concentrations resulting in biphenyl doses of 0, 50, 150, 300,
13 or 450 mg/kg-day (S0ndergaard and Blom, 1979). Other groups were administered diets
14 consisting of commercial chow and biphenyl at concentrations resulting in biphenyl doses of 0,
15 50, 150, 300, 500, or 1,000 mg/kg-day. The treatment period lasted for up to 21 days. The
16 numbers of male and female rats in each treatment group are specified in Table 4-13. Urine was
17 collected on days 4, 10, and 17 for urinalysis. At terminal sacrifice, absolute and relative kidney
18 weights were determined and kidney tissues were prepared for light and electron microscopic
19 assessment. Apparently, interim sacrifices (days 1, 2, 4, and 10) were performed in order to
20 assess the activity of AP in proximal tubules. Table 4-13 presents semiquantitative study results,
21 which include increases in urine volume/specific gravity and relative kidney weight, as well as
22 polycystic kidney changes. No changes in AP levels were seen as a result of biphenyl exposure.
23 The kidney effects of biphenyl appeared to be more pronounced when added to the semisynthetic
24 diet versus the commercial diet, with 50 mg/kg-day as a LOAEL for the onset of kidney changes.
25
47 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-13. 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
-
-
-
-
-
**
-/-
•/•
•/•
aDose for 14 days.
bAbsolute organ weight.
+ = statistically significant compared with controls (p < 0.05), as calculated by the authors (Student's t-test);
+++ = statistically significant compared with controls (p < 0.001), as calculated by the authors (Student's t-test);
* = less than one-third of the area; ** = less than two-thirds of the area; *** = greater than two-thirds of the area;
• = effect; - = no effect
Source: Sendergaard and Blom (1979).
1
2 Male F344 rats (20/group) were exposed to 0 or 0.5% (w/w) biphenyl in the diet for
3 24 weeks (Shibata et al., 1989a). After 4 weeks, 5 rats/group were injected with 100 mg/kg
4 5-bromo-2-deoxyuridine (BrdU) and sacrificed 1 hour later. One kidney from each rat was
5 processed for immune-histopathologic identification of BrdU as an index of cell proliferation,
6 while the second kidney was processed for light and scanning electron microscopic examination.
7 The remaining rats were sacrificed after 8, 16, and 24 weeks to monitor further development of
8 morphological alterations in the renal papilla and pelvis. Survival was unaffected by treatment
9 and biphenyl-treated animals showed no adverse clinical signs. The study authors reported that
10 treatment resulted in significantly lower mean body weight compared to controls; food
11 consumption was unaffected and water consumption was slightly higher than that of controls.
12 There were no significant treatment-related effects on labeling indices of cell proliferation (BrdU
13 incorporation) in renal papilla or pelvic epithelia, and no histopathologic lesions of the renal
14 papilla and pelvis were evident. Focal calcification of the renal medulla was observed in the
15 majority of the biphenyl-treated rats. The study authors stated that urinalysis demonstrated an
16 association between biphenyl exposure and microcalculi formation, but provided no additional
17 information regarding urinalysis results.
48
DRAFT - DO NOT CITE OR QUOTE
-------
1 In a similar study (Shibata et al., 1989b), a group of 10 male F344 rats received 0.5%
2 (w/w) biphenyl in the diet for up to 8 weeks. Based on U.S. EPA (1988) subchronic reference
3 values for body weight and food consumption in male F344 rats, the dose was estimated at
4 500 mg/kg-day. At 4 weeks, five rats/group were processed as described by Shibata et al.
5 (1989b) for assessment of BrdU incorporation, but in the urinary bladder rather than in the
6 kidney. During week 4, urine samples were taken for urinalysis. At terminal sacrifice, urinary
7 bladder tissues were processed for scanning electron microscopic examinations. There were no
8 treatment-related deaths or adverse clinical signs. Although food and water consumption were
9 similar to controls, biphenyl-treated rats showed a consistent reduction in average body weight
10 (229 versus 247 g after 4 weeks and 300 versus 327 g after 8 weeks, for treated versus controls,
11 respectively \p < 0.01]). A greater than fourfold increase in the BrdU labeling index was
12 observed in urinary bladder epithelium of the biphenyl-fed rats (mean percent labeling index of
13 0.58 ± 0.31 compared to 0.13 ± 0.09 in controls;/? < 0.05). Urinalysis revealed numerous
14 microcalculi in the urinary sediment of the biphenyl-treated rats. This condition, designated as
15 "severe" by the authors, was associated with histopathological lesions of the epithelium of the
16 urinary bladder that included simple hyperplasia with moderate severity (5/5 rats), moderate
17 pleomorphic microvilli (5/5 rats), moderate uniform microvilli (5/5 rats), and the occurrence of
18 ropey or leafy microridges (5/5 rats), the latter condition designated as severe. Scanning electron
19 microscope images of the luminal surface of bladder epithelial cells showed pleomorphic
20 microvilli that varied in size and shape and the formation of microridges.
21
22 4.4.3. Biphenyl as a Tumor Promoter
23 Male B6C3Fi mice (10-20/group) received the bladder carcinogen N-butyl-
24 N (4-hydroxybutyl)nitrosamine (BBN) at 0 or 0.05% in the drinking water for 4 weeks followed
25 by 0 or 1% biphenyl in the feed for 32 weeks (Tamano et al., 1993). The mice were observed for
26 clinical signs, and body weight and food consumption were monitored. At 37-week terminal
27 sacrifice, kidneys and urinary bladders were prepared for histopathological examination. No
28 treatment-related clinical signs were observed. Mean body weight of the BBN +1% biphenyl -
29 treated mice was significantly (p < 0.01) lower than that of mice receiving BBN treatment only
30 (32.2 ±1.8 versus 38.4 ± 2.6 g). Biphenyl treatment did not result in increased incidences of
31 simple hyperplasia or papillary or nodular dysplasia in the BBN-initiated mice. Administration
32 of 1% biphenyl in the feed to eight mice for 8 weeks did not significantly affect indices of cell
33 proliferation (BrdU incorporation) in urinary bladder epithelium.
34 In the initiation-promotion portion of a chronic toxicity study designed to assess the
35 ability of biphenyl to promote carcinogenesis by EHEN in the kidney (see Section 4.2.1.2.1 for a
36 detailed study description), male Wistar rats (25/group) received a basal diet with either 0 or
37 0.1% dietary EHEN for 2 weeks, followed by a basal diet containing either 0, 0.125, or 0.5%
38 biphenyl for 34 weeks (Shiraiwa et al., 1989). At terminal sacrifice, gross pathologic
49 DRAFT - DO NOT CITE OR QUOTE
-------
1 examinations were performed. Kidney and urinary bladder were fixed; kidneys were sectioned
2 transversely (10-12 serial slices) and urinary bladders were cut into 4-6 serial slices. The
3 authors used a computer-linked image analyzer to determine the incidence of kidney lesions and
4 dysplastic foci. The presence of stones in the kidney and urinary bladder was assessed
5 qualitatively using an infrared spectrophotometer.
6 Based on reported values for mean daily biphenyl intake (mg biphenyl/rat) and average
7 body weight (mean initial body weight + one-half the difference between mean initial and mean
8 final body weight) for each study group, doses of biphenyl at the 0.125 and 0.5% dietary levels
9 are estimated to have been 59.28 and 248.3 mg/kg-day, respectively, for rats on basal diet alone
10 for the first 2 weeks and 62.0 and 248.2 mg/kg-day, respectively, for rats on basal diet and
11 EHEN for the first 2 weeks. Stones were present in the kidney, ureter, and urinary bladder of
12 high-dose rats irrespective of whether animals were initially exposed to the basal or
13 EHEN-containing diet (combined incidences of 6/25 and 8/25, respectively). The incidence of
14 rats with renal cell tumors after EHEN and subsequent biphenyl administration was lower than
15 that of rats receiving EHEN followed by basal diet (7/25 and 13/25, respectively). This finding
16 indicates that biphenyl was not a promoter of renal cell tumors in male Wistar rats under the
17 conditions of the study.
18 Male F344 rats (25/group) were exposed to 0.05% BBN (a bladder carcinogen) in the
19 drinking water for 4 weeks followed by diets containing either 0 or 0.5% biphenyl for 32 weeks
20 (Kurata et al., 1986). One group of five rats received biphenyl without pretreatment with BBN.
21 The rats receiving biphenyl either with or without pretreatment with BBN gained less weight
22 than control rats or those receiving only BBN. Incidences of urinary bladder hyperplasia,
23 papilloma, and carcinoma were 17/18 (94%), 15/18 (83%), and 11/18 (61%), respectively, in the
24 group of rats that survived treatment of BBN followed by biphenyl, compared to 6/24 (25%),
25 3/24 (12%), and 0/24 (0%), respectively, in the rats receiving BBN only. These urinary bladder
26 lesions were not seen in any of the five rats receiving biphenyl without BBN pretreatment.
27 Urinary bladder calculi were found in 25% of the rats receiving BBN followed by biphenyl and
28 in 12% of the rats receiving BBN only. Biphenyl was considered a urinary bladder tumor
29 promoter in male F344 rats under the conditions of the study.
30 Biphenyl was negative for tumor promotion in a skin-painting experiment in which the
31 initiator was 0.3% 9,10-dimethyl-l,2-benzanthracene in benzene (Boutwell and Bosch, 1959). In
32 the 16/20 mice that survived the topical application of 20% biphenyl for 16 weeks, none had
33 developed papillomas or carcinomas as a result of treatment.
34 Six-week-old male F344 rats (20-30/group) were exposed to BBN in drinking water at
35 0.01 or 0.05% for 4 weeks, followed by 0.5% biphenyl in the feed for 32 weeks (Ito et al., 1984).
36 Controls receiving only BBN and controls receiving only biphenyl were included. After
37 sacrifice, urinary bladders were prepared for light microscopic assessment of neoplastic and
50 DRAFT - DO NOT CITE OR QUOTE
-------
1 cancerous lesions. The study authors reported that biphenyl exhibited moderate bladder cancer-
2 promoting activity, but data to support this finding were not included in the study report.
o
3
4 4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
5 ACTION
6 Studies have been conducted to investigate the mechanisms by which biphenyl induces
7 effects on the urinary tract, liver, and endocrine system. Other studies have looked at the
8 potential for biphenyl to induce apoptosis, to affect mitochondrial activity, and to induce genetic
9 changes. This literature is summarized in Appendix B. Mechanistic studies of biphenyl effects
10 on the urinary tract, a principal target of biphenyl toxicity, and genotoxic potential are briefly
11 discussed below.
12
13 4.5.1. Effects on the Urinary Tract of Rats
14 Mechanistic studies in F344/N rats have been performed to identify urinary metabolites
15 of biphenyl, to assess conditions leading to calculi formation, and to determine the composition
16 of urinary crystals and calculi (Ohnishi et al., 2001; 2000a: 2000b). Urinary calculi in male
17 F344/N rats exposed to 4,500 ppm biphenyl in the diet for 104 weeks have been shown to be
18 composed mainly of 4-hydroxybiphenyl-O-sulphate, whereas calculi in female rats were
19 composed primarily of 4-hydroxybiphenyl and potassium sulphate, the hydrolysis products of 4-
20 hydroxybiphenyl-O-sulphate (Ohnishi et al., 2000b). Using a study design that also involved the
21 addition of potassium bicarbonate, potassium chloride, or sodium bicarbonate to the diet for 13
22 weeks, Ohnishi et al. (2001) determined that a combination of high urinary pH and high
23 potassium levels was necessary to cause precipitation of biphenyl sulphate, and proposed that the
24 crystalline precipitate caused obstruction that led to damaged of the transitional epithelium in the
25 urinary bladder.
26
27 4.5.2. Genotoxicity
28 The overall weight of evidence for biphenyl genotoxicity from short-term tests is
29 negative or equivocal (see Appendix B, Table B-2). Biphenyl did not induce mutations in a
30 variety of bacterial test systems (in the absence or presence of exogenous metabolic activation),
31 but in vitro assays of genotoxicity in mammalian test systems yielded a mix of negative and
32 positive results, with positive results mostly in the presence of metabolic activation. In tests of
33 clastogenic effects in mammalian systems, biphenyl induced SCEs, CAs, and micronuclei in
34 cultured human peripheral blood lymphocytes (Rencuzogullari et al., 2008) and CAs in one
35 assay of CHL fibroblasts in the presence, but not the absence, of rat liver metabolic activation
36 (Sofuni et al., 1985). However, biphenyl did not induce clastogenic effects (in the presence of
37 metabolic activation) in other assays with Chinese hamster fibroblasts (Ishidate et al., 1984;
51 DRAFT - DO NOT CITE OR QUOTE
-------
1 Ishidate and Odashima, 1977) or CHO cells (Yoshida et al., 1978). In the only adequately
2 reported in vivo genotoxicity studies with biphenyl, single oral doses of 2,000 mg/kg of biphenyl
3 or 2-hydroxybiphenyl induced DNA damage in several organs of CD-I mice (including liver and
4 bladder), but it is uncertain if the damage was due to a direct effect on DNA by biphenyl or its
5 metabolites or indirectly due to cytotoxicity or ROS generated by redox cycling of a
6 hydroquinone metabolite of 2-hydroxybiphenyl (Sasaki et al., 2002; Sasaki etal., 1997).
7 The overall weight of evidence for 2-hydroxybiphenyl genotoxicity (see Appendix B,
8 Table B-3) suggests that oxidative DNA damage from redox cycling between 2,5-
9 dihydroxybiphenyl and phenylbenzoquinone is possible (Sasaki et al., 2002; Sasaki etal., 1997;
10 Pathak and Roy. 1993: Morimoto et al.. 1989). but no evidence for DNA adducts or DNA
11 binding in urinary bladder epithelium tissue was found in rats following short-term (Kwok et al.,
12 1999) or subchronic (Smith et al., 1998) oral exposure to 2-hydroxybiphenyl at high doses
13 associated with the formation of urinary bladder tumors. Increased micronuclei in urinary
14 bladder epithelium were detected in rats exposed to 2% 2-hydroxybiphenyl or its sodium salt in
15 the diet for 14 days (Balakrishnan et al., 2002). The mechanism of this clastogenic effect is
16 uncertain, but could involve micronuclei formation in secondary response to cytotoxicity or
17 regenerative cell proliferation, DNA damage from ROS generated from redox cycling of a
18 hydroquinone metabolite, or protein modifications leading to mitotic spindle interference or
19 inhibition of enzymes important in DNA replication.
20 4-Hydroxybiphenyl, the predominant metabolite of biphenyl, was not mutagenic in
21 bacterial testing at noncytotoxic concentrations (Narbonne et al., 1987; Hanada, 1977) (see
22 Appendix B, Table B-3). 2,5-Dihydroxybiphenyl (i.e., phenylhydroquinone) caused in vitro
23 damage to human DNA from plasmid pbcNI in the presence of Cu(II) (Inoue et al., 1990) and
24 DNA adducts when applied to mouse skin (Pathak and Roy, 1993), but did not cause DNA
25 damage when injected intravesically into the urinary bladder of F344 rats at a concentration of
26 0.05% nviorimoto et al., 1989).
27
28 4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS
29 Tables 4-14 and 4-15 include the major studies and the observed effects for oral and
30 inhalation exposure to biphenyl, respectively.
52 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Subchronic studies
Rat, Long-Evans
(female, 8/group)
Mouse, BDFj
(10/sex/group)
Diet
Diet
0, 10, 30, or 100
90 d
0, 93, 347, 747, 1,495,
1,868, or 2,989
13wks
ND
M: 747
F: 747
ND
M: 1,495
F: 1,495
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
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.
(1953)a
Umeda et al.
(2004b)
Chronic studies
Rat, F344
(50/sex/group)
Diet
M: 0,36.4, 110, or
378
F: 0, 42.7, 128, or
438
2yrs
M: 110
F: 42.7
M: 378
F: 128
M: Bladder tumors and
transitional cell
hyperplasia.
F: Nonneoplastic kidney
lesions (transitional cell
hyperplasia in the renal
pelvis and hemosiderin
deposits).
Umeda et al. (2002)
53
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Rat, Wistar
(50/sex/group)
Rat, Wistar
(male, 25/group)
Rat, albino
(weanling,
15/sex/group)
Rat, albino
(male, 8/group)
Rat, 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
atO, 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
groups.
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.
Decreased survival and
small number of
animals/group may
have impaired ability to
detect late-developing
tumors.
Reference
Shiraiwa et al.
(1989)
Shiraiwa et al.
(1989)
Ambrose et al.
(I960)
Pecchiai and
Saffiotti (19571
Dow Chemical Co.
(1953V
54
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Mouse, BDFj
(50/sex/group)
Mouse, ddY
(female, 34-37/group)
Mice, hybrid
(two strains,
1 8/sex/strain/group)
Dog, Mongrel
(males/group; 1
female/group)
Monkey, Rhesus
(2 males/group;
Ifemale/group)
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
1,050
F: 0, 134, 414, or
1,420
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
lyr
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 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.
Author considered an
increase in relative liver
weight in high-dose
monkeys to be possibly
compound-related
Reference
Umeda et al. (2005)
Imai et al. (1983)
Innes et al. (1969):
NCI (1968)
Monsanto (1946)"
Dow Chemical Co.
(1953)a
55
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. Summary of major studies evaluating effects of biphenyl after oral administration in rats and mice
Species, strain
Exposure route
Dose (mg/kg-d),
duration
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Effect(s) at the LOAEL
Comments
Reference
Reproductive and developmental studies
Rat, Wistar
(18-20 pregnant
females/group)
Rat, Long Evans
(3 males/group;
9 females/group)
Rat, albino
Experiment 1:3-5
males/group; 9-10
females/group.
Experiment 2: 3-4
males/group; 8-9
females/group.
Gavage in corn oil
Diet
Diet
0, 125, 250, 500, or
1,000 on CDs 6-1 5
M: 9, 89, or 887
F: 10, 101, or 1,006
Continuous breeding
0, 105, or 525
Experiment 1 : 60 days
prior to mating
Experiment 2: 11 days
prior to mating
Dam: 500
Offspring:
250
M: ND
F: 101
ND
Dam: 1,000
Offspring:
500
M: ND
F: 1,006
ND
Dam: Maternal toxicity
(increased mortality),
increased in dead or
resorbed fetuses.
Offspring: Increased
incidence of anomalous
fetuses and litters.
M: ND
F: Decreased fertility and
litter size; reduced
offspring body weight.
ND
The effects seen in the
high-dose group may be
associated with
unpalatability and
resultant decreased food
intake.
Authors presented
tabulated data and
concluded that the
compound had no
significant effect on
reproduction.
Khera et al. (1979)
Dow Chemical
Co.(1953V
Ambrose et al.
(I960)
aReport was not peer reviewed.
F = female; M = male; ND = not determined
Note: Other studies of subchronic duration that examined the effects of biphenyl on the urinary tract only (Shibata et al.. 1989a; Shibataetal.. 1989b) are summarized in
Section 4.4.2. Because these studies were designed to investigate the effects of biphenyl on the kidney and urinary bladder and the mode of action by which biphenyl
induces these effects, the studies were not useful for identifying NOAELs and LOAELs, and were not included in this table.
56
DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-15. Summary of major studies evaluating effects of biphenyl after inhalation exposure in rats, mice and
rabbits
Species, strain
Rabbit, albino
(3/group)
Rat, Sprague-Dawley
(10/group)
Rabbit, albino
(3/group)
Rat, Sprague-Dawley
(6/group)
Mice (12/group)
Rat, Sprague-Dawley
(4/group)
Mouse, GDI
(50/sex/group)
Dose (mg/m3), duration
300 mg/m3 (7 hrs/d, 5 d/wk)
64 d over 94-d period
40 mg/m3 (7 hrs/d, 5 d/wk)
46 d over 68-d period
5 mg/m3 (7 hrs/d, 5 d/wk)
62 d over 92-d period
0, 157.7, or 315.3 mg/m3 (7 hrs/d,
5 d/wk), 13 wks
NOAEL
(mg/m3)
Rabbit: ND
Rat: ND
Rabbit: ND
Rat: ND
Mouse: ND
Rat: ND
ND
LOAEL
(mg/m3)
Rabbit: ND
Rat: 300
Rabbit: ND
Rat: 40
Mouse: 5
Rat: ND
157.7
Effect(s) at the LOAEL
Rabbit: ND
Rat: Mortality (5/10), acute emphysema,
congestion, edema, bronchitis, lobular pneumonia,
and multiple pulmonary abscesses
Rabbit: ND
Rat: Mortality (1/6), acute emphysema, congestion,
edema, bronchitis, lobular pneumonia, and multiple
pulmonary abscesses
Mouse: Mortality (2/12); upper respiratory tract
irritation (acute emphysema, congestion, edema,
bronchitis, lobular pneumonia, and multiple
pulmonary abscesses)
Rat: ND
Histopathologic lung, liver, and kidney lesions
(congested and hemorrhagic lungs, trachea!
hyperplasia, and congestion and edema in the liver
and kidney) in both sexes.
References
Deichmann et al.
(1947): Monsanto
(1946)
Sun Company Inc.
(1977b)a
aReport was not published.
ND = not determined
57
DRAFT - DO NOT CITE OR QUOTE
-------
1 4.6.1. Oral
2 Biphenyl displays a relatively low acute oral toxicity, with LD50 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, consistent
6 findings indicated reduced body weight gain (Umeda et al., 2005; Umeda et al., 2004b: Umeda et
7 al., 2002; Ambrose et al., I960; Dow Chemical Co, 1953) and increased liver and kidney weights
8 (Umeda et al.. 2004b: Umeda et al.. 2002: S0ndergaard and Blom, 1979: Ambrose et al.. 1960:
9 Dow Chemical Co, 1953: Monsanto, 1946) in rodents. In some studies, reduced weight gain has
10 been attributed to low palatability of the feed by the authors (Ambrose et al., I960: Dow
11 Chemical Co, 1953): however, the feed intake data of Umeda et al. (2005) in mice did not
12 support this hypothesis. Signs of liver damage (increased serum activities of ALT, AST, AP,
13 and LDH) were observed in mice (Umeda et al., 2005).
14 Pathological effects on the urinary system have been reported in dogs (Monsanto, 1946),
15 rats (Umeda et al., 2002: Dow Chemical Co, 1953), and mice (Umeda et al., 2005). Increased
16 urine volume with increased specific gravity, polycystic changes, nephritis, and precipitation of
17 free 4-OH-biphenyl and its glucuronide in urine have been consistently reported following oral
18 exposure to biphenyl (Kluwe, 1982: S0ndergaard and Blom, 1979: Booth etal., 1961: Monsanto,
19 1946). Calculi appeared in the urine of male rats only (Umeda et al., 2002: Ohnishi et al., 2001:
20 Ohnishi et al., 2000a: Ohnishi et al., 2000b: Shibata et al., 1989b: Ambrose et al., 1960).
21 Urothelial hyperplasia with increased indices of cell proliferation have been described in rats but
22 not in mice and were attributed to irritation by calculi (Umeda et al., 2005: Umeda et al., 2002:
23 Shibata et al., 1989b). Tubular dilatation and morphological changes in papillae and pelvis,
24 kidney stones, obstructive pyelonephritis, tubular atrophy, fibrosis, and pelvic hyperplasia were
25 observed (Shibata et al., 1989b: Shibata et al., 1989a: Shiraiwa et al., 1989: Takita, 1983: Kluwe,
26 1982: Booth etal., 1961).
27 Increased incidences of fetuses with skeletal anomalies were reported following gavage
28 administration of biphenyl to Wistar rats during gestation (Kheraet al., 1979). A three-
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
58 DRAFT - DO NOT CITE OR QUOTE
-------
1 well as brain edema (Hakkinen et al., 1973; Hakkinen et al., 1971). A small cluster of
2 Parkinson's disease (PD) was reported at a facility manufacturing biphenyl-impregnated paper,
3 but other studies have not found a similar association (Wastensson et al., 2006). The workplace
4 conditions reported for these studies (Wastensson et al., 2006; Hakkinen et al., 1973; Hakkinen
5 et al., 1971) suggested that inhalation represented the predominant route of exposure, but dermal
6 absorption as well as oral uptake (hand to mouth) might have occurred at a 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 observable clinical toxicity (Sun, 1977b). In another study,
9 groups of rabbits, rats, or mice 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 at high concentrations. Mice were more sensitive than rats in these experiments, additionally
13 showing congestion and hemorrhage of the lungs (Deichmann et al., 1947). High incidences of
14 pneumonia and tracheal hyperplasia, and congestion and edema in the lungs, liver, and kidney
15 were reported in a 13-week inhalation study of biphenyl in mice that was limited by study
16 methodologic and reporting issues (Sun, 1977a). Reproductive or developmental studies using
17 the inhalation route of exposure were not identified.
18
19 4.6.3. Mode-of-Action Information
20 The studies described above have demonstrated that exposure to biphenyl may lead to a
21 variety of noncancer health effects (i.e., weight loss, liver toxicity, urinary tract toxicity).
22 However, the available information is insufficient to establish the mode of action for noncancer
23 health effects following exposure to biphenyl.
24 Weight loss or lack of weight gain has been consistently associated with oral exposure to
25 biphenyl QJmeda et al.. 2005: Umeda et al.. 2002: Shibata et al.. 1989b: Ambrose et al.. 1960:
26 Dow Chemical Co, 1953). The work of Nishihara (1985) provides a possible explanation for this
27 toxic effect. This author found that, in vitro, biphenyl can act as an uncoupler of respiration. It
28 may be speculated that long-term, high-dose exposure to biphenyl uncouples mitochondrial
29 respiration to a certain extent, resulting in a futile cycle that diverts the use of nutrients from
30 building body mass into maintaining necessary energy stores. It is not clear at what level of in
31 vivo exposure this effect might become operative.
32 Several of the oral animal studies (Umeda et al., 2005: Sun, 1977a: Pecchiai and Saffiotti,
33 1957: Dow Chemical Co, 1953: Deichmann et al., 1947) and the epidemiological study by
34 Hakkinen et al. (1973) provide evidence that the liver is a target for biphenyl toxicity. This
35 evidence consists of changes in blood parameters that are indicative of liver toxicity; however, in
36 animal studies, liver histopathology does not support or explain this finding. Evidence for
37 damage to the nervous system was suggested in epidemiology studies by Hakkinen et al. (1973)
38 and Seppalainen and Hakkinen (1975). The nervous system has not been identified as a target in
59 DRAFT - DO NOT CITE OR QUOTE
-------
1 chronic toxicity studies in rodents. The limited evidence for an estrogenic activity of
2 4,4'-dihydroxybiphenyl (Kitamura et al., 2003; Schultz et al., 2002) is insufficient to assign a
3 clear endocrine-disrupting effect to this major metabolite of biphenyl.
