EPA/63 5/R-08/004F
www.epa.gov/iris
f/EPA
TOXICOLOGICAL REVIEW
OF
NITROBENZENE
(CAS No. 98-95-3)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
January 2009
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
-------
CONTENTS—TOXICOLOGICAL REVIEW OF NITROBENZENE
(CAS No. 98-95-3)
LIST OF TABLES vi
LIST OF FIGURES x
LIST OF ACRONYMS xi
FOREWORD xiii
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiv
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION 3
3. TOXICOKINETICS 5
3.1. ABSORPTION 5
3.1.1. Gastrointestinal Tract Absorption Studies 5
3.1.2. Pulmonary Absorption Studies 6
3.1.3. Dermal Absorption Studies 7
3.2. DISTRIBUTION 8
3.3. METABOLISM 10
3.3.1. Microbial Reduction of Nitrobenzene (the Three-Step, Two-Electrons-per-
Step Transfer Process) 13
3.3.2. Hepatic and Erythrocytic Reduction of Nitrobenzene (the Six-Step, One-
Electron-per-Step Transfer Process) 17
3.3.3. Microsomal Oxidation of Nitrobenzene 22
3.4. ELIMINATION 23
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 25
4. HAZARD IDENTIFICATION 26
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY AND CASE REPORTS 26
4.1.1. Oral Exposure 26
4.1.2. Inhalation Exposure 28
4.1.3. Dermal Exposure 29
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION 31
4.2.1. Oral Exposure 31
4.2.1.1. Subchronic Studies 31
4.2.1.2. Chronic Studies 43
4.2.2. Inhalation Exposure 43
4.2.2.1. Subchronic Studies 43
4.2.2.2. Chronic Studies 46
4.2.3. Dermal Exposure 57
4.2.3.1. Subchronic Studies 57
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION... 61
4.3.1. Oral Exposure 61
4.3.2. Inhalation Exposure 68
4.4. OTHER STUDIES 73
4.4.1. Acute and Short-Term Toxicity Data 73
iii
-------
4.4.2. Structure-Activity Relationships 77
4.4.3. Immunotoxicity Studies 81
4.4.4. Neurotoxicity Studies 84
4.4.5. Genotoxicity Studies 86
4.4.6. Other Studies in Support of Mode of Action 92
4.5. SYNTHESIS OF MAJOR NONCANCER EFFECTS 92
4.5.1. Oral Exposure 93
4.5.2. Inhalation Exposure 97
4.5.3. Mode-of-Action Information 101
4.6. EVALUATION OF CARCINOGENICITY 103
4.6.1. Summary of Overall Weight of Evidence 103
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence 103
4.6.3. Mode-of-Action Information 105
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 107
4.7.1. Possible Childhood Susceptibility 107
4.7.2. Possible Gender Differences 107
4.7.3. Other 108
5. DOSE-RESPONSE ASSESSMENTS 109
5.1. ORAL REFERENCE DOSE (RfD) 109
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and
Justification 109
5.1.2. Method of Analysis—Including Models 112
5.1.3. RfD Derivation—Including Application of Uncertainty Factors (UFs) 115
5.1.4. Previous RfD Assessment 117
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 117
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and
Justification 117
5.2.2. Methods of Analysis—Including Models 121
5.2.3. Evaluation of Human Equivalent Concentrations 123
5.2.4. RfC Derivation—Including Application of Uncertainty Factors (UFs) 126
5.2.5. Previous RfC Assessment 127
5.3. CANCER ASSESSMENT 127
5.3.1. Choice of Principal Study and Target Organ—with Rationale and
Justification 127
5.3.2. Benchmark Concentration Modeling 130
5.3.3. Inhalation Dose Adjustments, Inhalation Unit Risk, and Extrapolation
Methods 130
5.3.4. Uncertainties in Cancer Risk Values 133
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE 134
6.1. HUMAN HAZARD POTENTIAL 134
6.1.1. Exposure Pathways 134
6.1.2. Toxicokinetics 134
6.1.3. Characterization of Noncancer Effects 135
6.1.4. Reproductive Effects and Risks to Children 137
6.1.5. Noncancer Mode of Toxic Action 138
6.1.6. Characterization of the Human Carcinogenic Potential 138
iv
-------
6.2. DOSE RESPONSE 139
6.2.1. OralRfD 139
6.2.2. Inhalation RfC 140
6.2.3. Oral Cancer Risk 141
6.2.4. Inhalation Cancer Risk 141
7. REFERENCES 143
APPENDIX A: SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION A-l
APPENDIX B: DOSE-RESPONSE MODELING B-l
APPENDIX B-l: Dose-Response Modeling for Derivation of an RfD for Nitrobenzene B-2
APPENDIX B-2: Dose-Response Modeling for Derivation of an RfC for Nitrobenzene B-24
APPENDIX B-3: Dose-Response Modeling of Carcinogenicity Data for Nitrobenzene B-36
v
-------
LIST OF TABLES
Table 3-1. Reduction of nitrobenzene by various rat tissue homogenates 13
Table 3-2. MetHb formation in the blood of rats dosed intraperitoneally with 200 mg/kg
nitrobenzene in corn oil 15
Table 3-3. Formation of metabolites of nitrobenzene in the presence of cecal contents in
vitro: influence of diet 15
Table 3-4. Urinary metabolites of [14C]-nitrobenzene excreted within 72 hours after gavage.... 16
Table 3-5. Enzyme systems in erythrocytes 20
Table 3-6. Recovery of radiolabel in F344 and CD rats and B6C3F1 mice 72 hours after
exposure to a single oral dose of [14C]-nitrobenzene 24
Table 3-7. Urinary excretion of nitrobenzene metabolites in male rats and mice gavaged
with a single oral dose of [14C]-nitrobenzene 25
Table 4-1. Cases of human poisoning following ingestion of nitrobenzene 29
Table 4-2. Cases of human poisoning with nitrobenzene following inhalation or dermal
exposure 31
Table 4-3. Changes in absolute and relative liver, kidney, and testis weights in male
F344 rats exposed to nitrobenzene by gavage for 90 days 33
Table 4-4. Changes in absolute and relative liver and kidney weights in female F344 rats
exposed to nitrobenzene by gavage for 90 days 33
Table 4-5. Hematologic parameters, reticulocytes, and metHb levels in male F344 rats
exposed to nitrobenzene via gavage for 90 days 34
Table 4-6. Hematologic parameters, reticulocytes, and metHb levels in female F344 rats
exposed to nitrobenzene via gavage for 90 days 34
Table 4-7. Selected histopathology findings in male F344 rats exposed to nitrobenzene
for 90 days via gavage 35
Table 4-8. Selected histopathology findings in female F344 rats exposed to nitrobenzene
for 90 days via gavage 36
Table 4-9. Changes in absolute and relative liver, kidney, and testis weights in male
B6C3F1 mice exposed to nitrobenzene by gavage for 90 days 37
Table 4-10. Changes in absolute and relative liver, kidney, and thymus weights in female
B6C3F1 mice exposed to nitrobenzene by gavage for 90 days 37
vi
-------
Table 4-11. Hematologic parameters, reticulocytes, and metHb levels in male B6C3F1
mice exposed to nitrobenzene via gavage for 90 days 38
Table 4-12. Hematologic parameters, reticulocytes, and metHb levels in female B6C3F1
mice exposed to nitrobenzene via gavage for 90 days 38
Table 4-13. Selected histopathology findings in male B6C3F1 mice exposed to
nitrobenzene for 90 days via gavage 39
Table 4-14. Selected histopathology findings in female B6C3F1 mice exposed to
nitrobenzene for 90 days via gavage 39
Table 4-15. Hematologic and clinical chemistry parameters in rats treated with
nitrobenzene for 28 days, with or without a recovery period of 14 days 40
Table 4-16. Summary of effects observed in oral dosing studies with nitrobenzene 42
Table 4-17. Concentrations of metHb in plasma of F344 and CD rats and B6C3F1 mice in
response to nitrobenzene inhalation 44
Table 4-18. Summary of effects observed in subchronic inhalation studies with
nitrobenzene 46
Table 4-19. Summary of neoplastic and nonneoplastic findings following 2-year inhalation
exposure to nitrobenzene 50
Table 4-20. Percentage metHb formation in response to inhaled nitrobenzene 52
Table 4-21. Selected noncancer histopathologic changes in rats as a result of exposure to
nitrobenzene via inhalation for 2 years 55
Table 4-22. Selected noncancer histopathologic changes in B6C3F1 mice as a result of
exposure to nitrobenzene via inhalation for 2 years 56
Table 4-23. Summary of effects observed from chronic inhalation with nitrobenzene at
terminal sacrifice 57
Table 4-24. Incidence of histopathologic lesions in male F344 rats exposed to
nitrobenzene for 90 days via dermal exposure 58
Table 4-25. Incidence of histopathologic lesions in female F344 rats exposed to
nitrobenzene for 90 days via dermal exposure 58
Table 4-26. Incidence of histopathologic lesions in male B6C3F1 mice exposed to
nitrobenzene for 90 days via dermal exposure 59
Table 4-27. Incidence of histopathologic lesions in female B6C3F1 mice exposed to
nitrobenzene for 90 days via dermal exposure 60
vn
-------
Table 4-28. Summary of effects observed in dermal dosing studies with nitrobenzene 60
Table 4-29. Hematologic findings in male Sprague-Dawley rats exposed via gavage to
nitrobenzene 64
Table 4-30. Relative organ weights of male Sprague-Dawley rats gavaged with
nitrobenzene 65
Table 4-31. Summary of effects observed in an oral reproductive study with nitrobenzene 66
Table 4-32. Incidence of skeletal variations in Sprague-Dawley fetuses exposed via
inhalation to nitrobenzene in utero 69
Table 4-33. Fertility indices for the FO, Fl, and recovery generations: number of
pregnancies per number of females mated 71
Table 4-34. Summary of effects observed in developmental inhalation studies with
nitrobenzene 72
Table 4-3 5. Percent metFIb in blood of rats exposed to nitrobenzene vapors 75
Table 4-36. Summary of effects observed in short-term inhalation studies with nitrobenzene .. 75
Table 4-37. Overview of properties and toxicities of nitrobenzenes 78
Table 4-38. Summary of toxicological studies with 1,3-dinitrobenzene 79
Table 4-39. Summary of toxicological studies with 1,3,5-trinitrobenzene 80
Table 4-40. Summary of studies on the direct mutagenicity/genotoxicity of nitrobenzene 91
Table 4-41. Neoplasms in F344 and CD rats and B6C3F1 mice exposed to nitrobenzene
via inhalation for 2 years 104
Table 5-1. Summary of effects in F344 rats associated with exposure to nitrobenzene by
gavage for 90 days 113
Table 5-2. Summary of noncancer BMD modeling of selected endpoints from F344 rats
exposed by gavage to nitrobenzene for 90 days, using NTP (1983a) bioassay
data 115
Table 5-3. Incidence of histopathologic lesions in mice following chronic nitrobenzene
inhalation 120
Table 5-4. Modeling results for bronchiolization of the alveoli and olfactory
degeneration in mice 122
Table 5-5. Hepatocellular neoplastic findings in F344 rats exposed to nitrobenzene via
inhalation for 2 years 128
viii
-------
Table 5-6. Selected cancer incidences in B6C3F1 mice, F344 rats, and CD rats following
2-year inhalation exposure to nitrobenzene 129
Table 5-7. Estimated BMCs and BMCLs based on tumor incidence data in male F344 rats
exposed to nitrobenzene via inhalation 130
Table 5-8. Cancer risk estimates from nitrobenzene tumor incidence in male F344 rats,
based on the slope to background from the BMC 132
Table 5-9. ITJRs for nitrobenzene, based on tumor incidence in male F344 rats, based on
the slope to background from the BMCL 132
Table B-l. 1. Summary of splenic congestion, reticulocyte count (%) and metFIb levels
(%) in male and female F344 rats exposed by gavage to nitrobenzene for 90
days B-2
Table B-l.2. Summary of PODs derived from BMD modeling of NTP (1983a) bioassay
data for male and female rats exposed by gavage to nitrobenzene for 90
days B-3
Table B-2.1. Modeling results for olfactory degeneration in mice; bioassay data from
Cattleyetal. (1994) and CUT (1993) B-24
Table B-2.2. Modeling results for bronchiolization in mice; bioassay data from Cattley et
al. (1994) and CUT (1993) B-24
Table B-3.1. Tumorigenic responses in experimental animals exposed to nitrobenzene
via inhalation for up to 2 years B-36
Table B-3.2. Summary of BMD modeling results for nitrobenzene cancer data B-37
Table B-3.3. Overall risk based on kidney, thyroid, or liver adenomas/carcinomas in
F344 rats, using simulated data derived from observed incidences (Bayesian
approach) B-49
Table B-3.4. Estimated incidence of male F344 rats with liver, thyroid, or kidney tumors
from separately tabulated site-specific incidences following nitrobenzene
exposure for 2 years B-50
IX
-------
LIST OF FIGURES
Figure 2-1. Chemical structure of nitrobenzene 3
Figure 3-1. Time course of covalently bound [14C]-nitrobenzene in RBCs and spleen of
rats and mice 9
Figure 3-2. Time-related changes in spleen weight in rats and mice following
nitrobenzene treatment 10
Figure 3-3. Outline of the metabolism of nitrobenzene: a substrate for oxidation and
reduction reactions 11
Figure 3-4. Type I nitroreductase activity in male Sprague-Dawley rats 13
Figure 3-5. Mechanism of bacterial nitrobenzene reduction 17
Figure 3-6. Type II nitroreductase activity of male Sprague-Dawley rats 17
Figure 3-7. Mechanism of microsomal nitrobenzene reduction 19
Figure 3-8. Cycling of nitrosobenzene and phenylhydroxylamine in RBCs, resulting in
the formation of metHb 21
Figure 5-1. Exposure-response array of selected subchronic and reproductive-
developmental toxicity effects by the oral route 110
Figure 5-2. Exposure-response array of selected subchronic, chronic, and reproductive
toxicity effects by the inhalation route 118
x
-------
LIST OF ACRONYMS
ADAF age-dependent adjustment factor
AGMK African green monkey kidney
AIC Akaike Information Criterion
ALT alanine aminotransferase
AST aspartate aminotransferase
ATSDR Agency for Toxic Substances and Disease Registry
BBMV brush border membrane vesicle
BMC benchmark concentration
BMCL 95% lower bound on the BMC
BMD benchmark dose
BMDL 95% lower bound on the BMD
BMDS benchmark dose software
BMR benchmark response
BRRC Bushy Run Research Center
BUN blood urea nitrogen
CASRN Chemical Abstracts Service Registry Number
CUT Chemical Industry Institute of Toxicology
con A concanavalin A
CREST calcinosis, Raynaud's phenomenon, esophageal motility disorders, sclerodactyly,
and telangiectasia
DMPO 5,5-dimethyl-l-pyrroline-l-oxide
DMSO dimethyl sulfoxide
DNA deoxyribonucleic acid
EC Enzyme Commission (only in combination with numbers [e.g., EC 1.6.99.1])
ECso median effective concentration
EPA Environmental Protection Agency
ESR electron spin resonance
FasL Fas ligand
G6PD glucose-6-phosphate dehydrogenase
GD gestation day
GLP good laboratory practice
GSH reduced glutathione
Hb hemoglobin
Hct hematocrit
HEC human equivalent concentration
IgG immunoglobulin G
IgM immunoglobulin M
i.p. intraperitoneal
IPCS International Programme on Chemical Safety
IRIS Integrated Risk Information System
IUBMB International Union of Biochemistry and Molecular Biology
IUR inhalation unit risk
KLH keyhole limpet hemocyanin
LOAEL lowest-observed-adverse-effect level
LPS lipopolysaccharide
MCHb mean corpuscular hemoglobin
MCV mean corpuscular volume
xi
-------
metHb methemoglobin
MOA mode of action
NAD(H) (reduced) nicotine adenine dinucleotide
NADP(H) (reduced) nicotinamide adenine dinucleotide phosphate
NLM National Library of Medicine
NOAEL no-observed-adverse-effect level
NRC National Research Council
NTP National Toxicology Program
OECD Organization for Economic Cooperation and Development
oxyHb oxyhemoglobin
PBPK physiologically based pharmacokinetic
PHA phytohemagglutinin
PLN popliteal lymph node
PND postnatal day
POD point of departure
RBC red blood cell
RfC reference concentration
RfD reference dose
RGDR regional gas dose ratio
S9 9000 x g microsomal supernatant fraction
SD standard deviation
SDS-PAGE sodium dodecyl sulfate-polyacrylamide gel electrophoresis
T3 triiodothyronine
T4 thyroxine
TSH thyroid-stimulating hormone
UF uncertainty factor
WBC white blood cell
xn
-------
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
nitrobenzene. It is not intended to be a comprehensive treatise on the chemical or toxicological
nature of nitrobenzene.
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).
Xlll
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER
Ghazi A. Dannan, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
TedBerner, M.S.
Ghazi A. Dannan, Ph.D.
Belinda Hawkins, Ph.D., DABT
Karen Hogan, M.S.
James W. Holder, Ph.D.
Ching-Hung Hsu, Ph.D., DABT
Leonid Kopylev, Ph.D.
Todd Stedeford, Ph.D., J.D., DABT
Jeanmarie Zodrow, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
George Holdsworth, Ph.D.
Lutz W. Weber, Ph.D., DABT
Oak Ridge Institute for Science and Education
Oak Ridge Associated Universities
Oak Ridge, TN
xiv
-------
REVIEWERS
This document has been reviewed by EPA scientists, interagency reviewers from other
federal agencies, and the public, and peer reviewed by independent scientists external to EPA. A
summary and EPA's disposition of the comments received from the independent external peer
reviewers and from the public is included in Appendix A.
INTERNAL EPA REVIEWERS
Lynn Flowers, Ph.D., DABT
Karen Hammerstrom
Samantha Jones, Ph.D.
Channa Kesheva, Ph.D.
Susan Rieth
John Whalan
Diana Wong, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
EXTERNAL PEER REVIEWERS
Mark Miller, Ph.D., Chair
Wake Forest University School of Medicine
Bruce Allen
Bruce Allen Consulting
Rudolph Jaeger, Ph.D.
Environmental Medicine Incorporated
Martin Philbert, Ph.D.
University of Michigan
Richard Pleus, Ph.D.
Intertox
David Pyatt, Ph.D.
Summit Toxicology, LLP
Lorenz Rhomberg, Ph.D.
Gradient Corporation
xv
-------
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
nitrobenzene. IRIS Summaries may include oral reference dose (RfD) and inhalation reference
concentration (RfC) values for chronic and other exposure durations, and a carcinogenicity
assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal of entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per |ig/m3 air breathed.
Development of these hazard identification and dose-response assessments for
nitrobenzene has followed the general guidelines for risk assessment as set forth by the National
Research Council (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, 1986a), Guidelines for
Mutagenicity Risk Assessment (U.S. EPA, 1986b), 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, 199 la), Interim Policy for Particle Size and Limit
1
-------
Concentration Issues in Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of
Inhalation Reference Concentrations and Application of Inhalation Dosimetry (U.S. EPA,
1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995),
Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998a), Science Policy Council Handbook: Risk
Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A
Framework for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA,
2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through March 2008.
-------
2. CHEMICAL AND PHYSICAL INFORMATION
Structurally, nitrobenzene consists of a benzene ring with a single substituted nitro group
(Figure 2-1). The compound is an oily yellow liquid with an odor of bitter almonds. Synonyms
for nitrobenzene include oil of mirbane, essence of mirbane, nitrobenzol, and solvent black 6.
Figure 2-1. Chemical structure of nitrobenzene.
Pertinent physical and chemical properties of nitrobenzene are listed as follows (National
Library of Medicine's [NLM] Hazardous Substances Data Bank, 2003; World Health
Organization [IPCS], 2003; Agency for Toxic Substances and Disease Registry [ATSDR],
1990):
Chemical formula
Molecular weight
Melting point
Boiling point
Density
Water solubility
Log KQW
Log Koc
Vapor pressure
Henry's law constant
Conversion factor
C6H5N02
123.11
5.7°C
210.8°C
1.2g/mL(at20°C)
l,900mg/L(at20°C)
1.87
1.56
0.15 mm Hg at 25°C (20 Pa at 20°C)
1.31 x 10~5 atm-mVmol
1 ppm = 5.04 mg/m3; 1 mg/m3 = 0.2 ppm
Nitrobenzene is manufactured by direct nitration of benzene with nitric acid, using
sulfuric acid as catalyst and dehydrating agent. The purified product is used extensively in
chemical manufacturing, especially in the synthesis of other industrial chemicals and
intermediates. Most important among these is aniline, which is predominantly used in the
manufacture of polyurethane (IPCS, 2003). Other chemical products of nitrobenzene include
benzidine, quinoline, and azobenzene (NLM, 2003). The compound has been used as a solvent
for cellulose ethers and acetates and in petroleum refining. Nitrobenzene is present in a number
3
-------
of commercial products, such as shoe and metal polishes and soaps. An estimated
2,133,800 tons of nitrobenzene were produced worldwide in 1994 (IPCS, 2003), about one-third
of which was produced in the U.S. U.S. production of nitrobenzene has been increasing in recent
years, from 435,000 tons in 1986 to 533,000 tons in 1990 to 740,000 tons in 1994 (IPCS, 2003).
The majority (97-98%) of the nitrobenzene produced is retained in closed manufacturing
systems. Losses to wastewater have been estimated at <2% of production (ATSDR, 1990);
releases to air, land, and water via industrial processes also occur. EPA's Toxics Release
Inventory reported total on- and off-site releases of nitrobenzene in 2002 as 320,836 pounds, the
majority of which (-75%) were on-site disposal to Class I underground injection wells (U.S.
EPA, 2004). The chemical industry reported -89% of the nitrobenzene releases in 2002 with the
largest reported in Louisiana (226,526 pounds), followed by Texas (45,480 pounds) and Nevada
(31,669 pounds). Since 1988, trends data have shown annual declines in total on- and off-site
releases of nitrobenzene, comprised of decreases in surface water discharges, underground
injection, and releases to land; total air emissions have varied between 25,529 and 81,297
pounds, with no evident upward or downward trend.
Nitrobenzene exposure is predominantly occupational via the inhalation and/or dermal
routes. For members of the general population, both inhalation (ambient air) and ingestion
(drinking water) exposures are possible and are likely to be highest for those individuals living
near industrial/manufacturing sources or hazardous waste sites. Potential exposure through the
use of consumer products is also possible, but data are lacking to quantify these exposures.
Little or no information is available on nitrobenzene in foods. Nitrobenzene has been
measured in fish samples in Japan (4 of 147 samples) at levels ranging from 11-26 ng/kg (IPCS,
2003). In a recent British study, no nitrobenzene was detected (<2 |J,g/kg) in 49 honey samples
collected from hives fumigated with "Frow mixture" containing petroleum-derived substances in
addition to nitrobenzene to treat hives against parasitic mites (Castle et al., 2004).
Nitrobenzene may be absorbed through the skin as either a liquid or a vapor and may be
an additional primary route of exposure, besides inhalation, for workers in industries in which
nitrobenzene is used. The general population may be exposed dermally to nitrobenzene through
the use of consumer products containing nitrobenzene, but information about these potential
exposures is lacking (IPCS, 2003).
-------
3. TOXICOKINETICS
3.1. ABSORPTION
3.1.1. Gastrointestinal Tract Absorption Studies
There are no quantitative data on the extent of absorption of nitrobenzene in humans via
the oral route; however, it has been shown that nitrobenzene is well absorbed into brush border
membrane vesicles (BBMVs) from the small intestines of Sprague-Dawley rats (in vitro).
Absorption assays with isolated BBMVs and nitrobenzene were independent of age, sex, or
segment (i.e., proximal third, middle third, or distal third) of small intestine, suggesting that
lipophilicity of the compound and lipid composition of the membrane are the determining factors
(Alcorn et al., 1991). These basic considerations may be applicable to humans as well.
The IPCS (2003) has cited reports of incidents where individuals have been poisoned by
ingesting nitrobenzene, either accidentally or intentionally. Some of these case reports provide
inferential evidence of the compound's ready passage across the intestinal absorption barrier.
For example, Myslak et al. (1971) reported the case of a 19-year-old female who ingested about
50 mL of nitrobenzene approximately 30 minutes prior to the appearance of symptoms. During
recovery, samples of her urine were analyzed and revealed the presence of high levels of
p-amino- and p-nitrophenol, metabolites of nitrobenzene (see section 3.4), demonstrating
absorption from the gastrointestinal tract.
Extensive intestinal absorption of nitrobenzene has been demonstrated in experimental
animals. For example, a total of six rabbits (sex and strain not stated) were administered
[14C]-nitrobenzene and unlabeled nitrobenzene at total doses of 200 mg/kg (two animals) and
250 mg/kg (three animals) by stomach tube. One animal was exposed to 400 mg/kg; however, it
died after 2 days (Parke, 1956). Animals were kept in metabolic cages for 30 hours after dosing
to permit the collection of feces, urine, and expired air. Exhaled derivatives were trapped in
ethanol and/or CO2 absorbers. Thereafter, the animals were housed in open cages so that their
urine and feces could be collected up to 10 days. By 4-5 days after dosing animals, the author
found that nearly 70% of the radioactivity had been eliminated from the body. This included 1%
of the radioactivity expired as CO2, 0.6% expired as nitrobenzene (up to 30 hours), 58% excreted
as metabolites in the urine (up to 4-5 days), and 9% eliminated in the feces (up to 4-5 days).
The action of bacteria normally present in the small intestine of the rat is an important
element in the formation of methemoglobin (metHb) resulting from nitrobenzene exposure.
Germ-free rats do not develop methemoglobinemia when intraperitoneally dosed with
nitrobenzene (Reddy et al., 1976). When nitrobenzene (200 mg/kg of body weight in sesame oil)
was intraperitoneally administered to normal Sprague-Dawley rats, 30-40% of the hemoglobin
(Fib) in the blood was converted to metFIb within 1-2 hours. When the same dose was
administered to germ-free or antibiotic-pretreated rats, there was no measurable metFIb
5
-------
formation, even when measured up to 7 hours after treatment. The nitroreductase activities of
various tissues (liver, kidney, gut wall) were not appreciably different in germ-free and control
rats, but the activity was negligible in gut contents from germ-free rats and high in control rats.
This led the authors to suggest that a nitrobenzene metabolite such as aniline (which is formed by
the bacterial reduction of nitrobenzene in the intestines of rats) is involved in metHb formation.
Confirming and extending the results of Reddy et al. (1976), Rickert et al. (1983) examined the
role of bacterial nitroreductases in the gastrointestinal tract in altering the absorption of
nitrobenzene. The authors utilized conventional animals and axenic (bacteria-free) animals.
Single oral doses of 22.5 or 225 mg/kg [14C]-labeled nitrobenzene were administered to male
F344 (CDF[F344]/CrlBR), CD (Crl:CD[SD]BR), and axenic CDF(F344)/CrlGN rats and to male
B6C3F1 (B6C3Fl/Crl/BR) mice (225 mg/kg only). Animals were housed in metabolic cages for
72 hours after dosing to collect urine, feces, and expired air. In the conventional rats, 56-65% of
the administered dose was recovered in the urine, with a maximum of 21.4% recovered in the
feces. Six metabolites were found in the bile of conventional rats. Since the metabolites were
absent from the bile of axenic rats, the authors concluded that the reduction of nitrobenzene at
the nitro group that produced metabolites in conventional rats must have been initiated in the
intestines. When corrected for overall recovery, these data provide intestinal absorption
estimates of 62-69% in conventional rats. The estimate from the mouse data was lower (43%).
Albrecht and Neumann (1985) gavaged female Wistar rats with [14C]-nitrobenzene
(25 mg/kg) in propylene glycol and collected blood, tissue, fecal, and urine samples at various
time intervals. Excretion in urine was the major route of elimination, with 50% of the
administered radioactivity excreted in the urine after 24 hours and 65% after 1 week. In contrast
to urine, cumulative fecal excretion of nitrobenzene reached no more than 15.5% of the
administered dose within the same time period. This study, taken together with the above
observations, indicates that nearly two-thirds of orally administered nitrobenzene is absorbed via
the gastrointestinal tract.
3.1.2. Pulmonary Absorption Studies
Several reports from the occupational and clinical research setting have addressed the
pulmonary absorption of nitrobenzene. Ikeda and Kita (1964) discussed the case of 47-year-old
woman who had been exposed via inhalation to paint that contained nitrobenzene. Although her
symptoms were less severe, they were nearly identical to an oral exposure case study by Myslak
et al. (1971) discussed above. The urinary metabolites p-amino and p-nitrophenol demonstrated
absorption of nitrobenzene from the lungs and indicated that these metabolites were formed in
humans after both oral and inhalation exposures. The report from Ikeda and Kita (1964)
suggests that pulmonary absorption of nitrobenzene had occurred, although it is likely that some
dermal absorption had also taken place.
-------
Quantitative estimates of nitrobenzene's pulmonary absorption were provided by
Salmowa et al. (1963), who administered a continuous 6-hour exposure of nitrobenzene
(5-30 ng/L; 1-6 ppm) to seven human research subjects (adult males, age unstated). Subjects
were exposed to nitrobenzene through a mask that also permitted expired air to be collected and
analyzed for nitrobenzene. The amount of nitrobenzene absorbed, estimated as the difference
between the amount inhaled and the amount exhaled, ranged from 8.4-67.6 mg. The retention of
nitrobenzene vapors in the lungs averaged 80%, varying from a mean value of 87% in the first
hour to 73% in the sixth hour.
Piotrowski (1967) also exposed four human research subjects (adult males, age unstated)
to a range of nitrobenzene concentrations in air (5-30 |J,g/L; 1-6 ppm). One was exposed for
6 hours daily for 4 successive days. The remaining three were subjected to longer exposures
lasting Monday through Saturday and, after a pause on Sunday, were exposed again on Monday
of the next week. The absorbed doses of nitrobenzene were estimated from measurements of the
concentrations in the air, the volume of the expired air, and the mean pulmonary retention time
of 80% as determined by Salmowa et al. (1963). The absorbed doses of nitrobenzene were then
compared with the cumulative appearance of nitrobenzene metabolites in the urine. Based on
these data, Beauchamp et al. (1982) determined that humans exposed to an airborne nitrobenzene
concentration of 10 mg/m3 for 6 hours would absorb 18.2-24.7 mg of nitrobenzene through the
lungs.
3.1.3. Dermal Absorption Studies
Data from a number of sources point to the capacity of nitrobenzene to penetrate the
dermal barrier in humans. For example, human research subjects were placed in an exposure
chamber containing nitrobenzene vapor for 6 hours, while receiving fresh air through a breathing
tube and mask (Piotrowski, 1967). The absorption rate per unit of concentration of nitrobenzene
was highly variable (0.23-0.30 mg/hour per ng/L), depending on the nitrobenzene concentration
in the chamber (5-30 ng/L) and whether the subject was dressed or naked. In naked subjects
exposed to a chamber concentration of 10 ng/L nitrobenzene, the absorbed dose ranged from 10-
19 mg compared with 8-16 mg in clothed subjects. Depending on the air concentration (5-
30 ng/L), normal working clothes reduced the overall absorption of nitrobenzene by 20-30%. In
another study involving human research subjects (age and sex not stated), the capacity of
21 organic compounds, including nitrobenzene, to penetrate the dermal barrier as liquid was
surveyed by Feldmann and Maibach (1970), who applied [14C]-labeled compounds in acetone
(4 jig/cm2) to a 13 cm2 circular area of the ventral forearm surface of six subjects. The skin site
was not protected and the subjects were asked not to wash the area for 24 hours. The authors
also examined the elimination of nitrobenzene following intravenous administration as a
comparison with the dermal absorption and elimination studies. For the skin absorption studies,
7
-------
the cumulative amounts of radiolabel measured in urine over 5 days amounted to approximately
1.53 ± 0.84% of the load. The highest rate of absorption was monitored in the first 24-hour
period after application, but excretion in the urine was still measurable between 96 and 120 hours
after application. The absorption rate (percent dose per hour) over the 120-hour period was as
follows: 0.022%/hour: 0-12 hours; 0.022%/hour: 12-24 hours; 0.013%/hour: 24-48 hours;
0.013%/hour: 48-72 hours; 0.011%/hour: 72-96 hours; and 0.006%/hour: 96-120 hours.
Continued excretion of [14C]-label at the later time points may have represented redistribution of
nitrobenzene or its metabolites from adipose tissue rather than continued absorption. Following
intravenous administration of [14C]-nitrobenzene, 60.5% of the radioactive label was detected in
the urine by 20 hours after administration. When corrected for the appearance of nitrobenzene in
urine following an intravenous injection, an overall dermal absorption factor of approximately
2.6% was determined for nitrobenzene.
3.2. DISTRIBUTION
Albrecht and Neumann (1985) exposed female Wistar rats to 25 mg/kg (0.20 mmol/kg)
[14C]-nitrobenzene in propanediol by gavage and reported the appearance of radiolabel,
predominantly in blood, liver, kidney, and lung, 1 and 7 days after dosing. These findings
suggest a wide distribution for nitrobenzene or its metabolites among the major organs and
tissues. Radioactive label (radioactivity in tissue [pmol/mg]/dose [|imol/kg]) recovered from
various tissues was blood (229 ± 48) > kidney (204 ± 27) » liver (129 ± 9.5) » lung (62 ± 14)
after 1 day of exposure. Only about 50 ± 10% of the nitrobenzene appeared in the urine. Seven
days after exposure, tissue levels from highest to lowest were blood (134 ± 19) » kidney
(48 ± 2.4) » lung (29 ± 4.1) ~ liver (26.5 ± 3.5). After seven days, urinary elimination of
nitrobenzene reached 65 ± 5.8%. [14C]-nitrobenzene metabolites were shown to bind with higher
affinity to Hb and plasma proteins than [14C]-acetanilide (0.15 mmol/kg), although in both cases
the reactive metabolite was thought to be nitrosobenzene, and the compound bound to protein
sulfhydryls via a sulfinic acid amide bond was identified as aniline. After 1 day, specific binding
of nitrobenzene to Hb (1,030 ± 137 pmol/mg/dose) and plasma proteins (136 ± 34) was much
higher than acetanilide binding to Hb (177 ± 14) and plasma proteins (70 ± 7). By 7 days
posttreatment, a marginal decrease in the nitrobenzene binding to Hb (1,024 ± 82 pmol/mg/dose)
and plasma proteins (101 ± 34) had occurred, as compared with acetanilide binding to Hb
(102 ± 24) and plasma proteins (14 ± 3). This is the only study of tissue distribution of
nitrobenzene that has been identified.
Goldstein and Rickert (1984) administered a single oral dose of 10 or 40 jiCi
[14C]-nitrobenzene in corn oil to male CDF (F344)/CrlBR rats and B6C3Fl/CrlBR mice with
sufficient carrier nitrobenzene to yield doses ranging from 75-300 mg/kg. The disposition of the
bound radiolabel in red blood cells (RBCs) and spleen proteins was then evaluated after lysates
-------
or homogenates (spleen) were dialyzed, solubilized, and then separated by using sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The amount of covalently bound
radiolabel increased dose dependently in RBCs and spleen for both species. Total and bound
levels of [14C]-label in RBCs from rats were approximately 6-13 times greater than those from
mice at all doses tested. A statistically significant difference between the rat and mouse was
observed, with time for nitrobenzene binding to RBCs and spleen (Figure 3-1). Spleen weights
in rats exposed to 200 mg/kg nitrobenzene increased by up to a factor of two by 168 hours after
dosing; however, there was no equivalent effect in mice (Figure 3-2).
Goldstein and Rickert (1984) used SDS-PAGE to investigate binding of [14C]-
nitrobenzene in the erythrocytes and spleen of rats and mice. SDS-PAGE of RBC lysates from
rats showed that most radioactivity coeluted with Hb. The radioactivity bound to spleen
homogenates coeluted with metFIb1 and an unidentified low molecular weight component. By
contrast, there was no sign of significant macromolecular binding of nitrobenzene-derived
radiolabel in mice. Goldstein and Rickert (1984) hypothesized that the degree of RBC damage
induced by nitrobenzene in mice was insufficient to induce splenic scavenging and clearance
from the systemic circulation.
RBCs
1= ^
W
500n
400-
300-
200-
100-
n
A\ ^B6C3F1
/ \
/ ^-^_^^
r ^~~*—- ^^
1 ^
f
J ~ ^ o
50 100 150
Time (hours)
200
C H
•8^
u ~3
•3.5
75-
50-
25-
Spleen
F344
0 50 100 150
Time (hours)
200
14
Figure 3-1. Time course of covalently bound [C]-nitrobenzene in RBCs and
spleen of rats and mice.
Note: Animals were administered 200 mg/kg [14C] -nitrobenzene and sacrificed at various time
points. Each point represents the mean ± standard error of the mean of three to four
1 MetHb (a greenish-brown to black pigment) may be formed from Hb, which is made of four globin polypeptide
chains, each of which has a single heme group (iron-containing porphyrin) capable of reversibly binding one oxygen
molecule. The ability of Hb to bind and transport oxygen depends on the heme valance (oxidation) state such that
the ferrous iron (Fe+2) in Hb may readily bind oxygen, while formation of metHb, due to loss of an electron from the
heme iron (becoming Fe+3 or ferric iron), causes heme to lose its ability to combine reversibly with and transport
oxygen (Smith, 1996). MetHb reduces tissue oxygenation by two mechanisms: iron in the ferric rather than the
ferrous form is unable to combine with oxygen, and consequently the oxygen-carrying capacity of the blood is
reduced and the presence of oxidized iron changes the heme tetramer in such a way as to reduce oxygen release in
the tissues (i.e., shifts the oxyhemoglobin dissociation curve to the left as in alkalosis) (Ellenhorn et al., 1997).
Additional information on metHb and methemoglobinemia, including clinical effects resulting from different metHb
levels, may be found in footnote 3 and in section 4.5.1.
9
-------
determinations. Statistically significant differences between the F344 rat and B6C3F1 mouse
were noted at all doses tested.
Source: Adapted from Goldstein and Rickert (1984).
Spleen Weight
1500n
ioo(H
a
F344
500r -^B6C3F1
0 50 100 150 200
Time (hours)
Figure 3-2. Time-related changes in spleen weight in rats and mice following
nitrobenzene treatment.
Note: All animals were administered an oral dose of 200 mg/kg nitrobenzene.
Source: Adapted from Goldstein and Rickert (1984).
3.3. METABOLISM
Metabolism of nitrobenzene in mammals involves both oxidation and reduction reactions.
Evidence for this has come from the identification of potential products of nitrobenzene
oxidation and reduction reactions in the urine of humans and animals that had been exposed to
the compound. Oxidation products of nitrobenzene include o-, m-, and p-nitrophenol; reduction
products of nitrobenzene include nitrosobenzene, phenylhydroxylamine, and aniline. The
metabolites from aniline include the following oxidative metabolites: o-, m-, and p-aminophenol,
nitrocatechols, and aniline (Parke, 1956; Robinson et al., 1951). For all metabolites,
involvement in phase II reactions is likely, and the formation and appearance of sulfated or
glucuronidated conjugates has been demonstrated (Figure 3-3) (Rickert, 1987).
The processes driving the metabolism of nitrobenzene in mammals display tissue
specificity. Three primary mechanisms have been identified: reduction to aniline by intestinal
microflora, reduction by hepatic microsomes and in erythrocytes, and oxidative metabolism by
hepatic microsomes.2 First, nitrobenzene has been shown to undergo a three-step, two-electrons-
per-step transfer reduction to aniline in intestinal microflora (Bryant and DeLuca, 1991; Reddy et
al., 1976). The intermediates in this process are nitrosobenzene and phenylhydroxylamine.
The metabolic competency of lung towards nitrobenzene has not been explored; however, as summarized in
sections 3.3.1 and 3.3.2, studies in whole tissues, including lung, are available on nitroreductases towards the
nitroaromatic drug nilutamide.
10
-------
Second, nitrobenzene undergoes a six-step, one-electron-per-step transfer reduction to aniline
that takes place in hepatic microsomes and erythrocytes (Levin and Dent, 1982; Reddy et al.,
1976).
OSO,H
OG1
phenylhvdroxylamine
OS03H
CONJUGATION
NH
p-aminophenol
p-hy dr oxy ac etanilide
and conjugates
OG1
Figure 3-3. Outline of the metabolism of nitrobenzene: a substrate for
oxidation and reduction reactions.
Sources: Adapted from IPCS (2003); Rickert (1987).
As illustrated by Holder (1999), intermediates in the latter process include a nitro anion
free radical, nitrosobenzene, an hydronitroxide free radical, phenylhydroxylamine, and a
theoretical amino-cation free radical. The reductive intermediates have been shown to reverse
chemically (i.e., aniline can oxidize back towards nitrobenzene or any step in between), with the
direction of flow depending on local redox potentials. The first intermediate in the chain, the
nitro anion free radical, may also react nonenzymatically with tissue oxygen to reform
11
-------
nitrobenzene. This "futile loop" generates a superoxide anion in the process (Sealy et al., 1978),
which may undergo dismutation by superoxide dismutase to molecular oxygen and hydrogen
peroxide (Holder, 1999; Mason and Holtzman, 1975a, b). Third, oxidative metabolism to the
nitrophenols takes place in hepatic microsomes, with probable involvement of the cytochrome
P450 family of enzymes. The intermediates in this process are p- and m-nitrophenols of which
the end products are conjugates of phase II enzymes. The process takes place at an even slower
rate than the six-step, one-electron-per-step microsomal reduction of nitrobenzene. Figure 3-3
shows the range of oxidative and reductive products of nitrobenzene that have been
demonstrated (Rickert, 1987).
The metabolic processes undergone by nitrobenzene are important because many of the
toxicological effects of the compound are likely to be triggered by metabolites of nitrobenzene.
For example, there is abundant evidence that methemoglobinemia3 is caused by the interaction of
Hb with the products of nitrobenzene reduction (i.e., nitrosobenzene, phenylhydroxylamine, and
aniline). The current understanding of how metHb is formed from Hb in the presence of these
components is discussed below. Similarly, the formation of a superoxide anion during the
microsomal reduction of nitrobenzene, with subsequent formation of hydrogen peroxide, may
disturb the redox balance of target cells, such as hepatocytes, potentially leading to oxidative
stress (Gutteridge, 1995) (see section 4.6.3).
Methemoglobinemia, which may be defined as a metHb concentration exceeding 2-3% of total Hb (Lee and
Ferguson, 2007; Smith, 1996; Harrison, 1977), arises when the rate of metHb formation exceeds the rate of
reduction of oxidized heme iron, and it can develop by three distinct mechanisms: genetic mutation resulting in the
presence of abnormal Hb, a deficiency of metHb reductase enzyme, and toxin-induced oxidation of Hb. Small
amounts of metHb are continually produced due to autoxidation of Hb during the normal respiratory function of
loading and unloading of oxygen by erythrocytes. A variety of xenobiotics, including nitrobenzene and aromatic
amines, can cause methemoglobinemia by accelerating the oxidation of Hb to metHb, which loses its ability to
combine reversibly with oxygen (also see footnote 1) (Percy et al., 2005; Smith, 1996, Harrison, 1977).
In normal erythrocytes, maintenance of metHb at low levels (<1% of total Hb) is achieved by the steady reduction
of metHb mainly by the NADH-dependent cytochrome b5 metHb reductase (also known as NADH-diaphorase).
Chronic congenital methemoglobinemia is a rare condition that may be caused by an inherited deficiency in this
enzyme, resulting in metHb levels of 15-30%, which makes these individuals particularly susceptible to metHb-
generating chemicals. Also, low levels of functional enzyme render premature babies susceptible to metHb-forming
chemicals. Normal erythrocytes also have another minor pathway, known as NADPH-diaphorase (or flavin
reductase), which can be enhanced by administration of methylene blue (MB) as an antidote for chemically induced
methemoglobinemia, where, acting as an artificial electron acceptor, MB is reduced to leukomethylen blue, which in
turn reduces metHb to Hb (Coleman and Coleman, 1996). The later metHb reduction pathway requires NADPH as
a cofactor, which is normally furnished by the intracellular hexose monophosphate shunt (also called pentose
phosphate shunt), with glucose-6-phosphate dehydrogenase (G6PD) being a key enzyme in this multistep pathway.
A deficiency in G6PD, due to a sex-linked inherited disorder among nearly 100 million people of African, Asian, or
Mediterranean origin, limits the supply of intracellular NADPH, which renders affected individuals more susceptible
to metHb-forming chemicals or drugs. Chemically induced methemoglobinemia in these individuals may not
respond to MB administration, which, if commenced, may even exacerbate possible ongoing hemolytic anemia due
to scarcity of NADPH (Bradberry, 2003; Bloom and Brandt, 2001).
Chemicals that cause methemoglobinemia vary, with some being able to oxidize hemoglobin both in vitro and in
vivo (e.g., sodium nitrite and phenylhydroxylamine). Other chemicals are only active in vivo (e.g., aniline and
nitrobenzene) due to requirement for enzymatic activation. A third group, typified by potassium ferricyanide, is
active in Hb solutions or blood lysates but not in intact cells in vitro (Smith, 1996).
12
-------
3.3.1. Microbial Reduction of Nitrobenzene (the Three-Step, Two-Electrons-per-Step
Transfer Process)
Reduction of nitroaromatic compounds by the two-electron reductive pathway is
catalyzed by a type I (oxygen-insensitive) nitroreductase (Enzyme Commission [EC] 1.6.99.1,
common name reduced nicotine adenine dinucleotide phosphate [NADPH] dehydrogenase).4
This enzyme catalyzes the following general reaction: NADPH + H+ + acceptor = NADP+ +
reduced acceptor (International Union of Biochemistry and Molecular Biology [IUBMB],
2005a). The enzymatic activity for type I nitroreductase is highest in the microflora of the
intestinal tract of male Sprague-Dawley rats; however, organ-specific activities have been
reported (Figure 3-4).
Type I Nitroreductase Activity
g 5°n
TS 40
£ ~S 30
(N / s
1120
-
o
10
S.I. Liver Heart Brain Cecum Lung L.I. Kidney
Figure 3-4. Type I nitroreductase activity in male Sprague-Dawley rats.
Note: Results are expressed as pmol of reduced nilutamide (R-NH2) formed per milligram protein
per minute (mean ± standard error of the mean; n > 4). S.I. = small intestine contents, L.I. = large
intestine contents.
Source: Adapted from Ask et al. (2004).
Some of the earliest evidence to suggest the importance of microbial nitrobenzene
reduction for toxicological outcomes such as metHb formation came from Reddy et al. (1976).
These researchers administered 200 mg/kg nitrobenzene in sesame oil intraperitoneally to four
groups of male Sprague-Dawley rats, either normal, bacteria-free, bacteria-free then acclimatized
in a normal room for 7 days, or normal pretreated with antibiotics. Methemoglobinemia
developed in normal rats and those bacteria-free animals that had been acclimatized in a normal
room (30-40% metHb within 1-2 hours of exposure). When nitrobenzene was given to bacteria-
EC numbers specify enzyme-catalyzed reactions, not specific enzymes.
13
-------
free rats or those pretreated with antibiotics, they did not develop methemoglobinemia. These
data emphasize the importance of microbial reduction of nitrobenzene to the onset of
methemoglobinemia. Reddy et al. (1976) showed the relative importance of exogenous versus
endogenous reductive nitrobenzene metabolism by comparing the rate of synthesis of aniline in
homogenates of liver, kidney, gut wall, and gut contents prepared from animals in various
treatment groups (Table 3-1). Nitroreductase activity was greatest in the gut contents of control
rats. By contrast, this activity was missing in the gut contents of bacteria-free animals.
Table 3-1. Reduction of nitrobenzene by various rat tissue homogenates
Tissue
Liver
Kidney
Gut wall
Gut contents
Aniline formation (nmol/mg protein/hour)3
Bacteria-free
2.0 ±0.2
0.5 ±0.1
2.0 ±0.4
0.2 ±0.0
Bacteria-free (acclimatized)
2.5 ±0.4
0.8 ±0.1
2.0 ±0.6
15.2 ±2.7
Control
3.3 ±0.4
0.7 ±0.4
2.4 ±1.0
11.1±3.3
"Results are means ± standard error of the means of determinations in three animals/group, with all
determinations in triplicate.
Source: Reddy et al. (1976).
Facchini and Griffiths (1981) demonstrated that little or no metHb was formed when
blood was incubated with nitrobenzene in vitro.5 Their results, taken together with their
in vivo findings with axenic animals (Table 3-2), confirm the importance of microbial
reductive metabolism in the formation of metHb, specifically through the formation of
nitrosobenzene, phenylhydroxylamine, or aniline.
While nitrobenzene is generally regarded as unable to directly oxidize Hb in whole blood cells, due to the absence
of nitroreductases normally present in bacteria, mitochondria, or microsomes, one study reported metHb formation
(10% above background oxidation) when human or rabbit hemolysates were incubated for 5 hours with 5 mM
nitrobenzene (Kusumoto and Nakajima, 1970 [based on Vasquez et al., 1995]). Though these findings may indicate
direct oxidation (i.e., no prior metabolic activation was required), it is noteworthy that several factors may
distinguish this in vitro finding from whole animal studies in that the process of metHb formation here seems to be
slow and it requires high concentrations of substrate, which, in itself, may have caused dissociation of the tetrameric
structure of Hb and/or possibly lowered its redox potential, resulting in a more direct access and permissible electron
transfer between nitrobenzene and the heme moiety.
A similar in vitro study utilizing bovine Hb (0.01-0.02 mM) and dinitrobenzene (DNB) isomers (o-, m-, or p-)
(120 uM) demonstrated direct formation of metHb from deoxyhemoglobin (dxHb) but not from oxy- or carboxy-
hemoglobin (Vasquez et al., 1995). The study concluded that DNB can oxidize dxHb to metHb directly without the
prior metabolic activation inferred from in vivo studies; however, the exact mechanism was not known, even though
lack of covalent binding of DNB to the heme or globin was considered suggestive of a direct redox reaction
(electron transfer) occurring between the heme-bound iron and DNB. It should also be noted that, very likely, DNB
has a higher redox potential and, therefore, a greater tendency to be reduced by Hb (resulting in direct metHb
formation) than nitrobenzene.
14
-------
Table 3-2. MetHb formation in the blood of rats dosed intraperitoneally with
200 mg/kg nitrobenzene in corn oil
Time after dosing (hours)
1
2.5
5
8
MetHb formation (%)a
Control rats
18.2 ±5.0
24.7 ±4.2
32.7 ±5.0
9.9 ±2.3
Antibiotic-treated rats
1.7 ±0.4
2.1 ±0.2
1.9 ±0.4
0.4 ±0.1
"Results are means ± standard error of the means, three animals/group.
Source: Facchini and Griffiths (1981).
Goldstein et al. (1984) fed male CDF(F344)/CrlBR rats diets containing pectin (a
carbohydrate with nutritional value for microflora) or cellulose (a metabolically inert
carbohydrate) for 28 days prior to administering a single 200 mg/kg dose of [14C]-nitrobenzene
via gavage. Levels of metHb were monitored in the blood 1, 2, 4, 8, and 24 hours after dosing.
Rats receiving the pectin-spiked diet had elevated metHb in the blood, with levels peaking at the
4-hour time point. However, no metHb was formed in the blood of animals receiving the
cellulose-containing diet. The authors correlated these findings with the greater numbers of
anaerobic bacteria present in the cecum of rats receiving the pectin-containing diets. As shown
in Table 3-3, [14C]-nitrobenzene was metabolized in vitro in the presence of gut contents from
animals exposed to the subject diets. Metabolites included aniline, nitrosobenzene, and
azoxybenzene, with larger amounts measured in those incubations containing pectin-enriched gut
contents.
Table 3-3. Formation of metabolites of nitrobenzene in the presence of cecal
contents in vitro: influence of diet
Diet
Nffl-07
AIN-76A
AIN-76A
Pectin (%)
8.4
5 (added)
0
Metabolite formation (percent total radioactivity)3
Aniline
36 ± 10b
11±4
3±1
Nitrosobenzene
7±0b
3±2
0±0
Azoxybenzene
7±lb
3±2
0±0
Nitrobenzene
34±llb
78 ±11
95 ± 2
"Values are means ± standard error of the means of four determinations.
bSignificantly different from AIN-76A.
Source: Goldstein etal. (1984).
Experiments of Levin and Dent (1982), pertaining to the influence of gut microflora on
the metabolism of nitrobenzene, included an in vivo protocol in which normal or antibiotic-
treated male F344 (COBS CDF/CrlBR) rats were gavaged with 225 mg/kg nitrobenzene
(containing 0.1 |iCi/mg [14C]-nitrobenzene). Rats were kept in metabolic cages for up to
15
-------
72 hours after treatment, during which urine, feces, and expired air were collected. To the extent
possible, the excretory products were characterized and measured by high-performance liquid
chromatography. As shown in Table 3-4, a statistically significant decrease in p-hydroxy-
acetanilide (a reductive metabolite of nitrobenzene) and a slight increase in p- and m-nitrophenol
(oxidative metabolites) were observed in antibiotic-treated rats versus controls. Antibiotic
pretreatment ameliorated the nitrobenzene-induced methemoglobinemia following a single oral
dose of 300 mg/kg. Moreover, antibiotic-treated animals exposed to 300 mg/kg nitrobenzene
had metHb concentrations of 2.1 ± 0.4%, 2.8 ± 0.5%, and 1.9 ± 1.9% at 6, 24, and 96 hours after
the dose. However, nitrobenzene-exposed vehicle-control rats still had elevated metHb
concentrations (20.0 ± 7.9%) 96 hours after the dose.
rl4x
Table 3-4. Urinary metabolites of [ C]-nitrobenzene excreted within
72 hours after gavage
Metabolite
p-Nitrophenol
m-Nitrophenol
p-Hydroxy-acetanilide
Unidentified peak I
Unidentified peak II
Total recovered
Percent of total"
Control rats
22.4 ±0.9
11.4 ±0.6
16.2 ± 1.7
4.5 ±0.3
3.7 ±0.6
58.2
Antibiotic-treated rats
26.5 ±3.8
16.1 ±2.0
0.9±0.0b
5.5 ±0.9
0.5±0.1b
49.5
"Values are means ± standard deviations for three animals/group.
bSignificantly different from controls.
Source: Levin and Dent (1982).
Collectively, the findings of the studies by Levine and Dent (1982) and by Goldstein et
al. (1984) indicate that lack of metHb formation in germ-free animals is not caused by lack of
absorption or bioavailability of nitrobenzene, due for instance to intestinal physiological changes
in the germ-free animals, but is rather due to absent gut bacteria that have the nitroreductases
needed to activate nitrobenzene to metHb forming metabolites.
Bryant and DeLuca (1991) purified and characterized an oxygen-insensitive nicotinamide
adenine dinucleotide (phosphate) (NAD[P]H)-dependent nitroreductase from Enterobacter
cloacae, which they considered to be typical of enteric bacterial nitroreductases that have been
identified in a number of microbial genera. This enzyme was shown to act through an obligatory
two-electron transfer mechanism. Figure 3-5 illustrates the three-step, two-electrons-per-step
reduction process for nitrobenzene in the intestinal microflora.
16
-------
O
2e-
HOH
2e-
2e-
Nitrobenzene
Nitrosobenzene
Phenylhydroxylamine
Aniline
Figure 3-5. Mechanism of bacterial nitrobenzene reduction.
Source: Adapted from Holder (1999).
3.3.2. Hepatic and Erythrocytic Reduction of Nitrobenzene (the Six-Step, One-Electron-
per-Step Transfer Process)
Reduction of nitroaromatic compounds by the one-electron reductive pathway is
catalyzed by a type II (oxygen-sensitive) nitroreductase (EC 1.6.99.3; common name NADH
dehydrogenase) (IUBMB, 2005b). A mitochondrial form of type II nitroreductase (EC 1.6.5.3;
common name NADH dehydrogenase [ubiquinone]) catalyzes a similar one-electron addition
(IUBMB, 2005c). Type II nitroreductases catalyze the following general reaction: NADH + H+
+ acceptor = NAD+ + reduced acceptor.
Type II nitroreductase activity is highest in the microflora of the intestinal tract of male
Sprague-Dawley rats; however, organ-specific activities have been reported (Figure 3-6).
300n
Brain Lung Heart L.I. Kidney Liver Cecum
Figure 3-6. Type II nitroreductase activity of male Sprague-Dawley rats.
Note: Results are expressed as pmol of reduced nirutamide (R-NH2) formed per milligram protein
per minute (mean ± standard error of the mean; n >_4). S.I. = small intestine; L.I. = large intestine.
Source: Adapted from Ask et al. (2004).
17
-------
The findings of Reddy et al. (1976) and Levin and Dent (1982) can be interpreted to
suggest differences in the kinetics and mechanisms of action of bacterial versus hepatic
microsomal nitroreductases. For example, when Levin and Dent (1982) incubated nitrobenzene
(100 |j,M) under aerobic or anaerobic conditions (e.g., oxygen-scavenging system used) with
microsomes or 9,000 g supernatant fractions prepared from the livers of phenobarbital-induced
male F344 (COBS CDF/CrlBR) rats, metabolism of nitrobenzene by hepatic microsomes was
extremely slow under aerobic conditions (0.022 nmol/minute-mg protein) compared to anaerobic
conditions (0.33 nmol/minute-mg protein). In contrast, the rate of reduction of nitrobenzene by
cecal microflora, which contains an oxygen-insensitive nitroreductase, was 150 times that in
microsomes when expressed as nmol of product/minute-g of liver (4.4 ± 0.1) or cecal contents
(668 ± 74). The masses of liver and cecal contents in a 200 g rat are approximately equal, so that
the cecal contents would represent the major site of reductive metabolism in vivo.
The use of electron spin resonance (ESR) spectrometry by Mason and Holtzman
(1975a, b) on the reaction products of in vitro incubations of rat hepatic microsomes,
mitochondria, or 165,000 g supernatants incubated with nitrobenzene or p-nitrobenzoic acid
demonstrated the formation of nitroaromatic radicals. The authors suggested that these
components were likely to be the first intermediates in the reduction of the respective substrates.
The appearance of nitroaromatic radicals would be consistent with a six-step, one-electron-per-
step reduction mechanism for the microsomal metabolism of nitroarenes, such as nitrobenzene.
Sealy et al. (1978) used the same incubation system as Mason and Holtzman (1975a, b) with the
substrates nitrofurantoin, nitrofurazone, misonidazole, or nitrobenzoate but added the spin traps
5,5-dimethyl-l-pyrroline-l-oxide (DMPO) or phenyl-N-t-butyl nitrone shortly before the
addition of microsomes. The resulting spectra were consistent with the formation and reaction of
superoxide anion with the spin traps to give relatively long-lived nitroxide adducts with a
characteristic ESR spectrum. These results suggested that compound-specific nitro anion
radicals had been rapidly converted by molecular oxygen to the parent nitroarene with the
formation of a superoxide anion. The reconversion to the nitroarenes was an experimental
demonstration of the futile cycle by which reduced coenzymes are expended in the presence of
endogenous nitrobenzene, with the concomitant production of superoxide radical and possibly
hydrogen peroxide. A metabolic chart in Holder (1999) summarizes the six one-electron
reduction step process for nitrobenzene reduction (Figure 3-7).
18
-------
superoxide free radical Qo"
futile reaction (reforms nitrobenzene again)
Nitroreductase
nitroanion free radical
Net Reduction
P45o Flavins
NAD(P)H
Aniline
fast\
Nitrosobenzene
aminocation
H,O
NH O
Phenylhydroxylamine
hydronitroxide free radical
Figure 3-7. Mechanism of microsomal nitrobenzene reduction.
Source: Adapted from Holder (1999).
The scheme captures the series of five intermediate compounds and/or radicals to form
aniline, with the additional potential for the first product of the process, the nitro anion free
radical, to be reoxidized to nitrobenzene with the formation of a superoxide anion. Superoxide
dismutase can rapidly convert superoxide anion to hydrogen peroxide, which in turn may be
converted to oxygen and water by catalase or conjugated with glutathione by glutathione
peroxidase, thereby forming glutathione disulfide and water (Table 3-5).
19
-------
Table 3-5. Enzyme systems in erythrocytes
Enzyme (reference)
Superoxide dismutase (IUBMB, 2005d)
Glutathione peroxidase (IUBMB, 2005e)
Catalase (IUBMB, 2005f)
Glutathione transferase (IUBMB, 2005g)
Glutathione reductase (IUBMB, 2005h)
NADPH-cytochrome c reductase (IUBMB, 2005i)
NADH-cytochrome bs reductase (IUBMB, 2005J)
EC number3
EC 1.15. 1.1
EC 1.11. 1.9
EC 1.11. 1.6
EC 2.5. 1.18
EC 1.8.1.7
EC 1.6.2.4
EC 1.6.2.2
Reaction
2 O2' + 2 H+ = O2 + H2O2
2 glutathione + H2O2
= glutathione disulfide + 2 H2O
2 H2O2 = O2 + 2 H2O
RX + glutathione = HX + R-S -glutathione
2 glutathione + NADP+
= glutathione disulfide + NADPH + H+
NADPH + H+ + n oxidized hemoprotein
= NADP+ + n reduced hemoprotein
NADH + H+ + 2 ferricytochrome bs
= NAD+ + 2 ferrocytochrome b5
aEC numbers specify enzyme catalyzed reactions, not specific enzymes (Bairoch, 2000).
Mason and Holtzman (1975a, b) discussed available information on the biochemical
characteristics of hepatic microsomal nitrobenzene reductases. The activities were thought to
consist of one or more flavoproteins that represent only single electron-to-electron acceptors.
The authors speculated that the microsomal flavoenzymes NADPH-cytochrome c reductase
(EC 1.6.2.4) and NADH-cytochrome bs reductase (EC 1.6.2.2) may be the enzymes responsible
for the reduction of nitroarenes to their anion radicals (Table 3-5).
Harada and Omura (1980) provided data that addressed this issue by monitoring the formation of
aniline, nitrosobenzene, and phenylhydroxylamine in hepatic microsomes that were incubated in
the presence of antibodies to NADPH-cytochrome c reductase, NADH-cytochrome b5 reductase,
cytochrome &5, or cytochrome P450 (subfamily not stated). When incubated with antibodies to
NADPH-cytochrome c reductase and cytochrome P450, the activities of NADPH- and NADH-
dependent nitrobenzene reductases were inhibited, with concomitant blockage of the
nitrosobenzene and phenylhydroxylamine formation. However, antibodies to NADH-
cytochrome bs reductase or cytochrome bs were ineffective. The initial step in nitrobenzene
reduction appeared to be catalyzed by NADPH-cytochrome c reductase, with cytochrome P450
playing a role in the final conversion of the intermediates to aniline.
In addition to the hepatic microsomal reduction of nitrobenzene, the reductive
metabolism in erythrocytes has been extensively studied due to the propensity of nitrobenzene
metabolites to form metHb. Mammalian RBCs are particularly susceptible to oxidative damage
because, being an oxygen carrier, they are exposed uninterruptedly to high oxygen tension,
RBCs have no capacity to repair damaged components, and Hb is susceptible to autoxidation
while its membrane components are susceptible to lipid peroxidation (Rice-Evans, 1990).
Several biochemical changes occur in the human RBC during its entire lifespan of about 120
days; for example, there are changes in lipid and protein content of the membrane and in enzyme
20
-------
activities, ion permeability, size, and deformability (Clark and Shohet, 1985; Westerman et al.,
1963). At the end of its life span, the erythrocyte is phagocytized by macrophages,
predominantly in the spleen. This latter event can lead to splenic congestion in rats following
acute treatment with nitrobenzene due to the increased fragility of RBCs and the ultimate
increase in splenic scavenging and clearance from the systemic circulation (Goldstein and
Rickert, 1984).
The particular redox chemistry associated with nitrobenzene metabolism in RBCs is of
special interest because of its association with the development of methemoglobinemia. The
work of Reddy et al. (1976) has pointed to an association of metHb formation with the reduction
of nitrobenzene to nitrosobenzene, phenylhydroxylamine, and aniline by nitroreductases present
within intestinal microflora. Moreover, in vitro incubation of RBCs with nitrobenzene does not
result in the formation of metHb (Facchini and Griffiths, 1981). Taken together, these findings
suggest that it is the presence and cycling of the reductive products of nitrobenzene within RBCs
that cause the conversion of oxyhemoglobin (oxyHb) to metHb (Figure 3-8).