4 Damage to the urinary tract has been observed consistently in animals. The work of
5 Ohnishi et al. (2001; 2000a: 2000b) provides evidence that, in the rat, this is due to the
6 precipitation in the urinary tract of crystals consisting mostly of 4-hydroxybiphenyl. These
7 crystals irritate the epithelia of ureters and bladder, leading to chronic inflammation and
8 obstruction of the urinary tract with subsequent hydronephrosis. The work of Ohnishi et al.
9 (2001; 2000b) has made it clear that, at least in their animal model, two conditions are required
10 for this event to occur: (1) the pH in the urine of the animals needs to be elevated and (2)
11 elevated potassium levels need to accompany the elevated pH because it is the potassium salt of
12 4-hydroxybiphenyl sulphate that has the lowest solubility in high-pH urine.
13
14 4.7. EVALUATION OF CARCINOGENICITY
15 4.7.1. Summary of Overall Weight of Evidence
16 Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the
17 database for biphenyl provides "suggestive evidence of carcinogenic potential" at
18 environmentally relevant exposure levels in humans where the formation of urinary bladder
19 tumors would not be expected to occur. This cancer descriptor is based on an increase in the
20 incidence of liver tumors (hepatocellular adenomas and carcinomas) in female BDFi mice
21 (Umeda et al., 2005) and urinary bladder tumors (transitional cell papillomas and carcinomas) in
22 male F344 rats (Umeda et al., 2002) exposed to biphenyl in the diet for 104 weeks, as well as
23 information on mode of carcinogenic action. Earlier chronic cancer bioassays in orally exposed
24 animals found no clear evidence of biphenyl-induced carcinogenicity in rats (Shiraiwa et al.,
25 1989: Ambrose et al.. 1960: Pecchiai and Saffiotti, 1957: Dow Chemical Co. 1953). mice (Imai
26 etal.. 1983: Innesetal.. 1969: NCL 1968). dogs (Monsanto. 1946). or Rhesus monkeys (Dow
27 Chemical Co, 1953). The findings from these earlier studies were less informative for the
28 carcinogenicity of biphenyl than Umeda et al. (2005, 2002) because of various study limitations.
29 With the exception of Imai et al. (1983), these limitations include small group sizes and shorter-
30 than-lifetime exposure durations due to design or decreased survival unrelated to tumor
31 development. Imai et al. (1983) found no evidence of carcinogenic responses in female mice of
32 a different species (ddY mice) (n = 34) exposed to 5,000 ppm biphenyl in the diet for 2 years
33 (Imai etal.. 1983).
34 The Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) ("Cancer
35 Guidelines") emphasize the importance of weighing all of the evidence in reaching conclusions
36 about the human carcinogenic potential of agents. Information on mode of action has been taken
37 into consideration in evaluating the weight of evidence for carcinogenicity. The induction of
38 urinary bladder tumors in F344 male rats by dietary biphenyl exposure is a high-dose
60 DRAFT - DO NOT CITE OR QUOTE
-------
1 phenomenon closely related to the formation of urinary bladder calculi. As discussed in more
2 detail in Section 4.7.3.1, the mode of action information is sufficient to conclude that urinary
3 bladder tumors will not occur without the development of calculi. While the proposed mode of
4 action for urinary bladder tumors in male rats is assumed to be relevant to humans, the available
5 evidence suggests that humans would be less susceptible to these tumors than rats (see discussion
6 in Section 4.7.3.1.4.2). Overall, the mode of action analysis supports the conclusion that
7 biphenyl should not pose a risk of urinary bladder tumors at environmentally relevant exposure
8 levels in humans.
9 The available data are insufficient to establish a mode of action for liver tumors in female
10 mice (see Section 4.7.3.2.2.1 for further discussion). In the absence of information to indicate
11 otherwise, the development of liver tumors in female mice with chronic exposure to biphenyl is
12 assumed to be relevant to humans. EPA acknowledges that some mouse strains are relatively
13 susceptible to liver tumors and the background incidence of this tumor can be high. For these
14 reasons, use of mouse liver tumor data in risk assessment has been a subject of controversy
15 (King-Herbert and Thayer, 2006). According to historical control data from the Japan Bioassay
16 Research Center, the institute that conducted the mouse bioassay published by Umeda et al.
17 (2005), the mean incidence of liver tumors (hepatocellular adenoma or carcinoma) in male and
18 female control BDFi mice is 32.2 and 7.1%, respectively, incidences consistent with the
19 concurrent controls in the mouse bioassay of biphenyl. The relatively low background incidence
20 of liver tumors in female control mice from Umeda et al. (2005) minimizes the possible
21 confounding of compound-related liver tumors in this sex.
22 Thus, when one takes into consideration information on the mode of action for biphenyl-
23 induced tumors, risk of female liver tumors only is operative at environmentally relevant
24 exposures. Accordingly, this assessment concludes that there is "suggestive evidence of
25 carcinogenic potential."
26 EPA's Cancer Guidelines (U.S. EPA, 2005a) indicate that for tumors occurring at a site
27 other than the initial point of contact, the cancer descriptor may apply to all routes of exposure
28 that have not been adequately tested at sufficient doses. An exception occurs when there is
29 convincing toxicokinetic data that absorption does not occur by other routes. Information
30 available on the carcinogenic effects of biphenyl demonstrates that tumors occur in tissues
31 remote from the site of absorption following chronic oral exposure (urinary bladder in male rats
32 and liver in female mice). No information on the carcinogenic effects of biphenyl via the
33 inhalation or dermal routes in humans and animals is available. Biphenyl is rapidly and
34 extensively absorbed by the oral route of exposure, but no studies of uptake following inhalation
35 or dermal exposure have been conducted; however, a case report of hepatic toxicity produced by
36 a probable combination of inhalation and dermal exposures in a worker in a biphenyl -
37 impregnated fruit wrapping paper production facility (Hakkinen et al., 1973) provides qualitative
38 evidence of absorption by these routes. Therefore, based on the observation of systemic tumors
61 DRAFT - DO NOT CITE OR QUOTE
-------
1 following oral exposure and assumed absorption by all routes of exposure, it is assumed that an
2 internal dose will be achieved regardless of the route of exposure. In the absence of information
3 to indicate otherwise, the database for biphenyl provides "suggestive evidence of carcinogenic
4 potential" by all routes of exposure.
5
6 4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence
7 Available human studies were not designed to evaluate associations between exposure to
8 biphenyl and occurrence of cancer (see Section 4.1).
9 As discussed in Section 4.2, carcinogenicity studies in animals are limited to the oral
10 exposure route. In well-designed cancer bioassays of F344 rats (Umeda et al., 2002) and BDFi
11 mice (Umeda et al., 2005), dietary exposure to biphenyl resulted in the occurrence of urinary
12 bladder tumors in male rats and significantly increased incidences in liver tumors in female mice.
13 Earlier chronic toxicity and carcinogenicity assessments found no clear evidence of
14 biphenyl-induced carcinogenicity in orally exposed rats, mice, dogs, or Rhesus monkeys.
15 However, these studies were generally limited in design, with the exception of a study reporting
16 no evidence of carcinogenic responses in female ddY mice (n = 34 mice versus 37 control mice)
17 exposed to 5,000 ppm biphenyl in the diet for 2 years (Imai et al., 1983). In a study of Wistar
18 rats, sufficient numbers of animals (50/sex/group) were exposed to biphenyl in the diet at
19 concentrations up to 5,000 ppm, but only for 75 weeks (Shiraiwa et al., 1989). Some of the male
20 rats exhibited urinary bladder calculi and simple or diffuse hyperplasia and papillomatosis of the
21 urinary bladder mucosa in the absence of neoplastic lesions, but the study may have been
22 terminated prior to eventual urinary bladder tumor development. Ambrose et al. (1960) exposed
23 albino rats (15/sex/exposure level) to biphenyl in the diet at concentrations up to 10,000 ppm for
24 2 years (10, 50, 100, 500, 1,000, 5,000, or 10,000 ppm); however, decreased survival in rats
25 exposed to 5,000 or 10,000 ppm, presumably from decreased food consumption, and the
26 relatively small numbers of animals in each exposure group may have impaired the ability to
27 detect late-developing tumors. In another study, groups of Sprague-Dawley rats (12/sex/group)
28 received biphenyl in the diet at concentrations up to 10,000 ppm for up to 2 years (Dow
29 Chemical Co, 1953). However, this study suffered from a pneumonia outbreak, particularly
30 among control males, and the relatively small numbers of animals and the decreased survival
31 may have impaired the ability to detect late-developing tumors. A study of male albino rats
32 included small numbers of rats (8/group) and a short (13 months) exposure period (Pecchiai and
33 Saffiotti, 1957). A study of B6C3Fi or B6AkFi mice exposed to biphenyl in the diet for only
34 18 months (Innes et al., 1969; NCI, 1968) included relatively small numbers of mice
35 (18/sex/group) and only one exposure level (517 ppm) that was similar to the concentration
36 (667 ppm) without carcinogenic effect in the Umeda et al. (2005) 24-month BDFi mouse
37 bioassay. The dog study included two males and one female, a high dose of 25 mg/kg-day, and
38 an exposure period of only 1 year (Monsanto, 1946). Rhesus monkeys (two males and one
62 DRAFT - DO NOT CITE OR QUOTE
-------
1 female) were exposed to biphenyl in the diet at a concentration of 10,000 ppm, but for only
2 1 year (Dow Chemical Co. 1953).
3 The overall weight of evidence for biphenyl genotoxicity from short-term tests is
4 negative or equivocal. Biphenyl did not induce mutations in a variety of bacterial test systems,
5 but both negative and positive results were obtained in mammalian in vitro test systems (see
6 Section 4.5.6). Single oral doses of 2,000 mg biphenyl/kg induced DNA damage (detected by
7 the Comet assay) in several organs of CD-I mice (including the liver and bladder), but it is
8 uncertain if the damage was due to a direct effect on DNA or was an indirect effect due to
9 cytotoxicity or ROS generated by redox cycling of phenylhydroquinone, a major urinary
10 metabolite of 2-hydroxybiphenyl and a minor metabolite of biphenyl in rats (Sasaki et al., 2002;
11 Smith etal.. 1998).
12 The overall weight of evidence for 2-hydroxybiphenyl genotoxicity suggests that
13 oxidative DNA damage from ROS from redox cycling between 2,5-dihydroxybiphenyl and
14 phenylbenzoquinone is possible. DNA damage was detected in liver and bladder of CD-I mice
15 exposed to 2,000 mg/kg of 2-hydroxybiphenyl (Sasaki et al., 2002, 1997) and in the urinary
16 bladder of male F344 rats fed the sodium salt of 2-hydroxybiphenyl at 1 or 2% in the diet for 3-
17 5 months (Morimoto et al., 1989). DNA adducts were detected by [32P]-post labeling in skin of
18 CD-I mice after topical application of the sodium salt of 2-hydroxybiphenyl or phenylhydro-
19 quinone (Pathak and Roy, 1993), and increased micronuclei were detected in urinary bladder
20 epithelium of male F344 rats exposed to 2,000 ppm 2-hydroxybiphenyl or 2,000 ppm NaCl plus
21 2,000 ppm 2-hydroxybiphenyl in the diet for 2 weeks (Balakrishnan et al., 2002). However,
22 increased binding of radioactivity to DNA was not detected in DNA extracted from urinary
23 bladder epithelium of male F344 rats exposed to single gavage doses of 2-hydroxybiphenyl as
24 high as 1,000 mg/kg (Kwok et al., 1999), and DNA adducts were not detected in urinary bladder
25 epithelium of male F344 rats exposed for 13 weeks to biphenyl dietary concentrations as high as
26 12,500 ppm (Smith et al., 1998). The mechanism by which 2-hydroxybiphenyl may induce
27 micronuclei in the urinary bladder epithelium is uncertain, but could involve micronuclei
28 generation as a secondary response to cytotoxicity or regenerative cell proliferation, DNA
29 damage from ROS from redox cycling of 2,5-dihydroxybiphenyl, or protein modifications
30 leading to mitotic spindle interference or inhibition of enzymes important in DNA replication
31 (Balakrishnan et al., 2002). The hydroxylation of biphenyl to produce 2-hydroxybiphenyl is a
32 minor pathway in rats and mice (Halpaap-Wood et al., 1981a, b; Meyer and Scheline, 1976).
33 2-Hydroxybiphenyl and 2,5-dihydroxybiphenyl collectively accounted for <2% of metabolites in
34 urine of rats administered single oral doses of 100 mg biphenyl/kg (Meyer and Scheline, 1976)
35 or single i.p. doses of 30 mg biphenyl/kg (Halpaap-Wood et al., 1981b). In mice given i.p. doses
36 of 30 mg biphenyl/kg, these metabolites accounted for <5% of urinary metabolites (Halpaap-
37 Wood etal., 1981b).
38
63 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.7.3. Mode-of-Action Information
2 4.7.3.1. Mode-of-Action Information for Bladder Tumors in Male Rats
3 4.7.3.1.1. Hypothesized mode of action. The best-supported hypothesis proposes a mode of
4 action whereby the formation of urinary bladder calculi (from the precipitation of
5 4-hydroxybiphenyl-O-sulphate) is a key event in the development of urinary bladder tumors in
6 male rats fed high levels of biphenyl in the diet for 2 years. According to this hypothesis, the
7 calculi (occurring in association with increased urinary pH and potassium, and predominantly
8 composed of 4-hydroxybiphenyl-O-sulphate) cause irritation to transitional epithelial cells of the
9 urinary bladder leading to sustained cell proliferation, which promotes the development of
10 initiated cells in the urinary bladder with progression to papillomas and carcinomas.
11
12 4.7.3.1.2. Experimental support for the hypothesized mode of action
13 Strength, consistency, and specificity of association, including support for the hypothesized
14 mode of action in male rats. The formation of urinary bladder calculi, predominantly composed
15 of potassium 4-hydroxybiphenyl-O-sulphate, is strongly, consistently, and specifically associated
16 with the formation of urinary bladder tumors in male rats chronically exposed to high dietary
17 concentrations of biphenyl. Several findings support this association. Urinary bladder calculi
18 were formed at a high prevalence (43/50; 86%) in a group of male rats exposed to biphenyl in the
19 diet at a concentration of 4,500 ppm, but were absent in male rats receiving diets containing 0,
20 500, or 1,500 ppm biphenyl (Umeda et al., 2002). These observations were consistent with the
21 detection of urinary bladder transitional cell papilloma (10/50; 20%), carcinoma (24/50; 48%),
22 and papilloma or carcinoma (31/50; 62%) in the 4,500 ppm group of male rats and total absence
23 of urinary bladder papilloma or carcinoma in the control, 500, or 1,500 ppm groups of male rats.
24 Bladder calculi were found in all 24 of the male rats with urinary bladder transitional cell
25 carcinoma and in 8/10 of the male rats with transitional cell papilloma.
26 The association between urinary bladder calculus formation and development of urinary
27 bladder tumors is both gender and species specific. Urinary bladder calculi, of similar size to
28 those observed in males, were observed at much lower incidence (8/50; 16%) in the 4,500 ppm
29 female rats, but they were of more uniform color (white and yellow versus white, yellow, brown,
30 gray, and black in males) and shape (spheroidal versus triangular, pyramidal, cubical, and
31 spheroidal in males) and primarily composed of 4-hydroxybiphenyl and potassium bisulphate
32 (which are hydrolysis products of potassium 4-hydroxybiphenyl-O-sulphate) (Umeda et al.,
33 2002: Ohnishi et al.. 2000b). No urinary bladder calculi were found in the 500 and 1,500 ppm
34 groups of female rats. Transitional cell hyperplasia was found in 10/50 4,500-ppm female rats,
35 but no urinary bladder transitional cell papillomas or carcinomas were seen in any of the
36 biphenyl-exposed groups of female rats. Furthermore, there was no evidence of biphenyl-
37 induced urinary bladder calculi or bladder tumors in male or female BDFi mice receiving dietary
38 biphenyl at concentrations as high as 6,000 ppm for 2 years (Umeda et al., 2005).
64 DRAFT - DO NOT CITE OR QUOTE
-------
1 Urinary bladder calculi in male rats were associated with significantly increased urinary
2 pH (average pH of 7.97 in the 4,500 ppm group at the final week of exposure compared to
3 7.66 in controls) (Umeda et al., 2002) and were composed primarily of potassium
4 4-hydroxybiphenyl-O-sulphate (Ohnishi et al., 2000b). The urine pH of female rats exposed to
5 4,500 ppm for 104 weeks (pH = 7.26) was not elevated compared with controls (pH = 7.29)
6 (Umeda et al., 2002). From these observations, it appears that the formation of the calculi results
7 from the precipitation of the potassium salt of the sulphate conjugate of 4-hydroxybiphenyl
8 under the elevated pH conditions of the male rat urine. The mechanism responsible for increased
9 urinary pH is unknown, although Ohnishi et al. (2001; 2000a: 2000b) proposed that gender
10 differences in urinary conditions, such as pH and potassium concentrations, and sulphatase
11 activities in kidneys, may be responsible for the gender differences in urinary calculi
12 composition and formation and the subsequent development of urinary bladder tumors in male,
13 but not female, F344 rats.
14 Relatively strong, consistent, and specific associations between calculi formation and
15 transitional cell hyperplasia and between transitional cell hyperplasia and the development of
16 transitional cell tumors in the urinary bladder have been shown in male F344 rats chronically
17 exposed to high concentrations of biphenyl in the diet. Urinary bladder transitional cell
18 hyperplasia (simple, nodular, papillary) occurred in 45/50 (90%) male rats receiving biphenyl in
19 the diet for 2 years at the same dietary concentration (4,500 ppm) at which high prevalences of
20 both urinary bladder calculi formation (43/50; 86%) and transitional cell tumors (31/50 62%)
21 were observed (Umeda et al., 2002). Forty-two of the 45 male rats with urinary bladder
22 transitional cell hyperplasia also exhibited urinary bladder calculi. In another study, evidence of
23 biphenyl-induced calculi formation (microcalculi in the urine) and increased indices of urinary
24 bladder transitional cell proliferation (greater than fourfold increase in BrdU incorporation) in
25 male F344 rats has been reported following as little as 4-8 weeks of dietary exposure to
26 5,000 ppm biphenyl (Shibata et aL 1989b).
27 The most convincing evidence that degenerative changes in the urinary bladder
28 epithelium lead to tumor formation is the site-concordance of associations between calculi
29 formation in the urinary bladder, transitional cell proliferation, transitional cell hyperplasia, and
30 transitional cell tumors (Umeda et al., 2002). In addition, the strong associations between
31 urinary tract calculi formation, ulcerations or inflammation, and subsequent hyperplasia
32 combined with repeated, high-level exposure to other chemicals that cause urinary bladder
33 tumors in rodents, including melamine, uracil, and the sodium salt of 2-hydroxybiphenyl (Capen
34 etal.. 1999: IARC. 1999a: IARC. 1999b: Cohen. 1998. 1995) provide further evidence that
35 degenerative changes are involved in the etiology of rodent urinary bladder tumors. It is not
36 unusual to see extensive proliferation or hyperplasia in bladder epithelium in response to urinary
37 calculi from other rodent bladder tumorigens without an associated ulceration or intense
38 inflammatory response. In male rats exposed to 4,500 ppm biphenyl, increasing numbers of rats
65 DRAFT - DO NOT CITE OR QUOTE
-------
1 with clinical hematuria were observed beginning at about the 40th week of exposure, and
2 histologic examinations at study termination revealed focal hyperplasia in 45/50 rats, providing
3 some evidence of calculi-induced bladder epithelial damage followed by cell proliferation
4 (Umeda et al., 2002). Over the course of the study, 94% of male rats with hematuria had bladder
5 or kidney calculi, but hematuria was not found in any biphenyl-exposed females. In addition,
6 with 8 weeks, but not 4 weeks, of exposure to 5,000 ppm biphenyl in the diet, moderate urinary
7 bladder epithelial hyperplasia and microcalculi in urine were observed in 5/5 male F344 rats, but
8 no descriptions of degenerative changes were provided; these observations are consistent with a
9 rapid repair response to epithelial damage from biphenyl-induced urinary tract calculi (Shibata et
10 al.. 1989b).
11 The ability of repeated biphenyl exposure to promote previously initiated urinary bladder
12 cells to bladder tumors is supported by results of a bladder tumor initiation-promotion study
13 (Kurata et al., 1986). Incidences of urinary bladder hyperplasia, papilloma, and carcinoma were
14 significantly increased in male F344 rats initiated with dietary BBN for 4 weeks followed by
15 5,000 ppm biphenyl in the diet for 32 weeks, compared with rats receiving BBN only for
16 4 weeks. For example, 94 and 83% of rats treated with BBN followed by biphenyl developed
17 urinary bladder hyperplasia and papillomas, respectively, compared with 25 and 12% of rats
18 exposed to BBN alone.
19 The hypothesis that the mode of action involves the development of urinary bladder
20 tumors in biphenyl-exposed male rats is further supported by an overall negative or equivocal
21 weight of evidence for the genotoxicity of biphenyl. As discussed earlier, there are consistently
22 negative results for biphenyl in bacterial mutation assays and inconsistent positive results for
23 biphenyl in in vitro mammalian assays mostly in the presence of metabolic activation. There is
24 evidence that 2,5-dihydroxybiphenyl (i.e., phenylhydroquinone), the principal urinary metabolite
25 in rats exposed to high doses of 2-hydroxybiphenyl, can undergo redox cycling to produce ROS
26 that may damage DNA and lead to tumor-initiating mutations; however, 2-hydroxybiphenyl is a
27 minor urinary metabolite of biphenyl in rats and 2,5-dihydroxybiphenyl was not detected in urine
28 of rats exposed to oral doses of 100 mg biphenyl/kg (Meyer and Scheline, 1976).
29
30 Dose-response concordance. Dose-response relationships for urinary bladder calculi formation,
31 transitional cell hyperplasia, and transitional cell tumor development show concordance in the 2-
32 year oral study of rats (Umeda et al., 2002). In male rats, urinary calculi, nonneoplastic lesions
33 (epithelial hyperplasia), and neoplastic lesions (papillomas and carcinomas) of the urinary
34 bladder were observed only at the highest exposure level (4,500 ppm); no urinary bladder calculi,
35 transitional cell hyperplasia, or transitional cell tumors were found in control, 500, or 1,500 ppm
36 male rats. Furthermore, urinary bladder calculi were found in 43/45 high-dose male rats, in all
37 24 male rats with transitional cell carcinoma, and in 8/10 of the male rats with transitional cell
38 papilloma.
66 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Temporal relationship. Results from the 2-year oral study in rats (Umeda et al., 2002) provide
3 some evidence of a progression from urinary bladder calculi formation to the development of
4 bladder tumors. Urinary bladder calculi were observed in the first 4,500 ppm male rat that died
5 (week 36), evidence of blood in the urine was observed in 4,500 ppm male rats by week 40, and
6 incidences of bladder calculi and bloody urine that paralleled increases in mortality and tumor
7 formation were observed throughout the remainder of the study. In addition, results of a short-
8 term oral study demonstrate that microcalculi can be detected in the urine of male rats after as
9 little as 4 weeks of dietary exposure to 5,000 ppm biphenyl and that hyperplasia of urinary
10 bladder epithelium can be detected at least by week 8 (Shibata et al., 1989b). Presumably, the
11 development of biphenyl-induced urinary bladder tumors requires a longer exposure period to
12 urinary calculi of sufficient size, shape, and composition to induce urinary bladder epithelial
13 damage and a sustained proliferative response.
14
15 Biological plausibility and coherence. The proposed mode of action is consistent with the
16 current understanding of cancer biology and is supported by the wide body of evidence that other
17 chemicals with primarily nongenotoxic profiles produce urinary bladder tumors in rodents at
18 high exposure levels by a mode of action involving calculi formation, ulceration or
19 inflammation, and regenerative cell proliferation (Capen et al., 1999; IARC, 1999a, b; Cohen,
20 1998, 1995). Additional information could strengthen the plausibility and coherence of the
21 proposed mode of action to explain the occurrence of biphenyl-induced urinary bladder tumors
22 in male rats. These additional data include results from investigations of earlier time points in
23 the proposed temporal progression from calculi formation to epithelial damage, regenerative cell
24 proliferation, and tumor development and further investigations into the factors underlying
25 gender-specific differences in precipitation of 4-hydroxybiphenyl-O-sulphate to form bladder
26 calculi in rats.
27
28 4.7.3.1.3. Other possible modes of action for bladder tumors in male rats. Although the
29 weight of evidence from short-term standard genotoxicity tests with biphenyl and
30 4-hydroxybiphenyl is predominantly negative, evidence is available that suggests that oral
31 exposure to high doses of 2-hydroxybiphenyl is associated with the development of urinary
32 bladder tumors in male rats. The induction of genotoxic effects in the urinary bladder epithelium
33 leading to tumor initiation is proposed to occur via redox cycling between 2,5-dihydroxy-
34 biphenyl and phenylbenzoquinone (Balakrishnan et al., 2002; Kwok et al., 1999; Pathak and
35 Roy, 1993; Morimoto et al., 1989). However, the strong, consistent, and specific association
36 between the occurrence of urinary bladder calculi composed of 4-hydroxybiphenyl-O-sulphate
37 and development of urinary bladder tumors in male but not female rats, the evidence that 2-
38 hydroxybiphenyl is a minor urinary metabolite of biphenyl and, finally, that
67 DRAFT - DO NOT CITE OR QUOTE
-------
1 2,5-dihydroxybiphenyl was not detected in the urine of biphenyl-exposed rats, demonstrate that
2 the support for a genotoxic mode of action involving key mutational events from biphenyl or its
3 metabolites in the urinary bladder leading to initiation of tumor cells is not compelling.
4 Additional support for a proposed genotoxic mode of action would come from studies showing
5 formation of 2,5-dihydroxybiphenyl and phenylbenzoquinone in the urinary bladder epithelium
6 of rats exposed to low doses of biphenyl.
7
8 4.7.3.1.4. Conclusions about the hypothesized mode of action for bladder tumors in male rats
9 Support for the hypothesized mode of action in rats. There is strong evidence that urinary
10 bladder tumors in male rats chronically exposed to biphenyl in the diet is a high-dose
11 phenomenon involving sustained occurrence of calculi in the urinary bladder leading to
12 transitional cell damage, sustained regenerative cell proliferation, and eventual promotion of
13 spontaneously initiated tumor cells in the urinary bladder epithelium.
14 To summarize, chronic exposure of male rats to a high dietary concentration of biphenyl
15 (4,500 ppm) caused increased urinary pH and high prevalence of urinary bladder calculi (from
16 the precipitation of 4-hydroxybiphenyl-O-sulphate in the urine), transitional cell hyperplasia, and
17 transitional cell tumors. Incidences of male rats with calculi and those with bladder tumors were
18 strongly correlated, and chronic exposure of male rats to lower dietary concentrations of
19 biphenyl (500 and 1,500 ppm) did not increase urinary pH and did not cause calculi formation,
20 transitional cell hyperplasia, or bladder tumor development. There were relatively strong
21 associations between incidences of rats with calculi and those with transitional cell hyperplasia
22 and between incidences of rats with transitional cell hyperplasia and bladder tumors. In contrast,
23 high concentrations of biphenyl in the diet of female rats had no effect on urinary pH, caused a
24 much lower prevalence of urinary bladder calculi of a different composition, and resulted in no
25 urinary bladder tumors. The urinary bladder calculi in the male rats were mainly composed of
26 the conjugated biphenyl metabolite, potassium 4-hydroxybiphenyl-O-sulphate, whereas those of
27 the female rats were predominantly composed of 4-hydroxybiphenyl and potassium bisulphate
28 (which are hydrolysis products of potassium 4-hydroxybiphenyl-O-sulphate). There was no
29 evidence of urinary bladder calculi formation or tumor development in male and female mice
30 exposed to similar dietary concentrations of biphenyl. Results of a tumor initiation-promotion
31 study in male rats support the proposal that biphenyl-induced sustained cell proliferation
32 promotes initiated tumor cells in the urinary bladder. Finally, results of genotoxicity tests with
33 biphenyl are predominantly negative or equivocal at best. The preponderance of evidence
34 supports a mode of action for biphenyl in male rats only involving urinary tract calculi
35 formation, urinary epithelium damage, sustained regenerative cell proliferation and hyperplasia,
36 and subsequent bladder tumor formation. There is evidence that 2,5-dihydroxybiphenyl can
37 undergo redox cycling to produce ROS that may damage DNA leading to tumor-initiating
38 mutations, but it was not detected in urine of rats exposed to oral doses of 100 mg biphenyl/kg
68 DRAFT - DO NOT CITE OR QUOTE
-------
1 and its metabolic precursor, 2-hydroxybiphenyl, is a minor urinary metabolite of biphenyl in rats
2 (Meyer and Scheline, 1976).
o
J
4 Relevance of the hypothesized mode of action to humans. Although there are no studies in
5 humans examining possible associations of biphenyl exposure with urinary bladder calculi
6 formation or cancer, urinary bladder calculi have been reported in humans following exposure to
7 other chemicals (Capen et al., 1999; Cohen, 1998, 1995). Urinary bladder calculi are, in general,
8 expected to be irritating and lead to reparative cell proliferation regardless of composition or
9 species; however, based on the anatomy of the urinary tract in humans and their upright, bipedal
10 stature, calculi are either quickly excreted in urine or cause obstruction leading to pain and
11 subsequent therapeutic removal of the calculi (Cohen, 1998, 1995). In contrast, the rodent
12 horizontal quadruped stature is expected to promote calculi residency time in the bladder without
13 causing obstruction (Cohen, 1998, 1995). In white populations, 95% of bladder tumors are
14 transitional cell carcinomas such as those found in male rats exposed to high concentrations of
15 biphenyl. Four case-control studies of urinary bladder cancer in white human populations found
16 RRs for an association between a history of urinary tract stones and bladder carcinomas ranging
17 from about 1.0 to 2.5 (Capen etal., 1999). Thus, the proposed mode of action is expected to be
18 relevant to humans at exposure levels sufficient to cause urinary bladder calculi in humans,
19 because: (1) calculi resulting from human exposure to other substances have been associated
20 with urinary bladder irritation, regeneration, and cancer (Capen etal., 1999; Cohen, 1998, 1995)
21 and (2) sulphate conjugation of hydroxylated biphenyl metabolites has been demonstrated in
22 human tissues (see Section 3.3).
23 The underlying physiological factors determining the precipitation of 4-hydroxybiphenyl-
24 O-sulphate in urine to form calculi in male rats, but not female rats, exposed to high dietary
25 biphenyl concentrations are unknown. Elevated urine pH appears to play a role in the induction
26 of urinary bladder tumors by biphenyl in the male rat (Umeda et al., 2002). Because humans on
27 average have a slightly more acidic urine than the rat (Cohen, 1995), it is possible that humans
28 might be less susceptible than the rat to the development of urinary bladder calculi. Given the
29 lack of understanding of physiological factors that influence susceptibility in rats and the absence
30 of specific human data on biphenyl-induced calculi or urinary stones, there is uncertainty in
31 extrapolation of the dose-response relationship for biphenyl-induced calculi formation in male
32 rats to humans.
33
34 Populations or lifestages particularly susceptible to the hypothesized mode of action. I ARC
35 (1999) noted that increased risks for bladder carcinoma in humans have been associated with
36 cigarette smoking, exposure to infectious agents, such as Shistosoma haematobium, causing
37 urinary tract inflammation, and a history for urinary tract infections in general. As such, people
38 with these types of exposure or history may be particularly susceptible to the formation of
69 DRAFT - DO NOT CITE OR QUOTE
-------
1 urinary calculi and urinary bladder cancer, but evidence supporting this inference is lacking. In
2 addition, there are conditions (bladder diverticuli, neurogenic bladder, and staghorn renal pelvic
3 calculi) that can increase the residency time of calculi in humans; thus, individuals with these
4 conditions may also be particularly susceptible to biphenyl-induced bladder tumors under the
5 hypothesized mode of action.
6
7 4.7.3.2. Mode-of-Action Information for Liver Tumors in Female Mice
8 Evidence that chronic oral exposure to biphenyl can cause liver tumors comes from the
9 2-year BDFi mouse bioassay by Umeda et al. (2005). Exposure to 2,000 or 6,000 ppm biphenyl
10 in the diet, but not to 667 ppm, produced increased incidences of hepatocellular adenomas or
11 carcinomas in female mice, but no carcinogenic response in male BDFi mice. Earlier studies
12 found no carcinogenic response in B6C3Fi or B6AkFi mice exposed to 517 ppm biphenyl in the
13 diet for 18 months (Innes et al., 1969; NCI, 1968) or in ddY female mice exposed to 5,000 ppm
14 biphenyl in the diet for 2 years (Imai et al., 1983). The only investigations into the mode of
15 action for biphenyl-induced liver tumors in mice involve examinations of indicators of
16 peroxisome proliferation following biphenyl exposure (Umeda et al., 2004b: Sunouchi et al.,
17 1999). Thus, a mode of action involving PPARs is proposed and an evaluation of the supporting
18 data follows.
19
20 4.7.3.2.1. Hypothesized mode of action for liver tumors in female mice. Proliferation of
21 peroxisomes is regulated by a class of ligand-activated transcription factors known as PPARs.
22 PPARa regulates induction of the peroxisome proliferation response in rodents and is thought to
23 mediate at least some of the responses for hepatocarcinogens, including initiation of cellular
24 events leading to transformation. Peroxisome proliferators (PPARa agonists) are a structurally
25 diverse group of non- or weakly mutagenic chemicals that induce a suite of responses including
26 the induction of tumors in rats and mice (Klaunig et al., 2003).
27 Klaunig et al. (2003) have proposed a mode of action for PPARa agonists involving the
28 following key events. PPARa agonists activate PPARa to transcribe genes involved in
29 peroxisome proliferation, cell cycling/apoptosis, and lipid metabolism. The changes in gene
30 expression lead to changes in cell proliferation and apoptosis, and to peroxisome proliferation.
31 Suppression of apoptosis coupled with increased cell proliferation allows transformed cells to
32 persist and proliferate, resulting in preneoplastic hepatic foci and ultimately promotion of tumor
33 growth via selective clonal expansion. Peroxisome proliferation may lead to oxidative stress,
34 which potentially contributes to the proposed mode of action by causing indirect DNA damage
35 and/or by causing cytotoxicity leading to reparative cell proliferation. PPARa agonists also
36 inhibit gap junction intercellular communication and stimulate non-parenchymal hepatic Kupffer
37 cells; these events are also thought to stimulate cell proliferation. Increases in the size and
38 number of peroxisomes and induction of peroxisome-related gene expression (e.g., palmitoyl-
70 DRAFT - DO NOT CITE OR QUOTE
-------
1 CoA oxidase and acyl-CoA oxidase) are regarded as indicators that the PPARa agonism mode of
2 action is operative.
3
4 4.7.3.2.2. Experimental support for the hypothesized mode of action for liver tumors in female
5 mice
6 Strength, consistency, and specificity of association, including support for the hypothesized
7 mode of action in mice. Support for a possible association between biphenyl-induced
8 proliferation of peroxisomes and liver tumor is limited to findings in female BDFi mice (which
9 developed liver tumors following dietary exposure to 2,000 or 6,000 ppm); male BDFi mice did
10 not develop liver tumors following exposure to concentrations as high as 6,000 ppm biphenyl.
11 Dietary exposure of female BDFi mice to 16,000 ppm biphenyl for 13 weeks induced
12 hepatocellular peroxisomes as evidenced by light microscopy detection of enlarged hepatocytes
13 filled with eosinophilic fine granules and electron microscopy confirmation that the granules
14 corresponded to increased numbers of peroxisomes (Umeda et al., 2004b). Significantly
15 increased activities were measured for potassium cyanide-insensitive palmitoyl CoA oxidation in
16 liver homogenate (up to 1.9-fold) and lauric acid 12-hydroxylation in liver microsomes (up to
17 3.8-fold) from female BDFi mice given oral doses up to 5.2 mmol/kg-day (800 mg/kg-day) for 3
18 days (Sunouchi et al.. 1999).
19 The available data do not demonstrate strong, consistent, or specific associations between
20 key events in the proposed mode of action and the development of liver tumors in female mice
21 exposed to biphenyl. Klaunig et al. (2003) proposed that an adequate data set to support a
22 PPARa agonism mode of action should meet the following criteria, most of which as noted in
23 parentheses have not been investigated for biphenyl or its metabolites: (1) activation of PPARa
24 (no data), (2) expression of peroxisomal genes including PPARa-mediated expression of cell
25 cycle, growth, and apoptosis, and nonperoxisomal lipid gene expression (no data),
26 (3) peroxisomal proliferation (limited data for biphenyl in mice as summarized in previous
27 paragraph) and perturbation of cell proliferation and apoptosis (no data for mouse liver),
28 (4) inhibition of gap junction intercellular communication (no data), (5) hepatocyte oxidative
29 stress (no data), (6) Kupffer cell-mediated events (no data), and (7) selective clonal expansion
30 (no data).
31
32 Dose-response concordance. The available data do not show concordance between the dose-
33 response relationships for liver tumors in female BDFi mice exposed for 2 years to biphenyl in
34 the diet (liver tumors at 2,000 or 6,000 ppm, but not 667 ppm) (Umeda et al., 2005) and liver
35 peroxisome proliferation, the only key event in the proposed mode of action that has been
36 investigated. Umeda (2004b) reported that, compared with controls, increased liver peroxisomes
71 DRAFT - DO NOT CITE OR QUOTE
-------
1 were detected in female BDFi mice exposed to 16,000 ppm biphenyl in the diet for 13 weeks, but
2 not in mice exposed to 500, 2,000, 4,000, 8,000, or 10,000 ppm.
o
4 Temporal relationship. Indicators of liver peroxisome proliferation were elevated in female
5 mice, but not male mice, with oral exposure durations of 3 days following exposure to 800
6 mg/kg-day (increased activities of potassium cyanide-insensitive palmitoyl CoA oxidation and
7 lauric acid 12-hydroxylation) (Sunouchi etal., 1999) and 13 weeks following exposure to 16,000
8 ppm in the diet (increased numbers of liver peroxisomes), but not at lower dietary concentrations
9 QJmeda et al.. 2004b).
10
11 Biological plausibility and coherence. The data are inadequate to evaluate the biological
12 plausibility and coherence of the proposed mode of action as it relates to liver tumors in female
13 mice exposed to biphenyl.
14
15 4.7.3.2.3. Other possible modes of action for liver tumors in mice. As discussed in
16 Section 4.5.6, the overall weight of evidence from short-term standard genotoxicity tests with
17 biphenyl and 4-hydroxybiphenyl is predominantly negative. A genotoxic mode of action for
18 biphenyl-induced liver tumors in mice could be proposed based on the large metabolic capacity
19 of the mouse liver to convert biphenyl to hydroxylated metabolites and evidence that metabolites
20 of 2-hydroxybiphenyl (2,5-dihydroxybiphenyl and 2,5'-benzoquinone) can produce DNA
21 damage (Tani et al.. 2007: Balakrishnan et al.. 2002: Sasaki et al.. 2002: Sasaki etal.. 1997:
22 Pathak and Roy, 1993: Morimoto et al., 1989). However, hydroxylation of biphenyl to produce
23 2-hydroxybiphenyl appears to be a minor metabolic pathway in mice administered single i.p.
24 doses of 30 mg biphenyl/kg (Halpaap-Wood et al., 1981b), and the available data are inadequate
25 to establish that this genotoxic mode of action operates in the biphenyl induction of liver tumors
26 in mice. There have been no in vitro or in vivo investigations of biphenyl-induced DNA adducts
27 or ROS generation in mouse liver cells or of possible gender differences in the production of
28 biphenyl-induced DNA adducts or other genotoxic events. Current mode-of-action information
29 is inadequate to provide plausible explanations for why female BDFi mice exposed to high
30 dietary concentrations of biphenyl develop liver tumors, but male BDFi mice exposed to
31 6,000 ppm and female ddY mice exposed to 5,000 ppm do not (Umeda et al., 2005: Imai et al.,
32 1983).
33
34 4.7.3.2.4. Conclusions about the hypothesized mode of action for liver tumors in mice.
35 A PPARa agonism mode of action for liver tumors in female mice exposed to 2,000 or
36 4,000 ppm biphenyl in the diet for 2 years is not adequately supported by the experimental data.
37 This is based on the lack of concordance between dose-response relationships for biphenyl-
38 induced liver tumors and proliferation of hepatocellular peroxisomes in female mice. Evidence
72 DRAFT - DO NOT CITE OR QUOTE
-------
1 for increased hepatocellular peroxisomes in female mice was only found with 13-week exposure
2 to 16,000 ppm biphenyl and not at several concentrations < 10,000 ppm (Umeda et al., 2004b).
3 Furthermore, a series of key events demonstrating PPARa agonism mode of action have not been
4 identified.
5 Available data are inadequate to support alternative modes of action that propose direct
6 or indirect genotoxic events from reactive biphenyl metabolites or ROS, respectively, as key
7 events. Results from standard short-term genotoxicity tests are mostly negative or equivocal for
8 biphenyl and 4-hydroxybiphenyl. Although there is some evidence for DNA damage from ROS
9 generated from redox cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone, there
10 are no investigations into the metabolic formation of 2-hydroxybiphenyl, 2,5-dihydroxybiphenyl,
11 and phenylbenzoquinone in livers of biphenyl-exposed mice exposed to a range of biphenyl
12 doses, no in vitro or in vivo investigations of biphenyl-induced DNA adducts or ROS generation
13 in mouse liver cells, and no investigations of possible gender differences in capability to produce
14 biphenyl-induced DNA adducts or other genotoxic events.
15
16 4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
17 4.8.1. Possible Childhood Susceptibility
18 No specific information was identified that would specifically suggest an early childhood
19 susceptibility for biphenyl toxicity. However, the developmental profiles of superoxide
20 dismutase and catalase in humans that were reported by McElroy et al. (1992) indicate that the
21 activities of both enzymes may be comparatively low before and at birth, placing humans in the
22 perinatal period at an increased risk of adverse effects elicited by quinoid metabolites of
23 biphenyl. Specifically, Buonocore et al. (2001) drew attention to the fact that the human brain
24 has relatively low superoxide dismutase activity at birth. Given the limited data on age-specific
25 ROS scavenging enzymes, any suggestions of childhood susceptibility to biphenyl is speculative.
26 Studies in animals provide evidence that biphenyl metabolism is mediated by CYP1A2
27 and CYP3 A4 (Haugen, 1981). Phase II enzymes, such as sulphotransferases (SULTs) and
28 UGTs, may be involved in conjugation activities with hydroxybiphenyls in mammalian tissues
29 (Pacifici et al., 1991; Bock et al., 1980). CYP1A2 expression is negligible in the early neonatal
30 period, but is significantly increased to 50% of adult levels by 1 year of age (Sonnier and
31 Cresteil, 1998). In general, SULTs and UGTs, depending on the isoforms, also exhibit
32 differential expression during human development (Duanmu et al., 2006; Strassburg et al., 2002).
33 To the extent that metabolism increases or reduces the toxicity of biphenyl, changes in the
34 expression of Phase I and II enzymes during development can influence susceptibility to
35 biphenyl toxicity. Specific isoforms of CYPs and Phase II enzymes have not been identified as
36 the principal catalyzers involved in biphenyl metabolism and the effect of differences in enzyme
37 expression on childhood susceptibility to biphenyl has not been established.
38
73 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.8.2. Possible Gender Differences
2 Benford and Bridges (1983) evaluated the sex- and tissue-specific induction of biphenyl
3 2-, 3-, and 4-hydroxylase activities in microsomal preparations or primary hepatocyte cultures
4 from male and female Wistar rats. No differences in biphenyl hydroxylase activities were
5 observed between the sexes. However, there were some sex differences in the way tissues
6 responded to the action of enzyme inducers. For example, the CYP1A inducer a-naphthoflavone
7 strongly induced 2-hydroxylase in male liver but had no effect on female liver. Betamethasone
8 induced 2-hydroxylase activity in female liver but inhibited it in male liver. The available
9 limited human data do not suggest that gender differences exist in the response to biphenyl
10 exposure. However, available animal data suggest gender-related differences in susceptibility to
11 tumors (i.e., bladder tumors in male, but not female, F344 rats and increased incidences of liver
12 tumors in female, but not male, BDFi mice administered biphenyl in the diet for a lifetime).
13
14 4.8.3. Other
15 The limited information on the specifics of biphenyl metabolism and toxic effects in
16 humans does not allow a meaningful assessment of populations that might be highly susceptible
17 to the adverse effects of biphenyl. For example, there is as yet no clear attribution of CYP
18 isozymes to the various biphenyl hydroxylases and no information on which sulphotransferases
19 and glucuronidases conjugate hydroxylated biphenyl metabolites. It is known that many CYP
20 isozymes, as well as glucuronidases, exist in polymorphic forms with catalytic activities that
21 differ from the wild type. In addition, such enzyme polymorphisms display specific distributions
22 across populations and ethnicities that might put some at increased risk and others at decreased
23 risk of adversity from biphenyl exposure. This lack of information represents a data gap.
74 DRAFT - DO NOT CITE OR QUOTE
-------
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Candidate Principal Studies and Candidate Critical Effects—with
Rationale and Justification
No information was located regarding possible associations between oral exposure to
biphenyl and health outcomes in humans.
As discussed in Section 4.6.1, the most sensitive targets of toxicity following oral
exposure to biphenyl are the liver, urinary system, body weight, and developing organism (see
Figure 5-1). In the rat, chronic oral studies identified the kidney and urinary bladder as critical
noncancer targets (see Figure 5-1 for LOAELs and NOAELs found in these studies). Kidney
effects observed include: renal pelvis transitional cell hyperplasia and hemosiderin deposits in
female F344 rats at doses >128 mg/kg-day and renal pelvis mineralization at 378 mg/kg-day
(Umeda et al., 2002): kidney stone formation and obstructive pyelonephritis with tubular
atrophy, tubular cysts, and fibrosis in male and female Wistar rats at 165 and 370 mg/kg-day,
respectively (Shiraiwa et al., 1989): renal lymphocytic infiltration, tubular atrophy, and tubular
cysts in male and female albino rats at doses >420 mg/kg-day (Ambrose et al., 1960): mild renal
tubular degeneration in male albino rats at 250 or 450 mg/kg-day (Pecchiai and Saffiotti, 1957:
not plotted in Figure 5-1 because quantitative data were not included in the study report); and
renal tubular dilatation in male and female Sprague-Dawley rats at 732 mg/kg-day (Dow
Chemical Co, 1953). An increased incidence of urinary bladder hyperplasia associated with
calculi or "stones" was observed in male and female F344 rats at 378 and 438 mg/kg-day, but not
at 110 and 128 mg/kg-day, respectively (Umeda et al., 2002). Elevated incidences of the same
lesion were observed in male and female Wistar rats at 353 and 370 mg/kg-day, respectively
(Shiraiwa et al., 1989). In contrast, urinary bladder hyperplasia and calculi were not observed in
male or female albino rats at doses as high as 840 mg/kg-day (Ambrose et al., 1960) or in male
or female Sprague-Dawley rats exposed to doses as high as 732 mg/kg-day (Dow Chemical Co,
1953).
75 DRAFT - DO NOT CITE OR QUOTE
-------
mnn
£? 800 -
"O
\j*
400 -
200 -
0 -I
Drop-lines span highest to lowest doses tested
|
i
A
i
:
A
i
Female Female Male Fern
BDFi ddY BDFi ddx
mouse mouse mouse mou
(1) (2) (1) (2
Liver3 Body weightb
i
ANOAEL aLOAEL
A
i
ale Female
{ BDFi
se mouse
(1)
Female
ddY
mouse
(2)
i
i
i
Female
F344 rat
(3)
1
Male
Wistar rat
(4)
'
I
i
' A
Male,
female
albino rat
(5)
I
k
Male,
female
Sprague-
Dawley rat
(6)
Kidney0
i
'
, L
i
t
Male
F344 rat
(3)
i
I
Male,
female
albino rat
(5)
Male
Wistar rat
i
i
i
Male,
female
Sprague-
Dawley rat
(6)
Urinary bladder01
i
t
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.
76
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 BDFi
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 and increased BUN levels were observed in BDFi mice
9 at >219 mg/kg-day (males) and >414 mg/kg-day (females), respectively (Umeda et al., 2005),
10 but exposure-related histological changes in the kidney were not found in female ddY mice at
11 885 mg/kg-day (Imai et al., 1983). Following the same pattern of apparent strain difference in
12 susceptibility to biphenyl toxicity, body weights were decreased by >10% at>291 mg/kg -day in
13 male BDFi mice and >414 mg/kg-day in females (Umeda et al., 2005), but body weights in
14 female ddY mice exposed to 885 mg/kg-day were similar to control values (Imai et al., 1983).
15 Shorter-duration oral exposure (13 weeks) of mice to biphenyl at higher dietary concentrations
16 (estimated doses > 1,500 mg/kg-day) has also been shown to affect body and/or liver weights in
17 mice (Umeda et al., 2004b).
18 In the only available oral developmental toxicity study (Khera et al., 1979), frank
19 maternal toxicity (increased mortality [5/20 versus 0/18 in controls] and decreased number of
20 dams with live fetuses [9/20 versus 16/18 in controls]) occurred at the highest dose
21 (1,000 mg/kg-day). Significantly increased incidences of fetuses with skeletal anomalies were
22 noted at doses >500 mg/kg-day. The NOAEL and LOAEL of 250 and 500 mg/kg-day for
23 delayed skeletal development are noted in Figure 5-1.
24 The 2-year dietary studies in F344 rats (Umeda et al., 2002) and BDFi mice (Umeda et
25 al., 2005) and the developmental study in Wistar rats (Khera et al., 1979) were selected as
26 candidate principal studies for deriving the RfD because they provide the best available data
27 (adequate number of dose groups and dose spacing, sufficient group sizes, comprehensive
28 endpoint assessment and quantitation of results) to describe dose-response relationships for the
29 critical effects in rats and mice associated with chronic or gestational exposure to biphenyl.
30 Candidate critical effects from the chronic study in F344 rats (Umeda et al., 2002) were:
31 (1) nodular or simple transitional cell hyperplasia in the renal pelvis of males and females,
32 (2) mineralization in the renal pelvis or renal papillary mineralization in males and females,
33 (3) renal hemosiderin deposits in females, and (4) transitional cell hyperplasia in the urinary
34 bladder of males. Candidate critical effects from the chronic study in BDFi mice (Umeda et al.,
35 2005) were: (1) decreased body weight in males and females, (2) mineralization of the renal
36 inner stripe-outer medulla in males and females, (3) BUN in males and females, and (4) serum
37 liver enzyme activities (AST [GOT], ALT [GPT], AP, and LDH) in females. The candidate
77 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
3
4
5
6
7
8
9
10
critical effect from the rat oral developmental toxicity study (Kheraetal., 1979) was litters with
fetal skeletal anomalies from Wistar rat dams exposed during gestation.
5.1.2. Methods of Analysis—Including Models (e.g., PBPK, BMD)
Datasets modeled included selected nonneoplastic lesions in the urinary system of male
and female F344 rats (Table 5-1) exposed to biphenyl in the diet for 2 years (Umeda et al.,
2002), mineralization in the kidney of male and female BDFi mice (Table 5-2) exposed to
biphenyl in the diet for 2 years (Umeda et al., 2005), and litters with skeletal anomalies from
Wistar rat dams (Table 5-3) administered biphenyl by gavage on GDs 6-15 (Kheraetal., 1979).
Table 5-1. Datasets employed in the BMD modeling 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).
11
78
DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-2. Datasets employed in the BMD modeling of body weight, selected
clinical chemistry results, and histopathological kidney effects in male and
female BDFi mice exposed to biphenyl in the diet for 2 years
Endpoint
Biphenyl concentration in the diet (ppm)
0
667
2,000
6,000
Males
Dose (mg/kg-d)
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
3
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.
Source: Umeda et al. (2005).
79
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: Khera et al. (1979).
1
2 Consistent with the EPA's draft Benchmark Dose Technical Guidance (U.S. EPA,
3 2000a), dose-response modeling was conducted using the U.S. EPA's benchmark dose (BMD)
4 software (BMDS, version 2.1.2.) to calculate potential points of departure (PODs) for deriving
5 the RfD by estimating the effective dose at a specified level of response (BMDx) and its 95%
6 lower bound (BMDLx).
7 All available dichotomous models in the EPA Benchmark Dose (BMD) Software
8 (BMDS) (version 2.1.2) were fit to the incidence data for each dataset. The multistage model
9 was run for all polynomial degrees up to n-1 (where n is the number of dose groups including
10 control). Adequate model fit was judged by three criteria: chi-square goodness-of-fit/>-value (p
11 > 0.1), visual inspection of the fit of the dose-response curve to the data points, and a value of <2
12 for the largest scaled residual for any data point in the dataset (including the control). Among all
13 of the models providing adequate fit to the data, the lowest BMDL (95% lower confidence limit
14 on the BMD) was selected as the potential point of departure (POD) when the difference
15 between the BMDLs estimated from these models was more than threefold; otherwise, the
16 BMDL from the model with the lowest Akaike's Information Criterion (AIC) was chosen as the
17 candidate POD. BMDs and BMDLs associated with an extra risk of 10% were calculated for all
18 models. In the absence of information to identify a biologically significant level of response for
19 an endpoint, a benchmark response (BMR) of 10% extra risk is typically chosen as an
20 appropriate response level for dichotomous data. A BMR of 10% is also recommended to
21 facilitate a consistent basis of comparison across assessments.