CD
Hemopathy -denatured
hemoglobins with RBC lysis
methemoglobin reductase
CD
normal RBCs
Hb-Cysteinesulphinamide
"methemoglobin v
metHb
CD
CD
CD
nitroBeHzene. exposure
o ° CD?"
some affected RBCs *
Phenvlhvdroxvlamme
GS_Nitrosobenzene
(GS_NOB)
\
Glutathionesulphinamide
(GSO_AN)
Figure 3-8. Cycling of nitrosobenzene and phenylhydroxylamine in RBCs,
resulting in the formation of metHb.
Note: GSH = reduced glutathione, GSSG = oxidized glutathione; GS = glutathionyl conjugate.
Source: Adapted from Holder (1999).
21
-------
The primary metabolic event in the formation of metHb (Fe3+) from oxyHb (Fe2+) as a
result of nitrobenzene exposure is the cycling between phenylhydroxylamine and
nitrosobenzene. As explained by Maples et al. (1990), nitrosobenzene can be reduced
nonenzymatically by endogenous reducing agents or enzymatically by NADH-cytochrome &s
reductase to reform phenylhydroxylamine. This completes the redox cycle with the overall
expenditure of NADH and the accumulation of metHb. Nitrosobenzene has been shown to
participate in a number of reactions that adversely affect the metabolic balance of RBCs. For
example, nitrosobenzene has a 14-fold higher binding affinity to the heme moiety of Fib than
does molecular oxygen (Eyer and Ascherl, 1987). It is also thought to promote the dissociation
of tetrameric Hb to its constituent dimers (Eyer and Ascherl, 1987). Nitrosobenzene can also
bind to peptides and proteins carrying cysteine residues, including Hb and reduced glutathione
(GSH) (Eyer, 1979). The consequences of the latter interaction potentially include the formation
of sulfhemoglobin, the formation of an oxidized dimer of glutathione with reformation of
phenylhydroxylamine, or rearrangement to form GSH sulfinamide. Furthermore, an overall
depletion of GSH may result from excessive cycling of nitrosobenzene.
Maples et al. (1990) used ESR to demonstrate the formation of a phenylhydronitroxide
free radical during the phenylhydroxylamine-initiated reduction of oxyHb. The use of DMPO as
a spin trap further demonstrated the transfer of a free electron to cysteine-carrying components,
such as GSH and Hb, with the formation of their respective thiyl radicals, GS» and HbSv These
moieties are likely to be highly reactive, with the capacity to transfer the unpaired electron to
other subcellular components. Continuous recycling of phenylhydroxylamine and
nitrosobenzene may lead to increased fragility of RBC membranes, premature scavenging, and
destruction within the reticuloendothelial system, followed by engorgement and sinusoid
congestion of the spleen (Chemical Industry Institute of Toxicology [CUT], 1993; Goldstein and
Rickert, 1984).
3.3.3. Microsomal Oxidation of Nitrobenzene
Oxidation of nitrobenzene can generally occur via hydroxylation of the benzene ring
(usually at positions 3 or 4) forming nitrophenols or after initial nitroreduction of the exocyclic
nitro group to the amine by oxidation to phenylhydroxylamine. These reactions are thought to be
mediated by microsomal enzymes.
The appearance of conjugated derivatives of nitrophenols in the urine of female giant
chinchilla rabbits having received an oral dose of nitrobenzene (0.5 g in 25 mL water by stomach
tube) implied that the compound can undergo oxidation reactions in addition to the more
extensively characterized reduction reactions that are discussed above (Robinson et al., 1951). A
greater range of both oxidation and reduction metabolites was formed when rabbits (strain and
sex not stated) were given a single oral dose of [14C]-nitrobenzene and unlabeled nitrobenzene at
22
-------
total doses of 200 mg/kg (two animals) and 250 mg/kg (three animals) (Parke, 1956). Although
the mechanism of microsomal oxidation of nitrobenzene has not been well characterized, the
involvement of members of the cytochrome P450 family is likely (IPCS, 2003). While it is
probable that not all active subcellular sites involved in nitrobenzene oxidation have been
identified, the overall rate of oxidative metabolism is thought to be very slow. However,
oxidation products of nitrobenzene such as p- and m-nitrophenol have been detected in the urine
of subjects exposed to nitrobenzene by inhalation (5-30 |ig/L; 1-6 ppm) for 6 hours, suggesting
that oxidation reactions do play a role in the metabolism of nitrobenzene in vivo (Salmowa et al.,
1963).
3.4. ELIMINATION
The major route of elimination for nitrobenzene in humans and animals is urine (Albrecht
and Neumann, 1985; Rickert et al., 1983), with the majority of the dose eliminated within
48 hours. For example, a subject who ingested about 50 mL of nitrobenzene, as reported by
Myslak et al. (1971), showed extensive excretion of the nitrobenzene metabolites,
p-aminophenol and p-nitrophenol, in the urine. These reached maximum levels on day 2 for
p-aminophenol (198 mg/day) and on day 3 for p-nitrophenol (512 mg/day). As discussed in
section 3.1, Ikeda and Kita (1964) detected the same compounds in the urine of a woman who
was exposed to nitrobenzene, primarily by inhalation, in an occupational setting. However,
Salmowa et al. (1963) detected p-nitrophenol, but not p-aminophenol, in the urine of human
research subjects exposed to nitrobenzene via inhalation.
p-Nitrophenol was also detected in the urine of subjects exposed to nitrobenzene through
the skin (Piotrowski, 1967). In a quantitative study using human research subjects (section 3.1),
Feldmann and Maibach (1970) applied [14C]-labeled nitrobenzene (50 jig dissolved in acetone)
to the forearm skin of six subjects. As noted earlier, an estimated 2.6% of the dose was absorbed
through the skin. Excretion of nitrobenzene-derived radiolabel in urine over 5 days was
1.5 ± 0.84% of the dose or about 58% of the absorbed dose. The highest rate of absorption
occurred during the first 24 hours after dosing, but radioactivity could be detected in urine for
96-120 hours after application. Following intravenous administration of [14C]-nitrobenzene,
60.5% of the radioactive label was detected in the urine by 20 hours after administration,
confirming the high rate of urinary excretion of nitrobenzene in humans.
Robinson et al. (1951) studied the metabolism of nitrobenzene in the giant chinchilla
rabbit. Their results demonstrated that urine was a major excretion pathway, with 45% of the
radioactivity following a [14C]-nitrobenzene dose excreted in urine within 72 hours. Parke
(1956), using [14C]-nitrobenzene, was able to demonstrate in rabbits that 0.6-0.7% of the
radioactivity from various doses was eliminated via exhaled air as parent compound, up to 1.2%
as CC>2, and a very small amount (0.04% at best) as aniline.
23
-------
As discussed in section 3.3, the study by Levin and Dent (1982) on nitrobenzene
metabolism in rats also determined levels of fecal, urinary, and exhalatory excretion. Values for
the recovery of radiolabel in feces and expired air were 16.4 ± 2.2% and 2.3 ± 0.5%,
respectively, for control rats and 12.5 ± 3.6% and 3.4 ± 1.5%, respectively, for antibiotic-treated
animals. The observed metabolites were present in urine as sulfate conjugates.
Rickert et al. (1983) exposed male F344 (CDF[F344]/CrlBR) rats, male CD
(Crl:CD[SD]BR) rats, and male B6C3F1 (B6C3Fl/Crl/BR) mice to single doses of 22.5 (oral) or
225 mg/kg (oral or intraperitoneal [i.p.]) nitrobenzene (containing 20 jiCi [14C]-nitrobenzene) in
corn oil. Samples of feces, urine, and expired air were collected at various time points up to
72 hours. Urinary metabolites of nitrobenzene were identified after incubation with
p-glucuronidase and/or sulfatase. The disposition of radiolabeled products among feces, urine,
and expired air 72 hours after dosing is shown in Table 3-6, corroborating urine as the primary
route of excretion in all exposed groups. Species and strain differences were evident in the
degree of conjugation exhibited by nitrobenzene metabolites (Table 3-7). In F344 rats, all
nitrobenzene metabolites were conjugated as sulfates, confirming the observation of Levin and
Dent (1982). By contrast, the urine of CD rats and B6C3F1 mice contained sulfate and
glucuronide conjugates as well as free product. p-Aminophenol was detected only in the urine of
mice.
Table 3-6. Recovery of radiolabel in F344 and CD rats and B6C3F1 mice
72 hours after exposure to a single oral dose of [14C]-nitrobenzene
Excretory
product
Urine
Feces
Expired air
Total
Percentage of dose recovered
F344 rat
225 mg/kg
oral
63.2 ±2.1
14.2 ±0.7
1.6 ±0.1
79.0 ±2.2
225 mg/kg
i.p.
56.8 ±0.9
13.7 ±1.8
1.4 ±0.1
71.9 ±2.6
22.5 mg/kg
oral
65.8 ±2.4
21.4±1.8a
1.0 ±0.6
88.2±1.8a
CD rat
225 mg/kg
oral
60.8 ±1.1
11.8±1.1
2.5 ±0.3
75.1 ±1.1
22.5 mg/kg
oral
64.5 ±0.8
11.5±0.1
0.8 ±0.2
76.8 ±1.0
B6C3F1 mouse
225 mg/kg
oral
34.7 ±4.8
18.8±0.4a
0.8 ±0.1
54.3 ± 4.7a
"Significantly different from F344 rats given 225 mg/kg orally.
Source: Rickert etal. (1983).
Albrecht and Neumann (1985) administered a single dose of 25 mg/kg nitrobenzene by
gavage to female Wistar rats. They found that 50% of the dose was eliminated via urine within
the first 24 hours and a total of 65% of the dose was excreted in urine within 1 week. Only
15.5% of the dose was eliminated in the feces within 1 week after dosing.
24
-------
Table 3-7. Urinary excretion of nitrobenzene metabolites in male rats and
14,
mice gavaged with a single oral dose of [ C]-nitrobenzene
Compound
p-Hydroxyacetanilide
p-Aminophenol
p-Nitrophenol
m-Nitrophenol
Unidentified peak I
Unidentified peak II
Free/
conjugate
Free
Glucuronide
Sulfate
Free
Glucuronide
Sulfate
Free
Glucuronide
Sulfate
Free
Glucuronide
Sulfate
Total
Total
Percentage of dose"
F344 rat (mg/kg)
225
b
-
19.0 ±0.9
-
-
-
-
-
19.9 ± 1.1
-
-
10.2 ±0.6
9.8 ±0.7
-
22.5
-
-
19.8 ±2.8
-
-
-
-
-
23.3 ±2.1
-
-
11.6±1.4
9.0 ±0.5
-
CD rat (mg/kg)
225
1.3 ±0.2
1.8 ±0.6
5. 8 ±1.2
-
-
-
2.2 ±0.6
0.5 ±0.1
10.3 ±2.9
1.2 ±0.4
0.5 ±0.2
6.2 ±1.7
25.3 ±1.2
5.7 ±4.0
22.5
0.9 ±0.2
l.liO.l
1.7 ±0.9
-
-
-
0.7 ±0.2
0.6 ±0.0
5.6 ±1.8
0.4 ±0.1
0.5 ±0.1
3. 8 ±1.2
31.1±2.1
16.4 ±5.6
B6C3F1 mouse
(mg/kg)
25
0.4 ±0.0
3.1±0.3
0.4 ±0.1
0.1 ±0.1
0.2 ±0.2
9.4 ±1.3
0.8 ±0.1
O.liO.l
6.3 ±1.1
O.liO.l
-
6.1 ±1.2
4.8 ±0.7
2.6 ±0.2
aValues are means ± standard error of the means for three animals/group over a 72-hour period.
b- = not detected.
Source: Rickertetal. (1983).
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
No studies were located that addressed the toxicokinetics of nitrobenzene as applicable to
physiologically based pharmacokinetic (PBPK) modeling of the compound.
25
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY AND CASE REPORTS
There are no reports of epidemiological studies of the human health impacts of
nitrobenzene exposure in the workplace or environment. However, a number of case reports of
nitrobenzene poisoning have been published in the biomedical literature. As described in the
following sections, nitrobenzene induces a suite of well-characterized toxicological responses
irrespective of the route of exposure—oral, inhalation, or dermal. Some toxicokinetic
information on nitrobenzene has also emerged from studies in which nitrobenzene was
administered to human research subjects (see section 3).
4.1.1. Oral Exposure
Schimelman et al. (1978) reported on a 48-year-old man who was taken to an emergency
room 10 minutes after consuming approximately 300 mL of Hoppe's Gunpowder Solvent #9
(30% denatured ethyl alcohol, 30% kerosene, 20-35% essential oils and fatty oil base, 3%
ammonia, and 2% nitrobenzene by volume). Upon arrival, the patient was cyanotic, and his
respiration was shallow and irregular. Blood was obtained and was dark brown in color, and
methylene blue was administered.6 MetHb level before and after treatment was 75%. The
patient underwent seven blood transfusions, after which the level of metHb in the blood
gradually declined. Six hours following arrival at the emergency room, his metHb level was
33%. Five days after admission, the patient continued follow-up for a mild poison-induced
hemolytic anemia.
Section 3.1 discusses a case report by Myslak et al. (1971) in which a 19-year-old female
consumed approximately 50 mL nitrobenzene. The resulting acute symptoms of toxicity
included cyanosis, unconsciousness, and severe methemoglobinemia (82% about 90 minutes
after consumption of nitrobenzene), and the patient initially had a distinct smell of bitter almonds
on the expired breath. This report is typical of accounts in which subjects have experienced
nitrobenzene-induced toxicosis through consuming nitrobenzene-containing substances.
Harrison (1977) described the case of a 19-year-old male who consumed a brown liquid
while pipetting that apparently contained nitrobenzene. The time between ingestion and hospital
admission was approximately 1.5-2 hours. On examination, the patient was unconscious, his
6 Therapeutic interventions for methemoglobinemia include the administration of redox scavengers, with ascorbic
acid and/or methylene blue (MB). Ascorbic acid infusion results in acidosis and a resultant shift of the oxygen
dissociation curve to the right, which improves oxygen delivery to the tissues. MB (CASRN 61-73-4) is the antidote
of choice for methemoglobinemia. The recommended dose is 1 mg/kg over a period of 5 minutes. At high levels of
metHb, MB reduces the half-life of metHb from 15-20 hours to 40-90 minutes. MB acts as a cofactorto increase
erythrocyte reduction of metHb to oxyHb in the presence of NADPH, utilizing the hexose monophosphate shunt
pathway. The MB is reduced to leucomethylene blue, which is the electron donor for the nonenzymatic reduction of
metHb to Hb (DiSanto and Wagner, 1972).
26
-------
lips, tongue, and mucous membranes were navy blue, almost black, and his skin was slate gray.
A strong smell similar to that of mothballs or bitter almond was noted. Profound signs of
methemoglobinemia were associated with an initial metHb level of 65% and the characteristic
chocolate brown coloration of the blood. The patient underwent gastric lavage and received
intravenous administration of methylene blue, ascorbic acid, methylprednisolone, and diazepam.
Analysis of gastric aspirate revealed the presence of aniline and nitrobenzene. Approximately
12 hours after admission and following exchange transfusion, the patient's metHb was 25%.
Seven days after admission, hemolytic anemia became apparent. Following blood transfusions,
the patient ultimately had an uneventful recovery and was discharged after 19 days.
The characteristic signs of acute nitrobenzene poisoning (coma, cyanosis, a smell of bitter
almonds on the breath) were evident in a 24-year-old female who had ingested an unreported
quantity of nitrobenzene (Ajmani et al., 1986). As in other cases, the patient was responsive to a
treatment protocol featuring gastric lavage, intravenous fluids, methylene blue, ascorbic acid,
and diuretics. During day 6 of the recovery phase, the subject developed mild jaundice and
anemia, yet fully recovered within 2 weeks.
Kumar et al. (1990) described a 21-year-old male who was taken to an intensive care unit
approximately 30 minutes after relatives said he consumed 30-40 mL of "varnish," a
nitrobenzene-containing dye used in screen printing. On arrival, the patient was in a deeply
comatose state with very shallow breathing. Blood samples were obtained that were dark brown
in color, and a diagnosis of methemoglobinemia was made, secondary to nitrobenzene
consumption, when there was no change in the blood sample color after it was placed on white
filter paper and bubbled with oxygen. Gastric lavage was performed, and ascorbic acid and
methylene blue were administered intravenously. A second dose of methylene blue was
administered after 50 minutes. The patient's metHb measurement was repeated 2 hours after the
second dose of methylene blue and was 5.7%. After the fifth day of admission, the patient was
discharged.
Abbinante et al. (1997) reported nine cases of nitrobenzene poisoning in Venezuela
between April and July 1993 in people ingesting bitter almond oil containing nitrobenzene. A
range of clinical manifestations was observed in affected subjects, including vomiting, dizziness,
cyanosis (oral, distal, or general), respiratory depression, convulsions, and generalized weakness.
Biochemical findings included anemia, hemolysis, and high levels of metHb. Nuclear magnetic
resonance and infrared spectroscopy were used to analyze the almond oil samples and positively
confirmed the presence of nitrobenzene.
Two articles by Chongtham et al. (1999, 1997) describe a 24-year-old female whose
metHb level was measured as 56.5% as a result of drinking nitrobenzene. The patient was
cyanotic and gasping and had a pulse of 120/minute. In response to the usual range of palliative
27
-------
and corrective measures (gastric aspiration, lavage, intravenous methylene blue, and ascorbic
acid), the subject's metHb level was 5% after 3 days of intensive treatment and care.
Wentworth et al. (1999) described the case of a 2-year-old girl who presented with toxic
methemoglobinemia, most likely as a result of consuming a nitrobenzene-containing product.
The patient was in shock, with marked cyanosis, a heart rate of 170 beats/minute, blood pressure
of 80/50 mm Hg, a respiration rate of 28/minute, and a grade II systolic murmur. While the
precise source of the toxicosis remained unknown, nitrobenzene ingestion was suspected and the
usual suite of palliative and remedial measures to reduce the patient's 41% metHb level were
undertaken. Gupta et al. (2000) reported the case of a 5-year-old boy who died as a result of
consuming some screen-printing material that contained nitrobenzene. The level of
methemoglobinemia was not reported. The patient showed an initial improvement as a result of
gastric lavage and oral administration of vitamin C (methylene blue was not given in this case).
However, the patient later died of cardiac arrest. Table 4-1 presents a chronological compilation
of the cases reported in this section.
4.1.2. Inhalation Exposure
As discussed in section 3.1, exposure of human research subjects to nitrobenzene vapor
resulted in an average absorption of 87% at the blood:gas barrier (Salmowa et al., 1963).
However, no case reports were identified that addressed the toxicity of nitrobenzene solely via the
inhalation route. For example, the incident described by Ikeda and Kita (1964) most likely also
involved dermal contact (section 3.1). The patient presented with a range of typical symptoms of
nitrobenzene toxicosis, including headache, nausea, weakness, hyperalgesia, and cyanosis. The
woman had been employed for 17 months in a small paint firm where she painted and polished
lids of pans with a red paint containing nitrobenzene as a solvent. The authors determined the
nitrobenzene content of the paint solvent to be 97.7% by gas chromatography. Apparently, the
workshop was remodeled, and the ventilation became quite poor. The patient started to complain
of severe headache, nausea, vertigo, and numbness in the legs approximately 2 months later.
After 5 days of bed rest, she returned to work. Nearly 3 months later, the patient experienced the
same bout of symptoms, and she was admitted to the hospital the following day. On physical
examination, she was emaciated and in a state of distress. Her lips and oral mucosa were
cyanotic and the sclerae were slightly jaundiced. The liver and spleen were palpable. During the
woman's 2-week stay in the hospital, the nitrobenzene metabolites p-amino- and p-nitrophenol
gradually disappeared from her urine.
28
-------
Table 4-1. Cases of human poisoning following ingestion of nitrobenzene
Subject(s)
Male, 5 years
Female, 2 years
Female, 24 years
Nine cases, adults
and children
Male, 21 years
Female, 24 years
Male, 48 years
Male, 19 years
Female, 19 years
Agent, dose
Screen-printing
material, unknown
quantity
Unknown substance,
unknown quantity
Nitrobenzene,
unknown quantity
Bitter almond oil,
unknown quantity
Screen-printing
varnish, 30-40 mL
Nitrobenzene,
unknown quantity
Gunpowder solvent
(2% nitrobenzene),
300 mL
Brown liquid,
unknown quantity
Nitrobenzene,
50 mL
Symptoms
Methemoglobinemia;
cardiac arrest and death
after initial improvement
Shock, cyanosis,
tachycardia, 41%
methemoglobinemia
Cyanosis,
labored breathing,
tachycardia
Vomiting, dizziness,
cyanosis, respiratory
depression, convulsions,
methemoglobinemia
Coma,
dark brown blood
Coma, cyanosis,
bitter almond breath,
mild jaundice
Cyanosis,
breathing problems,
75% methemoglobinemia
Unconsciousness,
cyanosis, bitter almond
breath, 65%
methemoglobinemia,
hemolytic anemia
Unconsciousness,
cyanosis, bitter almond
breath, 82%
methemoglobinemia
Treatment
Gastric lavage,
ascorbic acid
Methylene blue
Gastric lavage,
methylene blue,
ascorbic acid
Not stated
Gastric lavage,
methylene blue,
ascorbic acid
Gastric lavage,
methylene blue,
ascorbic acid,
intravenous fluids,
diuretics
Methylene blue,
blood transfusions
Gastric lavage,
methylene blue,
ascorbic acid,
methylprednisolone,
diazepam
Gastric lavage, 2%
thionine in glucose
intravenous, oxygen,
blood transfusions
Reference
Gupta et al.
(2000)
Wentworth et
al. (1999)
Chongtham et
al. (1999,
1997)
Abbinante et
al. (1997)
Kumar et al.
(1990)
Ajmani et al.
(1986)
Schimelman et
al. (1978)
Harrison
(1977)
Myslak et al.
(1971)
4.1.3. Dermal Exposure
A number of case reports exist in which at least a portion of the nitrobenzene dose was
absorbed via the dermal route. For example, Stevens (1928) discussed a case in which infant
twins were exposed to nitrobenzene contained in a disinfectant that had been applied to their
mattress to exterminate bed bugs. The subjects displayed marked cyanosis, rapid pulse rates, and
depressed respiration rates, and blood samples revealed the presence of methemoglobinemia.
Both subjects made a steady recovery when removed from the source of the contamination.
Levin (1927) discussed the case of a 2-year-old child who was dermally exposed when his
mother painted his shoes with a dye containing nitrobenzene. Cyanosis ensued, with rapid pulse
and depressed respiration, similar symptoms to those of the infant twins described by Stevens
(1928). A sample of blood was extremely dark in color, though metHb was not measured
29
-------
specifically. With the aid of bed rest and occasional oxygen administration, the child recovered
once the source of the poisoning had been removed.
Zeligs (1929) reported similar cases involving infants who had been dermally exposed to
nitrobenzene or aniline from a laundry mark that had been stamped on their cotton mattress pads.
The infants displayed the typical symptoms of cyanosis and discolored blood. They recovered
rapidly when oxygen was administered to aid the restoration of oxyHb levels.
Stevenson and Forbes (1942) reported a case in which an infant developed the
characteristic symptoms of nitrobenzene poisoning after the family's living quarters had been
treated with an insecticide containing 12.5% nitrobenzene and unstated amounts of kerosene,
turpentine, and oil of lilacine, which apparently contaminated the child's crib and mattress. As
with the other early cases, it is not clear whether exposure was via inhalation, dermal, or both
routes. The patient presented with marked cyanosis and methemoglobinemia, considerable
temperature fluctuations, and the appearance of a skin rash. The infant recovered steadily with
the aid of oxygen, an intravenous injection of 5% dextrose, and two blood transfusions.
A paper by Zeitoun (1959) discussed 21 cases of cyanotic infants and children who had
become sick after being rubbed with fake bitter almond oil that contained nitrobenzene. As in
other cases, a range of symptoms including hypoxia, weakness, shock, and, in some cases,
excitation or depression accompanied profound methemoglobinemia. Of the 21 cases, 2 subjects
died from complications associated with developing bronchopneumonia, while the remaining
19 subjects recovered completely.
A more recent example of methemoglobinemia induced through dermal penetration of
nitrobenzene occurred in a 2-month-old baby boy whose mother rubbed his skin with Oleum
dulcis, a topical hair oil containing about 1% nitrobenzene (Mallouh and Sarette, 1993). The
typical presentation of bluish coloration of the skin and lips was accompanied by a chocolate-
colored venous blood sample, in which the metHb level reached 31.5%. The patient was
observed without treatment and recovered. A chronological compilation of the case reports
involving inhalation and/or dermal exposure to nitrobenzene is presented in Table 4-2.
30
-------
Table 4-2. Cases of human poisoning with nitrobenzene following inhalation
or dermal exposure
Subject(s)
Male, 2 months
Female, 47 years
21 Infants
(15 males,
6 females)
Infant
Infants
Infant twins
Male, 2 years
16 Cases
Female, adult
Agent
Dermal application of
O. dulcis
(1% nitrobenzene)
Paint fumes containing
97.7% nitrobenzene
Dermal application of
false bitter almond oil
containing 2-10%
nitrobenzene
Insecticide containing
12.5% nitrobenzene
Laundry marking color
containing
nitrobenzene
Insecticide containing
nitrobenzene
Shoe polish fumes
Shoe dye fumes
Cleaning fluid
Symptoms
Cyanosis, 31.5%
methemoglobinemia
Cyanosis, headache,
nausea, jaundice,
hyperalgesia;
p-aminophenol and
p-nitrophenol in urine
Shock, tachycardia,
cyanosis, hypoxia, coma,
weakness,
methemoglobinemia;
two fatalities
Cyanosis,
methemoglobinemia,
skin rash
Cyanosis,
methemoglobinemia
Cyanosis, shallow
breathing, tachycardia,
methemoglobinemia
Cyanosis, shallow
breathing, tachycardia,
76% methemoglobinemia
Headache, nausea,
dizziness, malaise
Multiple neuritis,
contractures, weakness
Treatment
None
Glucose intravenous,
vitamins B 1 and B6,
iron preparations
Washing to remove
oil, methylene blue,
oxygen, ascorbic
acid, blood
transfusions
Oxygen,
5% glucose
intravenous, blood
transfusions
Oxygen
Removal from
exposure source
Oxygen, rest
NAa
NA
Reference
Mallouh and
Sarette (1993)
Ikeda and Kita
(1964)
Zeitoun (1959)
Stevenson and
Forbes (1942)
Zeligs (1929)
Stevens (1928)
Levin (1927)
Stifel(1919)
Adams (1912)
(as cited by
Hamilton
[1919])
aNA = data not available.
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
4.2.1. Oral Exposure
4.2.1.1. Subchronic Studies
The National Toxicology Program (NTP) sponsored a 90-day oral study (NTP, 1983a) of
nitrobenzene in which 10 F344 rats/sex/group received 0, 9.38, 18.75, 37.5, 75, and 150 mg/kg-
day and 10 B6C3F1 mice/sex/group received 0, 18.75, 37.5, 75, 150, and 300 mg/kg-day by
gavage in corn oil. The doses selected were based on the outcome of a 14-day range-finding
study in which 10 animals/sex/group received doses from 37.5-600 mg/kg. In the range-finding
study, all rats and mice receiving 600 mg/kg-day and all rats and a single mouse receiving
300 mg/kg died prior to planned termination. Toxicological responses to nitrobenzene among
31
-------
the survivors in the range-finding study included depressed body weight gain that was evident in
male mice receiving >37.5 mg/kg nitrobenzene and in female mice receiving >75 mg/kg. Other
toxicological endpoints included statistically significant increases in reticulocyte counts7 and
metHb levels. These responses exceeded control levels in treated rats (doses not specified), in
male mice at 75 mg/kg and above (reticulocytes) and 150 mg/kg and above (metHb), and in
female mice at 75 mg/kg and above (metHb). Histopathologic lesions were observed in brain,
liver, lung, kidney, and spleen in rats and mice, though at unstated dose levels.
In the main study, all animals were observed twice daily for clinical signs of toxicity, and
body weights and food consumption were monitored weekly. Blood samples were obtained at
term to measure hematologic parameters, reticulocyte count, and metHb levels, and the weights
of the brain, liver, right kidney, thymus, heart, lungs, and right testis were recorded. Necropsies
were performed on all animals that died prematurely or were sacrificed at term, and gross
examinations of a large suite of organs and tissues were carried out. Tissues were preserved in
formalin, and most of those listed were processed for histopathologic examination, primarily all
controls, rats at 75 and 150 mg/kg-day, and mice at the 300 mg/kg-day dose levels.
Additionally, putative target organs of nitrobenzene toxicity, such as liver, spleen, kidney, lung,
brain, bone marrow, testis, epididymis, and uterus, were examined from rats and mice exposed at
intermediate dose levels. There was no apparent autolysis among animals that were found dead
(all in the 150 mg/kg-day dose group); tissues from these animals were also examined
microscopically.
Nine male and three female rats at the 150 mg/kg-day dose level died prior to study
completion. The earliest deaths in the 150 mg/kg-day dose male and female rats were at day 67
(week 10) and day 38 (week 6), respectively. In the same group, six males also died on day 73
(week 11) and two more died on day 88 (week 13), while another female rat died on each of days
45 (week 7) and 60 (week 9). Clinical signs of toxicity, such as ataxia, head tilt, lethargy, and
trembling, were evident, mostly in animals receiving 150 mg/kg-day and, to a lesser extent,
75 mg/kg-day. Overall, there was little change in body weight gain between control and treated
groups, and the final body weights were not significantly different from controls at any dose
level. In fact, the only sign of treatment-related body weight reduction was in the single
surviving male rat receiving 150 mg/kg nitrobenzene. Organ weights appeared to have been
dose dependently affected by nitrobenzene exposure, most notably in the case of liver, kidney,
and testis (males). As shown in Tables 4-3 and 4-4, liver weights and their ratios to body weight
were dose dependently increased over control levels and achieved statistical significance
compared with controls at all dose levels. Right kidney weight was increased over controls at all
dose levels, and the ratio of kidney weight to final body weight was significantly increased over
7 Reticulocytes are immature RBCs that are unable to carry and deliver oxygen. Reticulocyte counts are expressed
as a percentage of circulating RBCs, which, if increased, may be considered adverse because it may signify loss of
RBCs due, for instance, to hemolytic anemia.
32
-------
controls at the 9.38, 18.75, and 75 mg/kg-day dose levels. Right testis weight and its ratio to
body weight were decreased in the 18.75-75 mg/kg dose range.
Table 4-3. Changes in absolute and relative liver, kidney, and testis weights
in male F344 rats exposed to nitrobenzene by gavage for 90 days
Dose
(mg/kg-day)
0
9.38
18.75
37.5
75
150
Organ weights (mean ± standard deviation)3
Liver
Absolute
(mg)
11,668 ±1,309
13,269 ±l,555b
14,567 ±l,168b
15,451 ±l,327b
15,679 ±2,117b
11,264
Relative
(x ID'2)
3.52 ±0.22
4.04±0.2b
4.37±0.14b
4.77 ± 0.22b
5.15±0.15b
4.79
Kidney
Absolute
(mg)
1,025 ± 108
1,085 ± 142
1,115 ± 83
1,070 ±153
1,083 ± 104
1,023
Relative
(x ID'3)
3. 10 ±0.2
3.30±0.2b
3.36±0.1b
3.30 ±0.38
3.44±0.23b
4.35
Testis
Absolute
(mg)
1,435 ± 96
1,435 ± 79
1,425 ± 104
1,406 ± 71
873 ± 476b
835
Relative
(x ID'3)
4.34 ±0.26
4.39 ±0.35
4.30 ±0.23
4.33 ±0.15
2.78±1.42b
3.55
an = 10 in all groups except the 150 mg/kg-day group with one surviving male.
bSignificantly different from control values, as calculated by the authors.
Source: NTP(1983a).
Table 4-4. Changes in absolute and relative liver and kidney weights in
female F344 rats exposed to nitrobenzene by gavage for 90 days
Dose
(mg/kg-day)
0
9.38
18.75
37.5
75
150
Organ weights (mean ± standard deviation)3
Liver
Absolute
(mg)
6413 ±613
7402 ± 279b
7481±702b
8436 ± 587b
9198±713b
9925 ± 436b
Relative
(x ID'2)
3.43 ±0.16
3.76±0.08b
3.95±0.18b
4.23±0.18b
4.88±0.22b
5.21±0.28b
Kidney
Absolute
(mg)
582 ± 56
615 ±47
627 ± 41
644 ± 52b
641±68b
666 ± 40b
Relative
(x 10-3)
3.11±0.14
3. 13 ±0.24
3.32±0.14b
3.24 ±0.30
3.39±0.25b
3.49±0.12b
an = 10 in all groups except the 150 mg/kg-day group with 7 surviving females.
bSignificantly different from control values, as calculated by the authors.
Source: NTP(1983a).
There were statistically significant changes in some hematologic parameters in rats
exposed to nitrobenzene via gavage. As shown in Tables 4-5 and 4-6, the principal effects were
dose-dependent decreases in hematocrit (Hct), Hb, and RBC count and dose-dependent increases
in reticulocyte counts and metHb. In males, these changes achieved statistical significance
compared with controls at a dose of 9.38 mg/kg-day for metHb and Hb and 18.75 mg/kg-day for
the other parameters. In females, the changes achieved statistical significance compared with
controls at 37.5 mg/kg-day and above for the RBC count and at 9.38 mg/kg-day for the other
33
-------
parameters. The authors reported little change in white blood cell (WBC) count and differential
except in those rats receiving 150 mg/kg-day, in which a marked leukocytosis appeared to be
accompanied by lymphocytosis and a greater number of polymorphonuclear cells.
Table 4-5. Hematologic parameters, reticulocytes, and metHb levels in male
F344 rats exposed to nitrobenzene via gavage for 90 days
Dose
(mg/kg-day)
0
9.38
18.75
37.5
75
150
Hb
(g/dL)a
16.24 ±0.42
15.73 ±0.29b
15.54 ±0.37b
14.72 ±0.30b
14.87 ±0.41b
16.2
Hct
(%)a
47.82 ±3.2
44. 19 ±4.98
41.84 ±1.88b
37.66 ±0.93b
38.08 ±1.96b
38
RBCs
(x 106)a
9.06 ±0.41
9.01 ±0.23
8.70±0.37b
7.97±0.34b
7.61±0.41b
6.31
Reticulocytes
(%)a
2.23 ±0.44
2.62 ±0.45
3.72±0.65b
4.75 ± 0.62b
6.84 ± 0.72b
15
MetHb
(%)a
1.13 ±0.58
2.75±0.58b
4.22±1.15b
5.62±0.85b
7.31±1.44b
12.22
"Values are means ± standard deviations, where n = 10 in each group except for the 150 mg/kg-day group with one
male.
bSignificantly different from controls, as calculated by the authors.
Source: NTP(1983a).
Table 4-6. Hematologic parameters, reticulocytes, and metHb levels in
female F344 rats exposed to nitrobenzene via gavage for 90 days
Dose
(mg/kg-day)
0
9.38
18.75
37.5
75
150
Hb
(g/dL)a
15.82 ±0.22
15.53 ±0.29b
15.49 ±0.39b
15.43 ±0.38b
14.86 ±0.52b
15.62 ±0.60
Hct
(%)a
42.27 ±3.41
39.37 ±1.26b
39.59 ±1.79b
38.95 ± 0.62b'c
37.52±l.llb
35.88±1.30b
RBCs
(x 106)a
8.39 ±0.49
8.05 ±0.28
8.01 ±0.35
7.83±0.35b
7.33±0.30b
5.86±0.35b
Reticulocytes
(%)a
2.60 ±0.37
3.69±0.32b
4.75±0.68b
6.28±0.90b
8.72±1.49b
32.07 ±3.56b
MetHb
(%r
0.94 ±0.03
2.06±0.45b
3.62±1.09b
5.27±0.76b
6.85±2.25b
12.77 ±1.83b
"Values are means ± standard deviations, where n = 10 in each group except for the 150 mg/kg-day group with 7
females.
bSignificantly different from controls, as calculated by the authors.
Statistics based on nine rats.
Source: NTP(1983a)
At necropsy, rats receiving 150 mg/kg-day nitrobenzene had enlarged spleens. Males at
this dose level had enlarged livers, and those receiving 75 mg/kg-day and 150 mg/kg-day
showed signs of testicular atrophy.
34
-------
Histopathologic examination of the major organs and tissues revealed compound-related
effects in the spleen, which appeared to be congested.8 Splenic corpuscles were small, and the
red pulp contained hemosiderin.9 The severity of splenic congestion was graded by the study
authors. Control rats had no splenic congestion or minimal splenic congestion (grade 1).
Congestion increased in severity up to moderate in the highest dose group. The incidence of
these and other histopathologic lesions in relation to dose is shown in Tables 4-7 and 4-8. Tables
4-7 and 4-8 also report the incidence of splenic congestion of grade 2 or higher.
Table 4-7. Selected histopathology findings in male F344 rats exposed to
nitrobenzene for 90 days via gavage
Tissue examined
Spleen
Congestion
Congestion > grade 2
Lymphoid depletion
Liver
Congestion
Testis
Atrophy
Hypospermatogenesis
Multinucleate giant cells
Brain stem
Hemorrhage
Vacuolization
Degeneration
Malacia
Nitrobenzene dose (mg/kg-day)
0
1/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
7/10
0/10
0/10
9.38
4/10
0/10
0/10
0/10
0/10
0/10
0/10
1/10
0/10
0/10
0/10
18.75
7/10
0/10
0/10
0/10
0/10
0/10
0/10
4/10
4/10
0/10
0/10
37.5
6/10
0/10
1/10
0/10
1/10
0/10
0/10
4/10
0/10
0/10
0/10
75
10/10
5/10
9/10
0/10
9/10
10/10
10/10
5/10
3/10
0/10
0/10
150a
10/10
10/10
10/10
6/10
9/9
9/9
8/9
2/10
0/10
4/10
4/10
"Includes tissue findings in nine rats that died between days 67 and 88.
Source: NTP(1983a).
Splenic congestion is an abnormality that leads to elevated splenic vein pressure, which in turn results in higher
sinusoidal pressure, and is commonly observed in laboratory animals in response to a variety of circumstances,
including agonal death, method of euthanasia, or exposure to chemicals. Administration to rodents of aromatic
amine-type chemicals (e.g., aniline) may cause splenic congestion and hemorrhage, which are accompanied by
hemosiderin deposition (brown intracellular pigmentation due to insoluble iron), fatty change, and extramedullary
hematopoiesis and fibrosis. These changes have been suggested to result from methemoglobinemia or accumulation
in erythrocytes of toxic metabolites that are released in the spleen when RBCs are broken down in the red pulp.
Sustained congestion causes the spleen to become more firm, enlarged, and fibrotic and renders the organ
susceptible to trauma. Spleen enlargement in humans may be caused by a variety of diseases and, in some instances,
is associated with increased workload (such as in hemolytic anemia) or hyperfunction in response to destruction of
abnormal RBCs, with symptoms of abdominal pain and early satiety (Greaves, 2007; Cotran et al., 1994).
9 The red pulp (also called splenic pulp), which may act as a reservoir for storing blood, is a soft mass of dark
reddish-brown color resembling coagulated blood, and it is made of a fine reticulum of fibers divided into splenic
sinuses and splenic cords. The splenic red pulp may undergo changes due to a variety of factors, including immune
stimulation, changes in circulation, accumulation of macrophages, and connective tissue or pigment, and in response
to increased demand for filtration of abnormal RBCs (Greaves, 2007; Guyton and Hall, 2000).
35
-------
Table 4-8. Selected histopathology findings in female F344 rats exposed to
nitrobenzene for 90 days via gavage
Tissue examined
Spleen
Congestion
Congestion > grade 2
Lymphoid depletion
Kidney
Pigmentation
Brain stem
Hemorrhage
Vacuolization
Degeneration
Malacia
Nitrobenzene dose (mg/kg-day)
0
2/10
0/10
0/10
0/10
4/10
6/10
0/10
0/10
9.38
5/10
1/10
0/10
0/10
2/10
3/10
0/10
0/10
18.75
10/10
3/10
2/10
0/10
3/10
1/10
0/10
0/10
37.5
10/10
5/10
4/10
0/10
1/10
1/10
0/10
0/10
75
10/10
8/10
8/10
5/10
1/10
1/10
0/10
0/10
150a
10/10
9/10
10/10
9/10
7/10
5/10
4/10
3/10
"Includes tissue findings in three rats that died between days 38 and 60.
Source: NTP(1983a).
It should be noted that the recorded histopathology lesions in the high-dose male and
female rats included the findings from animals that died prior to the full 90-day study duration
(days 67-88 in males and days 38-60 in females). The extent to which some observed
histopathologic effects in the liver were compound related is unclear, because hematopoietic foci
and hepatocellular necrosis were evident in both treated and control rats. Hyaline droplets were
noted in the cortical tubule cells of the kidney, and some pigmented granules were evident in the
cells of a few treated rats. There were obvious compound-related histopathologic effects on the
seminiferous tubules of the testis of male rats. In some cases, the tubules contained
spermatogonia and spermatocytes, while in others there were very few or no spermatids,
spermatozoa, and Sertoli cells. Some tubules appeared to contain only a lacy fibrinous material,
and others contained multinucleate giant cells. Histopathologic changes in the brains of treated
rats included hemorrhage, vacuolization, and a wide range of inconsistent degenerative changes.
Based on the changes in absolute and relative organ weights and the dose-dependent
increases in reticulocyte count and metHb concentration, all of which were evident at the lowest
administered dose, a lowest-observed-adverse-effect level (LOAEL) of 9.38 mg/kg-day is
identified for the subchronic oral effects of nitrobenzene in F344 rats in this study.
As with the nitrobenzene-exposed rats, the mice exhibited signs of toxicity reflective of
neurological impairment, increased liver and kidney weights, and decreased testis weight in male
mice or decreased thymus in female mice. Three male B6C3F1 mice receiving 300 mg/kg-day
died prior to study completion, most likely as a result of nitrobenzene exposure. Some surviving
animals at this dose level showed clinical signs of toxicity, including ataxia, hyperactivity, and
irritability. However, there were no compound-related changes in body weight gain at any dose
level. Absolute and relative organ weight changes were confined to liver, kidney, and testis in
36
-------
male mice and to the liver, kidney, and thymus in females. For example, liver weight and its
ratio to body weight were dose dependently increased in male mice, the increases achieving
statistical significance at the 150 and 300 mg/kg-day dose levels. Relative kidney weight was
significantly increased at 75 and 300 mg/kg-day in males. Absolute and relative testis weights
were decreased at dose levels of 300 mg/kg-day. Treatment-related increases in absolute liver
weights in female mice were evident at 18.75 mg/kg-day and above, with relative liver weights
achieving statistical significance at a dose level of 37.5 mg/kg-day and above. Absolute and
relative thymus weights were also elevated in nitrobenzene-receiving female mice. These
changes are documented in Tables 4-9 and 4-10.
Table 4-9. Changes in absolute and relative liver, kidney, and testis weights
in male B6C3F1 mice exposed to nitrobenzene by gavage for 90 days
Dose
(mg/kg-day)
0
18.75
37.5
75
150
300
n
10
10
10
9
10
7
Organ weights in mg (mean ± standard deviation)
Liver
Absolute
(mg)
1527 ± 286
1597 ±137
1591 ±129
1709 ± 245
1871 ± 172a
2223 ± 126a
Relative
(x ID'2)
4.71 ±0.44
4.78 ±0.27
4.74 ±0.33
5.02 ±0.51
5.49±0.33a
6.53±0.55a
Kidney
Absolute
(mg)
272 ± 35
276 ± 23
288 ± 22
300 ± 19a
294 ± 20
312±28a
Relative
(x ID'3)
8.44 ±0.39
8.27 ±0.45
8.59 ±0.52
8.84±0.30a
8.61 ±0.31
9.14±0.58a
Testis
Absolute
(mg)
116 ±7.9
111±12
120±8.3b
113 ±9.7
113 ±16
84 ± 14a
Relative
(x ID'3)
3.66 ±0.60
3.32 ±0.34
3.60±0.30b
3.35 ±0.31
3. 33 ±0.52
2.45 ± 0.42a
aSignificantly different from control values, as calculated by the authors.
bSummary statistics reflect 9 samples, not 10.
Source: NTP(1983a).
Table 4-10. Changes in absolute and relative liver, kidney, and thymus
weights in female B6C3F1 mice exposed to nitrobenzene by gavage for 90
days
Dose
(mg/kg-day)
0
18.75
37.5
75
150
300
n
9
9
10
10
10
10
Organ weights (mean ± standard deviation)
Liver
Absolute
(mg)
1179 ± 58
1278±113a
1276 ± 74a
1256 ± 75a
1374 ± 51a
1566 ± 124a
Relative
(x ID'2)
4.41 ±0.22
4.64 ±0.32
4.79±0.32a
4.69±0.19a
5.05±0.14a
5.79±0.28a
Kidney
Absolute
(mg)
175 ± 14
179 ± 22
180 ±11
166 ±15
181 ±17
189 ±19
Relative
(x 10-3)
6.53 ±0.27
6.52 ±0.71
6.74 ±0.46
6.19 ±0.44
6.65 ±0.56
7.00±0.41a
Thymus
Absolute
(mg)
44.14 ±7.82
51.22 ±9.94
47.06 ± 9.47
50.41 ±8.97
47.21 ±13.2
51.45 ±9.19
Relative
(x 10-3)
1.65 ±0.26
1.87 ±0.39
1.76 ±0.35
1.89 ±0.38
1.73 ±0.46
1.91 ±0.37
"Significantly different from control values, as calculated by the authors.
Source: NTP(1983a).
37
-------
Hematologic responses observed in mice were similar to those in rats, with dose-
dependent increases in reticulocytes and metHb and progressively lower levels of Hb, Hct, and
RBCs. These changes are documented in Tables 4-11 and 4-12.
Table 4-11. Hematologic parameters, reticulocytes, and metHb levels in male
B6C3F1 mice exposed to nitrobenzene via gavage for 90 days
Dose
(mg/kg-day)
0
18.75
37.5
75
150
300
n
10
10
10
9
10
7
Hb
(g/dL)a
15.20 ±0.66
14.59 ±0.66
15.02 ±0.92
14.63 ±0.35b
14.44 ± 0.47b
15.45 ±0.52d
Hct
(%)'
41.77 ±2.29
39.76 ±2.89
41. 13 ±3.48
39.56 ±2.66
37.62 ±1.94b
36.26 ±3.30b'd
RBCs
(x 106)a
9.27 ±0.75
8.87 ±0.50
9.17 ±0.76
8.68 ±0.52
8.25±0.37b
7.79 ± 0.29b'e
Reticulocytes
(%)'
5.02 ±1.0
5.81±0.88C
6.95 ± 0.82bc
7.85 ± 0.74b
9.30 ± 1.12b
10.45 ± 1.58b
MetHb
(%)'
1.07 ±0.32
2.16±0.32b'c
3.42±0.61b'c
4.75±1.03b
5.98±0.97b
6.72 ± 1.28b
"Values are means ± standard deviations.
bSignificantly different from controls, as calculated by the authors.
'Summary statistics represent nine samples.
dSummary statistics represent six samples.
eSummary statistics represent five samples.
Source: NTP(1983a).
Table 4-12. Hematologic parameters, reticulocytes, and metHb levels in
female B6C3F1 mice exposed to nitrobenzene via gavage for 90 days
Dose
(mg/kg-day)
0
18.75
37.5
75
150
300
n
9
9
10
10
10
10
Hb
(g/dL)a
15.66 ±0.61
15.70 ±0.60
15.24 ±0.83
14.98 ±0.50b
14.96 ±0.33b
15.99 ±0.59
Hct
(%)a
44.33 ±3.41
44.24 ±2.32
43. 86 ±2.30
41.66 ±1.71b
40.98 ± 2.24b
38.66 ±2.69b
RBCs
(x 106)a
9.54 ±0.67
9.52 ±0.35
9.21 ±0.60
9.06 ± 0.44
8.81±0.35b
8.11±0.61b
Reticulocytes
(%)a
4.17 ±0.35
5.54±0.51b
6.29±0.61b
6.72 ± 0.60b
7.31±0.48b
11.08 ±1.96b
MetHb
(%)a
0.87 ±0.23
1.20±0.22b
1.45±0.34b
1.82±0.30b
2.25 ± 0.40b
3.54 ± 1.39b
aValues are means ± standard deviations.
bSignificantly different from controls, as calculated by the authors.
Source: NTP(1983a).
There were few signs of treatment-related lesions in the mice at necropsy, although some
evidence of a darkening in coloration of such organs as kidney, lung, spleen, adrenal, and lymph
nodes was noted in animals exposed to 300 mg/kg-day nitrobenzene. As summarized in Tables
4-13 and 4-14, there were not many histopathologic changes, and those that were observed may
have been unrelated to the effects of the compound. However, enlargement of hepatocytes in the
centrilobular zone in male and female mice exposed to 300 mg/kg-day was noteworthy.
38
-------
Table 4-13. Selected histopathology findings in male B6C3F1 mice exposed
to nitrobenzene for 90 days via gavage
Tissue examined
Spleen
Lymphoid depletion
Liver
Cytomegaly
Testis
Atrophy
Hypospermatogenesis
Multinucleate giant cells
Brain stem
Hemorrhage
Degeneration
Nitrobenzene dose (mg/kg-day)
0
0/10
0/10
0/10
0/10
0/10
3/10
0/10
18.75
0/10
0/10
3/10
0/10
0/10
1/10
0/10
37.5
0/10
0/10
2/10
0/10
0/10
3/10
0/10
75
0/10
1/10
0/10
0/10
0/10
0/10
0/10
150
0/10
2/10
5/10
0/10
0/10
0/10
0/10
300
1/10
10/10
5/10
4/10
2/10
2/10
1/10
Source: NTP(1983a).
Table 4-14. Selected histopathology findings in female B6C3F1 mice exposed
to nitrobenzene for 90 days via gavage
Tissue examined
Spleen
Lymphoid depletion
Liver
Cytomegaly
Adrenal
Fatty change
Brain stem
Hemorrhage
Nitrobenzene dose (mg/kg-day)
0
0/10
0/10
0/10
2/10
18.75
0/10
0/10
0/10
2/10
37.5
0/10
0/10
0/10
1/10
75
0/10
0/10
0/10
2/10
150
2/10
0/10
0/10
0/10
300
5/10
8/10
8/10
3/10
Source: NTP(1983a).
The statistically significant increase in metHb concentration observed in both sexes of
B6C3F1 mice at the lowest dose level tested points to a dose of 18.75 mg/kg-day as LOAEL for
the subchronic effects of nitrobenzene in this species when administered via the oral route.
Support for this designation is provided by the clear-cut trend in increased reticulocytes, which
was statistically significantly different from controls in females receiving 18.75 mg/kg-day.
While the increase in reticulocytes did not achieve statistical significance at the lowest dose level
in males, the value appeared to be part of a dose-dependent trend toward a statistical significance
that was evident at higher dose levels. This supports the choice of 18.75 mg/kg-day as LOAEL
for this response in B6C3F1 mice.
Shimo et al. (1994) gavaged six F344 rats/sex/group with 0, 5, 25, and 125 mg/kg-day
nitrobenzene for 28 days. An additional set of control and 125 mg/kg rats were allowed to
recover for 14 days after the completion of treatment. As determined from the English data
39
-------
tables, animals were evaluated for generalized signs of toxicity, and body weight changes and
food consumption were monitored in all groups. Blood samples were taken at term for
hematologic and clinical chemistry parameters. Major organs were weighed at term, and tissue
samples were fixed and processed for histopathologic examination.
Clinical signs in high-dose rats included decreased movement, pale skin, and abnormal
gait. Additionally, the authors plotted the body weight changes against time and showed a
marked treatment-related reduction in body weight increase, even though food consumption was
little changed among the groups. Striking changes in hematologic parameters were evident in
nitrobenzene-treated rats, with dose-dependent reductions in RBC count, Hct, and Hb
concentration and a dose-dependent increase in mean corpuscular volume (MCV). By contrast,
the WBC count increased dramatically with dose. However, these changes were not noted in
those animals allowed to recover for 14 days after dosing (Table 4-15).
Table 4-15. Hematologic and clinical chemistry parameters in rats treated
with nitrobenzene for 28 days, with or without a recovery period of 14 days
Parameter
28-Day dosing study"
Control
5 mg/kg
25 mg/kg
125 mg/kg
14-Day recovery group"
Control
125 mg/kg
Males
RBC (x 104/mm3)
Hb (g/dL)
Hct (%)
MCV (fL)
WBC (x 102/mm3)
BUNd (mg/dL)
ASTd (IU/L)
ALTd (IU/L)
761 ±117
16.9 ±0.6
41.6 ±6.3
54.7 ±0.8
44 ±14
17.8 ± 1.1
111±14
40 ±6
670 ± 54
16.6 ±0.6
35.6 ±3.3
53.0 ±0.9
45 ±8
16.1±1.5C
81±6b
43 ±5
524 ± 36b
14.5±0.5b
32.3±2.4C
61.3±2.7b
122 ± 44b
14.1±2.4b
86±6b
38 ±4
412±54b
14.2±0.5b
34.9±3.4C
84.8±5.5b
1426±521b
12.7 ± 1.2b
105 ±17
47 ±9
727 ± 93
16.7 ±0.7
38.2 ±4.9
52.7 ±1.4
46 ±5
16.8 ±2.5
89 ±9
35 ±7
724 ± 100
17.7±0.6C
45.7±6.6C
63.0 ±1.4
40 ±16
17.5 ±1.1
94 ±10
37 ±4
Females
RBC (x 104/mm3)
Hb (g/dL)
Hct (%)
MCV (fL)
WBC (x 102/mm3)
BUN (mg/dL)
AST (IU/L)
ALT (IU/L)
708 ± 63
17.5 ±0.9
38.1 ±3.2
53.8 ±1.2
40 ±12
17.5 ±2.2
96±5
39 ±5
718 ± 129
16.3 ±1.0
37.8 ±6.5
52.7 ±0.5
43 ±8
14.2±1.0b
79±5b
36 ±4
635 ± 126
15.5±0.6b
37.7 ±7.4
59.5 ± 1.6b
73 ±44
12.8±2.2b
85±9C
42 ±8
458±43b
14.5±0.8b
35.4 ±3.4
77.2±1.6b
1990 ± 298b
12.3±3.4C
94 ±10
53 ± 14C
694 ± 79
16.8 ±0.4
36.7 ±4.6
52.8 ±0.8
42 ±4
18.9 ±3. 9
77 ±5
40 ±2
674 ± 86
18.0 ±1.2
39.5 ±5.1
58.3 ± 5.2C
47 ±6
16.8 ±1.9
79 ±5
31±5b
"Values are means ± standard deviations for six animals/group, except for the 125 mg/kg-day female group with five
animals. The limited information available did not clarify the disposition of the additional female that apparently
started on study in the 125 mg/kg-day group.
bp < 0.01 versus controls, as calculated by the authors.
cp < 0.05 versus controls, as calculated by the authors.
dBUN = blood urea nitrogen; AST = aspartate aminotransferase; ALT = alanine aminotransferase.
Source: Shimo etal. (1994).
40
-------
Treatment-related changes in clinical chemistry parameters were also evident; a
consistent dose-dependent decrease in blood urea nitrogen (BUN) was evident in both males and
females (Table 4-15). Serum transaminase activities (e.g., aspartate aminotransferase [AST] and
alanine aminotransferase [ALT]) were inconclusive. AST activity was statistically significantly
decreased in male and female rats receiving either 5 or 25 mg/kg nitrobenzene; however, the
biological relevance of the observed decrease is questionable, especially since no change was
observed between the high-dose animals (125 mg/kg) and controls. Similarly, a statistically
significant increase in ALT was observed only in female rats receiving the high dose. This
finding was also of questionable relevance since the average value was only 14% higher than in
controls. Hematology parameters and serum BUN concentrations returned to control levels after
a 14-day recovery period.
Absolute changes in organ weights exhibited similar trends between male and female rats
with increases noted for the spleen, liver, and kidney and decreases found with the thymus and
adrenals of both sexes and with the testis in males. A strong dose-dependent increase in absolute
spleen weight was observed with males and females with a nearly fourfold increase at the highest
dose for both sexes. Absolute liver weight increased dose dependently in female rats up to 80%
with the highest dose, whereas a 19% increase was observed in male rats at the highest dose. In
contrast to the spleen and liver, increases in absolute kidney weights did not exhibit clear dose-
dependent responses. In male rats, an 8% increase was observed with the 25 mg/kg group;
however, kidney weights from high-dose animals (125 mg/kg) were consistent with those in
controls. In contrast, absolute kidney weight in female rats was only increased (13%) at the
highest dose with all other groups being similar to controls. Following the 14-day recovery
period, the absolute spleen weights for male and female rats were still increased by 37% in males
and 26% in females, whereas absolute liver and kidney weights returned to control values.
Decreases in absolute thymus weight occurred with both male (27% |) and female (30% |) rats
at the highest dose but returned to control values at the end of the 14-day recovery period.
Absolute testis weights were statistically significantly reduced (70% |) in high-dose males and
remained reduced by 46% at the end of the recovery period.
Histopathologic evaluation of tissues was used to corroborate changes in tissue weight
and clinical chemistry with severity of response (grade: no change < moderate < severe). In
male rats, graded responses for splenic congestion, increased brown pigmentation in red pulp,
and increased extramedullary hematopoiesis exhibited a dose-dependent increase in grade, with
100% of animals being scored as severe at the highest dose tested. Female rats exhibited a
similar dose-dependent increase in severity of scores for the above indices, with 100% of
animals being graded as severe for splenic congestion and increased extramedullary
hematopoiesis. Increased brown pigmentation in red pulp was graded as severe in two and
moderate in three female rats. Liver scores were graded as no change in all groups, except for
the high-dose animals in both sexes. In high-dose males and females, increased extramedullary
41
-------
hematopoiesis was moderate in five males and two females and exhibited no change in one male
and three females. Brown pigmentation in Kupffer cells was moderate in five males and four
females and severe in one male and one female. Absolute kidney weights in males were
inconsistent with the histopathologic finding. In the highest dose group, brown pigmentation in
tubular epithelium was reported as moderate in five animals and severe in one animal, but no
change in absolute kidney weight was reported at the highest dose. In contrast, female
histopathology of the kidney correlated with the absolute weight in that 100% of animals were
graded with moderate brown pigmentation in tubular epithelium at the highest dose, and a 3%
increase in absolute kidney weight was also observed at the highest dose. All other animals were
consistent with controls (no change). Decreased absolute testis weight correlated with severe
degeneration of seminiferous tubular epithelium and severe atrophy of seminiferous tubule in
100% of male rats receiving 125 mg/kg nitrobenzene. A synopsis of the oral toxicity studies in
animals is presented in Table 4-16.
Table 4-16. Summary of effects observed in oral dosing studies with
nitrobenzene
Species,
strain
Rat,
F344
Rat,
F344
Mouse,
B6C3F1
Number
6/sex
10/sex
10/sex
Dosing
0, 5, 25, 125
mg/kg-day,
gavage, 4 wk
0, 9.4, 18.8, 37.5,
75, 150 mg/kg-
day, gavage, 90 d
0, 18.8,37.5,75,
150, 300 mg/kg-
day gavage, 90 d
Effect3
RBC |, Hb |, Hct |,
MCVf
WBCt
Liver weight t
Kidney weightf
MetHb t & Hb |
Reticulocytes t
Liver weight t
MetHb t
Hb|
Reticulocytes t
Liver cytomegaly
NOAELb'c
(mg/kg-day)
5 (M, F)
NA
NA(M)
9.4 (F)
NA
9.4 (M)
NA(F)
75 (M)
NA(F)
NA (M, F)
37.5 (M, F)
18.8 (M)
NA(F)
150 (M, F)
LOAELbc
(mg/kg-day)
25 (M, F)
9.4 (M, F)
9.4 (M)
18.8 (F)
9.4 (M, F)
18.8 (M)
9.4 (F)
150 (M)
18.8 (F)
18.8 (M, F)
75 (M, F)
37.5 (M)
18.8 (F)
300 (M, F)
Reference
Shimo et al.
(1994)
NTP (1983a)
"Only endpoints with evident dose responses were selected. J, or f = decrease or increase in the respective endpoint.
bNo-observed-adverse-effect levels (NOAELs) and LOAELs determined by nitrobenzene assessment authors.
°M = male; F = female; NA = not applicable.
42
-------
4.2.1.2. Chronic Studies
No studies were identified that addressed the chronic toxicity of nitrobenzene
administered via the oral route.
4.2.2. Inhalation Exposure
4.2.2.1. Subchronic Studies
CUT (1984) reported a subchronic study in which F344 rats, CD rats, and B6C3F1 mice,
10/sex/group, were exposed via inhalation to 0, 5, 16, or 50 ppm nitrobenzene, 6 hours/day,
5 days/week for 90 days. During the in-life phase of the 90-day study, behavioral signs were
observed twice daily, and body weights were monitored weekly. At the end of the 90-day
exposure period, animals were fasted overnight and then sacrificed following an i.p. injection
with pentobarbital. Samples of blood were taken to measure hematologic and clinical chemistry
parameters. Animals were examined for gross abnormalities at necropsy, and the weights of
certain key target organs, such as the spleen, liver, kidney, testes, and brain, were recorded.
Eight-hour urine samples were obtained from all animals after 60 days of exposure. Among the
parameters assessed were color, turbidity, specific gravity, pH, protein, glucose, ketones,
bilirubin, blood, and the presence of cells, casts, and crystals. Histopathologic examination was
carried out in a full range of excised organs and tissues, including the epithelium lining the air
passages of the nose and lungs.
There were no compound-related effects on body weight, mortality, or the occurrence of
behavioral signs in the subchronic 90-day study. However, increases in spleen weights were
evident in all strains and sexes of rats and mice exposed to nitrobenzene at the high concentration
and at 16 ppm in male F344 and CD rats. By contrast, there was a statistically significant
reduction in testis weight in male F344 rats exposed to 50 ppm nitrobenzene. Examination of the
internal organs of exposed animals at necropsy confirmed that the liver, spleen, and testis were
the primary target organs of nitrobenzene. For example, in high-concentration rats of either
strain, males presented with testicular atrophy, enlarged spleens, and the presence of irregular
blotches on the surface of the liver. Similarly, both sexes of B6C3F1 mice had enlarged spleens
in response to nitrobenzene at 50 ppm.
A number of statistically significant changes occurred in the hematologic parameters
under investigation, but all were not obviously related to exposure concentration. However, in
the rats, there was an increased incidence of hemolytic anemia in response to increased
concentrations of nitrobenzene. Most marked among the potential compound-related changes in
hematologic or clinical chemistry parameters were the increased concentrations of serum metHb
(Table 4-17) and a 50% increase in the concentration of bilirubin in male F344 rats receiving 16
and 50 ppm nitrobenzene. Histologic sections of organs and tissues of nitrobenzene-receiving
rats and mice demonstrated treatment-related lesions in the spleen, testis, liver, epididymides,
kidney, and bone marrow, plus other possible target organs of nitrobenzene, such as the adrenals,
43
-------
lymph nodes, and lungs. For example, in F344 rats, lesions in the spleen consisted of acute
sinusoidal congestion, proliferative capsular lesions, and increases in extramedullary
hematopoiesis. These effects were dose dependent with 10/10 animals of either sex affected in
F344 rats exposed to 50 ppm.
Table 4-17. Concentrations of metHb in plasma of F344 and CD rats and
B6C3F1 mice in response to nitrobenzene inhalation
Strain/
species
F344 rat
CD rat
B6C3F1
mouse
Sex
Males
Females
Males
Females
Males
Females
Concentration of nitrobenzene (ppm)
0
5
16
50
Concentration of metHb in plasma (%)a
1.2 ±0.4
1.6 ±0.8
0.6 ±0.2
2.1 ±1.2
0.7 ±0.6
1.3 ±0.9
3.0 ± 1.0b
3.2 ±0.9
0.9 ±0.6
2.3 ±0.6
1.6 ±0.4
0.8 ±0.5
4.4 ± 1.3b
3.9±1.3b
3.2±0.7b
3.7 ±0.2
2.1 ±1.3
2.0 ±0.6
10.1±1.2b
10.5±1.5b
10.1±2.0b
9.6±2.5b
5.8±1.7b
5.1±0.8b
aValues are means ± standard deviations, where n = 5 except for the 16 ppm F344 rat female group with 4 animals.
bp < 0.05, as calculated by the authors.
Source: CUT (1984).
Histopathologic effects of nitrobenzene on the liver in F344 rats included disorganization
of the hepatic cord architecture and centrilobular degeneration of the hepatocytes in 7/10 high-
concentration males but only in 1/10 high-concentration females. Other histopathologic effects
evident in F344 rats included basophilia of the medullary cells of the adrenal in 5/10 high-
concentration males and in 3/10 high-concentration females, plus an increased incidence of
bronchial hyperplasia in both sexes receiving the highest dose. All male F344 rats displayed
degeneration of tubular epithelial cells in the testis. The condition was described by the authors
as representing a cessation of maturation at the level of primary and secondary spermatocytes
and was usually accompanied by interstitial edema and hyperplasia of Ley dig cells. There were
no mature sperm in the epididymis of these F344 rats. Instead, the presence of some apparently
proteinaceous material was noted in the ducts. Kidney effects of nitrobenzene in F344 rats were
characterized by a toxic nephrosis associated with an accumulation of droplets in the cytoplasm
of proximal tubular epithelial cells. The droplets were described in the report as hyaline and
eosinophilic, and the lesions increased in incidence and severity with dose in both sexes of F344
rats. The report makes no mention of whether or not kidney sections were stained for the male
rat-specific protein, a2U-globulin. In the absence of this information and in view of the
appearance of kidney lesions in both sexes of F344 rat, the kidney responses cannot be assigned
to a2u-globulin-associated nephropathy (U.S. EPA, 1991b).
44
-------
Many of the target organs indicative of nitrobenzene toxicity in F344 rats also were target
organs in CD rats, including spleen, liver, kidney, epididymis, bone marrow, and nasal
turbinates. For example, the splenic lesions consisted of sinusoidal congestion, increased
extramedullary hematopoiesis, and numbers of hemosiderin-laden macrophages infiltrating the
red pulp. An increase in the thickness of the splenic capsule was noted in 4/10 male and 3/10
female CD rats exposed to 50 ppm nitrobenzene. CD rats also displayed a marked bilateral
testicular atrophy in response to nitrobenzene, as indicated by a loss of seminiferous epithelium,
interstitial cell hyperplasia, edema, and the absence of mature sperm in the epididymal lumen.
These features were evident in 1/10 subjects receiving 5 ppm nitrobenzene, 2/10 receiving
16 ppm, and 9/10 receiving nitrobenzene at the highest concentration. Toxic effects of
nitrobenzene were particularly apparent in the nasal passages of CD rats. These lesions were
characterized by the occurrence of lymphoid hyperplasia, inflammation, and the presence of
interstitial and granulomatous pneumonitis, together with the presence of macrophages and
lymphocytes in perivascular areas. In a manner similar to F344 rats, CD rats displayed dose-
dependent toxic nephrosis, with 10/10 male and 5/10 female rats exposed to 50 ppm
nitrobenzene displaying this condition.
The adrenal gland, liver, and spleen were also target organs of nitrobenzene in B6C3F1
mice, as judged by the range of histopathologic lesions observed in the study. In the liver,
instances of centrilobular hyperplasia were noted in mid- (4/9) and high-concentration (9/9)
males, compared with 7/9 high-concentration females displaying these lesions. Table 4-18
provides a summary of the identified LOAELs for rats and mice.
45
-------
Table 4-18. Summary of effects observed in subchronic inhalation studies
with nitrobenzene
Species,
strain
Rat,
F344
Rat,
CDC
Mouse,
B6C3F1
Number
10/sex
Dosing
0, 5, 16, 50 ppm,
6 hr/d, 5 d/wk,
90 d
Effect3
Methemoglobinemia |
Organ weight t
Splenic congestion f
Testicular pathology t
Methemoglobinemia t
Spleen weight t
Splenic congestion f
Liver weight t
Testicular pathology t
Methemoglobinemia t
Splenic congestion f
NOAELb'c
(ppm)
NA(M)
5(F)
5 (M, F)
NA (M, F)
NA
5(M)
16(F)
16 (M)
5(F)
NA
5 (M, F)
5(M)
16 (M, F)
5(M)
NA(F)
LOAELbc
(ppm)
5(M)
16 (F)
16 (M, F)
5 (M, F)
5
16 (M)
50 (F)
50 (M)
16 (F)
5 (M, F)
16 (M, F)
16 (M)
50 (M, F)
16 (M)
5(F)
Reference
CUT (1984)
"Only endpoints with evident dose responses were selected, t = an increase in the respective endpoint.
bNo-observed-adverse-effect levels (NOAELs) and LOAELs determined by nitrobenzene assessment authors.
°M = male; F = female; CD = Sprague-Dawley; NA = not applicable.
4.2.2.2. Chronic Studies
A chronic inhalation study of nitrobenzene was conducted in F344 rats, Sprague-Dawley
(CD) rats, and B6C3F1 mice (Cattley et al., 1994; CUT, 1993). A total of 70 male and female
F344 rats and 70 male Sprague-Dawley (CD) rats were exposed to 0, 1, 5, or 25 ppm
nitrobenzene, and a total of 70 male and female B6C3F1 mice were exposed to 0, 5, 25, or
50 ppm nitrobenzene, 6 hours/day, 5 days/week, excluding holidays, for 2 years, resulting in a
total of 505 exposures. Animals were observed for clinical signs twice daily, with body weights
determined weekly for the first 13 weeks and twice weekly thereafter. Ten rats/sex/strain/group
were terminated 15 months into the study to provide samples for an interim evaluation of
hematologic parameters. For the scheduled interim and final sacrifices, animals were fasted
overnight, weighed, and then anesthetized using an i.p. injection of pentobarbital prior to
exsanguination. Mice were evaluated at study termination but not at 15 months. Among the
hematologic parameters evaluated were WBC counts, RBC counts, Hb, Hct, MCV, mean
corpuscular Fib (MCUb), red cell distribution width, and platelet count. In addition, a percentage
metUb value was determined, and the relative and absolute differential cell counts were
46
-------
determined microscopically. A wide range of tissues from high-concentration and control
animals and all gross lesions were processed for histopathologic examination. Tissues
considered to be specific targets of nitrobenzene, such as liver, spleen, and nose, were examined
microscopically in all exposure groups. Additional tissues were examined where significant
findings of toxicity had become evident in the high-dose group.
Effects of nitrobenzene on clinical signs, body weight changes, and survival appeared to
be sporadic and unrelated to dose. For example, during the first 2 weeks of exposure, nine
B6C3F1 mice died (across all exposure groups) and were replaced with substitutes of the same
age and shipment. Animals that died after the first 2 weeks of the study were not replaced. As
noted by the authors, the probability of surviving to term in the control groups was 60 and 45%
in male and female mice, 75 and 80% in male and female F344 rats, and 40% in the male CD
rats, thus providing sufficient statistical power to support conclusions about the incidence of any
late developing neoplastic lesions that became apparent at necropsy (Cattley et al., 1994; CUT,
1993). The probability of survival was not statistically significantly affected by exposure to
nitrobenzene since the actual percentage of mice available during final euthanasia was 55.7,
62.9, 65.7, and 68.6% among males and 44.3, 54.3, 64.3, and 47.1% among females in the
control, 5, 25, and 50 ppm nitrobenzene exposure groups, respectively. After rejecting autolyzed
specimens from animals that were found dead, all available tissues, including those from animals
that were sacrificed in a moribund state, were fixed and evaluated microscopically. The numbers
of examined lungs in mice were 68/70, 67/70, 65/70, and 66/70 in males and 53/70, 60/70, 64/70,
and 62/70 in females of the 0, 5, 25, and 50 ppm nitrobenzene exposure groups, respectively.
Other target organs in mice or rats also had similar or identical numbers of tissues examined per
group as specified here for the mouse lung (Cattley et al., 1994; CUT, 1993).
A summary of the positive findings of tumor formation in the study in animals with two
years of exposure is shown in Table 4-19. Animals sacrificed at 15 months (interim) were not
included in the analysis because they were deliberately removed from the study, rather than
being removed due to nitrobenzene-induced effects. In male F344 rats, the incidence of
combined adenomas and carcinomas in liver displayed a statistically significant trend and an
increased incidence with dose (16/46 in males receiving 25 ppm compared with 1/43 in
controls). However, this effect was not apparent in female F344 rats. Similarly, statistically
significant trends for dose-dependent increases in combined adenomas and carcinomas in kidney
and thyroid were observed in male F344 rats but not in females. However, there was a dose-
dependent trend and statistically significant increase in the incidence of endometrial polyps in
female F344 rats (19/49 in rats exposed to 25 ppm versus 9/48 in controls). The only compound-
related tumorigenic effect in CD rats was in males that showed statistically significant increases
in the incidences of combined adenomas and carcinomas in liver (5/23 in 25 ppm rats versus
0/23 in controls). As set forth in Table 4-19, there was a possible compound-related increase in
47
-------
the incidence of combined adenomas and carcinomas in the follicular cells of the thyroid in male
B6C3F1 mice. Other neoplastic responses to nitrobenzene observed in the mice included the
formation of adenocarcinomas of the mammary gland and an increased incidence of combined
adenomas and carcinomas of the lungs in males.
A number of noncarcinogenic responses to nitrobenzene were observed in the study
(Cattley et al., 1994; CUT, 1993). Both male and female F344 rats in the 25 ppm group
displayed treatment-related statistically significant reductions in RBCs, Hct, and Hb
concentration, with mean levels that were lower in animals sacrificed at term compared with
animals sacrificed at 15 months. Concentrations of metHb increased with increasing
nitrobenzene exposure, though time-related trends in this parameter were less clear-cut. Most
notable among the hematologic responses in CD rats were the increases in metHb in the
15-month interim blood samples, as shown in Table 4-20. These achieved statistical significance
(p < 0.01) versus controls at all dose concentrations employed in the study. No histopathology
was performed on the spleens of CD rats at interim or final sacrifice to determine if effects in the
spleen accompanied the statistically significant increase in metHb levels. It should be noted,
however, that, at final sacrifice, metHb levels were only increased in the 25 ppm exposure group,
which may indicate a compensatory response to metHb formation.10 It should also be pointed
out that background metHb levels in both strains of control rats were consistently higher at 24
months than at 15 months, resulting in apparently less pronounced relative changes at 24 months
than at 15 months among exposed animals (Table 4-20). Furthermore, metHb levels were much
lower among control rats in the 90-day inhalation or gavage studies (Tables 4-5, 4-6, and 4-17)
than among control rats at 15 or 24 months (Table 4-20). Collectively, these findings may
There is no known information in the literature on a specific possible compensatory response to
methemoglobinemia following extended exposure to metHb-forming chemicals. However, two enzyme systems,
one in the liver and the other in erythrocytes, may help attenuate metHb formation. Recently, Kurian et al. (2006,
2004) described a liver microsomal reductive pathway in human liver, known as NADH cytochrome b5 reductase,
that metabolizes and eliminates arylhydroxylamines, which are known to be rodent carcinogens and may be linked
to some human tumors. According to these studies, interindividual variability in expression of this enzyme system
is thought to account for cancer susceptibility to arylhydroxylamines. Though not explored in these studies, it is
possible that the same liver enzyme system may also partake indirectly in attenuating methemoglobinemia by
reducing phenylhydroxylamine (PHA) to aniline, thereby disallowing PHA from undergoing oxidation to
nitrosobenzene along with the concomitant conversion of Hb to metHb (Figure 3-8). In another study on a common
drug (dapsone), a similar enzyme seems to play a key role in mitigating methemoglobinemia by catalyzing the
reduction of the hydroxylamine back to dapsone (Tingle et al., 1997). According to the report by Kurian et al.
(2006), a similar or identical soluble enzyme system to the one reported in their studies on arylhydroxylamines is
expressed in erythrocytes to maintain Hb in its reduced state (Hultquist and Passon, 1971). This erythrocyte enzyme
system is likely NADH-dependent cytochrome b5 methemoglobin reductase that regenerates hemoglobin from
metHb as described earlier (footnote 2).
It can be hypothesized that both the liver and erythrocyte NADH-cytochrome b5 reductases are likely critical in
attenuating methemoglobinemia. On the one hand, the hepatic enzyme may help keep down the levels of PHA, a
key metHb-forming metabolite of nitrobenzene, while on the other hand, metHb is reduced to hemoglobin by the
erythrocytic enzyme. While this may apply to healthy individuals, others with congenital deficiency in these
enzymes may suffer from worsening methemoglobinemia over extended exposure to nitrobenzene (refer to footnote
2 and section 4.7.3).
48
-------
indicate a correlation between metHb and age of rats. However, in a 24-months dietary feeding
chronic toxicity study of aniline (CUT, 1982), metHb levels were variable among control F344
rats during the course of the study with no discernible correlation with age. In this study, control
levels (% metHb) at 26, 52, 78, and 104 weeks were 1.08, 1.87, 0,96, and 1.39 in males and 2.06,
1.12, 1.95, and 2.72 in females, respectively (CUT, 1982). Among the high-dose animals (100
mg/kg-day), the levels of metHb were 236, 129, 245, and 261% higher than among male controls
and 143, 146, 122, and 111% higher than among female controls at the respective intervals.
Some of the variations may be age-dependent changes, but it is also likely, at least partly, that
they are a consequence of study-to-study variation or are due to other artifacts. Nonetheless,
though it is hard to pinpoint a compensatory response, it can be concluded that
methemoglobinemia did not worsen as a function of exposure time to nitrobenzene.
49
-------
Table 4-19. Summary of neoplastic and nonneoplastic findings following 2-year inhalation exposure to
nitrobenzene
Site of increased
tumorigenicity
Sex with positive
carcinogenic response
Comments on neoplastic and/or nonneoplastic lesions"
B6C3F1 mouse
Lung: A/B adenoma or
carcinoma
Thyroid: follicular cell
adenoma
Mammary gland
Liver: hepatocellular
adenoma
M
M
F
F
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic r?
A significantly positive nitrobenzene exposure-related trend in incidence for A/B hyperplasia and
bronchiolization was observed.13
Nonneoplastic $
A significantly positive nitrobenzene exposure-related trend in incidence for bronchiolization was
observed.13
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic c^
A significantly positive nitrobenzene expo sure -related trend in incidence for follicular cell hyperplasia was
observed.13
Nonneoplastic $
A significantly positive nitrobenzene expo sure -related trend in incidence for follicular cell hyperplasia was
observed.13
Neoplastic $
Statistically significant difference in incidence for 50 ppm group versus controls when interim sacrifice
animals included0; 25 ppm and 5 ppm groups were not examined.
Neoplastic $
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic $
A significantly positive nitrobenzene exposure-related trend in incidence for centrilobular
hepatocytomegaly was observed.13
Nonneoplastic r?
A significantly positive trend in incidence of centrilobular hepatocytomegaly and multinucleated
hepatocytes was observed.13
F344/Nrat
Liver: hepatocellular
adenoma or carcinoma
M
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic r?
A significantly positive trend in incidence of eosinophilic foci and centrilobular hepatocytomegaly was
observed.13
50
-------
Table 4-19. Summary of neoplastic and nonneoplastic findings following 2-year inhalation exposure to
nitrobenzene
Site of increased
tumorigenicity
Thyroid: follicular cell
adenoma or
adenocarcinoma
Kidney: tubular
adenoma or carcinoma
Endometrial stromal
polyp
Sex with positive
carcinogenic response
F
M
M
F
Comments on neoplastic and/or nonneoplastic lesions"
Neoplastic $
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic $
A significantly positive trend in incidence of eosinophilic foci was observed13
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic r?
A significantly positive trend in incidence of follicular cell hyperplasia was observed.13
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence.13
Nonneoplastic r?
A significantly positive trend in incidence of tubular hyperplasia was observed.13
Nonneoplastic $
Significantly positive nitrobenzene exposure-related trend in incidence.13
Sprague-Dawley rat
Liver: hepatocellular
adenoma or carcinoma
M
Neoplastic r?
Significantly positive nitrobenzene exposure-related trend in incidence13
Nonneoplastic r?
A significantly positive trend in incidence of eosinophilic foci and centrilobular hepatocytomegaly was
observed.13
aThe sex of the animal is the same as the sex that exhibited a positive carcinogenic response, unless indicated otherwise (male $ or female
bCochran-Armitage trend test, p < 0.05, as calculated by the study authors.
°Fisher's exact test, p < 0.05, as calculated by the study authors.
Sources: Cattley et al. (1994); CUT (1993).
51
-------
In mice, RBCs and Hct were statistically significantly lower in 50 ppm males than in
controls (8.70 ± 0.12 versus 9.61 ± 0.29 x 106 cells/uL and 41.64 ± 0.52 versus 45.06 ± 1.15%,
respectively). In common with the rats, there were statistically significant increases in metHb
concentrations in high-dose mice of both sexes compared with controls (Table 4-20).
Table 4-20. Percentage metHb formation in response to inhaled nitrobenzene
Treatment group
MetHb (%)
Interim sacrifice (15 months)
Males
Females
Terminal sacrifice (24 months)
Males
Females
B6C3F1 mice
0
5
25
50
NAa
NA
NA
NA
NA
NA
NA
NA
1.97 ±0.24
1.94 ±0.34
3.02 ±0.41
3.97±0.48C
1.39 ±0.20
1.37 ±0.18
2.22 ± 0.26b
2.79 ± 0.24C
F344 rats
0
1
5
25
2.90 ±0.31
3.21 ±0.18
3. 18 ±0.43
4.73±0.52C
2.35 ±0.36
3.33 ±0.40
3. 17 ±0.39
5.90±0.96C
3.88 ±0.33
3.31 ±0.32
4.19 ±0.53
5.27±0.33C
2.68 ±0.37
2.13 ±0.16
2.54 ±0.30
5.00±0.45C
CD rats
0
1
5
25
1.18 ±0.34
4.08±0.80C
6.22±1.60C
5.85±0.83C
NA
NA
NA
NA
2.75 ±0.52
2.87 ±0.34
2.35 ±0.32
4.60±0.53C
NA
NA
NA
NA
aNA = not applicable.
bp < 0.05.
Source: Cattley et al. (1994).
Numerous noncancerous histopathologic lesions resulted from nitrobenzene inhalation,
though some of these responses were not clear-cut because of a high incidence of the same effect
in controls, which left the possibility that the response might be a nonspecific lesion due to age.
For example, chronic nephropathy and extramedullary hematopoiesis of the spleen occurred in
controls and at all concentration levels in both sexes of F344 rats and in male Sprague-Dawley
rats. However, a number of histopathologic effects of nitrobenzene appeared to be compound
related, including those in the nose, spleen, liver, kidney, and testis (Table 4-21).
Pigmentation of the olfactory epithelium was dose-dependently increased in male and
female rats, with incidences of 99% in male F344 rats versus 60% of controls, 95% in male CD
rats versus 67% of controls, and 100% in female F344 rats versus 55% of controls in the high-
exposure groups. An increased incidence of focal inflammation and hypertrophy of the
submucosal glands in areas lined by respiratory epithelium was observed in the nasal region of
52
-------
high-exposure male and female F344 rats. In CD rats, exposure-related lesions in nasal sections
consisted of a slight increase in the incidence and severity of inflammatory changes in the
anterior section of the nose. Splenic pigmentation was assessed in male and female F344 rats.
In male F344 rats, an exposure-related increase was observed (100% of 25 ppm exposed animals
versus 80% of controls). In contrast, 99% of female rats were found with this endpoint in the
highest exposure group compared to 90% of controls. Liver effects exhibited a mixed response
with respect to exposure-dependent changes. Hepatic eosinophilic foci were observed in a dose-
dependent manner in 81 and 23% of male and female F344 rats at the highest dose (25 ppm)
compared with 38 and 8.6% of controls, respectively. Male F344 rats exhibited an exposure-
dependent increase in spongiosis hepatis (83% of animals at 25 ppm versus 36% of controls),
whereas this endpoint was observed with only the high-exposure groups in 57% of male CD rats
compared to 40% of controls and 9% of female F344 rats versus 0% of controls. The number of
male rats presenting with centrilobular hepatocytomegaly at necropsy was increased at 5 and
25 ppm nitrobenzene, with 81% of F344 rats and 60% of CD rats afflicted at the highest
exposure level compared with 0 and 5% of controls, respectively; however, this endpoint was not
detected in female F344 rats, regardless of exposure level. Changes in the kidney were restricted
to the high-exposure group in male F344 rats, with less clear exposure-related changes in female
F344 rats. Tubular hyperplasia was detected in 19% of male F344 rats versus 3% of controls,
only 3% of female F344 rats at 5 and 25 ppm nitrobenzene, and none of the controls. Testicular
changes were assessed in male CD rats. Clear exposure-dependent changes were observed for
bilateral atrophy of the testis (57% at the highest dose; 18% of controls) and bilateral
hypospermia of the epididymis (54% at the highest dose; 13% of controls).
In mice, tissue sites displaying increased incidence of nonneoplastic lesions included
lung, olfactory epithelium, and, in the males, thyroid follicular cells and hepatocytes (Table 4-
22). Histopathologic endpoints for the lung included hyperplasia and alveolar bronchiolization.11
In male mice, a clear exposure-dependent increase in hyperplasia was found, up to 20% in high-
exposure animals versus 1.5% of controls. In contrast, female mice displayed a mixed response,
with findings of hyperplasia in 3% of animals at 5 ppm, 8% at 25 ppm, and 2% at 50 ppm versus
controls. Bronchiolization of the alveoli was increased at all exposure levels (male mice: 5 ppm,
87%; 25 ppm, 89%; and 50 ppm, 94%; female mice: 5 ppm, 92%; 25 ppm, 98%; and 50 ppm,
100%). This endpoint was not detected in any controls. Additional effects of nitrobenzene on
the respiratory tract were noted with statistically significant increases in the number of animals
11 According to Nettesheim and Szakal (1972), bronchiolization of alveoli are lesions that are thought to arise from
the ' 'colonization'' of alveolar walls with bronchiolar epithelium either via cell migration through alveolar pores or
from the transformation (metaplasia) of alveolar type II cells into bronchiolar-type epithelium. The pathology
summary in the nitrobenzene study characterized bronchiolization of the alveolar walls as "a pronounced change in
the alveolar epithelium in the region of the terminal bronchioles from a simple squamous to tall columnar epithelium
resembling that of the terminal bronchioles" (CUT, 1993). Section 4.5.2 has additional details on bronchiolization
of alveoli in the nitrobenzene study and in other studies of animals exposed to other agents as well as a discussion of
the relevance of this finding to humans.
53
-------
presenting with pigmentation and degeneration of the olfactory epithelium in the nasal region.
Pigmented olfactory epithelium was detected in 74 and 48% of high-dose male and female mice,
respectively. Similarly, an exposure-dependent increase in degenerated olfactory epithelium
occurred in mice of both sexes (male mice: control, 1%; 5 ppm, 2%; 25 ppm, 49%; and 50 ppm,
62%; female mice: control, 0%; 5 ppm, 32%; 25 ppm, 75%; and 50 ppm, 69%). Lesions noted
in nasal sections increased in severity with increasing dose (CUT, 1993); however, severity
scores were not reported. A differential response was observed between male and female mice
with histopathologic endpoints in the thyroid and liver. In the thyroid, an exposure-dependent
increase in follicular cell hyperplasia, up to 19% at 50 ppm, was found in male mice versus 2%
of controls, whereas this effect was only observed in females up to 13% compared to 4% of
controls at the highest exposure (50 ppm). In the liver, male mice presented with exposure-
dependent changes in centrilobular hepatocytomegaly and multinucleated hepatocytes, up to
89 and 88%, respectively. In contrast, centrilobular hepatocytomegaly was undetectable in
female mice, except for the highest dose (11% above controls), as were multinucleated
hepatocytes (3% above controls).
54
-------
Table 4-21. Selected noncancer histopathologic changes in rats as a result of exposure to nitrobenzene via
inhalation for 2 years
Target tissue
Exposure concentration (ppm)
Males
0
1
5
25
Females
0
1
5
25
F344 rats
Liver
Eosinophilic foci
Centrilobular hepatocytomegaly
Spongiosis hepatic
Kidney
Tubular hyperplasia
Nose
Pigmented olfactory epithelium
Spleen
Pigmentation
26/69
0/69
25/69
2/69
40/67
55/69
25/69
0/69
24/69
2/68
53/67
63/69
44/703
8/70a
33/70
2/70
67/70
64/70
57/703
57/703
58/703
13/703
68/69a
70/703
6/70
0/70
0/70
0/70
37/67
62/69
9/66
0/66
0/66
0/66
54/65
61/66
13/66
0/66
0/66
2/66
60/65
60/66
16/703
0/70
6/70a
2/70
66/66a
68/69a
CD rats
Liver
Centrilobular hepatocytomegaly
Spongiosis hepatic
Nose
Pigmented olfactory epithelium
Testis
Bilateral atrophy
Epididymis
Bilateral hypospermia
3/63
25/63
42/63
11/62
8/60
1/67
25/67
49/64
17/66
13/65
14/703
25/70
60/66
22/70
15/67
39/65a
37/65a
58/6 la
35/613
32/59a
"Statistically significantly different from control values, as calculated by the authors.
Sources: Cattley et al. (1994); CUT (1993).
55
-------
Table 4-22. Selected noncancer histopathologic changes in B6C3F1 mice as a result of exposure to nitrobenzene
via inhalation for 2 years
Target tissue
Liver
Centrilobular hepatocytomegaly
Multinucleated hepatocytes
Lung
Hyperplasia
Bronchiolization
Thyroid
Follicular cell hyperplasia
Nose
Pigmented olfactory epithelium
Degenerated olfactory epithelium
Exposure concentration (ppm)
Males
0
1/68
2/68
1/68
0/68
1/65
0/67
1/67
5
15/65
14/65
2/67
58/67a
4/65
7/66
1/66
25
44/65a
45/65a
8/65a
58/65a
7/65a
46/65a
32/65a
50
57/64a
56/64a
13/66a
62/66a
12/64a
49/66a
41/66a
Females
0
0/51
0/51
0/53
0/53
2/49
0/52
0/52
5
0/61
0/61
2/60
55/60a
1/59
6/60a
19/60a
25
0/64
0/64
5/64a
63/64a
1/61
37/63a
47/63a
50
7/62a
2/62a
1/62
62/62a
8/61
29/6 la
42/6 la
""Statistically significantly different from control values, as calculated by the authors.
Sources: Cattley et al. (1994); CUT (1993).
56
-------
A synopsis of the effects observed from chronic nitrobenzene inhalation in animals is
presented in Table 4-23.
Table 4-23. Summary of effects observed from chronic inhalation with
nitrobenzene at terminal sacrifice
Species,
strain
Rat,
F344
Rat,
CD
Mouse,
B6C3F1
Number
70/sex
70 males
70/sex
Dosing
0, 1, 5, 25 ppm
6 hr/d, 5 d/wk,
2y
0, 5, 25, 50 ppm
6 hr/d, 5 d/wk,
2y
Effect3
Methemoglobinemia |
Liver, eosinophilic foci t
Methemoglobinemia |
Hepatocytomegaly t
Methemoglobinemia t
Bronchiolization |
NOAELb'c
(ppm)
5 (M, F)
1 (M), 5 (F)
5
1
25 (M, F)
NA
LOAELbc
(ppm)
25 (M, F)
5 (M), 25 (F)
25
5
50 (M, F)
5 (M, F)
Reference
CUT (1993)
aOnly endpoints with evident dose responses were selected, t = increase in the respective endpoint.
bNo-observed-adverse-effect levels (NOAELs) and LOAELs determined by nitrobenzene assessment authors.
°M = male; F = female; CD = Sprague-Dawley.
4.2.3. Dermal Exposure
4.2.3.1. Subchronic Studies
NTP sponsored a 90-day skin painting toxicological study (NTP, 1983b) with
nitrobenzene in F344 rats and B6C3F1 mice. The authors treated F344 rats and B6C3F1 mice
(10 animals/sex/group) with 50, 100, 200, 400, and 800 mg/kg-day nitrobenzene in acetone, the
responses being compared with those in animals painted with acetone alone. At 800 mg/kg-day,
all rats and 9/10 male and 8/10 female mice died before the end of the experiment. Furthermore,
surviving animals in the other exposure groups (dose levels not stated) displayed profound
clinical signs of acute toxicity, including ataxia, dyspnea, circling, lethargy, and insensitivity to
pain. Only female mice showed a dose-related increase in metHb concentration. Among the
histopathologic findings, there was a marked degeneration of the testes in the males of both
species and all nitrobenzene-receiving rats displayed congestion of the spleen. The incidence of
congestion of the lungs was dose-dependently increased in males and females of both species.
Vacuolization of the brain or brain stem was another characteristic histopathologic finding, the
effects becoming apparent in rats exposed to nitrobenzene at 100 mg/kg or higher, in male mice
exposed to 800 mg/kg, and in female mice exposed to 400 and 800 mg/kg nitrobenzene.
Tables 4-24, 4-25, 4-26, and 4-27 document these histopathologic changes.
57
-------
Table 4-24. Incidence of histopathologic lesions in male F344 rats exposed to
nitrobenzene for 90 days via dermal exposure
Target tissue
Lung
Congestion
Spleen
Congestion
Hematopoiesis
Lymphoid atrophy
Liver
Congestion
Kidney
Congestion
Testis
Atrophy
Hypospermatogenesis
Multinucleate giant cells
Brain
Hemorrhage
Dose (mg/kg-day)
0
1/10
0/10
10/10
0/10
0/10
0/10
0/10
0/10
0/10
1/10
50
1/10
10/10
10/10
0/10
1/10
0/10
0/10
0/10
0/10
4/10
100
7/10
10/10
10/10
7/10
0/10
0/10
0/10
0/10
0/10
0/10
200
4/10
10/10
10/10
7/10
0/10
0/10
0/10
0/10
0/10
0/10
400
4/10
10/10
10/10
10/10
0/10
0/10
10/10
10/10
9/10
2/10
800
10/10
10/10
10/10
10/10
6/10
7/10
10/10
10/10
10/10
2/10
Source: NTP(1983b).
Table 4-25. Incidence of histopathologic lesions in female F344 rats exposed
to nitrobenzene for 90 days via dermal exposure
Target tissue
Lung
Congestion
Spleen
Congestion
Hematopoiesis
Lymphoid atrophy
Liver
Congestion
Kidney
Congestion
Uterus
Atrophy
Brain
Hemorrhage
Cerebrum
White matter vacuolization
Cerebellum
White matter vacuolization
Brain stem
Hemorrhage
Vacuolization
Dose (mg/kg-day)
0
1/10
8/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
50
1/10
10/10
10/10
0/10
0/10
0/10
0/10
1/10
0/10
0/10
1/10
0/10
100
3/10
10/10
10/10
0/10
0/10
0/10
0/10
5/10
10/10
8/10
1/10
10/10
200
1/10
9/10
10/10
1/10
0/10
0/10
0/10
2/10
10/10
4/10
4/10
8/10
400
6/10
10/10
10/10
9/10
0/10
4/10
0/10
1/10
4/10
7/10
7/10
4/10
800
9/10
10/10
10/10
10/10
4/10
4/10
6/10
2/10
3/10
6/10
6/10
3/10
Source: NTP(1983b).
58
-------
Table 4-26. Incidence of histopathologic lesions in male B6C3F1 mice
exposed to nitrobenzene for 90 days via dermal exposure
Target tissue
Lung
Congestion
Spleen
Congestion
Hematopoiesis
Lymphoid atrophy
Liver
Congestion
Pigmentation
Thymus
Atrophy
Testis
Atrophy
Hypospermatogenesis
Multinucleate giant cells
Brain
Hemorrhage
Brain stem
Hemorrhage
Degeneration
Skin
Inflammation
Dose (mg/kg-day)
0
2/10
0/10
1/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
1/10
1/10
0/10
0/10
50
6/10
0/10
3/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
1/10
1/10
0/10
0/10
100
4/10
0/10
3/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
3/10
2/10
0/10
0/10
200
4/10
0/10
9/10
0/10
1/10
0/10
0/10
0/10
0/10
0/10
1/10
1/10
0/10
0/10
400
10/10
0/10
9/10
0/10
10/10
0/10
0/10
5/10
2/10
0/10
0/10
1/10
0/10
8/10
800
9/10
10/10
10/10
3/10
10/10
6/10
111
10/10
10/10
4/10
2/10
6/10
3/10
3/10
Source: NTP(1983b).
59
-------
Table 4-27. Incidence of histopathologic lesions in female B6C3F1 mice
exposed to nitrobenzene for 90 days via dermal exposure
Target tissue
Lung
Congestion
Spleen
Congestion
Hematopoiesis
Lymphoid atrophy
Liver
Cytomegaly
Thymus
Atrophy
Ovary
Atrophy
Uterus
Atrophy
Adrenal cortex
Fatty change
Brain
Hemorrhage
Brain stem
Hemorrhage
Degeneration
Skin
Inflammation
Dose (mg/kg-day)
0
4/10
0/10
7/10
0/10
0/10
0/10
0/10
0/10
0/10
0/10
1/10
0/10
0/10
50
3/10
0/10
4/10
0/10
0/10
0/10
0/10
0/10
6/10
1/10
0/10
0/10
0/10
100
2/10
1/10
3/10
1/10
0/10
0/10
0/10
0/10
9/10
0/10
0/10
0/10
0/10
200
4/10
0/10
7/10
0/10
0/10
0/10
0/10
1/10
10/10
1/10
0/10
0/10
0/10
400
8/10
2/10
10/10
0/10
0/10
0/10
0/10
1/10
8/10
3/10
2/10
1/10
9/10
800
10/10
9/10
9/10
3/10
8/10
9/9
3/10
5/10
2/10
2/10
4/10
3/10
7/10
Source: NTP(1983b).
A summary of the animal toxicity studies with nitrobenzene following dermal
administration is presented in Table 4-28.
Table 4-28. Summary of effects observed in dermal dosing studies with
nitrobenzene
Species,
strain
Rat, F344
Mouse,
B6C3F1
Number
10/sex
Dosing
0, 50, 100,
200, 400,
800 mg/kg-
day, 90 d
Effect3
Splenic congestion f
Lung congestion f
Brain pathology t
Testicular pathology t
Splenic hematopoiesis t
Testicular pathology t
Mortality t
NOAELb'c
(mg/kg-day)
NA
50 (M, F)
50 (F)
200 (M)
100 (M, F)
200 (M)
NA
LOAELbc
(mg/kg-day)
50 (M, F)
100 (M, F)
100 (F)
400 (M)
200 (M, F)
400 (M)
800 (M, F)
Reference
NTP
(1983b)
aOnly endpoints with evident dose responses were selected, t = an increase in the respective endpoint.
bNo-observed-adverse-effect levels (NOAELs) and LOAELs determined by nitrobenzene assessment authors.
°M = male; F = female; NA = not applicable.
60
-------
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
4.3.1. Oral Exposure
Levin et al. (1988) investigated adverse effects of nitrobenzene on spermatogenesis that
might be associated with impaired testicular function by surgically routing the vas deferens of
male F344 rats to the bladder. This permitted spermatogenesis to be continually monitored
during and after exposure to nitrobenzene. Six rats/group were subjected to this surgical
procedure and, after a recovery period of 6 weeks, gavaged with a single dose of 300 mg/kg
nitrobenzene in corn oil. Controls received corn oil alone. Animals were housed in metabolic
cages and assessed for the release of sperm to the urine for up to 100 days. Two other groups of
rats, 45 exposed and 30 controls, were gavaged in a manner similar to the surgically altered
subjects. These were serially sacrificed for histopathologic examination at various time points,
up to 100 days. Output of sperm held steady after nitrobenzene administration for about 20 days
then dropped to zero within 12 days and persisted at this level until day 48. Fifty days after
treatment, sperm began to reappear in the urine of treated animals, ultimately achieving about
78% of control levels. Histopathologically, treated animals displayed degeneration of the
seminiferous epithelium within 3 days of treatment, an effect characterized by the appearance of
pachytene-derived giant cells and loss of the more mature elements of the seminiferous
epithelium. As discussed by the authors, the pachytene spermatocytes (found in stages VI-XIII)
were the most sensitive to the effects of the compound. Clear histopathologic signs of
regeneration were apparent at about 21 days after treatment. However, at least some signs of the
abnormal cellular architecture and tubular organization described above always remained. For
example, approximately 10% of the tubules examined showed little evidence of spermatogenesis
even at 8 weeks posttreatment, with mature spermatids rarely apparent. The authors interpreted
their results in accordance with the known processes and time frame by which spermatogenesis
occurs in F344 rats and presented a nomogram that correlated the spermatogenic cycle of the rats
with the proposed chronology of nitrobenzene-induced lesions.
Bond et al. (1981) administered a single oral dose of 0, 50, 75, 110, 165, 200, 300, or
450 mg/kg nitrobenzene in corn oil to six male F344 rats/group. Three rats at each dosage were
sacrificed 2 and 5 days following nitrobenzene administration. Samples of blood were obtained
by cardiac puncture to measure metHb, and 25 tissues and organs were excised for
histopathologic examination. The liver, testes, and brain from all animals in the study were
examined histopathologically, whereas histologic sections of other tissues were examined only in
the high-dose and control groups. Hepatic centrilobular necrosis appeared inconsistently in rats
given various doses of nitrobenzene, while hepatocellular nucleolar enlargement was
consistently detected in rats given doses of nitrobenzene as low as 110 mg/kg. Lesions occurred
in the seminiferous tubules of the testicles, with marked necrosis of primary and secondary
spermatocytes following a single oral dose of 300 mg/kg (Bond et al., 1981). Furthermore,
within 3 days of nitrobenzene administration, multinucleated giant cells were observed, and
61
-------
decreased numbers of spermatozoa were observed in the epididymis. Histopathologic analyses
indicated that nitrobenzene had no apparent effects on spermatogonia or the epididymal
epithelium. In parallel to the observed histopathologic lesions in liver and testes,
methemoglobinemia was increased to 25% immediately after dosing at 300 mg/kg, with a
subsequent slow decline over the next 10 days. In a control experiment, the administration of
sodium nitrite also induced methemoglobinemia but had no histopathologic effects on the testes
and liver, suggesting that the histopathologic effects of nitrobenzene occurred through a direct
action of the compound or its metabolites at the tissue site rather than as a secondary effect of
metHb formation.
Two further studies confirmed the association between orally administered nitrobenzene
and the onset of toxic effects in the testes and epididymides. In the first study, Matsuura et al.
(1995) gavaged 10-week-old male Sprague-Dawley rats with 30 or 60 mg/kg nitrobenzene,
5 days/week for 3 weeks. Parameters evaluated included the weights and histopathology of the
testes and epididymides, together with an analysis of the count, motility, viability, and
morphology of the sperm. Nitrobenzene at the high dose (60 mg/kg) induced a relative decrease
in the weight of the epididymis, decreases in sperm motility and viability, and an increase in the
incidence of morphologically abnormal sperm. Degeneration and decreases in spermatids and
pachytene spermatocytes were specified as primary effects of nitrobenzene at this dose level. In
the second study, Koida et al. (1995) gavaged several groups of five male Sprague-Dawley rats
of different ages (6, 8, 10, and 40 weeks old) with 50 mg/kg-day nitrobenzene in sesame oil for
2 or 4 weeks. All subjects were examined for changes in testis and epididymis weights
(compared with controls), differential morphology and histopathology, and altered sperm counts.
In general, treatment was associated with reduced sperm counts and depressed sperm activity,
with some histopathologic changes evident in the reproductive organs of younger animals.
Kawashima et al. (1995a) administered nitrobenzene (60 mg/kg-day in sesame oil by
gavage) to male Sprague-Dawley rats for periods of time from 7-70 days, after which the
animals were mated with untreated females and then terminated the following day. Comparative
changes in testicular and epididymal weights, sperm count, motility, and viability were
evaluated, along with the fertility and copulation indices of treated groups. Large reductions in
testicular (>50%) and epididymal weights, sperm count, and motility were observed in those
animals exposed to nitrobenzene for 14 days, while sperm viability and the fertility index were
severely reduced in those males exposed to nitrobenzene for 21 days or more. There was a
concomitant increase in the incidence of abnormal sperm. While the copulation indices of
treated males appeared unchanged with duration of exposure, the numbers of virgin females
becoming pregnant by treated males declined markedly with duration of exposure. No mating
females became pregnant in groups that were mated with males treated for 28 days or longer, an
effect that appeared to result from the production of sperm with poor motility and reduced
viability.
62
-------
Kawashima et al. (1996, 1995b) used computer-imaging systems to evaluate the motility
of sperm from rats gavaged with nitrobenzene. For example, they described an experimental
protocol in which, in the first study, male Sprague-Dawley rats were gavaged with 60 mg/kg-day
nitrobenzene for up to 2 weeks (Kawashima et al., 1995b). Sperm from treated and control rats
were evaluated in an image processor that used motion analysis software to quantify such
parameters as curvilinear distance, curvilinear velocity, and amplitude of lateral head
displacement. The values of each motility parameter were lower in the sperm of nitrobenzene-
exposed rats. These researchers also used computer-assisted sperm analysis to evaluate sperm
motility in Sprague-Dawley rats exposed to up to 60 mg/kg-day by gavage for up to 28 days
(Kawashima et al., 1996). All sperm motility parameters in rats exposed to 30 and 60 mg/kg-day
were lower than in controls, irrespective of exposure duration. Such parameters as curvilinear
velocity, straight-line velocity, and motility rate were lower in rats exposed at the lowest dose
level (15 mg/kg-day) for 28 days.
Other abstracts of studies attested to the impact of nitrobenzene on sperm viability and
motility when administered to rats via the oral route (Kito et al., 1999, 1998; Kato et al., 1995).
In one example, Kato et al. (1995) exposed rats (number and strain not stated) to nitrobenzene at
concentrations up to 60 mg/kg and used a vital dye (ethidium homodimer) to show loss of sperm
viability compared with equivalent samples from untreated rats. Viable sperm from
nitrobenzene-receiving animals showed reduced motility. In a more recent full-length research
report (Ban et al., 2001), nitrobenzene was used as one of several recognized testicular toxicants
to evaluate the utility of different parameters in sperm motion analysis. Curvilinear velocity and
mean amplitude of lateral head movement were considered to be among the more sensitive
indicators of impaired sperm motility.
Linder et al. (1992) had likewise included nitrobenzene as a positive control in a survey
of compounds for spermatotoxic effects in male Sprague-Dawley rats. The experimental
protocol featured oral administration of the compound as a single dose of 300 mg/kg. A number
of well-characterized spermatotoxic tests were employed, including counts of sperm heads,
sperm velocity, sperm morphology, and the histopathology of the testis and epididymis. Marked
changes observed in nitrobenzene-receiving rats included degenerating and missing pachytene
spermatocytes in stages VII-XIV, some multinucleated giant cells, the existence of testicular
debris, and an increase in the number of morphologically abnormal sperm.
Mitsumori et al. (1994) reported a reproductive toxicity study on nitrobenzene that
employed a complex protocol proposed by the Organization for Economic Cooperation and
Development (OECD). Ten Sprague-Dawley rats/sex/group were gavaged with 0, 20, 60, or
100 mg/kg-day nitrobenzene in sesame oil for a 14-day premating period, a mating period of up
to 14 days, a gestation period of 22 days, and a subsequent lactation period of 4 days, making a
potential overall dosing period of 54 days, at which point all animals (males, females, and pups)
were necropsied. Because the observed mating period was no more than a single day for most
63
-------
mating pairs, the actual dosing duration for males and females was 40-41 days but could have
lasted as long as 54 days for some. Clinical signs were observed daily, and body weights and
food consumption were monitored weekly. A complete range of hematologic and clinical
chemistry parameters was measured in blood and serum samples collected from the males prior
to termination. At necropsy, weights of liver, kidneys, thymus, adrenals, spleen, testes,
epididymides, and ovaries were noted. The numbers of corpora lutea and implantation sites were
counted in females. Excised pieces of brain, heart, liver, kidneys, adrenals, spleen, ovaries,
testes, and epididymides were fixed and processed for histopathologic examination.
High-dose animals displayed a number of clinical signs as a result of nitrobenzene
administration, including piloerection, salivation, emaciation, and an apparent anemia from day
13 onward. A number of behavioral/neurological signs were evident and body weight and food
consumption were reduced by 17% in the high-dose males from day 21 onwards. Male rats
displayed profound dose-related changes in the levels of some hematologic parameters, including
decreases in RBCs, Hb, and Hct and increases in metHb, MCHb, WBCs, reticulocytes, and
erythroblasts. For a number of these parameters, statistically significant differences from
controls were observed in the low-dose group (Table 4-29). At necropsy, the relative liver,
kidney, and spleen weights were statistically significantly increased, and those of testes and
epididymides were significantly decreased in the 60 and 100 mg/kg-day animals compared with
controls. However, in rats exposed to 20 mg/kg-day nitrobenzene there was a slight upward
fluctuation in relative testis and epididymis weights compared with controls (Table 4-30).
Table 4-29. Hematologic findings in male Sprague-Dawley rats exposed via
gavage to nitrobenzene
Parameter
RBC (1012/L)
Hb(g/L)
MetHb (%)
Packed cell volume (%)
Mean cell volume (fL)
MCHb (pg)
Reticulocytes (per 1,000 RBCs)
Erythroblasts (per 200 WBCs)
WBCs (109/L)
Dose (mg/kg-day)a
0
8.96 ±0.23
15.3 ±0.6
0.70 ±0.69
45.0 ±1.8
50.2 ±1.1
17.1 ±0.4
34.1±21.1
2.3 ±2.6
4.65 ± 1.49
20
7.75 ± 0.40b
13.6±0.6b
3.64 ±3.14°
40.7±1.8b
52.5 ±1.7
17.5 ±0.5
64.2 ±23.0
7.0 ±4.9
4.69 ±1.0
60
6.44 ± 0.44b
13.3±0.7b
4.79 ± 1.09b
38.5±2.2b
59.8±2.4b
20.8±0.8b
116.6±24.4b
18.7 ±16.6°
4.12 ±1.28
100
5.28 ± 0.44b
12.9±1.0b
6.76 ± 2.07b
36.5±2.3b
69.3 ± 5.2b
24.5 ± 1.0b
223.0 ±60.9b
19.6±14.6b
16.42 ± 7.70°
"Values are means ± standard deviations.
bp < 0.01 versus controls, as calculated by the authors.
cp < 0.05 versus controls, as calculated by the authors.
Source: Mitsumorietal. (1994).
64
-------
Table 4-30. Relative organ weights of male Sprague-Dawley rats gavaged
with nitrobenzene
Organ3
Liver
Kidney
Spleen
Testes
Epididymides
Dose (mg/kg-day)
0
2.87 ±0.24
0.64 ± 0.04
0.18 ±0.01
0.79 ±0.04
0.28 ±0.02
20
3.38±0.17b
0.67 ±0.05
0.29 ± 0.04b
0.83 ±0.07
0.31 ±0.04
60
3.94±0.30b
0.73±0.05b
0.51±0.07b
0.32±0.04b
0.23±0.05b
100
4.15±0.20b
0.84 ± 0.07b
0.67±0.14b
0.37±0.07b
0.20 ± 0.02b
aValues are grams per 100 grams body weight, means ± standard deviations.
bp < 0.01 versus controls, as calculated by the authors.
Source: Mitsumorietal. (1994).
A wide range of histopathologic consequences of nitrobenzene treatment was observed,
especially in animals receiving 60 and 100 mg/kg-day of the compound. These included atrophy
of the seminiferous tubules, hyperplasia of Ley dig cells, and loss of intraluminal sperm in the
epididymides. Such histopathologic lesions as centrilobular swelling of hepatocytes,
hemosiderin deposition in Kupffer cells, and increased extramedullary hematopoiesis in the liver
and spleen were seen in all exposed groups. Neuronal necrosis/gliosis in the cerebellar medulla
was evident in rats exposed to 60 and 100 mg/kg-day nitrobenzene.
Among the reproductive/developmental parameters that were evaluated, there were no
statistically significant differences from controls in the copulation and fertility indices at any
dose level. However, among the dams, only two of nine pregnant females in the high-dose group
survived to term, with the subsequent deaths of the two survivors (and their reduced litters)
occurring on days 1 and 3 of lactation. In the remaining offspring, pup body weights were
statistically significantly decreased at day 0 for both males and females by approximately 10% in
the 60 mg/kg-day group. At day 4, body weights in male pups were statistically significantly
decreased by about 5% in the 20 mg/kg-day group and by about 25% in both male and female
pups in the 60 mg/kg-day group.
A synopsis of no-observed-adverse-effect levels (NOAELs) and LOAELs, as identified
by the nitrobenzene assessment authors, from Mitsumori et al. (1994) is presented in Table 4-31.
65
-------
Table 4-31. Summary of effects observed in an oral reproductive study with
nitrobenzene
Species,
strain
Rat,
Sprague-
Dawley
Number
10/sex
Dosing
0, 20, 60,
100 mg/kg-
day, gavage,
up to 54 d
Effect3
Organ weights |.
Testicular pathology |.
Copulation, fertility.
Developmental toxicity:
day 4 male pup body weights |.
Mortality.
NOAELb
(mg/kg-day)
NA
20
100
NA
60 (M, Fo)
LOAELb
(mg/kg-day)
20 (M, Fo)
60 (M, Fo)
NA
20
100 (M, Fo)
aOnly endpoints with evident dose responses were selected, t = increase in the respective endpoint. J, = decrease in
the respective endpoint.
bM = male; F0= parental generation; NA = not applicable.
Source: Mitsumorietal. (1994).
Sertoli cells control spermatogenesis via the secretion of different proteins varying
cyclically according to the stage of spermatogenesis. In order to assess the possibility of
identifying chemical-induced, stage-specific changes in protein secretion, McLaren et al. (1993a)
employed a novel experimental approach to examine the in vivo effects of nitrobenzene (single
oral dose of 300 mg/kg) and m-dinitrobenzene, using seminiferous tubules from male Wistar rats
at different stages of the spermatogenic cycle. Tissue extracts then were cultured in vitro for
24 hours with [35S]-methionine. Incorporation of [35S]-methionine served as a marker for the
secretion of newly formed polypeptides in response to challenges with nitrobenzene or
m-dinitrobenzene, a well-characterized Sertoli cell toxicant. In other experiments, seminiferous
tubules were exposed to nitrobenzene and m-dinitrobenzene in vitro in the presence of [35S]-
methionine. Using two-dimensional SDS-PAGE or isoelectric focusing, the authors were able to
identify six marker proteins, normally produced in the tubules, whose secretion was changed as a
result of exposure to nitrobenzene or m-dinitrobenzene. For the most part, the abundance of
these marker proteins was reduced in response to nitrobenzene, as compared with controls. One
component, however, MP-4, a structural protein in Sertoli cells, had not been apparent
previously in the secretions of seminiferous tubule cells from control animals but appeared in
detectable amounts in the polypeptide secretions from nitrobenzene-exposed seminiferous
tubules. Further work demonstrated that the toxicological effects of nitrobenzene, such as those
outlined above, did not occur in isolates from immature rats, thus suggesting an age specificity of
the nitrobenzene- and m-dinitrobenzene-induced responses (McLaren et al., 1993b).
Morrissey et al. (1988) evaluated rodent sperm, vaginal cytology, and reproductive organ
weight data from a series of NTP 13-week gavage studies, one of which was on nitrobenzene
(NTP, 1983a). As tabulated by Morrissey et al. (1988), the effects of nitrobenzene on the
reproductive organs and the incidence of abnormal sperm were assessed at dose levels of 0, 9.4,
37.5, and 75 mg/kg in rats and at 0, 18.75, 75, and 300 mg/kg in mice. Though no dose-specific
66
-------
data were provided in the report, the authors stated that the absolute and relative weights of
epididymides and testes were reduced in animals receiving nitrobenzene. In addition, sperm
motility was adversely affected, and the incidence of abnormal sperm was increased.
A number of experimental approaches have been used to determine the mechanism by
which nitrobenzene induces testicular toxicity. For example, Allenby et al. (1990) used in vitro
experimental protocols to investigate possible mechanisms for how nitrobenzene may affect
spermatogenesis. The effects of incubating Sertoli cell isolates or cocultures from Alpk:AP
(Wistar derived) rats with a range of concentrations of nitrobenzene or m-dinitrobenzene (the
latter compound being a well-characterized Sertoli cell toxicant serving as a positive control)
were investigated. A number of parameters were monitored, including the exfoliation of germ
cells, the secretion to the medium of lactate, pyruvate, inhibin (a gonadal glycoprotein hormone
that inhibits pituitary follicle-stimulating hormone secretion), and, in general, any apparent
changes in cellular morphology. Vacuolization of the Sertoli cells was observed in the presence
of 1 mM nitrobenzene, with lower concentrations of the compound stimulating the release of
lactate and pyruvate, indicators of cell damage. Similarly, the release of inhibin was enhanced in
the presence of low concentrations of nitrobenzene, allowing the conclusion that the compound
is a Sertoli cell toxicant, though less effective than m-dinitrobenzene. The same scientists
(Allenby et al., 1991) also compared the ability of nitrobenzene and m-dinitrobenzene to induce
inhibin release from seminiferous tubule cultures obtained from rats of the Sprague-Dawley -
derived strain or Sertoli cell cultures obtained from AlpK: APpSD (Wistar derived) rats. Adult
Sprague-Dawley rats (approximately 70 days old) were used for in vivo experiments.
Nitrobenzene and m-dinitrobenzene caused a statistically significant increase in the release of
inhibin from isolated seminiferous tubules and, more variably, from isolated Sertoli cells. When
animals were administered a single dose of either nitrobenzene (300 mg/kg), m-dinitrobenzene
(25 mg/kg), or methoxyacetic acid (650 mg/kg), levels of inhibin were detectable in the testicular
interstitial fluid 1 to 3 days postexposure, although a statistically significant decrease in testicular
weight was not apparent until 3 days, suggesting that inhibin release may serve as an early
indicator of impairment of spermatogenesis.
Shinoda et al. (1998) used terminal deoxynucleotidyl transferase-mediated deoxyuridine
triphosphate nick end-labeling and deoxyribonucleic acid (DNA) gel electrophoresis to
investigate the extent to which germ cell degeneration represented necrosis or apoptosis. The in-
life phase of the experiment featured a single oral dose of 250 mg/kg nitrobenzene to male
Sprague-Dawley rats, the subjects being terminated at various time points up to 7 days
posttreatment. Germ cell degeneration was evident as early as 24 hours after dosing, and
electron micrographs showed spermatocytes undergoing changes thought to be characteristic of
apoptosis. Degenerating spermatocytes contained fragmented DNA. Linking their data to those
of Allenby et al. (1991, 1990), Shinoda et al. (1998) speculated that nitrobenzene exposure could
alter secretion of one or more Sertoli cell factors that might trigger germ cell apoptosis.
67
-------
Richburg and Nafiez (2003) studied molecular mechanisms of nitrobenzene-induced
testicular toxicity via the Fas/Apo-l/CD95 and Fas ligand (FasL) signaling system, in which
FasL activates Fas. Following the engagement of FasL with Fas, an intrinsic apoptotic program
is initiated in the target cell. In testis, Sertoli cells express FasL and select germ cells express
Fas. This is a paracrine signaling system12 by which Sertoli cells can initiate killing of Fas-
expressing germ cells (Richburg and Boekelheide, 1996). Two mouse spontaneous mutations,
Ipr and gld, are loss-of-function mutations of Fas and FasL, respectively (Takahashi et al., 1994;
Watanabe-Fukunaga et al., 1992). In the study by Richburg and Nafiez (2003), similar mice
(CBA/K\1ms-Tnfrsf6Jpr-cg [lprcg] and B6.SMNC3H-FasgW'gW \gld\) were utilized to determine
the role of Fas and FasL at initiating germ cell apoptosis at 0, 6, 12, and 24 hours following a
challenge with nitrobenzene (8-week-old mice, 800 mg/kg; 4-week-old mice, 600 mg/kg). The
authors found that lprcg and gld mice still displayed nitrobenzene-induced apoptosis of germ cells
and concluded that nitrobenzene-induced germ cell apoptosis was not mediated by the Fas and
FasL system but more likely by an autocrine pathway within the germ cells.
Kawaguchi et al. (2004) investigated differences in fertility and sperm motion in male
rats treated with a-chlorohydrin, known to produce spermatotoxicity, and nitrobenzene, known
to produce testicular toxicity. Ten-week-old male Crj :CD(SD) IGS rats were treated with either
saline solution or 60 mg/kg-day nitrobenzene by gavage for 3 or 18 days. Male rats were mated
with 8-week-old female rats, same strain, on day 3 and days 14-17. In the 18-day treated group,
but not the 3-day group, a statistically significant decrease in absolute and relative weights of
both testes and epididymides was observed. No histopathologic lesions were observed in the
3-day group; however, in the 18-day group, nitrobenzene caused severe atrophy of the
seminiferous tubules, along with decreased concentrations of sperm and prominent cellular
debris in the tubular lumina of the caput/corpus and cauda epididymidis. A statistically
significant increase in the number of detached sperm heads was observed in the cauda
epididymis of 18-day treated animals. The movement of sperm in the 18-day nitrobenzene group
was less vigorous than at other time periods and was attributed to the marked decrease of
spermatogenesis in the testes. The fertility index was not affected by nitrobenzene treatment.
The authors concluded that the full adverse effect on male fertility (viz., complete absence of
sperm in the cauda epididymis) could be detected only after a full spermatogenic cycle (i.e., 21-
28 days after treatment).
4.3.2. Inhalation Exposure
Tyl et al. (1987) exposed 26 pregnant female Sprague-Dawley rats/group to gaseous
nitrobenzene at 0, 1, 10, or 40 ppm, 6 hours/day on gestation days (GDs) 6-15. Clinical signs
were monitored daily, and maternal body weights were recorded on GDs 0, 6, 9, 12, 15, 18,
12 Paracrine signaling involves communication between cell "A" releasing a signal and nearby cell "B" receiving
the signal. Autocrine signaling involves the release of a signal by cell "A" that is received within cell "A."
68
-------
and 21. All dams were terminated on GD 21 and subjected to a gross necropsy. The range of
evaluated maternal and fetal reproductive and developmental parameters included the numbers
of corpora lutea, maternal liver and uterine weights, the numbers of live and dead fetuses, the
numbers of resorption sites, fetal weights, and sex distribution, the incidence of fetal
malformations, and visceral and skeletal abnormalities.
The results showed that there were no compound-related clinical signs, although maternal
body weight gain was reduced by 19% in the high-dose group compared with controls between
GDs 6 and 15. However, this parameter had returned to control values by GD 21. Spleen
weights increased dose dependently from 0.60 g in controls to 0.84 g in 40 ppm dams, achieving
statistical significance in the 10 and 40 ppm dose groups. Gestational parameters, such as the
numbers of corpora lutea, resorptions and dead fetuses, live fetuses per litter, the pre- or
postimplantation loss rates (as a percent), sex ratio, or fetal body weights, were all unaffected by
treatment. Similarly, there were no indications of concentration-dependent developmental
toxicity or teratogenicity. The incidence of skeletal variations also did not indicate fetal toxicity.
The single exception was a statistically significant increase in the incidence of parietal skull
plates with an area of nonossification in the 40 ppm group, as shown in Table 4-32. However, it
is unclear whether this isolated effect represents a teratogenic effect of nitrobenzene or whether it
is a consequence of maternal toxicity observed in the high-concentration group. In general, the
reproductive and developmental toxicity effects of nitrobenzene on Sprague-Dawley rats
appeared to be mild, at least to the extent of their effects on female reproductive physiology.
Table 4-32. Incidence of skeletal variations in Sprague-Dawley fetuses
exposed via inhalation to nitrobenzene in utero
Nitrobenzene
concentration
(ppm)
0
1
10
40
Incidence by fetus and litter
Parietal skull plates
(non-ossification)3
9/167 (f) 8/25 (1)
15/172 (f) 9/25 (1)
2 1/174 (f) 11/25(1)
29/181 (f) 19/26 (l)b
Bilobed thoracic
centrum 9a
6/167 (f) 6/25 (1)
3/172 (f) 3/25 (1)
3/174 (f) 3/25 (1)
1/181 (f) 1/26 (l)b
Split anterior arch of
atlas3
1/167 (f) 1/25 (1)
7/172 (f) 7/25 (l)b
5/174 (f) 5/25 (1)
6/181 (f) 5/26 (1)
Poorly ossified
premaxillary"
3/167 (f) 3/25 (1)
19/172 (f) ll/25(l)b
13/174 (f) 7/25 (1)
12/181 (f) 6/26 (1)
a(f) = incidence among all fetuses of one dose group; (1) = litters affected per all litters of one dose group.
bp < 0.05.
Source: Tyl etal. (1987).
Dodd et al. (1987) carried out a two-generation reproductive/developmental toxicity
study on nitrobenzene in which, initially, 30 Sprague-Dawley rats/sex/group were exposed to 0,
1, 10, or 40 ppm nitrobenzene, 6 hours/day, 5 days/week for 10 weeks via inhalation prior to a
mating period of up to 2 weeks. This study also has appeared as a Toxic Substances Control Act
69
-------
Test Submission (Bushy Run Research Center [BRRC], 1985). After mating, the FO males were
sacrificed, while the pregnant females were exposed to nitrobenzene through GD 19 and again
after delivery on postnatal days (PNDs) 5-20 at which point the pups were weaned. The
FO females were sacrificed prior to necropsy on PND 21. On this day, 30 pups/sex/group
(Fi generation) were selected (one male and one female from each litter, where possible) and
allowed a 2-week growth period during which no nitrobenzene was administered. Subsequently,
a repeat of the F0 exposure and treatment protocol was undertaken, with the exception that, after
mating, some FI males from the 40 ppm nitrobenzene group were not sacrificed. These males
were allowed to enter a recovery phase, and after 9 weeks of nonexposure they were mated with
virgin, unexposed Sprague-Dawley females to examine potential reversibility of effects on the
male gonads. The results of this mating and all associated reproductive and developmental
parameters of this offspring and the F2 progeny were noted, as described below. During the in-
life phase of the study, clinical signs of all rats were observed daily, while body weights were
recorded weekly. After parturition, litters were examined for the numbers of pups, their sexes,
the numbers of stillbirths and live births, the appearance of external abnormalities, and all
incidences of toxicity and/or mortality. Pup weights were noted on a litter basis on PND 0, then
individually on PNDs 4, 7, 14, and 21. The 30 animals/sex/group that were entered into the FI
mating study were weighed weekly. FI males selected for the recovery phase and subsequent
mating were weighed every 2 weeks. At termination, all animals were subjected to a full
necropsy, and the weights of putative target organs, such as the testis and epididymis, were
recorded. Tissues preserved for histopathologic examination from the 40 ppm and control
animals included the vagina, uterus, ovaries, testis, epididymides, seminiferous tubules, prostate,
and all tissues with gross lesions. Sections of the testis were examined in males exposed at all
concentration levels.
As indicated in Table 4-33, there were marked reductions in the fertility indices as a
result of matings among the 40 ppm animals compared with controls. Most notably, this
reduction was also apparent in the matings that involved unexposed females with those high-
concentration FI males that had been allowed a 9-week period of recovery. In all matings that
resulted in live offspring, gestational parameters, such as the number of uterine implantations,
resorptions, and postimplantation losses, were unaffected by nitrobenzene in either generation.
However, marked spermatocyte degeneration and atrophy of the seminiferous tubules were
observed in both generations of high-concentration males, including those that entered the
9-week recovery period. Morphologically, the lesions were characterized by severe multifocal
and diffuse atrophy of the seminiferous tubules in 14/30 animals in the 40 ppm group and by
the appearance of giant syncytial spermatocytes in the seminiferous tubules of 22/30 subjects of
the FO generation. Giant syncytial spermatocytes were much less evident in FI males (1/30),
and the active stages of spermatocyte degeneration in the seminiferous tubules were less
frequent. However, the epididymides of 40 ppm males in the F0 and FI generations displayed
70
-------
degenerative spermatocytes and a reduced number of spermatids. By contrast, there were no
apparent lesions in the histopathology of the female reproductive organs at this concentration.