22 A BMR of 10% extra risk was selected to derive the POD for developmental effects from
23 the Khera et al. (1979) study because the endpoint was characterized as affected litters. A BMR
24 of 5% extra risk has typically been used for reproductive and developmental studies when data
25 are reported as affected pups within litters (U.S. EPA, 2000a). Since this level of reporting was
26 not available in Khera et al. (1979), nested models could not be used. Thus, a BMR of 10%
27 extra risk among affected litters was employed in order to better approximate a 5% extra risk in
80
DRAFT - DO NOT CITE OR QUOTE
-------
1 affected offspring and to recognize the litter as the experimental unit. BMDs and BMDLs
2 associated with extra risk of 5% for all endpoints were also calculated for comparison.
3 When standard models failed to provide adequate fit to the data, modifications of these
4 standard models (i.e., parameter restriction adjustments, specification of initial parameter values,
5 or use of alternative models) were attempted in an effort to achieve adequate fit. If these
6 modifications failed to achieve adequate fit, the highest dose was dropped, and the entire
7 modeling procedure was repeated. If an adequate fit could not be achieved after dropping the
8 highest dose, then the dataset was determined to be unsuitable for BMD modeling.
9 For continuous data, all continuous models available in the EPA BMDS (version 2.1.2)
10 were first applied to the data while assuming constant variance. If the data were consistent with
11 the assumption of constant variance (p > 0.1), then the fit of all the continuous models to the
12 mean were evaluated while assuming constant variance. In the absence of information to
13 indicate a biologically significant level of response, BMDs and BMDLs were calculated based
14 on a BMR representing a change of 1 SD from the control. BMDs and BMDLs for decreased
15 body weight were also calculated for a BMR of 10% decrease from the control (i.e., 10% relative
16 deviation [RD]) because a 10% decrease in body weight is generally considered to represent a
17 minimally biologically significant effect. For serum enzyme activities (AST, ALT, AP, LDH),
18 BMDs and BMDLs were also calculated for a BMR of 100% increase from the control (i.e.,
19 twofold or 1 RD; BMDiRD and BMDLiRo). Several expert organizations, particularly those
20 concerned with early signs of drug-induced hepatotoxicity, have identified an increase in liver
21 enzymes (AST, ALT, AP) compared with concurrent controls of two- to fivefold as an indicator
22 of concern for hepatic injury (EMEA, 2006; Boone et al., 2005). Because LDH, like liver
23 enzymes, is one of the more specific indicators of hepatocellular damage in most animal species
24 and generally parallels changes in liver enzymes in toxicity studies where liver injury occurs, a
25 similar twofold increase in LDH is considered indicative of liver injury in experimental animals.
26 A similar approach was taken for BUN.
27 Adequate model fit was judged by three criteria: goodness-of-fit^-value (p > 0.1), visual
28 inspection of the dose-response curve, and a value of <2 for the largest scaled residual for any
29 data point in the data set (including the control). Among all of the models providing adequate fit
30 to the data, the lowest BMDL was selected as the potential POD when the BMDLs estimated
31 from these models varied by more than threefold; otherwise, the BMDL from the model with the
32 lowest AIC was chosen as the candidate POD. When the test for constant variance failed, all
33 models were run again while applying the power model integrated into the BMDS to account for
34 nonhomogeneous variance. When the nonhomogeneous variance model provided an adequate fit
35 (p > 0.1) to the variance data, the models were evaluated using the nonhomogeneous variance
36 model. Model fit and POD selection proceeded as described earlier. When both tests for
37 variance (constant and nonhomogeneous) provided inadequate fit to the variance data, model
38 restriction adjustments were attempted in an effort to achieve adequate fit. If these adjustments
81 DRAFT - DO NOT CITE OR QUOTE
-------
1 failed to achieve better fit, then the highest dose was dropped and the entire modeling procedure
2 was repeated. If an adequate fit could not be achieved after dropping the highest dose, then the
3 dataset was determined to be unsuitable for BMD modeling.
4 Summary modeling results are presented in Table 5-4 and Figure 5-2; more detailed
5 modeling results are presented in Appendix C (Tables C-4 through C-24 and respective model
6 output files). The BMDs and BMDLs shown in Table 5-4 and Figure 5-2 are those from the
7 best-fitting models for each endpoint. BMDs and BMDLs for serum AST levels in female mice
8 and for serum BUN levels in male mice were derived after dropping the data from the highest
9 dose groups. For datasets to which no model could be fit, NOAELs and LOAELs were
10 considered for candidate PODs.
11
82 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
BMR
Benchmark result
(mg/kg-d)
BMD
BMDL
Females
Best fitting
model
BMR
Benchmark
result (mg/kg-d)
BMD
BMDL
F344 rats (Umeda et al., 2002); biphenyl in the diet for 2 yrs
Kidney
Renal pelvis
Transitional cell
nodular hyperplasia
Transitional cell simple
hyperplasia
Mineralization
Multistage
3 -degree
Gamma
Log-probit
10%
10%
10%
193
314
208
127
113
138
Multistage
2-degree
Gamma
Multistage
1 -degree
10%
10%
10%
274
71
88
212
52
56
Kidney - other
Hemosiderin deposit
Papillary mineralization
Not selected13
Multistage
1 -degree
10%
10%
-
92
-
58
Dichotomous-
Hill
Logistic
10%
10%
45
292
23
219
Bladder
Transitional cell
hyperplasia
Gamma
10%
205
147
Not selected13
10%
-
-
BDFi mice (Umeda etaL 2005); biphenyl in the diet for 2 yrs
Kidney
Mineralization
Log-logistic
10%
721
276
Log-logistic
10%
233
122
Clinical chemistry
AST
ALT
LDH
AP
BUN
Not selected13
Not selected13
Not selected13
Not selected13
Linear
1RD
1RD
1RD
1RD
ISO
-
-
-
-
415a
-
-
-
-
267a
Power
No adequate fit0
No adequate fit0
No adequate fit0
No adequate fit0
1RD
1RD
1RD
1RD
ISO
190a
-
-
-
-
122a
-
-
-
-
Body weight
Terminal body weight
No adequate
fit0
0.1RD
-
-
Linear
0.1RD
583
511
Wistar rats (Khera et al., 1979); biphenyl by gavage to dams on GDs 6-15
Litters with fetal skeletal anomalies
Log-logistic
10%
57
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 BMR: i.e., 10 = dose associated with 10% extra risk; 1RD = 100% RD from
control mean value; O.IRD = 10% RD from control mean value; 1SD = 1 SD from control mean value)
83
DRAFT - DO NOT CITE OR QUOTE
-------
Ann -
4 E
M
1 en .
en .
Female
mouse (1)
Terminal
Body
weight
MDL = 5
1
<
1
>
Female
mouse (1)
AST
Liver
11 mg/kc
I
|
*
i
Male
rat (2)
TC
hyperplasia
Bladder
-d
<
1 BIVIU - /z i mg/Kg-a
1
>
<
1
<
Male
mouse (1)
I
• BMD
»BMDL
1
'
1
>
Female
mouse (1)
Mineralization
Male
mouse (1)
BUN
<
1
>
1
,
,
Male
rat (2)
Female
rat (2)
Nodular hyperplasia
•
1
•
<
>
Male
rat (2)
J
Female
rat (2)
Simple hyperplasia
i
<
1
>
>
Male
rat (2)
T
Female
rat (2)
Renal pelvis
mineralization
T
Female
rat (2)
Hemo-
siderin
1
<
1
>
.1
Male Female
rat (2) | rat (2)
Papillary mineralization
Kidney
Anomalous
litters (3)
Develop-
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.
84
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)
14 and 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 In EPA's guidance document entitled, Recommended Use of Body Weight3'4 as the
20 Default Method in Derivation of the Oral Reference Dose (U.S. EPA, 2011), the Agency
21 endorses a hierarchy of approaches for converting doses administered orally to laboratory animal
22 species to human equivalent oral exposures in deriving the RfD, with the preferred approach
23 being physiologically-based toxicokinetic modeling. An alternate approach includes using
24 chemical-specific information in the absence of a complete physiologically-based toxicokinetic
25 model. In lieu of a toxicokinetic model or chemical-specific data to inform the generation of
26 human equivalent oral exposures, EPA endorses body weight scaling to the % power (i.e., BW3 4)
27 as a default to extrapolate lexicologically equivalent doses of orally administered agents from
28 laboratory animals to humans for the purpose of deriving an RfD. When BW3/4 scaling is used in
29 deriving the RfD, EPA also advocates a reduction in the interspecies uncertainty factor from 10
30 to 3, as BW3/4 scaling addresses predominantly toxicokinetic (and some toxicodynamic) aspects
31 of the UF A.
32 Statements in the guidance raise some important uncertainties in applying allometric
33 scaling, and more specifically BW3/4 scaling, when trying to extrapolate across different sized
34 individuals within a species (e.g., between neonates and adults) or across individuals in different
35 lifestages between species (e.g., between fetal rats and adult humans). Furthermore, the data on
36 which to base a default allometric scaling factor for converting the administered dose in a
37 laboratory animal in a different lifestage to a comparable dose in an adult human are sparse, and
38 thus more uncertain. For these reasons, a BW3/4 scaling factor has not been applied as a default
85 DRAFT - DO NOT CITE OR QUOTE
-------
1 approach (in combination with a reduced default UF for interspecies extrapolation) when
2 extrapolating from developmental effects in laboratory animals to adult humans when deriving
3 the RfD.
4
5 5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs)
6 Consideration of available dose-response data led to the selection of the developmental
7 study (Khera et al., 1979) and fetal skeletal anomalies in litters from biphenyl-treated pregnant
8 Wistar rats as the principal study and critical effect, respectively, for RfD derivation. The
9 uncertainty factors, selected based on EPA's A Review of the Reference Dose and Reference
10 Concentration Processes (U.S. EPA, 2002; Section 4.4.5), address five areas of uncertainty
11 resulting in a composite UF of 100. This composite uncertainty factor was applied to the selected
12 POD to derive an RfD.
13
14 • An UF of 10 was applied to account for interspecies variability in extrapolation from
15 laboratory animals (rats) to humans because information is not available to quantitatively
16 assess toxicokinetic or toxicodynamic differences between animals and humans.
17
18 • An UF of 10 was applied to account for intraspecies variability in susceptibility to
19 biphenyl, as quantitative information for evaluating toxicokinetic and toxicodynamic
20 differences among humans are not available.
21
22 • An UF of 1 was applied for use of data from a subchronic study to assess potential effects
23 from chronic exposure because the POD is from a developmental toxicity study.
24 Consistent with EPA practice (U.S. EPA, 1991), an uncertainty factor was not applied to
25 account for the extrapolation from less than chronic exposure because developmental
26 toxicity resulting from a narrow period of exposure was used as the critical effect. The
27 developmental period is recognized as a susceptible life stage when exposure during a
28 time window of development is more relevant to the induction of developmental effects
29 than lifetime exposure.
30
31 • An UF of 1 was applied for extrapolation from a LOAEL to a NOAEL because the
32 current approach is to address this factor as one of the considerations in selecting a BMR
33 for BMD modeling. In this case, a BMR of 10% increase in incidence of litters with
34 skeletal anomalies was selected under an assumption that it represents a minimal
35 biologically significant change.
36
37 • An UF of 1 to account for database deficiencies was applied. The biphenyl database
38 includes chronic toxicity studies in rats (Umeda et al., 2002; Shiraiwa et al., 1989;
39 Ambrose et al., I960; Pecchiai and Saffiotti, 1957; Dow Chemical Co, 1953) and mice
40 (Umeda et al., 2005; Imai etal., 1983): subchronic toxicity studies in rats (Shibataet al.,
41 1989b: Shibata et al.. 1989a: Kluwe. 1982: S0ndergaard and Blom. 1979: Booth et al..
42 1961) and mice (Umeda et al., 2004b): a developmental toxicity study in rats (Khera et
43 al., 1979): and one- and three-generation reproductive toxicity studies in rats (Ambrose et
44 al., 1960: Dow Chemical Co, 1953). Epidemiological studies provide some evidence that
45 biphenyl may induce functional changes in the nervous system at concentrations in
86 DRAFT - DO NOT CITE OR QUOTE
-------
1 excess of occupational exposure limits. Seppalainen and Hakkinen (1975) reported small
2 increases in anomalies in nerve conduction, EEG, and ENMG signals in workers exposed
3 to biphenyl during the production of biphenyl-impregnated paper at concentrations that
4 exceeded the occupational limit by up to 100-fold, and Wastensson et al. (2006) reported
5 a cluster of PD in a Swedish factory manufacturing biphenyl-impregnated paper. No
6 other clusters of PD have been reported in biphenyl exposed populations, and
7 Wastensson et al. (2006) acknowledged that chance is an alternative explanation for this
8 cluster. Studies in experimental animal models have not identified effects on the nervous
9 system following biphenyl exposure. Accordingly, these epidemiologic studies do not
10 suggest that the nervous system is a sensitive target of biphenyl toxicity, and therefore,
11 the lack of nervous system-specific studies is not considered a gap in the biphenyl
12 toxicity database.
13
14 The RfD for biphenyl was calculated as follows:
15 RfD = BMDLio - UF
16 = 20 mg/kg-day - 100
17 =0.2 mg/kg-day
18
19 5.1.4. Previous RfD Assessment
20 The previous IRIS assessment for biphenyl, posted to the IRIS database in 1987, derived
21 an oral RfD of 0.05 mg/kg-day based on kidney damage in albino rats administered biphenyl for
22 2 years at dietary levels >0.5% (Ambrose et al., 1960). U.S. EPA considered the dietary level of
23 0.1% (50 mg/kg-day using a food factor of 0.05/day) to represent a NOAEL due to the
24 following: (1) uncertainty in the significance of effects observed at lower doses as compared to
25 the more certain adverse effect level of 0.5% in the diet and (2) supportive findings of 0.1%
26 biphenyl as a NOAEL in an unpublished report of a subchronic rat feeding study and a three-
27 generation rat reproduction study performed by Stanford Research Institute (Dow Chemical Co,
28 1953). The NOAEL of 50 mg/kg-day was divided by a total UF of 1,000 (10 for extrapolation
29 from animals to humans, 10 for protection of sensitive human subpopulations, and a modifying
30 factor of 10 to account for intraspecies variability demonstrated in the threshold suggested by the
31 data in the chronic animal study).
32
33 5.2. INHALATION REFERENCE CONCENTRATION (RfC)
34 5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
35 Human data are limited to assessments of possible associations between occupational
36 exposure to biphenyl and health outcomes where inhalation is presumed to have been the major
37 exposure route. Clinical signs and abnormal electrophysiological test results among workers
38 exposed to biphenyl during the production of biphenyl-impregnated fruit wrapping paper provide
39 evidence of biphenyl-induced neurological effects (Seppalainen and Hakkinen, 1975; Hakkinen
40 et al., 1973; Hakkinen et al., 1971). Case reports include an account of periodic loss of strength
41 and eventual signs of chronic hepatitis in a woman during a 25-year period of employment at a
87 DRAFT - DO NOT CITE OR QUOTE
-------
1 fruit-packing facility where biphenyl-impregnated paper was used (Carella and Bettolo, 1994)
2 and a cluster of five cases of PD (0.9 cases expected) at a facility manufacturing biphenyl -
3 impregnated paper (Wastensson et al., 2006). None of these studies provided air monitoring data
4 adequate to characterize workplace exposures to biphenyl. Therefore, data from the available
5 human studies could not be used for dose-response analysis and derivation of an RfC.
6 Limited information is available regarding the effects of inhaled biphenyl in laboratory
7 animals. In mice, repeated airborne exposure to biphenyl (7 hours/day, 5 days/week for 2 weeks)
8 at concentrations as high as 54.75 ppm (345.5 mg/m3) appeared to cause no symptoms (Sun,
9 1977b). In a series of studies that included repeated inhalation exposure of rabbits, rats, and
10 mice to air containing biphenyl for periods of 68-94 days (Deichmann et al., 1947; Monsanto,
11 1946), rabbits exhibited no signs of exposure-related adverse effects at concentrations as high as
12 300 mg/m3. Irritation of mucous membranes was observed in rats at concentrations of 40 and
13 300 mg/m3. Mice were the most sensitive to inhaled biphenyl; irritation of the upper respiratory
14 tract was noted at a concentration of 5 mg/m3 (Deichmann etal., 1947; Monsanto, 1946), but
15 other biphenyl concentrations were not tested in this experiment. The limitations of a single
16 exposure level and poorly reported study details preclude the use of this study for RfC
17 derivation.
18 Repeated exposure of mice to biphenyl at vapor concentrations of 25 or 50 ppm
19 (157.75 or 315.5 mg/m3) for 13 weeks resulted in high incidences of pneumonia and tracheal
20 hyperplasia, and high incidences of congestion and edema in the lungs, liver, and kidney (Sun,
21 1977a). The following study limitations and lack of supporting data preclude the usefulness of
22 this study for deriving an RfC for biphenyl. Measured biphenyl exposure concentrations varied
23 greatly during the first half of the 13-week exposure period; for example, in the high
24 concentration group (target concentration of 50 ppm), the measured concentrations ranged from
25 5 to 102 ppm during the first 45 exposure sessions. High mortality after 46 exposures (as a result
26 of accidental overheating of the chambers) necessitated the use of 46 replacement animals; these
27 replacement animals received the same total number of exposure sessions as the surviving
28 animals from the original groups, but exposures were not concurrent. Histopathological findings
29 were reported only for males and females combined. Reports of lung congestion and
30 hemorrhagic lungs in some control mice were not confirmed histopathologically, and congestion
31 in the lung, liver, and kidney were considered by the study pathologist a likely effect of the
32 anesthetic used for killing the mice. The severity of reported histopathologic lesions was not
33 specified.
34 Given these deficiencies, the Sun Company Inc. (1977a) 13-week inhalation mouse
35 study, the only available study that employed at least subchronic-duration exposure and multiple
36 biphenyl exposure levels, is considered inadequate for RfC derivation. An RfC was not derived
37 due to the significant uncertainty associated with the inhalation database for biphenyl, and route-
38 to-route extrapolation was not supported in the absence of a physiologically based
88 DRAFT - DO NOT CITE OR QUOTE
-------
1 pharmacokinetic (PBPK) model. Although an RfC cannot be derived, it should be noted that the
2 available inhalation data provides some evidence that inhalation exposure to biphenyl could
3 induce respiratory or systemic lesions.
4
5 5.2.2. Previous RfC Assessment
6 No RfC was derived in the previous (1985) IRIS assessment.
7
8 5.3. UNCERTAINTIES IN THE RfD AND RfC
9 Risk assessments should include a discussion of uncertainties associated with the derived
10 toxicity values. To derive the oral RfD, the UF approach (U.S. EPA. 2002. 1994b) was applied
11 to a POD of 20 mg/kg-day (see Section 5.1). Factors were applied to the POD to account for
12 extrapolating from responses observed in an animal bioassay to a diverse human population of
13 varying susceptibilities. Uncertainties associated with the data set used to derive the biphenyl
14 RfD are more fully described below. The available database was determined to be inadequate
15 for deriving a chronic inhalation RfC for biphenyl (see Section 5.2).
16 Selection of the critical effect for RfD determination. The critical effect selected for
17 derivation of the RfD was skeletal anomalies in fetuses from rat dams administered biphenyl by
18 gavage during GDs 6-15. An increased incidence of these anomalies was reported at doses >500
19 mg/kg-day; frank maternal toxicity, including death, was observed at the highest dose level
20 (1,000 mg/kg-day). There is some degree of uncertainty regarding the toxicological significance
21 of the reported skeletal anomalies (wavy or extra ribs and delayed ossification most commonly
22 observed) and the influence of gavage dosing in the developmental toxicity study on human
23 exposures. Supporting developmental toxicity studies are not available.
24 Dose-response modeling. BMD modeling was used to estimate the POD for the biphenyl
25 RfD. BMD modeling has advantages over a POD based on a NOAEL or LOAEL because, in
26 part, the latter are a reflection of the particular exposure concentration or dose at which a study
27 was conducted. A NOAEL or LOAEL lacks characterization of the entire dose-response curve,
28 and for this reason, is less informative than a POD obtained from BMD modeling. The selected
29 model (i.e., the log-logistic model) provided the best mathematical fit to the experimental data
30 set (as determined by the lowest AIC), but does not necessarily have greater biological support
31 over the various other models included in BMDS. Other models in BMDS yielded estimates of
32 the POD higher than the POD derived using the log-logistic model (by up to 5.8-fold).
33 Inadequate data to support RfC derivation. The available data do not support RfC
34 derivation (see Section 5.2.1). Nevertheless, limited findings from human reports and from
35 inhalation toxicity studies in experimental animals suggest that exposure to sufficiently high
36 concentrations of biphenyl can potentially target the lungs, liver, and kidney. The lack of
37 adequate data to derive an RfC represents a significant uncertainty for the evaluation of risks
38 from exposure to inhaled biphenyl.
89 DRAFT - DO NOT CITE OR QUOTE
-------
1 5.4. CANCER ASSESSMENT
2 As noted in Section 4.7.1, EPA concluded that there is "suggestive evidence of
3 carcinogenic potential" for biphenyl. The Guidelines for Carcinogen Risk Assessment (U.S.
4 EPA, 2005a) state: "When there is suggestive evidence, the Agency generally would not attempt
5 a dose-response assessment, as the nature of the data generally would not support one; however,
6 when the evidence includes a well-conducted study, quantitative analyses may be useful for
7 some purposes, for example, providing a sense of the magnitude and uncertainty of potential
8 risks, ranking potential hazards, or setting research priorities. In each case, the rationale for the
9 quantitative analysis is explained, considering the uncertainty in the data and the suggestive
10 nature of the weight of evidence. These analyses generally would not be considered Agency
11 consensus estimates."
12 In this case, the carcinogenicity of biphenyl has been evaluated in two well-conducted 2-
13 year bioassays in rats and mice (Umeda et al., 2005, 2002) that provide evidence of increased
14 incidences of liver tumors in female BDFi mice and urinary bladder tumors in male F344 rats.
15 Considering these data and uncertainty associated with the suggestive nature of the tumorigenic
16 response, EPA concluded that quantitative analyses may be useful for providing a sense of the
17 magnitude of potential carcinogenic risk. Based on the weight of evidence, a dose-response
18 assessment of the carcinogenicity of biphenyl is deemed appropriate.
19
20 5.4.1. Choice of Study/Data—with Rationale and Justification
21 No information was located regarding possible associations between oral exposure to
22 biphenyl and cancer in humans. Two animal bioassays found statistically significant
23 associations between lifetime oral exposure to biphenyl and tumor development. Biphenyl was
24 associated with urinary bladder tumors in male, but not female, F344 rats (Umeda et al., 2002)
25 and liver tumors in female, but not male, BDFi mice (Umeda et al., 2005). Tumor data for these
26 two sites were selected for dose-response analysis.
27 No studies were identified that examined the association between inhalation exposure to
28 biphenyl and cancer in humans or animals.
29
30 5.4.2. Dose-Response Data
31 The dose-response data for urinary bladder tumor formation resulting from lifetime oral
32 exposure of male and female F344 rats (Umeda et al., 2002) are shown in Table 5-5. The dose-
33 response data for liver tumor formation resulting from lifetime oral exposure of male and female
34 BDFi mice (Umeda et al., 2005) are shown in Table 5-6. The datasets selected for dose-response
35 analysis include urinary bladder transitional cell papilloma or carcinoma (combined) in the male
36 F344 rats and liver adenoma or carcinoma (combined) in the female BDFi mice.
37
90 DRAFT - DO NOT CITE OR QUOTE
-------
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 incidence1"
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
10/49C
24/49c
31/49C
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).
bOne 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.
Significantly different from control group (p < 0.01) according to Fisher's exact test.
Source: Umeda et al. (2002).
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)
Dietary concentration of biphenyl (ppm)
Males
0
0
667
97
2,000
291
6,000
1,050
Females
0
0
667
134
2,000
414
6,000
1,420
Tumor incidence"
Adenoma
Carcinoma
Adenoma or carcinoma
8/50
8/50
16/50
6/49
8/49
12/49
7/49
5/49
9/49
3/50
4/50
7/50
2/48
1/48
3/48
3/50
5/50
8/50
12/49b
7/49b
16/49C
10/48b
5/48
14/48b
aOne low-dose, one mid-dose male, two control, one mid-dose, and two high-dose female mice were excluded from
the denominators because they died prior to week 52. It is assumed that they did not have tumors and were not
exposed for a sufficient time to be at risk for developing a tumor. Umeda et al. (2005) did not specify the time of
appearance of the first tumor.
bSignificantly different from controls (p < 0.05) according to Fisher's exact test as reported by Umeda et al. (2005).
Significantly different from controls (p < 0.01) according to Fisher's exact test as reported by Umeda et al. (2005).
Source: Umeda et al. (2005).
91
DRAFT - DO NOT CITE OR QUOTE
-------
1 5.4.3. Dose Adjustments and Extrapolation Method(s)
2 5.4.3.1. Bladder Tumors in Male Rats
3 There is strong evidence that the occurrence of urinary bladder tumors in male rats
4 chronically exposed to biphenyl in the diet is a high-dose phenomenon involving occurrence of
5 calculi in the urinary bladder leading to transitional cell damage, sustained regenerative cell
6 proliferation, and eventual promotion of spontaneously initiated tumor cells in the urinary
7 bladder epithelium (see Section 4.7.3.1 for a detailed discussion of the hypothetized mode of
8 action for urinary bladder tumors in biphenyl-exposed male rats). No increased risk of bladder
9 tumors is expected as long as the exposure to biphenyl is below the dose needed to form calculi
10 (Cohen and Ellwein, 1992). As noted in the EPA Guidelines for Carcinogen Risk Assessment
11 (U.S. EPA, 2005a), a nonlinear approach to dose-response analysis is used when there are
12 sufficient data to ascertain the mode of action and conclude that it is not linear at low doses and
13 the agent does not demonstrate mutagenic or other activity consistent with linearity at low doses.
14 Therefore, consistent with the Cancer Guidelines, a nonlinear extrapolation approach for
15 biphenyl-induced urinary bladder tumors was selected.
16 Based on the proposed mode of action, the available evidence indicates that doses below
17 the oral RfD would not result in the sequence of events that includes calculus formation,
18 consequent epithelial cell damage, and sustained regenerative cellular proliferation.
19 Accordingly, the RfD of 0.2 mg/kg-day derived for noncancer effects of biphenyl was judged to
20 be protective against an increased risk of biphenyl-induced urinary bladder cancer.