Table 4-33. Fertility indices for the FO, Fl, and recovery generations:
number of pregnancies per number of females mated
Groups
Fo
Fi
Fj /recovery
Fertility index
Exposure groups (ppm)
0
30/30
30/30
29/30
1
27/30
27/30
NDb
10
29/30
26/30
NDb
40
16/303
3/30a
14/303
ap < 0.01 compared with control.
bND = not determined.
Source: Doddetal. (1987).
Dodd et al. (1987) considered the histopathologic lesions to be less striking in the
Fl males of the recovery group compared with other high-concentration males and correlated
this finding with the higher fertility index in their matings compared with those of the regular
Fl males. From the authors' data, a NOAEL of 10 ppm for the reproductive and fertility
effects of nitrobenzene in Sprague-Dawley rats was suggested.
Biodynamics Inc. (1983) carried out a reproductive/developmental study in which
12 pregnant female New Zealand white rabbits were exposed to nitrobenzene at 0, 10, 40, or
80 ppm, 6 hours/day on GDs 7-19. All dams were terminated on GD 20. The weights of livers
and kidneys of all subjects were recorded, and fertility data, such as the number of corpora
lutea, live and dead fetuses, late or early resorptions, and implantation sites, were monitored.
There were no maternal effects of nitrobenzene, including dose-related changes in body weight
or observable clinical signs. The absolute and relative weights of kidneys were similar among
all groups, while any increases in liver weights were not statistically significant. One of the
few findings of any toxicological importance in the study was the statistically significant
increase in the concentration of metHb on GDs 13 and 19, a well recognized effect of
nitrobenzene. However, the study did not indicate any nitrobenzene-related changes in any of
the fertility parameters measured.
Biodynamics Inc. (1984) carried out a follow-up study in which 22 pregnant female New
Zealand white rabbits were exposed to nitrobenzene concentrations of 0, 10, 40, and 100 ppm,
6 hours/day on days 7-19 of gestation. All surviving dams were sacrificed on GD 30, and, as in
the range-finding experiment of Biodynamics Inc. (1983), the suite of reproductive and
developmental toxicity parameters evaluated included such fertility data as the numbers of
corpora lutea, implantation sites, resorptions, and live fetuses. However, in this experiment,
71
-------
recovered fetuses were given a gross external examination, and all were evaluated for either soft
tissue malformations or skeletal malformations and variations. Maternal toxicity was evidenced
by some upward fluctuations in relative liver weight (to about 12%) and 40 and 60% increases in
mean metHb levels in the 40 and 100 ppm groups, respectively. However, the only evidence of
any reproductive or developmental toxicity effects was in the slightly higher incidence of
resorptions in high-concentration dams (11 litters with resorptions versus 7 in controls). These
high-dose resorption data were stated to be at or near the historical value observed in New
Zealand white rabbits for this testing laboratory. No teratological effects of nitrobenzene were
observed.
BRRC (1984) carried out a reproductive, developmental, and toxicological study of the
effects of inhaled nitrobenzene in 26 pregnant female CD rats/group. Exposure to nitrobenzene
vapor was at nominal concentrations of 0, 1, 10, or 40 ppm, 6 hours/day on GDs 6-15. All dams
were sacrificed on GD 21. The weights of the liver, kidney, spleen, and uterus of all subjects
were recorded, and fertility data, such as the numbers of corpora lutea, live and dead fetuses, late
or early resorptions, and implantation sites, were monitored. Recovered fetuses were given a
gross external examination, and all were evaluated for either soft tissue malformations or skeletal
malformation and variations. The authors reported some evidence of maternal toxicity, including
transient fluctuations in body weight and elevated absolute and relative spleen weights in mid-
and high-dose dams. However, all reproductive, developmental, and teratological parameters
were unaffected by treatment.
A synopsis of developmental toxicity studies with nitrobenzene following inhalation
exposure is presented in Table 4-34.
Table 4-34. Summary of effects observed in developmental inhalation studies
with nitrobenzene
Species,
strain
Rat,
S-DC
Rat,
S-D
Rabbit,
New
Zealand
Number
26
pregnant
30/sex
two-
generation
12 (22)
pregnant
Dosing
0, 1, 10, 40 ppm,
f. \\riA nn« f, i s
sacrifice onGD 21
0, 1, 10, 40 ppm,
10 wk before
mating & through
mating, gestation
0, 10, 40, 80 (100)
ppm, 6 hr/d, GDs
7-19, sacrifice on
GD 20 (30)
Effect3
Fertility |
Skull non-ossification f
Testicular pathology t
Fertility |
Developmental toxicity
Fertility |
Developmental toxicity
NOAELb'c
(ppm)
40
10
10 (M, FO
10 (M, FO
40
80 (100)
80(100)
LOAELbc
(ppm)
NA
40
40 (M, FO
40 (M, FO
NA
NA
NA
Reference
Tyl et al.
(1987)
BRRC
(1985); Dodd
etal. (1987)
Biodynamics
Inc. (1984,
1983)
aOnly endpoints with evident dose responses were selected. J, or f = a decrease or increase in the respective
endpoint.
bNOAELs and LOAELs determined by nitrobenzene assessment authors.
°M = male; F = female; F! = first filial generation; S-D = Sprague-Dawley; NA = not applicable.
72
-------
4.4. OTHER STUDIES
4.4.1. Acute and Short-Term Toxicity Data
DuPont (1981) reported a short-term inhalation study in which 16 male Crl:CD rats/group
were restrained and exposed (head only) 6 hours/day, 5 days/week for 2 weeks to either 0, 12,
39, or 112 ppm nitrobenzene. A subset of the exposed animals was terminated directly at the
completion of dosing (10 exposures), whereas others were allowed to recover for 14 days after
treatment. Blood was obtained from the tail vein on the day of the final exposure and at the end
of the recovery period. A wide range of hematologic parameters was monitored, along with such
clinical chemistry parameters as the activities of alkaline phosphatase, glutamate pyruvate
transaminase, and AST and the concentrations of BUN, creatinine, total protein, and cholesterol.
Depending on the dose level, a number of the animals displayed clinical signs of exposure to
nitrobenzene. Signs were severe, reflecting a degree of toxicity that led to death among animals
of the high-concentration group. For example, rats in the mid- and high-concentration groups
were cyanotic, and, from day 7 onward, high-concentration males appeared semi-prostrated
during exposure, with labored breathing, hind-limb ataxia, and reduction in body weight. In fact,
after the scheduled 10 total exposures, the high-concentration group was reduced to three
survivors, of which only one survived through the recovery period. Among the hematologic
responses, the mid- and high-concentration animals displayed statistically significant reductions
in Hb concentration and RBC count, while the platelet count, MCV, and MCHb were increased.
MetHb was markedly and dose-dependently higher in nitrobenzene-receiving rats versus
controls, with mean percentage values of 0.86, 1.7, 4.1, and 18.1 for rats exposed to 0, 12, 39,
and 112 ppm, respectively. Urinalysis indicated a decrease in osmolarity, but there was a
treatment-related increase in urine volume and urobilinogen concentration, a breakdown product
of Hb. After the 14-day recovery period, many of these symptoms were found to persist.
Among the histopathologic responses, there was a dose-dependent increase in the deposition of
hemosiderin in the spleen of mid- and high-concentration animals. High-dose rats displayed
hemorrhage of the brain plus lesions of the spinal cord, atrophy of the germinal cells, a range of
histopathologic effects in the testis and epididymis, pulmonary edema, and lymphoid cell
atrophy. In evaluating their data, the authors noted a trend toward increases in the organ/body
weight ratios for such organs as spleen, liver, kidney, and heart, though they considered these
changes to be unrelated to the toxic effects of nitrobenzene. By contrast, there were large
reductions in the testis and epididymis weights that appeared to be related to treatment and that
persisted in those animals allowed to undergo a period of recovery.
Sprague-Dawley (CD) rats and B6C3F1 mice were more sensitive to the effects of
inhaled nitrobenzene than F344 rats in a 2-week exposure study reported by Medinsky and Irons
(1985). Ten rats and mice of both sexes were exposed to concentrations of 0, 10, 35, or 125 ppm
nitrobenzene, 6 hours/day, 5 days/week for 2 weeks. Five animals of each species, strain, and
sex were sacrificed at 3 and 14 days after the last exposure, though many of the B6C3F1 mice
73
-------
and Sprague-Dawley rats in the high-exposure groups either died or were moribund prior to the
end of the exposure period. A total of 24 organs and tissues were examined for signs of gross
lesions, and the spleen, left kidney, liver, testes, and brain were weighed. Hematologic
parameters and clinical chemistry measurements were also evaluated.
In the 125 ppm group, it was necessary to sacrifice all mice of both sexes after 2-4 days
of exposure, and all Sprague-Dawley rats were sacrificed after 5 days of exposure. By contrast,
all F344 rats of both sexes survived the full 2-week exposure period with minimal signs of
distress. Concentration-dependent increases in relative liver, kidney, and spleen weights were
observed in both sexes of F344 rats, and increased relative spleen weights were observed in
Sprague-Dawley rats. Statistically significant increases in relative liver and kidney weights in
F344 male rats and relative spleen weights in Sprague-Dawley rats were observed even in the
low-concentration (10 ppm) groups. Decreased testis weights were observed in the high-
concentration (125 ppm) F344 rats, a response that persisted throughout the 14-day recovery
period. The cause of death in the high-concentration Sprague-Dawley rats was presumably due
to perivascular hemorrhage, accompanied by edema and malacia in the cerebellar peduncle.
Similar lesions were found in high-concentration group B6C3F1 mice. Histopathologic lesions
were observed in the brain, liver, kidney, lung, and spleen of Sprague-Dawley rats and B6C3F1
mice exposed to nitrobenzene. As tabulated by Medinsky and Irons (1985), these lesions
included, in the brain, cerebellar perivascular hemorrhage; in the liver, centrilobular necrosis,
centrilobular hydropic degeneration, and necrosis of hepatocytes; in the lung, bronchial epithelial
hyperplasia, vascular congestion, and perivascular edema; in the kidney, hydropic degeneration
of cortical tubular cells; in the testis, testicular degeneration, dysspermiogenesis, and the
appearance of multinucleated giant cells; and in the spleen, acute congestion, extramedullary
hyperplasia, and the appearance of hemosiderin-laden macrophages in red pulp. Histopathologic
lesions observed in F344 rat tissues as a result of exposure to 125 ppm nitrobenzene included, in
the spleen, acute congestion, extramedullary hyperplasia, focal capsular hyperplasia, and the
appearance of hemosiderin-laden macrophages in red pulp; in testis, edema, increased numbers
of multinucleated giant cells, Sertoli cell hyperplasia, and severe dysspermiogenesis; and, in the
kidney, a hyaline nephrosis that was especially marked in male rats. Testicular degeneration was
observed in the high-concentration mice and in one animal in the 35 ppm concentration group.
The most sensitive organ, based on the histopathology findings, was the spleen. Lesions in the
spleen were observed in all animals of all concentration groups. In F344 rats, there was a
concentration-dependent increase in the number of hemosiderin-laden macrophages infiltrating
the red pulp, increased extramedullary hematopoiesis, and acute sinusoidal congestion 3 days
after the last exposure. Similar lesions were observed in Sprague-Dawley rats and B6C3F1
mice. A concentration-dependent increase in blood metHb was noted in F344 rats 3 days after
the end of exposure, but this effect was not observed after 14 days (Table 4-35). Blood metHb
ranged from 13-31% in B6C3F1 mice that were sacrificed early.
74
-------
Table 4-35. Percent metHb in blood of rats exposed to nitrobenzene vapors
Group
F344 rats
Male
Female
Sprague-Dawley rats
Male
Female
Sacrifice at term + 3 days
Control
lOppm
35 ppm
125 ppm
0
1.9 ±0.7
6.6 ±0.2
11.7 ±1.2
3.6 ±2.2
4.8 ±0.8
6.6 ±0.8
13.4±2.1
6.9 ±1.3
6.1 ±0.5
8.7 ±1.0
14.0 ±1.3
4.8 ±0.7
6.3 ±0.6
7.3 ±1.4
31.3±2.5a
Sacrifice after recovery period
Control
10 ppm
35 ppm
125 ppm
4.5 ±0.3
4.1±0.1
5.6 ±2.2
4.8 ±1.9
4.1 ±0.5
3.1±0.3
5.1 ±1.9
4.5 ± 1.5
4.6 ±0.3
9.2 ± 1.6
5.8 ±0.9
b
5.6 ±0.6
5.2 ± 1.0
5.0 ±0.5
b
aRats were euthanized after 5 days of exposure.
bNo high-concentration rats survived in this group.
Note: Statistical significance was not provided by the authors.
Source: Medinsky and Irons (1985).
A synopsis of the acute inhalation studies with nitrobenzene is presented in Table 4-36.
Table 4-36. Summary of effects observed in short-term inhalation studies
with nitrobenzene
Species,
strain
Rat,
Crl:CD
Rat,
S-D
Rat,
F344
Mouse,
B6C3F1
Number
16 male
10/sex
Dosing
0, 12, 39, 112 ppm,
6 hr/d, 5 d/wk, 2 wk
0, 10, 35, 125 ppm,
6 hr/d, 5 d/wk, 2 wk
Effect3
Methemoglobinemia
Mortality
Spleen weight t
Mortality
Organ weights t
Testis weight J,
Testicular pathology
NOAELb'c
(ppm)
NA
39
NA
35
NA
35
35
LOAELbc
(ppm)
12
112
10
125
10
125
125
Reference
DuPont(1981)
Medinsky and
Irons (1985)
aOnly endpoints with evident dose responses were selected. J, or f = decrease or increase in the respective endpoint.
bNOAELs and LOAELs determined by nitrobenzene assessment authors.
°NA = not applicable.
Few data are available for the oral median lethal dose for nitrobenzene, although Lewis
(1992) reported a value of 590 mg/kg in mice. NLM (2003) gives values of 600-640 mg/kg
nitrobenzene in rats. DuPont (1981) reported a 4-hour median lethal concentration of 556 ppm
in male Sprague-Dawley rats exposed (head only) to nitrobenzene vapor.
A number of research reports describe the use of acute or short-term exposure regimens
to examine sublethal toxicological effects of nitrobenzene. Those addressing the absorption,
75
-------
distribution, metabolism, and excretion of the compound and its metabolites have been described
in section 3. Other toxicological responses of experimental animals to short-term nitrobenzene
exposure are described in the following paragraphs.
As discussed in section 4.3.1, the single oral dose experiments of Bond et al. (1981)
resulted in histopathologic lesions in liver, testes, and brain and in the immediate development
and subsequent slow decline of methemoglobinemia in male F344 rats at a dose of 300 mg/kg.
Morgan et al. (1985) extended the observations of Bond et al. (1981) on the histopathologic
effects of nitrobenzene on the brain by a light and electron microscopic study of male F344 rats
receiving single oral doses of 550 mg/kg [14C]-labeled nitrobenzene. Administration of
nitrobenzene induced petechial hemorrhages in the brain stem and cerebellum and bilateral
symmetric degeneration (malacia) in the cerebellum and cerebellar peduncle. Ultrastructural
studies suggested that edematous swelling of a membrane-bounded tissue compartment in the
region of the vestibular nuclei and other nuclei lying near the lateral margins of the fourth
ventricle were responsible for the malacia. Hemorrhages were found throughout the brain stem,
but there was little evidence of vascular degeneration, and no ultrastructural abnormalities were
found in the blood vessel walls. Heinz bodies were observed in the erythrocytes in the
hemorrhages, consistent with induction of metHb by nitrobenzene. However, it could not be
established whether tissue anoxia due to metHb formation could have contributed to the
neurotoxicity of nitrobenzene. Whole body autoradiography indicated that only a small portion
of the administered nitrobenzene dose actually penetrated the blood-brain barrier. Radiotracer
studies indicated that approximately 0.02% of the total nitrobenzene dose was present in the
cerebellum 12 hours after administration. However, no nitrobenzene metabolites could be
detected, and the mechanism of nitrobenzene neurotoxicity could not be determined from these
studies. Though, quantitatively, the brain appeared not to be a primary target organ of
nitrobenzene deposition, a range of marked histopathologic effects of nitrobenzene was
identified, including bilateral symmetrical degeneration of the cerebellum and instances of
neuronal degeneration.
NTP sponsored a 14-day skin painting toxicological study (NTP, 1983b) with
nitrobenzene in F344 rats and B6C3F1 mice. In the study, dose levels ranged from 200-
3200 mg/kg, the higher doses (1600 and 3200 mg/kg) inducing death or morbidity before the end
of the experiment. Among surviving animals, statistically significantly reduced weight gain
(>10%) was observed in all but the low-dose groups. Reticulocyte counts and metHb
concentrations were increased significantly, most conspicuously in mice where these effects
were seen in the low-dose males. RBCs and Hb concentrations were reduced. Histopathologic
changes were evident in brain, liver, spleen, and testis.
Shimkin (1939) demonstrated the ability of nitrobenzene to penetrate the skin and induce
toxic effects in female C3H and male A strain mice. In these experiments, nitrobenzene was
brushed onto the shaved abdomen of C3H mice, covering less than one-tenth of the body surface.
76
-------
Because of the method of application, the applied dose was unknown. Treatment-related clinical
signs, morbidity and mortality, along with associated evidence of incipient methemoglobinemia
and other hematologic perturbations, were observed. One hour after application, 15/18 female
C3H mice were in partial collapse, but all recovered within 24 hours. After a second application,
three animals died, and after a third application nine more animals died. Approximately
30 minutes after vigorously brushing nitrobenzene over the unshaved abdomens of 10 male
A strain mice for 20 seconds, all the mice were in partial collapse and 8/10 died within 3 days.
One to 3 hours after application, the skin became dark gray-blue, the blood became chocolate
colored and viscous, and the urine was orange with an odor of nitrobenzene. Spectrographic
analysis of blood showed a strong absorption band characteristic of metHb. The hematologic
data in Shimkin's report emphasized the variability of the cell counts with a normal differential
count but a greater than 50% reduction in WBCs (5,000 cells/mm3, reduced from 11,000-
14,000 cells/mm3 in controls). However, while RBC numbers were unaffected, smears indicated
hypochromia and hemolysis. Among the necropsy findings, the most susceptible target organ
was the liver, which demonstrated diffuse necrosis, especially in the outer portions of the liver
lobules. There was a large amount of dark, brownish pigment in the Kupffer cells; the pigment
was more prominent in the necrotic portions of the lobules. Among secondary sites, the kidney
showed evidence of enlargement of the glomeruli and tubular epithelium. However, other
potential target organs, such as the spleen, lungs, and testis, displayed no morphologic changes.
4.4.2. Structure-Activity Relationships
Nitroaromatic compounds related to nitrobenzene include four structurally similar
compounds that vary based on the number and position of the nitro group (Table 4-37). A large
body of toxicological information is available on 1,3-dinitrobenzene and 1,3,5-trinitrobenzene.
Toxicity data on these compounds in experimental animals have revealed a similar spectrum of
toxicological effects to those seen with nitrobenzene (e.g., metHb formation and splenomegaly)
(Tables 4-38 and 4-39) (Salice and Holdsworth, 2001).
For example, the male reproductive toxicity expressed by nitroaromatics is greatly
influenced by the structure of the compound. Of the three dinitrobenzene isomers listed in Table
4-37, only 1,3-dinitrobenzene, not 1,2-dinitrobenzene or 1,4-dinitrobenzene, is a potent testicular
toxicant that targets the Sertoli cell. However, 1,4-dinitrobenzene, but not 1,2-dinitrobenzene,
has a potency similar to that of 1,3-dinitrobenzene in producing cyanosis and splenic
enlargement in male Alpk/AP (Wistar derived) rats, indicating that different mechanisms are
probably responsible for these two toxic effects (Blackburn et al., 1988). Similarly, the
cerebellar neurotoxicity ascribed to 1,3-dinitrobenzene and 1,3,5-trinitrobenzene is not observed
in animals dosed with 1,4-dinitrobenzene (Chandra et al., 1999; Romero et al., 1995; Morgan et
al., 1985).
77
-------
Table 4-37. Overview of properties and toxicities of nitrobenzenes
IUPACa name
1,2-
Dinitrobenzeneb
1,3-
Dinitrobenzenee'f
1,4-
Dinitrobenzene8
1,3,5-
Trinitrobenzene11'1
CASRN
528-29-0
99-65-0
100-25-4
99-65-0
Chemical
formula
C6H4N204
C6H4N204
C6H4N204
C6H3N306
Structural formula
I* I
N 1
° '.; j
i
r i
i
° ; °
1
r i
' T
O O
° , °
M
1
0 . ---- -:•--- , °
1
0 0
LOAEL
No data
Drinking
water:
8 ppm
No data
Dietary
study:
13.31
mg/kg-
day
NOAEL
TLV0:
0.15 ppm
(as TWAd)
(skin)
Drinking
water:
3 ppm
(0.40
mg/kg-
day)
TLV:
0.15 ppm
(as TWA)
(skin)
Dietary
study: 2.68
mg/kg-day
Critical effect
Liver impairment,
metliemoglobinemia,
anemia
Increased splenic
weight
Liver impairment,
metliemoglobinemia,
anemia
Metliemoglobinemia
and spleen-erythroid
cell hyperplasia
aIUPAC = International Union for Pure and Applied Chemistry.
bhttp://pubchem.ncbi.nlm.nih.gov/sunimary/summary.cgi?sid=36593;
http://www.incheni.org/docunients/icsc/icsc/eics0460.htm; http://www.epa.gov/iris/subst/0633.htm.
CTLV = threshold limit value.
dTWA = time-weighted average.
ehttp://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?sid=48779; http://www.epa.gov/iris/subst/0318.htm; Cody
etal. (1981).
Conversion factors: drinking water concentrations converted to dosages by Cody et al. (1981).
8http://pubchem.ncbi.nlm.nih.gov/sunimary/summary.cgi?sid=15738;
http://www.inchem.org/documents/icsc/icsc/eics0692.htni.
hhttp://pubchem.ncbi.nlm.nih.gov/sunimary/summary.cgi?sid=48734; http://www.epa.gov/iris/subst/0316.htni;
Reddy etal. (1996).
'Conversion factors and assumptions: based on food consumption data, Reddy et al. (1996) calculated the intake of
trinitrobenzene from dietary concentrations of 0, 5, 60, and 300 ppm as 0, 0.23, 2.68, and 13.31 mg/kg-day
(females) and 0, 0.22, 2.64, and 13.44 mg/kg-day (males).
78
-------
Table 4-38. Summary of toxicological studies with 1,3-dinitrobenzene
Study
Linder et al.
(1986);
Perreault et
al. (1989)
Philbert et al.
(1987)
Reddy et al.
(1994a)
Reddy et al.
(1994b)
Cody et al.
(1981)
Cody et al.
(1981)
Species
(strain)
Rat
(Sprague-
Dawley)
Rat
(F344)
Rat
(F344)
Rat
(F344)
Rat
(Carworth
Farms)
Rat
(Carworth
Farms)
Test duration
12 weeks
5 days
90 days
14 days
8 weeks
16 weeks
NOAEL
(mg/kg-day)a
0.54
NA
NA
0.07
0.39
0.21
0.8
1.98
NA
0.48
1.13
LOAEL
(mg/kg-day)
1.1
0.54
20
0.35
1.73
0.8
1.98
5.77
4.72
1.32
2.64
Effects observed
at the LOAEL
Reduced spermatid head count
Large reduction in reproductive
performance (pups/litter)
Ataxia in all male rats
Methemoglobinemia and an
increase in reticulocytes
Reduction in RBCs and in other
hematological responses;
changes in spleen and testicular
histopathology
Methemoglobinemia
Splenomegaly
Nephropathy associated with
hyaline droplet formation;
testicular degeneration
Splenomegaly, fluctuation in Hb
levels, atrophy and
histopathologic lesions of the
testes
Splenomegaly
Depleted spermatogenesis
aNA = not applicable.
79
-------
Table 4-39. Summary of toxicological studies with 1,3,5-trinitrobenzene
Study
Reddy et al.
(2001,
1996)
Reddy et al.
(1998,
1994a)
Reddy et al.
(1994b)
Kinkead et
al. (1995,
1994a)
Kim et al.
(1997)
Narayanan
etal. (1995)
Kinkead et
al. (1994b)
Chandra et
al. (1995a)
Chandra et
al. (1995b)
Chandra et
al. (1997)
Cooper and
Caldwell
(1995)b
Reddy et al.
(1995)
Reddy et al.
(2000)
Species
(strain)
Rat
(F344)
Rat
(F344)
Rat
(F344)
Rat
(Sprague-
Dawley)
Rat
(F344)
Rat
(Sprague-
Dawley)
Rat
(Sprague-
Dawley)
Rat
(F344)
Rat
(F344)
Rat
(F344)
Rat
(Sprague-
Dawley)
Mouse
(Peromyscus
leucopus)
Shrew
(Cryptotis
parva)
Test
duration
2 years
90 days
14 days
90 days
10, 20,
and 30
days
90 days
7 weeks
10 weeks
10 days
10 days
10 days
CDs 6-15
90 days
14 days
NOAEL
(mg/kg-day)a
2.68
NA
4.29
NA
4.52
2.0
NA
NA
NA
23
4
NA
35.5
NA
45
67.4
23.5
10.75
10.68
LOAEL
(mg/kg-day)
13.31
3.91
22.73
4.54
16.85
9.0
2.0
35.5
3.0
51
23
35.5
71
35.5
90
113.5
67.4
21.60
22.24
Effects observed at the LOAEL
Methemoglobinemia, spleen erythroid cell
hyperplasia, decreased body weight
Nephropathy, a2u-globulin-associated
hyaline droplet formation in males at all
doses
Methemoglobinemia, spleen erythroid cell
hyperplasia in high- and mid-dose groups
(males and females)
Reduced RBC count and Hct in all female
groups
Histopathologic changes to the kidney in
males
Sperm motility /seminiferous tubular
degeneration of the testes
Nephropathy, hyaline droplet formation in
males at all doses
Nephropathy, a2u-globulin-associated
hyaline droplet formation in males at all
doses tested
Increase in tissue concentrations of various
neurotransmitters in several brain regions,
potentially associated with neurological
disorders and histopathologic lesions
Testicular degeneration and sperm
depletion in males
Encephalitis in females
Hematologic deficits and metHb formation
Histopathologic lesions in the brain of
males
Testicular degeneration
Developmental deficits among the pups
Testicular degeneration in high-dose males
Erythroid hyperplasia, increase in
reticulocyte count in mid- and high-dose
males
Decrease in liver and body weight
Increase in spleen weight of females
aNA = not applicable
bAs cited in Reddy et al. (1997).
80
-------
4.4.3. Immunotoxicity Studies
Burns et al. (1994) carried out a 14-day gavage study of nitrobenzene in corn oil in which
female B6C3F1 mice were administered 0, 30, 100, and 300 mg/kg of the compound. The
primary focus of the study was the immunotoxicity of the compound, although some
characteristic responses of nitrobenzene's acute toxicity to B6C3F1 mice at these exposure levels
were reported. For example, 17 of 200 high-dose mice died during the period of exposure, and
others displayed typical signs of lexicologically stressed animals, such as ataxia, lethargy, and
circling. Eight distinct investigations of the immunotoxicological effects of nitrobenzene were
carried out among the exposed mice, while some nonimmunotoxicological parameters were
monitored in all animals.
Examination of the mice at autopsy 24 hours after the final exposure showed
hepatomegaly and splenomegaly in the mid- and high-dose groups, although the overall liver
changes were slight. The affected spleens were dark red in color, with mild congestion in the red
pulp areas and the appearance of occasional nucleated erythrocytes. Hemosiderin pigment was
noted in the red pulp areas, a response thought to be indicative of erythrocyte dysfunction.
However, white pulp areas of the spleen appeared to be normal. Compound-related changes in
organ weights were noted, including dose-dependent increases in the absolute and relative
weights of liver, spleen, and kidney. A number of apparently compound-related effects in
hematologic responses to nitrobenzene were observed, consistent with the concept of the
erythrocyte as a primary target organ of nitrobenzene toxicity. The changes included decreases
in erythrocyte number (7.64 ±0.15 x 106 cells/uL in controls versus 6.94 ± 0.14 x 106 cells/uL in
mice exposed to 300 mg/kg-day nitrobenzene) but increases in MCV (56 ± 1 fL in controls
versus 63.7 ± 1.4 fL in mice receiving 300 mg/kg-day) and MCHb (18.1 ± 0.3 pg in controls
versus 20.6 ± 0.6 pg in animals receiving 300 mg/kg). However, there were no treatment-related
changes in Hb concentration or Hct. Although no treatment-related differences in leukocyte
differentials were observed after 14 days, there were striking changes in the percentage of
circulating reticulocytes as a result of treatment (4.57 ± 0.48% in mice receiving 300 mg/kg
versus 1.03 ± 0.9% in controls). MetHb was not evaluated.
Burns et al. (1994) also observed some treatment-related changes in clinical chemistry
parameters, including a dose-dependent increase in the activity of AST (80 ± 9 ITJ/mL in controls
versus 128 ± 16 ITJ/mL in high-dose mice) and ALT (27 ± 1 ITJ/mL in controls versus 74 ± 11
ITJ/mL in high-dose animals). Other dose-dependent effects of nitrobenzene on clinical
chemistry parameters included apparent increases in the levels of bilirubin and albumin but
decreases in glucose concentration.
In light of the changes observed in the spleen and hematologic parameters, Burns et al.
(1994) examined the bone marrow for cell number, status of DNA synthesis, and the number of
macrophage and granulocyte-monocyte progenitor cells. DNA synthesis was measured by the
incorporation of [3H]-thymidine over a 3-hour incubation period. Progenitor cells were
81
-------
measured by incubating bone marrow cells with 10% colony stimulating factors isolated from
either mouse fibroblast L-929 cells or mouse lung-conditioned medium. Colonies were counted
after 8 days. The number of nucleated cells/femur was increased dose dependently to a level of
62% above controls, with statistical significance seen in the low-dose group. Overall rates of
DNA synthesis also were increased up to 80% above those of controls. As described by the
authors, the number of colony-forming unit (granulocyte-monocyte) stem cells was the same as
in controls when calculated per 105 bone marrow cells. However, the number of cells/femur and
the number of colony-forming unit (granulocyte-monocyte) stem cells/femur were increased
twofold in association with nitrobenzene treatment (Burns et al., 1994).
Burns et al. (1994) determined spleen immunoglobulins G and M (IgG and IgM)
antibody responses to T-dependent sheep RBCs in mice exposed to nitrobenzene by using a
modified hemolytic plaque assay. Animals receiving nitrobenzene were sensitized to sheep
RBCs by intravenous injection on day 11 of exposure, and spleen cells were harvested at term.
Suspended cells were incubated with guinea pig complement, sheep RBCs, and warm agar.
Rabbit anti-mouse IgG-developing serum was added when IgG plaques were evaluated, and cell
and plaque counts were obtained after a 3-hour incubation at 37°C.
Although there was a dose-dependent increase in spleen weight and spleen cell number
4 days after exposure to nitrobenzene, there was no difference in the splenic IgG responses to
sheep erythrocytes as a result of nitrobenzene exposure. By contrast, nitrobenzene exposure
caused a dose-dependent decrease in the IgM response to sheep erythrocytes on day 4 (40 and
34% for the mid- and high-dose nitrobenzene groups, respectively). According to the authors,
this suppression could be accounted for by the observed compound-induced splenomegaly
(Burns et al., 1994). However, treated mice recovered their ability to mount an IgM response
within 20 days.
The capacity of spleen cells to undergo a proliferative response to the T cell mitogens
(phytohemagglutinin [PHA], concanavalin A [con A], and the B cell mitogen, lipopolysaccharide
[LPS]) was investigated. Cells were isolated from excised spleen tissue after 15 days of
nitrobenzene exposure and cultured for 3 days in the presence of four concentrations of the
above mitogens. The amount of [3H]-thymidine incorporated into the cells over the last 18 hours
of the incubation was taken as a measure of spleen cell proliferation. The effects of nitrobenzene
on the response to PHA and con A appeared to be dose related, with a marked suppression of
[3H]-thymidine incorporation following exposure to 100 and 300 mg/kg nitrobenzene (106,152 ±
10,326 cpm/culture in control cultures of spleen cells incubated with 5 |ig/mL con A versus
59,602 ±5189 cpm/culture in cultures of spleen cells from high-dose mice incubated with the
same concentration of mitogen). However, there were no effects of nitrobenzene on the response
to the B cell mitogen, LPS.
The impact of nitrobenzene exposure on the onset of delayed hypersensitivity of keyhole
limpet hemocyanin (KLH) was assessed by administering a subcutaneous injection of 100 jig
82
-------
KLH on days 1 and 8 of nitrobenzene exposure. On the last day of nitrobenzene exposure,
mononuclear cells were labeled in vivo by intravenous injection of [125I]-5-iododeoxyuridine
(2 jiCi) per mouse. On day 15, animals were challenged in the central portion of the left ear with
an intradermal injection of 30 jig KLH, and ear biopsies were radioassayed 24 hours later. As
expressed by a stimulation index, no effect of nitrobenzene on a delayed hypersensitivity
response to KLH was observed. Similarly, in another sequence of observations, there were no
differences in serum complement levels between nitrobenzene-exposed and control groups.
Burns et al. (1994) investigated the comparative uptake and organ distribution of injected
radiolabeled sheep erythrocytes in control and nitrobenzene-exposed mice. Compared with the
vehicle control group, nitrobenzene-receiving mice (39.4 ± 1.8 in controls versus 55.3 ± 1.7 in
high-dose animals) showed a dose-dependent increase in particle uptake into the livers.
However, this effect was considered to be a consequence of liver enlargement in nitrobenzene-
receiving groups.
In other experimental approaches, Burns et al. (1994) monitored the number of cells that
could be harvested by lavage from the peritoneal cavity of nitrobenzene-challenged mice,
examined the ability of isolated macrophages to take up fluorescent beads (0.85 jam), and
determined the effect of nitrobenzene on natural killer cell activity in the spleen. In the latter
case, natural killer cell function was assessed by monitoring the capacity of spleen cells to lyse
[51Cr]-labeled YAC-1 target cells in vitro. Nitrobenzene exposure caused a dose-dependent
decrease in lytic activity at all effectortarget cell ratios tested.
The same research report describes a series of experiments to evaluate the effect of
nitrobenzene on host resistance to infection with Plasmodium berghei, Listeria monocytogenes,
Streptococcus pneumoniae, herpes simplex type 2 virus, and the metastatic pulmonary tumor,
B16F10. Mice treated with nitrobenzene were no more susceptible to S. pneumoniae or
P. berghei than were control animals. However, a challenge with 6 x 103 L. monocytogenes per
mouse killed 13% of the control mice and 57% of those receiving 300 mg/kg nitrobenzene.
Similar differences were observed for different liters in mice exposed to 100 mg/kg
nitrobenzene. As pointed out by the authors, host resistance of L. monocytogenes is mediated by
T lymphocytes, macrophages, and complement activity. Nitrobenzene exposure did not impair
host resistance to herpes simplex virus, as measured by percent mortality or time to death. Host
resistance to B16F10 melanoma involves T-lymphocytes and macrophages. Nitrobenzene
somewhat impaired host resistance at the highest level, indicating a modest depression of T-cell
immunity.
In seeking to explain their results, Burns et al. (1994) considered that most of the effects
of nitrobenzene on the immune system could be explained by the increased cellularity of the
spleen. However, the perturbation of the bone marrow in mice exposed to nitrobenzene was
pronounced, manifested in these studies by increases in cells/femur, DNA synthesis, and colony-
forming units (granulocytes/monocytes)/femur. These results were thought to indicate that the
83
-------
principal target of nitrobenzene toxicity was bone marrow, with consequent hematologic and
immunotoxicological impacts.
Wulferink et al. (2001) presented findings that nitrosobenzene (but not nitrobenzene,
aniline, or p-aminophenol) stimulated the production of antigen-specific T-cells in female
C57BL/6J mice. The study analyzed primary and secondary popliteal lymph node (PLN)
response, an assay that detects the immunostimulatory capacity of low molecular weight
substances. For the primary PLN response, animals received a single subcutaneous injection
(50 |jL) into the left hind footpad. After 6 days, the PLNs from the treated and untreated sides
were removed and cell numbers were counted. Cell counts from nitrobenzene-, aniline-, or
p-aminophenol-treated animals (0.2 |j,mol/mouse) were indistinguishable from controls;
however, nitrosobenzene caused a statistically significant increase in cell counts at 0.1 and
0.2 |j,mol/mouse. For the secondary PLN response, animals were primed with a single
subcutaneous injection (50 |jL) of aniline or nitrosobenzene. Thirteen weeks later (the time
period it takes for PLNs to return to normal size and cellularity), a second subcutaneous injection
(50 |jL) containing a suboptimal dose (a dose too low to stimulate a primary PLN response;
0.005 |j,mol/mouse of either aniline or nitrosobenzene) was administered to the same footpad.
After four days, the PLNs from the treated and untreated sides were removed. Cell counts from
animals primed with aniline and challenged with either aniline or nitrosobenzene were consistent
with controls. Similarly, the cell counts from animals primed with nitrosobenzene and
subsequently challenged with aniline were not statistically significantly different from controls.
In contrast, when animals were primed with nitrosobenzene and also challenged with
nitrosobenzene, a statistically significant increase in cellularity was observed compared with
controls. Hopkins et al. (2005) reported similar findings that dermal application of
nitrosobenzene (100 jiL; 0.02%, weight/volume, in 5% dimethyl sulfoxide [DMSO]) for three
consecutive days on the nape of the neck of female BALB/c mice caused a statistically
significant increase in lymph node cellularity and proliferation 5 days after the initial application.
4.4.4. Neurotoxicity Studies
Signs and symptoms of neurotoxicity following exposure to nitrobenzene have been
reported as early as the 1900s. No epidemiological studies have been conducted on
occupationally exposed cohorts; however, numerous case reports indicate neurological
involvement following accidental or intentional exposure to nitrobenzene. Abbinante et al.
(1997) identified dizziness, generalized weakness, and convulsions as the most frequent
neurological manifestations from nine individuals intoxicated with nitrobenzene (levels of
exposure unknown). Similarly, Stifel (1919) reported 16 cases of nitrobenzene poisoning from
shoe dye. Many of the patients complained of headache, nausea, dizziness, and general malaise.
In a more comprehensive report, Ikeda and Kita (1964) presented findings from a woman
who was occupationally exposed to nitrobenzene. Seventeen months after starting a new
84
-------
position, the woman's workplace was remodeled and the ventilation became quite poor. After
about 6 weeks of working under these conditions, the woman presented with severe headache,
nausea, vertigo, and numbness in her legs. After 5 days of bed rest, her condition improved and
she returned to work. Nearly 3 months later, the woman presented with similar symptoms. In
addition, she experienced hyperalgesia to pinprick on the backs of her hands and feet, which
suggested degenerative changes in the peripheral nerves. She was discharged after 39 days in the
hospital with only residual hyperalgesia in the hands and feet.
Adams (1912) (as cited in Hamilton [1919]) presented observations of a middle-aged
woman who was chronically exposed (18-year observation period) to nitrobenzene through its
use as an ingredient in cleaning fluid. The symptoms, which progressed very slowly, were those
of a multiple neuritis, which finally resulted in contractures and almost complete powerlessness.
Interestingly, 1,3-dinitrobenzene, a compound structurally similar to nitrobenzene, has been
reported to cause numbness in the distal portions of the limbs in humans (Lazerev and Levina,
1976, as cited in Philbert et al. [1987]).
Obvious shortcomings of the above studies are the lack of quantitative estimates for
exposure and effects and the fact that they are primarily anecdotal. However, similar
manifestations of toxicity have been reported in nitrobenzene poisonings of experimental
animals. Matsumaru and Yoshida (1959) treated male and female rabbits (strain not stated) with
nitrobenzene injection via the ear vein or by topical application to the skin of the back.
Neurotoxicity was manifest with paralysis of the limbs, elevated sensitivity, and general
convulsion. When acute in nature, intoxication was evident mainly as convulsion, whereas
chronic intoxication resulted in paralysis. Central nervous system effects were evident with an
enormous number of well-defined round vacuoles occurring in the medulla, which was more
marked in those animals in the high-dose intravenous group and those treated dermally for a
prolonged term (time period not stated) compared with those in a low-dose (and control) group
and treated for shorter term(s), respectively.
Bond et al. (1981) described a lesion consisting of a bilateral malacic area and reactive
gliosis in the cerebellar peduncles. However, this lesion was observed with only one rat
(F344[CDF/CrlBR]) 5 days following oral administration of 450 mg/kg nitrobenzene. Marked
methemoglobinemia was excluded as the precipitating factor, since administration of sodium
nitrite to rats for 3 days resulted in a prolonged methemoglobinemia of severity similar to that
produced by nitrobenzene but showed no evidence of toxicity to the brain.
Shimo et al. (1994) treated F344 rats with nitrobenzene at the doses of 0, 5, 25, and
125 mg/kg-day for 28 days via intragastric administration. Absolute brain weights of male rats
revealed an increasing trend (up to 4.5% above control) that became statistically significant in
the 25 mg/kg group, and absolute brain weights in female rats followed a similar trend that
resulted in statistically significant increases in the 125 mg/kg group. Histopathology revealed
moderate to severe spongiform changes and brown pigmentation in the perivascular region of the
85
-------
cerebellum in male and female rats treated with 125 mg/kg. Following a 14-day recovery period,
brain weights of treated animals (males and females) were consistent with those of controls;
however, moderate to severe spongiotic changes persisted in five of six male rats and four of six
female rats, whereas moderate brown pigmentation in the perivascular region was present in
three of three male rats and two of four female rats.
Morgan et al. (1985) administered a single oral dose (550 mg/kg) of nitrobenzene to male
F344 (CDF/CrlBR) rats. Within 24 hours after dosing, the rats were lethargic and ataxic but
responsive to external stimuli (tail pinch). By 36-48 hours, several rats displayed moderate to
severe ataxia and loss of righting reflex and no longer responded to external stimuli.
Microscopic analysis revealed variable numbers of small hemorrhages scattered throughout the
brain stem and cerebellum. Many neurons and areas adjacent to malacia, both lateral and dorsal
to the fourth ventricle, showed moderate to severe fine, foamy vacuolation of the perikarya and
nuclear condensation. The affected areas exhibited numerous vacuoles, some of which could be
identified as distended myelin sheaths of large axons. Swelling of myelin sheaths was also
observed in white matter tracts adjacent to areas of malacia.
Burns et al. (1994) treated female B6C3F1 mice with nitrobenzene at 0, 30, 100, or
300 mg/kg-day for 14 consecutive days. Neurotoxicity was manifest in the 300 mg/kg-day
group only with animals exhibiting marked ataxia, lethargy, and circling. One animal was
observed with bobbing head movements. Absolute brain weights for all treatment groups were
consistent with controls. Histopathologic changes in the liver of the high-dose group consisted
of very mild hydropic degeneration around focal central veins with elevated levels of serum
transaminases and bilirubin.
4.4.5. Genotoxicity Studies
The mutagenicity/genotoxicity of nitrobenzene has been addressed in a number of studies
using standard Ames test protocols. For example, in the multicenter survey of compounds that
was carried out for the U.S. National Institute of Environmental Health Sciences, nitrobenzene
was found to be negative for reverse mutation with or without 9000 x g liver microsomal
supernatant fraction (S9) in all of the Salmonella typhimurium tester strains that were used
(Haworth et al., 1983). Similarly, in a survey of nitroaromatic compounds that were evaluated
for mutageni city (without S9) in nine tester strains of S. typhimurium, nitrobenzene was negative
for reverse mutation at all concentrations in every strain tested (Vance and Levin, 1984).
Furthermore, several studies from different laboratories (Assmann et al., 1997; Dellarco and
Prival, 1989; Shimizu et al., 1983; Ho et al., 1981; Anderson and Styles, 1978; Chiu et al., 1978;
Garner and Nutman, 1977) have reported essentially similar findings for nitrobenzene in this
experimental system, irrespective of the presence of an Aroclor-1254-induced S9 liver
preparation or added flavin mononucleotide (Dellarco and Prival, 1989).
86
-------
In contrast with the studies recounted above, two studies by Suzuki et al. (1987, 1983)
reported positive findings for a mutagenic action of nitrobenzene in the Ames test with the tester
strain TA 98 plus S9, in the presence of the comutagen norharman (9H-pyrido[3,4-b]indole).
None of the compounds was mutagenic without norharman in strains TA98 or TA100. In the
presence of S9 and norharman, nitrobenzene induced reverse mutations in TA 98 but not in
TA 100. Because norharman-containing controls were negative for reverse mutation in this
tester strain, the authors concluded that nitrobenzene could induce reverse mutations in the
presence of a comutagen. In a further series of experiments, Suzuki et al. (1987) demonstrated
that the nitroreductase-deficient isolate TA 98NR was negative for reverse mutations even in the
presence of S9 and norharman. These data are considered to be consistent with the concept that
metabolic activation by S9, and norharman, were unrelated to the induction of nitroreductase, but
presence of the reductase was required to elicit nitrobenzene mutagenicity.
In general, available data on the mutagenicity of nitrobenzene using the Ames assay
demonstrate no effects on reverse mutations (Table 4-40). This conclusion may be tempered by
the limited range of tests that have been employed for nitrobenzene and the inferential evidence
of the compounds' mutagenicity in S. typhimurium TA 98 in the presence of the comutagen,
norharman (Suzuki et al., 1983). In addition, Clayson and Garner (1976) speculated that the
electrophilic nitrenium ion (NH2+) is the ultimate carcinogen from aromatic amino and nitro
compounds and not enough is known about the capability of S. typhimurium to create this
reactive intermediate from nitrobenzene.
Kligerman et al. (1983) exposed male CDF(F344)/CrlBR rats to doses of 0, 5, 16, or
50 ppm nitrobenzene for 6 hours/day, 5 days/week for 21 days during a 29-day period. The
authors assessed the ability of inhaled nitrobenzene to induce cytogenetic damage in the
lymphocytes of isolated spleen or peripheral blood. No statistically significant increases in sister
chromatid exchanges were observed at any doses tested. Similarly, nitrobenzene did not induce
unscheduled DNA synthesis in an in vivo-in vitro hepatocyte DNA repair test (Mirsalis et al.,
1982).
In contrast to the above results, nitrobenzene was weakly positive for the induction of
chromosome aberrations in cultured human peripheral lymphocytes (Huang et al., 1996, 1995).
However, the compound did not induce structural chromosome aberrations in human
spermatozoa incubated with 500 |j,g/mL nitrobenzene for 120 minutes in the absence of S9
fraction (Tateno et al., 1997).
The six-step, one-electron-per-step transfer reduction sequence that has been proposed
for intracellular metabolism of nitrobenzene suggests that nitrobenzene may act as a promoter,
since the reactive intermediates generated during nitrobenzene metabolism may have the
potential to initiate, promote, and/or accelerate the progression of nonneoplastic or neoplastic
changes in cells (Figure 3-7) (Dreher and Junod, 1996; Feig et al., 1994; Guyton and Kensler,
1993; Kensler et al., 1989). Ohkuma and Kawanishi (1999) induced DNA damage in vitro using
87
-------
calf thymus DNA; nitrosobenzene, a primary metabolite of nitrobenzene (5-20 uM); Cu2+ ions
(20 uM); and NADH in a nonenzymatic reaction. Other metal ions, such as Fe2+, Fe3+, or Mn2+,
were ineffective. Bathocuproine, an agent that binds Cu+, chelating agents, or catalase, an
enzyme that destroys H2O2, prevented DNA damage, suggesting that adduct formation
proceeded via an oxidative process requiring the presence of both Cu+ and H2O2. Superoxide
anion or free radical scavengers did not suppress DNA damage. The authors found that NADH
plus Cu2+ caused damage mostly to thymidine and cytosine residues, whereas the »OH radical
attacked DNA in a nonspecific fashion. Therefore, they suggested that Cu2+ binds in a site-
specific manner to DNA, where it is reduced to Cu+ by NADH plus nitrosobenzene, with the
release of H2O2. The latter then forms a DNA-Cu+-H2O2 complex that releases »OH and attacks
the nucleotide at which it was formed. The authors stated that the concentration of NADH used
was well within the physiological range, but they did not elaborate on physiological Cu2+
concentrations. Further work may be needed to see if these in vitro findings are relevant in
whole animal or tissue systems and if a mechanism like this could play a role in organ-specific
carcinogenesis by nitrobenzene.
Bonacker et al. (2004) recently demonstrated the induction of micronuclei in
V79 hamster lung fibroblast cells following exposure to nitrobenzene possibly by affecting
tubulin assembly and the spindle apparatus. To further delineate the mechanism by which the
micronuclei were formed, the authors used primary syndrome of calcinosis, Raynaud's
phenomenon, esophageal motility disorders, sclerodactyly, and telangiectasia (CREST)
antibodies that bind to kinetochore proteins at chromosomal centromeres and detect
aneugenicity. CREST syndrome is a disorder of the skin and connective tissue that leads to
hardening of the skin's surface; its cause is unknown (Schuler et al., 1997; Miller and Adler,
1990). Following an 18-hour incubation, a doubling of micronuclei was observed at 1, 10, and
100 |jM nitrobenzene versus solvent (DMSO) controls. Nitrobenzene (up to 10 |jM) was shown
to induce mostly kinetochore-positive micronuclei, indicative of an aneugenic effect. To
determine the possible effect of nitrobenzene on the cellular spindle apparatus, temperature-
dependent assembly (at 37°C) and disassembly (at 4°C) of tubulin were determined in the
presence of nitrobenzene in vitro. A slight inhibitory effect was observed with 1 mM
nitrobenzene in the absence of DMSO; however, in the presence of 1% DMSO, nitrobenzene
exerted no detectable effect on tubulin assembly up to the solubility limit of about 15 mM. A
functional analysis of the tubulin-kinesin motor system revealed that nitrobenzene had a clear
dose-dependent effect on the gliding velocity of microtubules with a minimal degree of
inhibition above 7.5 |jM to complete inhibition at 30 |jM (Bonacker et al., 2004).
Li et al. (2003a, b), using the ultrasensitive method of accelerator mass spectrometry,
demonstrated recently that nitrobenzene forms adducts with Hb and with hepatic DNA in male
Kunming mice. [14C]-Nitrobenzene was administered intraperitoneally in corn oil at doses of
0.1-100 |ig/kg and 10 mg/kg, and animals were sacrificed 2 hours after treatment. The authors
88
-------
found that both Hb and hepatic DNA adducts occurred with similar dose-response relationships
within 2 hours of exposure over the whole range of doses. Regressions of log dose versus log
adduct per gram Hb or DNA resulted in straight lines with regression coefficients of 0.998 and
0.993, respectively. In addition, a time-course experiment was conducted in which the mice
received 4.1 |ig/kg nitrobenzene and were sacrificed between 4 hours and 21 days after dosing.
This study revealed a biphasic pattern of adduct elimination, with adduct levels in hepatic DNA
attaining peak levels at 4 hours after dosing then declining with a half-life of 10 hours for the
initial 3 days. Thereafter, for up to 21 days, adducts disappeared with a half-life of 6.5 days.
Although the findings of Li et al. (2003a, b) appear to point to a genotoxic potential of
nitrobenzene, they are disputable. The binding level was extremely low, and any biological
significance at such levels of DNA binding is unclear. Also, the DNA adducts were neither
characterized nor identified. Further independent confirmation is warranted to elucidate the
toxicological meaning of these observations.
More recently, however, Robbiano et al. (2004) reported in vivo and in vitro findings that
suggest a genotoxic potential for nitrobenzene. Male Sprague-Dawley rats were administered a
single dose of nitrobenzene (300 mg/kg) by gavage and euthanized 20 hours later. A statistically
significant increase in DNA damage, measured by the comet assay, and broken or detached
chromosomes separated from the spindle apparatus, measured by the micronucleus assay, were
observed. The in vitro findings with primary cultures of kidney cells from male Sprague-Dawley
rats and human kidney cells obtained from patients with kidney cancer were consistent with the
in vivo results. Cells were treated with 0, 0.062, 0.125, 0.25, or 0.5 mM nitrobenzene. This dose
range was based on preliminary studies with concentrations that produced a lower than 30%
reduction of relative survival. Nitrobenzene caused a statistically significant increase in DNA
damage in rat primary kidney cells (0.125-0.5 mM) and human kidney cells (0.062-0.25 mM),
following 20-hour incubation with the compound. A statistically significant increase in
clastogenic effects was observed in rat primary kidney cells (0.0125-0.5 mM) and human kidney
cells (0.250-0.5 mM), following a 48-hour incubation with the compound.
Mattioli et al. (2006) provided in vitro and in vivo evidence of a non-genotoxic mode of
action (MOA) for nitrobenzene. The authors treated primary human thyroid cells with 1.25, 2.5,
or 5 mM nitrobenzene for 20 hours. A dose-dependent increase in DNA fragmentation and
unscheduled DNA synthesis was observed; however, the amount of DNA fragmentation at 5 mM
nitrobenzene was eightfold lower than 0.075 mM methyl methenesulfonate, a monofunctional
alkylating agent used as a positive control. In the companion in vivo studies, the authors treated
rats with a single dose of nitrobenzene (310 mg/kg, by mouth) and examined the degree of DNA
fragmentation 16 hours later in the kidney, liver, and thyroid. The findings showed that the
amount of DNA fragmentation was as follows from highest to lowest: liver ~ kidney > thyroid.
Although the results support the ability of nitrobenzene to generate alkali labile sites, as
measured by the comet assay, these findings need to be viewed cautiously with regard to other
89
-------
effects that may be operational at much lower doses. The in vitro and in vivo findings support
the notion that a redox couple can be established, with the subsequent generation of reactive
oxygen species and DNA fragmentation. However, liver hypertrophy has been shown to occur at
doses as low as 9.38 mg/kg-day in F344 rats (NTP, 1983a), an effect that may alter the
hypothalamic-pituitary-thyroid axis by increasing the clearance of thyroxine (T4) (via
glucuronidation) and triiodothyronine (T3) (via sulfation) in rodents. Such an effect may cause a
compensatory increase in circulating thyroid-stimulating hormone (TSH) and ultimately
follicular cell activation (U.S. EPA, 1998b).
In conclusion, results of genotoxicity testing are mixed. Nitrobenzene appears to be at
most weakly genotoxic. This determination is based on the almost exclusively negative results
in salmonella assays (Ames tests; the only exception is TA98 in the presence of a comutagen), as
well as negative clastogenic findings from in vivo assays of sister chromatid exchange,
unscheduled DNA synthesis, and chromosomal aberrations. In vitro chromosome aberration
results were mixed, as were the DNA breakage and micronucleus data. For instance,
nitrobenzene was weakly positive for the induction of chromosome aberrations in cultured
human peripheral lymphocytes but negative in human spermatozoa. Nitrobenzene induced weak
DNA fragmentation but no DNA strand breaks. In addition, nitrobenzene did not cause cell
transformation in these cell systems. A summary of the genotoxic findings on nitrobenzene is
presented in Table 4-40.
90
-------
Table 4-40. Summary of studies on the direct mutagenicity/genotoxicity
of nitrobenzene
Test system
Cell/strain
Result3
(+/- S9)
Reference
Comments'"
Bacteria
S. typhimurium
TA 98, TA 100, TA 1535, TA
1537
TA 98, TA 98NR, TA 100, TA
100NR, TA 97a, TA 1535, TA
1537, TA 1537NR, TA 1538
TA 98, TA 100, TA 1535, TA
1538
TA 98, TA 100
TA 98, TA 100
TA 98, TA 100, TA 1535, TA
1537, TA 1538
TA98
TA 1538
TA 98, TA 100
TA 98, TA 100
TA98C
TA 100C
TA 98NRC
-/-
-/ND
ND/-
-/-
-/ND
-/-
ND/-
-/-
ND/-
-/-
+/-
-/-
-/-
Haworthetal. (1983)
Vance and Levin (1984)
Anderson and Styles
(1978)
Assmannetal. (1997)
Chiuetal. (1978)
Shimizuetal. (1983)
Hoetal. (1981)
Garner and Nutman (1977)
Dellarco and Prival (1989)
Suzuki etal. (1987, 1983)
Reverse
mutations
Positive in the
presence of
norharman as
comutagen
Mammalian cells in vitro
Human
lymphocytes
Human
spermatozoa
Hamster lung
fibroblasts
Human
hepatocarcinoma
Syrian hamster
kidney cells
Human diploid
lung fibroblasts
Human
hepatocytes
Rat hepatocytes
Human thyroid
cells
V79
SMMC-7721
BHK-21 C13
WI-38
+
+
-
-
-
-
-
+
Huang etal. (1996, 1995)
Tateno etal. (1997)
Bonacker et al. (2004)
Han etal. (2001)
Styles (1978)
Styles (1978)
Butterworth etal. (1989)
Butterworth et al. (1989)
Mattioli et al. (2006)
CA
CA
MN
DNA damage
Cell
transformation
Cell
transformation
UDS
UDS
DNA damage
and UDS
In vivo tests
F344 rats
F344 rats
F344 rats
Kunming mice
Sprague-Dawley
rats
Male Sprague-
Dawley rats
Peripheral blood lymphocytes
Isolated spleen lymphocytes
Hepatocytes
Primary rat kidney cells
-
-
+
+
+
Kligerman etal. (1983)
Kligerman etal. (1983)
Mirsalis etal. (1982)
Li et al. (2003a, b)
Robbiano et al. (2004)
Mattioli et al. (2006)
SCE and CA
SCE
UDS
DNA binding
DNA damage
andMN
DNA damage
and UDS
91
-------
Table 4-40. Summary of studies on the direct mutagenicity/genotoxicity
of nitrobenzene
Test system
Human kidney
cells
Male and female
B6C3F1 mice
Cell/strain
Kidney cell isolates discarded
during surgery
Bone marrow
Result3
(+/- S9)
+
-
Reference
Robbiano et al. (2004)
BASF (1995), as cited in
IPCS (2003)
Comments'"
DNA damage
andMN
MN
aND = no data.
bCA = chromosomal aberration; MN = micronuclei; UDS = unscheduled DNA synthesis; SCE = sister chromatid
exchange.
°In the presence of norharman.
4.4.6. Other Studies in Support of Mode of Action
Han et al. (2001) exposed a human hepatocarcinoma cell line, SMMC-7721, in culture to
nitrobenzene. According to the English translation of the Chinese article, they found that
concentrations at or above 8 mM caused cell death but no DNA strand breaks. They also
observed that typical reactive oxygen scavengers, such as superoxide dismutase, hydrogen
peroxidase, or mannitol, provided partial protection from nitrobenzene-induced cell death. The
authors concluded that nitrobenzene causes cellular damage by reactive oxygen species and that
nitrobenzene was a non-genotoxic agent.
Hong et al. (2002) studied the nephrotoxic potential of nitrobenzene in vitro using renal
cortical slices from male F344 rats. Nitrobenzene was tested at concentrations of 0, 1, 2, 3, 4, or
5 mM for a 2-hour exposure. The authors reported that nitrobenzene was capable of causing a
statistically significant change in cellular function, as measured by a decrease in pyruvate-
stimulated gluconeogenesis, at 1 mM; however, overt cytotoxicity, as measured by an increase in
lactate dehydrogenase release, did not occur at any of the tested concentrations. In contrast to
these findings, Mochida et al. (1986) reported that nitrobenzene was more toxic in comparison to
two established nephrotoxicants (i.e., 1,2-dichloroethane and carbon disulfide) in two cell lines.
The authors exposed a human epidermoid carcinoma cell line (KB) and African green monkey
(Cercopithecus aethiops) kidney (AGMK) cells with doses of nitrobenzene up to 300 |ig/mL for
72 hours. A dose-dependent decrease in cell viability was observed. The concentration of
nitrobenzene reducing cell viability to 50% of control values during the 72-hour exposure period
(ECso or median effective concentration) was calculated to be 42 and 30 |ig/mL in KB and
AGMK cells, respectively.
4.5. SYNTHESIS OF MAJOR NONCANCER EFFECTS
The toxicological effects of nitrobenzene in experimental studies are characterized by a
broad spectrum of noncancer impacts. In general terms, these include the onset of cyanosis and
methemoglobinemia, changes in hematologic parameters, histopathologic lesions of key target
92
-------
organs, such as the spleen, liver, adrenal, kidney, and brain, and testicular atrophy with
associated functional deficits in the male reproductive system, although species-specific
differences with respect to these latter endpoints occur depending on the route of exposure. For
example, oral administration of nitrobenzene induces methemoglobinemia and histopathologic
lesions in the liver (bile stasis, fatty degeneration, centrilobular necrosis, and hepatocellular
nucleolar enlargement), brain (malacia of the cerebellar peduncle), and testes (necrosis of
primary and secondary spermatocytes, multinucleated giant cells) in male F344 rats but not in
male B6C3F1 mice (Morgan et al., 1985; Bond et al., 1981). Unlike oral exposures, however,
hepatic, splenic, and testicular lesions were observed in B6C3F1 male mice following short-term
inhalation exposure to nitrobenzene (Medinsky and Irons, 1985). In addition, inhalation studies
have shown that male and female B6C3F1 mice are more susceptible to developing
histopathologic lesions in the nasal passages and lungs compared with male and female F344 rats
(CUT, 1993). A summary of the MOA for noncancer effects following oral and inhalation
exposures is provided below.
4.5.1. Oral Exposure
The formation of metHb in the blood of human beings and animals appears to be a
consistent feature of almost all case-control or experimental studies on the toxicity of
nitrobenzene. That this response and potentially associated histopathologic responses such as
congestion of the spleen are a primary toxicological effect of nitrobenzene is indicated by their
potential to be triggered at lower doses than most of the other responses to the compound.
Holder (1999) hypothesized how interconversion between nitrobenzene and the primary
metabolites nitrosobenzene, phenylhydroxylamine, and aniline are intimately associated with the
oxidation of the Hb prosthetic group to the ferric state (see Figure 3-8). The consequent anemia
is caused by depleted oxygen-carrying capacity, globin chains altered by binding to thiol-
containing amino acids, and RBC lysis.
The discussion of a case report by Schimelman et al. (1978) pointed out that nitrobenzene
is but one of a wide range of toxicants that can induce methemoglobinemia. Toxic
methemoglobinemia is likely to occur if the rapid formation of metHb overwhelms the capacity
of the protective enzyme systems (i.e., NADH-cytochrome b5 reductase [major pathway] and
NADPH-cytochrome c reductase [minor pathway]) (see Table 3-5) (Jaffe, 1981). The NADH-
cytochrome bs reductase pathway in RBCs may reduce metHb to Hb at a rate of approximately
15% per hour in healthy individuals, assuming no ongoing metHb production (Finch, 1947).
Small amounts of methemoglobin are continually produced due to autoxidation of
hemoglobin during the normal respiratory function of loading and unloading of oxygen by
erythrocytes. Under normal conditions, the level of metHb in RBCs is kept below 1% of total
Hb (Harrison, 1977); however, a normal range in healthy humans is generally considered to be
0-3% (Lee and Ferguson, 2007). (See footnotes 1 and 3 under section 3 for additional
93
-------
information on metHb and methemoglobinemia.) Tissue hypoxia may develop due to the
presence of excessive amounts of metHb, which is incapable of transporting oxygen in the body.
Methemoglobin reduces tissue oxygenation by two mechanisms: iron in the ferric rather than the
ferrous form is unable to combine with oxygen and consequently the oxygen-carrying capacity
of the blood is reduced, and the presence of oxidized iron changes the heme tetramer in such a
way as to reduce oxygen release in the tissues (i.e., shifts the oxyHb dissociation curve to the left
as in alkalosis) (Ellenhorn et al., 1997).
There appears to be a progression of incrementally more severe symptoms in humans
with increasing metHb concentration. Most patients tolerate low levels (<10%) of metHb fairly
well despite appearance of cyanosis (bluish color) of lips and ears at levels as low as 6% in those
with light complexions; however, some otherwise normal patients may experience difficulty
tolerating metHb levels between 10 and 15% (Coleman and Coleman, 1996). In general, a
metHb level of 20% may be considered too high for a patient to be considered asymptomatic,
although patient discomfort may largely depend on the rapidity with which metHb accumulates.