21
22 5.4.3.2. Liver Tumors in Female Mice
23 In the study report of their 2-year bioassay in BDFi mice, Umeda et al. (2005) provided
24 average food consumption and biphenyl dose estimates for each exposure group (Table 1 of
25 (Umeda et al., 2005). The study report did not include average body weights for the exposure
26 groups. Therefore, the biphenyl concentration in the food was multiplied by the corresponding
27 average daily food consumption value to determine the average daily biphenyl intake. Dividing
28 this average daily biphenyl intake by the author-calculated daily dose yielded the average body
29 weight that would have been used by the study authors to calculate the average daily biphenyl
30 dose. Scaling factors were calculated using U.S. EPA (1988) reference body weight for humans
31 (70 kg) and the average body weight for each dose group of female mice: (average body
f\ r\c
32 weight/70) = scaling factor. The human equivalent dose (HED) was calculated as:
33 HED = scaling factor x reported dose (Table 5-7).
34
92 DRAFT - DO NOT CITE OR QUOTE
-------
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
1
2
o
3
4
5
6
7
8
9
10
11
12
13
a(Biphenyl concentration in food [mg/kg food] x reported average food consumption [kg/day]) -^ reported average
daily dose of biphenyl (mg/kg-day) = calculated average mouse body weight (kg).
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).
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)
HED (mg/kg-d)
0
0
667
19
2,000
59
6,000
195
Tumor incidence
Adenoma or carcinoma (combined)
3/48a
8/50
16/49**
14/48*-°
aTwo control, one mid-dose, and two high-dose female mice were excluded from the denominators because they
died prior to week 52. It is assumed that they did not have tumors and were not exposed for a sufficient time to be
at risk for developing a tumor. Umeda et al. (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).
14
93
DRAFT - DO NOT CITE OR QUOTE
-------
1 The multistage-cancer model in the EPA BMDS (version 2.1.2), using the extra risk
2 option, was fit to the female mouse liver tumor incidence data. The multistage model has been
3 used by EPA in the vast majority of quantitative cancer assessments because it is thought to
4 reflect the multistage carcinogenic process and it fits a broad array of dose-response patterns.
5 The multistage-cancer model was run for all polynomial degrees up to n-1 (where n is the
6 number of dose groups including control). An extra risk of 10% tumor incidence was selected as
7 the BMR. A 10% response is generally at or near the limit of sensitivity in most cancer
8 bioassays, and in the case of biphenyl, corresponded to a POD near the lower end of the
9 observed range in the Umeda et al. (2005) bioassay data. Adequate model fit was judged by
10 three criteria: chi-square goodness-of-fit/>-value (p > 0.05), visual inspection of the fit of the
11 dose-response curve to the data points, and a value of <2 for the largest scaled residual for any
12 data point in the dataset (including the control). If an adequate fit to the data was not achieved
13 using the protocol above, then the other dichotomous models were fit to the data. If none of the
14 models achieved an adequate fit for the full dataset, then the highest dose was dropped and the
15 entire modeling procedure was repeated.
16 When liver tumor incidence data for all dose groups were modeled, none of the models in
17 BMDS, including the multistage model, provided an adequate fit of the data (see Appendix D,
18 Table D-2). The animals in the highest dose group, while exhibiting a statistically significantly
19 increased incidence in liver tumors compared with controls, did not show a monotonic increase
20 in tumor response compared with the responses at the lower doses. To better estimate responses
21 in the low-dose region, the high-dose group was excluded as a means of improving the fit of the
22 model in the region of interest. When the high-dose group was dropped, the multistage model
23 provided an adequate fit to the data (see Appendix D, Table D-2). The BMDnEDio and
24 BMDLnEDio using this latter dataset were 18.7 and 12.2 mg/kg-day, respectively. See
25 Appendix D for more information.
26
27 5.4.4. Oral Slope Factor and Inhalation Unit Risk
28 A low-dose linear extrapolation approach results in calculation of an oral slope factor that
29 describes the cancer risk per unit dose of the chemical at low doses. The oral slope factor was
30 calculated by dividing the risk (i.e., BMR of 10% extra risk) at the POD by the corresponding
31 BMDL (0. l/BMDLHEDio)- Using linear extrapolation from the BMDLnEDio, the human
32 equivalent oral slope factor of 8.2 x 10"3 (mg/kg-d)"1 (rounded to one significant figure,
33 8 x 10~3 (mg/kg-d)"1) was derived for liver tumors in female BDFi mice (Table 5-9).
34
94 DRAFT - DO NOT CITE OR QUOTE
-------
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
BMDLHED10
(mg/kg-d)
12.2
Slope factor" (risk per
[mg/kg-d])
8.2 x ID'3
"Human equivalent slope factor = 0.1/BMDL10HED; see Appendix C for details of modeling results.
1
2 This slope factor should not be used with exposures >12.2 mg/kg-day (the POD for this
3 dataset), because above the POD, the fitted dose-response model better characterizes what is
4 known about the carcinogenicity of biphenyl (i.e., the slope factor may not approximate the
5 observed dose-response relationship adequately at exposure exceeding 12.2 mg/kg-day).
6 An inhalation unit risk for biphenyl was not derived in this assessment. The potential
7 carcinogenicity of inhaled biphenyl has not been evaluated in human or animal studies, and
8 route-to-route extrapolation was not possible in the absence of a PBPK model.
9
10 5.4.5. Uncertainties in Cancer Risk Values
11 5.4.5.1. Oral Slope Factor
12 A number of uncertainties underlie the cancer unit risk for biphenyl. Table 5-10
13 summarizes the impact on the assessment of issues such as the use of models and extrapolation
14 approaches (particularly those underlying the Guidelines for Carcinogen Risk Assessment (U.S.
15 EPA, 2005a), the effect of reasonable alternatives, the decision concerning the preferred
16 approach, and its justification.
95
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
J, slope factor 1.5 -fold
if BMD10 used rather
thanBMDLjo
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 Guidelines 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.
BMDLi o = 95% lower confidence limits on the doses associated with a 10% extra risk of cancer incidence.
96
DRAFT - DO NOT CITE OR QUOTE
-------
1 The uncertainties presented in Table 5-10 have a varied impact on risk estimates. Some
2 suggest risks could be higher than was estimated, while others would decrease risk estimates or
3 have an impact of an uncertain direction. Several uncertainties are quantitatively characterized
4 for the significantly increased rodent tumors. These include the statistical uncertainty in the
5 multistage modeling estimate. Due to limitations in the data, particularly regarding the MOA
6 and relative human sensitivity and variability, the quantitative impact of other uncertainties of
7 potentially equal or greater impact has not been explored. As a result, an integrated quantitative
8 analysis that considers all of these factors was not undertaken.
9
10 5.4.5.2. Inhalation Unit Risk
11 The potential carcinogen!city of inhaled biphenyl has not been assessed. Therefore, a
12 quantitative cancer assessment for biphenyl by the inhalation pathway was not performed.
13
14 5.4.6. Previous Cancer Assessment
15 In the previous IRIS cancer assessment (U.S. EPA, 1991), biphenyl was listed in Group
16 D; not classifiable as to human carcinogenicity based on no human data and inadequate studies
17 in mice and rats. Neither an oral slope factor nor inhalation unit risk was derived in the previous
18 cancer assessment.
97 DRAFT - DO NOT CITE OR QUOTE
-------
1 6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
2 RESPONSE
o
J
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 etal., 1976a: Meyer etal., 1976b). Data for absorption, distribution,
11 and elimination are not available for inhaled or dermally applied biphenyl. Metabolism to a
12 range of hydroxylated metabolites has been demonstrated in in vitro systems with rat and human
13 cells and tissues. Human metabolism of biphenyl appears to be qualitatively similar to
14 metabolism in the rat, although some reports of quantitative differences are available (Powis et
15 al.. 1989: Powis etal.. 1988: 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. In chronically exposed rats,
21 nonneoplastic kidney lesions (simple transitional cell hyperplasia in the renal pelvis and
22 hemosiderin deposits) were found in females at >1,500 ppm biphenyl in the diet (128 mg/kg-
23 day), and urinary bladder tumors, associated with urinary bladder calculi and transitional cell
24 hyperplasia, were found in males, but not females, at the highest tested concentration, 4,500 ppm
25 (378 mg/kg-day) (Umeda et al., 2002). Several other rat studies provide supporting evidence
26 that the kidney and other urinary tract regions are sensitive targets for biphenyl in rats (Shiraiwa
27 et al.. 1989: Ambrose et al.. I960: Pecchiai and Saffiotti, 1957: Dow Chemical Co. 1953). In
28 chronically exposed BDFi mice, increased incidence of nonneoplastic effects on the kidney
29 (mineralization) and liver (increased activities of plasma ALT and AST) were found in females
30 exposed to >2,000 ppm biphenyl in the diet (414 mg/kg-day) (Umeda et al., 2005). In contrast,
31 no exposure-related nonneoplastic or neoplastic effects on the liver or kidney were found in
32 female ddY mice exposed to 5,000 ppm biphenyl in the diet for 2 years (Imai et al., 1983) or in
33 B6C3Fi and B6AKFi mice exposed to 517 ppm biphenyl in the diet for 18 months (Innes et al.,
34 1969: NCI, 1968). In the only available developmental toxicity study for biphenyl, increased
35 incidences of litters with fetuses showing skeletal anomalies were reported following exposure of
36 pregnant rats to gavage doses >500 mg/kg-day on GDs 6-15 (Kheraet al., 1979).
37 Biphenyl effects on reproductive function in rats have been reported at a higher exposure
38 level than the lowest exposure levels associated with urinary tract, liver, or developmental
98 DRAFT - DO NOT CITE OR QUOTE
-------
1 toxicity. No exposure-related effect on the number of dams with litters was found following
2 exposure of male and female albino rats to up to 5,000 ppm biphenyl in the diet (525 mg/kg-day)
3 for 11 or 60 days prior to mating (Ambrose et al., 1960). In a three-generation rat study,
4 decreased fertility, decreased number of pups/litter, and decreased pup body weight were
5 observed at 10,000 ppm biphenyl in the diet (947 mg/kg-day), but not at <1,000 ppm (Dow
6 Chemical Co. 1953).
7 No chronic inhalation toxicity studies in animals are available. In subchronic inhalation
8 toxicity studies, respiratory tract irritation and increased mortality following exposure to dusts of
9 biphenyl (7 hours/day, 5 days/week for up to about 90 days) were reported in mice exposed to
10 5 mg/m3 and in rats exposed to 300 mg/m3, but not in rabbits exposed to 300 mg/m3 (Deichmann
11 et al., 1947; Monsanto, 1946). Congestion or edema of the lung, kidney, and liver, accompanied
12 by hyperplasia with inflammation of the trachea, was reported in CD-I mice exposed to biphenyl
13 vapors at 25 or 50 ppm (158 or 315 mg/m3) for 13 weeks (Sun. 1977a).
14
15 6.1.2. Cancer
16 No assessments are available regarding possible associations between exposure to
17 biphenyl and increased risk of cancer in humans.
18 In a 2-year study of F344 rats administered biphenyl in the diet, significantly increased
19 incidences of urinary bladder tumors in males were observed at the highest dose level
20 (378 mg/kg-day). There is strong evidence that the occurrence of urinary bladder tumors in the
21 male rats is a high-dose phenomenon involving occurrence of calculi in the urinary bladder
22 leading to transitional cell damage, sustained regenerative cell proliferation, and eventual
23 promotion of spontaneously initiated tumor cells in the urinary bladder epithelium. Urinary
24 bladder calculi in the high-dose (438 mg/kg-day) female rats were observed at much lower
25 incidence and were different in physical appearance and chemical composition; furthermore,
26 there were no urinary bladder tumors in any of the biphenyl-exposed female rats.
27 In a 2-year study of BDFi mice administered biphenyl in the diet, the incidence of liver
28 tumors in female mice was significantly increased at doses >414 mg/kg-day, but not in males at
29 doses up to and including 1,050 mg/kg-day. Available data are insufficient to establish a mode
30 of action for liver tumors in female mice.
31 Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the
32 database for biphenyl provides "suggestive evidence of carcinogenic potential" at
33 environmentally relevant exposure levels in humans where the formation of urinary bladder
34 tumors would not be expected to occur. This cancer descriptor is based on an increase in the
35 incidence of liver tumors (hepatocellular adenomas and carcinomas) in female BDFi mice
36 (Umeda et al., 2005) and urinary bladder tumors (transitional cell papillomas and carcinomas) in
37 male F344 rats (Umeda et al., 2002) exposed to biphenyl in the diet for 104 weeks, as well as
38 information on mode of carcinogenic action.
99 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 etal.. 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 UF was applied to account for the
10 lack of information regarding the range of responses to biphenyl in the human population.
11 The overall confidence in the RfD assessment is medium to high. Confidence in the
12 principal study (Khera et al., 1979) is medium to high. The design, conduct, and reporting of this
13 developmental toxicity study in Wistar rats were adequate; however, only litter average data
14 were available that did not permit a nested analysis based on individual fetal data. Confidence in
15 the database is high. The database is robust in that it includes chronic-duration oral exposure
16 studies in several rat and mouse strains, a developmental toxicity study in Wistar rats, and one-
17 and three-generation reproductive toxicity studies in rats.
18
19 6.2.2. Noncancer/Inhalation
20 No inhalation RfC was derived due to the lack of studies of biphenyl toxicity following
21 chronic exposure and studies involving subchronic exposure that were inadequate for RfC
22 derivation. Repeated exposure of mice to biphenyl vapors for 13 weeks resulted in high
23 incidences of pneumonia and tracheal hyperplasia, and high incidences of congestion and edema
24 in the lungs, liver, and kidney (Sun, 1977a): however, study limitations and lack of supporting
25 data preclude the use of this study for deriving an RfC for biphenyl. Study limitations include
26 highly variable biphenyl exposure concentrations during the first half of the study, high mortality
27 after 46 exposures in one group of biphenyl-exposed mice due to an overheating event and
28 cannibalization that necessitated the use of replacement animals, and limitations in the reporting
29 of histopathological findings.
30
31 6.2.3. Cancer/Oral
32 The oral slope factor of 8 x 10"3 per mg/kg-day is based on the tumor response in the liver
33 of female BDFi mice exposed to biphenyl in the diet for 2 years (Umeda et al., 2005). The slope
34 factor was derived by linear extrapolation from a human equivalent BMDLio of 12.2 mg/kg-day
35 for liver adenomas or carcinomas.
100 DRAFT - DO NOT CITE OR QUOTE
-------
1 A nonlinear extrapolation approach for biphenyl-induced urinary bladder tumors in male
2 rats was used because the available mode of action information indicates that the induction of
3 urinary bladder tumors is a high-dose phenomenon involving occurrence of calculi in the urinary
4 bladder leading to transitional cell damage, sustained regenerative cell proliferation, and eventual
5 promotion of spontaneously initiated tumor cells in the urinary bladder epithelium. As long as
6 the dose is below that which is needed to form calculi, no increased risk of bladder tumors is
7 expected. Therefore, the RfD of 0.2 mg/kg-day derived for noncancer effects of biphenyl was
8 judged to be 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.
14
15
101 DRAFT - DO NOT CITE OR QUOTE
-------
7. REFERENCES
3 Abe. S: Sasaki M. (1977). Chromosome aberrations and sister chromatid exchanges in Chinese hamster cells
4 exposed to various chemicals. J Natl Cancer Inst 58: 1635-1641.
5 ACGIH. (American Conference of Governmental Industrial Hygienists). (2001). 1,1-Biphenyl. In Documentation of
6 the threshold limit values and biological exposure indices (7th ed.). Cincinnati, OH.
7 Ambrose. AM: Booth. AN: DeEds. F: Cox. AJJ. (1960). A lexicological study of biphenyl, a citrus fungistat. Food
8 Res 25: 328-336. http://dx.doi.0rg/10.llll/i.1365-2621.1960.tb00338.x.
9 Balakrishnan. S: Uppala. PT: Rupa. PS: Hasegawa. L: Eastmond. DA. (2002). Detection of micronuclei, cell
10 proliferation and hyperdiploidy in bladder epithelial cells of rats treated with o-phenylphenol. Mutagenesis
11 17: 89-93. http://dx.doi.0rg/10.1093/mutage/17.l.89.
12 Benford. DJ: Bridges. JW: Boobis. AR: Kahn. GC: Brodie. MJ: Davies. PS. (1981). The selective activation of
13 cytochrome P-450 dependent microsomal hydroxylases in human and rat liver microsomes. Biochem
14 Pharmacol 30: 1702-1703. http://dx.doi.org/10.1016/0006-2952(81)90401-9.
15 Benford. DJ: Bridges. JW. (1983). Tissue and sex differences in the activation of aromatic hydrocarbon
16 hydroxylases in rats. Biochem Pharmacol 32: 309-313. http://dx.doi.org/10.1016/0006-2952(83)90560-9.
17 Bianco. PJ: Jones. RS: Parke. DV. (1979). Effects of carcinogens on biphenyl hydroxylation in isolated rat
18 hepatocytes. Biochem Soc Trans 7: 639-641. http://dx.doi.org/10.1042/bst0070639.
19 Billings. RE: McMahon. RE. (1978). Microsomal biphenyl hydroxylation: The formation of 3- hydroxybiphenyl and
20 biphenyl catechol. Mol Pharmacol 14: 145-154.
21 Bock. KW: von Clausbruch. UC: Kaufmann. R: Lilienblum. W: Oesch. F: Pfeil. H: Platt KL. (1980). Functional
22 heterogeneity of UDP-glucuronyltransferase in rat tissues. Biochem Pharmacol 29: 495-500.
23 http://dx.doi.org/10.1016/0006-2952(80)90368-8.
24 Boehncke. A: Koennecker. G: Mangelsdorf. I: Wibbertmann. A. (1999). Concise international chemical assessment
25 document 6: Biphenyl. Geneva, Switzerland: World Health Organization.
26 http://www.who.int/ipcs/publications/cicad/en/cicad06.pdf.
27 Boone. L: Meyer. D: Cusick. P: Ennulat D: Bolliger. AP: Everds. N: Meador. V: Elliott G: Honor. D: Bounous. D:
28 Jordan. H. (2005). Selection and interpretation of clinical pathology indicators of hepatic injury in
29 preclinical studies. Vet ClinPathol 34: 182-188. http://dx.doi.Org/10.llll/i.1939-165X.2005.tb00041.x.
30 Booth. AN: Ambrose. AM: DeEds. F: Cox. AJJ. (1961). The reversible nephrotoxic effects of biphenyl. Toxicol
31 Appl Pharmacol 3: 560-567. http://dx.doi.org/10.1016/0041-008X(61)90046-l.
32 Bos. RP: Theuws. JL: Jongeneelen. FJ: Henderson. PT. (1988). Mutagenicity of bi-, tri- and tetra-cyclic aromatic
33 hydrocarbons in the "taped-plate assay" and in the conventional salmonella mutagenicity assay. Mutat Res
34 204: 203-206. http://dx.doi.org/10.1016/0165-1218(88)90090-0.
35 Boutwell RK: Bosch. DK. (1959). The tumor-promoting action of phenol and related compounds for mouse skin.
36 Cancer Res 19: 413-424.
37 Brams. A: Buchet JP: Crutzen-Favt MC: De Meester. C: Lauwerys. R: Leonard. A. (1987). A comparative study,
38 with 40 chemicals, of the efficiency of the Salmonella assay and the SOS chromotest (kit procedure).
39 Toxicol Lett 38: 123-133.
40 Brouns. RE: Foot. M: de Vrind. R: von Hoek-Kon. T: Henderson. PT: Kuyper. CMA. (1979). Measurement of
41 DNA-excision repair in suspensions of freshly isolated rat hepatocytes after exposure to some carcinogenic
42 compounds: Its possible use in carcinogenicity screening. Mutat Res Environ Mutagen Relat Subj 64: 425-
43 432. http://dx.doi.org/10.1016/0165-1161(79)90112-2.
44 Buonocore. G: Perrone. S: Bracci R. (2001). Free radicals and brain damage in the newborn. Biol Neonate 79: 180-
45 186. http://dx.doi.org/10.1159/000047088.
46 Burke. MD: Bridges. JW. (1975). Biphenyl hydroxylations and spectrally apparent interactions with liver
47 microsomes from hamsters pre-treated with phenobarbitone and 3-methylcholanthrene. Xenobiotica 5: 357-
48 376. http://dx.doi.org/10.3109/00498257509056106.
49 Capen. CC: Dybing. E: Rice. JM: Wilbourn. JD. (1999). Species differences in thyroid, kidney and urinary bladder
50 carcinogenesis. In IARC Scientific Publications (Vol. 147). Lyon, France: International Agency for
102 DRAFT - DO NOT CITE OR QUOTE
-------
1 Research on Cancer.
2 Carella. G: Bettolo. PM. (1994). Reversible hepatotoxic effects of diphenyl: Report of a case and a review of the
3 literature [Review Article]. J Occup Med 36: 575-576.
4 Charles River Laboratories. (Charles River Laboratories International Inc.). (1999). B6D2F1 (BDF1) Mouse.
5 Wilmington, MA. http://www.criver.com/en-
6 US/ProdServ/BvTvpe/ResModOver/ResMod/Pages/B6D2FlMouse.aspx.
7 Chung. KT: Adris. P. (2002). Growth inhibition of intestinal bacteria and mutagenicity of aminobiphenyls, biphenyl
8 and benzidine [Abstract]. Abstracts of the General Meeting of the American Society for Microbiology 102:
9 10.
10 Chung. KT: Adris. P. (2003). Growth inhibition of intestinal bacteria and mutagenicity of 2-, 3-, 4-aminobiphenyls,
11 benzidine, and biphenyl [Review Article]. Toxicol In Vitro 17: 145-152. http://dx.doi.org/10.1016/S0887-
12 2333(02)00131-5.
13 Cline. JC: McMahon. RE. (1977). Detection of chemical mutagens: Use of concentration gradient plates in a high
14 capacity screen. Res Comm ChemPatholPharmacol 16: 523-533.
15 Cohen. SM: Ellwein. LB. (1992). Risk assessment based on high-dose animal exposure experiments. Chem Res
16 Toxicol 5: 742-748. http://dx.doi.org/10.1021/tx00030a002.
17 Cohen. SM. (1995). Cell proliferation in the bladder and implications for cancer risk assessment. Toxicology 102:
18 149-159. http://dx.doi.org/10.1016/0300-483X(95)03044-G.
19 Cohen. SM. (1998). Cell proliferation and carcinogenesis. Drug Metab Rev 30: 339-357.
20 http://dx.doi.org/10.3109/03602539808996317.
21 Deichmann. WB: Kitzmiller. KV: Dierker. M: Witherup. S. (1947). Observations on the effects of diphenyl, O- and
22 P-aminodiphenyl, O- and P- nitrodiphenyl and dihydroxyoctachlorodiphenyl upon experimental animals. J
23 IndHyg Toxicol 29: 1-13.
24 Dow Chemical Co. (Dow Chemical Company). (1939). Toxicity of diphenyl and diphenyl oxide (sanitized). (EPA
25 Document No. 86-890001205S). Midland, MI.
26 http://www.ntis.gov/search/product.aspx? ABBR=OTS0520717.
27 Dow Chemical Co. (Dow Chemical Company). (1953). Toxicological study of diphenyl in citrus wraps with cover
28 letter. Menlo Park, CA: Stanford Research Institute.
29 http://www.ntis.gov/search/product.aspx? ABBR=OTS0206456.
30 Dow Chemical Co. (Dow Chemical Company). (1983). Partition coefficients of biphenyl, diphenyl oxide and
31 dowtherm a between 1-octanol and water-another look. (EPA/OTS Doc No. 878213735). Midland, MI.
32 http://www.ntis.gov/search/product.aspx? ABBR=OTS0206456.
33 Duanmu. Z: Weckle. A: Koukouritaki. SB: Hines. RN: Falany. JL: Falany. CN: Kocarek. TA: Runge-Morris. M.
34 (2006). Developmental expression of aryl, estrogen, and hydroxysteroid sulfotransferases in pre- and
35 postnatal human liver. J Pharmacol Exp Ther 316: 1310-1317. http://dx.doi.org/10.1124/ipet.105.093633.
36 EMEA. (European Medicines Agency). (2006). Draft guidelines on detection of early signals of drug-induced
37 hepatoxicity in non-clinical studies. London, United Kingdom: Committee for Medicinal Products for
38 Human Use.
39 Fujita. H: Kojima. A: Sasaki. M: Higara. K. (1985). [Mutagenicity test of antioxidants and fungicides with
40 Salmonella typhimurium TA97a, TA102]. Tokyo-toritsu Eisei Kenkyusho Kenkyu Nenpo 36: 413-417.
41 Garberg. P: Akerblom. E-L: Bolcsfoldi G. (1988). Evaluation of a genotoxicity test measuring DNA-strand breaks
42 in mouse lymphoma cells by alkaline unwinding and hydroxyapatite elution. Mutat Res 203: 155-176.
43 http://dx.doi.org/10.1016/0165-1161(88)90101-X.
44 Garrett. NE: Stack. HF: Waters. MD. (1986). Evaluation of the genetic activity profiles of 65 pesticides. Mutat Res
45 168: 301-325.
46 Glatt H: Anklam. E: Robertson. LW. (1992). Biphenyl and fluorinated derivatives: Liver enzyme-mediated
47 mutagenicity detected in Salmonella typhimurium and Chinese hamster V79 cells. Mutat Res 281: 151-156.
48 Gombar. VK: Borgstedt HH: Enslein. K: Hart. JB: Blake. BW. (1991). A QSAR model of teratogenesis. QSAR
49 Comb Sci 10: 306-332. http://dx.doi.org/10.1002/asar.19910100404.
50 Hakkinen. I: Vikkula. E: Hernberg. S. (1971). The clinical picture of diphenyl poisoning [Abstract]. Scand J Clin
51 Lab Invest 27: 53.
103 DRAFT - DO NOT CITE OR QUOTE
-------
1 Hakkinen. I: Siltanen. E: Hernberg. S: Seppalainen. AM: Karli. P: Vikkula. E. (1973). Diphenyl poisoning in fruit
2 paper production: A new health hazard. Arch Environ Health 26: 70-74.
3 Halpaap-Wood. K: Horning. EC: Horning. MG. (1981a). The effect of 3-methylcholanthrene, Aroclor 1254, and
4 phenobarbital induction on the metabolism of biphenyl by rat and mouse 9000g supernatant liver fractions.
5 Drug Metab Dispos 9: 103-107.
6 Halpaap-Wood. K: Horning. EC: Horning. MG. (1981b). The effect of phenobarbital and beta-naphthoflavone
7 induction on the metabolism of biphenyl in the rat and mouse. Drug Metab Dispos 9: 97-102.