Fatigue, dyspnea, headache, nausea, and tachycardia may occur at metHb levels of 30 to 50%,
while lethargy, stupor, and deteriorating consciousness occur as levels approach 55%. Higher
levels may cause cardiac arrhythmia, circulatory failure, and neurological depression, while
levels at or above 70% are usually fatal (Coleman and Coleman, 1996).
NTP (1983a) is the single oral study in which experimental animals were exposed to
nitrobenzene for a sufficient duration to permit dose-response analysis. In the study, 10 F344
rats/sex/group received 0, 9.38, 18.75, 37.5, 75, and 150 mg/kg-day and 10 B6C3F1
mice/sex/group received 0, 18.75, 37.5, 75, 150, and 300 mg/kg-day by gavage in corn oil for
90 days. There was good consistency in the range of adverse effects attributable to the
compound among rats and mice. These included mortality in some animals at the highest doses
(150 mg/kg-day in rats and 300 mg/kg-day in mice), dose-dependent increases in absolute and
relative weights of the liver and kidney, but a progressive decrease in absolute and relative testis
weights. Hematologic parameters of F344 rats and mice were markedly affected by nitrobenzene
in this study. For example, Hb concentrations, RBC counts, and Hct were dose-dependently
reduced in both species, whereas percent reticulocytes and metHb concentration were dose-
dependently increased. For the reticulocyte and metHb effects, statistical significance compared
to controls was achieved at all dose levels. Histopathologic lesions were observed in the spleen,
testis, and brain in both exposed species. In addition, liver lesions were observed in B6C3F1
mice, while kidney effects were observed in F344 rats. The congestion of the spleen (especially
in F344 rats) was noteworthy since it may be associated with the presence of metHb in the RBCs
of exposed animals.
Among studies where nitrobenzene was administered for shorter durations to laboratory
animals via the oral route, Bond et al. (1981) observed a dose-dependent increase in metHb
formation in male F344 rats, with the increases becoming statistically significantly different in
94
-------
the 110 mg/kg and higher dose groups. An increase in metHb in response to orally administered
nitrobenzene in Sprague-Dawley rats also was observed in the OECD-protocol
reproductive/developmental toxicity study conducted by Mitsumori et al. (1994). These studies
revealed a statistically significant increase in blood metHb at the lowest dose level employed
(20 mg/kg-day). These findings contrast with those of Burns et al. (1994), who, while reporting
a number of hematologic perturbations in B6C3F1 mice as a result of nitrobenzene exposure (up
to 300 mg/kg for 14 days), did not report any compound-related increases in metHb formation.
This may be consistent with the observation that mice are more resistant than rats to the metHb-
forming properties of nitrobenzene (IPCS, 2003).
Of particular note in the NTP (1983 a) study of nitrobenzene, the F344 rats experienced
significant mortality at a dose that also produced a 12-13% increase in metHb levels. It is not
clear what process led to these deaths. Further support for mortality being a nitrobenzene-related
effect at comparatively low metHb levels was seen in the Mitsumori et al. (1994) study, with
20% of the male Sprague-Dawley rats administered 100 mg/kg-day nitrobenzene dying (within
35 days of exposure) at an average metHb level of 6.8% in the surviving rats and 100% of the
pregnant female rats in the 100 mg/kg-day group dying by 41 days of exposure (metHb not
measured).
Closely related to the formation of metHb in nitrobenzene-treated rodents (especially
rats) is the range of changes induced in other hematologic parameters. These are likely to be part
of the same metHb-induced continuum of RBC-related toxicological consequences of
nitrobenzene reduction and the uptake of its metabolites by RBCs. The reproductive or
developmental toxicity study on nitrobenzene by Mitsumori et al. (1994) identified statistically
significant changes compared with controls in a number of hematologic parameters, including
reductions in RBCs, Hb, and Hct and increases in erythroblast, reticulocyte, and WBC counts as
a result of oral administration of nitrobenzene to male and female Sprague-Dawley rats for
approximately 41 days (Table 4-29). Burns et al. (1994) documented a similar suite of
hematologic effects in B6C3F1 mice that were orally exposed to nitrobenzene for 14 days.
Increases in reticulocyte counts were especially marked in this species.
The male reproductive system—testis, epididymis, and seminiferous tubules—comprises
an important target for nitrobenzene toxicity in rodents. Impairment of this system due to
nitrobenzene has become apparent through the formation of histopathologic lesions, the
production of sperm with reduced motility and/or viability, and, in some studies, functional
deficits such as reduction in fertility. For example, the 90-day oral gavage study in F344 rats and
B6C3F1 mice sponsored by the NTP (1983a) showed a dose-dependent atrophy of the testis and
the appearance of a range of treatment-related histopathologic lesions. In the single-exposure
oral study carried out by Bond et al. (1981) in F344 rats, a number of distinct histopathologic
effects in the testes and seminiferous tubules were apparent at a dose level of 300 mg/kg or
greater. The lesions were marked by necrosis of spermatogenic cells, the appearance of
95
-------
multinucleated giant cells, and an associated decrease in sperm count. A single dose of
300 mg/kg was also effective in temporarily abolishing spermatogenesis in male F344 rats, in
parallel to a marked degeneration of the seminiferous epithelium (Levin et al., 1988).
Reestablishment of sperm generation appeared in concert with the partial restoration of normal
cellular architecture.
Other short-term oral exposure studies that centered on the effects of nitrobenzene on the
male reproductive system include those of Koida et al. (1995) and Matsuura et al. (1995), both of
which demonstrated a relative decrease in epididymal weight, reduced sperm motility and
viability, histopathologic and morphologic abnormalities, and degeneration of the spermatids and
pachytene spermatocytes (in Sprague-Dawley rats exposed by gavage to nitrobenzene at doses of
30-60 mg/kg for up to 4 weeks). Kawashima et al. (1995a) observed similar changes in
testicular and epididymal responses in male Sprague-Dawley rats exposed orally to 60 mg/kg
nitrobenzene for up to 70 days and demonstrated that, for males exposed to nitrobenzene for
21 days or more before mating, there was a reduction in the fertility index of their (unexposed)
breeding partners. This was considered to be a consequence of the nitrobenzene-induced
production of sperm with low motility and viability.
Notwithstanding the appearance of profound histopathologic effects in the testes and
epididymides, Mitsumori et al. (1994) did not observe impaired fertility as a result of exposing
Sprague-Dawley rats to up to 100 mg/kg nitrobenzene for 14 days prior to mating,
reemphasizing the importance of the spermatogenic cycle to reproductive performance. Taken
together, the data of Kawashima et al. (1995a) and Mitsumori et al. (1994) point to the ability of
nitrobenzene to disrupt spermatogenesis by causing the production of sperm with reduced
motility and viability. This will result in reduced fertility if the males are mated at the point
when the deficient sperm are released.
As set forth in section 4.3, nitrobenzene has been included as a positive control in studies
aimed at refining experimental techniques for evaluating the spermatotoxic effects of potentially
harmful chemical agents (Ban et al., 2001; Linder et al., 1992; Allenby et al., 1991, 1990). As
reported in a number of meeting abstracts, oral exposure of rats for 14 days resulted in
histopathologic changes in the testes and epididymides and in the production of an increased
proportion of abnormal sperm (Kito et al., 1999, 1998; Kato et al., 1995). Morphologically
normal sperm from rats undergoing these treatments displayed reduced motility.
Other target organs of nitrobenzene toxicity following oral administration to rodents
include the liver, kidney, thyroid, and brain, as indicated by changes in relative organ weights
and the appearance of histopathologic lesions. For example, the 28-day oral gavage study of
Shimo et al. (1994) in F344 rats noted a characteristic brown coloration of the perivascular
region of the cerebellum, increased medullary hematopoiesis of the liver, and brown
pigmentation of the renal tubular epithelium. The latter symptoms are a likely result of
deposition of metHb and/or degradation products.
96
-------
4.5.2. Inhalation Exposure
In general, long-term studies of the toxicology of nitrobenzene in experimental animals
have employed inhalation as the route of administration. As with oral exposures to nitrobenzene,
inhalation exposures result in the formation of metHb. However, in contrast to the two-electron
additions that occur in the intestinal lumen of experimental animals following oral exposures to
nitrobenzene, metabolism of nitrobenzene from inhalation exposures is expected to occur via
one-electron additions, with the resultant formation of the nitro anion free radical. As depicted
by Holder (1999), the nitro anion free radical can be further reduced in RBCs to nitrosobenzene
and phenylhydroxylamine, both of which participate in the formation of metHb. However, the
nitro anion free radical may also be oxidized back to the parent compound with the subsequent
formation of the superoxide free radical.
The most comprehensive of these studies was a 2-year investigation of the inhalation
effects of nitrobenzene in male and female F344 rats, male Sprague-Dawley (CD) rats, and male
and female B6C3F1 mice (Cattley et al., 1994; CUT, 1993). Included in a wide range of cancer
and noncancer effects were the dose-dependent increases in metHb that achieved statistical
significance in each species and strain under test. For example, in male Sprague-Dawley rats,
statistically significant differences in this parameter were observed at all exposure levels after
15 months, compared with controls (interim blood samples). The lowest concentration
administered to male Sprague-Dawley rats (1 ppm) is a chronic exposure LOAEL (unadjusted)
for metHb formation, measured at the interim sacrifice (15 months). This suggests that male
Sprague-Dawley rats may form metHb more readily than F344 rats of either sex, male B6C3F1
mice, or female B6C3F1 mice, for which 1, 25, and 5 ppm, respectively, are NOAELs for metHb
measured at study termination (24 months) and for metHb in F344 rats at interim sacrifice.
MetHb levels in mice were not measured at 15 months. This study also reported
bronchiolization of the alveoli in both male and female B6C3F1 mice. Unlike the systemic
effects, this portal-of-entry effect was detectable in >87% of mice at the lowest dose tested
(5 ppm) and nearly 100% of animals at 50 ppm. Bronchi olizati on of the alveoli was not
detectable in controls. Pulmonary effects have also been observed in subchronic inhalation
studies in both F344 rats and B6C3F1 mice (CUT, 1984). In male F344 rats, 60% of the animals
in the 50 ppm group exhibited bronchiolar epithelium hyperplasia, whereas 20% of females were
found with this lesion. In B6C3F1 mice, bronchial mucosa hyperplasia was observed in 78% of
males and 100% of females at 50 ppm.
According to Nettesheim and Szakal (1972), bronchi olizati on of alveoli are lesions that
may arise from the "colonization" of alveolar walls with bronchiolar epithelium either via cell
migration through alveolar pores or from the transformation (metaplasia) of alveolar type II cells
into bronchiolar-type epithelium. The pathology summary in CUT (1993) characterized
bronchi olizati on of the alveolar walls as "a pronounced change in the alveolar epithelium in the
region of the terminal bronchioles from a simple squamous to tall columnar epithelium
97
-------
resembling that of the terminal bronchioles" (CUT, 1993). Bronchiolization was dose related in
severity such that this epithelial change tended to be located almost entirely in the region of the
terminal bronchioles in animals of the low-dose exposure group, while the lesions were more
florid and involved a large proportion of the lung parenchyma in mice that were exposed to the
mid or high concentration of nitrobenzene. In contrast, alveolar/bronchiolar hyperplasia was
characterized by a more discrete lesion than that described for bronchi olizati on of the alveolar
walls. In hyperplasia "the epithelium involved was similar in appearance but tended to
proliferate into the alveolar spaces in multiple fronds, forming a space-occupying lesion which
did not exhibit compression of adjacent tissues" (CUT, 1993).
Alveolar bronchiolization does not appear to be a pre-neoplastic event or a prerequisite to
lung neoplasia (Cattley et al., 1994; CUT, 1993). While incidences of alveolar bronchiolization
were high among all nitrobenzene-exposed male and female mice, only male mice had increased
incidences of combined adenomas and carcinomas of the lungs. Furthermore, there were no
incidences of alveolar bronchiolization among nonexposed control mice, but incidences of
combined adenomas and carcinomas of the lungs were 19 and 13% in the control males and
females, respectively. Therefore, there is no apparent association in the nitrobenzene
carcinogenicity study between findings of lung neoplasia and bronchiolization in mice. Similar
findings and conclusions were derived from another study following a lifetime exposure to
p-nitrotoluene (NTP, 2002).
In an NTP dietary feeding study of p-nitrotoluene (at 0, 1,250, 2,500, and 5,000 ppm), the
incidence and severity of alveolar bronchiolization was increased dose dependently in mice
(males: 0/50, 20/50, 30/50, and 48/50; females: 0/50, 33/50, 41/50, and 49/50) but not in rats.
Also, the combined incidence of alveolar/bronchi olar adenoma or carcinoma was significantly
increased in the 5,000 ppm male mice and exceeded the historical control range while female
mice had no increased lung neoplasia. Bronchiolization was considered a metaplastic change
and not a pre-neoplastic lesion based on lack of relationship between incidences of lung
neoplasia and bronchiolization (NTP, 2002).
Alveolar bronchiolization does not seem to be mouse specific nor does it seem to be
tightly linked to lung neoplasia (as for instance being a pre-neoplastic event) since the same
pathology has been described in other species, including humans, in the absence or presence of
lung neoplasia. In rats or mice, continued inhalation of low levels of toxic oxidant gases, such as
ozone or synthetic smog, or exposure to particulate irritants, such as silica or calcium chromate,
may result in bronchiolization of alveoli (Friemann et al., 1999; Muhle et al., 1995; Pinkerton et
al., 1993; Nettesheim and Szakal, 1972). In several instances, bronchiolization was evidenced in
the absence of lung neoplastic or pre-neoplastic changes (Friemann et al., 1999; Pinkerton et al.,
1993; Nettesheim and Szakal, 1972). In addition to bronchiolization, increased incidences of
bronchoalveolar hyperplasia and lung tumors were also found in rats (Muhle et al., 1995).
98
-------
Nonetheless, being a cancer precursor, per se, does not necessarily disqualify an endpoint from
being used to develop an RfC.
The findings of another study indicate that alveolar bronchiolization may not be
dependent on hyperplasia or DNA proliferation following a single intratracheal instillation of
quartz in rats (Friemann et al., 1999). Bronchiolization, described as an outgrowth of
bronchiolar epithelium to alveolar walls, was increased (more than 30-fold) at 12 and 18 months
(but not at 3 or 6 months) after a single quartz instillation. In contrast, bronchoalveolar
hyperplasia was increased between 6 and 12 months but not at 18 months, while DNA
proliferation was increased (without achieving statistical significance) throughout the 18 months
and seemed to peak at 12 months, followed by a drop at 18 months. One of the conclusions in
this study was that "alveolar bronchiolization was not regularly associated topographically with
the development of fibroproliferative lesions, which is usually observed after exposure to
chrysotile asbestos" (Friemann et al., 1999).
Bronchiolization of alveoli was also found in bleomycin-induced pulmonary fibrosis in
baboons (Collins et al., 1982) and in human lung cancer biopsies (Jensen-Taubman et al., 1998)
as well as in cancer-free patients with idiopathic pulmonary fibrosis/usual interstitial pneumonia
(Chilosi et al., 2003). While most biopsied tissues came from lung cancer patients (specifically,
non-small-cell lung cancer), some specimens were from non-lung cancer patients with findings
of alveolar bronchiolization in 12 and 8% of the specimens, respectively (Jensen-Taubman et al.,
1998).
In addition to bronchiolization of the alveoli in mice in the 2-year inhalation bioassay by
CUT (1993), nitrobenzene also resulted in dose-related changes in the nasal passages in mice and
rats. In male and female mice, nitrobenzene caused an increased incidence of olfactory epithelial
degeneration at all exposure concentrations in female mice and at mid and high concentrations in
male mice. CUT (1993) noted that these changes were consistent with degenerative changes in
the olfactory bulbs of rats following inhalation of nitrobenzene as described by Beauchamp et al.
(1982). In the CUT (1993) inhalation study in rats, olfactory changes in the nasal passages in
both F344 and CD rat strains were less marked than in mice, consisting of an increased incidence
of focal inflammation and hypertrophy of the submucosal glands in areas lined by respiratory
epithelium in high-exposure F344 rats, a slight treatment-related increase in the incidence and
severity of inflammatory changes in the anterior section of the nose in CD rats, and pigment
accumulation within the olfactory region in both rat strains.
The 2-year inhalation study of nitrobenzene noted statistically significant reductions in
RBCs, Hct, and Hb in those rats exposed to nitrobenzene at the highest dose of 25 ppm (Cattley
et al., 1994; CUT, 1993). This would identify an unadjusted NOAEL of 5 ppm for the onset of
effects on these hematologic parameters. Although a statistically significant increase in the
incidence of extramedullary hematopoiesis in the spleen was noted in F344 rats exposed to
1 ppm nitrobenzene for 2 years, the extent of the difference from controls was not particularly
99
-------
striking because of the high background incidence in aging rodent spleens (Cattley et al., 1994;
CUT, 1993). Adverse effects on the spleen, however, were more apparent in younger animals
exposed to nitrobenzene for 90 days (CUT, 1984; NTP, 1983a). In both sexes of F344 and CD
rats and in B6C3F1 mice, exposure to nitrobenzene at 50 ppm was associated with increases in
absolute and relative spleen weights at necropsy, an obvious enlargement of the organ, the
appearance of histopathologic lesions characterized by acute sinusoidal congestion, and
increased extramedullary hematopoiesis. Other features of altered spleen histopathology
included an increase in the number of macrophages infiltrating the red pulp and a proliferation of
capsular lesions. Although the effects on the spleen were less severe at lower concentrations,
extramedullary hematopoiesis was observed even in the low-concentration (5 ppm) group. An
unadjusted LOAEL of 5 ppm would apply to this effect from the data in the study. Short-term
inhalation studies of nitrobenzene toxicity in experimental animals also have resulted in metHb
formation (Medinsky and Irons, 1985; CUT, 1984).
Male CD rats exposed to nitrobenzene by inhalation developed histopathologic lesions of
the spleen in mid- and high-dose (39 and 112 ppm) groups (DuPont, 1981). Similarly, pregnant
female Sprague-Dawley rats exposed via inhalation to 0, 1, 10, or 40 ppm nitrobenzene on
GDs 6-15 displayed an increase in the relative spleen weight in the mid- and high-concentration
groups (Tyl et al., 1987).
The 2-year and 90-day inhalation studies on the toxicological effects of nitrobenzene in
rodents noted a range of histopathologic effects on the reproductive organs (Cattley et al., 1994;
CUT, 1993, 1984). For example, in the 2-year study the development of bilateral hypertrophy of
the testis in CD rats was considered to be compound related because of the concentration-related
incidence of the lesion among exposed groups and its statistically significant increase, 35/61 at
the highest exposure level (25 ppm) versus 11/62 in controls. This suggests that the mid-
concentration level of 5 ppm would represent an unadjusted NOAEL for this effect in CD rats.
Reductions in testicular weight and associated histopathologic changes also were features of the
90-day study (CUT, 1984). The effects were noted in F344 rats, CD rats, and B6C3F1 mice at
the highest dose of 50 ppm. Bilateral testicular atrophy was observed in 10/10 male CD rats
exposed to 50 ppm nitrobenzene but in only 2/10 animals exposed to 16 ppm. This
concentration, therefore, would constitute an unadjusted NOAEL for this effect, based on the
data in the 90-day study.
The two-generation reproductive study in Sprague-Dawley rats reported the well-
recognized effects of nitrobenzene on the histopathology of the male reproductive system, with
reduced fertility resulting from exposed F0 males mating with exposed F0 females, exposed FI
males mating with FI females, and "recovered" FI males mating with virgin females (Dodd et al.,
1987). The authors suggested a NOAEL of 10 ppm for reproductive toxicity in F344 rats.
In contrast to the effects of nitrobenzene on the male reproductive system, nitrobenzene
administered to pregnant rats and rabbits displayed few effects on reproductive, developmental,
100
-------
or teratological parameters under the conditions of the studies (BRRC, 1985, 1984; Biodynamics
Inc., 1984, 1983).
In the 2-year and 90-day inhalation studies of nitrobenzene (Cattley et al., 1994; CUT
1993, 1984), nonneoplastic lesions of the liver included both morphologic and histopathologic
effects. For example, in the 2-year study (Cattley et al., 1994; CUT, 1993), an increase in the
incidence of eosinophilic foci in the livers of male F344 rats was observed at the mid-
concentration level of 5 ppm, while centrilobular hepatocytomegaly was observed in the males of
both strains of rat at the 5 and 25 ppm levels. In addition, in the 90-day study, the formation of
histopathologic lesions identified as basophilic hepatocytes was observed in all male B6C3F1
mice exposed to nitrobenzene at the high-concentration level, whereas these lesions were absent
from female mice in all dose groups.
4.5.3. Mode-of-Action Information
As set forth in section 3.3, plausible schemes have been developed that link nitrobenzene
metabolism in the gastrointestinal lumen and tissues with the biochemical, physiological, and
toxicological changes observed in target organs (e.g., liver and lung). Phase I metabolism occurs
mostly by intestinal microflora following oral exposure and, at a lower rate, in the tissues after
gastrointestinal absorption or following internalization by any other route of exposure. The
extent to which the route of administration determines target organ toxicity is uncertain. It is,
however, likely that the metabolites produced by intestinal microflora, such as o-, m-, and p-
nitrophenols, o-, m-, and p-aminophenols, and aniline, can undergo further metabolism inside the
mammal organism to form a variety of reactive, mostly short-lived intermediates, such as
nitrosobenzene, phenylhydroxylamine (Figure 3-3), and the benzene nitrenium ion. These may
be formed by the action of microsomal NADPH-cytochrome c reductase, by mitochondrial and
cytosolic nitroreductases, and by hydroxylases poorly characterized with respect to nitrobenzene.
Some of these reactions, such as formation of the nitro anion free radical, are reversed
immediately in a nonenzymatic process, leading to futile redox cycling with the regeneration of
the parent compound (i.e., nitrobenzene) and the concurrent formation of superoxide anion. A
similar type reaction occurs with the production of pulmonary toxicity—that is, redox cycling
with the generation of superoxide anion with paraquat, a prototypical pulmonary toxicant
(Parkinson, 2000). Since the activity of nitroreductase type II is the predominant form in the
respiratory system, generation of the nitro anion free radical with subsequent futile cycling may
explain the respiratory effects observed in rats and particularly mice following inhalation
exposures to nitrobenzene. In addition, the nitroso derivatives can enter redox processes that
result in the formation of reactive oxygen species (nitro anion, nitroxide, and superoxide free
radical) (Figure 3-3). Phase II metabolism appears to involve acetylation at the amino group or
conjugation with sulfate, glucuronic acid, and, predominantly, GSH. GSH conjugates may be
split to reenter the futile redox cycle. Further support of the protective effects of GSH
101
-------
conjugation came from the studies by Nystrom and Rickert (1987) with three dinitrobenzene
isomers (e.g., 1,2-, 1,3-, and 1,4-dinitrobenzene). The authors showed that 1,3-dinitrobenzene
was the only isomer that is not conjugated with GSH. The relevance of this finding is that
1,3-dinitrobenzene is the only isomer to cause testicular toxicity. Therefore, they speculated that
the testicular toxicity of this compound may be related to the ease of its reduction to a nitroso
compound plus the lack of its removal via conjugation. Ellis and Foster (1992) investigated the
metabolism of the same three isomers in subcellular fractions from rats of the Alpk:AP (Wister-
derived) strain. They found that the soluble fraction from testis homogenate (but not
microsomes) contains a powerful nitroreductase that works under aerobic conditions,
transforming 1,3-dinitrobenzene to m-nitrosonitrobenzene. The authors did not investigate
whether this enzyme works on 1,2- or 1,4-dinitrobenzene as well, which might have provided
more information on the unique testicular toxicity of the 1,3-isomer. Still, assuming that this
enzyme activity is high in testis, as compared with other organs, and that it is able to reduce
nitrobenzene to nitrosobenzene, provides a reasonable explanation for the pronounced testicular
toxicity of nitrobenzene.
Skeletal variations following gestational exposure of Sprague-Dawley rats to
nitrobenzene were observed only at doses toxic to the mother, thus suggesting strongly that the
effect was due to maternal toxicity rather than direct embryotoxicity.
While the details are not understood, there is evidence linking the interconversion of
nitrobenzene and its metabolites to the formation of metHb and to the possible binding of
nitrosobenzene to important thiol-containing macromolecules, such as Hb and GSH. Other
intracellular proteins containing cysteine residues also would be expected to undergo such
interactions (IPCS, 2003; Holder, 1999). Changes in blood chemistry values and splenic
pathology observed after nitrobenzene intoxication are the likely consequences of metHb
formation, Hb destruction, and the deposition of degradation products in these tissues. Splenic
toxicity is likely related to erythrocyte toxicity, because a primary function of the spleen is to
scavenge senescent or damaged RBCs. Splenic injury may arise from the deposition of massive
amounts of iron or other RBC breakdown products, with an added potential for reactive
metabolites of nitrobenzene to take part in additional intracellular reactions.
The six-step, one-electron-per-step transfer reduction sequence that has been proposed
for intracellular metabolism of nitrobenzene (Figure 3-7) may result in reactive intermediates
that can react with cells or tissues where this sequence is operative, leading to gross and
microscopic changes. As demonstrated by Ohkuma and Kawanishi (1999), reactive oxygen
species formed in the metabolic processing of nitrobenzene and its derivatives can cause damage
to DNA; however, the bulk of experimental evidence from genotoxicity assays has provided
negative results. Reactive oxygen species, in general, have the potential to initiate, promote,
and/or accelerate the progression of nonneoplastic or neoplastic changes in cells (Dreher and
Junod, 1996; Feig et al., 1994; Guyton and Kensler, 1993; Kensler et al., 1989).
102
-------
4.6. EVALUATION OF CARCINOGENICITY
4.6.1. Summary of Overall Weight of Evidence
According to the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
nitrobenzene is characterized as "likely to be carcinogenic to humans" for any route of exposure.
Nitrobenzene has been shown to be a carcinogen in rats and mice (see Table 4-19). Adenomas
and/or carcinomas with a pronounced dose-response relationship were found in livers of male
F344 and male CD rats and in thyroids of male F344 rats. Less pronounced dose-related trends
were observed for kidney tumors in male F344 rats, endometrial polyps in female F344 rats,
cancers of the lung and thyroid in male B6C3F1 mice, and cancers in the mammary gland in
female B6C3F1 mice. In all cases the incidences at the highest dose were elevated statistically
significantly compared to controls. While there are no human carcinogenicity data on
nitrobenzene, the cancer characterization is based on evidence of the compound's tumorigenicity
in a single well-conducted study in two animal species (Cattley et al., 1994; CUT, 1993).
Furthermore, the 2005 cancer guidelines (U.S. EPA, 2005a) state that when tumors occur at a
site other than the point of initial contact, the descriptor generally applies to all exposure routes
that have not been adequately tested at sufficient doses. An exception occurs when there is
convincing information (e.g., toxicokinetic data that absorption does not occur by another route).
Thus, nitrobenzene is "likely to be carcinogenic to humans by any route of exposure." This
decision is based on the observations that nitrobenzene is absorbed via all routes and reductive
and oxidative metabolites of nitrobenzene are produced following inhalation, oral, or dermal
exposures. The carcinogenic action of nitrobenzene may be related to these intermediates or the
oxygen free radicals they may produce; however, there is no experimental evidence supporting
oxidative stress as a carcinogenic MOA (refer to section 4.6.3).
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence
The carcinogenicity of nitrobenzene has been evaluated in male and female mice
(B6C3F1), male rats of two strains (F344/N and Sprague-Dawley), and female rats of one strain
(F344/N). When administered to mice and rats by inhalation, nitrobenzene caused statistically
significant increased incidences of tumors at multiple tissue sites in both species. Exposure to
nitrobenzene caused lung and thyroid tumors in male B6C3F1 mice and mammary gland tumors
in female B6C3F1 mice (Cattley et al., 1994; CUT, 1993). Exposure to nitrobenzene caused
liver tumors in male rats of the F344/N and Sprague-Dawley strains, kidney tumors in male
F344/N rats, and endometrial polyps in female F344/N rats. In addition, statistically significant
increasing trends in the incidences of liver tumors in female mice and female F344/N rats and
thyroid tumors in male F344/N rats were observed with increasing nitrobenzene exposure levels
(Cattley et al., 1994; CUT, 1993). A summary of the carcinogenicity results is presented in
Table 4-41.
103
-------
Table 4-41. Neoplasms in F344 and CD rats and B6C3F1 mice exposed to
nitrobenzene via inhalation for 2 years
Species, sex, strain
Mouse, male, B6C3F1
Mouse, female,
B6C3F1
Target organ,
tumor type"
Lung, bronchio-alveolar
adenoma or carcinoma
Thyroid, follicular
cell adenoma
Liver,
hepatocellular adenoma
Mammary gland,
adenocarcinoma
Thyroid, follicular cell
adenoma
Rat, female, F344
Rat, male, F344
Rat, male, CD
Rat, female, F344
Rat, male, F344
Liver,
hepatocellular adenoma
or carcinoma
Uterus, endometrial
stromal polyp
Kidney, tubular
adenoma or carcinoma
Thyroid, follicular cell
adenoma or carcinoma
Nitrobenzene concentration (ppm)
0
8/42
(19.0%)
0/41
(0%)
4/31
(12.9%)
0/30
(0%)
1/30
(3.3%)
5
16/44 (36.4%)
4/44
(9.1%)
4/38
(10.5%)
Not evaluated
0/37
(0%)
25
20/45
(44.4%)
1/45
(2.2%)
5/46
(10.9%)
Not evaluated
2/45
(4.4%)
50
21/48
(43.8%)
6/46
(13.0%)
11/34
(32.4%)
2/34b
(6%)
2/34
(5.9%)
Nitrobenzene concentration (ppm)
0
0/49
(0%)
1/43
(2.3%)
0/23
(0%)
9/48
(18.8%)
0/43
(0%)
1/43
(2.3%)
1
2/50
(4.0%)
4/50
(8.0%)
0/23
(0%)
15/50
(30.0%)
0/50
(0%)
1/50
(2.0%)
5
0/59
(0%)
5/47
(10.6%)
1/25
(4.0%)
14/50
(28.0%)
0/47
(0%)
5/47
(10.6%)
25
3/49
(6.1%)
16/46
(34.8%)
5/23
(21.7%)
19/49
(38.8%)
6/46
(13.0%)
8/46
(17.4%)
aAll tumor incidences in this table displayed statistically significant (p < 0.05), dose-related trends in the
Cochran-Armitage test.
bThe incidence among all high-dose female mice (including interim sacrifice at 15 months) was statistically
significantly higher than in the control group (5/60 [10%] versus 0/48, respectively), while the incidences
between these two groups at the terminal sacrifice only (in table) were not statistically significantly
different.
Sources: Cattley et al. (1994); CUT (1993).
While no evidence exists to directly address the issue of the carcinogenicity of
nitrobenzene in humans, the "likely" weight-of-evidence descriptor is chosen because the
compound was shown to be carcinogenic in a 2-year inhalation experiment that resulted in the
dose-related formation of tumors at multiple tissue sites in both species of animals employed in
the study (Cattley et al., 1994; CUT, 1993). In this study the strongest individual carcinogenic
response to nitrobenzene was the dose-dependent increase in the incidence of hepatocellular
tumors in male F344 rats, for which the incidence and trend data showed statistically significant
effects in the formation of both adenomas and carcinomas. These data constitute sufficient
evidence of carcinogenicity, and hepatocellular tumors may be considered to be the primary
carcinogenic effect of the compound. This overall conclusion is strengthened by the
nitrobenzene-induced formation of hepatocellular adenomas and carcinomas in male CD rats,
104
-------
though these tumors were predominantly benign. The observations of hepatocellular neoplasia
in rats strengthen the relevance of the animal studies to humans because the spontaneous
incidence of hepatocellular neoplasia was lower in F344 or CD rats than in B6C3F1 mice and the
exposure concentration of nitrobenzene was lower for rats than for mice. Neoplastic effects were
also observed in other organs, such as the endometrium in female F344 rats, thyroid and kidney
in male F344 rats, lung and thyroid in male B6C3F1 mice, and mammary gland in female
B6C3F1 mice. It should be noted that, although the thyroid and kidney tumors observed in male
rats and thyroid tumors in male mice are suggestive of rodent-specific MO As, the experimental
data do not satisfy the criteria set forth in EPA's technical reports on Assessment of Thyroid
Follicular Cell Tumors (U.S. EPA, 1998b) and Alpha2U-Globulin: Association with Chemically
Induced Renal Toxicity and Neoplasia in the Male Rat (U.S. EPA, 1991b) to make this
determination. Other evidence that supports the classification of nitrobenzene as a likely human
carcinogen is the known carcinogenicity of aniline, a metabolite of nitrobenzene (U.S. EPA,
1994c). A recent study by Bonacker et al. (2004) pointed to an aneugenic potential of
nitrobenzene. Studies by Li et al. (2003a, b) showed that nitrobenzene is capable of binding to
hepatic DNA; however, further research will be needed to characterize the DNA adducts and
their toxicological relevance.
No information is available on the carcinogenic effects of nitrobenzene via the oral route.
However, the available information from subchronic oral studies suggests that the compound
could be carcinogenic via the oral route. This conclusion is based on the ready absorption of the
compound at the intestinal absorption barrier and the fact that, in the 2-year inhalation study,
tumors were formed in tissues remote from the site of absorption. These findings suggest that
nitrobenzene or its metabolites can cause tumor formation at multiple sites following passage
into the general circulation. Such a capability would be expected to apply to nitrobenzene when
administered orally. However, the issue of the carcinogenicity of nitrobenzene by the oral route
constitutes a data gap.
4.6.3. Mode-of-Action Information
Based on the studies discussed in section 4.4.5, nitrobenzene appears to be, at most,
weakly genotoxic. This determination is based on the almost exclusively negative results in
salmonella assays (Ames tests; the only exception is TA98 in the presence of a comutagen), as
well as negative clastogenic findings from in vivo assays of sister chromatid exchange,
unscheduled DNA synthesis, and chromosomal aberrations. In vitro chromosome aberration
results were mixed, as were the DNA breakage and micronucleus data. For instance,
nitrobenzene was weakly positive for the induction of chromosome aberrations in cultured
human peripheral lymphocytes but negative in human spermatozoa. Nitrobenzene induced weak
DNA fragmentation but no DNA strand breaks. In addition, nitrobenzene did not cause cell
transformation in these cell systems.
105
-------
Using a weight-of-evidence approach of the mutagenicity study findings, a mutagenic
MO A is not considered a significant contributor to the carcinogenic potential of nitrobenzene.
As discussed in section 3.3, nitrobenzene undergoes reductive and oxidative metabolism,
including generation of free radicals (e.g., nitro anion and superoxide) and propagation of redox
cycling. It is possible that tumors may arise from oxidative stress resulting from nitrobenzene
metabolism if the cellular defenses are overwhelmed as proposed recently by Hsu et al. (2007).
Under oxidative stress conditions, there may be several possible scenarios by which reactive
chemical species (including oxygen radicals) could facilitate tumor development, including
direct DNA oxidative damage, lipid peroxidation, protein damage (including DNA repair
enzymes), or modulation of DNA methylation (Halliwell, 2007). However, there is no
experimental evidence linking any of these processes to nitrobenzene exposure and tumor
formation. Also lacking are actual studies and information on the status of the in situ antioxidant
defenses, especially under similar nitrobenzene exposure conditions that gave rise to tumors.
Demonstrating a correlation among exposure to nitrobenzene, status of antioxidant defenses, and
changes in specific toxicity endpoints characteristic of oxidative stress would be critical to
establishing a link between nitrobenzene-induced carcinogenicity and oxidative stress.
Other possible MOAs by which nitrobenzene may cause tumors, including cytotoxicity
followed by increased cell proliferation resulting in promotion of initiated cells, formation of
DNA adducts, or disruption of intercellular communication, also remain unexplored. Due to the
paucity of information on the MOA of carcinogenicity, the MOA framework described in the
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) has not been applied, the
observed tumors are deemed relevant to humans and a linear low-dose extrapolation as a default
option is applied.
The thyroid and kidney tumors observed in experimental animals may be suggestive of
rodent-specific MOAs; however, as discussed below, experimental data required by EPA's
guidance for excluding these tumors are lacking (U.S. EPA, 1998b; U.S. EPA, 1991b). Tumors
of the rodent thyroid may develop by the following hypothesized MOA. A sustained increase in
conjugating enzymes alters the hypothalamic-pituitary-thyroid axis by increasing the clearance
of T4 (via glucuronidation) and T3 (via sulfation) in rodents, which causes a compensatory
increase in circulating TSH and ultimately follicular cell activation (U.S. EPA, 1998b). Since
the levels of T3 and T4 are tightly regulated in humans, chemicals that cause tumors of the
thyroid via this MOA are not relevant to humans. However, specific data are not available to
support this MOA, such as studies determining the effects of nitrobenzene on circulating blood
levels of TSH, T4, and T3. Since these data are not available for nitrobenzene, the thyroid
tumors are considered relevant for assessing carcinogenic risk to humans (U.S. EPA, 1998b).
Similarly, tubule tumors of the male rat kidney are hypothesized to occur by the following MOA.
After chronic exposure to some chemicals, a,2u-globulin-induced nephropathy may result from
sustained target cytotoxicity and necrosis that leads to increased cell proliferation followed by
106
-------
promotion of spontaneously initiated cells. EPA has determined that the toxicity observed in
rodents via this MOA is not relevant for assessing human risk. However, relevant data are not
available for nitrobenzene.
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.7.1. Possible Childhood Susceptibility
Fetal Hb is more easily oxidized to metHb than adult Hb (Seger, 1992; Goldstein and
Rickert, 1984). The switch ("Hb switching") in the globin chain composition from fetal to adult
Hb (i.e., 0,272 to o^) is nearly complete by 30 weeks postnatal age (Nienhuis and
Stamatoyannopoulos, 1978; Wood, 1976). Therefore, the time period of heightened
susceptibility to methemoglobinemia due to the globin chain composition of Hb spans from
about 6 weeks postconceptual age to about 30 weeks postnatal age (Miller, 2002). However, the
susceptibility of infants and young children persists past this period due to reduced levels of
NAD(P)H, the cofactors for NADPH-cytochrome c reductase, and NADH-cytochrome b5
reductase (Seger, 1992). Wentworth et al. (1999) suggested that newborns are susceptible
because the activity of NADH-cytochrome &s reductase in the RBCs of children is only about
60% that of adults, slowing the reduction of metHb to Hb. Finally, the blood of newborns is low
in glucose-6-phosphate dehydrogenase (G6PD) activity, an enzyme that is crucial for
replenishing NADPH reducing equivalents (see Table 3-5) (Goldstein et al., 1969). Although the
available developmental studies with nitrobenzene were generally negative, metHb levels were
not examined in the offspring (BRRC, 1985, 1984; Biodynamics Inc., 1984, 1983). Hence,
uncertainty exists as to the susceptibility of the test species' Hb to oxidation compared to that of
developing humans.
As indicated by Pinkerton and load (2000), approximately 80% of the human alveoli
develop after birth and continue to develop through early adulthood. This time period for the
developing respiratory system may predispose infants and children to adverse pulmonary effects
from nitrobenzene.
4.7.2. Possible Gender Differences
Nitrobenzene has been shown to cause endometrial polyps in female F344/N rats and
mammary tumors in female B6C3F1 mice. It is not known whether these findings reflect gender
specificity or whether estrogen-responsive tissues (e.g., endometrium and mammary gland) are
targets due to a disturbance of estrogen homeostasis.
In male rats (F344/N and CD) and mice (B6C3F1), nitrobenzene exposure via the
inhalation and oral routes has been shown to cause testicular atrophy, including a dramatic
decrease in sperm count with ensuing loss of fertility. This suggests that nitrobenzene is a male-
specific reproductive toxicant.
107
-------
4.7.3. Other
A review by Harrison (1977) stressed the fundamental difference between hereditary and
chemically induced forms of metHb. There are at least two inherited diseases that affect an
organism's susceptibility to metHb formation (Goldstein et al., 1969). First, genetic deficiency
of NADH cytochrome bs reductase, the enzyme that restores to Hb the small amount of metHb
always being formed in RBCs, imparts a comparatively higher susceptibility to affected
populations upon nitrobenzene exposure. In addition, there is G6PD deficiency (see section
4.7.1), more commonly known because it imparts intolerance to the antimalarial primaquine.
Because the gene for the enzyme is located on the X chromosome, females are usually
heterozygotes and thus not affected by the deficiency. A high frequency of variants of G6PD
deficiency is found in African, Mediterranean, and Asiatic populations (Porter et al., 1962).
Within the U.S., about 13% of African-Americans are affected with the condition. Second,
chemically induced methemoglobinemia can occur from much lower levels of exposure in
patients with comorbidities, such as anemia, cardiovascular disease, lung disease, sepsis, or the
presence of abnormal Hb species (e.g., carboxyhemoglobin, sulfhemoglobin, or sickle cell Hb)
(Goldfrank et al., 1998).
108
-------
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
The RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a
daily oral exposure to the human population (including sensitive subgroups) that is likely to be
without an appreciable risk of deleterious effects during a lifetime. It can be derived from a
NOAEL, LOAEL, or benchmark dose (BMD), with uncertainty factors (UFs) generally applied
to reflect limitations of the data used.
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
There are no lifetime nitrobenzene exposure studies by the oral route, and no human
studies other than case reports of accidental nitrobenzene poisonings are known. Of the animal
studies of oral exposure to nitrobenzene, the 90-day gavage study (10 animals/dose/sex)
conducted by NTP (1983a) is the most relevant study for deriving an RfD for nitrobenzene,
because it is the longest duration study available and used several dose levels. Several other
studies are available but are less suitable for developing a reference value (e.g., reproductive
toxicity studies using a one-time administration or a single dose level (Kawashima et al.,
1995a, b; Levin et al., 1988; Bond et al., 1981) or relatively short duration (Koida et al., 1995;
Matsuura et al., 1995). A 28-day toxicity study (Shimo et al., 1994), a reproductive study
(Mitsumori et al., 1994), and an immunotoxicology study (Burns et al., 1994) were also
considered, as summarized below.
When a well-characterized PBPK model is available, route-to-route extrapolation from a
suitable chronic inhalation study can inform the oral database. As described in sections 4.2.2.2
and 5.2, there is a 2-year inhalation bioassay (Cattley et al., 1994) (see section 5.3). However,
the absence of well-characterized PBPK modeling for nitrobenzene precludes a route-to-route
extrapolation to derive an RfD.
Figure 5-1 is an exposure-response array that presents NOAELs, LOAELs, and the dose
range tested corresponding to selected health effects observed in relevant subchronic and
reproductive oral toxicity studies.
109
-------
350
300
250
(0
•
I, 200
0)
(/)
150
100
50
0
a. Shimo et al.
(1994)
b. NTP (1983a)
c. Mitsumori et
al. (1994)
"ro
il
Subchronic
ro
1
• NOAEL
• LOAEL
The vertical lines represent the range
of doses tested in a given study.
8
8
to
1
8
9
Reproductive
Developmental
«
Figure 5-1. Exposure-response array of selected subchronic and reproductive-developmental toxicity effects by the oral route.
110
-------
Shimo et al. (1994) conducted a 28-day gavage toxicity study of nitrobenzene in F344
rats (six/sex/group) at doses of 0, 5, 25, and 125 mg/kg-day. Animals were evaluated for
generalized signs of toxicity, body and organ weight changes, food consumption, histopathology,
and hematologic and clinical chemistry parameters. There were clinical signs as well as a
marked reduction in body weight increase in the high-dose rats. Organ weight changes were also
observed (spleen and liver were increased and testes were decreased) and were corroborated by
histopathology findings (e.g., spleens had increased congestion, brown pigmentation in red pulp,
and extramedullary hematopoiesis, while testis had epithelium degeneration and atrophy of
seminiferous tubules). However, the most sensitive changes were dose-dependent hematologic
changes, including reductions in RBC count, Hct, and Hb concentration, in addition to increases
in MCV and the WBC count; metHb concentrations were not reported. The NOAEL and
LOAEL for these reported changes were 5 and 25 mg/kg-day, respectively.
Mitsumori et al. (1994) conducted a reproductive toxicity study of nitrobenzene in male
and female Sprague-Dawley rats (10 animals/dose/sex), using 0, 20, 60, and 100 mg/kg for up to
54 days. Because of the experimental protocol used, total nitrobenzene exposure time for most
animals was only 40-41 days. Some effects were observed at the lowest dose (cf. Tables 4-29,
4-30, and 4-31). While this dose was more than twice the lowest dose used in the NTP (1983a)
study, the dose-response relationships of the effects common to both studies were consistent with
those in the NTP study.
Burns et al. (1994) assessed the immunotoxic potential of nitrobenzene for select
immunologic and host resistance responses over a 14-day treatment period. The doses used, 30-
300 mg/kg, were higher than in the NTP (1983a) study, and essentially confirmed the effect of
nitrobenzene on the spleen and hematology parameters. However, effects on the immune system
were mild.
The NTP (1983a) study included both sexes and two species, the F344 rat and the
B6C3F1 mouse; five dose groups plus controls (0, 9.38, 18.75, 37.5, 75, and 150 mg/kg-day for
rats and 0, 18.75, 37.5, 75, 150, and 300 mg/kg-day for mice); and 10 animals/sex/dose group.
In rats, there were seven survivors among the highest dose females but only one survivor among
the highest dose males. In mice, there were no deaths among the highest dose females but three
deaths among the highest dose males. The study reported multiple potentially biologically
significant endpoints, including changes in absolute and relative organ weights, changes in
hematologic parameters, and histopathologic outcomes. The nitrobenzene-induced pathological
changes were much less pronounced in mice than in rats. Since the mice were treated with
higher doses and generally more resistant to nitrobenzene toxicity, the mouse data were not
considered further for RfD evaluation. The similarity of endpoints in both species, however, had
considerable bearing on the choice of critical effect.
Organ weights affected by subchronic nitrobenzene exposure included liver and kidney
(increase) in both sexes and testis (decrease) in male F344 rats (Tables 4-3 and 4-4). The
111
-------
statistically significant increases in liver and kidney weights were generally not supported by
other tissue-specific findings, such as histopathology. Therefore, changes in liver weight were
not considered further. Moreover, kidney weight increases were not considered for risk
evaluation because of the lack of confirmatory tests (e.g., histopathology) and the absence of
kidney effects in nitrobenzene-exposed humans.
There is evidence that nitrobenzene is a male reproductive toxicant (see section 4.3).
However, decreases in testis weight (-40%) were generally seen only at the two highest doses in
rats (75 and 150 mg/kg-day), accompanied by an up to 90% lethality (NTP, 1983a). Similarly, a
decrease in testis weight (-30%) was only observed with the highest dose in male mice
(300 mg/kg-day) with an accompanying 30% mortality. Because of the high doses required to
demonstrate testicular toxicity and the lack of this response in the available human exposure or
poisoning data, this endpoint was not used in the RfD assessment for nitrobenzene, since more
relevant endpoints were identified at lower levels of exposure.
A number of dose-dependent hematologic changes were observed in both species in the
NTP (1983a) study, including hematology-related histopathologic splenic congestion and
increased reticulocyte count. It was assumed that these changes reflected primary or secondary
effects of the nitrobenzene-induced methemoglobinemia (cf Tables 4-5, 4-6, 4-11, and 4-12).
Because methemoglobinemia and splenic congestion have been observed with animal studies
and methemoglobinemia has been observed with most human poisonings, these outcomes were
considered candidate critical effects. Of the other hematology endpoints that were affected with
increasing exposure, reticulocyte levels showed the greatest change, increasing by about 42% in
female rats at the lowest exposure level. The remaining hematology parameters were not
considered further for deriving the RfD because they were considered manifestations of the same
toxicity response and seem to be less sensitive than the endpoints already highlighted.
In summary, the following considerations were used for selecting the 90-day gavage
administration study (NTP, 1983a) from among other toxicity studies based on the approach
outlined in U.S. EPA (2002):
• Route-to-route extrapolation from the inhalation study is not possible in the absence of
well-characterized PBPK modeling for nitrobenzene.
• The NTP (1983a) study has the longest continuous exposure duration via the oral route.
• It included both sexes of rats and mice with five dose groups, plus control, spaced from
9.4-150 mg/kg-day in rats and from 18-300 mg/kg-day in mice.
5.1.2. Method of Analysis—Including Models
BMD software (BMDS), version 1.4. Ic (U.S. EPA, 2007), was used to estimate a point
of departure (POD) for deriving an RfD for nitrobenzene. Although splenic congestion and
changes in reticulocyte counts are considered secondary to the formation of metHb, data on
112
-------
metHb, splenic congestion, and reticulocyte counts were modeled for purposes of comparison
(Table 5-1).
Table 5-1. Summary of effects in F344 rats associated with exposure to
nitrobenzene by gavage for 90 days
Dose
(mg/kg-day)
na
Reticulocytes
(%)"
MetHb
(%)"
Splenic congestion,
severity greater than minimal0
Males
0
9.38
18.75
37.5
75
150
10
10
10
10
10
1
2.23 ±0.44
2.62 ±0.45
3.72±0.65d
4.75 ± 0.62d
6.84 ± 0.72d
15
1.13 ±0.58
2.75±0.58d
4.22 ± 1.15d
5.62±0.85d
7.3 1± 1.44d
12.22
0/10
0/10
0/10
0/10
5/10
10/10
Females
0
9.38
18.75
37.5
75
150
10
10
10
10
10
7
2.60 ±0.37
3.69±0.32d
4.75 ± 0.68d
6.28±0.90d
8.72±1.49d
32.07 ±3.56d
0.94 ±0.03
2.06±0.45d
3.62±1.09d
5.27 ± 0.76d
6.85±2.25d
12.77 ± 1.83d
0/10
1/10
3/10
5/10
8/10
9/10
"Number of animals surviving at terminal sacrifice.
bValues are means ± standard deviations.
°A11 animals, including early deaths, underwent histopathologic examination.
dSignificantly different from controls, as calculated by the authors.
Source: NTP(1983a).
Consistent with the U.S. EPA (2000b) BMD technical guidance, biologically relevant
response levels were considered for these endpoints. There is considerable information in the
literature concerning management and treatment of methemoglobinemia in humans, while
characterization of levels at which cyanosis and other clinical symptoms become apparent vary
across the available literature (see section 4.5.1), commonly falling in the range of 6-10%.
Unfortunately, little information exists concerning the biological significance of particular
metHb levels in rodents and what would correspond to humans, at least regarding relative
biological significance. Information that is available suggests that the normal range in rats may
not necessarily parallel that in humans. Compared to human erythrocytes, rat erythrocytes are
known to have a higher activity of NADH-metHb reductase, the enzyme that spontaneously
regenerates Hb from metHb (Smith, 1996). This difference may indicate that rat erythrocytes are
more efficient at regenerating Hb from metHb, especially following exposure to metHb-forming
agents, although the extent to which this translates into differences in biologically significant
metHb levels in rats and humans is unknown. Of particular note in the NTP (1983a) study of
nitrobenzene, the F344 rats experienced significant mortality when metHb levels increased above
113
-------
10% (see both 150 mg/kg-day groups in Table 5-1). Therefore, projecting metHb levels
associated with clinical symptoms from humans to rats was considered not to be appropriate in
this assessment.
For metHb, a 1 standard deviation (SD) benchmark response (BMR) was considered first.
As detailed in the BMD technical guidance (U.S. EPA, 2000b), a 1 SD BMR provides the
exposure level at which 10% of those exposed would be expected to exceed the 98th (or 2nd)
percentile of the control group's responses. Second, a BMR corresponding to the normal range
of 0-3% in humans was also considered (see section 4.5.1), under the unverified assumption that
this could describe a comparable range in rats. This second approach was implemented by
estimating the exposure level at which 10% of those exposed would be expected to exceed 3%
metHb and would be the preferred approach to selecting a BMR if a normal upper limit for
metHb in rats had been established at 3% (U.S. EPA, 2000b). In addition, a BMR of 2 SDs was
included for comparison.
No information concerning minimally significant increases of splenic congestion or
reticulocytes in rats was found. Because increased reticulocyte counts and splenic congestion are
understood to be sequelae of methemoglobinemia, BMR selection focused primarily on metHb
levels and characterization of normal ranges. BMD modeling carried out for increased
reticulocyte counts and splenic congestion used BMRs of 1 SD and 10% extra risk, respectively,
for comparison with the results from modeling metHb levels.
The endpoints were modeled in terms of administered dose. In the NTP (1983a) study,
the animals were gavaged 7 days/week, thus no adjustment from intermittent to continuous
exposure was required.
The BMD modeling results for metHb levels, reticulocyte count, and splenic congestion
in male and female F344 rats are provided in Appendix B-l and summarized in Table 5-2. For
male rats, the 95% lower bound on the BMD (BMDL) corresponding to 10% of an exposed
population exceeding a 3% metHb level was nearly twice as high as the BMDL corresponding to
a 1 SD increase in the control mean. The BMDL2SD was twofold higher than the BMDLiso
Results for the female rats were in a similar range. As noted in Appendix B-l, however, the
female rat metHb data were somewhat unusual in that there was very little variability in the
control values, and an estimate of the SD was assumed to be the same as that associated with the
low-dose females. Therefore, the female rat metHb results are shown for consistency but were
not considered suitable for deriving an RfD.
114
-------
Table 5-2. Summary of noncancer BMD modeling of selected endpoints from
F344 rats exposed by gavage to nitrobenzene for 90 days, using NTP (1983a)
bioassay data
Endpoint
MetHb
Reticulocyte
count
Splenic
congestion
Sexa
Mb
Fb
Mb
Fb
M
F
Model used
Hill
Hill
Linear
Hill
Multistage
Multistage
p value
0.42
0.41
0.16
0.58
1.0
1.0
BMR
10% extra risk of
exceeding 3% metHb
1 SD
2SD
10% extra risk of exceeding
3% metHb
1 SD
2SD
1 SD
10% extra risk of mild or
moderate congestion
BMD
(mg/kg-day)
4.9
3.0
5.7
8.2
4.7
7.2
9.4
2.7
54.6
7.8
BMDL
(mg/kg-day)
3.2
1.8
3.9
6.3
3.1
5.2
7.9
1.8
37.8
5.6
aM = male, F = female.
bHighest dose not included in BMD modeling.
The results of modeling reticulocyte counts and splenic congestion fall in a range similar
to that of the metHb results, reinforcing the latter results, with the exception of the male rat
splenic congestion. The BMDL was roughly 10-fold higher than that for splenic congestion in
the female rats and for the other endpoints (BMDLisos) but is consistent with splenic congestion
being secondary to methemoglobinemia.
Based on the considerations above, the POD for developing an RfD was the BMDLiso
derived from the male rat metHb data. The BMDL corresponding to 10% of an exposed
population exceeding 3% metHb was roughly twofold higher but not clearly linked to biological
significance for rats. The use of a 1 SD BMR in developing an RfD for humans engenders an
assumption of analogous variability between rats and humans, i.e., that humans would be
expected to demonstrate a 10% extra risk of exceeding the normal range of an unexposed
population. This does not necessarily imply that the mean metHb level would be shifted by an
amount equal to the rat SD (0.5%) in a human population exposed at the BMDL.
5.1.3. RfD Derivation — Including Application of Uncertainty Factors (UFs)
The RfD for nitrobenzene was calculated as follows:
BMDL -T- UF = RfD
1.8 mg/kg-day •*• 1,000 = 2 x 10~3 mg/kg-day
The composite UF of 1,000 follows from considering these areas of uncertainty and
variability:
115
-------
An intraspecies UF of 10 was applied to account for human variability and to protect
potentially sensitive humans (e.g., G6PD deficiency or chronic congenital
methemoglobinemia) and life stages (e.g., children). The default value was selected in
the absence of information, indicating the degree to which humans might vary in
susceptibility to nitrobenzene toxicity.
An interspecies UF of 10 was applied for extrapolation from animals to humans. No
suitable data on the toxicity of nitrobenzene to humans exposed by the oral route were
identified. Insufficient information is currently available to assess rat-to-human
differences in nitrobenzene toxicokinetics or toxicodynamics.
A UF to account for the extrapolation from a LOAEL to a NOAEL was not applied
because the BMR of 1 SD was determined to represent a minimal biologically significant
level of change.
A subchronic-to-chronic UF of 3 was applied to account for less-than-lifetime exposure
in the principal study. A chronic oral study is not available. In studies by the inhalation
route, the severity of hematologic effects (e.g., metHb, reticulocyte count, and splenic
congestion) did not increase between subchronic (CUT, 1984) and chronic (CUT, 1993)
exposure durations (see section 4.5.2). Nonetheless, other toxicity endpoints may result
from chronic oral exposure due to route-specific differences in metabolism,
pharmacokinetics, and/or pharmacodynamics that were not observed in the subchronic
oral study or the inhalation studies. In particular, several studies of gut bacterial
metabolic activation (nitro reduction) support the possibility of higher relative
concentrations of active metHb-forming metabolites than would be expected following
exposure by the inhalation route (see section 3.3.1).
An UF of 3 for database deficiencies was applied. The database of oral studies includes
the principal study, a 90-day gavage study in two species and both sexes (NTP, 1983b); a
reproductive/developmental study (Mitsumori et al., 1994) and two male reproductive
toxicity studies (Morrissey et al., 1988; Bond et al., 1981); structure-activity relationship
studies with dinitro- and trinitrobenzene; and a multidose immunological study in mice
(Burns et al., 1994). Due to the lack of an oral multigeneration reproductive toxicity
study and evidence of male reproductive toxicity, a factor of 3 is warranted. There is a
two-generation reproductive toxicity study (Dodd et al., 1987) via inhalation exposure,
but possible route-specific differences in metabolism, pharmacokinetics, and/or
pharmacodynamics suggest uncertainty in the potential for transgenerational effects from
longer-term oral exposures.
116
-------
5.1.4. Previous RfD Assessment
The previous IRIS assessment based the RfD for nitrobenzene of 5 x 1CT4 mg/kg-day on a
90-day inhalation study in F344 rats and B6C3F1 mice (CUT, 1984). Critical endpoints included
methemoglobinemia and histopathologic lesions to the adrenal gland, kidney, and liver. A route-
to-route extrapolation was performed, and the LOAEL-NOAEL approach was used to derive the
RfD. A POD of 25 mg/m3 (LOAEL) was identified and converted to an equivalent oral dose of
4.6 mg/kg-day by using default assumptions for mouse breathing rate and body weight. A
combined UF of 10,000 was applied, resulting in an RfD of 5 x 10^ mg/kg-day.
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
There are no studies in humans that investigate outcomes of long-term inhalation
exposure to nitrobenzene combined with quantitative measures of exposure. However, there are
animal studies that examine inhalation effects of nitrobenzene in rats and mice with short-term
exposure (Medinsky and Irons, 1985; DuPont, 1981), subchronic exposure (CUT, 1984), and 2-
year chronic exposure (Cattley et al., 1994; CUT, 1993). Noncancer effects of inhalation
exposure to nitrobenzene were generally similar to those observed following oral exposure
(methemoglobinemia, altered hematology with signs of hemolytic anemia, damage to the male
reproductive system, changes in relative organ weights, and pigment deposition in organs). In
addition, portal-of-entry effects following chronic inhalation exposure to nitrobenzene included
bronchiolization of the alveoli and olfactory degeneration in both male and female B6C3F1 mice
(CUT, 1993). Pulmonary effects were also observed in subchronic inhalation studies in both
F344 rats and B6C3F1 mice (CUT, 1984). The DuPont (1981) and Medinsky and Irons (1985)
studies were not considered as principal studies for RfC derivation because both studies had
short exposure times (14 days) and comparatively high levels of exposure (10-125 ppm
nitrobenzene). Figure 5-2 is an exposure-response array that presents NOAELs, LOAELs, and
the dose range tested corresponding to selected health effects observed in chronic, subchronic,
and reproductive inhalation toxicity studies of nitrobenzene.
A 90-day subchronic study was conducted by using both sexes of F344 and CD rats as
well as B6C3F1 mice (CUT, 1984). Exposure concentrations were 0, 5, 16, or 50 ppm,
6 hours/day, 5 days/week. The treatments had no effect on body weights, but spleen weights
were increased and testis weights were decreased in rats. Signs of hemolytic anemia were
evident in rats while methemoglobinemia was consistently observed in both species
(Table 4-18). Pulmonary effects were also observed in F344 rats and B6C3F1 mice. In male
F344 rats, 60% of the animals in the 50 ppm group exhibited bronchiolar epithelial hyperplasia,
whereas 20% of females were found with this lesion. In B6C3F1 mice, bronchial mucosal
hyperplasia was observed in 78% of males and 100% of females at 50 ppm.
117
-------
100 n
90
80
Q. 70
Q.
I 60
"ro
1 50
o
8 40
30
20
10
0
• NOAEL
• LOAEL
The vertical lines represent the range
of doses tested in a given study.
Subchronic Chronic
Subchronic
4
> 4
|
4
4
> 1
» 4
1 1
»
1 4
•
|
• J
• 4
ii
§
984) |
QQO\ -&
yyoj o
ill. (1987) |
r^S^-rs-I£ £ I s
I
Developmental
Chronic
> 4
- i
i
i
9 •=;
^, ^
4
1
»
1
4
k
|
1
1 1
1 1
1
1 1
1
. . ^ ^nrn-i-'O^^
1 1 1 I | I | 1 I 1 1 1 1 | 1 I 1 1 i 1 I i |
t 1 ! I 1 1 1 ! ! Itftill ? | I s ! i 1
d. BRRC(1985);
Doddetal.(1987)
e. Biodynamics
(1983,1984)
^ ,
|! ^ £
S S J.
,
g
ro
_Q
ro
_Q
Figure 5-2. Exposure-response array of selected subchronic, chronic, and reproductive toxicity effects by the inhalation route.
118
-------
The 2-year study, also conducted by CUT (Cattley et al., 1994; CUT, 1993), is the most
suitable study for derivation of an RfC because of the chronic exposure duration and large group
sizes (70 animals/sex/group). The study used B6C3F1 mice and F344 rats of both sexes and
male CD rats. Rats were exposed to 0, 1, 5, or 25 ppm nitrobenzene and mice to 0, 5, 25, or
50 ppm nitrobenzene for 6 hours/day, 5 days/week (for details see section 4.2.2.2). Animals
were sacrificed at 24 months of exposure, and blood analyses and complete necropsies were
performed. Ten rats/sex/strain/group were terminated 15 months into the study to provide
samples for an interim evaluation of hematologic parameters. According to the study authors,
and as noted under section 4.2.2.2, the proportion of study animals surviving to study termination
was not statistically significantly affected by exposure to nitrobenzene.
Cattley et al. (1994) identified the following target tissues: thyroid, spleen, nose, and liver
in all strains and species; kidney in rats only; and respiratory tissues in mice only. Testis and
epididymis were target tissues in male CD rats. A statistically significant difference in the
incidence of centrilobular hepatocytomegaly was observed in a concentration-dependent fashion
in both strains of male rats but not in female rats. The incidence of renal tubular hyperplasia in
male F344 rats showed a statistically significant positive trend. Chronic nephropathy and tubular
hyperplasia were observed in both males and females. Bilateral testicular atrophy was reported
with effects appearing in the high-concentration group only in both male CD and F344 rats.
Bilateral hypospermia was observed in high-concentration male CD rats.
At interim sacrifice, a statistically significant increase in metFIb was observed at all
exposure levels in male CD rats and only at the highest concentration with male and female F344
rats. At terminal sacrifice, a statistically significant increase in metFIb was observed with both
sexes of mice at the highest concentrations tested. An approximate twofold increase in metFIb
was observed with male and female B6C3F1 mice, female F344 rats, and male CD rats, whereas
an approximate 1.5-fold increase was observed with male F344 rats (Cattley et al., 1994; CUT,
1993). Hct and Fib were reduced only in female mice, being statistically significantly different at
the 5 ppm concentration and lower concentrations, albeit still statistically significantly reduced at
25 ppm but not at 50 ppm. Since this effect occurred only in female mice and did not exhibit
concentration dependency, it was considered to not be treatment related because of the lack of a
dose response.