8 Hanada. S. (1977). Studies on food additives, diphenyl (biphenyl) and o-phenyl phenol from the view point of public
9 health: Part 2. On the toxicities of diphenyl and o-phenyl phenol. Nagoya-shiritsu Daigaku Igakkai Zasshi
10 28: 983-995.
11 Haugen. DA. (1981). Biphenyl metabolism by rat liver microsomes: Regioselective effects of inducers, inhibitors,
12 and solvents. Drug Metab Dispos 9: 212-218.
13 Haworth. S: Lawlor. T: Mortelmans. K: Speck. W: Zeiger. E. (1983). Salmonella mutagenicity test results for 250
14 chemicals. EnvironMutagen 5: 3-142. http://dx.doi.org/10.1002/em.2860050703.
15 Hellmer. L: Bolcsfoldi G. (1992). An evaluation of the E. coli K-12 uvrB/recA DNA repair host-mediated assay: I.
16 In vitro sensitivity of the bacteria to 61 compounds. Mutat Res 272: 145-160.
17 http://dx.doi.org/10.1016/0165-1161(92)90043-L.
18 Houk. VS: Schalkowsky. S: Claxton. LD. (1989). Development and validation of the spiral Salmonella assay: An
19 automated approach to bacterial mutagenicity testing. Mutat Res 223: 49-64.
20 htto://dx.doi.org/10.1016/0165-1218(89)90062-l.
21 HSDB. (Hazardous Substances Data Bank). (2005). Biphenyl - CASRN 92-52-4 [Data Set]. Bethesda, MD: National
22 Library of Medicine. Retrieved from http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSDB
23 Hsia. MTS: Kreamer. BL: Dolara. P. (1983a). Quantitation of chemically induced DNA damage and repair in
24 isolated rat hepatocytes by a filter elution method. In Developments in the science and practice of
25 toxicology: Proceedings of the Third International Congress on Toxicology held in San Diego, California,
26 USA, August 28-September 3, 1983 (Vol. 11). New York, NY: Elsevier Science.
27 Hsia. MTS: Kreamer. BL: Dolara. P. (1983b). A rapid and simple method to quantitate chemically induced
28 unscheduled DNA synthesis in freshly isolated rat hepatocytes facilitated by DNA retention of membrane
29 filters. Mutat Res 122: 177-185. http://dx.doi.org/10.1016/0165-7992(83)90057-X.
30 IARC. (International Agency for Research on Cancer). (1999a). Melamine. In Some chemicals that cause tumours of
31 the kidney or urinary bladder in rodents and some other substances (Vol. 73). Lyon, France.
32 http://monographs.iarc.fr/ENG/Monographs/vol73/mono73.pdf.
33 IARC. (International Agency for Research on Cancer). (1999b). ortho-Phenylphenol and its sodium salt. In Some
34 chemicals that cause tumours of the kidney or urinary bladder in rodents and some other substances (Vol.
35 73). Lyon, France. http://monographs.iarc.fr/ENG/Monographs/vol73/index.php.
36 Imai S: Morimoto. J: Sekigawa. S. (1983). Additive toxicity test of thiabendazole and diphenyl in mice. Nara Igaku
37 Zasshi 34: 512-522.
38 limes. JRM: Ulland. BM: Valerio. MG: Petrucelli L: Fishbein. L: Hart. ER: Pallotta. AJ: Bates. RR: Falk. HL: Gait
39 JJ: Klein. M: Mitchell I: Peters. J. (1969). Bioassay of pesticides and industrial chemicals for
40 tumorigenicity in mice: A preliminary note. J Natl Cancer Inst 42: 1101-1114.
41 Inoue. S: Yamamoto. K: Kawanishi. S. (1990). DNA damage induced by metabolites of o-phenylphenol in the
42 presence of copper(II) ion. Chem Res Toxicol 3: 144-149. http://dx.doi.org/10.1021/tx00014a010.
43 Ishidate. MJ: Odashima. S. (1977). Chromosome tests with 134 compounds on Chinese hamster cells in vitro: A
44 screening for chemical carcinogens. Mutat Res 48: 337-353. http://dx.doi.org/10.1016/0027-
45 5107(77)90177-4.
46 Ishidate. MJ: Sofuni. T: Yoshikawa. K: Havashi. M: Nohmi. T: Sawada. M: Matsuoka. A. (1984). Primary
47 mutagenicity screening of food additives currently used in Japan. Food Chem Toxicol 22: 623-636.
48 http://dx.doi.org/10.1016/0278-6915(84)90271-0.
49 Ito. N: Fukushima. S: Shirai T: Hagiwara. A: Imaida. K. (1984). Drugs, food additives and natural products as
50 promoters in rat urinary bladder carcinogenesis. In Models, mechanisms and etiology of tumour promotion:
51 Proceedings of a symposium held in Budapest on 16-18 May 1983 (Vol. 56). Lyon, France: International
104 DRAFT - DO NOT CITE OR QUOTE
-------
1 Agency for Research on Cancer.
2 Kawachj T: Yahagi T: Kada. T: Tazima. Y: Ishidate. M: Sasaki. M: Sugiyama. T. (1980). Cooperative programme
3 on short-term assays for carcinogenicity in Japan. In Molecular and cellular aspects of carcinogen screening
4 tests: Proceedings of a meeting; June 1979; Hanover, Federal Republic of Germany (Vol. 27). Lyon,
5 France: International Agency for Research on Cancer.
6 Khera. KS; Whalen. C; Angers. G; Trivett G. (1979). Assessment of the teratogenic potential of piperonyl butoxide,
7 biphenyl, and phosalone in the rat. Toxicol Appl Pharmacol 47: 353-358. http://dx.doi.org/10.1016/0041-
8 008X(79)90330-2.
9 King-Herbert. A; Thaver. K. (2006). NTP workshop: Animal models for the NTP rodent cancer bioassay: Stocks
10 and strains - Should we switch? Toxicol Pathol 34: 802-805.
11 http://dx.doi.org/10.1080/01926230600935938.
12 Kitamura. S; Sanoh. S; Kohta. R; Suzuki. T; Sugihara. K; Fujimoto. N; Ohta. S. (2003). Metabolic activation of
13 proestrogenic diphenyl and related compounds by rat liver microsomes. J Health Sci 49: 298-310.
14 Klaunig. JE: Babich. MA: Baetcke. KP: Cook. JC: Gorton. JC: David. RM: DeLuca. JG: Lai. DY: McKee. RH:
15 Peters. JM; Roberts. RA; Fenner-Crisp. PA. (2003). PPARalpha agonist-induced rodent tumors: Modes of
16 action and human relevance [Review Article]. Crit Rev Toxicol 33: 655-780.
17 http://dx.doi.org/10.1080/713608372.
18 Kluwe. WM. (1982). Development of resistance to nephrotoxic insult: Changes in urine composition and kidney
19 morphology on repeated exposures to mercuric chloride or biphenyl. J Toxicol Environ Health 9: 619-635.
20 http://dx.doi.org/10.1080/15287398209530191.
21 Kojima. A; Hiraga. K. (1978). Mutagenicity of citrus fungicides in the microbial system. Tokyo-toritsu Eisei
22 Kenkyusho Kenkyu Nenpo 29: 83-85.
23 Kokel D; Xue. D. (2006). A class of benzenoid chemicals suppresses apoptosis in C. elegans. Chembiochem 7:
24 2010-2015. http://dx.doi.org/10.1002/cbic.200600262.
25 Kurata. Y; Asamoto. M; Hagiwara. A; Masui T; Fukushima. S. (1986). Promoting effects of various agents in rat
26 urinary bladder carcinogenesis initiated by N-butyl-N-(4-hydroxybutyl)nitrosamine. Cancer Lett 32: 125-
27 135.
28 Kwok. ES; Buchholz. BA; Vogel JS; Turteltaub. KW; Eastmond. DA. (1999). Dose-dependent binding of ortho-
29 phenylphenol to protein but not DNA in the urinary bladder of male F344 rats. Toxicol Appl Pharmacol
30 159: 18-24. http://dx.doi.org/10.1006/taap.1999.8722.
31 Maronpot RR. (2009). Biological basis of differential susceptibility to hepatocarcinogenesis among mouse strains. J
32 Toxicol Pathol 22: 11-33. http://dx.doi.org/10.1293/tox.22.11.
33 Matanoskj GM; Elliott. EA. (1981). Bladder cancer epidemiology. Epidemiol Rev 3: 203-229.
34 Matsubara. T; Prough. RA; Burke. MD; Estabrook. RW. (1974). The preparation of microsomal fractions of rodent
3 5 respiratory tract and their characterization. Cancer Res 34:2196-2203.
36 McElroy. MC; Postle. AD; Kelly. FJ. (1992). Catalase, superoxide dismutase and glutathione peroxidase activities
37 of lung and liver during human development. Biochim Biophys Acta 1117: 153-158.
38 http://dx.doi.org/10.1016/0304-4165(92)90073-4.
39 Meyer. T; Scheline. RR. (1976). The metabolism of biphenyl II Phenolic metabolites in the rat. Basic Clin
40 Pharmacol Toxicol 39: 419-432. http://dx.doi.0rg/10.llll/i.1600-0773.1976.tb03193.x.
41 Meyer. T; JChr. L; Hansen. EV; Scheline. RR. (1976a). The metabolism of biphenyl III Phenolic metabolites in the
42 pig. Basic Clin Pharmacol Toxicol 39: 433-441. http://dx.doi.org/10.1111/i. 1600-0773.1976.tb03194.x.
43 Meyer. T; Aarbakke. J; Scheline. RR. (1976b). The metabolism of biphenyl. I. Metabolic disposition of 14C-
44 biphenyl in the rat. Acta Pharmacol Toxicol 39: 412-418.
45 Meyer. T. (1977). The metabolism of biphenyl. IV. Phenolic metabolites in the guinea pig and the rabbit. Acta
46 Pharmacol Toxicol 40: 193-200. http://dx.doi.org/10.1111/i. 1600-0773.1977.tb02068.x.
47 Millburn. P; Smith. RL; Williams. RT. (1967). Biliary excretion of foreign compounds. Biphenyl, stilboestrol and
48 phenolphthalein in the rat: molecular weight, polarity and metabolism as factors in biliary excretion.
49 BiochemJ 105: 1275-1281.
50 Monsanto. (Monsanto Company). (1946). Final report on the physiological response of experimental animals to the
51 absorption of diphenyl, and several resins, elastomers and plastics. St. Louis, MO.
105 DRAFT - DO NOT CITE OR QUOTE
-------
1 Morimoto. K: Sato. M: Fukuoka. M: Hasegawa. R: Takahashi. T: Tsuchiya. T: Tanaka. A: Takahashi. A: Havashi.
2 Y^ (1989). Correlation between the DNA damage in urinary bladder epithelium and the urinary 2-phenyl-
3 1,4-benzoquinone levels from F344 rats fed sodium o-phenylphenate in the diet. Carcinogenesis 10: 1823-
4 1827. http://dx.doi.org/10.1093/carcin/10.10.1823.
5 Nakao. T: Ushiyama. K: Kabashima. J: Nagai. F: Nakagawa. A: Ohno. T: Ichikawa. H: Kobavashi. H: Hiraga. K.
6 (1983). The metabolic profile of sodium o-phenylphenate after subchronic oral administration to rats. Food
7 Chem Toxicol 21: 325-329. http://dx.doi.org/10.1016/0278-6915(83)90068-6.
8 Narbonne. JF: Cassand. P: Alzieu. P: Grolier. P: Mrlina. G: Calmon. JP. (1987). Structure-activity relationships of
9 the N-methylcarbamate series in Salmonella typhimurium. MutatRes 191: 21-27.
10 htto://dx.doi.org/10.1016/0165-7992(87)90165-5.
11 NCI. (National Cancer Institute). (1968). Evaluation of carcinogenic, teratogenic and mutagenic activities of
12 selected pesticides and industrial chemicals. Volume I. Carcinogenic study. (NCI-DCCP-CG-1973-1-1).
13 Bethesda, MD: National Institutes of Health.
14 https://ntrl.ntis.gov/search/TRLProductDetail.aspx?ABBR=PB223159.
15 Nishihara. Y. (1985). Comparative study of the effects of biphenyl and Kanechlor-400 on the respiratory and energy
16 linked activities of rat liver mitochondria. Occup Environ Med 42: 128-132.
17 Nishioka. H: Ogasawara. H. (1978). Mutagenicity testing for diphenyl derivatives in bacterial systems [Abstract].
18 Mutat Res Environ Mutagen Relat Subj 54: 248-249. http://dx.doi.org/10.1016/0165-1161(78)90102-4.
19 NRC. (National Research Council). (1983). Risk assessment in the federal government: Managing the process.
20 Washington, DC: National Academies Press.
21 Ohnishi. M:. H: Takemura. T: Yamamoto. S: Matsushima. T: Ishii. T. (2000a). Characterization of hydroxy-
22 biphenyl-O-sulfates in urine and urine crystals induced by biphenyl and KHCO3 administration in rats. J
23 Health Sci 46: 299-303.
24 Ohnishi. M: Yajima. H: Yamamoto. S: Matsushima. T: Ishii. T. (2000b). Sex dependence of the components and
25 structure of urinary calculi induced by biphenyl administration in rats. Chem Res Toxicol 13: 727-73 5.
26 http://dx.doi.org/10.1021/tx0000163.
27 Ohnishi. M: Yajima. H: Takeuchi. T: Saito. M: Yamazaki. K: Kasai. T: Nagano. K: Yamamoto. S: Matsushima. T:
28 Ishii. T. (2001). Mechanism of urinary tract crystal formation following biphenyl treatment. Toxicol Appl
29 Pharmacol 174: 122-129. http://dx.doi.org/10.1006/taap.2001.9192.
30 Pacifici. GM: Vannucci. L: Bencini. C: Tusini. G: Mosca. F. (1991). Sulphation of hydroxybiphenyIs in human
31 tissues. Xenobiotica 21: 1113-1118.
32 Pagano. G: Esposito. A: Giordano. GG: Vamvakinos. E: Quinto. I: Bronzetti. G: Bauer. C: Corsi. C: Nieri. R:
33 Ciajolo. A. (1983). Genotoxicity and teratogenicity of diphenyl and diphenyl ether: A study of sea urchins,
34 yeast, and Salmonella typhimurium. Teratog Carcinog Mutagen 3: 377-393.
35 http://dx.doi.org/10.1002/1520-6866(1990)3:4<377::AID-TCM1770030407>3.0.CO:2-6.
36 Pagano. G: Cipollaro. M: Corsale. G: Delia Morte. R: Esposito. A: Giordano. GG: Micallo. G: Quinto. I: Staiano. N.
37 (1988). Comparative toxicity of diphenyl, diphenyl ester, and some of their hydroxy derivatives. Medecine
38 Biologie Environnement 16: 291-297.
39 Parkinson. A: Ogilvie. BW. (2008). Biotransformation of xenobiotics. In Casarett & Doull's toxicology: The basic
40 science of poisons (7th ed.). New York, NY: McGraw-Hill Companies, Inc.
41 Paterson. P: Fry. JR. (1985). Influence of cytochrome P-450 type on the pattern of conjugation of 4-
42 hydroxybiphenyl generated from biphenyl or 4-methoxybiphenyl. Xenobiotica 15: 493-502.
43 http://dx.doi.org/10.3109/00498258509045023.
44 Pathak. DN: Roy. D. (1993). In vivo genotoxicity of sodium ortho-phenylphenol: Phenylbenzoquinone is one of the
45 DNA-binding metabolite(s) of sodium ortho-phenylphenol. Mutat Res-Fundam Mol Mech Mutagen 286:
46 309-319. http://dx.doi.org/10.1016/0027-5107(93)90196-M.
47 Pecchiai. L: Saffiotti. U. (1957). [Study of the toxicity of biphenyl, oxydiphenyl and their mixture (Dowtherm)].
48 Med Lav 48: 247-254.
49 Powis. G: Jardine. I: Van Dyke. R: Weinshilboum. R: Moore. D: Wilke. T: Rhodes. W: Nelson. R: Benson. L:
50 Szumlanski. C. (1988). Foreign compound metabolism studies with human liver obtained as surgical waste.
51 Relation to donor characteristics and effects of tissue storage. Drug Metab Dispos 16: 582-589.
106 DRAFT - DO NOT CITE OR QUOTE
-------
1 Powis. G: Melder. DC: Wilke. TJ. (1989). Human and dog, but not rat, isolated hepatocytes have decreased foreign
2 compound-metabolizing activity compared to liver slices. Drug Metab Dispos 17: 526-531.
3 Probst GS: McMahon. RE: Hill LE: Thompson. CZ: Epp. JK: Neal SB. (1981). Chemically-induced unscheduled
4 DNA synthesis in primary rat hepatocyte cultures: A comparison with bacterial mutagenicity using 218
5 compounds. Environ Mutagen 3:11-32. http://dx.doi.org/10.1002/em.2860030103.
6 Purchase. IFH: Longstaff. E: Ashbv. J: Styles. JA: Anderson. D: Lefevre. PA: Westwood. FR. (1978). An evaluation
7 of 6 short-term tests for detecting organic chemical carcinogens. Br J Cancer 37: 873-903.
8 http://dx.doi.org/10.1038/264624aO.
9 Reitz. RH: Fox. TR: Quast JF: Hermann. EA: Watanabe. PG. (1983). Molecular mechanisms involved in the
10 toxicity of orthophenylphenol and its sodium salt. Chem Biol Interact 43: 99-119.
11 http://dx.doi.org/10.1016/0009-2797(83)90107-2.
12 Rencuzogullari. E: Parlak. S: Ila. HB. (2008). The effects of food protector biphenyl on sister chromatid exchange,
13 chromosome aberrations, and micronucleus in human lymphocytes. Drug Chem Toxicol 31: 263-274.
14 http://dx.doi.org/10.1080/01480540701873285.
15 Sasaki. YF: Saga. A: Akasaka. M: Yoshida. K: Nishidate. E: Su. YQ: Matsusaka. N: Tsuda. S. (1997). In vivo
16 genotoxicity of ortho-phenylphenol, biphenyl, and thiabendazole detected in multiple mouse organs by the
17 alkaline single cell gel electrophoresis assay. Mutat Res Genet Toxicol Environ Mutagen 395: 189-198.
18 http://dx.doi.org/10.1016/S1383-5718(97)00168-X.
19 Sasaki. YF: Kawaguchj S: Kamava. A: Ohshita. M: Kabasawa. K: Iwama. K: Taniguchi K: Tsuda. S. (2002). The
20 comet assay with 8 mouse organs: results with 39 currently used food additives. Mutat Res 519: 103-119.
21 http://dx.doi.org/10.1016/S1383-5718(02)00128-6.
22 Schultz. TW: Sinks. GD: Cronin. MT. (2002). Structure-activity relationships for gene activation oestrogenicity:
23 Evaluation of a diverse set of aromatic chemicals. Environ Toxicol 17: 14-23.
24 http://dx.doi.org/10.1002/tox.10027.
25 Seppalainen. AM: Hakkinen. I. (1975). Electrophysiological findings in diphenyl poisoning. J Neurol Neurosurg
26 Psychiatry 38: 248-252.
27 Shibata. M-A: Tanaka. H: Yamada. M: Tamano. S: Fukushima. S. (1989a). Proliferative response of renal pelvic
28 epithelium in rats to oral administration of ortho-phenylphenol, sodium ortho-phenylphenate and diphenyl.
29 Cancer Lett 48: 19-28. http://dx.doi.org/10.1016/0304-3835(89)90198-5.
30 Shibata. MA: Yamada. M: Tanaka. H: Kagawa. M: Fukushima. S. (1989b). Changes in urine composition, bladder
31 epithelial morphology, and DNA synthesis in male F344 rats in response to ingestion of bladder tumor
32 promoters. Toxicol Appl Pharmacol 99: 37-49. http://dx.doi.org/10.1016/0041-008X(89)90109-9.
33 Shiraiwa. K: Takita. M: Tsutsumi. M: Kinugasa. T: Denda. A: Takahashi. S: Konishi. Y. (1989). Diphenyl induces
34 urolithiasis does not possess the ability to promote carcinogenesis by N-ethyl-N-hydroxyethylnitrosamine
35 inkidneys of rats. J Toxicol Pathol 2: 41-48.
36 Smith. RA: Christenson. WR: Bartels. MJ: Arnold. LL: St John. MK: Cano. M: Garland. EM: Lake. SG: Wahle. BS:
37 McNett DA: Cohen. SM. (1998). Urinary physiologic and chemical metabolic effects on the urothelial
38 cytotoxicity and potential DNA adducts of o-phenylphenol in male rats. Toxicol Appl Pharmacol 150: 402-
39 413. http://dx.doi.org/10.1006/taap.1998.8435.
40 Snyder. RD: Matheson. DW. (1985). Nick translation~a new assay for monitoring DNA damage and repair in
41 cultured human fibroblasts. Environ Mutagen 7: 267-279. http://dx.doi.org/10.1002/em.2860070304.
42 Sofuni T: Havashi. M: Matsuoka. A: Sawada. M: Hatanaka. M: M Jr. I. (1985). [Mutagenicity tests on organic
43 chemical contaminants in city water and related compounds. II. Chromosome aberration tests in cultured
44 mammalian cells]. Kokuritsu lyakuhin Shokuhin Eisei Kenkyusho Hokoku 103: 64-75.
45 Sondergaard. D: Blora L. (1979). Polycystic changes in rat kidney induced by biphenyl fed in different diets. Arch
46 Toxicol 2: 499-502.
47 Sonnier. M: Cresteil. T. (1998). Delayed ontogenesis of CYP1A2 in the human liver. Eur J Biochem 251: 893-898.
48 Strassburg. CP: Strassburg. A: Kneip. S: Barut A: Tukev. RH: Rodeck. B: Manns. MP. (2002). Developmental
49 aspects of human hepatic drug glucuronidation in young children and adults. Gut 50: 259-265.
50 http://dx.doi.0rg/10.1136/gut.50.2.259.
51 Stuehmeier. G: Legrum. W: Netter. KJ. (1982). Does cobalt pretreatment of mice induce a phenobarbitone-type
107 DRAFT - DO NOT CITE OR QUOTE
-------
1 cytochrome P-450. Xenobiotica 12: 273-282. http://dx.doi.org/10.3109/00498258209052467.
2 Sun. (Sun Company, Inc.). (1977a). 90-day inhalation toxicity study of biphenyl (99 + % purity) in GDI mice.
3 Radnor, PA.
4 Sun. (Sun Company, Inc.). (1977b). Acute inhalation toxicity of biphenyl. Radnor, PA.
5 Sunouchj M: Miyaiima. A: Ozawa. S: Ohno. Y. (1999). Effects of diphenyl on hepatic peroxysomal enzyme and
6 drug-metabolizing enzyme activities inBDF 1 mice. J Toxicol Sci 24: 333.
7 Takita. M. (1983). Urolithiasis induced by oral administration of diphenyl in rats. J Nara Med Assoc 34: 565-584.
8 Tamano. S: Asakawa. E: Boomyaphiphat P: Masui. T: Fukushima. S. (1993). Lack of promotion of N-butyl-N-(4-
9 hydroxybutyl)nitrosamine-initiated urinary bladder carcinogenesis in mice by rat cancer promoters. Teratog
10 Carcinog Mutagen 13: 89-96. http://dx.doi.org/10.1002/tcm. 1770130205.
11 Tan. Y: Yamada-Mabuchi. M: Arya. R: St Pierre. S: Tang. W: Tosa. M: Brachmann. C: White. K. (2011).
12 Coordinated expression of cell death genes regulates neuroblast apoptosis. Development 138: 2197-2206.
13 http://dx.doi.org/10.1242/dev.058826.
14 Tani. S: Yonezawa. Y: Morisawa. S: Nishioka. H. (2007). Development of a new E. coli strain to detect oxidative
15 mutation and its application to the fungicide o-phenylphenol and its metabolites. Mutat Res Genet Toxicol
16 Environ Mutagen 628: 123-128. http://dx.doi.0rg/10.1016/i.mrgentox.2006.12.006.
17 U.S. EPA. (U.S. Environmental Protection Agency). (1978). Polychlorinated Biphenyls (PCBs): Laws and
18 Regulations, http://www.epa.gov/wastes/hazard/tsd/pcbs/pubs/laws.htm.
19 U.S. EPA. (U.S. Environmental Protection Agency). (1986a). Guidelines for mutagenicity risk assessment.
20 (EPA/630/R-98/003). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
21 http://www.epa. gov/iris/backgrd.html.
22 U.S. EPA. (U.S. Environmental Protection Agency). (1986b). Guidelines for the health risk assessment of chemical
23 mixtures. (EPA/630/R-98/002). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment
24 Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=22567.
25 U.S. EPA. (U.S. Environmental Protection Agency). (1988). Recommendations for and documentation of biological
26 values for use in risk assessment. (EPA/600/6-87/008). Cincinnati, OH: U.S. Environmental Protection
27 Agency, Environmental Criteria and Assessment Office.
28 http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=34855.
29 U.S. EPA. (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk assessment.
30 (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
31 http://www.epa. gov/iris/backgrd.html.
32 U.S. EPA. (U.S. Environmental Protection Agency). (1994a). Interim policy for particle size and limit concentration
33 issues in inhalation toxicity studies. Washington, DC: U.S. Environmental Protection Agency, Health
34 Effects Division, Office of Pesticide Products.
35 http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid= 186068.
36 U.S. EPA. (U.S. Environmental Protection Agency). (1994b). Methods for derivation of inhalation reference
37 concentrations and application of inhalation dosimetry. (EPA/600/8-90/066F). Research Triangle Park, NC:
38 U.S. Environmental Protection Agency, Office of Research and Development, Office of Health and
39 Environmental Assessment, Environmental Criteria and Assessment Office.
40 http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=71993.
41 U.S. EPA. (U.S. Environmental Protection Agency). (1995). The use of the benchmark dose approach in health risk
42 assessment. (EPA/630/R-94/007). Washington, DC: U.S. Environmental Protection Agency, Risk
43 Assessment Forum, http://www.epa.gov/raf/publications/useof-bda-healthrisk.htm.
44 U.S. EPA. (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk assessment.
45 (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
46 http://www.epa.gov/raf/publications/pdfs/REPRO51.PDF.
47 U.S. EPA. (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk assessment.
48 (EPA/630/R-95/00IF). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
49 http://www.epa.gov/raf/publications/pdfs/NEUROTOX.PDF.
50 U.S. EPA. (U.S. Environmental Protection Agency). (2000a). Benchmark dose technical guidance document
51 [external review draft]. (EPA/630/R-00/001). Washington, DC: U.S. Environmental Protection Agency,
108 DRAFT - DO NOT CITE OR QUOTE
-------
1 Risk Assessment Forum, http://www.epa.gov/raf/publications/benchmark-dose-doc-draft.htm.
2 U.S. EPA. (U.S. Environmental Protection Agency). (2000b). Science policy council handbook: Risk
3 characterization. (EPA 100-B-00-002). Washington, D.C.: U.S. Environmental Protection Agency, Office
4 of Research and Development, Office of Science Policy, http://www.epa.gov/osa/spc/pdfs/rchandbk.pdf.
5 U.S. EPA. (U.S. Environmental Protection Agency). (2000c). Supplementary guidance for conducting health risk
6 assessment of chemical mixtures. (EPA/630/R-00/002). Washington, DC: U.S. Environmental Protection
7 Agency, Risk Assessment Forum, http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=20533.
8 U.S. EPA. (U.S. Environmental Protection Agency). (2002). A review of the reference dose and reference
9 concentration processes. (EPA/630/P-02/0002F). Washington, DC.
10 http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=51717.
11 U.S. EPA. (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen risk assessment.
12 (EPA/630/P-03/00IF). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
13 http://www.epa.gov/cancerguidelines/.
14 U.S. EPA. (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for assessing susceptibility
15 from early-life exposure to carcinogens. (EPA/630/R-03/003F). Washington, DC: U.S. Environmental
16 Protection Agency, Risk Assessment Forum, http://www.epa.gov/cancerguidelines/guidelines-carcinogen-
17 supplement, htm.
18 U.S. EPA. (U.S. Environmental Protection Agency). (2006a). A framework for assessing health risk of
19 environmental exposures to children. (EPA/600/R-05/093F). Washington, DC.
20 http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=158363.
21 U.S. EPA. (U.S. Environmental Protection Agency). (2006b). Science policy council handbook: Peer review.
22 (EPA/100/B-06/002). Washington, DC. http://www.epa.gov/OSA/spc/2peerrev.htm.
23 U.S. EPA. (U.S. Environmental Protection Agency). (2011). Recommended use of body weight374 as the default
24 method in derivation of the oral reference dose. (EPA/100/R11/0001). Washington, DC.
25 http://www.epa.gov/raf/publications/interspecies-extrapolation.htm.
26 Umeda. Y: Arito. H: Kano. H: Ohnishi. M: Matsumoto. M: Nagano. K: Yamamoto. S: Matsushima. T. (2002). Two-
27 year study of carcinogenicity and chronic toxicity of biphenyl in rats. J Occup Health 44: 176-183.
28 Umeda. Y: Matsumoto. M: Yamazaki. K: Ohnishi. M: Arito. H: Nagano. K: Yamamoto. S: Matsushima. T. (2004a).
29 Carcinogenicity and chronic toxicity in mice and rats administered vinyl acetate monomer in drinking
30 water. J Occup Health 46: 87-99.
31 Umeda. Y: Aiso. S: Arito. H: Nagano. K: Matsushima. T. (2004b). Short communication: Induction of peroxisome
32 proliferation in the liver of biphenyl-fed female mice. J Occup Health 46: 486-488.
33 Umeda. Y: Aiso. S: Yamazaki. K: Ohnishi. M: Arito. H: Nagano. K: Yamamoto. S: Matsushima. T. (2005).
34 Carcinogenicity of biphenyl in mice by two years feeding. J Vet Med Sci 67: 417-424.
35 Union Carbide. (Union Carbide Corporation). (1949). Range finding tests on diphenyl tables of protocols attached
36 with cover letter. (878213680). Danbury, CT: Union Carbide Corp.
37 http://www.ntis.gov/search/product.aspx? ABBR=OTS0206426.
38 Wangenheim. J: Bolcsfoldi. G. (1986). Mouse lymphoma tk+/- assay of 30 compounds [Abstract]. EnvironMutagen
39 8:90.
40 Wangenheim. J: Bolcsfoldi. G. (1988). Mouse lymphoma L5178Y thymidine kinase locus assay of 50 compounds.
41 Mutagenesis 3: 193-205. http://dx.doi.0rg/10.1093/mutage/3.3.193.
42 Wastensson. G: Hagberg. S: Andersson. E: Johnels. B: Barregard. L. (2006). Parkinson's disease in diphenyl-
43 exposed workers-- A causal association? Parkinsonism Relat Disord 12: 29-34.
44 http://dx.doi.0rg/10.1016/i.parkreldis.2005.06.010.
45 Waters. MD: Sandhu. SS: Simmon. VF: Mortelmans. KE: Mitchell. AD: Jorgenson. TA: Jones. DC: Valencia. R:
46 Garrett. NE. (1982). Study of pesticide genotoxicity. In Genetic toxicology an agricultural perspective
47 (Vol. 21). New York, NY: Plenum Press.
48 Westinghouse Electric Corporation. (1977). Potential carcinogenicity testing of PCB replacements using the Ames
49 test with cover letter. (OTS0206616). Pittsburgh, PA.
50 http://www.ntis.gov/search/product.aspx? ABBR=OTS0206616.
51 Wiebkin. P: Fry. JR: Jones. CA: Lowing. R: Bridges. JW. (1976). The metabolism of biphenyl by isolated viable rat
109 DRAFT - DO NOT CITE OR QUOTE
-------
1 hepatocytes. Xenobiotica 6: 725-743. http://dx.doi.org/10.3109/00498257609151390.
2 Wiebkin. P; Fry. JR: Jones. CA: Lowing. RK: Bridges. JW. (1978). Biphenyl metabolism in isolated rat
3 hepatocytes: effect of induction and nature of the conjugates. Biochem Pharmacol 27: 1899-1907.
4 http://dx.doi.org/10.1016/0006-2952(78)90003-5.
5 Wiebkin. P: Schaeffer. BK: Longnecker. PS: Curphev. TJ. (1984). Oxidative and conjugative metabolism of
6 xenobiotics by isolated rat and hamster acinar cells. Drug Metab Dispos 12: 427-431.
7 Williams. G. (1980). DNA repair and mutagenesis in liver cultures as indicators in chemical carcinogen screening.
8 In Mammalian cell transformation by chemical carcinogens (Vol. 1). Princeton Junction, NJ: Senate Press.
9 Williams. GM: Mori. H: McQueen. CA. (1989). Structure-activity relationships in the rat hepatocyte DNA-repair
10 test for 300 chemicals. Mutat Res 221: 263-286. http://dx.doi.org/10.1016/0165-1110(89)90039-0.
11 Yoshida. S: Masubuchj M: Hiraga. K. (1978). Cytogenetic studies of antimicrobials on cultured cells. Tokyo-toritsu
12 Eisei Kenkyusho Kenkyu Nenpo 29: 86-88.
13
14
15
16
110 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]
6
7
A-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 APPENDIX B. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
2 MODE OF ACTION
o
3
4 B.I. EFFECTS ON THE URINARY TRACT OF RATS
5 Urinary tract effects in male rats chronically exposed to biphenyl in the diet are
6 associated with the formation of urinary bladder calculi. Mechanistic studies performed by
7 Ohnishi and coworkers (2001; 2000a: 2000b) were designed to identify urinary metabolites of
8 biphenyl, to assess conditions leading to calculi formation, and to determine the composition of
9 urinary crystals and calculi. Ohnishi et al. (2000a) identified sulphate conjugates of mono- and
10 dihydroxy biphenyl metabolites in urine and urinary crystals from F344 rats treated with
11 biphenyl and KHCCb (to elevate the pH and K+ concentration of the urine). Male F344 rats (five
12 per group) were administered a diet containing 1.6% biphenyl and 5% potassium bicarbonate for
13 7 days (Ohnishi et al., 2000a). Urine was collected on days 6 and 7 and pooled. Urinary crystals
14 (i.e., precipitates) were collected, dissolved in acetonitrile, and analyzed by HPLC to identify
15 metabolites or by inductively coupled plasma spectroscopy to identify inorganic elements. As
16 shown in Table B-l, biphenyl sulphate conjugates in the urine consisted primarily of
17 3,4-dihydroxybiphenyl-3-O-sulphate (40.9% of the total biphenyl sulphate conjugates) and
18 3-hydroxybiphenyl (23.4%). No bisulphates were observed (Ohnishi et al., 2000a). In contrast,
19 about 90% of sulphate conjugates in urinary crystals were 4-hydroxybiphenyl-O-sulphate, and
20 only 3.9 and 1.06% were 3,4-dihydroxybiphenyl-3-O-sulphate and 3-hydroxybiphenyl,
21 respectively. In a follow-up study, Ohnishi et al. (2000b) evaluated the composition of urinary
22 calculi in male and female rats exposed to 4,500 ppm biphenyl in the diet for 104 weeks.
23 Urinary calculi in chronically exposed male rats were composed mainly of
24 4-hydroxybiphenyl-O-sulphate, whereas calculi in female rats were composed primarily of
25 4-hydroxybiphenyl and potassium sulphate, the hydrolysis products of 4-hydroxybiphenyl-
26 O-sulphate (Ohnishi et al.. 2000b).
27
B-l DRAFT - DO NOT CITE OR QUOTE
-------
Table B-l. 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 -Hydroxybiphenyl-O-sulphate
4-Hydroxybiphenyl-O-sulphate
4,4 ' -Dihydroxybiphenyl-O-sulphate
2,5 -Dihydroxybiphenyl-O-sulphate
3 ,4 -Dihy droxybiphenyl-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
aThe component fraction (%) for each of the sulphate conjugates was estimated from the ratio of the liquid
chromatography tandem MS peak area of the sulfate to the total area.
Source: Ohnishi et al. (2000a).
1
2 Using the same experimental protocol as that described in Ohnishi et al. (2000a), but
3 adding potassium bicarbonate (5%), potassium chloride (5%), or sodium bicarbonate (8%) to the
4 diet for 13 weeks, Ohnishi et al. (2001) reported hydronephrosis and blood in the urine only in
5 those animals receiving biphenyl plus potassium bicarbonate. Feed consumption was not
6 affected by the dietary additions, while water intake was greatly increased in all groups of
7 animals that received biphenyl and/or salts. Neither high urinary potassium levels alone, as
8 induced by co-feeding of potassium chloride, nor high urinary pH alone, as induced by co-
9 feeding of sodium bicarbonate, were sufficient to cause kidney damage. It was concluded that a
10 combination of high urinary pH and high potassium levels was necessary to cause precipitation
11 of biphenyl sulphate. It was proposed that the crystalline precipitate caused obstruction that led
12 to hydronephrosis or damaged the transitional epithelium in the bladder causing hyperplasia.
13
14 B.2. EFFECTS ON THE LIVER OF MICE
15 Based on findings of biphenyl-induced liver tumors in female BDFi mice administered
16 high dietary concentrations of biphenyl for 2 years (Umeda et al., 2005) (see Section 4.2.1.2.2), a
17 13-week oral study was performed to assess whether peroxisome proliferation might be induced
18 (Umeda et al., 2004b). Groups of male and female BDFi mice (10/sex/group) were administered
19 biphenyl in the diet at six different concentrations ranging from 500 to 16,000 ppm. Biphenyl
20 concentrations >8,000 ppm resulted in significantly decreased final body weights of males and
21 females. Significantly increased liver weights were noted in the 8,000 and 16,000 ppm groups of
22 female mice. Evidence of peroxisome proliferation was restricted to the 16,000 ppm group of
23 female mice and included light microscopy findings of clearly enlarged hepatocytes filled with
24 eosinophilic fine granules and electron microscopy confirmation that the granules corresponded
B-2
DRAFT - DO NOT CITE OR QUOTE
-------
1 to increased numbers of peroxisomes. Light microscopy of livers from rats exposed to
2 concentrations <8,000 ppm showed no indications of proliferation of peroxisomes. There were
3 no indications of other biphenyl-induced liver effects in any of the groups of male mice.
4
5 B.3. ESTROGENIC EFFECTS
6 Several biphenyl derivatives display estrogenic activity. Schultz et al. (2002) used the
7 Saccharomyces cerevisiae/LacZ reporter assay to study the estrogenic activity of 120 chemicals
8 to identify chemical structures that impart estrogenic activity to a molecule. Chemicals without a
9 hydroxy group, among them biphenyl, were inactive in this assay. The estrogenic activities of
10 biphenyl metabolites in this assay were 4,4'-dihydroxybiphenyl (median effective concentration
11 = (2.6 x 10'7 M) > 4-hydroxybiphenyl (1.2 x 10'6 M) > 3-hydroxybiphenyl (9.2 x 10'6 M)
12 > 2-hydroxybiphenyl (1.8 x 10"5 M). Estrogenic activities of the corresponding hydroxylated di-,
13 tri-, or tetrachlorobiphenyl metabolites were approximately two orders of magnitude higher,
14 provided there were no chlorines and hydroxy groups on the same ring.
15 Kitamura et al. (2003) used MCF-7 cells transfected with an estrogen receptor-luciferase
16 reporter construct to test biphenyl and its metabolites for estrogenic activity. The starting point
17 for this investigation was the structural similarity between hydroxylated metabolites of biphenyl
18 and of 2,2-diphenyl propane, the 4,4'-dihydroxy metabolite of which is bisphenol A, a known
19 endocrine disrupter. Biphenyl per se displayed no estrogenic activity in this assay. Metabolites
20 of biphenyl formed by liver microsome preparations were identified after solvent extraction from
21 reaction media by HPLC-MS. The compounds were also tested in an in vitro competitive
22 estrogen receptor binding assay. The biphenyl metabolites, 2-, 3-, 4-hydroxybiphenyl, and
23 4,4'-dihydroxybiphenyl, all exhibited estrogenic activity when the cell culture contained
24 microsomes from 3-methylcholanthrene-induced rat livers and to a lesser extent, phenobarbital-
25 induced rat livers, in the presence of NADPH. In the competitive estrogen receptor binding
26 assay, 4,4'-dihydroxybiphenyl displayed weak binding affinity, while biphenyl and its
27 monohydroxy metabolites did not show any activity. 4,4'-Dihydroxybiphenyl is one of two
28 major biphenyl metabolites in rats and mice (Halpaap-Wood et al., 198la, b; Meyer and
29 Scheline, 1976), suggesting that high doses of biphenyl, in the form of this metabolite, might
30 induce some minor estrogenic effect.
31
32 B.4. EFFECTS ON APOPTOSIS
33 Kokel and Xue (2006) tested a series of benzenoid chemicals (including mesitylene,
34 cyclohexane, benzene, toluene, and biphenyl) for their ability to suppress apoptosis in the
35 nematode, Caenorhabditis elegans, a model suitable for the characterization of carcinogens that
36 act by way of apoptosis inhibition. The study included wild type and three strains of C. elegans
37 mutants; the ced-3(n2438) mutant (which carries a partial loss-of-function mutation in the ced-
B-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 3 gene), the ced-3(n2273) mutant (also partly defective in cell death), and the ced-(n2433)
2 mutant (a strong loss-of-function ced-3 mutant). Effects on apoptosis were assessed by counting
3 the numbers of cells that should have died during embryogenesis, but inappropriately survived.
4 The results indicated that these chemicals did not significantly affect apoptosis in wild type
5 C. elegans. However, inhibition of apoptosis was apparent in mutant strains ced-3(n2438) and
6 ced-3(n2273) exposed to benzene, toluene, or biphenyl. The study authors interpreted these
7 results as indicative of apoptosis-inhibitory activity that does not depend on mutations in a
8 specific cell-death gene. A lack of apparent apoptosis-inhibitory activity in the strong loss-of-
9 function ced-3(n2433) mutant was interpreted as indicative that inhibition of apoptosis, rather
10 than transformation of cell fates, caused the increase in extra cell observed in the other two
11 mutant strains. All three chemicals also displayed embryotoxicity. Biphenyl and naphthalene
12 were both shown to suppress apoptosis in C. elegans mutant strain ced-3(n2438) by causing
13 overexpression of the CED-3 caspase. The authors proposed that benzenoid chemicals that can
14 form quinones suppress apoptosis in C. elegans via this reactive intermediate, although this was
15 proven only for benzene, toluene, and naphthalene.
16 Regulation of apoptosis during embryogenesis is critical, and a recent study by Tan et al.
17 (2011) showed that inhibition of apoptosis during this stage of development may have
18 detrimental effects on the nervous system. No literature was identified, however, that
19 specifically supports an association between inhibition of apoptosis by biphenyl and effects on
20 embryogenesis.
21
22 B.5. MITOCHONDRIAL EFFECTS
23 Nishihara (1985) assessed the effects of biphenyl on the respiratory and energy linked
24 activities of rat liver mitochondria that had been isolated from male Wistar rats. Biphenyl (5-
25 60 ug/mL in acetone solvent) was added to liver mitochondria and effects on rates of succinate
26 oxidation and a-ketoglutarate/malate oxidation were assessed by measuring oxygen
27 consumption. Solvent controls were included in the study. Biphenyl significantly inhibited
28 state 3 respiration at concentrations >20 ug/mL. The inhibition was greater for
29 a-ketoglutarate/malate oxidation than for succinate oxidation. State 4 respiration was
30 significantly stimulated by biphenyl; the effect was greater in magnitude for succinate than for
31 a-ketoglutarate/malate oxidation. Biphenyl also altered mitochondrial membrane permeability,
32 as evidenced by the instantaneous release of endogenous K+, leading to instantaneous dissipation
33 of the mitochondrial membrane potential. Inhibition of state 3 respiration is generally considered
34 to reflect an interference with electron transport. The study author suggested that the biphenyl-
35 induced stimulation of state 4 respiration may be explained by an uncoupling action on
36 respiration.
37
B-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 B.6. GENOTOXICITY
2 Biphenyl. The results of genotoxicity studies of biphenyl are summarized in Table B-2.
3 Reverse mutation assays using Salmonella typhimurium and Escherichia coli provide
4 consistently negative results both with and without the addition of a mammalian metabolic
5 activation system (rat S9 mix). Biphenyl was not genotoxic in a host-mediated deoxyribonucleic
6 acid (DNA) repair assay of E. coli in the presence of S9 (Hellmer and Bolcsfoldi, 1992). In rec
7 assays of Bacillus subtilis, two studies reported negative results both with and without S9
8 (Garrett et al., 1986; Kojima and Hiraga, 1978), one study reported negative results without S9
9 (Kawachi etal., 1980), and one study reported equivocal results with S9 (Hanada, 1977).
10 Biphenyl was reported to induce mitotic recombination both with and without S9 in
11 Saccharomyces cerevisiae strain D3 (1988), but not in S. cerevisiae strain Diploid D7 (Garrett et
12 al., 1986).
13
B-5 DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. Genotoxicity test results for biphenyl
Organism
Strain or test
system
Endpoint
Test substance
concentrations
Metabolic
activation"
+S9
-S9
Reference
Bacterial and prokaryotic 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 WP2tiyTicr~
B/y WP2tryTicr~
K-12 uvrB/recA+
K-12 uvrB/recA"
Mutation
Mutation
Host-mediated
DNA repair
Not specified
Not specified
Not specified
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
Bos et al. (1988)
Purchase et al. (1978)
Kawachi et al. (1980)
Brams et al. (1987)
Narbonne et al. (1987)
Chung and Adris
(2003. 2002)
Cline and McMahon
(1977)
Garrett et al. (1986):
Waters et al. (1982)
Glatt et al. (1992)
Hanada (1977)
Kojima and Hiraga
(1978)
Haworth et al. (1983)
Houk et al. (1989)
Ishidate et al. (1984)
Pagano et al. (1988:
1983)
Probst et al. (1981)
Westinghouse (1977)
Brams et al. (1987)
Cline and McMahon
(1977)
Probst et al. (1981)
Garrett et al. (1986)
Hanada (1977)
Kojima and Hiraga
(1978)
Hellmer and Bolcsfoldi
(1992)
B-6
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. 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
Not specified
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)
Garrett et al. (1986)
Hanada (19771
Kojima and Hiraga
(1978)
Garrett et al. (1986)
Pagano et al. (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
SCEs
CAs
Cell
transformation
CAs
SCEs
CAs
Micronuclei
Cell
transformation
DNA repair
UDS
UDS
Excision
repair
DNA repair
HGPRT
mutation
5-100 ug/mL
100 ug/mLb
0.1-lmM
Not specified
Not specified
Up to 25 ug/mL
Up to 60 ug/mL
75-125 ug/mL
0.1-lmM
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
-
-
-
-
-
—
Glatt et al. (1992)
Abe and Sasaki (1977)
Kawachi et al. (1980)
Kawachi et al. (1980)
Ishidate et al. (1984)
Ishidate and Odashima
(1977)
Sofuni et al. (1985)
Abe and Sasaki (1977)
Purchase et al. (1978)
Glatt et al. (1992)
Yoshida et al. (1978)
Rencuzogullari et al.
(2008)
Purchase et al. (1978)
Purchase et al. (1978)
Snyder and Matheson
(1985)
Garrett et al. (1986):
Waters et al. (1982)
Hsia et al. (1983a. b)
Probst et al. (1981)
Brouns et al. (1979)
Williams et al. (1989)
Williams (1980)
B-7
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. 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
10 uMa
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
SCEs
CAs
DNA damage,
Comet assay
DNA damage,
Comet assay
Mutation
Not specified
Not specified
10-2,000 mg/kg
2,000 mg/kg
Not specified
-
-
+
+
-
Kawachi et al. (1980)
Sasaki et al. (2002)
Sasaki et al. (1997)
Kawachi et al. (1980)
aLowest concentration resulting in cytotoxicity.
bLowest concentration resulting in precipitation.
'Highest concentration not causing cytotoxicity.
dPositive result only at cytotoxic concentrations.
CA = chromosomal aberration; CHL = Chinese hamster lung; CHO = Chinese hamster ovary;
HGPRT = hypoxanthine guanine phosphoribosyl transferase; NT = not tested; +/- = weakly positive or equivocal
result; empty cell = no information available; SCE = sister chromatid exchange; UDS = unscheduled DNA
synthesis
1
2 Assays of biphenyl-exposed cultured mammalian cells provide mixed genotoxicity
3 results. In the absence of exogenous metabolic activation, biphenyl produced negative results for
4 sister chromatid exchanges (SCEs) and/or chromosomal aberrations (CAs) in the DON Chinese
5 hamster cell line (Abe and Sasaki, 1977) or Chinese hamster lung (CHL) fibroblasts (Sofuni et
6 al.. 1985: Kawachi et al.. 1980): unscheduled DNA synthesis (UDS) (Garrett et al.. 1986: Hsia et
7 al.. 1983a. b) (Waters et al.. 1982: Probst etal.. 1981). excision repair (Brovms et al.. 1979). and
8 DNA repair in rat hepatocytes (Williams et al., 1989): and hypoxanthine guanine phosphoribosyl
9 transferase (HGPRT) mutation in rat immortalized liver epithelial cells (Williams, 1980) In the
10 presence of S9 mix, biphenyl produced negative results for CAs in CHL fibroblasts (Ishidate et
11 al., 1984: Ishidate and Odashima, 1977) or Chinese hamster ovary (CHO) cells (Yoshida et al.,
12 1978): DNA repair in human HSBP diploid lung fibroblasts (Snyder and Matheson. 1985): cell
B-S
DRAFT - DO NOT CITE OR QUOTE
-------
1 transformations in Chinese hamster kidney cells (Purchase et al., 1978) and human diploid lung
2 fibroblasts (Purchase et al., 1978): and UDS in human lung WI-38 lung fibroblasts (with or
3 without S9) (Garrett et al.. 1986).
4 Positive results were obtained for CAs in CHL fibroblasts (Sofuni et al., 1985) and
5 mutations and transformations in Chinese hamster V79 cells (Glatt et al., 1992) in the presence,
6 but not absence, of S9. Biphenyl induced forward mutations in mouse L5178Y/TK+/" lymphoma
7 cells with and without S9 (Wangenheim and Bolcsfoldi, 1988, 1986): another study provided
8 similar results in the presence, but not the absence, of S9 (Garberg et al., 1988). Significant
9 increases in SCEs (less than twofold higher than solvent controls), CAs (two- to fourfold higher
10 than solvent controls), and micronuclei (approximately 2.5-fold higher than solvent controls)
11 were reported in human peripheral blood lymphocytes exposed to biphenyl for 24-48 hours at
12 concentrations >50 |iL/mL (Rencuzogullari et al., 2008).
13 Evaluations of the potential genotoxicity of biphenyl in vivo have been performed in rats,
14 mice, and silkworms. Biphenyl did not induce SCEs or CAs in bone marrow cells of rats or
15 mutations in silkworms, but limited information is available for these studies (Kawachi et al.,
16 1980). In a Comet assay, positive results were reported for DNA damage in stomach, blood,
17 liver, bone marrow, kidney, bladder, lung, and brain cells of CD-I mice administered single
18 doses of 2,000 mgbiphenyl/kg (Sasaki et al., 2002: Sasaki et al., 1997). It is unknown if the
19 DNA damage was caused by direct reaction with biphenyl or its metabolites, or by indirect
20 damage from cytotoxicity or ROS generated from redox cycling of hydroquinone metabolites.
21 Biphenyl metabolites. Table B-3 summarizes results from genotoxicity tests of several
22 biphenyl metabolites, 2-hydroxybiphenyl (also known as o-phenylphenol), 4-hydroxybiphenyl
23 (the principal metabolite of biphenyl), and 2,5-dihydroxybiphenyl. 2-Hydroxybiphenyl and its
24 sodium salt have received the most research attention because they are used as fungicides and
25 anti-bacterial agents and have been found to cause urinary bladder tumors in male F344 rats with
26 chronic exposure to high concentrations in the diet (Balakrishnan et al., 2002: Kwok et al., 1999)
27 for review).