Exposure-related degeneration and loss of olfactory epithelium were observed in mice of
both sexes, with the females being more sensitive than the males. At the highest concentration
tested (50 ppm), the incidence was 62% in males and 69% in females. Bronchiolization of the
alveoli was also observed at all concentrations in both sexes, with 94% incidence in males and
100% incidence in females at the highest concentration tested. Follicular cell hyperplasia of the
thyroid was observed in both sexes of mice, with males being more sensitive than females. At
the highest concentration, this response was reported in 19% of the males. Exposure-related
hepatocellular changes (e.g., centrilobular hepatocytomegaly) were observed in males, with
119
-------
incidence up to 89% at the highest concentration, and occurred in 11% of females only at the
highest concentration. Hypercellularity of the bone marrow, an effect secondary to hemolytic
anemia, was recorded for males in a concentration-dependent fashion with low incidence; in
females, only animals exposed at the highest concentrations were examined for this effect, and
the response was even lower than in males. There was also evidence for testicular toxicity in
males, but only the high-concentration animals were examined.
The most sensitive effects observed following nitrobenzene exposure were degeneration
and loss of the olfactory epithelium and bronchiolization of the alveoli in mice. Degeneration
and loss of the olfactory epithelium occurred in a concentration-dependent manner, with high
incidences (>62%) at the highest exposure in both males and females, while females were more
sensitive than males at the lowest exposure, with 19/60 females responding versus 1/66 males
(Table 5-3). The study report (CUT, 1993) indicated that the severity increased with increasing
exposure but provided no further details.
Table 5-3. Incidence of histopathologic lesions in mice following chronic
nitrobenzene inhalation
Histopathologic lesion
Olfactory epithelium
degeneration, lossb
Bronchiolization of the
alveolib
Sexa
M
F
M
F
Exposure level (ppm)
0
1/67
0/52
0/68
0/53
5
1/66
19/60
58/67
55/60
25
32/65
47/63
58/65
63/64
50
41/66
42/61
62/66
62/62
aM = male, F = female.
bSignificant positive trend by Armitage-Cochran test in both sexes, p < 0.05.
Sources: Cattley et al. (1994); CUT (1993).
As discussed in section 4.5.2, exposure-related olfactory changes in the nasal passages
were also observed in both F344 and CD rat strains in the CUT (1993) study, although the
changes were less marked than in mice, and in the olfactory bulbs of rats exposed to
nitrobenzene by inhalation in Beauchamp et al. (1982). Olfactory degeneration was considered
as a candidate critical effect for the derivation of the RfC.
Bronchiolization of the alveoli occurred with high incidence (>87%) in both males and
females in the exposed groups (Table 5-3). The lesions were characterized by a pronounced
change in the alveolar epithelium in the region of the terminal bronchioles from a simple
squamous to a tall columnar epithelium resembling that of the terminal bronchioles. According
to the report, the change was concentration related in severity; however, no additional
information on severity was reported (CUT, 1993). In the low-concentration exposed animals,
bronchi olization was located almost entirely in the region of the terminal bronchioles. In the
mid- and high-concentration animals, the lesions were more florid and involved a large
120
-------
proportion of the lung parenchyma, with animals in the mid-concentration group being slightly
less affected than the high-concentration animals.
As discussed in section 4.5.2, bronchiolization of the alveoli is a histologically distinct
lesion that has been seen in various species, including mice and humans, and that may indicate a
variety of pathological conditions, including inflammation, chemical irritation, or exposure to
carcinogens. Bronchiolization was also considered as a candidate critical effect for the
derivation of the RfC.
Methemoglobinemia was not chosen as a candidate critical endpoint for the inhalation
RfC. As explained in section 4.2.2.2, the biological significance of the hematologic findings,
including methemoglobinemia, in the chronic inhalation study (Cattley et al., 1994) is unclear.
In several instances, the differences between the dosed groups and the controls were minimal or
decreased with increasing length of exposure. In most instances, methemoglobinemia was
notably increased only at the highest nitrobenzene exposure, while time-related trends were not
clear-cut due to a possible compensatory response among all exposed rat groups (Table 4-20).
5.2.2. Methods of Analysis—Including Models
BMD modeling (U.S. EPA, 2000b) was used to analyze the incidence data for
bronchiolization of the alveoli and olfactory degeneration from CUT (1993) as shown in
Table 5-3. All of the available dichotomous models in U.S. EPA's BMDS (version 1.4.1c) were
fit to the incidence data for bronchiolization of the alveoli and olfactory degeneration.
Consistent with U.S. EPA (2000b) BMD technical guidance, consideration was given to
identifying biologically relevant response levels for developing RfDs. Insufficient information
was available to identify minimally adverse levels of response for either bronchiolization of the
alveoli or olfactory degeneration. A BMR of 10% is generally used to facilitate a consistent
basis of comparison across assessments and was used for these endpoints in the absence of
information regarding the level of change considered to be biologically significant.
The BMD modeling results for bronchiolization of the alveoli and olfactory epithelium
degeneration are summarized in Table 5-4. Detailed model output is in Appendix B-2.
121
-------
Table 5-4. Modeling results for bronchiolization of the alveoli and olfactory
degeneration in mice
Endpoint
Olfactory epithelium
degeneration, loss
Bronchiolization of
the alveoli
Sex
Males
Females3
Males
Females
Model
Probit
Gamma, multistage (1°), Weibull
Log-logistic3
Log-probit
Log-logistic
p Value
0.38
0.50
0.028
0.12
0.80
BMC10
(ppm)b
12.3
1.75
0.13
0.40
0.18
BMCL10
(ppm)b
10.0
1.42
0.083
0.28
0.022
"High-dose group excluded.
bBMC = benchmark concentration; BMCL = 95% lower bound on the BMC.
Data sources: Cattley et al. (1994); CUT (1993).
Dose-response modeling provided satisfactory descriptions of the olfactory epithelium
degeneration data (with adequate goodness-of-fit/> values > 0.1 and low chi-squared residuals).
The greater sensitivity in female mice noted above was estimated to correspond to an
approximate 10-fold difference between the 95% lower bound on the benchmark concentrations
(BMCLios) for male and female mice in this study, at 10 ppm for males versus 1.4 ppm for
females.
Modeling was not quite as successful for the bronchiolization data. The male mice data
only supported a model with a goodness-of-fit/? value of 0.02 for the best fit. The female mice
data supported two adequate but somewhat equivocal fits. That is, the slightly better fitting
model (log-logistic; better fit owing to a lower chi-squared residual at the lowest dose) led to a
BMCLio (0.022 ppm) about 10-fold lower than its benchmark concentration (BMCio)
(0.18 ppm), while the only other model providing an adequate fit to these data (the log-probit
model) yielded a more precise BMCLio/BMCio range of 0.28/0.40 ppm, both measures slightly
higher than the first model's BMCio.
The male mice bronchiolization data were problematic to fit with available models
because the response at the low end of the plateau of responses was underestimated. Further,
there is no way to tell how far into the lower exposures the plateau really extends, since most of
the dose-response relationship has not been captured. It is therefore possible that a BMCLgo, for
example, more closely fitting the observed data could be overestimated, aside from the accuracy
of any extrapolations to lower doses. In addition, despite the relatively better fitting models for
the female mice, these data are as uncertain as for the males regarding the extent of the response
plateau.
On the other hand, the study report noted that the severity of bronchiolization was dose
related, increasing with increasing exposure. Because no data on the quantification of severity
was provided in the study report, it is unknown what impact the consideration of such data might
have on the dose-response relationship, although it is plausible that the reported low-exposure
122
-------
response was overstated relative to that at the higher exposures. Given these two divergent
possibilities, use of the modeled response appears to be a reasonable compromise.
Because of the uncertainty evident in modeling the bronchiolization data, the NOAEL/
LOAEL approach was also considered for estimating the POD. The highest BMCLio estimated,
0.28 ppm, is about twofold lower than the LOAEL (5 ppm) divided by the LOAEL-to-NOAEL
UF of 10, or 0.5 ppm. A LOAEL at an 87% response level, however, is arguably less
informative than a BMCLio for characterizing exposure response in the low-exposure region,
since it is not clear that a 10-fold UF applied to such a high response can provide an adequate
estimate of a minimally biologically significant response level.
Therefore, the BMD approach was used for characterizing a POD for the bronchiolization
data. An average of the three BMCLi0s— 0.1 ppm—was considered as one POD for developing
the RfC, while acknowledging the uncertainty associated with the somewhat extreme degree of
extrapolation to estimate a BMCLio from either the male or female mice bronchiolization data.
5.2.3. Evaluation of Human Equivalent Concentrations
While the dose-response modeling shows bronchiolization to be a more sensitive
response in mice than olfactory degeneration, the BMCLioS for both endpoints were considered
for extrapolation to human equivalent concentrations (HECs), since this process can have
different impacts depending on the respiratory sites affected. This process involves two main
steps, adjustment to equivalent continuous lifetime exposures, followed by adjustment to human
equivalents.
Because the RfC is a metric that addresses continuous human exposure for a lifetime,
adjustments need to be made to animal data obtained from intermittent and/or less-than-lifetime
exposure scenarios, as supported in the Methods for Derivation of Inhalation Reference
Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b). The first step is
adjustment of the intermittent inhalation exposure to continuous exposure, based on the
assumption that the product of exposure concentration and exposure time is constant, in the
absence of information to the contrary (U.S. EPA, 2002). In the chronic studies (Cattley et al.,
1994; CUT, 1993), animals were exposed for 6 hours/day, 5 days/week. Therefore, the POD
adjusted for continuous exposure (PODAoj) for inhalation of nitrobenzene is as follows:
PODADJ = POD (in ppm) x 6 hours/24 hours x 5 days/7 days (5-1)
Furthermore, because the RfC is expressed in mg/m3, the POD in units of ppm needs to be
converted to mg/m3 by using the conversion factor for nitrobenzene of 1 ppm = 5.04 mg/m3.
EPA guidance for RfC derivation provides procedures for determining an HEC from the
PODADJ obtained from animal data (U.S. EPA, 1994b). The approach considers the
physicochemical characteristics of the gas or vapor in question as well as the toxicological
123
-------
specifics of the target tissue (respiratory versus systemic and, in the former case, extrathoracic,
thoracic, tracheobronchial, or pulmonary). The effects considered, bronchiolization and
olfactory degeneration, were pulmonary and extrathoracic effects, respectively. Nitrobenzene
qualifies as a category 2 gas: moderately water soluble, reactive in respiratory tissue, and
lexicologically active at remote sites (U.S. EPA, 1994b). For category 2 gases, HEC values are
calculated by using methods for category 1 gases for portal-of-entry effects and category 3
methods for systemic effects (U.S. EPA, 1994b). The olfactory degeneration is more clearly a
portal-of-entry effect; hence, the method for category 1 gases was used to derive its HEC.
Because the bronchiolization occurred on the boundary with systemic circulation, HECs
consistent with this endpoint being either a portal-of-entry or a systemic effect were estimated.
EPA's Methods for Derivation of Inhalation Reference Concentrations and Application
of Inhalation Dosimetry (U.S. EPA, 1994b) suggests that HECs be estimated by applying to the
duration-adjusted exposure level, here the PODAoj, a factor that is specific for the affected region
of the respiratory tract and the breathing characteristic of the species to be compared. This
factor, the regional gas dose ratio (RGDR), as detailed in the RfC guidance (U.S. EPA, 1994b) is
determined for the pulmonary and extrathoracic regions as follows:13
RGDRpu or ET = (MVa/Sa,PU or EX) - (MVh/Sh,PU or EX) (5-2)
where
MVa = minute volume for mice = 0.06 mVday (see Appendix B-2)
MVh = minute volume for humans = 20 m3/day
Sa,pu = default pulmonary surface area for mice = 0.05 m2
Sh,pu = default pulmonary surface area for humans = 54 m2
SO,EX = default extrathoracic surface area for mice = 3 cm2
Sh,Ex = default extrathoracic surface area for humans = 200 cm2
Finally, the PODHEc is derived as follows:
PC-DHEC = PODADJ x RGDR (5-3)
13 The equation for category 1 gases for portal-of-entry effects in the pulmonary region is more complicated, but the
additional factors for extrathoracic and tracheobronchial regions are very close to 1.
124
-------
Olfactory Degeneration
The PODHEc for olfactory degeneration using the methods described above is estimated
as follows:
PODAoj = 1.42 ppm x 6 hours/24 hours x 5 days/7 days x 5.04 (mg/m3)/ppm
= 1.275mg/m3
Substituting the PODADJ and appropriate values for surface area for the extrathoracic
region into equation 5-2, the RGDRPU is calculated as follows:
RGDRpu = (0.06 m3/day)/(3 cm2) •*• (20 m3/day)/(200 cm2) = 0.2
Finally, the PODHEc is calculated as follows:
PODHEc = 1.275 mg/m3 x 0.2 = 0.26 mg/m3
Bronchiolization of the Alveoli
The PODHEc for bronchiolization of the alveoli using the methods described above is
estimated as follows:
PODAoj = 0.1 ppm x 6 hours/24 hours x 5 days/7 days x 5.04 (mg/m3)/ppm
= 0.09 mg/m3
Substituting the PODAoj and appropriate values for surface area for the pulmonary region
into equation 5-2, the RGDRPU is calculated as follows:
RGDRpu = (0.06 m3/day)/(0.05 m2) •*• (20 m3/day)/(54 m2) = 3.24
Finally, the PODHEc is derived as follows:
PODHEc = 0.09 mg/m3 x 3.24 = 0.29 mg/m3
Alternatively, under the assumption that bronchi olizati on of the alveoli might be a
systemic effect, interspecies extrapolation to an FtEC involves consideration of the nitrobenzene
airblood partition coefficients for humans and rats (U.S. EPA, 1994b). In the absence of such
data, as in this case, the ratio of animal to human airblood partition coefficients is assumed to be
unity. The PODHEc = 0.09 mg/m3.
125
-------
5.2.4. RfC Derivation—Including Application of Uncertainty Factors (UFs)
Because the PODHEcs for the two respiratory effects—bronchiolization of the alveoli and
olfactory degeneration—were similar in value, at 0.29 and 0.26 mg/m3, respectively, the effects
were considered co-critical effects for purposes of deriving the RfC and the lower of the two
values, 0.26 mg/m3, was chosen as the POD.
The RfC was calculated based on the PODnEc values for these co-critical effects by
application of UFs as follows:
RfC = PC-DHEC - UF (5-4)
RfC = 0.26 mg/m3 •*• 30 = 0.0085 mg/m3 = 9 x 10~3 mg/m3
The composite UF of 30 follows from considering these areas of uncertainty and
variability:
• An intraspecies UF of 10 was applied to account for human variability and to protect
potentially sensitive humans and life stages (e.g., children). The default value was
selected in the absence of information indicating the degree to which humans might vary
in susceptibility to nitrobenzene toxicity.
• A UF of 3 was applied to account for uncertainty in extrapolating from laboratory
animals to humans. This value is adopted by convention, where a dosimetric adjustment
from an animal-specific PODADJ to a PODnEc already has been incorporated.
Application of a full UF of 10 would depend on two areas of uncertainty (i.e.,
toxicokinetic and toxicodynamic uncertainties). In this assessment, the toxicokinetic
component is mostly addressed by the determination of an FtEC as described in the RfC
methodology (U.S. EPA, 1994b). The toxicodynamic uncertainty is also accounted for to
a certain degree by the use of the applied dosimetry method.
• A UF to account for extrapolation from a LOAEL to a NOAEL was not used because the
current approach is to address this extrapolation as one of the considerations in selecting
a BMR for BMD modeling. In this case, a BMR of a 10% change in either of the
respiratory effects was selected under an assumption that it represents a minimal
biologically significant change.
• A subchronic-to-chronic UF for extrapolation to lifetime exposure was not applied since
the data used originated from a 2-year (lifetime) chronic study.
• A UF of 1 was applied to account for database deficiencies. The inhalation database is
considered complete because it includes developmental toxicity studies in rats (Tyl et al.,
1987) and rabbits (Biodynamics Inc., 1984), a two-generation reproduction study in rats
(Dodd et al., 1987), and a 2-year toxicity study in mice and two strains of rats (Cattley et
126
-------
al., 1994; CUT, 1993) in addition to short-term toxicity studies in mice and two strains of
rats (Medinsky and Irons, 1985; DuPont, 1981).
5.2.5. Previous RfC Assessment
An inhalation assessment was not provided in the previous IRIS evaluation of
nitrobenzene.
5.3. CANCER ASSESSMENT
No studies exist on the carcinogenicity of nitrobenzene in humans. In animals, there is
no cancer bioassay available following oral administration of nitrobenzene, but there is a single
chronic inhalation cancer bioassay that supported quantitative cancer assessment.
5.3.1. Choice of Principal Study and Target Organ—with Rationale and Justification
A 2-year inhalation cancer bioassay (CUT, 1993, published as Cattley et al. [1994]) was
used for development of an inhalation unit risk (IUR) for nitrobenzene. In this study, both sexes
of F344 rats and B6C3F1 mice, along with male CD rats, were exposed to nitrobenzene for 2
years via inhalation (CUT, 1993, published as Cattley et al. [1994]). Ranges of exposure
concentrations were selected based on a subchronic study by the same route. Rats were exposed
to 0, 1, 5, or 25 ppm nitrobenzene and mice to 0, 5, 25, or 50 ppm nitrobenzene for 6 hours/day,
5 days/week (see section 4.2.2.2 for additional details).
Increased incidences of neoplasms with increasing nitrobenzene exposure were observed
in both mice and rats. Adenoma and carcinoma incidences within each site were combined by
counting animals with either of these responses. This practice was performed under the
assumption that adenomas and carcinomas originating from the same cell type represent stages
along a continuum of carcinogenic effects resulting from the same mechanism, as recommended
by the EPA cancer guidelines (U.S. EPA, 2005a).
For example, hepatocellular adenomas or carcinomas were consistently seen in males of
both rat strains (i.e., F344 and CD) and also in female F344 rats. The incidence of these
neoplasms in male CD rats was lower than in male F344 rats. Table 5-5 summarizes the
incidences of hepatocellular neoplasms, by type and combined, among terminally sacrificed
F344 rats. The hepatocellular adenomas were described as being spherical, sharply demarcated,
and compressed to surrounding normal parenchyma, and they consisted of well-differentiated
hepatocytes arranged in sheets or irregular cords. Hepatocellular carcinomas were generally
larger and more irregular than the hepatocellular adenomas, caused a marked compression of
surrounding parenchyma, and had a distinctive feature of a trabecular (rod-shaped) pattern. The
carcinomas were composed of pleomorphic cells arranged in sheets or thickened disorganized
cords (Cattley et al., 1994). Based on the overlapping characteristics of the two forms of
hepatocellular tumors, it was judged reasonable to analyze both forms together.
127
-------
Table 5-6 presents an overview of the tumor incidence data from Cattley et al. (1994).
Nitrobenzene caused an increased incidence of neoplasms in the respiratory tract and in follicular
cells of the thyroid in male B6C3F1 mice, as well as an elevated incidence of liver neoplasia in
female B6C3F1 mice. A slightly elevated incidence of thyroid neoplasia, without strong
evidence of a dose response, was also observed in female B6C3F1 mice. Significant dose-
related trends (atp < 0.05 in the Cochran-Armitage test) were observed for lung adenomas or
carcinomas and thyroid follicular cell adenomas in male B6C3F1 mice and for hepatocellular
adenomas in female B6C3F1 mice.
Table 5-5. Hepatocellular neoplastic findings in F344 rats exposed to
nitrobenzene via inhalation for 2 years
F344 rat
Adenoma3
Carcinoma3
Adenoma3 or carcinoma
Nitrobenzene concentration (ppm)
Male
0
1/43
0/43
1/43
1
3/50
1/50
4/50
5
3/47
2/47
5/47
25
15/46
4/46
16/46
Female
0
0/49
0/49
0/49
1
2/50
0/50
2/50
5
0/50
0/50
0/50
25
3/49
2/49
4/49
3Liver tumor incidences are based on terminal sacrifice data.
Sources: Cattley et al. (1994); CUT (1993).
128
-------
Table 5-6. Selected cancer incidences in B6C3F1 mice, F344 rats, and CD
rats following 2-year inhalation exposure to nitrobenzene
Species, sex,
strain
Mouse, male,
B6C3F1
Mouse, female,
B6C3F1
Target organ,
tumor type a
Lung, bronchio-alveolar
adenoma or carcinoma
Thyroid, follicular
cell adenoma
Liver,
hepatocellular adenoma
Rat, male, F344
Rat, male, CD
Rat, female, F344
Rat, male, F344
Liver,
hepatocellular adenoma or
carcinoma
Uterus, endometrial
stromal polyp
Kidney, tubular
adenoma or carcinoma
Thyroid, follicular cell
adenoma or carcinoma
Nitrobenzene concentration (ppm)
0
8/42
(19.0%)
0/41
(0%)
4/31
(12.9%)
5
16/44
(36.4%)
4/44
(9.1%)
4/38
(10.5%)
25
20/45
(44.4%)
1/45
(2.2%)
5/46
(10.9%)
50
21/48
(43.8%)
6/46
(13.0%)
11/34
(32.4%)
Nitrobenzene concentration (ppm)
0
1/43
(2.3%)
0/23
(0%)
9/48
(18.8%)
0/43
(0%)
1/43
(2.3%)
1
4/50
(8.0%)
0/23
(0%)
15/50
(30.0%)
0/50
(0%)
1/50
(2.0%)
5
5/47
(10.6%)
1/25
(4.0%)
14/50
(28.0%)
0/47
(0%)
5/47
(10.6%)
25
16/46
(34.8%)
5/23
(21.7%)
19/49
(38.8%)
6/46
(13.0%)
8/46
(17.4%)
aAll incidences shown have significant dose response trends atp < 0.05 (Cochran-Armitage test).
Source: CUT (1993).
Statistically significant increasing trends were reported for hepatocellular adenomas or
carcinomas in male and female F344 rats and male CD rats, endometrial stromal polyps in
female F344 rats, and kidney and thyroid follicular cell adenomas or carcinomas in male F344
rats. However, kidney tubular adenomas or carcinomas in male F344 rats were observed only at
the highest dose. Moreover, no corresponding renal neoplasia occurred in female F344 rats or
male CD rats in the same study. The incidence data for uterine endometrial stromal polyps in
female F344 rats, a common benign lesion in this rat strain (NTP historical controls = 11.6%),
displayed a high incidence in controls (18.8%), but there was still some evidence of a dose
response. However, these data were not modeled because there was no evidence or concern that
nitrobenzene exposure would be associated with more severe neoplasia at this site.
The strongest dose response for hepatocellular adenomas or carcinomas occurred in male
F344 rats; therefore, this data set and the data sets for kidney and thyroid adenomas or
carcinomas in male F344 rats were chosen for cancer dose-response assessment. In addition,
thyroid and lung adenomas or carcinomas in male B6C3F1 mice were considered for cancer
dose-response assessment. As indicated in section 4.6.2, the relevance of hepatocellular
neoplasia in rats to the overall assessment of nitrobenzene carcinogenicity in humans is
strengthened by the facts that the spontaneous incidence of hepatocellular neoplasia is higher in
129
-------
B6C3F1 mice than in F344 or CD rats (Cattley et al., 1994) and that B6C3F1 mice have been
known to more likely develop hepatocarcinogenicity in response to chemical exposures than do
rats (Goodman et al., 1985). While the liver tumors in the male F344 rat appear to be the most
sensitive cancer endpoint, ITJRs were also calculated for the thyroid and kidney tumors in order
to characterize potential total risk (see sections 4.6.2 and 4.6.3 concerning the human relevance
of these tumors).
5.3.2. Benchmark Concentration Modeling
Because there are no biologically based dose-response models suitable for the tumor data
identified above, these data were modeled by using the multistage model, as implemented by
BMDS 1.4.1c (U.S. EPA, 2007). This model has the following form:
P(d) = 1 exp[-(q0 + qid + q2d2
where P(d) represents the lifetime risk (probability) of cancer at dose d (i.e., human equivalent
exposure in this case), and q; > 0 (for i = 0, 1, ..., k) are parameters estimated in fitting the model.
A 10% BMR was used with each tumor type (U.S. EPA, 2005a). BMC modeling results
are shown in Appendix B-3.
Male rats were slightly more sensitive to nitrobenzene carcinogenicity, displaying lower
values for the 95% lower bound at 10% extra risk (BMCLio) than male mice. Consequently, the
modeling results for male rats form the basis for IUR derivation, with support from the male
mouse data. Table 5-7 shows the estimated BMCi0s, BMCLi0s, and chi-square/> values derived
for the three tumor types modeled for male F344 rats.
Table 5-7. Estimated BMCs and BMCLs based on tumor incidence data in
male F344 rats exposed to nitrobenzene via inhalation
Target
organ
Kidney
Thyroid
Liver
Tumor type"
Tubular adenoma or carcinoma
Follicular cell adenoma or carcinoma
Hepatocellular adenoma or carcinoma
BMC10
(ppm)
22.8
13.6
6.8
BMCL10
(ppm)
16.8
7.8
4.4
%2p value for
lack of fit
1.0
0.37
0.63
aSee Appendix B-3.3 for modeling results.
5.3.3. Inhalation Dose Adjustments, Inhalation Unit Risk, and Extrapolation Methods
The current Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) stipulate that
the method used to characterize and quantify cancer risk from a chemical is determined by what
is known about the MOA of the carcinogen and the shape of the cancer dose-response curve at
low dose. The dose response is assumed to be linear in the lowest dose range, when evidence
130
-------
supports a mutagenic MOA because of DNA reactivity or if another MOA that is anticipated to
be linear is applicable. An assumption of nonlinearity is appropriate when the MOA
theoretically has a threshold (e.g., when the carcinogenic action is secondary to another toxic
effect that itself has a threshold). Low-dose extrapolation could also include other nonlinear
modeling approaches if indicated by the MOA or other data. If the MOA of carcinogenicity is
not adequately understood, a linear dose-response relationship at low doses is assumed and the
linear extrapolation is used (U.S. EPA, 2005a).
The available evidence suggests that nitrobenzene is not, or is at most weakly, mutagenic
(see section 4.4.5). In addition, nitrobenzene has been shown to undergo redox cycling (see
section 3.3) with the possibility that it may cause oxidative stress (see section 4.6.3). This
process can cause DNA damage and is also thought to be cytotoxic. However, as described in
section 4.6.3, the data available on the role of redox cycling and oxidative stress generated
during the metabolism of nitrobenzene are not complete enough to substantiate these phenomena
as the carcinogenic MOA. Accordingly, the low-dose linear approach is used for the derivation
of carcinogenic potency.
In order to derive an inhalation risk unit (IUR), the BMCLs from inhalation exposure to
nitrobenzene reported in Table 5-8 were converted to mg/m3 (1 ppm = 5.04 mg/m3 under
0.15 mm Hg at 25°C) and adjusted for continuous exposure as follows:
BMCL (adjusted) = BMCL x 5.04 (mg/m3)/ppm x 6/24 hours x 5/7 days
The tumor types associated with nitrobenzene exposure in rats are systemic effects. As
was discussed for deriving an RfC from the bronchiolization data, interspecies extrapolation to a
HEC involves consideration of the nitrobenzene airblood partition coefficients for humans and
rats (U.S. EPA, 1994b). Since these coefficients were not available, the ratio of animal to human
airblood partition coefficients is assumed to be unity.
The slope of the dose response from the BMCio/HEc for each site (derived by dividing the
BMR [e.g., BMR = 10% or 0.10] by the BMCio/HEc) is provided to illustrate statistical
uncertainty. Estimates of these slopes based on kidney, thyroid, or liver tumors in male F344
rats are shown in Table 5-8.
131
-------
Table 5-8. Cancer risk estimates from nitrobenzene tumor incidence in male
F344 rats, based on the slope to background from the BMC
Target organ/tumor type
Kidney tubular adenoma or carcinoma
Thyroid follicular cell adenoma or carcinoma
Hepatocellular adenoma or carcinoma
BMC10
(ppm)
22.8
13.6
6.8
BMClo/HEC
(mg/m3)
20.5
12.2
6.1
Slope from the BMC10/HEcb
(jig/m3)-1
5 x KT6
8x KT6
2 x 1(T5
aHEC = BMC x 5.04 mg/m3 x 5/7 x 6/24, assuming ratio of animal to human airblood partition coefficients is 1.
bSlope from the BMCHEc = BMR (O.iyBMCuEc.
Estimated lURs are calculated by dividing the BMR (e.g., BMR = 10% or 0.1) by the
BMCLHEc. Estimates of the lURs based on kidney, thyroid, or liver tumors in male F344 rats are
shown in Table 5-9.
Table 5-9. lURs for nitrobenzene, based on tumor incidence in male F344
rats, based on the slope to background from the BMCL
Target organ/tumor type
Kidney tubular adenoma or carcinoma
Thyroid follicular cell adenoma or carcinoma
Hepatocellular adenoma or carcinoma (combined)
BMCL
(ppm)
16.8
7.8
4.5
BMCLHEc"
(mg/m3)
15.1
7.0
4.1
Estimated IURb'c
(Hg/m3)-1
7 x 10~6
1 x 10~5
2 x 10~5
C (adjusted) = BMCL x 5.04 mg/m3 x 5/7 x 6/24 assumes ratio of animal to human airblood partition
coefficients is 1.
bIUR = BMR (0. l)/BMCLHEc.
These lURs should be used with caution at exposure concentrations above the BMCLHEcS, because above these
levels the responses are not expected to continue linearly (i.e., responses cannot exceed 100%).
The highest IUR estimated for nitrobenzene was from liver adenomas or carcinomas, at
3 x 1Q~5 (iig/m3)"1. The IUR associated with thyroid tumors was about threefold lower than that
for liver tumors, and the IUR associated with kidney tumors was about fourfold lower than that
for liver tumors.
With a multiplicity of tumors, as is the case for nitrobenzene, there is a concern that a
potency or risk estimate based solely on one tumor site (e.g., hepatocellular adenomas or
carcinomas) may underestimate the overall cancer risk associated with exposure to this chemical.
Two approaches were considered in which a composite IUR for tumors of the liver, kidney, and
thyroid was developed (see Appendix B-3.3). The two approaches supported each other. The
composite IUR, rounded to one significant figure, is 4 x 10~5 (jig/m3)"1, two fold higher than the
IUR based on liver adenomas or carcinomas, 2 x 10~5 (jig/m3)"1. The recommended upper bound
estimate on human extra cancer risk from continuous lifetime inhalation exposure to
132
-------
nitrobenzene is 4 x 10 5 (|j,g/m3) l, reflecting the exposure-response relationships for liver,
thyroid, and kidney cancer.
5.3.4. Uncertainties in Cancer Risk Values
Extrapolation of study data to estimate potential risks to human populations from
exposure to nitrobenzene has engendered some uncertainty in the results. The uncertainty falls
into two major categories: model uncertainty and parameter uncertainty. Model uncertainty
"refers to a lack of knowledge needed to determine which is the correct scientific theory on
which to base a model," whereas parameter uncertainty "refers to a lack of knowledge about the
values of a model's parameters" (U.S. EPA, 2005a). In the absence of a biologically based
model, a multistage model was the preferred model because it has some concordance with the
multistage theory of carcinogenesis and serves as a benchmark for comparison with other cancer
dose-response analyses. It is unknown how well this model or the linear low-dose extrapolation
predicts low-dose risks for nitrobenzene. Also, while the male mice did not appear to have as
strong a carcinogenic response as the male rats, it is not known which species is more relevant
for extrapolation of risk to humans.
Parameter uncertainty can be assessed through confidence intervals and probabilistic
analysis. Each description of parameter uncertainty assumes that the underlying model and
associated assumptions are valid. Some uncertainty in the animal dose-response data can be
assessed through the ratio of BMDs to their BMDLs. For the tumors evaluated here, the ratio
was below a factor of 2, which is a typical degree of statistical uncertainty.
133
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
6.1.1. Exposure Pathways
At room temperature nitrobenzene is a liquid with a vapor pressure high enough to allow
human exposure to occur via inhalation. It is also able to penetrate human skin, both as liquid
and as vapor. Most serious poisonings with nitrobenzene appear to have happened in domestic
settings via either accidental or intentional ingestion or by dermal and inhalation exposure from
its use in pesticides. Nitrobenzene is also used in significant amounts as an intermediate in
chemical syntheses and as a solvent in products, such as paint, printing ink, and shoe polish, or
as a scenting agent in soap. There are no epidemiological studies of the health effects of
nitrobenzene in humans.
6.1.2. Toxicokinetics
The lipophilicity of nitrobenzene and the composition of membranes in the human body
are the main determinants for systemic absorption. Reports from accidental poisonings (Myslak
et al., 1971), studies in human volunteers (Piotrowski, 1967; Salmowa et al., 1963), and
occupational studies (Ikeda and Kita, 1964) indicate that nitrobenzene is absorbed well from the
human gastrointestinal tract as well as from the lungs. In addition, Feldmann and Maibach
(1970) demonstrated that nitrobenzene is absorbed through the skin. Although their data pointed
to a rather insignificant amount penetrating the skin, poisoning cases in children seem to indicate
that at least young humans are at risk from dermal exposure to nitrobenzene. Beauchamp et al.
(1982) calculated that, in adults, about equal parts of a dose originating from exposure to
nitrobenzene vapor are due to inhalation and dermal absorption, respectively. Animal
experiments have supported the findings in humans.
Although nitrobenzene is rather lipophilic, it does not display a high affinity for fatty
tissues. The only study on the distribution of nitrobenzene in animals (Albrecht and Neumann,
1985) showed that highest levels after an oral dose to female Wistar rats were present in the
blood 1 or 7 days after administration, followed by kidney, with lower levels in liver and lung.
The tendency of nitrobenzene to associate with blood has been confirmed by Goldstein and
Rickert (1984). The main targets are RBCs, which are chemically modified by binding with
nitrobenzene metabolites, and the spleen.
Nitrobenzene is metabolized via reduction of the nitro group to aniline and/or by
hydroxylation of the aromatic ring to phenolic compounds. Reduction of the nitro group appears
to be the dominant process. Two processes have been described for the reduction of the nitro
group with reductase enzymes as catalysts and NAD(P)H as the cofactor: an aerobic three-step,
134
-------
two-electrons-per-step process in intestinal microflora that operates at a high metabolic rate and
an anaerobic six-step, one-electron-per-step process in mammalian cells that is much less
effective because it is inhibited by normal tissue levels of oxygen. RBCs command a set of
enzymes that force nitrobenzene into a futile redox cycle between the nitrobenzene metabolite,
nitrosobenzene, and phenylhydroxylamine (Holder, 1999). This pathway also can result in the
formation of glutathione conjugates. Redox cycling of nitrobenzene is thought to contribute to
the development of methemoglobinemia and to DNA damage caused by reactive oxygen species
(Levin and Dent, 1982).
Nitrobenzene is eliminated in humans and animals, mostly via urine, independent of the
route of exposure. Ortho-, meta-, and para-variants of both nitrophenol and aminophenol have
been identified in the urine of nitrobenzene-exposed experimental animals (Parke, 1956;
Robinson et al., 1951) and humans (Myslak et al., 1971; Feldmann and Maibach, 1970;
Piotrowski, 1967). Experiments with specific pathogen-free animals suggest that more than half
of the urinary nitrobenzene metabolites are formed by intestinal microflora (Reddy et al., 1976).
Fecal and exhalatory elimination also have been observed in rats and mice, with about 1/6 of a
dose of [14C]-labeled nitrobenzene excreted via feces and about 1/40 exhaled in air (Rickert et
al., 1983; Levin and Dent, 1982). Elimination of nitrobenzene from the human or rodent
organism is not a rapid process. In rats, it took about 3 days to eliminate 80% of a 22.5 mg/kg
dose of nitrobenzene (Rickert et al., 1983). In some of the human poisoning cases, it took about
a week to overcome the clinical signs of methemoglobinemia.
6.1.3. Characterization of Noncancer Effects
The database of studies of nitrobenzene effects in animals is considerably more robust
than that of studies in humans. Case reports dealing with acute poisonings via ingestion or
dermal exposure indicate that the hallmark effect of nitrobenzene exposure in humans is
methemoglobinemia. This condition can be treated with blood transfusions or with reducing
agents, such as vitamin C and methylene blue, that return the iron in metHb from iron (III) to its
normal, oxygen-carrying iron (II) state. Severe cases have been known to have a fatal outcome,
particularly in children. Splenic pathology can be traced to the role that the spleen plays in
scavenging RBCs damaged by nitrobenzene metabolites.
There is a considerably more detailed database for nitrobenzene effects in animals. In
animals, methemoglobinemia and other signs of acute toxicity can be observed, including signs
of neurotoxicity, likely due to a lack of oxygen and possibly due to a general solvent effect. A
90-day oral gavage study (NTP, 1983a) found dose-dependent increases in liver, kidney, and
spleen weights (both absolute and relative) in both sexes of mice and rats and a decrease in testis
weight in male F344 rats. By the end of the study, animals surviving the highest dose had
methemoglobinemia (>12%) and considerable blood pathology (decreased Hb, Hct, and RBC
count and increased reticulocyte count), all compatible with hemolytic anemia caused by metHb
135
-------
formation. Histopathologic evaluation revealed congestion and lymphoid depletion of the
spleen, pigment (hemosiderin) deposition in the kidney and brain, and testicular atrophy in
males. Splenic congestion and effects on some hematologic values were observed at low doses.
The splenic pathology, too, can be traced to metHb formation and subsequent RBC hemolysis.
Generally, similar pathology was observed in male and female B6C3F1 mice in the oral
subchronic NTP (1983a) study and in a 28-day gavage study in F344 rats (Shimo et al., 1994).
In that study, some of the animals were allowed a 14-day recovery period. While most of the
pathology observed tended to return to normal within 2 weeks, testicular atrophy in male rats
treated with the highest dose, 125 mg/kg-day, showed little tendency for improvement.
Several studies were conducted with inhalation exposure of experimental animals,
including 14-day studies (Medinsky and Irons, 1985; DuPont, 1981), a 90-day subchronic study
(CUT, 1984), and a 2-year chronic study (CUT, 1993, published as Cattley et al. [1994]). The
chronic study (CUT, 1993) was conducted in compliance with good laboratory practice (GLP)
and contemporary requirements for chronic studies. Both the 90-day and the 2-year studies were
carried out by using both sexes of F344 rats and B6C3F1 mice; in addition, the 90-day study
included both sexes of CD rats, while the chronic study included only male CD rats. Several of
the same target tissues as in the oral study were identified following inhalation exposure, with
the addition, at lower exposures, of the degeneration of the olfactory epithelium of the nasal
turbinates and bronchiolization of the alveoli in mice. Other pathologies following 90-day or 2-
year inhalation exposure to nitrobenzene common to both species were changes in target organ
weights, blood pathology, and methemoglobinemia.
NTP (1983b) also conducted a 90-day dermal study with nitrobenzene in F344 rats and
B6C3F1 mice of both sexes. Again, the pathological effects were very similar to those observed
in the gavage study (NTP, 1983a), but, in addition, congestion of the lung was observed at higher
doses (>100 mg/kg-day) as was uterine atrophy in female rats at the highest dose, 800 mg/kg-
day.
In summary, the major effects of mid- to long-term exposure to nitrobenzene,
independent of the route of exposure, appear to be increases in liver, kidney, and spleen weights
and methemoglobinemia with subsequent hemolytic anemia and splenic congestion.
Administration of nitrobenzene via inhalation additionally elicited olfactory degeneration and
bronchi olizati on of the alveoli as effects specific for this route of exposure. The olfactory
degeneration occurred in a concentration-dependent manner, and bronchi olizati on of the alveoli
occurred in >86% of male and female mice at the lowest concentration tested. Effects on the
male reproductive system, which are also potentially critical effects, are discussed in the
following section.
136
-------
6.1.4. Reproductive Effects and Risks to Children
As young children are more susceptible to methemoglobinemia (a toxic effect of
nitrobenzene) than are adults, they may be more susceptible to this aspect of nitrobenzene
toxicity. There are several reasons for this. First, newborns still have fetal Hb, which is more
susceptible to metHb formation than adult Hb (Goldstein et al., 1969). Next, the activity of
NADH-cytochrome &s reductase, an enzyme required for the conversion of ferric iron to ferrous
iron in Hb, is not fully developed in infants and very young children (Wentworth et al., 1999)
and neither is G6PD activity, an enzyme required to replenish NADPH (Goldstein et al., 1969).
Additionally, the observation of more accidental fatal poisonings in children exposed dermally
indicates a potentially greater sensitivity to dermal nitrobenzene exposures.
There is no information available concerning potential reproductive toxicity of
nitrobenzene in humans. In rodents, however, nitrobenzene is a moderately effective male
reproductive toxicant. A single 300 mg/kg dose of nitrobenzene to male F344 rats caused sperm
production to decrease 20 days after administration, eventually dropping to zero by 50 days
(Levin et al., 1988). By 100 days after treatment, sperm production had returned to 78% of
control levels. This time course reflects the normal spermatogenic cycle of rats. In another
experiment, the same dose was found to cause lesions to seminiferous tubules and marked
necrosis of spermatogenic cells (Bond et al., 1981), as well as decreases in sperm mobility and
viability and morphologically abnormal sperm (Koida et al., 1995; Matsuura et al., 1995).
Dosing with 60 mg/kg-day nitrobenzene for 7-70 days had no effect on the copulatory behavior
of male Sprague-Dawley rats, but their fertility decreased dramatically with exposure times
longer than 14 days. By 4 weeks of dosing, the males were effectively sterile (Kawashima et al.,
1995a, b).
In a reproductive toxicity study (Mitsumori et al., 1994) with 20, 60, and 100 mg/kg-day
nitrobenzene administered orally to Sprague-Dawley rats for 14 days preceding mating, no effect
on fertility or the offspring was observed (dosing was continued throughout pregnancy and the
first 4 days of lactation). Mortality was high among the high-dose females (7/9 and 2/2 died
during gestation and lactation, respectively). In a two-generation study where Sprague-Dawley
rats were exposed to 1, 10, and 40 ppm nitrobenzene via inhalation, starting 10 weeks before
mating, a strong, dose-dependent reduction in fertility was observed that was more marked in the
FI generation than in the F0 generation (Dodd et al., 1987; BRRC, 1985).
In a study with inhalation exposure of pregnant Sprague-Dawley rats to 1, 10, or 40 ppm
nitrobenzene on GDs 6-15, no effects on number of implantations, resorptions, or stillbirths were
observed (Tyl et al., 1987). There were no typical signs of teratogenicity in the offspring,
although some effects on ossification were observed. However, the authors were uncertain
whether those observations were compound related. Several other reproductive/developmental
inhalation studies in New Zealand rabbits (Biodynamics Inc., 1984, 1983) and in CD rats (Tyl et
al., 1987) also produced no indication of a teratogenic action of nitrobenzene. In summary, there
137
-------
is strong evidence for nitrobenzene to act as a male reproductive toxicant, although at higher
exposures than those eliciting other effects, but there is no indication that nitrobenzene affects
female fertility or acts as a developmental toxicant.
6.1.5. Noncancer Mode of Toxic Action
Nitrobenzene elicits an array of toxic effects, and, for any of these to occur, it appears
that metabolic activation or conversion of the parent compound may be involved. A prominent
critical effect identified here is methemoglobinemia. This effect requires metabolism, which is
mostly carried out by intestinal microflora (Reddy et al., 1976). The active metabolite appears to
be nitrosobenzene, which is taken up into RBCs, where it binds with high affinity to Hb (Holder,
1999; Kiese, 1966). The exact mechanism is not completely understood, but it is likely that
redox cycling of nitrosobenzene via phenylhydroxylamine results in oxidation of Fe2+ to Fe3+ in
Hb and thus formation of metHb. This leads to destruction of the RBCs, with resulting
hemolysis, anemia, and splenic congestion.
There is, as yet, no hypothesis concerning the development of olfactory degeneration,
bronchiolization of the alveoli, or potential for immunotoxicity from nitrobenzene. Humans are
facultative nose breathers, while rodents are obligatory nose breathers. Olfactory degeneration
observed following long-term nitrobenzene inhalation in rodents may, therefore, not be relevant
for humans, but supportive or refuting evidence is not available. However, bronchiolization of
the alveoli is of relevance to both facultative and obligatory nose breathers. It has been proposed
that metabolism of nitrobenzene involves the formation of reactive oxygen species (Han et al.,
2001) that can be the cause of damage to point-of-entry tissues, provided they command suitable
activities of metabolizing enzymes.
The male reproductive toxicity of nitrobenzene affects the Sertoli cells (Allenby et al.,
1990). Shinoda et al. (1998) demonstrated that the loss of germ cells following nitrobenzene
exposure was due to apoptosis, and they speculated that factor(s) released from Sertoli cells
might be responsible. Another potent testicular toxicant, mono-(2-ethylhexyl) phthalate, caused
apoptosis in germ cells via the Fas/Jun/AP-1 system, but nitrobenzene-induced testicular toxicity
did not proceed via this pathway (Richburg and Nafiez, 2003). The action of reactive oxygen
species cannot be excluded as a causative factor here. In summary, the noncancer MOA of
nitrobenzene requires metabolism of the parent compound and may involve reactive oxygen
species but otherwise is not well elucidated.
6.1.6. Characterization of the Human Carcinogenic Potential
Under the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), nitrobenzene
is "likely to be carcinogenic to humans." This descriptor is based on the induction of cancers in
two species of laboratory animals, rats and mice, in both sexes, in two strains of rats (F344 and
138
-------
male CD), and in multiple sites in a 2-year inhalation bioassay. There are no studies that
document the carcinogenicity of nitrobenzene in humans.
There are no nitrobenzene exposure data or studies in humans from which to assess a
potential mechanism of action for cancer. Nitrobenzene caused neoplasia in a 2-year chronic
inhalation study (Cattley et al., 1994; CUT, 1993) in a dose-related manner in the livers of male
F344 rats and the lungs of male B6C3F1 mice. Increased incidences of neoplasia with
statistically significant, positive dose trends were also observed as kidney and thyroid adenomas
and carcinomas in male F344 rats, endometrial polyps in female F344 rats, hepatocellular
adenomas and carcinomas in male CD rats, and kidney neoplasia in male B6C3F1 mice.
Although the probable human carcinogen, aniline, is a metabolite of nitrobenzene, there is no
evidence that it is a causative agent (U.S. EPA, 1994c).
Based on the results of genotoxicity tests, nitrobenzene does not seem to induce tumor
formation via a mutagenic MO A. This determination is based on the almost exclusively negative
results in salmonella assays (Ames tests; the only exception is TA 98 in the presence of a
comutagen), as well as negative clastogenic findings from in vivo assays of sister chromatid
exchange, unscheduled DNA synthesis, and chromosomal aberrations. Other MO As, including
oxidative stress, formation of DNA adducts, disruption of intercellular communication, or
cytolethality with subsequent regenerative hyperplasia, a promotion-based MO A, have not been
experimentally validated in connection with nitrobenzene. Also, experimental data for excluding
thyroid and kidney tumors observed in experimental animals based on rodent-specific MO As are
lacking (U.S. EPA, 1998b, 1991b). Additionally, it is not known whether there are any specific,
qualitative, or quantitative differences in nitrobenzene metabolism between rodents and humans,
and there is no reason to assume that a cancer MOA exists in animals that might not be relevant
to humans. Therefore, a final conclusion on a cancer MOA cannot be determined at this time.
This is reflected in the use of a linear approach as a default option in extrapolating the
carcinogenic potential of nitrobenzene.
6.2. DOSE RESPONSE
A few studies have been conducted with nitrobenzene in human research subjects.
However, they were of short duration, used nontoxic doses, and only examined clinical signs.
All dose-response assessments are therefore based on animal data obtained from chronic or
subchronic studies.
6.2.1. OralRfD
The only study in which nitrobenzene was administered orally for an extended period of
time, 90 days (NTP, 1983a), was conducted in a well-controlled fashion in accordance with GLP
guidelines valid at that time. The NTP (1983a) study included both sexes and two species, the
F344 rat and the B6C3F1 mouse; 10 animals per sex and dose group; and five dose groups plus
139
-------
controls (0, 9.38, 18.75, 37.5, 75, and 150 mg/kg-day for rats and 0, 18.75, 37.5, 75, 150, and
300 mg/kg-day for mice). The study reported an abundance of toxic endpoints, including
changes in absolute and relative organ weights, changes in hematologic parameters, and
histopathologic outcomes. Methemoglobinemia, splenic congestion, and reticulocyte count in
male F344 rats were considered as potential critical effects. Dose-response data were evaluated
using BMDS (version 1.4.1c), with 10% extra risk as the BMR for splenic congestion of grade 2
(mild) or higher, and with 1 SD as the BMR for reticulocyte count and metHb levels. Splenic
congestion and increased reticulocyte count were considered to be sequelae of
methemoglobinemia, and the BMD modeling of these endpoints generally supported that
assumption. The resulting POD, based on the exposure at which 10% of an exposed population
would be expected to exceed the 98% upper limit of an unexposed population's metHb levels,
was 1.8 mg/kg-day. After application of a composite UF of 1,000, the oral RfD was identified as
2 x 10~3 mg/kg-day.
The composite UF consists of an interspecies UF of 10 for extrapolation from animals to
humans, an intraspecies UF of 10 to adjust for sensitive subpopulations (most importantly small
children), a subchronic-to-chronic UF of 3 to correct for the less-than-lifetime exposure duration
in the principal study, and a database deficiency UF of 3 to account for lack of an oral
multigeneration reproductive study.
The overall confidence in the RfD is medium. The critical effect on which the RfD is
based is well supported by several other oral gavage studies over time periods of up to 70 days
(Kawashima et al., 1995a, b). Nitrobenzene also displayed toxicity in reproductive and
immunological studies but at doses higher than those used in the principal study. On the basis of
these considerations, confidence in the principal study is high. Confidence in the database is
medium because there is no 2-year oral study, no NOAEL in the 90-day gavage study, and no
multigeneration reproductive/developmental oral study. The medium confidence rating is driven
by such deficits in the database.
6.2.2. Inhalation RfC
A few studies with nitrobenzene in human research subjects have been conducted that
were of short duration with nontoxic doses, and their target was not pathological evaluation.
There are four animal studies available dealing with inhalation toxicity of nitrobenzene, ranging
in duration from acute to chronic. A 90-day subchronic study was conducted by using F344 and
CD rats as well as B6C3F1 mice of both sexes (CUT, 1984). Exposure concentrations were 0, 5,
16, and 50 ppm for 6 hours/day, 5 days/week. This study identified a variety of hematologic
endpoints, above all methemoglobinemia, with several other outcomes secondary to hemolytic
anemia. The 2-year study, also conducted by CUT (Cattley et al., 1994; CUT, 1993), used
B6C3F1 mice and F344 rats of both sexes and male CD rats. Rats were exposed to 0, 1, 5, and
25 ppm nitrobenzene and mice to 0, 5, 25, and 50 ppm nitrobenzene for 6 hours/day,
140
-------
5 days/week (except holidays). This study identified a range of noncancer endpoints, of which
bronchiolization of the alveoli was the most sensitive endpoint in both male and female mice.
Olfactory degeneration was a sensitive endpoint in female mice and to a lesser extent in male
mice. Bronchi olizati on of the alveoli and olfactory degeneration were chosen as co-critical
effects for deriving the RfC, over methemoglobinemia, because of the greater extent and
increasing severity of both endpoints with increasing concentration compared to the lack of a
clear concentration-dependent response for methemoglobinemia at final sacrifice.
The effects selected for RfC derivation, bronchi olizati on of the alveoli and olfactory
degeneration in male and female B6C3F1 mice in the chronic study, were considered portal-of-
entry effects. The PODADJ for bronchiolization, derived by BMD modeling, was 0.9 mg/m3,
which was converted to a PODHEc of 0.29 mg/m3. The PODAoj for olfactory degeneration, also
derived by BMD modeling, was 1.27 mg/m3, which was converted to a PODnEc of 0.26 mg/m3,
essentially the same as the PODHEc derived for bronchiolization. The lower POD, 0.26 mg/m3,
was chosen as the POD. A composite UF of 30 was applied to the PODnEc, resulting in an RfC
of 9 x 10~3 mg/m3. The composite UF included an interspecies UF of 3 for animal-to-human
pharmacodynamics extrapolation, because an HEC pharmacokinetic adjustment had already been
incorporated. An intraspecies UF of 10 was applied to allow for sensitive human populations,
and a database deficiency UF of 1 was used.
Confidence in the principal study is high because it was a 2-year bioassay with a
sufficient number of animals, and it is reasonable to assume that the endpoint is relevant to
humans. Confidence in the database is rated high due to the existence of a 2-year inhalation
study, a two-generation reproductive and developmental toxicity study, and a subchronic
inhalation study. The overall confidence in the RfC evaluation is medium due to a concern that a
NOAEL was not identified for the incidence of bronchiolization of the alveoli in all exposure
groups.
6.2.3. Oral Cancer Risk
The lack of available data precludes an assessment of a potential cancer risk for humans
following oral exposure to nitrobenzene. Since a PBPK model for nitrobenzene is not available,
a quantitative comparison of the IRIS drinking water unit risk for aniline with the levels of
aniline produced from metabolism of inhaled nitrobenzene cannot be made (U.S. EPA, 1994c).
6.2.4. Inhalation Cancer Risk
The mode of carcinogenic action of nitrobenzene remains poorly understood but, based
on available studies, is not likely due to mutagenicity. Nitrobenzene was inactive in all bacterial
mutagenicity assays and gave equivocal results in both in vivo and in vitro mammalian assay
systems. There is limited experimental evidence that nitrobenzene can form DNA adducts or
cause oxidative DNA damage. According to the Guidelines for Carcinogen Risk Assessment
141
-------
(U.S. EPA, 2005a), the default approach in such a case is to use a linear dose extrapolation
approach. Nitrobenzene caused tumors in multiple organs in both sexes in two species (rat and
mouse) and in two different strains of rats in a 2-year inhalation study (Cattley et al., 1994; CUT,
1993). Nitrobenzene increased the incidence of lung adenomas and carcinomas in male B6C3F1
mice only, providing minimal evidence for point-of-entry carcinogenesis.
Male F344 rats appeared to be the most sensitive animal and presented with tumors of the
liver, kidney, and thyroid. ITJRs were developed for each of these tumor types, and liver tumors
appear to be the most sensitive tumor type. The IUR for liver tumors is 2 x 1CT5 (jig/m3)"1.
Composite ITJRs that account for all tumor types were also developed using two approaches—a
Bayesian approach and a tumor-bearing animal approach. Both approaches provided an IUR of
4 x 1CT5 (iig/m3)"1, two fold higher than the IUR based on liver tumors alone. The recommended
upper bound estimate on human extra cancer risk from continuous lifetime inhalation exposure
to nitrobenzene is 4 x icr5 (jig/m3)"1, an estimate that reflects the exposure-response
relationships for liver, thyroid, and kidney cancer.
142
-------
7. REFERENCES
Abbinante, A; Zerpa, R; Pasqualatto, D. (1997) Intoxication due to ingestion of bitter almond oil contaminated with
nitrobenzene: clinical experience. Toxicologist 36:43.
Ajmani, A; Prakash, SK; Jain, SK; et al. (1986) Acquired methaemoglobinaemia following nitrobenzene poisoning.
J Assoc Physicians India 34:891-892.
Albrecht, W; Neumann, HG. (1985) Biomonitoring of aniline and nitrobenzene. Hemoglobin binding in rats and
analysis of adducts. Arch Toxicol 57:1-5.
Alcorn, CJ; Simpson, RJ; Leahy, D; etal. (1991) In vitro studies of intestinal drug absorption. Determination of
partition and distribution coefficients with brush border membrane vesicles. Biochem Pharmacol 42:2259-2264.
Allenby, G; Sharpe, RM; Foster, PM. (1990) Changes in Sertoli cell function in vitro induced by nitrobenzene.
Fundam Appl Toxicol 14:364-375.
Allenby, G; Foster, PM; Sharpe, RM. (1991) Evaluation of changes in the secretion of immunoactive inhibinby
adult rat seminiferous tubules in vitro as an indicator of early toxicant action on spermatogenesis. Fundam Appl
Toxicol 16:710-724.
Anderson, D; Styles, JA. (1978) The bacterial mutation test. Six tests for carcinogenicity. Br J Cancer 37:924-930.
Ask, K; Decologne, N; Asare, N; et al. (2004) Distribution of nitroreductase activity toward nilutamide in rat.
Toxicol Appl Pharmacol 201:1-9.
Assmann, N; Emmrich, M; Kampf, G; et al. (1997) Genotoxic activity of important nitrobenzenes and nitroanilines
in the Ames test and their structure-activity relationship. Mutat Res 395:139-144.
ATSDR (Agency for Toxic Substances and Disease Registry). (1990) Toxicological profile for nitrobenzene. Public
Health Service, U.S. Department of Health and Human Services, Atlanta, GA. Available online at
http://www.atsdr.cdc.gov/toxpro2.html.
Bairoch, A. (2000) The enzyme database in 2000. Nucleic Acids Res 28:304-305.
Ban, Y; Naya, M; Nishimura, T; et al. (2001) Collaborative study on rat sperm motion analysis using CellSoft Series
4000 semen analyzer. J Toxicol Sci 26:9-24.
Beauchamp, RO, Jr; Irons, RD; Rickert, DE; et al. (1982) A critical review of the literature on nitrobenzene toxicity.
CritRev Toxicol 11:33-84.
Biodynamics Inc. (1983) Range-finding study to evaluate the toxicity of nitrobenzene in the pregnant rabbit—draft
final report with cover letter dated 092683. Produced Biodynamics Inc., East Millstone, NJ for the Nitrobenzene
Association, Wilmington, DE; Project No. 83-2723. Submitted under TSCA Section 4A; EPA Document No. 40-
8324347; NTIS No. OTS0509345.
Biodynamics Inc. (1984) An inhalation teratology study in rabbits with nitrobenzene (final report). Produced
Biodynamics Inc., East Millstone, NJ for the Nitrobenzene Association, Wilmington, DE; Project No. 83-2725.
Submitted under TSCA Section 4; EPA Document No. 40-8424492; NTIS No. OTS0510651.
Blackburn, DM; Gray, AJ; Lloyd, SC; et al. (1988) A comparison of the effects of three isomers of dinitrobenzene
on the testes in the rat. Toxicol Appl Pharmacol 92:54-64.
Bloom, JC; Brandt, JT. (2001) Toxic responses of the blood. In: Klaassen, CD; ed. Casarett and Doull's toxicology:
the basic science of poisons. 6th edition. New York, NY: McGraw-Hill; pp. 389-417.
143
-------
Bonacker, D; Stoiber, T; Bohm, KJ; et al. (2004) Chromosomal genotoxiciry of nitrobenzene and benzonitrile. Arch
Toxicol 78:49-57.
Bond, JA; Chism, JP; Rickert, DE; et al. (1981) Induction of hepatic and testicular lesions in Fischer 344 rats by
single oral doses of nitrobenzene. Fundam Appl Toxicol 1:389-394.
Bradberry, SM. (2003) Occupational methemoglobinemia: mechanisms of production, features, diagnosis, and
management including the use of methylene blue. Toxicol Rev 22:13-27.
BRRC (Bushy Run Research Center). (1984) Teratogenicity evaluation of inhaled nitrobenzene in the CD rat (final
report). Produced by Bushy Run Research Center, Export, PA for the Nitrobenzene Association, Wilmington, DE;
Project Report 47-522. Submitted under TSCA Section 4; EPA Document No. 40-8424493; NTIS No.
OTS0510652.
BRRC (Bushy Run Research Center). (1985) Potential effects of nitrobenzene inhalation on reproductive
performance and fertility in rats. Produced by Bushy Run Research Center, Export, PA for the Nitrobenzene
Association, Wilmington, DE; Project Report 47-524. Submitted under TSCA Section 4; EPA Document No. 40-
8524494; NTIS No. OTS0510653.
Bryant, C; DeLuca, M. (1991) Purification and characterization of an oxygen-insensitive NAD(P)H nitroreductase
fromEnterobacter cloacae. JBiol Chem266:4119-4125.
Burns, LA; Bradley, SG; White, KL, Jr; et al. (1994) Immunotoxicity of nitrobenzene in female B6C3F1 mice.
Drug Chem Toxicol 17:271-315.
Butterworth, BE; Smith-Oliver, T; Earle, L; et al. (1989) Use of primary cultures of human hepatocytes in
toxicology studies. Cancer Res 49:1075-1084.
Castle, L; Philo, MR; Sharman, M. (2004) The analysis of honey samples for residues of nitrobenzene and
petroleum from the possible use of Frow mixture in hives. Food Chem 84(4):643-649.
Cattley, RC; Everitt, JI; Gross, EA; et al. (1994) Carcinogenicity and toxicity of inhaled nitrobenzene in B6C3F1
mice and F344 and CD rats. Fundam Appl Toxicol 22:328-340.
Chandra, AM; Quails, CW, Jr; Reddy, G; et al. (1995a) Hematological effects of 1,3,5-trinitrobenzene (TNB) in rats
in vivo and in vitro. J Toxicol Environ Health 46:57-72.
Chandra, AM; Quails, CW, Jr; Reddy, G. (1995b) 1,3,5-Trinitrobenzene-induced encephalopathy in male Fischer-
344 rats. Toxicol Pathol 23:527-532.
Chandra, AM; Quails, CW, Jr; Campbell, GA; et al. (1997) Testicular effects of 1,3,5-trinitrobenzene (TNB). II.
Immunolocalization of germ cells using proliferating cell nuclear antigen (PCNA) as an endogenous marker. J
Toxicol Environ Health 50:379-387.
Chandra, AM; Campbell, GA; Reddy, G; et al. (1999) Neurotoxicity of 1,3,5-trinitrobenzene (TNB):
immunohistochemical study of cerebrovascular permeability. Vet Pathol 36:212-220.
Chilosi, M; Poletti, V; Zamo, A; et al. (2003) Aberrant Wnt/p-catenin pathway activation in idiopathic pulmonary
fibrosis. Am J Pathol 162:1495-1502.
Chiu, CW; Lee, LH; Wang, CY; et al. (1978) Mutagenicity of some commercially available nitro compounds for
Salmonella typhimurium. Mutat Res 58:11-22.
Chongtham, DS; Phurailatpam, J; Singh, MM; et al. (1997) Methaemoglobinaemia in nitrobenzene poisoning. J
Postgrad Med 43:73-74.
Chongtham, DS; Phurailatpam, J; Singh, MM; et al. (1999) Methaemoglobinaemia in nitrobenzene poisoning—a
case report. J Indian Med Assoc 97:469-470.
144
-------
CUT (Chemical Industry Institute of Toxicology). (1982) 104-Week chronic toxicity study in rats: aniline
hydrochloride [unpublished]. Performed by Hazleton Laboratories America, Inc., Vienna, VA for the Chemical
Industry Institute of Toxicology, Research Triangle Park, NC; Project No. 2010-101.
CUT (Chemical Industry Institute of Toxicology). (1984) Ninety day inhalation toxicity study of nitrobenzene in
F344 rats, CD rats, and B6C3F1 mice. Chemical Industry Institute of Toxicology. Research Triangle Park, NC;
Docket No. 12634. Submitted under TSCA Section 8D; EPA Document No. 878214291; NTIS No. OTS0206507.
CUT (Chemical Industry Institute of Toxicology). (1993) Initial submission: a chronic inhalation toxicity study of
nitrobenzene inB6C3Fl mice, Fischer 344 rats and Sprague-Dawley (CD) rats. Chemical Industry Institute of
Toxicology. Research Triangle Park, NC. EPA Document No. FYI-OTS-0794-0970; NTIS No. OTS0000970.
Clark, MR; Shohet, SB. (1985) Red cell senescence. ClinHaematol 14:223-257.
Clayson, DB; Garner, RC. (1976) Carcinogenic aromatic amines and related compounds. In: Searle, CE; ed.
Chemical carcinogens. American Chemical Society monograph 173. Washington, DC: American Chemical Society;
pp. 366-461.
Cody, TE; Witherup, S; Hastings, L; et al. (1981) 1,3-Dinitrobenzene: toxic effect in vivo and in vitro. J Toxicol
Environ Health 7(5):829-847.
Coleman MD; Coleman, NA. (1996) Drug-induced methemoglobinemia. Treatment issues. Drug Safety 14(6):394-
405.
Collins, JF; Orozco, CR; McCullough, B; et al. (1982) Pulmonary fibrosis with small-airway disease; a model in
nonhuman primates. Exp Lung Res 3:91-108.
Cooper, K. R. and D. J. Caldwell. 1995. Developmental toxicity of 1,3,5-trinitrobenzene in Sprague-Dawley rats.
Final Report, Wright-Patterson AFB.
Cotran, RS; Kumar, V; Robbins, SL. (1994) Robbins pathologic basis of disease. 5th edition. Philadelphia, PA: WB
Saunders Co.; p. 670.
Dellarco, VL; Prival, MJ. (1989) Mutagenicity of nitro compounds in Salmonella typhimurium in the presence of
flavin mononucleotide in a preincubation assay. Environ Mol Mutagen 13:116-127.
DiSanto, AR; Wagner, JG. (1972) Pharmacokinetics of highly ionized drugs. II. Methylene blue—absorption,
metabolism, and excretion in man and dog after oral administration. J Pharm Sci 61:1086-1090.
Dodd, DE; Fowler, EH; Snellings, WM; et al. (1987) Reproduction and fertility evaluations in CD rats following
nitrobenzene inhalation. Fundam Appl Toxicol 8:493-505.
Dreher, D; Junod, AF. (1996) Role of oxygen free radicals in cancer development. Eur J Cancer 32A:30-38.
DuPont. (1981) Inhalation median lethal concentration (LC50) with cover letter. Performed by Haskell Laboratory
for Toxicology and Industrial Medicine, Vienna, VA for E.I. du Pont de Nemours and Company, Newark, DE.
Submitted under TSCA Section 8D; EPA Document No. 878220423; NTIS No. OTS0215040.
Ellenhorn, MJ; Schoenwald, S; Ordog, G; eds. (1997) Antidotes. In: Ellenhorn's medical toxicology: diagnosis and
treatment of human poisoning. 2nd edition. Baltimore, MD: Williams & Wilkins; pp. 89-105.
Ellis, MK; Foster, PM. (1992) The metabolism of 1,3-dinitrobenzene by rat testicular subcellular fractions. Toxicol
Lett 62:201-208.
Eyer, P. (1979) Reactions of nitrosobenzene with reduced glutathione. Chem Biol Interact 24:227-239.
Eyer, P; Ascherl, M. (1987) Reactions of para-substituted nitrosobenzenes with human hemoglobin. Biol Chem
Hoppe Seyler 368:285-294.
145
-------
Facchini, V; Griffiths, LA. (1981) The involvement of the gastro-intestinal microflora in nitro-compound-induced
methaemoglobinaemia in rats and its relationship to nitrogroup reduction. Biochem Pharmacol 30:931-935.
Feig, DI; Reid, TM; Loeb, LA. (1994) Reactive oxygen species in tumorigenesis. Cancer Res 54 (Suppl. 7): 1890s-
1894s.
Feldmann, RJ; Maibach, HI. (1970) Absorption of some organic compounds through the skin in man. J Invest
Dermatol 54:399-404.
Finch, C. (1947) Treatment of intracellular methemoglobinemia. Bull N Engl Med Center 6:241-245.
Friemann, J; Albrecht, C; Breuer, P; et al. (1999) Time-course analysis of type II cell hyperplasia and alveolar
bronchiolization in rats treated with different particulates. Inhal Toxicol 11:837-854.
Garner, RC; Nutman, CA. (1977) Testing of some azo dyes and their reduction products for mutagenicity using
Salmonella typhimurium TA 1538. MutatRes 44:9-19.
Goldfrank, L; Flomenbaum, N; Lewin, N; et al.; eds. (1998) Goldfrank's toxicologic emergencies. 6th edition.
Stamford, CT: Appleton & Lange.
Goldstein, RS; Rickert, DE. (1984) Macromolecular covalent binding of [14C]nitrobenzene in the erythrocyte and
spleen of rats and mice. Chem Biol Interact 50:27-37.
Goldstein, A; Aronow, L; Kalman, SM. (1969) Principles of drug action: the basis of pharmacology. New York,
NY: Harper and Row Publishers; pp. 274^52.
Goldstein, RS; Chism, JP; Sherrill, JM; et al. (1984) Influence of dietary pectin on intestinal microfloral metabolism
and toxicity of nitrobenzene. Toxicol Appl Pharmacol 75:547-553.
Goodman, DG; Boorman, GA; Strandberg, JD. (1985) Selection and use of the B3C3F1 mouse and F344 rat in long-
term bioassays for carcinogenicity. In: Milman, HA; Weisburger, EK; eds. Handbook of carcinogen testing. Park
Ridge, NJ: Noyes Publications; pp. 282-325.
Greaves, P. (2007) Histopathology of preclinical toxicity studies. 3rd edition. Amsterdam, The Netherlands: Elsevier
Science; p. 118.
Gupta, G; Poddar, B; Salaria, M; et al. (2000) Acute nitrobenzene poisoning. Indian Pediatr 37:1147-1148.
Gutteridge, JM. (1995) Lipid peroxidation and antioxidants as biomarkers of tissue damage. Clin Chem 41:1819-
1828.
Guyton, AC; Hall, JE. (2000) Blood cells, immunity, and blood clotting. Chapter 32 in: Textbook of medical
physiology, pp. 382-391. Philadelphia, PA: WB Saunders.
Guyton, KZ; Kensler, TW. (1993) Oxidative mechanisms in carcinogenesis. Br Med Bull 49:523-544.
Halliwell, B. (2007) Oxidative stress and cancer: have we moved forward? Biochem J 401:1-11.
Hamilton, A. (1919) Industrial poisoning by compounds of the aromatic series. J Ind Hyg 1:200-212.
Han, C; Wang, Q; Wu, P. (2001) [A study on mechanism for cytotoxicity of nitrobenzene to hepatocarcinoma cell
line]. Zhonghua Yu Fang Yi Xue Za Zhi 35:48-50.
Harada, N; Omura, T. (1980) Participation of cytochrome P-450 in the reduction of nitro compounds by rat liver
microsomes. J Biochem 87:1539-1554.
Harrison, MR. (1977) Toxic methaemoglobinaemia. A case of acute nitrobenzene and aniline poisoning treated by
exchange transfusion. Anaesthesia 32:270-272.
146
-------
Haworth, S; Lawlor, T; Mortelmans, K; et al. (1983) Salmonella mutagenicity test results for 250 chemicals.
Environ Mutagen 5(Suppl. 1): 1-142.
Ho, CH; Clark, BR; Guerin, MR; et al. (1981) Analytical and biological analysis of test materials from the synthetic
fuel technologies. MutatRes 85:335-345.
Holder, JW. (1999) Nitrobenzene carcinogenicity in animals and human hazard evaluation. Toxicol Ind Health
15:445-457.
Hong, SK; Anestis, DK; Ball, JG; et al. (2002) In vitro nephrotoxicity induced by chloronitrobenzenes in renal
cortical slices from Fischer 344 rats. Toxicol Lett 129:133-141.
Hopkins, JE; Naisbitt, DJ; Humphreys, N; et al. (2005) Exposure of mice to the nitroso metabolite of
sulfamethoxazole stimulates interleukin 5 production by CD4(+) T-cells. Toxicology 206:221-231.
Hsu, CH; Stedeford, T; Okochi-Takada, E; et al. (2007) Framework analysis for the carcinogenic mode of action of
nitrobenzene. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev 25(2): 155-184.
Huang, Q; Wang, L; Han, S. (1995) The genotoxicity of substituted nitrobenzenes and the quantitative structure-
activity relationship studies. Chemosphere 30:915-923.
Huang, QG; Kong, LR; Liu, YB; et al. (1996) Relationships between molecular structure and chromosomal
aberrations in in vitro human lymphocytes induced by substituted nitrobenzenes. Bull Environ Contam Toxicol
57:349-353.
Hultquist, DE; Passon, PG. (1971) Catalysis of methaemoglobin reduction by erythrocyte cytochrome B5 and
cytochrome B5 reductase. Nat New Biol 229(8):252-4.
Ikeda, M; Kita, A. (1964) Excretion of />-nitrophenol and^-aminophenol in the urine of a patient exposed to
nitrobenzene. Br J Ind Med 21:210-213.
IPCS (International Programme on Chemical Safety ). (2003) Nitrobenzene. Environmental health criteria. Vol.
230. World Health Organization, Geneva, Switzerland. Available online at
http://www.inchem.org/documents/ehc/ehc/ehc230.htm.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005a)EC 1.6.99.1-NADPH
dehydrogenase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/6/99/l.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005b) EC 1.6.99.3-NADH dehydrogenase.
Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/6/99/3.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005c) EC 1.6.5.3-NADH dehydrogenase
(ubiquinone). Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/6/5/3.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005d)EC 1.15.1.1-Superoxide dismutase.
Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/15/l/l.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005e)EC 1.11.1.9-Glutathione
peroxidase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/l l/l/9.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (20051) EC 1.11.1.6-Catalase. Available
online at http://www.chem.qmul.ac.uk /iubmb/enzyme/ECl/11/1/6.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005g) EC 2.5.1.18-Glutathione
transferase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/EC2/5/l/18.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005h)EC 1.8.1.7-Glutathione-disulfide
reductase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/8/l/7.html.
147
-------
IUBMB (International Union of Biochemistry and Molecular Biology). (2005i) EC 1.6.2.4-NADPH—hemoprotein
reductase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/6/2/4.html.
IUBMB (International Union of Biochemistry and Molecular Biology). (2005j)EC 1.6.2.2-Cytochrome-b5
reductase. Available online at http://www.chem.qmul.ac.Uk/iubmb/enzyme/ECl/6/2/2.html.
Jaffe, ER. (1981)Methemoglobinpathophysiology. Prog ClinBiol Res 51:133-151.
Jensen-Taubman, SM; Steinberg, SM; Linnoila, RI. (1998) Bronchiolization of the alveoli in lung cancer: pathology,
patterns of differentiation and oncogene expression. Int J Cancer 75:489-496.
Kato, M; Kimura, H; Hayashi, H; et al. (1995) Sperm viability in rats treated with nitrobenzene and alpha-
chlorohydrin. Teratology 52:44B.
Kawaguchi, T; Kawachi, M; Morikawa, M; et al. (2004) Key parameters of sperm motion in relation to male fertility
in rats given alpha-chlorohydrin or nitrobenzene. J Toxicol Sci 29:217-231.
Kawashima, K; Usami, M; Sakemi, K; et al. (1995a) Studies on the establishment of appropriate spermatogenic
endpoints for male fertility disturbance in rodent induced by drugs and chemicals. I. Nitrobenzene. J Toxicol Sci
20:15-22.
Kawashima, K; Momma, J; Takagi, A; et al. (1995b) Examination of sperm motility defects by nitrobenzene with an
image processor. Teratology 52:37B.
Kawashima, K; Momma, J; Kitajima, S; et al. (1996) Sperm test using CASA (HTM-IVOS): II. Nitrobenzene and
its analog. Teratology 54:41 A.
Kensler, TW; Egner, PA; Taffe, BG; et al. (1989) Role of free radicals in tumor promotion and progression. Prog
Clin Biol Res 298:233-248.
Kiese, M. (1966) The biochemical production of ferrihemoglobin-forming derivatives from aromatic amines, and
mechanisms of ferrihemoglobin formation. PharmacolRev 18:1091-1161.
Kim, S; Quails, CW, Jr; Reddy, G; etal. (1997) l,3,5-Trinitrobenzene-inducedalpha-2u-globulinnephropathy.
Toxicol Pathol 25:195-201.
Kinkead, ER; Wolfe, RE; Flemming, CD; et al. (1994a) Reproductive toxicity screen of 1,3,5-trinitrobenzene
administered in the diet of Sprague-Dawley rats—final report for the period September 1993 through June 1994.
ManTech Environmental Technology, Inc., Dayton, OH. Available from the National Technical Information
Service, Springfield, VA; AD-A298 912.
Kinkead, ER; Wolfe, RE; Salins, SA; et al. (1994b) Range-finding study for a reproductive assessment of 1,3,5-
trinitrobenzene administered in the diet of Sprague-Dawley rats. ManTech Environmental Technology, Inc.,
Dayton, OH. Available from the National Technical Information Service, Springfield, VA; AD-A299 032.
Kinkead, ER; Wolfe, RE; Fleming, CD; et al. (1995) Reproductive toxicity screen of 1,3,5-trinitrobenzene
administered in the diet of Sprague-Dawley rats. Toxicol Ind Health 11(3):309-323.
Kito, Y; Hamamatsu, Y; Naya, M. (1998) Effects of nitrobenzene on sperm motility and fertility in rats (2).
Teratology 57:29A.
Kito, Y; Hamamatsu, Y; Naya, M. (1999) Application of Crj:CD(SD)IGS rats to reproductive and developmental
toxicity study: effects of nitrobenzene on sperm examination. Teratology 59:39A-40A.
Kligerman, AD; Erexson, GL; Wilmer, JL; et al. (1983) Analysis of cytogenetic damage in rat lymphocytes
following in vivo exposure to nitrobenzene. Toxicol Lett 18:219-226.
148
-------
Koida, M; Nakagawa, T; Irimura, K; et al. (1995) Effects on the sperm and testis of rats treated with nitrobenzene:
age and administration period differences. Teratology 52:39B.
Kopylev, L; Chen, C; White, P. (2007) Towards quantitative uncertainty assessment for cancer risks: central
estimates and probability distributions of risk in dose-response modeling. Regul Tox Pharmacol 49:203-207.
Kumar, A; Chawla, R; Ahuja, S; et al. (1990) Nitrobenzene poisoning and spurious pulse oximetry. Anaesthesia
45:949-951.
Kurian, JP; Bajad, SU; Miller, JL; et al. (2004) NADH cytochrome b5 reductase and cytochrome b5 catalyze the
microsomal reduction of xenobiotic hydroxylamines and amidoximes in humans. J Pharmacol Exp Ther 311:1171-
1178.
Kurian, JP; Chin, NA; Longlais, BJ; et al. (2006) Reductive detoxification of arylhydroxylamine carcinogens by
NADH cytochrome b5 reductase and cytochrome b5. Chem Res Toxicol 19:1366-1373.
Kusumoto, S; Nakajima, T. (1970) Methemoglobin formation by nitrobenzene in vitro. Naunyn-Schmiedeberg Arch
Pharmacol 266:113-118.
Lazerev, NV; Levina, EN. (1976) o-, m-, />-DNB. In: Harmful substances in industry II. Leningrad, Russia:
Khiimya Press; pp. 724-727. (as cited in Philbert et al., 1987).
Lee, DC; Ferguson, KL. (2007) Methemoglobinemia (last modified August 7, 2007). e-Medicine from WebMD.
Available from http://www.emedicine.com/emerg/TOPIC313.HTM.
Levin, SJ. (1927) Shoe-dye poisoning—relation to methemoglobin formation. JAMA 89:2178-2180.
Levin, AA; Dent, JG. (1982) Comparison of the metabolism of nitrobenzene by hepatic microsomes and cecal
microflora from Fischer 344 rats in vitro and the relative importance of each in vivo. Drug Metab Dispos 10:450-
454.
Levin, AA; Bosakowski, T; Earle, LL; etal. (1988) The reversibility of nitrobenzene-induced testiculartoxicity:
continuous monitoring of sperm output from vasocystotomized rats. Toxicology 53:219-230.
Lewis, PJ, Sr; ed. (1992) Sax's dangerous properties of industrial materials. 8th edition. New York, NY: Van
Nostrand Reinhold.
Li, H; Cheng, Y; Wang, H; et al. (2003a) Inhibition of nitrobenzene-induced DNA and hemoglobin adductions by
dietary constituents. Appl Radiat Isot 58:291-298.
Li, H; Wang, H; Sun, H; et al. (2003b) Binding of nitrobenzene to hepatic DNA and hemoglobin at low doses in
mice. Toxicol Lett 139:25-32.
Linder, RE; Hess, RA; Strader, LF. (1986) Testicular toxicity and infertility in male rats treated with 1,3-
dinitrobenzene. J Toxicol Environ Health 19(4):477-489.
Linder, RE; Strader, LF; Slott, VL; et al. (1992) Endpoints of spermatotoxicity in the rat after short duration
exposures to fourteen reproductive toxicants. Reprod Toxicol 6:491-505.
Mallouh, AA; Sarette, WO. (1993) Methemoglobinemia induced by topical hair oil. Ann Saudi Med 13:78-80.
Maples, KR; Eyer, P; Mason, RP. (1990) Aniline-, phenylhydroxylamine-, nitrosobenzene-, and nitrobenzene-
induced hemoglobin thiyl free radical formation in vivo and in vitro. Mol Pharmacol 37:311-318.
Mason, RP; Holtzman, JL. (1975a) The mechanism of microsomal and mitochondria! nitroreductase. Electron spin
resonance evidence for nitroaromatic free radical intermediates. Biochemistry 14:1626-1632.
149
-------
Mason, RP; Holtzman, JL. (1975b) The role of catalytic superoxide formation in the O2 inhibition of nitroreductase.
Biochem Biophys Res Commun 67:1267-1274.
Matsumaru, H; Yoshida, T. (1959) Experimental studies of nitrobenzol poisoning. Kyushu J Med Sci 10:259-264.
Matsuura, I; Hoshino, N; Wako, Y; et al. (1995) Sperm parameter studies on three testicular toxicants in rats.
Teratology 52:39B.
Mattioli, F; Martelli, A; Gosmar, M; et al. (2006) DNA fragmentation and DNA repair synthesis induced in rat and
human thyroid cells by chemicals carcinogenic to the rat thyroid. Mutat Res 609:146-153.
McLaren, TT; Foster, PM; Sharpe, RM. (1993a) Identification of stage-specific changes in protein secretion by
isolated seminiferous tubules from the rat following exposure to either /w-dinitrobenzene or nitrobenzene. Fundam
ApplToxicol 21:384-392.
McLaren, TT; Foster, PM; Sharpe, RM. (1993b) Effect of age on seminiferous tubule protein secretion and the
adverse effects of testicular toxicants in the rat. Int J Androl 16:370-379.
Medinsky, MA; Irons, RD. (1985) Sex, strain, and species differences in the response of rodents to nitrobenzene
vapors. In: Rickert, DE; ed. Toxicity of nitroaromatic compounds. New York, NY: Hemisphere Publishing
Corporation; pp. 35-51.
Miller, JL. (2002) Hemoglobin switching and modulation: genes, cells, and signals. Curr Opin Hematol 9:87-92.
Miller, BM; Adler, ID. (1990) Application of antikinetochore antibody staining (CREST staining) to micronuclei in
erythrocytes induced in vivo. Mutagenesis 5:411-415.
Mirsalis, JC; Tyson, CK; Butterworth, BE. (1982) Detection of genotoxic carcinogens in the in vivo-in vitro
hepatocyte DNA repair assay. Environ Mutagen 4:553-562.
Mitsumori, K; Kodama, Y; Uchida, O; et al. (1994) Confirmation study, using nitrobenzene, of the combined repeat
dose and reproductive/developmental toxicity test protocol proposed by the Organization for Economic Cooperation
and Development (OECD). J Toxicol Sci 19:141-149.
Mochida, K; Ito, Y; Saito, K; et al. (1986) Cytotoxic effects of 1,2-dichloroethane, nitrobenzene, and carbon
disulfide on human KB and monkey AGMK cells. J Pharm Sci 75:1190-1191.
Morgan, KT; Gross, EA; Lyght, O; et al. (1985) Morphologic and biochemical studies of a nitrobenzene-induced
encephalopathy in rats. Neurotoxicology 6:105-116.
Morrissey, RE; Schwetz, BA; Lamb, JC, IV; et al. (1988) Evaluation of rodent sperm, vaginal cytology, and
reproductive organ weight data from National Toxicology Program 13-week studies. Fundam Appl Toxicol 11:343-
358.
Muhle, H; Kittel, B; Ernst, H; et al. (1995) Neoplastic lung lesions in rat after chronic exposure to crystalline silica.
Scan J Work Environ Health 21 (Suppl. 2):27-29.
Myslak, Z; Piotrowski, JK; Musialowicz, E. (1971) Acute nitrobenzene poisoning. A case report with data on
urinary excretion of />-nitrophenol and/>-aminophenol. Arch Toxikol 28:208-213.
Narayanan, L; Caldwell, DJ; Miller, CR. (1995) Alteration in neurotransmitters and their metabolite levels in 1,3,5-
trinitrobenzene-treated Sprague-Dawley rats. Performed by Geo-Centers Inc., Newton Centre, MA for the U.S.
Army Medical Research Department, Walter Reed Army Institute of Research, Wright-Patterson AFB, OH;
AL/OET-TR-1995-0133. Available online at http://handle.dtic.mi1/100.2/ADA362513.
Nettesheim, P; Szakal, MS. (1972) Morphogenesis of alveolar bronchiolization. Lab Invest 26:210-219.
Nienhuis, AW; Stamatoyannopoulos, G. (1978) Hemoglobin switching. Cell 15:307-315.
150
-------
NLM (National Library of Medicine). (2003) Nitrobenzene. HSDB (Hazardous Substances Data Bank). National
Institutes of Health, U.S. Department of Health and Human Services, Bethesda, MD. Available online at
http://toxnet.nlm.nih.gov.
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press.
NRC (National Research Council). (1994) Science and judgment. Washington, DC: National Academy Press.
NTP (National Toxicology Program). (1983a) Report on the subchronic toxicity via gavage of nitrobenzene
(C60082) in Fischer 344 rats and B6C3F1 mice [unpublished]. Prepared by the EG&G Mason Research Institute,
Worcester, MA, for the National Toxicology Program, National Institute of Environmental Health Services, Public
Health Service, U.S. Department of Health and Human Services, Research Triangle Park, NC; MRI-NTP 08-83-19.
NTP (National Toxicology Program). (1983b) Report on the subchronic dermal toxicity of nitrobenzene (C60082) in
Fischer 344 rats and B6C3F1 mice [unpublished]. Prepared by the EG&G Mason Research Institute, Worcester,
MA, for the National Toxicology Program, National Institute of Environmental Health Services, Public Health
Service, U.S. Department of Health and Human Services, Research Triangle Park, NC; MRI-NTP 06-83-13.
NTP (National Toxicology Program). (1989) Toxicology and carcinogenesis studies of para-chloroaniline
hydrochloride (CAS No. 20265-96-7) in F344/N rats and B6C3F1 mice (gavage studies). Public Health Service,
U.S. Department of Health and Human Services; NTP TR 351. Available from the National Institute of
Environmental Health Sciences, Research Triangle Park, NC and online at
http://ntp.niehs.nih.gov/ntpweb/index.cfm?objectid=D16D6C59-FlF6-975E-7D23D1519B8CD7A5.
NTP (National Toxicology Program). (2002) NTP technical report on the toxicology and carcinogenesis studies of
^-nitrotoluene (CAS No. 99-99-0) in F344 rats and B6C3F1 mice (feed studies). Public Health Service, U.S.
Department of Health and Human Services; NTP TR 498; NIH publication no. 02-4432. Available from the
National Institute of Environmental Health Sciences, Research Triangle Park, NC and online at
http://ntp.niehs.nih.gov/ntpweb/index.cfm?objectid=D16D6C59-F!F6-975E-7D23D1519B8CD7A5.
Nystrom, DD; Rickert, DE. (1987) Metabolism and excretion of dinitrobenzenes by male Fischer 344 rats. Drug
Metab Dispos 15:821-825.
Ohkuma, Y; Kawanishi, S. (1999) Oxidative DNA damage by a metabolite of carcinogenic and reproductive toxic
nitrobenzene in the presence of NADH and Cu(II). Biochem Biophys Res Commun 257:555-560.
Parke, DV. (1956) Studies in detoxication. 68. The metabolism of [14C]nitrobenzene in the rabbit and guinea pig.
Biochem 162:339-346.
Parkinson, A. (2000) Biotransformation of xenobiotics. In: Klaassen, CD; ed. Casarett & Doull's toxicology: the
basic science of poisons. 6th edition. New York, NY: McGraw-Hill; pp. 133-224.
Percy, MJ; McFerran, NV; Lappin, TR. (2005) Disorders of oxidised haemoglobin. Blood Rev 19:61-68.
Perreault, SD; Linder, RE; Strader, LF; et al. (1989) The value of multiple endpoint data in male reproductive
toxicology: revelations in the rat. Prog Clin Biol Res 302:179-192.
Philbert, MA; Gray, AJ; Connors, TA. (1987) Preliminary investigations into the involvement of the intestinal
microflora in CNS toxicity induced by 1,3-dinitrobenzene in male F-344 rats. Toxicol Lett 38(3):307-314.
Pinkerton, KE; Joad, JP. (2000) The mammalian respiratory system and critical windows of exposure for children's
health. Environ Health Perspect 108(Suppl. 3):457-462.
Pinkerton, KE; Dodge, DE; Cederdahl-Demmler, J; et al. (1993) Differentiated bronchiolar epithelium in alveolar
ducts of rats exposed to ozone for 20 months. Am J Pathol 142:947-956.
Piotrowski, J. (1967) Further investigations on the evaluation of exposure to nitrobenzene. Br J Ind Med 24:60-65.
151
-------
Porter, IH; Schulze, J; McKusick, VA. (1962) Genetical linkage between the loci for glucose-6-phosphate
dehydrogenase deficiency and colour-blindness in American Negroes. Ann Hum Genet 26:107-122.
Reddy, BG; Pohl, LR; Krishna, G. (1976) The requirement of the gut flora in nitrobenzene-induced
methemoglobinemia in rats. Biochem Pharmacol 25:1119-1122.
Reddy, TV; Daniel, FB; Robinson, M; et al. (1994a) Subchronic toxicity studies on 1,3,5-trinitrobenzene, 1,3-
dinitrobenzene and tetryl in rats: subchronic toxicity evaluation of 1,3,5-trinitrobenzene in Fischer 344 rats.
Prepared by the U.S. Environmental Monitoring Systems Laboratory, Cincinnati, OH, for the U.S. Army Medical
Research and Development Command, Ft. Detrick, Frederick, MD. Available from the National Technical
Information Service, Springfield, VA; ADA283 663/3
Reddy, TV; Daniel, FB; Robinson, M; et al. (1994b) Subchronic toxicity studies on 1,3,5-trinitrobenzene, 1,3-
dinitrobenzene and tetryl in rats: 14-day toxicity evaluation of 1,3,5-trinitrobenzene in Fischer 344 rats. Prepared by
the U.S. Environmental Monitoring Systems Laboratory, Cincinnati, OH, for the U.S. Army Medical Research and
Development Command, Ft. Detrick, Frederick, MD. Available from the National Technical Information Service,
Springfield, VA; ADA283664.
Reddy, TV; Torsell, J; Daniel, FB; et al. (1995) Ninety-day toxicity evaluation of 1,3,5-trinitrobenzene (TNB) in
Peromyscus leucopus. Second SET AC (Society of Environmental Toxicology and Chemistry World Congress);
November 5-9, Vancouver, British Columbia, Canada. Pensacola, FL: Society of Environmental Toxicology and
Chemistry; p. 189.
Reddy, TV; Daniel, FB; Olson, GR; et al. (1996) Chronic toxicity studies on 1,3,5-trinitrobenzene in Fischer 344
rats [final report]. Prepared by the Environmental Monitoring Systems Laboratory, U.S. Environmental Protection
Agency, Cincinnati, OH, for the U.S. Army Medical Research and Development Command, Fort Detrick, MD.
Available from the National Technical Information Service, Springfield, VA; ADAS 15 216/2.
Reddy, G; Reddy, TV; Choudhury, H; et al. (1997) Assessment of environmental hazards of 1,3,5-trinitrobenzene. J
Toxicol Environ Health 52:447-460.
Reddy, TV; Olson, GR; Wiechman, B; et al. (1998) Subchronic toxicity of 1,3,5-trinitrobenzene in Fischer 344 rats.
IntJ Toxicol 17:393-411.
Reddy, G; Reddy, TV; Daniel, FB; et al. (2000) Fourteen-day toxicity evaluation of 1,3,5-trinitrobenzene (TNB) in
shrew (Cryptotis parva). Twenty-first meeting of the American College of Toxicology. November 12-15, San
Diego, CA. IntJ Toxicol 19(6):425-461.
Reddy, TV; Olson, GR; Wiechman, B; et al. (2001) Chronic toxicity of 1,3,5-trinitrobenzene in Fischer 344 rats. Int
J Toxicol 20:59-67.
Rice-Evans, C. (1990) Iron-mediated oxidative stress and erythrocytes. In: Harris, JR; ed. Blood cell biochemistry.
New York, NY: Plenum Press; pp. 429-453.
Richburg, JH; Boekelheide, K. (1996) Mono-(2-ethylhexyl) phthalate rapidly alters both Sertoli cell vimentin
filaments and germ cell apoptosis in young rat testes. Toxicol Appl Pharmacol 137:42-50.
Richburg, JH; Nanez, A. (2003) Fas- or FasL-deficient mice display an increased sensitivity to nitrobenzene-induced
testicular germ cell apoptosis. Toxicol Lett 139:1-10.
Rickert, DE. (1987) Metabolism of nitroaromatic compounds. Drug Metab Rev 18:23-53.
Rickert, DE; Bond, JA; Long, RM; et al. (1983) Metabolism and excretion of nitrobenzene by rats and mice.
Toxicol Appl Pharmacol 67:206-214.
Robbiano, L; Baroni, D; Carrozzino, R; et al. (2004) DNA damage and micronuclei induced in rat and human
kidney cells by six chemicals carcinogenic to the rat kidney. Toxicology 204:187-195.
152
-------
Robinson, D; Smith, JN; Williams, RT. (1951) Studies in detoxication. 40. The metabolism of nitrobenzene in the
rabbit; o-, m- and^-nitrophenols, o-, m- and^-aminophenols and 4-nitrocatechol as metabolites of nitrobenzene.
Biochem 150:228-23 5.
Romero, IA; Lister, T; Richards, HK; et al. (1995) Early metabolic changes during m-dinitrobenzene neurotoxicity
and the possible role of oxidative stress. Free Radic Biol Med 18:311-319.
Salice, CJ; Holdsworth, G. (2001) Wildlife toxicity assessment for 1,3,5-trinitrobenzene (1,3,5-TNB). U.S. Army
Center for Health Promotion and Preventative Medicine (USACHPPM), Aberdeen Proving Ground, MD; Project
No. 39-EJ1138-01B. Available online athttp://chppm-www.apgea.army.mil/erawg/tox/TNB(FINAL).pdf.
Salmowa, J; Piotrowski, J; Neuhorn, U. (1963) Evaluation of exposure to nitrobenzene. Absorption of nitrobenzene
vapour through lungs and excretion of />-nitrophenol in urine. Br J Ind Med 20:41-46.
Schimelman, MA; Soler, JM; Muller, HA. (1978) Methemoglobinemia: nitrobenzene ingestion. JACEP 7:406-408.
Schuler, M; Rupa, DS; Eastmond, DA. (1997) A critical evaluation of centromeric labeling to distinguish
micronuclei induced by chromosomal loss and breakage in vitro. MutatRes 392:81-95.
Sealy, RC; Swartz, HM; Olive, PL. (1978) Electron spin resonance-spin trapping. Detection of superoxide formation
during aerobic microsomal reduction of nitro-compounds. Biochem Biophys Res Commun 82:680-684.
Seger, DL. (1992) Methemoglobin-forming chemicals. In: Sullivan, JB; Krieger, GR; eds. Hazardous materials
toxicology: clinical principles of environmental health. Baltimore, MD: Williams & Wilkins; pp. 800-806.
Shimizu, M; Yasui, Y; Matsumoto, N. (1983) Structural specificity of aromatic compounds with special reference to
mutagenic activity in Salmonella typhimurium—a series of chloro- or fluoro-nitrobenzene derivatives. Mutat Res
116:217-238.
Shimkin, MB. (1939) Acute toxicity of mononitrobenzene in mice. Proc Soc Exp Biol Med 42:844-846.
Shimo, T; Onodera, H; Matsushima, Y; et al. (1994) [A 28-day repeated dose toxicity study of nitrobenzene inF344
rats]. Eisei Shikenjo Hokoku 112:71-81.
Shinoda, K; Mitsumori, K; Yasuhara, K; et al. (1998) Involvement of apoptosis in the rat germ cell degeneration
induced by nitrobenzene. Arch Toxicol 72:296-302.
Smith, RP. (1996) Toxic responses of the blood. In: Klaassen, CD; ed. Casarett and Doull's toxicology: the basic
science of poisons. 5th edition. New York, NY: McGraw-Hill; pp. 335-354.
Spiegelhalter, D; Thomas, A; Best, N. (2003) WinBugs user manual version 1.4, January 2003. MRC Biostatistics
Unit, Institute of Public Health, Cambridge, UK and Department of Epidemiology and Public Health, Imperial
College of Medicine, London, UK. Available online at http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/manuall4.pdf.
Stevens, A. (1928) Cyanosis in infants from nitrobenzene. JAMA 90:116.
Stevenson, A; Forbes, RP. (1942) Nitrobenzene poisoning: report of a case due to exterminator spray. J Pediat
21:224-228.
Stifel, RE. (1919) Methemoglobinemia due to poisoning by shoe dye. JAMA 72:395-396.
Styles, JA. (1978) Mammalian cell transformation in vitro. Six tests for carcinogenicity. Br J Cancer 37:931-936.
Suzuki, J; Koyama, T; Suzuki, S. (1983) Mutagenicities of mono-nitrobenzene derivatives in the presence of
norharman. MutatRes 120:105-110.
153
-------
Suzuki, J; Takahashi, N; Kobayashi, Y; et al. (1987) Dependence on Salmonella typhimurium enzymes of
mutagenicities of nitrobenzene and its derivatives in the presence of rat-liver S9 and norharman. Mutat Res
178:187-194.
Takahashi, T; Tanaka, M; Brannan, CI; et al. (1994) Generalized lymphoproliferative disease in mice, caused by a
point mutation in the Fas ligand. Cell 76:969-976.
Tateno, H; lijima, S; Asaka, A; et al. (1997) Evaluation of clastogenicity of chemical agents using in vitro assay
with human spermatozoa. Mutat Res 379:889.
Tingle, MD; Mahmud, R; Maggs, JL; et al. (1997) Comparison of the metabolism and toxicity of dapsone in rat,
mouse, and man. J Pharmacol Exp Ther 283:817-823.
Tyl, RW; France, KA; Fisher, LC; et al. (1987) Development toxicity evaluation of inhaled nitrobenzene in CD rats.
Fundam Appl Toxicol 8:482-492.
U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
mixtures. Federal Register 51(185):34014-34025. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment. Federal
Register 51(185):34006-34012. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1988) Recommendations for and documentation of biological
values for use in risk assessment. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH; EPA/600/6-87/008. Available from the National Technical Information
Service, Springfield, VA, PB88-179874/AS, and online at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
U.S. EPA (Environmental Protection Agency). (1991a) Guidelines for developmental toxicity risk assessment.
Federal Register 56(234):63798-63826. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1991b) Alpha 2u-globulin: association with chemically induced renal
toxicity and neoplasia in the male rat. Risk Assessment Forum, Washington, DC; EPA/625/3-91/019F. Available
online at http://www.epa.gov/nscep.
U.S. EPA (Environmental Protection Agency). (1994a) Interim policy for particle size and limit concentration issues
in inhalation toxicity: notice of availability. Federal Register 59(206):53799. Available online at
http://www.epa.gov/EPA-PEST/1994/October/Day-26/pr-ll.html.
U.S. EPA (Environmental Protection Agency). (1994b) Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of
Health and Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F. Available from the National Technical
Information Service, Springfield, VA, PB2000-500023, and online at
http://cfpub.epa. gov/ncea/raf/recordisplay.cfm?deid=71993.
U.S. EPA (Environmental Protection Agency). (1994c) Aniline. Integrated Risk Information System (IRIS).
National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.
U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk
assessment. Risk Assessment Forum, Washington, DC; EPA/630/R-94/007. Available from the National Technical
Information Service, Springfield, VA, PB95-213765, and online at
http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment. Federal
Register 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1998a) Guidelines for neurotoxicity risk assessment. Federal
Register 63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
154
-------
U.S. EPA (Environmental Protection Agency). (1998b) Assessment of thyroid follicular cell tumors. Risk
Assessment Forum, Washington, DC; EPA/630/R-97/002. Available from the National Technical Information
Service, Springfield, VA, PB98-133119, and online at http://nepis.epa.gov/EPA/html/Pubs/pubtitleORD.htm.
U.S. EPA (Environmental Protection Agency). (1999) Benchmark dose software (BMDS) version 1.3.2 (last
modified May 23, 2003). Available online at http://www.epa.gov/ncea/bmds/index.html.
U.S. EPA (Environmental Protection Agency). (2000a) Science policy council handbook: risk characterization.
Office of Science Policy, Office of Research and Development, Washington, DC. EPA/100-B-00-002. Available
online at http://www.epa.gov/OSA/spc/pdfs/prhandbk.pdf.
U.S. EPA (Environmental Protection Agency). (2000b) Benchmark dose technical guidance document [external
review draft]. Risk Assessment Forum, Washington, DC; EPA/630/R-00/001. Available online at
http://cfpub.epa.gov/ncea/cfm/
nceapublication.cfm?ActType=PublicationTopics&detype=DOCUMENT&subject=BENCHMARK+DOSE&subjty
pe=TITLE&excCol=Archive.
U.S. EPA (Environmental Protection Agency). (2000c) Supplementary guidance for conducting health risk
assessment of chemical mixtures. Risk Assessment Forum, Washington, DC; EPA/630/R-00/002. Available online
at http://cfpub.epa. gov/ncea/raf/chem_mix. cfm.
U.S. EPA (Environmental Protection Agency). (2002) A review of the reference dose concentration and reference
concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-02/002F. Available online at
http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (Environmental Protection Agency). (2004) Toxics release inventory. Office of Pollution Prevention and
Toxics, Washington, DC. Available online at http://www.epa.gov/triexplorer/.
U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Federal Register
70(66): 17765-18717. Available online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2006a) Science policy council handbook: peer review. 3rd edition.
Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-06/002. Available
online at http://www.epa.gov/OSA/spc/2peerrev.htm.
U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risk of environmental
exposures to children. National Center for Environmental Assessment, Washington, DC; EPA/600/R-05/093F.
Available online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363.
U.S. EPA (Environmental Protection Agency). (2007) Benchmark dose software (BMDS) version 1.4. Ic (last
modified November 9, 2007). Available online at http://www.epa.gov/ncea/bmds/index.html.
Vance, WA; Levin, DE. (1984) Structural features of nitroaromatics that determine mutagenic activity in Salmonella
typhimurium. Environ Mutagen 6:797-811.
Vasquez, GB; Reddy, G; Gilliland, GL; et al. (1995) Dinitrobenzene induces methemoglobin formation from
deoxyhemoglobin in vitro. Chem Biol Interact 96:157-171.
Watanabe-Fukunaga, R; Brannan, CI; Copeland, NG; et al. (1992) Lymphoproliferation disorder in mice explained
by defects in Fas antigen that mediates apoptosis. Nature 356:314-317.
Wentworth, P; Roy, M; Wilson, B; et al. (1999) Toxic methemoglobinemia in a 2-year-old child. Lab Med 30:311-
315.
155
-------
Westerman, MP; Pierce, LE; Jensen, WN. (1963) Erythrocyte lipids: a comparison of normal young and normal old
populations. J Lab Clin Med 62:394-400.
Wood, WG. (1976) Haemoglobin synthesis during human fetal development. Br Med Bull 32:282-287.
Wulferink, M; Gonzalez, J; Goebel, C; et al. (2001) T cells ignore aniline, a prohapten, but respond to its reactive
metabolites generated by phagocytes: possible implications for the pathogenesis of toxic oil syndrome. Chern Res
Toxicol 14:389-397.
Zeitoun, MM. (1959) Nitrobenzene poisoning in infants due to inunction with false bitter almond oil. J Trop Pediatr
5:73-75.
Zeligs, M. (1929) Aniline and nitrobenzene poisoning in infants. Arch Pediat 46:502-506.
156
-------
APPENDIX A: SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
The Toxicological Review of Nitrobenzene has undergone formal external peer review
performed by scientists in accordance with EPA guidance on peer review (U.S. EPA, 2006a).
The external peer reviewers were tasked with providing written answers to general questions on
the overall assessment and on chemical-specific questions in areas of scientific controversy or
uncertainty. A summary of significant comments made by the external reviewers and EPA's
responses to these comments arranged by charge question follow. In many cases the comments
of the individual reviewers have been synthesized and paraphrased in development of
Appendix A.
EXTERNAL PEER REVIEWER COMMENTS
A. General Comments
1. Is the Toxicological Review logical, clear and concise? Has EPA objectively and
transparently represented and synthesized the scientific evidence for noncancer and cancer
hazard?
Comment: Most reviewers agreed that the presentation, for the most part, is logical,
clear, and transparent. Some reviewers commented that editorial corrections of typographical
errors will be needed, including explanation of certain qualitative descriptors, such as
"significantly" or "substantial amounts."
Response: Typographical errors were corrected and the use of qualitative descriptors,
such as "significant(ly)" or "substantial amounts," was explained according to the context in the
respective cited literature.
Comment: One reviewer commented that "readability of the document" would be
improved by providing, at the end of each section, summaries of key points and comparisons
across studies in addition to graphical presentation of tabular data to help follow dose-response
patterns.
Response: In addition to the study-by-study tabular presentation of findings, summaries
by type of study, duration, and/or route of exposure, including dosing regimens, key findings,
and NOAEL/LOAEL, were already provided throughout section 4 (see for instances Tables 4-16,
4-18, 4-19, 4-23, 4-28, 4-34, and 4-36). Sections 4.5 and 4.6 are summaries of the noncancer
-------
and cancer studies presented in the earlier parts of section 4. Graphical exposure-response arrays
were added to section 5 as Figures 5-1 and 5-2.
Comment: One reviewer commented on the need to have additional discussion or a
checklist of criteria for selecting principal studies.
Response: A list of reasons for selecting the 90-day gavage study for deriving the RfD
was added at the end of section 5.1.1. As specified under sections 5.2.1 and 5.3, the 2-year
inhalation study was chosen for the RfC and cancer assessments because it was the only lifetime
study by this route and it included both sexes of two species with large group sizes and properly
spaced exposure levels.
Comment: Another reviewer commented that the document had no information on
exposures in the workplace or in the environment and recommended reviewing some of the
"workplace biomonitoring and somewhat older foreign studies," including epidemiologic data on
other structurally related nitro aromatics, and describing concentration levels that were found at
different sites and in different matrices. No specific information on the recommended studies
was provided.
Response: Information on environmental releases of nitrobenzene and potential exposure
through food was added to section 2 of the Toxicological Review. Information on exposures
through air and water in the workplace or the environment and epidemiologic data on related
nitroaromatic compounds were either not found or lacked data on health effects associated with
the exposures, making it unsuitable for drawing conclusions about health hazards. A number of
published case reports of nitrobenzene poisoning are summarized in section 4 of the
Toxicological Review; however, there are no reports of epidemiologic studies of the human
health impacts of nitrobenzene exposure in the workplace or environment.
2. Are you aware of additional studies that should be considered in the assessment of the
noncancer and cancer health effects of nitrobenzene?
Comment: Three reviewers said they were not aware of additional studies but one of the
three urged the authors to check on and provide a discussion about additional
workplace/epidemiologic literature studies from foreign countries as was recommended by
another reviewer. Another commented that some information on use patterns and potential
exposures might be useful.
A-2
-------
Response: The scope of IRIS assessments generally includes only the hazard
identification and dose-response portions of the risk assessment, which users combine with
exposure information to conduct a risk assessment. Biomonitoring data and ambient
concentrations are usually included in IRIS assessments only insofar as they are part of an
epidemiology study that also provides response data, or they illustrate specific aspects of the
hazard identification section such as toxicokinetic features of the chemical (e.g., measurements
of metabolites in urine) or susceptible subpopulations (e.g., measurements in breast milk
indicating high infant exposure). Consequently, much of the recommended literature is outside
the scope of this assessment. EPA obtained a bibliography from the reviewer but did not find
any relevant studies on the list that were not already cited in the Toxicological Review except for
some early studies from the 1950s and 1960s that are not available in English.
Comment: One reviewer suggested having additional coverage of background
information on metHb production (not specifically for nitrobenzene) and the health effects
associated with varying levels of metHb in the blood.
Response: Additional information on metHb formation and methemoglobinemia was
added to the Toxicological Review as footnotes 1 and 3 in sections 3.2 and 3.3. Information on
clinical effects in humans associated with metHb formation is summarized in sections 4.1 and
4.5.1.
Comment: Another reviewer cited the ability of nitroaromatics to directly produce
metHb without activation by microsomes or reducing/anaerobic environment as described in a
reference on dinitrobenzene by Vasquez et al. (1995).
Response: A discussion of the possibility of direct metHb generation by nitroaromatics
was added to the Toxicological Review as footnote 5 in section 3.3.1.
Comment: One reviewer raised additional issues, including possible alterations in
absorption due to changes in gut morphology in the germ-free antibiotic-treated animal (see
Heneghan [1984]), and questioned the validity of the assumption that reactive intermediates pass
through intestinal membranes rather than preferentially react with the gut contents.
Response: The issue of absorption (bioavailability) of nitrobenzene in germ-free animals
is discussed in text added to section 3.3.1 of the Toxicological Review. Information on gut
morphology and function of the alimentary tract in relation to germ colonization was evaluated
in the cited reference (Heneghan, 1984) but was considered not to be informative with regards to
possible alteration in nitrobenzene absorption in the germ-free animal.
A-3
-------
As discussed in section 3.2, nitrobenzene or its metabolites were widely distributed
among the major organs and tissues following [14C]-nitrobenzene administration in rats where
the highest levels were in blood, apparently due to covalent binding to hemoglobin and plasma
proteins (Albrecht and Neumann, 1985; Goldstein and Rickert, 1984). While reactive
intermediates of nitrobenzene may also likely react with gut contents, these findings as well as
others on systemic toxicity and metabolism in germ-free and conventional animals clearly show
that nitrobenzene and its metabolites are appreciably absorbed following administration by the
oral route.
B. Oral Reference Dose (RfD)
1. Is the selection of the NTP (1983) study as the principal study scientifically justified? Is the
rationale transparently and objectively described?
Comment: Despite some reservations, all reviewers agreed that the NTP (1983a) study is
the most relevant and is well justified. Two reviewers indicated that the transparency and
rationale for the selection could be improved by comparing/discussing (in section 5.1) this study
with other candidate studies, including the lower doses used in a shorter-term (28-day) study.
Other reviewers wanted a more focused discussion of how the study was selected.
Response: The considerations listed for evaluating studies outlined in U.S. EPA (2002a)
were used to select the principal study for deriving an RfD for nitrobenzene. The rationale for
selecting the 90-day gavage administration study (NTP, 1983a) was added at the end of section
5.1.1. A discussion of Shimo et al. (1994), the 28-day study, was added to section 5.1.1. The
LOAELs and NOAELs reported for selected endpoints in Shimo et al. (1994) and NTP (1983a)
are provided in Table 4-16 and Figure 5-1.
Comment: One reviewer expressed concern about using "a one-time bolus dose" in a
gavage study as being unrepresentative of what may be observed following actual continuous
exposures as in the 2-year inhalation bioassay. However, the reviewer conceded that route-to-
route extrapolation from the inhalation study would be very difficult in the absence of well-
characterized PBPK modeling for nitrobenzene. Therefore, using the NTP (1983a) 90-day oral
study is scientifically justified, especially since there was good agreement between routes of
exposure and toxicity outcomes.
Response: EPA agrees with the comment on the difficulty of route-to-route extrapolation
from the inhalation study. These comments are captured under the material that was added to
section 5.1.1 on selection of the principal study.
A-4
-------
Comment: One reviewer thought that there may be room for additional coverage of the
influence of route of exposure or organ on nitrobenzene metabolism and distribution including
organ-specific metabolizing enzymes and reactions.
Response: As summarized in section 3.3 of the Toxicological Review, information on
nitrobenzene metabolism is derived from nitrobenzene metabolite identification in the urine from
humans and animals and from in vitro metabolism by liver microsomes or erythrocytes. The
metabolite findings from these studies indicate that oxidative (possibly cytochrome P450
dependent) and reductive pathways, including conjugation, are involved but do not yield
information on tissue-specific roles of different metabolizing enzymes. The Toxicological
Review also covers in vitro studies on types I and II (oxygen-insensitive and oxygen-sensitive)
nitroreductases in various tissue homogenates by following nilutamide (a nitroaromatic drug)
reduction (R-NO2 —»• R-NEk) under aerobic and anaerobic conditions. Although nitrobenzene
was not studied, the findings on nilutamide demonstrated that the highest levels of
nitroreductases are in intestinal microflora, but various organs also have the same activities.
However, there are no specific studies on the influence of route of administration (e.g., oral
versus inhalation) on formation of nitrobenzene metabolites and on route-dependent pathways or
involvement of other metabolizing enzymes.
2. Splenic congestion (increased by 10%), methemoglobin levels (increased by 1 SD), and
reticulocyte count (increased by 1 SD) relative to control values serves as the basis for the RfD.
Is the selection of the splenic congestion, metHb levels, and reticulocyte count as the co-critical
effects for deriving the RfD scientifically justified? Has the rationale for selection of these
critical effects been transparently and objectively described? Is it appropriate to derive point of
departure by averaging BMDLs across sexes and co-critical effects?
Comment: Three reviewers agreed with the three selected endpoints, including rationale,
justification, and averaging of BMDLs. One reviewer agreed with the choice of endpoints but
commented that biological justification for the selection of BMR should be provided and that
more explanation is needed for averaging the BMDLs. Another reviewer provided extensive
comments on data presentation and BMD modeling used to derive the POD.
Response: The rationale for the selection of the three endpoints was expanded to clarify
the likely interdependence of the observed effects. The data considered for RfD derivation were
also summarized in section 5.1 and in the relevant appendices.
The individual animal data were obtained to determine whether splenic congestion (and
other histopathologic lesions) had been only noted as present (as analyzed for the Toxicological
Review} or graded, since this could affect the BMR. Splenic congestion had been graded; control
A-5
-------
rats of both sexes had none or only minimal congestion (grade 1), while congestion increased in
severity up to moderate in the most highly exposed rats, including those that died early in the
study. Footnotes were added to clarify how many animals with this finding died early. The
BMR for splenic congestion was reconsidered in terms of having at least mild congestion (grade
2). The ToxicologicalReview was revised to take into consideration the severity of splenic
congestion.
Individual hematology data were examined to assist in addressing the comments
regarding heterogeneous variances of the continuous variables. There were no apparent reasons
to consider excluding any of the individual values for the highlighted endpoints as outliers. The
modeling was revised to consider only monotonic dose-response shapes and the reviewer
suggestions for modeling variances. Model results were generated for all cases previously
missing results.
EPA disagrees with the comment on the necessary degree of consistency of the fitted
models with limited data in the high-dose group(s). Even with increased consistency
accomplished by using the monotonic dose-response shapes, it is more important to fit the data at
the lower end of the dose response, where mortality did not have as obvious an impact on the
responses. Consequently, a number of model fits dropped the high dose. These instances were
identified more clearly.
Regarding biological justifications for BMR selection, it is not clear what level of effect
constitutes a minimal biologically significant degree of change for the metHb and reticulocyte
effects in rodents and how changes in these effects relate quantitatively to humans. The NTP has
regarded similar changes in rat metHb to be biologically significant (e.g., p-chloroaniline [NTP,
1989]). However, additional analyses and text were added in section 5.1 to address the
biological significance of metHb formation and issues related to endogenous levels in humans.
Methemoglobin changes in male rats were identified as the most sensitive effect. Accordingly,
the critical effect was changed to reflect a reliance on metHb data in male rats with a POD based
on a 1SD increase from the control mean.
Comment: One reviewer stated that it would be useful to include "a clear and transparent
definition and description of the adverse consequences of splenic congestion," since the
pathophysiological consequences of "brown pigmentation" and "red pulp" mentioned in
section 4.2 are not clear. This reviewer also requested further explanation of how increased
metHb and reticulocyte count are considered adverse.
Response: These comments were addressed by providing appropriate discussion or
footnotes in section 4.2.1.1.
A-6
-------
Comment: One reviewer suggested consideration of the reversibility of the selected
endpoints upon reduction or discontinuation of exposure and the compensatory response
following extended exposure.
Response: In the NTP (1983a) study, there is no information on reversibility of the
endpoints that were selected for the RfD. However, changes in hematologic parameters,
including methemoglobinemia, as well as increased spleen weight were reversed after
discontinuation of exposure in other studies, including the 28-day gavage administration study
(Shimo et al., 1994) and the 2-week inhalation studies by DuPont (1981) and Medinsky and
Irons (1985). The following is a further discussion of the issues of reversibility and possible
compensatory response to methemoglobinemia while keeping in mind that other related findings,
including hemolytic anemia, reticulocytosis, and spleen congestion (or splenomegaly), are
interrelated in the sense that they are all indicative of erythrocyte toxicity by nitrobenzene or
other nitroaromatic and aromatic aniline chemicals.
As discussed in the ToxicologicalReview, metHb, which is an abnormal form of Hb with
diminished affinity for oxygen, is formed and metabolized back to Hb at roughly equal rates
under physiological conditions. Methemoglobinemia may develop when generation of metHb,
following for instance exposure to a chemical, exceeds the normal compensatory physiological
reductive capacity within erythrocytes mainly carried out by the NADH-dependent cytochrome
b5 metHb reductase. After discontinuation of exposure, methemoglobinemia may be reversed as
evidenced, for instance in the study by Medinsky and Irons (1985), after 14 days of
discontinuation of exposure to nitrobenzene (section 4.4.1, including Table 4-35). Reversibility
of methemoglobinemia also seems to take place in humans following discontinuation of
exposure and treatment with methylene blue and ascorbic acid as summarized in some of the
case reports (section 4.1). However, reversibility may be hampered in some individuals due to
congenital deficiencies in the reductase enzyme or in G6PD (see sections 3.3 and 4.7.3).
The issue of a possible compensatory response to methemoglobinemia was addressed in
section 4.2.2.2 and under footnote 10 of the Toxicological Review, where it is concluded that it is
difficult to convincingly identify a compensatory response in the chronic inhalation study, and
there is no known specific information in the published literature that addresses this topic. A
similar comment on the possibility of a compensatory response to methemoglobinemia in
relation to selecting a UF of 3 for less-than-lifetime exposure in the principal oral study was also
raised and responded to in the first comment under RfD question 4 (Q4).
Finally, an RfD is an estimate of a daily oral exposure that is likely to be without
appreciable risks to health during a lifetime, which implies that the RfD includes continuous
lifetime exposures.
A-7
-------
3. Are the uncertainty factors applied to the point of departure for the derivation of the RfD
scientifically justified and transparently and objectively described?
Comment: Two reviewers agreed that the type and value of the selected UFs appeared to
be transparently and objectively described but one considered the total value of 1,000 to be
relatively large and suggested that considering data on relative animal-to-human toxicity for
structurally similar compounds might help mitigate some of the uncertainties. Two reviewers
questioned applying the UF for subchronic-to-chronic exposure or using an interspecies 10-fold
UF, which implies that humans are more sensitive to metHb than are rodents. Some of the
suggestions included (1) providing a rationale with direct dose-response comparison between
rodents and humans regarding sensitivity to metHb by estimating the ingested dose from
nitrobenzene human ingestion studies or utilizing clinical data on oxidatively damaging drugs
and (2) including information on what is known about drug sensitivity due to G6PD deficiency.
Response: In the absence of comparative chemical-specific data, humans are generally
assumed to be more sensitive than animals on a mg/kg-day basis, based on their relative size.
Application of a full UF of 10 is warranted to account for uncertainty in extrapolating from
laboratory animals to humans based on two areas of uncertainty, namely toxicokinetic and
toxicodynamic differences. There are no available in vivo data that compare human sensitivity
relative to that of rodents towards metFIb induction by nitrobenzene or related nitroaromatic
chemicals. Also, it is not feasible to compare doses from existing human toxicity reports based
on a single ingestion exposure on the order of grams to chronic doses of nitrobenzene in the
microgram to milligram range used in animal studies. EPA's practice is that, unless data support
the conclusion that the test species is more or equally as susceptible to the pollutant as humans
and in the absence of any other specific toxicokinetic or toxicodynamic data, a default factor of 3
(in conjunction with HEC derivation) or 10 is applied (U.S. EPA, 2002a).
In a comparative study using in vivo and in vitro techniques, Tingle et al. (1997)
concluded that humans are more sensitive than rats or mice to hematotoxicity (mainly metFIb
formation) from dapsone (diaminodiphenyl sulfone), an antibacterial and anti-inflammatory drug
used to treat leprosy and certain skin conditions.
Species differences were also described in the activity of red cell spontaneous (NADH)
metFIb reductase with rat, mouse, guinea pig, and rabbit having higher activity than humans,
while horse and pig have lower rates than humans (Smith, 1996). Based on this review, rodents
are more apt to repair metHb than are humans, since the relative erythrocyte metFIb reductase
activity between rats and humans ranged from 1.3- to 5.0-fold and between mice and humans
was 9.5-fold. Also, the rat is nearly twice as efficient as humans in the methylene blue
stimulated (NADPH) erythrocyte metHb reductase activity (Smith, 1996). In the absence of
actual data, other pharmacokinetic differences between rodents and humans may also exist.
A-8
-------
Therefore, the 10-fold UF for extrapolating from rats to humans was retained.
The issue of applying a subchronic to chronic UF of 3 in relation to lack of apparent
worsening methemoglobinemia was also raised and responded to in the last comment under RfD
Q2 and in the first comment under RfD Q4.
Comment: One reviewer suggested factoring in "the species differences among rodents
and the dependence on gut flora," and questioned the rationale for applying a factor of 10-fold
for animal-to-human extrapolation for the oral RfD and only threefold for the inhalation RfC.
Response: The response to the previous comment addressed the issue of applying a 10-
fold factor for animal-to-human extrapolation for deriving an RfD. Regarding possible species
differences in gut microflora, there is no information that indicates that metabolism of
nitroaromatics by gut flora is specific to rodents.
As stated in section 5.2.4, the interspecies UF of 10 to extrapolate from animals to
humans would account for two areas of uncertainty, namely toxicokinetic and toxicodynamic
differences, each of which comprises 10°5 of the total UF. Based on the applied RfC
methodology (U.S. EPA, 1994), the toxicokinetic component was addressed by the HEC
dosimetric adjustment from an animal-specific LOAELAoj to a human LOAELHEc. Therefore, a
UF of 3 was applied to account for the toxicodynamic uncertainty in derivation of the RfC.
There is little information to inform nitrobenzene toxicokinetics via the oral route; consequently,
the toxicokinetic portion of the UF is retained.
Comment: Another reviewer answered "no" and suggested providing documentation of
guidelines or criteria of how UFs were selected in order to improve the transparency.
Response: A review document on the RfD and RfC by the Risk Assessment Forum
Technical Panel (U.S. EPA, 2002a) discusses derivation of reference values, which takes into
consideration and accounts for five areas of uncertainty or variability in the available data by
assigning a factor (UF) for each. The exact value of each UF depends on the quality of the
studies, the extent of the database, and scientific judgment. The five areas of
uncertainty/variability for deriving an RfD and RfC were identified in section 5 (sections 5.1.3
and 5.2.4, respectively) and justification was provided for each selected UF, including revisions
based on the external peer review comments.
4. An uncertainty factor of 3 was selected to account for less-than-lifetime exposure in the
principal oral study. Is the choice of this UF scientifically justified and transparently and
objectively described?
A-9
-------
Comment: One reviewer provided no comments and another indicated insufficient
familiarity with the use of UFs to comment while two reviewers agreed that using a UF for less-
than-lifetime exposure is reasonable or is consistent with Agency policy. Two of the four
reviewers suggested discussing the basis for the selection or describing the guidelines that were
followed. Three reviewers questioned the need for a UF for lifetime extrapolation from a
subchronic study, indicating that the extent of methemoglobinemia seemed to be independent of
exposure duration in the inhalation studies, with one reviewer finding the selection rationale "to
be a bit problematic." The same reviewer also cited an apparent selectivity in presenting the
rationale for applying the factor based on cited route-specific metabolite formation (under
section 5.1.3) compared to formation of nitrosobenzene from inhaled nitrobenzene in Figure 3-8.
This reviewer also questioned why a similar compensatory response to methemoglobinemia, as
the one seen in the inhalation studies (discussed in Appendix B-2), would not be operative when
nitrobenzene is administered orally. The second reviewer stated that "it doesn't seem likely that
additional toxicity not identified in the sub-chronic studies would occur in the chronic
bioassays." The third reviewer stated, "There is the experience with the inhalation studies that
suggests no such factor is needed."
Response: Considerations for deriving reference values and selection of UFs are
discussed in the response to a similar comment under the previous question (RfD, Q3).
EPA agrees with the comment that formation of the nitro anion free radical and
nitrosobenzene are not route specific; however, inadequate data exist from the oral route to
evaluate the formation of these metabolites and greater concentrations may be generated
following exposure by the oral route compared to exposure via inhalation due to gut bacterial
microflora activation (nitro reduction). A similar comment was also made by a reviewer under
RfD Q5. Section 5.1.3 was modified to resolve apparent inconsistencies.
As explained in section 4.2.2.2 and in the response to a comment on a compensatory
response under charge question 2 of the section on the oral RfD, the extent of
methemoglobinemia as a function of extended exposure in humans remains largely unknown.
According to the data in Table 4-20, metHb levels were not increased when exposure to
nitrobenzene via inhalation was extended from 15 months (interim sacrifice) to 24 months
(terminal sacrifice) in both rat strains. Additionally, other toxicity endpoints may result from
chronic oral exposure due to route-specific differences in metabolism, pharmacokinetics, and/or
pharmacodynamics that were not observed in the subchronic oral study or in the inhalation
studies. Therefore, a subchronic-to-chronic UF of 3 is appropriate to account for the uncertainty
regarding the compensatory response and for potential additional effects, other than the above-
mentioned hematologic endpoints, that might occur with chronic exposure.
A-10
-------
5. An oral database uncertainty factor of 3 was applied. The database of oral studies includes
the principal study (NTP, 1983b) (sic)., a 90-day gavage study in two species and both sexes;
high quality reproductive/ developmental studies (Mitsumori et al., 1994; Morrissey et al., 1988;
Bond et al., 1981); structure-activity relationship studies comparing nitrobenzene to dinitro- and
trinitrobenzene; and a multidose immunological study in mice (Burns et al., 1994). However,
due to a lack of an oral multigeneration reproductive toxicity study and in light of evidence of
male reproductive toxicity, a factor of 3 was applied. Is the choice of an UF of 3 scientifically
defensible given the available oral and inhalation databases? Does the available data suggest that
the oral exposures may result in new adverse effects at oral doses equivalent to or lower than the
inhalation concentrations used in the multigeneration reproductive and developmental study by
Doddetal. (1987)?
Comment: One reviewer indicated that weighing scientific information and presenting a
number of reasons for choosing a UF of 3 strengthened EPA's discussion on deriving an oral
RfD. Another reviewer stated that a UF may be justified based on enhanced gut bacterial
metabolic activation (nitro reduction), resulting in higher concentrations of different active
metabolites than would be possible by inhalation exposure, which makes it hard to predict oral
versus inhalation potential long-term effects on reproduction. Both of these reviewers, however,
reiterated the need to specify the guidance used by EPA for selecting the UFs.
Response: Refer to the response given under RfD Q3 on a similar comment on EPA
guidance for selecting UFs.
Comment: One reviewer stated: "At most, I would suggest a factor of 3 for these two
UFs combined (i.e., subchronic to chronic extrapolation and database deficiency), pending more
complete investigation of the route-to-route extrapolation that might allow a fuller integration of
the oral and inhalation databases." This reviewer also cited a lack of major differences in
metabolism or systemic response between oral and inhalation exposures.