28
B-9 DRAFT - DO NOT CITE OR QUOTE
-------
Table B-3. 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/y WP2try~
WP2 lacking
catalase and
superoxide
dismutase
WP2, WP2 uvrA~,
CM571, WP100
Not given
H17 (rec+)
M45 (reel
CHL
CHO
Liver DNA
Mutation
Streptomycin
resistance
mutation
DNA repair
Rec assay
CAs
DNA adducts,
[32P]-post
labeling
method
Not specified
3.3-250 ug/plate
1-1,000 ug/plate
1-100 ug/plate
Up to 0.5 mg/plate
Not specified
Units provided in
Japanese
1-1,000 ug/mL
1,000 ug/mLa
0-10 uM
Not specified
10-10,000 mg/plate
Units provided in
Japanese
Not specified
Up to 0.05 mg/niL
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
Kawachi et al. (1980)
Haworthetal. (1983)
Kojima and Hiraga (1978)
Fujita et al. (1985)
Ishidate et al. (1984)
Nishioka and Ogasawara
(1978)
Hanada (1977)
Kojima and Hiraga (1978)
Tani et al. (2007)
Nishioka and Ogasawara
(1978)
Kawachi et al. (1980)
Kojima and Hiraga (1978):
Hanada (1977)
Kawachi et al. (1980)
Ishidate et al. (1984)
Yoshida et al. (1978)
Pathak and Roy (1993)
2-Hydroxybiphenyl in vivo tests
Rat
Bone marrow
F344/bladder
epithelium
SCEs
Micronuclei
Hyperdiploidy/
hypodiploidy
Cell
proliferation
Not specified
2,000 ppm in diet,
14 days
-
+
+
Kawachi et al. (1980)
Balakrishnan et al. (2002)
B-10
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-3. Genotoxicity test results for biphenyl metabolites
Organism
Rat
Mouse
Mouse
Mouse
Mouse
Mouse
Rat
Rat
Silkworm
Strain or test
system
F344/bladder
epithelium
CD-I/stomach,
colon, liver,
kidney, bladder,
lung
CD-1/brain, bone
marrow
CD-I/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
Not specified
Metabolic
activation"
+S9 -S9
+
+
-
+
-
+
+
-
Reference
Morimoto et al. (1989)
Sasaki et al. (2002)
Sasaki et al. (2002)
Sasaki et al. (1997)
Sasaki et al. (1997)
Pathak and Roy (1993)
Smith et al. (1998)
Kwok et al. (1999)
Kawachi et al. (1980)
4-Hydroxybiphenyl in vitro tests
S. typhimurium
B. subtilis
TA98
TA1535
TA1535, 1536,
1537-1, 1538-1
H17 (rec+)
M45 (reel
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
Rat
DNA fragments
from plasmid
pbcNI
F344/bladder
epithelium
DNA damage,
Comet assay
DNA damage,
alkaline elution
assay
0.1 mM
0.05% injected
intravesically into
bladder wall
NT +d
e
Inoue et al. (1990)
Morimoto et al. (1989)
B-ll
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-3. Genotoxicity test results for biphenyl metabolites
Organism
Mouse
Strain or test
system
CD-1/skin
Endpoint
DNA adduct,
[32P]-post
labeling
method
Test substance
concentrations
10 or 20 mg applied
to skin
Metabolic
activation"
+S9
H
-S9
Reference
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%.
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 B-3 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 without S9 in CHL fibroblasts in one
13 study (Kawachi etal.. 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 B-3).
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; Smith et al., 1998). DNA damage was also detected in the urinary bladder
23 of male or female rats intravesically injected with 0.05 or 0.1% phenylbenzoquinone, but not
24 with injections of 0.05% 2-hydroxybiphenyl or 2,5-dihydroxybiphenyl, although DNA damage
25 was found in urinary bladders from male F344 rats fed the sodium salt of 2-hydroxybiphenyl in
B-12
DRAFT - DO NOT CITE OR QUOTE
-------
1 the diet for 3 months at 1,000 or 2,000 ppm, but not at 500 or 250 ppm (Morimoto et al.. 1989).
2 Topical application of 10 or 20 mg of the sodium salt of 2-hydroxybiphenyl or 5 mg of
3 2,5-dihydroxybiphenyl to the skin of female CD-I mice produced several DNA adducts in the
4 skin that were detected by the [32P]-post labeling technique (Pathak and Roy, 1993). Similar
5 adducts were formed in vitro when DNA was incubated with 2-hydroxybiphenyl (1 mM) in the
6 presence metabolic activation from rat skin homogenates, a CYP system, or a prostaglandin
7 synthase system (Pathak and Roy, 1993). In contrast, Smith et al. (1998), using a similar
8 technique to that used by Pathak and Roy (1993), were unable to detect exposure-related DNA
9 adducts in bladder epithelial tissue from male F344 rats fed 800, 4,000, 8,000, or 12,500 ppm 2-
10 hydroxybiphenyl in the diet for 13 weeks. In this experiment, increased bladder cell epithelium
11 proliferation (i.e., increased BrdU incorporation) was observed at 8,000 and 12,500 ppm, dietary
12 concentrations associated with the development of urinary bladder tumors in chronically exposed
13 rats (Smith et al., 1998). Kwok et al. (1999) found no evidence of binding of radioactivity to
14 DNA extracted from the bladder epithelium of male F344 rats given single gavage doses of
15 [14C]-labeled 2-hydroxybiphenyl at 15, 50, 250, 500, or 1,000 mg/kg, but increased protein
16 binding occurred with increasing doses of 250, 500, and 1,000 mg/kg. Kwok et al. (1999) noted
17 that the increase in protein binding increased with increasing dose levels of 250, 500, and 1,000
18 mg/kg, in parallel with increasing incidence of bladder epithelial lesions (hyperplasia,
19 papillomas, and carcinomas) in rats chronically exposed to 2-hydroxybiphenyl in the diet at 0,
20 269, and 531 mg/kg.
21 Increased micronuclei (about threefold increase over controls) and increased cell
22 proliferation (>200-fold increased incorporation of BrdU in DNA) were found in the bladder
23 epithelium of male F344 rats exposed to 2% (2,000 ppm) 2-hydroxybiphenyl in the diet for
24 2 weeks, without evidence for hypo- or hyperploidy as assayed by fluorescence in situ
25 hybridization with a DNA probe for rat chromosome 4 (Balakrishnan et al., 2002). Similar
26 exposure to 2% NaCl or 2% 2-hydroxybiphenyl + 2% NaCl, produced about two- or six-fold
27 increases of micronuclei in the bladder epithelium, respectively, but neither treatment stimulated
28 bladder epithelium cell proliferation to the same degree as 2% 2-hydroxybiphenyl in the diet
29 Balakrishnan. 2-Hydroxybiphenyl reportedly did not induce SCEs in the bone marrow of rats,
30 but exposure parameters were not specified in the report by Kawachi et al. (1980). The
31 mechanism of 2-hydroxybiphenyl-induced micronuclei is not understood, but, as discussed by
32 Balakrishan et al. (2002), possible mechanisms include: (1) DNA damage from ROS from redox
33 cycling between 2,5-dihydroxybiphenyl and phenylbenzoquinone, (2) interference of the mitotic
34 spindle through covalent modification of proteins, (3) inhibition of enzymes regulating DNA
35 replication, or (4) micronuclei generation as a secondary response to cytotoxicity or regenerative
36 hyperplasia.
37 Bacterial mutation assays of the major biphenyl metabolite, 4-hydroxybiphenyl, yielded
38 negative results in all but one case that was accompanied by overt cytotoxicity (Narbonne et al.,
B-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 1987). 2,5-Dihydroxybiphenyl (i.e., phenylhydroquinone) caused in vitro damage to human
2 DNA from plasmid pbcNI in the presence of Cu(II) (Inoue et al., 1990), DNA adducts when
3 applied to mouse skin (Pathak and Roy, 1993), but did not cause DNA damage when injected
4 intravesically into the urinary bladder of F344 rats at a concentration of 0.05% (Morimoto et al.,
5 1989).
B-14 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX C. BENCHMARK DOSE CALCULATIONS FOR THE REFERENCE
DOSE
Datasets used for modeling incidences of nonneoplastic effects in the urinary tract of
male and female F344 rats exposed to biphenyl in the diet for 2 years (Umeda et al., 2002) are
shown in Table C-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 C-2. The dataset for incidence of
litters with fetal skeletal anomalies from Wistar rat dams administered biphenyl by gavage on
GDs 6-15 (KheraetaL 1979) is shown in Table C-3.
Table C-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) (U.S. EPA. 1988. Table 1-6).
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).
C-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
3
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.
Source: Umeda et al. (2005).
C-2
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
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: Khera et al. (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 C-4 through C-24. Each table of modeled results for a particular effect is followed by the
information from the output file of the best-fitting model for that effect.
Table C-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
Weibullb
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
"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 ; 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).
C-3
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
0.6
0.5
0.4
0.3
0.2
0.1
50 100 150 200 250 300 350
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*doseAl-beta2*doseA2-beta3*doseA3)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 4
Total number of specified parameters = 0
Degree of polynomial = 3
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
C-4
DRAFT - DO NOT CITE OR QUOTE
-------
Model
Full model
Fitted model
Reduced model
of Fit
Est. Prob.
is a 90% two-sided confidence interval for the BMD
Table C-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)
BMD5
200.54
277.38
203.45
188.92
191.47
253.65
207.16
BMDL5
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
"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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-5
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
.2
•5
2
50 100 150 200 250 300 350 400 450
11:4801/132011
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/renalnodularhyper/female/mst_nodhypFrev_M;
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/renalnodularhyper/female/mst_nodhypFrev_M;
Thu Jan 13 11:48:49 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl-beta2*doseA2)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
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
C-6
DRAFT - DO NOT CITE OR QUOTE
-------
Model
Full model
Fitted model
Reduced model
AIC:
Test d.f.
P-value
Est. Prob.
d.f. =3
Taken together, (211.518, 351.444) is a 90% two-sided confidence interval for the BMD
Table C-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)d
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
"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 differed by less than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-7
DRAFT - DO NOT CITE OR QUOTE
-------
Gamma Multi-Hit Model with 0.95 Confidence Level
I
••5
S
0.5
0.4
0.3
0.2
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.pit
Thu Jan 13 11:55:07 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 = 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
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
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 inequality constraint and thus
has no standard error.
DRAFT - DO NOT CITE OR QUOTE
-------
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model
Fitted model
Reduced model
AIC: 184.407
Goodness of Fit
Dose Est._Prob. Expected Observed Size
P-value = 0.6558
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 313.755
BMDL = 113.219
Table C-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
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the gamma and Weibull models took the form of a 1-degree polynomial multistage model and
produced identical goodness of fit statistics and BMD values; the model with the lowest AIC was selected because
BMDL values for models providing adequate fit differed by less than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-9
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
o
't5
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.plt
Thu Jan 13 14:01:13 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 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
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.057038 * * *
0.00148135 * * *
P-value
C-10
DRAFT - DO NOT CITE OR QUOTE
-------
Fitted model
Reduced model
AIC
Goodness of Fit
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 C-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
%2/7-valuea
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
BMDL5
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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-ll
DRAFT - DO NOT CITE OR QUOTE
-------
LogProbit Model with 0.95 Confidence Level
0.5
0.4
0.3
0.2
0.1
0 50 100 150 200 250 300 350
dose
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.plt
Thu Jan 13 15:38:28 2011
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
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
-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 inequality constraint and thus
has no standard error.
C-12
DRAFT - DO NOT CITE OR QUOTE
-------
Model
Full model
Fitted model
Reduced model
Test d.f.
P-value
P-value = 0.6434
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 207.879
BMDL = 138.127
Table C-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
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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-13
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
I
c
o
't5
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage
BMDU 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.plt
Thu Jan 13 16:24:18 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 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
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0.228898 * * *
Beta(l) 0.0012018 * * *
* - Indicates that this value is not calculated.
C-14
DRAFT - DO NOT CITE OR QUOTE
-------
Full model
Fitted model
Reduced model
AIC:
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 87.669
BMDL = 56.2773
BMDU = 172.188
Taken together, (56.2773, 172.188) is a 90% two-sided confidence
Table C-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^-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
aValues <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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-15
DRAFT - DO NOT CITE OR QUOTE
-------
Dichotomous-Hill Model with 0.95 Confidence Level
I
0.7
0.6
0.5
0.4
0.3
0.2
0.1
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
Variable
C-16
DRAFT - DO NOT CITE OR QUOTE
-------
intercept
slope
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
Est. Prob.
d.f.
Benchmark Dose Computation
Table C-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
O.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
"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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-17
DRAFT - DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
o
••5
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.plt
Fri Jan 14 11:25:01 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl) ]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 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
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.168634 * * *
0.00114846 * * *
P-value
C-18
DRAFT - DO NOT CITE OR QUOTE
-------
Fitted model
Reduced model
Goodness of Fit
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 C-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
X2/>-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
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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-19
DRAFT - DO NOT CITE OR QUOTE
-------
Logistic Model with 0.95 Confidence Level
•5
I
•5
2
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
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
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
Log(likelihood) # Param's Deviance Test d.f.
-65.6458 4
-67.0198 2 2.74796 2
-71.3686 1 11.4455 3
C-20
DRAFT - DO NOT CITE OR QUOTE
-------
Est. Prob.
d.f.
Table C-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)"1
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
"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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2002).
C-21
DRAFT - DO NOT CITE OR QUOTE
-------
Gamma Multi-Hit Model with 0.95 Confidence Level
0.8
0.6
0.4
0.2
Gamma Multi-Hit
BMDL
0 50 100 150 200 250 300 350
dose
14:1501/142011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Gamma Model. (Version: 2.15; Date: 10/28/2009)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/bladdercombinedhyper/male/gam_bladcomhypMrev_gamma.(d
)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/bladdercombinedhyper/male/gam_bladcomhypMrev_gamma.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
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
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0 NA
Slope 0.0624215 0.00323795 0.0560752 0.0687677
Power 18 NA
NA - Indicates that this parameter has hit a bound implied by some inequality constraint and thus
has no standard error.
C-22
DRAFT - DO NOT CITE OR QUOTE
-------
Model
Full model
Fitted model
Reduced model
AIC:
Test d.f.
P-value
Goodness of Fit
Est. Prob.
d.f. =3
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 205.404
BMDL = 146.733
Table C-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
X2/>-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
BMDLS
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
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
kpower 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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2005).
C-23
DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
•5
2
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
Log-Logistic
BMDL
BMD
200
800
1000
400 600
dose
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
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
Variable
background
intercept
slope 1
* - Indicates that this value is not calculated.
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.185925 * * *
-8.77824 * * *
C-24
DRAFT - DO NOT CITE OR QUOTE
-------
Full model
Fitted model
Reduced model
AIC:
of Fit
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 721.275
BMDL = 276.216
Table C-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
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bPower restricted to >1.
'Selected model; the model with the lowest BMDL10 was selected because BMDL values for models providing
adequate fit differed by more than threefold.
dBetas restricted to >0.
BMD = maximum likelihood estimate of the dose associated with the selected BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Umeda et al. (2005).
C-25
DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
o
••s
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Log-Logistic
BMDL
BMD
0 200 400 600 800 1000 1200 1400
dose
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/2011/BMD/mice/minmedulla/female/lnl_minmedullF_loglogistic.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/minmedulla/female/lnl_minmedullF_loglogistic.plt
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 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
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 -0.48 0.44
intercept -0.48 1 -0.99
slope 0.44 -0.99 1
Variable
background
intercept
slope
* - Indicates that this value is not calculated.
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0.05773 * * *
-8.90345 * * *
199QQQ •*• •*• •*•
C-26
DRAFT - DO NOT CITE OR QUOTE
-------
Full model
Fitted model
Reduced model
AIC:
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 233.39
BMDL = 122.401
Table C-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.0001
O.0001
0.0001
O.0001
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.0001
0.0001
0.0001
O.0001
-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
aValues O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD 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).
C-27
DRAFT - DO NOT CITE OR QUOTE
-------
The constant variance models did not fit the variance data. The nonconstant variance
models did not fit the means data. Therefore, none of the models provided an adequate fit to the
data on serum LDH activity in female mice exposed to biphenyl in the diet for 2 years.
Table C-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.0001
O.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
O.OOOl
0.0001
0.0001
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.0001
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
O.OOOl
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
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
Selected model; only model providing adequate fit to modeled variance and means.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD 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).
C-28
DRAFT - DO NOT CITE OR QUOTE
-------
Power Model with 0.95 Confidence Level
8.
0
n:
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 (1RD), and are in units of mg/kg-
day.
Power Model. (Version: 2.16; Date: 10/28/2009)
Input Data File: C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/AST/pow_ASTFHDD_power.plt
Tue Jan 18 10:47:11 2011
The form of the response function is: Y[dose] = control + slope * doseApower
Dependent variable = mean
Independent variable = dose
The power is restricted to be greater than or equal to 1
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
lalpha = 10.765
rho = 0
control = 75
slope = 0.369536
power = 0.980467
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -power have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
lalpha rho control slope
Variable
lalpha
rho
control
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
-12.9059 4.06805 -20.8791 -4.93268
4.54893 0.905641 2.7739 6.32395
74.0253 5.21212 63.8097 84.2409
C-29
DRAFT - DO NOT CITE OR QUOTE
-------
slope 0.38893 0.113823
power 1 NA
NA - Indicates that this parameter has hit a bound implied by some inequality constraint and thus
has no standard error.
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij) Var{e(ij)
Model A2: Yij = Mu(i) + e(ij) Var{e(ij)
Model A3:
Likelihoods of Interest
Model Log(likelihood) # Param's
Al -410.240404 4
A2 -350.033965 6
A3 -350.072753 5
fitted -350.666161 4
R -412.701435 2
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 adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
Test -2*log(Likelihood Ratio) Test df
Test 1 125.335 4
Test 2 120.413 2
Test 3 0.0775771 1
Test 4 1.18681 1
The p-value for Test 1 is less than .05. There appears to be a difference between response
and/or variances among the dose levels. It seems appropriate to model the data
The p-value for Test 3 is greater than .1. The modeled variance appears to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems to adequately describe the
data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Relative risk
Confidence level = 0.95
BMD = 190.33
BMDL = 121.534
C-30 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
0.0001
0.0001
NA
0.55
9.61 x 1Q-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
0.0001
0.0001
0.0001
-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
O.0001
O.0001
O.0001
0.79
NA
NA
-0.22
4.25 x 10'7
-3.00 x 10'9
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
O.0001
O.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
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models did not fit the means data. Therefore, none of the models provided
an adequate fit to the data on serum ALT activity in female mice exposed to biphenyl in the diet
for 2 years.
C-31
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
O.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
O.0001
O.0001
0.0001
O.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
0.0001
0.0001
O.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
O.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
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models fit the variance data, but did not fit the means data. Therefore,
none of the models provided an adequate fit to the data on serum AP activity in female mice
exposed to biphenyl in the diet for 2 years.
C-32
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
"Values <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
C-33
DRAFT - DO NOT CITE OR QUOTE
-------
Linear Model with 0.95 Confidence Level
25
24
23
22
21
20
19
Linear
BMDL
BMI)
0 50 100 150 200 250 300 350 400
dose
11:0301/192011
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:
::/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/BUN/male/lin_BUNMHDD_linear.(d)
Gnuplot Plotting File:
::/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/BUN/male/lin_BUNMHDD_linear.plt
Wed Jan 19 11:03:37 2011
The form of the response function is: Y[dose] = beta_0 + beta_l*dose + beta_2*doseA2 + ...
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 1 = 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~l 3.2e-008 -0.74 1
Variable
alpha
beta 0
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
11.6911 20.0904
19.4657 21.6863
C-34
DRAFT - DO NOT CITE OR QUOTE
-------
beta_l 0.0096108 0.00317579 0.00338636 0.0158352
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A3 uses any fixed variance parameters that were specified by the user
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model Log(likelihood) # Param's
Al -206.630664 4
A2 -205.915695 6
A3 -206.630664 4
fitted -207.115525 3
R -211.514015 2
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 adequately 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
Test 1 11.1966 4
Test 2 1.42994 2
Test 3 1.42994 2
Test 4 0.969721 1
The p-value for Test 2 is greater than .1. A homogeneous variance model appears to be
appropriate here
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 414.775
BMDL = 266.77
C-3 5 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
0.0001
0.0001
NA
0.38
-3.45 x 1Q-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
0.0001
-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
O.0001
0.50
0.82
NA
-0.57
-0.18
-2.11 x 10'10
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
O.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
aValues O.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be positive.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; IRD = dose associated with a 100% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit the variance data. The nonconstant variance
models fit the variance data, but failed to fit the means data. When the data from the highest
dose group were dropped, the constant variance models did not fit the variance data. The
nonconstant variance models fit the variance data, but did not fit the means data. Therefore,
none of the models provided an adequate fit to the data on BUN levels in female mice exposed to
biphenyl in the diet for 2 years.
C-36
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
^-value"
Largest
residual
AIC
Benchmark result (mg/kg-d)
BMD1SD
BMDL1SD
BMD01RD
BMDL01RD
All doses
Constant variance
Hillb
Linear0, Powerd
Polynomial (3 -degree)0
O.0001
O.0001
0.0001
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
aValues <0.10 fail to meet conventional goodness-of-fit criteria.
bRestrictn> 1.
°Coefficients restricted to be negative.
dRestrictpower>l.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; CURD = dose associated with a 10% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
The constant variance models did not fit either the variance data or the means data. The
nonconstant variance models failed to fit the variance data. When the data from the highest dose
group were dropped, the constant variance models did not fit either the variance data or the
means data. The nonconstant variance models did not fit the means data. Therefore, none of the
models provided an adequate fit to the data on mean terminal body weight in male mice exposed
to biphenyl in the diet for 2 years.
C-37
DRAFT - DO NOT CITE OR QUOTE
-------
Table C-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
BMD01RD
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.
eRestrict power >1.
BMDL = 95% lower confidence limit on the BMD (subscripts denote BMR: i.e., iSD = dose associated with 1 SD
from control mean value; 0 IRD = dose associated with a 10% RD from control mean value); CF = computation
failed; NA = not applicable
Source: Umeda et al. (2005).
Linear Model with 0.95 Confidence Level
36
34
32
8 30
28
26
24
Linear
BMDL
BMD
0 200 400 600 800 1000 1200 1400
dose
09:2001/202011
BMD and BMDL indicated are associated with a 10% decrease from control (0.1 RD), and are in units of mg/kg-
day.
C-38
DRAFT - DO NOT CITE OR QUOTE
-------
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.plt
Thu Jan 20 09:20:01 2011
BMDS Model Run
The form of the response function is: Y[dose] = beta_0 + beta_l*dose + beta_2*doseA2 + ...
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_l = -0.00571961
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho have been estimated at a boundary point, or have been
specified by the user, and do not appear in the correlation matrix )
alpha beta_0 beta_l
alpha 1 -9.6e-009 9.1e-009
beta_0 -9.6e-009 1 -0.67
beta 1 9.1e-009 -0.67 1
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
8.27818 14.2255
32.6505 34.346
-0.0068114 -0.00467385
31
Est Std Dev
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij) Var{e(ij)
Model A2: Yij = Mu(i) + e(ij) Var{e(ij)
Model A3: Yij = Mu(i) + e(ij) Var{e(ij)
Model A3 uses any fixed variance parameters that were specified by the user
Model R: Yi = Mu + e(i) Var{e(i)} = SigmaA2
Model
Al
A2
A3
fitted
R
Likelihoods of Interest
# Param's
Test 1:
Test 2:
Test 3:
Test 4:
Explanation of Tests
Do responses and/or variances differ among Dose levels? (A2 vs.
Are Variances Homogeneous? (Al vs A2)
Are variances adequately modeled? (A2 vs. A3)
Does the Model for the Mean Fit? (A3 vs. fitted)
R)
Tests of Interest
C-39
DRAFT - DO NOT CITE OR QUOTE
-------
The p-value for Test 3 is greater than .1. The modeled variance appears to be appropriate here
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Relative risk
Confidence level = 0.95
BMD = 583.327
BMDL = 510.848
Table C-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
BMDL5
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 BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., 10 = dose associated with 10% extra risk; 5 = dose
associated with 5% extra risk)
Source: Khera et al. (1979).
C-40
DRAFT - DO NOT CITE OR QUOTE
-------
Log-Logistic Model with 0.95 Confidence Level
I
0.9
0.8
0.7
0.6
0.5
0.4
0.3
02 HMDL
Log-Logistic
BMD
0 200 400 600 800 1000
dose
16:0601/142011
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:
::/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/develop/anomlitt/lnl_anomlitt_loglogistic.(d)
Gnuplot Plotting File:
::/Storage/USEPA/IRIS/biphenyl/2011/BMD/rat/develop/anomlitt/lnl_anomlitt_loglogistic.plt
Fri Jan 14 16:06:43 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 = 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
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
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
background 0.503241 * * *
intercept -6.24131 * * *
C-41
DRAFT - DO NOT CITE OR QUOTE
-------
slope 1 * *
* - Indicates that this value is not calculated.
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -49.327 5
Fitted model -50.6629 2 2.67182 3
Reduced model -52.2232 1 5.79233 4
Est. Prob.
d.f. =3
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 57.0591
BMDL = 20.2399
C-42 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX D. BENCHMARK MODELING FOR THE ORAL SLOPE FACTOR
The mouse liver tumor dataset from Umeda et al. (2005) for which dose-response
modeling was performed is shown in Table D-l.
Table D-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
14/48^
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 D-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.
D-l
DRAFT - DO NOT CITE OR QUOTE
-------
Table D-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, Weibull0
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 O.05 fail to meet conventional goodness-of-fit criteria.
'"Betas restricted to >0.
°Power restricted to >1.
dSelected model.
BMD = maximum likelihood estimate of the dose associated with the selected BMR; BMDL = 95% lower
confidence limit on the BMD (subscripts denote BMR: i.e., HEDIO = HED associated with 10% extra risk
Source: Umeda et al. (2005).
BMDS graph of multistage (1-degree) model that includes data from the highest dose group:
0.5
0.4
0.3
0.2
0.1
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
BMDL
3MD
50
100
dose
150
200
14:01 09/192011
The BMDS graph of multistage (1-degree) model that includes data from the highest dose group.
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-
day.
D-2
DRAFT - DO NOT CITE OR QUOTE
-------
BMDS graph of multistage (1-degree) model with the highest dose group dropped:
Multistage Cancer Model with 0.95 Confidence Level
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Li near extrapolation
BMDL BMD
10
20
30
dose
40
50
60
09:3302/032011
BMD and BMDL indicated are associated with an extra risk of 10%, and are in units of mg/kg-day.
Multistage Cancer Model. (Version: 1.9; Date: 05/26/2010)
Input Data File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l.
(d)
Gnuplot Plotting File:
C:/Storage/USEPA/IRIS/biphenyl/2011/BMD/mice/livertumor/female/revised_n/msc_livtumFrev2HDD_MS_l.
pit
Thu Feb 03 09:33:34 2011
The form of the probability function is: P[response] = background + (1-background)*[1-EXP(-
betal*doseAl)]
The parameter betas are restricted to be positive
Dependent variable = incidence
Independent variable = dose
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: 2.22045e-016
Parameter Convergence has been set to: 1.49012e-008
**** We are sorry but Relative Function and Parameter Convergence are currently unavailable in
this model. Please keep checking the web site for model updates which will eventually
incorporate these convergence criterion. Default values used. ****
D-2
DRAFT - DO NOT CITE OR QUOTE
-------
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0.0630397 * * *
Beta(l) 0.00562948 * * *
* - Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -64.1585 3
Fitted model -64.1595 2
Reduced model -70.107 1
AIC: 132.319
Goodness of Fit
Est._Prob. Expected Observed
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 18.7158
BMDL = 12.1518
BMDU = 36.3895
Taken together, (12.1518, 36.3895) is a 90% two-sided confidence interval for the BMD
Multistage Cancer Slope Factor = 0.00822924
D-4 DRAFT - DO NOT CITE OR QUOTE
------- |