Response: Route-to-route extrapolation from the inhalation study would be difficult in
the absence of a well-characterized PBPK model for nitrobenzene. It is also difficult to calculate
dose equivalency between oral and inhalation routes in the absence of an equivalent metric for
the measured changes between the routes. Two reviewers offered a similar opinion (under RfD
Ql and RfD Q5) that comparison of dose equivalency between the oral and inhalation exposure
routes was not feasible.
Furthermore, there are knowledge gaps in route-specific comparative metabolism or
systemic responses that prevent a determination of whether differences exist following chronic
exposure to nitrobenzene by the oral versus inhalation routes. The database and subchronic-to-
A-ll
-------
chronic UFs are intended to address two different uncertainties. The database UF is intended to
account for the potential for deriving an underprotective RfD/RfC as a result of an incomplete
characterization of the chemical's toxicity Another UF is applied when a study of subchronic
duration is used for a chronic exposure scenario. It accounts for the uncertainty associated with
the possibility that other or more severe effects might have been observed if the duration of
exposure was longer.
Comment: One reviewer questioned the need for a database UF since reproductive
toxicity occurred at a dose greater than that used for setting a POD.
Response: EPA agrees that in the one-generation oral study by Mitsumori et al. (1994)
reproductive toxicity (specifically testicular pathology and male fertility) was a less sensitive
endpoint than other endpoints (such as spleen or liver weights or hematology). However, when
an RfD or RfC is based on animal data, EPA generally applies a database UF of 3 if either a
developmental (prenatal) toxicity study or a two-generation reproduction study is missing or a
factor of 10 if both are missing (U.S. EPA, 2002a). EPA is not aware of specific data to suggest
that reproductive effects may worsen with extended mating in a two-generational study;
however, EPA similarly has no data to establish that reproductive effects would not occur at
lower doses following multigeneration exposure. In light of this uncertainty, EPA considers a
database UF of 3 to be appropriate.
Comment: A reviewer commented that it is unlikely that oral exposures may result in
new adverse effects at oral doses equivalent to or lower than the inhalation concentrations used
in the multigeneration reproductive and developmental study by Dodd et al. (1987). This
reviewer also commented that "it is difficult, if not impossible, to calculate dose equivalence
between oral and inhalation routes without a precise basis and understanding for the mechanism
of action for the biologic effect and for an equivalent metric for the measured changes between
the routes."
Response: The issue of the feasibility of route-to-route extrapolation from the inhalation
study in the absence of well-characterized PBPK modeling for nitrobenzene was discussed
earlier (see comment under RfD Ql). It is not possible to determine oral equivalent doses to the
inhalation concentrations of the multigeneration reproductive study. Also, in the absence of a
multigeneration reproductive study by the oral route, it is difficult to predict the type or extent of
adverse outcomes in parental animals or their offspring.
A-12
-------
C. Inhalation Reference Concentration (RfC)
1. Is bronchiolization of the alveoli the most scientifically justifiable endpoint on which to base
the RfC? Have the rationale and justification for this selection been transparently and
objectively described? Are there any other studies that you believe would be justified
scientifically as the basis for the RfC?
Comment: None of the reviewers suggested other studies for deriving the RfC. Two
reviewers were concerned that bronchi olizati on of alveoli may not be the most justifiable
endpoint or that scientific justification was not provided. Another reviewer disapproved (saying
"no") of the choice, adding that "this endpoint is highly species-specific." Two other reviewers
considered the selection appropriate but recommended further discussion of possible species
specificity and relevance to humans. Additional specific concerns raised by one or more of these
five reviewers were difficulty in distinguishing bronchiolization as a noncancer endpoint,
likelihood that it is a pre-neoplastic event, unknown relevance to humans, lack of information on
whether the effect on lung tissue was the result of metabolism and thus requiring information on
metabolic competencies of lung cell types, and whether a similar finding may occur in humans.
Of the remaining two reviewers, one stated that the chosen endpoint is valid based on the
route-specific respiratory outcome and the last reviewer indicated that the discussion of the
reasoning for choosing bronchiolization was good but still found it "hard to completely dismiss
the methemoglobinemia that is concordant with effects seen in the oral studies."
Response: There is no available information on metabolism of nitrobenzene in mice and
human lung tissue. Alveolar bronchiolization, which may be a metaplastic change or a
colonization process of alveolar walls with bronchiolar epithelium, has been identified in
response to particulate irritants and oxidant gases in several studies with various species,
including humans. Highlights of these findings, including lack of association between alveolar
bronchiolization and lung neoplasia, were added to section 4.5.2. This endpoint was retained for
deriving an RfC, and additional language was incorporated under section 5.2.1. The revised
Sections 4.5.2 and 5.2.1 address the concerns that were raised regarding the relevance and
justification of bronchiolization as an endpoint for RfC.
Additional internal review clarified that the olfactory epithelial degeneration and loss in
female mice was at least as sensitive an endpoint, in terms of expected human equivalent
concentration, as bronchiolization. The RfC derivation has been modified to include olfactory
degeneration as a co-critical endpoint.
2. If bronchiolization of the alveoli is the most scientifically justifiable endpoint on which to
base the RfC, is the LOAEL-to-NOAEL approach the best method for deriving the RfC?
A-13
-------
Comment: Two reviewers disagreed with the approach, citing difficulty or inability to
compare bronchiolization to other endpoints in the analysis with one suggesting that BMD
modeling can still be done, since not all doses gave a 100% response. Another reviewer
commented that, although using a LOAEL seemed odd, it was difficult to see how BMD
modeling could be used and recommended some discussion about the flat dose-response pattern
and about considering bronchiolization possibly to be a systemic effect.
Response: On further review, neither approach is clearly superior for the
bronchiolization data. Consequently, both BMD modeling and the NOAEL/LOAEL approach
are considered in the document. The BMDLs for male and female mice were lower than
LOAEL/10 (5 ppm/10, or 0.5 ppm) at about 0.1-0.3 ppm. Neither approach validates the other,
but for the purpose of developing a reference value they do not suggest radically different
starting points. Considering the possibly greater crudeness of deriving a suitable POD by
dividing a LOAEL by 10, the assessment has been revised to use BMD modeling to characterize
the bronchiolization data.
Second, the bronchiolization data were problematic to fit with available models because
the response at the low end of a plateau of responses like this one is often underestimated. This
was true for the male mice data, while some of the models did fit the female mice data
adequately. Regardless, there is no way to tell how far into the lower exposures the plateau may
really extend for either data set. On the other hand, the study report did note that the severity of
bronchiolization was dose related, increasing with increasing exposure. Since no data clarifying
severity grading were provided, it is not clear how much of an impact considering severity would
have on the dose-response relationship, although it is possible that the reported low exposure
response was overstated relative to that at the higher exposures. Given these two divergent
possibilities, use of the modeled response is a reasonable compromise. This discussion has been
summarized in the assessment.
If it can be assumed that bronchiolization is a systemic effect rather than a site-of-contact
effect, then EPA's current RfC methodology leads to a lower RfC based on bronchiolization than
developed in the external peer review draft. Nitrobenzene does have systemic effects, so this
assumption is reasonable to consider. If nitrobenzene's impact on the lung is directly
proportional to administered concentration, and due only to systemic distribution of a metabolite,
then the RfC would be estimated at about threefold lower than an RfC based on site-of-contact
effects. This discussion has been summarized in the assessment.
Comment: Another reviewer cited several issues for further discussion, including the
observance of bronchiolization in rats, loss of animals and its impact on the assessment,
providing a better definition for bronchiolization (see Appendix R of CUT [1993]), and including
information on anatomical or species differences based on its occurrence in mice.
A-14
-------
Response: Bronchiolization or lung tumors have not been observed in two strains of rats.
Relevant issues on bronchiolization, including the anatomical and species differences in mice,
were addressed in the response to the comments of the previous charge question (RfC Ql). A
discussion was added to the ToxicologicalReview (sections 4.2.2.2 and 5.2.1) on animal survival
to term. A new definition of bronchiolization was also added (section 4.2.2.2).
Comment: Of the remaining three reviewers, one approved of the LOAEL-to-NOAEL
approach "based on the data and Agency policy"; another cited a lack of sufficient expertise in
modeling analysis to comment on this question but referred to the earlier comment that
"bronchiolization was not the best endpoint for these analyses"; and the last reviewer stated that
"the only comment would be that the dose response for this endpoint was pretty quirky (plateau).
As such, the extrapolation might involve even more uncertainty than usual."
Response: The comment that bronchiolization was not the best endpoint to use was
addressed under an earlier question (RfC Ql) as was the issue regarding the dose-response curve
for bronchiolization.
3. A database UF of 1 was applied in deriving the RfC because the database includes a two-year
(lifetime) chronic inhalation study with an interim (15-month) sacrifice, two-generation
reproductive and developmental inhalation studies, a subchronic (10-week) inhalation
neurotoxicity study, and two 90-day inhalation studies. Is the application of a database UF of 1
scientifically defensible and transparently and objectively described given the available data for
nitrobenzene?
Comment: All reviewers agreed that the selected UF of 1 is reasonable, appropriate, or
justified, but one reviewer recommended additional discussion and justification for asserting that
the literature is complete when the analyses are based on a few limited studies. Another
reviewer commented on "the inter-connectedness of the database UF for inhalation and for oral
routes of exposure" and indicated that the UF choices will be improved if, to the extent possible,
these can be integrated. One reviewer recommended discussing the confidence in choosing a UF
of 1 in addition to the already discussed rationale for having a complete inhalation database.
Response: Section 5.2.4 was modified, and additional discussion was added regarding
completeness of the database as well as rationale and confidence in selecting a database UF of 1.
A-15
-------
D. Carcinogenicity of Nitrobenzene
1. Under EPA's 2005 Guidelines for Carcinogen Risk Assessment (www.epa.gov/iris/backgr-
d.htm), nitrobenzene is classified as likely to be carcinogenic to humans. Have the rationale and
justification for this designation been transparently and objectively described? Do the available
data support the conclusion that nitrobenzene is a likely human carcinogen? If the weight of the
evidence supports the descriptor likely to be carcinogenic to humans, is it appropriate to describe
nitrobenzene as a case that lies on the low end of the range of this descriptor?
Comment: Most reviewers agreed with the classification of carcinogenicity but two
reviewers recommended having additional discussion and justification, including description of
criteria and explanation of why "suggestive" was not used instead. One reviewer provided no
comment on this question.
One reviewer concurred with the statement that nitrobenzene "lies at the low end of the
range" for the "likely to be carcinogenic in humans" descriptor. Four reviewers remarked that
the meaning of the statement was not clear, citing the need for additional information on the
basis of ranking and how it was arrived at, including examples and comparisons with other
chemicals. One reviewer questioned the designation "at the low end of the range" based on
uncertainties about the MO A and findings of tumors at multiple sites in two species with some
showing clear dose-response relationship.
Response: The considerations used to assign the descriptor are listed in section 2.5 of the
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). The carcinogenicity descriptor
"likely to be carcinogenic to humans" was used rather than "suggestive evidence of
carcinogenicity" because nitrobenzene induced tumors in two species and sexes of laboratory
animals, two strains of rats (F344 and male CD), and at multiple sites. The "suggestive evidence
of carcinogenicity" descriptor covers a spectrum of evidence associated with lower levels of
concern for carcinogenicity such as a single positive cancer result (in one species) in an
extensive database that includes negative studies in other species. Upon further evaluation, the
text indicating that the designation lies at the low end of the range of the "likely descriptor" has
been removed. The text in sections 4.6.1 and 6.1.6 has been augmented to better describe the
designation.
2. The two-year inhalation cancer bioassay (CUT, 1993; published as Cattley et al., 1994) was
used for development of an inhalation unit risk (IUR). Is this study the most appropriate
selection for the principal study? Has the rationale for this choice been transparently and
objectively described?
A-16
-------
Comment: All reviewers that commented on this question agreed that the selection was
appropriate, with some reviewers indicating that it was the only viable choice due to lack of
alternatives. Three reviewers stated that the rationale and/or justification were well documented
or well described. One reviewer requested more information on how the study was selected,
while another indicated that including additional information from the original study (e.g., design
and animal loss) would improve transparency.
Response: EPA has followed the guidelines for evaluating cancer bioassays presented in
section 2 of the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). The 2-year
inhalation cancer bioassay, conducted by CUT (Cattley et al., 1994; CUT, 1993) is the only study
available. It is a well-conducted lifetime exposure study that evaluated two species of both sexes
in laboratory animals. The study design and reporting are stated as meeting acceptable GLP
standards. Ranges of exposure concentrations were properly selected based on a subchronic
study by the same route. The study design is described in section 4.2.2.2. Additional
information on study design including animal survival has been added to the Toxicological
Review (sections 4.2.2.2 and 5.2.1).
Comment: A reviewer indicated that further discussion is needed for "the lack of
consistency with the oral studies, even for systemic endpoints."
Response: There are no oral cancer bioassays available. In general, there seems to be
good agreement in the noncancer toxicity outcomes following oral and inhalation exposure of
rats and mice to nitrobenzene. For instance, irrespective of the exposure route, among the most
common findings in the 90-day inhalation and gavage studies were methemoglobinemia,
hemolytic anemia, and target organ toxicity in spleen, liver, kidney, and testis (CUT, 1984; NTP,
1983a). However, there were portal-of-entry findings specific to exposure via the inhalation
route, including bronchial hyperplasia among F344 rats (both sexes) and nasal lymphoid
hyperplasia/inflammation in CD rats (CUT, 1984).
3. Data on hepatocellular tumors in F344 rats were used to estimate the IUR. Are the reasons
for basing the quantitative assessment on hepatocellular tumors in male F344 rats scientifically
justified and transparently described? For calculating the IUR, adenomas and carcinomas were
combined. Has EPA's justification for this approach been objectively and transparently
presented? Is combining adenomas and carcinomas the most scientifically justifiable approach
for these tumors? Please suggest any other scientifically justifiable approaches for calculating
the IUR.
A-17
-------
Note: This charge question overlaps somewhat with Q4 below in that both questions
concerned interpreting multiple tumor types in rats for extrapolation to humans and asked if the
risk estimate should be based on hepatocellular tumors only or if there are other methods for
deriving the IUR. For clarity, issues concerning the interpretation of the hepatocellular tumors
alone are addressed here, while those concerning total risk, involving additional tumor types, are
addressed in the comments and responses for Q4.
Comment: The reviewers generally agreed that hepatocellular tumors in rats were the
most biologically significant result and generally agreed with combining hepatocellular
adenomas and carcinomas for this assessment. One reviewer expressed no opinion, and, among
the rest, three recommended providing additional justification or other information in the
document. One of these three reviewers recommended clarifying why male rat hepatocellular
tumors were selected by stating that this choice is the most health protective and recommended
presenting only the combined modeling results instead of displaying separate modeling for each
tumor type. Another of these reviewers recommended providing more of the raw data on the
histology of hepatocellular tumors provided in the CUT (1993) report, to help understand
whether it is scientifically feasible to combine adenomas and carcinomas and to achieve greater
scientific transparency. The third reviewer provided no specific comments.
Response: The document has been clarified to indicate that, when there are multiple
significant tumor types and when there are no data to the contrary (such as, lack of human
relevance for particular MO As), total risk is one of the considerations weighed in recommending
a risk value. The incidence data for the separate hepatocellular tumor types have been clarified
and the modeling results have been removed.
Key nonneoplastic and neoplastic liver histopathology findings in F344 rats are included
in Tables 4-21 and 4-41 of the Toxicological Review. Information has been added to section
5.3.1 on the histology of hepatocellular tumors, including an additional summary table
(Table 5-5) of incidences of hepatocellular neoplasms, by type and combined, among terminally
sacrificed F344 rats.
Comment: One reviewer critiqued the selection of different BMRs for different
endpoints and suggested having some consistency across data sets by picking a BMR level that
"corresponds to a response level likely to be within the range of the experimental doses."
Response: Although the use of different BMRs was intended to provide estimates of the
low-dose slope most relevant for IUR derivation, there was no difference in ITJRs within each of
these data sets whether the choice of BMR was 5 or 10%. The document has been revised to
indicate that the highest IUR rather than the lowest BMDL denoted the most sensitive response,
A-18
-------
which was hepatocellular tumors. However, all IUR derivations were revised to use a 10% BMR
for consistency.
Comment: One reviewer suggested a different approach to quantifying cancer risk,
commenting that the decision to use linear low-dose extrapolation appeared to be inconsistent
with earlier statements in the document regarding the weak genotoxicity of nitrobenzene and that
the predominately negative genotoxicity evidence was ignored. For this reason, he questioned
the validity of estimating cancer risk based on calculating an IUR and concluded that, instead of
using an IUR for estimating inhalation cancer risk, a POD approach should be used with a UF of
30-100. This reviewer also suggested that evidence on possible DNA damage (in section 4.6.3)
from the in vitro study by Ohkuma and Kawanishi (1999) using calf thymus DNA and
nitrosobenzene in the presence of Cu2+ was misrepresented and stated that, based on the study's
findings, "the conditions under which nitrobenzene would induce DNA damage in vivo are
unlikely to occur." This reviewer added that discussion of the study should probably be under
the genotoxicity study description section (section 4.4.5).
Response: Following the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
EPA uses linear low-dose extrapolation when evidence is insufficient to establish the MOA for a
tumor site, as in this assessment. EPA also notes that there is no empirical indication of low-
dose nonlinearity in the observed range of rat hepatocellular tumors.
The summary (under section 4.6.3) of the in vitro study by Ohkuma and Kawanishi
(1999) has been moved to section 4.4.5 and has been replaced with new language on the MOA.
4. The IUR was calculated from hepatocellular tumors in male F344 rats. The recommended
upper bound estimate on human extra cancer risk from continuous lifetime exposure to
nitrobenzene was calculated to be 3 x 10"5 (jig/m3)"1. Is it scientifically defensible to base the
IUR on liver tumors alone? Have the rationale and justification for this analysis been
transparently and objectively described? Is it more appropriate to calculate the IUR using
combined tumor incidence of liver, thyroid, and kidney tumors in male F344 rats as done in the
alternate derivation of the IUR in the Appendix? If summing of tumors is scientifically justified,
is the method used to sum the tumors supported by the science and the data? If not, what
alternative method should be used?
Comment: Three reviewers indicated a preference for using the most sensitive tumor site
(male rat liver tumors) for the derivation of the IUR over combining the risks from three tumor
sites. Two other reviewers did not reject considering total risk but criticized the estimation
method used, listing pitfalls of combining unit risks from separate tumor types. One of these
reviewers recommended an alternative analysis involving likelihood profile methodology. The
A-19
-------
remaining two reviewers did not respond to this question, with one offering the same response as
for Q3 regarding adenomas and carcinomas.
The rationale given for not combining different tumor types included that it may not
make sense if they have different mechanisms and/or relevance to humans and that a lack of
consistency in effects among rodent species and sexes would not support relevance to humans.
Several reviewers noted that the additional tumor data did not increase the risk estimate when
rounded to one significant digit, 3 xlO"5 (jig/m3)"1.
Response: EPA appreciates these comments and the opportunity to improve the
transparency of the dose-response assessment. The charge question's unfortunate use of the
phrases "summing tumors" and "combined incidence" may have caused confusion based on the
responses to this question. "Summing tumors," suggesting a crude summing of the incidences
before developing the quantitative result, was not a part of this analysis. "Combined incidence,"
which suggests that the incidence of animals with liver, thyroid, or kidney tumors was the basis
of the quantitation, analogous to the consideration of hepatocellular adenomas and carcinomas
discussed in Q3 above, also was not used for this analysis. As described in the external review
draft assessment, it was the extra risk estimates for each site that were summed.
In cases of multisite carcinogens, it has been EPA's practice to characterize total risk, the
risk that an individual could develop any of the critical tumors, not just the most sensitive tumor
type. For example, see the final IRIS assessments for pentachlorophenol (U.S. EPA, 1993),
bromate (U.S. EPA, 2001), 1,3-butadiene (U.S. EPA, 2002b), and 1,2-dibromoethane (U.S. EPA,
2004). However, none of the reviewers considered this aspect of cancer risk assessment in their
comments. Aside from nitrobenzene-specific issues, the panel's discussion for this charge
question was not sufficient to set aside this general practice.
The reviewers who responded negatively to considering nitrobenzene's multiple tumor
types primarily cited lack of relevance of the other two types, kidney and thyroid tumors. EPA
agrees that it would be inappropriate to include risks from tumor types not expected to be
relevant to humans. As discussed in sections 4.6.2 and 4.6.3, however, although the thyroid and
kidney tumors observed in male rats and thyroid tumors in male mice may be suggestive of
rodent-specific MO As, the experimental data do not satisfy the criteria set forth in EPA's
technical reports on Assessment of'ThyroidFollicular Cell Tumors (U.S. EPA, 1998b) and Alpha
2u-globulin: Association with Chemically Induced Renal Toxicity andNeoplasia in the Male Rat
(U.S. EPA, 1991b) to make this determination. Furthermore, EPA does not require site
concordance between species before raising concern for human carcinogenic potential (U.S.
EPA, 2005a), given that some human carcinogens have produced tumors in rodents at different
sites than were observed in humans. Also, mice and rats often do not respond with the same set
of tumor types to any chemical exposure, so lack of consistency between laboratory species does
not inform human relevance. Without data to the contrary, the thyroid and kidney tumors are
A-20
-------
considered relevant to humans. EPA agrees that the justification in section 5.3 was brief and has
provided cross-references to the relevant discussions in chapter 4.
EPA does not agree that it is inappropriate to combine risks if there are different
mechanisms involved. Different mechanisms increase the possibility that the different tumor
types occur independently of each other (e.g., that the occurrence of a kidney tumor, for instance,
is independent of the occurrence of a liver tumor in an individual). The possibility of different
mechanisms for different tumor types is exactly why risks from different tumor types should be
combined, so that the total potential risk—the risk that in an individual could develop a tumor by
any of the possible mechanisms—is not underestimated. This is the basis for using a different
quantitative method than that used for adenomas and carcinomas resulting from the same
mechanism, as for the hepatocellular tumors.
As discussed in the assessment, the National Research Council (NRC, 1994) concluded
that an approach based on counts of animals with one or more tumors (or tumor-bearing
animals), as used by EPA previously, would tend to underestimate overall risk when tumor types
occur independently and that an approach based on combining the risk estimates from each
separate tumor type should be used. The quantitative method used in the total risk analysis in the
nitrobenzene external review draft has been routinely used (e.g., final IRIS assessments for
bromate, 1,3-butadiene, and 1,2-dibromoethane) and received favorable external peer review as
recently as March 2008 (1,2,3-trichloropropane). Meanwhile, EPA has evaluated methods for
combining risks across tumor sites in response to the NRC (1994) recommendations. EPA
applied a Bayesian approach, consistent with the NRC recommendation and consistent with the
likelihood profile method that was suggested by one of the external peer reviewers, to the site-
specific risk estimates for nitrobenzene. Note that, while the analysis recommended by the
reviewer could not be readily carried out without software development, the comparable
Bayesian analysis was straightforward to implement with available software.
The estimate of the IUR for total risk (3.8 x 10"5 (jig/m3)"1) is slightly higher than that in
the external peer review draft (3.49 x 10"5 (jig/m3)"1). EPA also conducted an analysis using the
tumor-bearing animal approach, which yielded an IUR of 4.1 x 10"5 (jig/m3)"1. These approaches
are discussed and results provided in Appendix B-3. All approaches considered for estimating
overall risk provided a higher IUR than liver tumors alone (2.5 x 10"5 (jig/m3)"1), and the
Toxicological Review has been revised to recommend an IUR of 4 x 10"5 (jig/m3)"1, replacing the
IUR of 3 x 10"5 (iig/m3)"1, which was based on liver tumors alone.
Comment: One reviewer remarked that some consideration may be warranted of possible
enhanced susceptibility from early-life exposure to carcinogens, as per U.S. EPA (2005b)
guidance and based on specific examples with other chemicals. This reviewer also noted five
review publications on early-life exposure and susceptibility to carcinogens for consideration in
the Toxicological Review.
A-21
-------
Response: In general, differences in susceptibility with respect to early life stages may
not be fully accounted for because cancer slope factors are usually based on effects observed
following exposures to adult humans or sexually mature animals. The Supplemental Guidance
for Assessing Susceptibility from Early-life Exposure to Carcinogens (U.S. EPA, 2005b)
provides an approach for developing lifetime cancer risk estimates that account for possible age-
related differences in susceptibility and exposure. The guidance takes into consideration the
MO A, if known, information on chemical-specific susceptibility differences between adults and
juveniles, if available, and the methodology (linear versus nonlinear) used for estimating cancer
risk, in addition to applying age-dependent adjustment factors (ADAFs) for mutagenic
carcinogens. According to the guidance, unless chemical-specific data are available, ADAFs are
currently only applied to carcinogens that are known to act via a mutagenic MO A.
All five review articles suggested by the peer reviewer are not specific to nitrobenzene.
Among the covered issues and examples were genotoxic carcinogens as well as factors that
influence susceptibility at different prenatal and/or neonatal stages of development, including
some of the known molecular mechanisms, age-related differences in drug metabolizing
enzymes, oncogenes, tumor-suppressor genes, and certain growth regulatory genes.
A-22
-------
PUBLIC COMMENTS
No public comments were submitted on this assessment.
A-23
-------
References to Appendix A
Albrecht, W; Neumann, HG. (1985) Biomonitoring of aniline and nitrobenzene. Hemoglobin binding in rats and
analysis of adducts. Arch Toxicol 57:1-5.
Bond, JA; Chism, JP; Rickert, DE; et al. (1981) Induction of hepatic and testicular lesions in Fischer 344 rats by
single oral doses of nitrobenzene. Fundam Appl Toxicol 1:389-394.
Burns, LA; Bradley, SG; White, KL, Jr; et al. (1994) Immunotoxicity of nitrobenzene in female B6C3F1 mice.
Drug Chem Toxicol 17:271-315.
Cattley, RC; Everitt, JI; Gross, EA; et al. (1994) Carcinogenicity and toxicity of inhaled nitrobenzene in B6C3F1
mice and F344 and CD rats. Fundam Appl Toxicol 22:328-340.
CUT (Chemical Industry Institute of Toxicology). (1984) Ninety day inhalation toxicity study of nitrobenzene in
F344 rats, CD rats, and B6C3F1 mice. Chemical Industry Institute of Toxicology, Research Triangle Park, NC;
Docket No. 12634. Submitted under TSCA Section 8D; EPA Document No. 878214291; NTIS No. OTS0206507.
CUT (Chemical Industry Institute of Toxicology). (1993) Initial submission: a chronic inhalation toxicity study of
nitrobenzene in B6C3F1 mice, Fischer 344 rats and Sprague-Dawley (CD) rats. Chemical Industry Institute of
Toxicology, Research Triangle Park, NC. EPA Document No. FYI-OTS-0794-0970; NTIS No. OTS0000970.
Dodd, DE; Fowler, EH; Snellings, WM; et al. (1987) Reproduction and fertility evaluations in CD rats following
nitrobenzene inhalation. Fundam Appl Toxicol 8:493-505.
DuPont. (1981) Inhalation median lethal concentration (LC50) with cover letter. Haskell Laboratory for
Toxicology and Industrial Medicine, E.I. du Pont de Nemours and Company, Newark, DE. Submitted under
TSCA Section 8D; EPA Document No. 878220423; NTIS No. OTS0215040.
Goldstein, RS; Rickert, DE. (1984) Macromolecular covalent binding of [14C]nitrobenzene in the erythrocyte and
spleen of rats and mice. Chem Biol Interact 50:27-37.
Heneghan, JB. (1984) Physiology of the alimentary tract. In: Coates, ME; Gustafsson, BE; eds. The germ-free
animal inbiomedical research. London, UK: Laboratory Animals, Ltd.; pp 169-191.
Medinsky, MA; Irons, RD. (1985) Sex, strain, and species differences in the response of rodents to nitrobenzene
vapors. In: Rickert, DE; ed. Toxicity of nitroaromatic compounds. New York, NY: Hemisphere Publishing
Corporation; pp. 35-51.
Mitsumori, K; Kodama, Y; Uchida, O; et al. (1994) Confirmation study, using nitrobenzene, of the Combined
Repeat Dose and Reproductive/Developmental Toxicity Test protocol proposed by the Organization for Economic
Cooperation and Development (OECD). J Toxicol Sci 19:141-149.
Morrissey, RE; Schwetz, BA; Lamb, JC, IV; et al. (1988) Evaluation of rodent sperm, vaginal cytology, and
reproductive organ weight data from National Toxicology Program 13-week studies. Fundam Appl Toxicol
11:343-358.
NRC (National Research Council). (1994) Science and judgment. Washington, DC: National Academy Press.
NTP (National Toxicology Program). (1983) Report on the subchronic toxicity via gavage of nitrobenzene
(C60082) in Fischer 344 rats and B6C3F1 mice [unpublished]. Prepared by the EG&G Mason Research Institute,
Worcester, MA, for the National Toxicology Program, National Institute of Environmental Health Services,
Public Health Service, U.S. Department of Health and Human Services, Research Triangle Park, NC; MRI-NTP
08-83-19.
NTP (National Toxicology Program). (1989) Toxicology and carcinogenesis studies of para-chloroaniline
hydrochloride (CAS No. 20265-96-7) in F344/N rats and B6C3F1 mice (gavage studies). Public Health Service,
U.S. Department of Health and Human Services; NTP TR351. Available from the National Institute of
A-24
-------
Environmental Health Sciences, Research Triangle Park, NC and online at
http://ntp.niehs.nih.gov/ntpweb/index.cfm?objectid=D16D6C59-FlF6-975E-7D23D1519B8CD7A5.
Ohkuma, Y; Kawanishi, S. (1999) Oxidative DNA damage by a metabolite of carcinogenic and reproductive toxic
nitrobenzene in the presence of NADH and Cu(II). Biochem Biophys Res Commun 257:555-560.
Shimo, T; Onodera, H; Matsushima, Y; et al. (1994) [A 28-day repeated dose toxicity study of nitrobenzene in
F344 rats]. Eisei Shikenjo Hokoku 112:71-81.
Smith, RP. (1996) Toxic responses of the blood. In: Klaassen, CD; ed. Casarett and Doull's toxicology: the basic
science of poisons. 5th edition. New York, NY:McGraw-Hill; pp. 335-354.
Tingle, MD; Mahmud, R; Maggs, JL; et al. (1997) Comparison of the metabolism and toxicity of dapsone in rat,
mouse, and man. J Pharmacol Exp Ther 283:817-823.
U.S. EPA (Environmental Protection Agency). (1991) Alpha 2u-globulin: association with chemically induced renal
toxicity and neoplasia in the male rat. Risk Assessment Forum, Washington, DC; EPA/625/3-91/019F. Available
online at http://www.epa.gov/nscep.
U.S. EPA (Environmental Protection Agency). (1993) IRIS Summary for pentachlorophenol. Available online at
http://www.epa.gov/ncea/iris/subst/0086.htm.
U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F. Available from the National Technical
Information Service, Springfield, VA, PB2000-500023, and online at
http://cfpub.epa. gov/ncea/raf/recordisplay. cfm?deid=71993.
U.S. EPA (Environmental Protection Agency). (1998) Assessment of thyroid follicular cell tumors. Risk
Assessment Forum, Washington, DC; EPA/630/R-97/002. Available from the National Technical Information
Service, Springfield, VA, PB98-133119, and online at http://nepis.epa.gov/EPA/html/Pubs/pubtitleORD.htm.
U.S. EPA (Environmental Protection Agency). (2001) Toxicological review of bromate. Available online at
http://www.epa. gov/ncea/iris/toxreviews/1002-tr.pdf.
U.S. EPA (Environmental Protection Agency). (2002a) A review of the reference dose concentration and reference
concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-02/002F. Available online at
http://cfpub.epa. gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (Environmental Protection Agency). (2002b) Health assessment of 1,3-butadiene. Available online at
http://www.epa.gov/ncea/iris/supdocs/buta-sup.pdf.
U.S. EPA (Environmental Protection Agency). (2004) Toxicological review of 1,2-dibromoethane. Available
online at http://www.epa.gov/ncea/iris/toxreviews/0361-tr.pdf.
U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Federal Register
70(66): 17765-18717. Available online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2006) Science policy council handbook: peer review. 3rd edition.
Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-06/002. Available
online at http://www.epa.gov/OSA/spc/2peerrev.htm.
Vasquez, GB; Reddy, G; Gilliland, GL; et al. (1995) Dinitrobenzene induces methemoglobin formation from
deoxyhemoglobin in vitro. Chem Biol Interact 96:157-171.
A-25
-------
APPENDIX B: DOSE-RESPONSE MODELING
APPENDIX B-l: Dose-Response Modeling for Derivation of an RfD for Nitrobenzene
APPENDIX B-2: Dose-Response Modeling for Derivation of an RfC for Nitrobenzene
APPENDIX B-3: Dose-Response Modeling of Carcinogenicity Data for Nitrobenzene
B-l
-------
APPENDIX B-l: Dose-Response Modeling for Derivation of an RfD for Nitrobenzene
B-l.l. METHODS
The models in U.S. EPA's BMDS (version 1.4.1c) were fit to multiple data sets presented
in a 90-day study of gavage exposure to nitrobenzene in F344 rats and B6C3F1 mice (NTP,
1983a). The endpoints considered for modeling were splenic congestion, methemoglobin
(metHb), and reticulocyte levels in male and female rats (Table B-l.l).
Table B-l.l. Summary of splenic congestion, reticulocyte count (%) and
metHb levels (%) in male and female F344 rats exposed by gavage to
nitrobenzene for 90 days
Dose
(mg/kg-day)
Splenic congestion, severity
greater than minimal3
Hematology endpoints
nb
Reticulocytes
(%)°
MetHb
(%)°
Males
0
9.38
18.75
37.5
75
150
0/10
0/10
0/10
0/10
5/10
10/10
10
10
10
10
10
1
2.23 ± 0.44
2.62 ±0.45
3.72±0.65d
4.75 ± 0.62d
6.84 ± 0.72d
15
1.13 ±0.58
2.75±0.58d
4.22±1.15d
5.62±0.85d
7.31±1.44d
12.22
Females
0
9.38
18.75
37.5
75
150
0/10
1/10
3/10
5/10
8/10
9/10
10
10
10
10
10
7
2.60 ±0.37
3.69±0.32d
4.75±0.68d
6.28±0.90d
8.72 ± 1.49d
32.07 ±3.56d
0.94 ±0.03
2.06±0.45d
3.62±1.09d
5.27 ± 0.76d
6.85±2.25d
12.77 ± 1.83d
aAll animals, including early deaths, were subject to histopathologic examination
bNumber of surviving animals at termination.
'Values are means ± SD.
""Significantly different from controls, as calculated by the authors.
Source: NTP(1983a).
The dose levels used were the administered doses reported in the study, since there was
no information supporting other dose metrics. In accordance with the U.S. EPA (2000b) BMD
technical guidance, biologically relevant response levels were considered for developing RfDs
where possible. Insufficient information was available to identify minimally adverse levels of
response for increased reticulocyte counts and splenic congestion. For methemoglobinemia,
BMRs based on 3% metHb being the upper limit of a normal range for humans was considered,
as well as a 2 SD increase in the control mean (see section 5.1.2), BMRs of a 10% increase in
B-2
-------
extra risk and a change in the mean equal to 1 SD from the control mean were estimated for
dichotomous and continuous data, respectively, as standard points of comparison.
Models were run using the default restrictions on parameters provided in BMDS,
primarily to avoid biologically implausible dose-response shapes. Specifically, the restrictions
included limiting shape parameters to be ^1 for the gamma, log-logistic, log-probit, and Weibull
(dichotomous) models and power and Hill (continuous) models and allowing only monotonic fits
of the multistage (dichotomous) and polynomial (continuous) models.
For continuous data, the assumption of constant variance across dose groups was
evaluated for each model. If the variances were statistically homogenous (p > 0.05), then fit of
the various models to the means was evaluated while assuming a constant variance. If variances
were heterogeneous, then variance was modeled as a power function of the mean. For data sets
where the fit of the variance model was adequate (p > 0.05), the fit to the means of the various
models was evaluated contingent on the use of the variance model.
B-1.2. RESULTS
The BMD modeling results are summarized in Table B-1.2. This table shows the final
BMDs and BMDLs derived for each endpoint modeled for male and female rats. The remainder
of this section shows detailed summaries of the modeling results for splenic congestion, metHb
concentration, and reticulocyte count.
Table B-1.2. Summary of PODs derived from BMD modeling of
NTP (1983a) bioassay data for male and female rats exposed by gavage to
nitrobenzene for 90 days
Endpoint
(data type)
Spleen congestion
(dichotomous)
MetHb (%)
(continuous)
Reticulocytes (%)a
(continuous)
Sex
M
F
Ma
Fa
Ma
Fa
Model
Gamma, multistage
Multistage, log-probit
Hill
Hill
Linear
Hill
BMR
10% extra risk of mild
or moderate congestion
1 SD = 0.5%
Point BMR = 2%b
2 SD = 1.0%
1 SD = 0.4%
Point BMR =1.9%c
0.8%
1 SD = 0.6%
1 SD = 0.3%
BMD
(mg/kg-day)
54.6
7.8
3.0
4.9
5.7
4.7
8.2
7.2
9.4
2.7
BMDL
(mg/kg-day)
37.8
5.6
1.8
3.2
3.9
3.1
6.3
5.2
7.9
1.8
Output
I
II
III
IV
V
VI
aHigh-dose group (150 mg/kg-day) dropped.
bPoint BMR = the mean of a distribution of metHb with 10% of the (exposed) population exceeding 3% metHb.
B-3
-------
B-l.2.1. Male F344 rat spleen congestion
Adequate fit (p > 0.1) with all of the models.
Gamma and multistage gave best fits (low Akaike Information Criteria [AICs], most biologically
plausible dose-response shapes); gamma output provided.
BMD = 51.9-54.6 mg/kg-day
BMDL = 35.0-37.8 mg/kg
Model fit to means
Gamma (power >1)
Logistic
Log-logistic (slope >1)
1 -Stage multistage (5th degree)
Probit
Log-probit (slope >1)
Weibull (power >1)
x2
0.05
0.01
O.01
0.23
O.01
0.01
O.01
df
5
4
5
5
4
4
4
p Value
for model
fit
1.00
1.00
1.00
1.00
1.00
1.00
1.00
AIC
15.96
17.86
15.86
16.3
17.86
17.86
17.86
BMD
(mg/kg-day)
54.6
70.6
66.4
51.9
66.5
64.4
66.6
BMDL
(mg/kg-day)
37.8
41.6
40.6
35.0
38.9
39.4
36.9
OUTPUT I. Male F344 rat spleen congestion
Gamma Model. (Version: 2.7; Date: 01/18/2007)
Input Data File: G:\_BMDS\NB_RAT_SPLEEN.(d)
Gnuplot Plotting File: G:\_BMDS\NB_RAT_SPLEEN.plt
Tue Feb 12 10:50:13 2008
BMDS MODEL RUN
The form of the probability function is:
Dependent variable = sp cng m
Independent variable = mg kg d
Power parameter is restricted as power ^1
Total number of observations = 6
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Parameter Estimate
B-4
-------
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -6.93147 6
-6.98023 1 0.0975094 5
-33.7401 1 53.6173 5
P-value
Fitted model
Reduced model
Goodness of Fit
Dose
Prob.
d.f. =
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
BMDL =
0.8
c
o
••§ 0.4
ro
0.2
Gamma Multi-Hit Model with 0.95 Confidence Level
Gamma Multi-Hit
BMDL
jBiyip
20
40
60
80
dose
100
120
140
10:5002/122008
B-5
-------
B-l.2.2. Female F344 rat spleen congestion
Adequate fit (p > 0.1) with all of the models, all have plausible shapes and small residuals.
Multistage and log-probit models gave best fits (lowest AICs, essentially the same); BMDs and
BMDLs averaged.
BMD = 7.8mg/kg-day
BMDL = 5.6 mg/kg-day
Model fit to means
Gamma (power >1)
Logistic
Log-logistic (slope >1)
Multistage, 1-stage
Probit
Log-probit (slope >1)
Weibull (power >1)
T2
0.60
0.14
5.80
0.59
5.99
0.23
0.60
df
4
4
4
5
4
5
4
p Value for
model fit
0.96
1.00
0.21
0.99
0.20
1.00
0.96
AIC
53.69
53.24
58.99
51.71
59.65
51.31
53.70
BMD
(mg/kg-day)
6.5
8.7
16.5
5.9
16.7
9.7
6.1
BMDL
(mg/kg-day)
4.2
3.0
11.5
4.2
12.1
6.9
4.2
OUTPUT II. Female F344 rat spleen congestion
The form of the probability function is:
P[response] = backgr
-beta'
The parameter betas
-ound + (1-background) * [1-EXP (
-betal*dose^l-beta2*dose^2-beta3*dose^3-beta4*dose^4-beta5*dose^5)]
Dependent variable = sp cong f
Independent variable = mg kg d
Total number of observations = 6
Total number of records with missing values = 0
Total number of parameters in model = 6
Total number of specified parameters = 0
Degree of polynomial = 5
Default Initial
Background =
Beta(l) =
Beta(2) =
Beta(3) =
Beta(4) =
Beta(5) =
Parameter Values
0.0671025
0.0160015
0
0
0
0
B-6
-------
Beta(1)
Variable Estimate Std. Err.
Background
Beta(l)
Beta(2)
Beta(3)
Beta(4)
Beta(5)
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
Param's Deviance Test d.f.
P-value
Dose
9.312
d.f. =
P-value =
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0. 95
5.90543
4.17053
11.3289
Taken together, (4.17053, 11.3289) is a 90
interval for the BMD
% two-sided confidence
B-7
-------
Multistage Model with 0.95 Confidence Level
T3
0.8
0.6
o
'•0 0.4
CO
0.2
Multistage
&
3Mp,L BMD ,
o
14:0302/01 2008
20
40
60 80
dose
100
120
140
Probit Model. (Version: 2.9; Date: 09/23/2007)
Input Data File: C:\BMDS\UNSAVED1.(d)
Gnuplot Plotting File: C:\BMDS\UNSAVEDl.plt
Fri Feb 01 14:05:29 200E
The form of the probability function is:
where CumNorm(.) is the cumulative normal distribution function
Total number of observations = 6
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0
intercept = -3.60482
slope = 1
symptotic Correlation Matrix of Parameter Estimates
( ^^^ The model parameter(s) -background -slope
B-8
-------
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-24.5458 6
-24.6565 1 0.221419 5
-41.0539 1 33.0162 5
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 9.68021
BMDL = 6.88503
B-9
-------
Probit Model with 0.95 Confidence Level
T3
0.8
0.6
o
'•0 0.4
CO
0.2
Probit
,. IT,
BMDU BMD
0
14:0502/01 2008
20
40
60 80
dose
100 120
140
B-10
-------
B-l.2.3. Male F344 rat metHb
Nonhomogenous variance.
No reasonable fits, including high dose, which only had one observation; dropped high-dose
group.
Adequate fit (p > 0.1) to means with Hill model.
Model fit to means
(without high dose
group)
Linear and higher order
polynomial models
(coefficients restricted to
be positive);
Power model (power >1)
Hill (power >1)
df
3
1
p Value
for
variance
model
0.27
0.27
p Value
for
overall
model
fit
<0.01
0.42
AIC for
fitted
model
66.09
46.38
BMR
1 SD
1 SD = 0.5%a
Point BMR = 2%b
2 SD = 1.0%
BMD
(mg/kg-day)
5.5
3.0
4.9
5.7
BMDL
(mg/kg-day)
4.6
1.8
3.2
3.9
aAs estimated for the control group by the fitted variance model. Note that the Hill model run used 0.5% as an
absolute deviation from the control mean rather than a BMR of 1 SD because of a software issue that prevented
estimation of a BMDL; the approaches are equivalent.
bThe BMR corresponds to the mean metHb level at which 10% (of those exposed) would be expected to exceed 3%
metHb, an upper limit on the normal range in humans.
OUTPUT III. Male F344 rat metHb
Hill Model. (Version: 2.11; Date: 01/18/2007)
Input Data File: G:\_BMDS\NB_RMALE_METHG.(d)
Gnuplot Plotting File: G:\_BMDS\NB_RMALE_METHG.plt
Tue Feb 12 17:04:56 2008
Total number of dose groups = 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
Default Initial Parameter Values
lalpha = -0.044823
rho = 0
intercept = 1.131
v = 6.176
n = 7.43183
B-ll
-------
rho intercept v n k
lalpha 1 -0.88 -0.27 0.087 -0.085 0.055
rho -0.88 1 0.25 -0.099 0.098 -0.063
intercept -0.27 0.25 1 -0.15 0.18 -0.031
-0.099 -0.15 1 -0.94 0.98
0.098 0.18 -0.94 1 -0.94
-0.063 -0.031 0.98 -0.94 1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
lalpha -1.47644 0.434447 -2.32794
rho 0.924156 0.300902 0.3344
intercept 1.14391 0.159504 0.831293
v 8.33828 2.62299 3.1973
n 1.17121 0.377134 0.432036
k 31.363 19.3683 -6.59809
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(1)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)))
Model A3 uses any fixed variance parameters that
were specified by the user
Likelihoods of Interest
B-12
-------
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
-2*log(Likelihood Ratio) Test df
Test 1
Test 2
Test 3
Test 4
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
Benchmark Dose Computation
Risk Type
Confidence level = 0.95
BMD = 2.99149
BMDL = 1.84409
Confidence level = 0.95
BMD = 4.92646
BMDL = 3.18476
Confidence level = 0.95
BMD = 5.71947
BMDL = 3.93129
8
7
6
CO
Hill Model with 0.95 Confidence Level
0BMPL,
10
20
30
40
dose
50
60
70
B-13
-------
B-l.2.4. Female F344 rat metHb
Nonhomogenous variance.
Adequate fit (p > 0.1) to means with Hill model after dropping high dose and transforming data
(see below).
Model fit to means
Linear and higher order
polynomial models
(coefficients restricted
to be positive); power
(power >1)
Hill (power >1)
df
4
2
p Value
for
variance
model fit
0.01
0.01
High dose group dro
Linear and higher order
polynomial models
(coefficients restricted
to be positive); power
(power >1)
Hill (power >l)a
1
0.1
p Value
for
overall
model fit
0.01
0.02
AIC for
fitted
model
23.2
36.2
BMR
0.4%a
0.4%a
BMD
(mg/kg-day)
4.7
3.0
BMDL
(mg/kg-day)
3.1
2.3
pped; constant (0.9) subtracted from each datapoinf
0.41
46.23
0.4%a
0.4%a
Point BMR =1.9%c
0.8%
4.5
8.2
7.2
2.6
6.3
5.2
aThe SD of the control group was unusually low, at 0.03%, compared with other data sets. A value closer to the SD
of the low-dose group was used for the BMR, as an absolute deviation from the control mean.
bDropping the high dose alone did not sufficiently improve the fit of the variance model or the continuous data
models. The variance model depends on the values of the means; transformation of the data by subtracting a
constant value of 0.9 from each mean produced an adequate fit of the variance model while preserving the overall
dose-response relationship. The p value for the variance model fit was 0.097 and reproduced the observed SDs
sufficiently well in the lower dose groups.
°As for the male metHb BMR, this BMR corresponds to the mean metHb level at which 10% (of those exposed)
would be expected to exceed 3% metHb, an upper limit on the normal range in humans. Because 0.9% was
subtracted from all metHb values in this analysis, the point estimate for the run was 1.0% (1.9-0.9).
OUTPUT IV. Female F344 rat metHb
Hill Model. (Version: 2.11; Date: 01/18/2007)
Input Data File: G:\_BMDS\NB_NTP83_FRATS_METHG_WO6.(d)
Gnuplot Plotting File: G:\_BMDS\NB_NTP83_FRATS_METHG_WO6.plt
Thu Feb 07 11:10:38 2008
B-14
-------
Total number of dose groups = 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
Default Initial Parameter Values
lalpha = 0.339891
0
0.041
5. 91
0.338927
Asymptotic Correlation Matrix of Parameter Estimates
lalpha rho intercept v
lalpha 1 -0.29 -0.4 -0.047
rho -0.29 1 0.56 -0.079
intercept -0.4 0.56 1 -0.035
v -0.047 -0.079 -0.035 1
n 0.018 0.011 0.012 -0.87
k 0.011 -0.036 -0.022 0.97
Variable
lalpha
rho
intercept
Lower Conf. Limit
1.37929
0.023501
4.12763
0.870274
8.61301
Upper Conf. Limit
-1.3519
1.90506
0.0594456
10.1359
2.24933
45.8014
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)))
Model A3 uses any fixed variance parameters that
were specified by the user
B-15
-------
Model Log(likelihood) # Param's
Al -30.863270 6
A2 13.937796 10
A3 10.783003 7
fitted 10.109497 6
R -68.956119 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
-2*log(Likelihood Ratio) Test df
Test 1
Test 2
Test 3
Test 4
The p-value for Test 2 is less than .1. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is less than .1. You may want to consider a
different variance model
Benchmark Dose Computation
Risk Type
Confidence level = 0.95
BMD = 4.4529
BMDL = 2.56818
Confidence level = 0.95
BMD = 8.24303
BMDL = 6.34379
Risk Type
Confidence level = 0.95
BMD = 7.22256
BMDL = 5.2537
B-16
-------
Hill Model with 0.95 Confidence Level
8
7
6
o)5
84
12
1
BI/IDl .
10
20
30
40
dose
50
60
70
B-17
-------
B-l.2.5. Male F344 rat reticulocytes
Homogenous variance (p = 0.43).
Adequate fit (p > 0.1) to means with linear model.
Model fit to means
Linear and higher order
polynomial models; power model
Hill model
df
3
3
p Value for
model fit
0.16
0.10
AIC for fitted
model
2.52
4.08
BMD
(mg/kg-day)
9.43
9.39
BMDL
(mg/kg-day)
7.94
6.07
OUTPUT V. Male F344 rat reticulocytes
Polynomial Model. (Version: 2.11; Date: 01/18/2007)
Input Data File: G:\_BMDS\UNSAVED1.(d)
Gnuplot Plotting File: G:\_BMDS\UNSAVEDl.plt
Tue Feb 12 17:30:59 2008
Dependent variable = MEAN
Independent variable = COLUMN1
rho is set to 0
Signs of the polynomial coefficients are not restricted
A constant variance model is fit
Total number of dose groups = 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
Default Initial Parameter Values
alpha = 0.34428
rho = 0 Specified
beta_0 = 2.2845
beta 1 = 0.0621333
Asymptotic Correlation Matrix of Parameter Estimates
alpha
beta_0
beta 1
Parameter Estimates
B-18
-------
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
alpha 0.343138 0.0686276 0.20863 0.477646
beta_0 2.2845 0.120762 2.04781 2.52119
beta 1 0.0621333 0.00312416 0.0560101 0.0682566
Model Descriptions for likelihoods calculated
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(1)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model
Al
A2
A3
fitted
R
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Test 1
Test 2 3.82122 4
Test 3 3.82122 4
Test 4
B-19
-------
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 = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 9.42779
BMDL = 7.93875
Linear Model with 0.95 Confidence Level
5
o:
ro 4
0) ^
Linear
- r"'
BMDL
0
<
.
BMD , , , , , , :-
10 20 30 40 50 60 70
dose
17:3002/122008
B-20
-------
B-l.2.6. Female F344 rat reticulocytes
Nonhomogenous variance; BMDS variance model had adequate fit.
BMR=1 SD = 0.3%.
Model fit to means
Linear
4 Degree polynomial (betas >0)
Power (power >1)
Hill
df
4
1
3
2
p Value for
variance model
0.29
0.29
0.29
0.01
p Value for
model fit
<0.01
0.01
O.01
0.96
AIC for
fitted model
77.30
55.59
76.82
38.2
BMD
(mg/kg)
2.36
4.68
3.50
6.8
BMDL
(mg/kg)
2.86
3.15
2.17
4.9
Dropped high-dose group
Linear, polynomial
Hill3
3
1
0.354
0.354
0.05
0.58
19.2
15.77
3.65
2.69
2.62
1.80
aNote that, in order to obtain an adequate fit from this model, starting parameters were adjusted from those supplied
by the software package (see "User Inputs Initial Parameter Values" in the output below).
OUTPUT VI. Female F344 rat reticulocytes
Hill Model. (Version: 2.12; Date: 02/20/2007)
Input Data File: C:\BMDS\UNSAVED1.(d)
Gnuplot Plotting File: C:\BMDS\UNSAVEDl.plt
Wed Feb 20 21:45:52 200E
The form of the response function is:
Y[dose] = intercept + v^dose^n/(k^n + dose^n)
Dependent variable = MEAN
Independent variable = mg_kg_d
Power parameter restricted to be greater than 1
The variance is to be modeled as Var(i) = exp(lalpha + rho ^ ln(mean(i)))
Total number of dose groups = 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 Inputs Initial Parameter Values
lalpha = 2
rho = 1
intercept = 0.1
v = 7
n = 2
k = 50
Asymptotic Correlation Matrix of Parameter Estimates
lalpha rho intercept v
lalpha 1 -0.96 -0.17 0.092
B-21
-------
rho
intercept
-0.099 0.078
-0.14 0.24
1 -0.95
1
1 -0.97
95.0% Wald Confidence Interval
Variable
lalpha
rho
intercept
v
n
k
Lower Conf. Limit
-6.19188
1.55334
2.42703
-8.58515
0.52933
Upper Conf. Limit
-3.30114
3.33228
2.7 9853
37.9369
1.591
371.142
Model Descriptions for likelihoods calculated
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)))
Model A3 uses any fixed variance parameters that
were specified by the user
Model R: Yi = Mu + e(i)
Var{e(i)} = S i gma ^ 2
Model
Al
A2
A3
fitted
R
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)
B-22
-------
Tests of Interest
Test -2*log(Likelihood Ratio) Test df
The p-value for Test 3 is greater than .1. The modeled variance 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 = 2.69076
BMDL = 1.79887
Hill Model with 0.95 Confidence Level
10
9
8
O.
CO _
-------
APPENDIX B-2: Dose-Response Modeling for Derivation of an RfC for Nitrobenzene
B-2.1. METHODS
Data considered were exposure-related increases in olfactory degeneration and alveolar
bronchiolization following 2-year exposure in mice (Cattley et al., 1994; CUT, 1993). See
section 5.2 for the data and more information on endpoint selection; the data are also repeated in
the model outputs in section B-2.3. See section B-l.l for additional information regarding
modeling methods.
B-2.2. RESULTS
The BMD modeling results for olfactory epithelium degeneration are summarized in
Table B-2.1, and for bronchiolization are summarized in Table B-2.2. Detailed model output for
both endpoints is in Section B-2.3.
Table B-2.1. Modeling results for olfactory degeneration in mice; bioassay
data from Cattley et al. (1994) and CUT (1993)
Sex
Males3
Females3
Model
Probit
Gamma, multistage (1 -stage), Weibull
p Value
0.38
0.50
BMC10
(ppm)
12.3
1.75
BMCL10
(ppm)
10.0
1.42
aHigh-dose group excluded.
Table B-2.2. Modeling results for bronchiolization in mice; bioassay data
from Cattley et al. (1994) and CUT (1993)
Sex
Males
Females
Model
Log-logistic3
Log-probit
Log-logistic
p Value
0.0276
0.12
0.80
AIC
105.1
49.4
49.4
BMC10
(ppm)
0.134
0.40
0.18
BMCL10
(ppm)
0.083
0.278
0.022
3High-dose group excluded.
B-24
-------
B-2.2.1. Olfactory degeneration in male mice following chronic nitrobenzene
inhalation
None of the models fit (p < 0.05). Dropping the high dose left only one dose level with a
response greater than control, at approximately 50%. All models fit (p > 0.1), with all BMDLioS
>8 ppm. Output from the best-fitting model (lowest AIC), probit, is provided below.
Probit Model. (Version: 2.9; Date: 09/23/2007)
Input Data File: C:\BMDS\NB_CIIT_MICE_RFC_WO_GP4.(d)
Gnuplot Plotting File: C:\BMDS\NB_CIIT_MICE_RFC_WO_GP4.plt
Fri Feb 22 15:43:37 200E
The form of the probability function is:
P[response] = CumNorm(Intercept + Slope*Dose) ,
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = olf degen m
Independent variable = ppm
Slope parameter is not restricted
Total number of observations = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-00<:
Parameter Convergence has been set to: le-008
Asymptotic Correlation Matrix of Parameter Estimates
intercept
1
-0. 9
Parameter Estimates
Variable Estimate
intercept -2.4247
slope 0.0958901
B-25
-------
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
J72
Est. Prob.
1
1
32
d.f. =1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 12.3212
BMDL = 9.99197
0.6
0.5
1 0.4
C 0.3
O
£ 0.2
0.1
0
Probit
Probit Model with 0.95 Confidence Level
BMD
MD
10 15
dose
20
25
15:4302/222008
B-26
-------
B-2.2.2. Olfactory degeneration in female mice following chronic nitrobenzene
inhalation
Gamma, multistage, and Weibull models shared the lowest AIC; multistage output is provided
below.
BMCio= 1.75ppm
BMCLio= 1.42ppm
Model used3
Gamma
Multistage, 1 -stage
Weibull
Log-probit
Log-logistic
p Value
0.50
0.50
0.50
0.13
1.00
AIC
149.64
149.64
149.64
152.06
150.32
BMC10
(ppm)
1.75
2.82
1.44
BMCL10
(ppm)
1.42
2.29
0.79
"High-dose group excluded.
Multistage Model. $Revision: 2.1 $ $Date: 2000/08/21 03:38:21 $
Input Data File: C:\BMDS\UNSAVED1.(d)
Gnuplot Plotting File: C:\BMDS\UNSAVEDl.plt
Fri Jun 16 12:34:52 2006
Observation # < parameter # for Multistage model.
The form of the probability function is:
Dependent variable = Incidence
Independent variable = Concentration
Total number of observations = 3
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
B-27
-------
Default Initial Parameter Values
Background = 0
Beta(l) = 0.0814877
Beta(2) = 0
Beta(3) = 0
k The model parameter(s) -Background -Beta(2) -Beta(3)
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
Beta(1)
1
Parameter Estimates
Model Log(likelihood) Deviance Test DF
Full model -73.159
Fitted model -73.8205 1.3229 2
Reduced model -115.963 85.6088 2
i :
i :
i :
1
0.0000
2
5.0000
3
25 . 0000
0.
0.
0 .
.0000
.2601
.7783
0.000 0
15.607 19
49.031 47
52
60
63
0.
0.
-0 .
.000
.294
.187
Chi-square = 1.38 DF = 2 P-value = 0.502J
B-28
-------
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
BMDL =
Multistage Model with 0.95 Confidence Level
.
"o
I
j
CO
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Multistage
BMD Lower Bound
BMDLl |BMD
0
12:3406/162006
10 15
dose
20
25
B-29
-------
B-2.2.3. Alveolar bronchiolization in male mice
The best fitting model for each subset considered below did not meet the usual p > 0.1 goodness-
of-fit criterion; only the model results for the best fitting model in each case is reported (AICs
are not shown because they are not comparable between nonidentical datasets.)
The fit for the subset excluding the high dose was considered best overall because it came closest
to fitting the low dose response.
Conditions
All dose groups
High-dose group dropped
Highest two groups dropped
Model
Log-logistic3
Log-logistic3
Multistage, 1°
p Value
0.0245
0.0276
b
BMC10
0.156
0.134
0.262
BMCL10
0.102
0.083
0.204
aAll other models had goodness-of-fit/> values <0.01.
bFit is "perfect" because there were only two points; all other models were unsuitable, with more
parameters than this reduced data set could support.
Logistic Model. (Version: 2.8; Date: 01/18/2007)
Input Data File: G:\_BMDS\NB_CIIT_M_MICE_WO4TH_GP.(d)
Gnuplot Plotting File: G:\_BMDS\NB_CIIT_M_MICE_WO4TH_GP.plt
Thu Mar 06 13:14:52 2008
Dependent variable = bronch
Independent variable = ppm
Slope parameter is restricted as slope >= 1
Total number of observations = 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
User has chosen the log transformed model
intercept
intercept 1
B-30
-------
Variable
background
intercept
slope
Estimate
0
-0.186463
1
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-48.6418 3
-51.5597 1 5.83582 2
-136.058 1 174.833 2
P-value
Specified effect =
Risk Type =
Confidence level =
BMD =
BMDL =
0.8
0.6
0.2
0
0.1
Extra risk
0. 95
0.133887
0.083223
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
BMDL
BMD
0
13:1403/062008
10 15
dose
20
25
B-31
-------
B-2.2.4. Alveolar bronchiolization in female mice
All models hadp values <0.01 except the log-logistic and log-probit models.
Model
Log-logistic
Log-probit
p value
0.80
0.12
AIC
49.4
49.4
BMC10
0.18
0.40
BMCL10
0.022
0.278
The log-logistic model had a smaller chi-squared residual at the low exposure.
Logistic Model. (Version: 2.8; Date: 01/18/2007)
Input Data File: M:\_BMDS\NB_CIIT_F_MICE.(d)
Gnuplot Plotting File: M:\_BMDS\NB_CIIT_F_MICE.plt
Wed Oct 01 15:35:13 2008
BMDS MODEL RUN
The form of the probability function is:
P[response] = background+(1-background)/[1+EXP(-intercept-slope*Log(dose))\
Dependent variable = bronch
Independent variable = ppm
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0
intercept = 0.699779
slope = 1.0606
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
intercept
slope
intercept
1
-0.94
slope
-0.94
1
Variable
background
intercept
slope
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
0 * * *
0.170022 * * *
1.37009 * * *
* - Indicates that this value is not calculated.
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
Log(likelihood)
-22.3612
-22.6898
-133.568
# Param's
4
2
1
Deviance Test d.f.
0.657286
222.413
P-value
0.7199
<.0001
B-32
-------
AIC:
49.3797
Goodness of Fit
Dose
0.0000
5.0000
25.0000
50.0000
Est. Prob.
0.0000
0.9149
0.9899
0.9960
Expected
0.000
54 .895
63 .350
61.755
Observed
0
55
63
62
Size
53
60
64
62
Scaled
Residual
0.000
0.049
-0.437
0.496
Chi 2 = 0.44
d.f. = 2
P-value = 0.8028
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.177674
BMDL = 0.0221528
0.8
lo.6
o 0.4
'
0.2
Log-Logistic Model with 0.95 Confidence Level
Log-Logistic
BMPL BMP
0
10
20 30
dose
40
50
B-33
-------
Probit Model. (Version: 2.7; Date: 01/18/2007)
Input Data File: M:\_BMDS\NB_CIIT_F_MICE.(d)
Gnuplot Plotting File: M:\_BMDS\NB_CIIT_F_MICE.plt
Wed Oct 01 15:39:45 2008
BMDS MODEL RUN
The form of the probability function is:
P[response] = Background
+ (1-Background) * CumNorm(Intercept+Slope*Log(Dose)),
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = bronch
Independent variable = ppm
Slope parameter is restricted as slope >= 1
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0
intercept = -1.09733
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -background -slope
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
intercept
intercept 1
Parameter Estimates
Variable
background
intercept
slope
Estimate
0
-0.363696
1
Std. Err.
NA
0.209203
NA
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-0.773726
0.0463335
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Analysis of Deviance Table
Model
Full model
Fitted model
Reduced model
Log(likelihood)
-22.3612
-23.6802
-133.568
# Param's
4
1
1
Deviance Test d.f.
P-value
2.63796
222.413
0.4509
<.0001
AIC:
49.3603
Goodness of Fit
Dose
0.0000
5.0000
25.0000
50.0000
Est. Prob.
0.0000
0.8936
0.9978
0.9998
Expected
0.000
53 .614
63 .862
61.988
Observed
0
55
63
62
Size
53
60
64
62
Scaled
Residual
0.000
0.580
-2 .327
0.110
= 5.76
d.f. = 3
P-value = 0.1237
B-34
-------
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.399375
BMDL = 0.278169
0.8
0.6
o
'•g 0.4
CO
0.2
Probit
BMDLBMD
0
16:2810/01 2008
Probit Model with 0.95 Confidence Level
10
20 30
dose
40
50
B-35
-------
APPENDIX B-3: Dose-Response Modeling of Carcinogenicity Data for Nitrobenzene
B-3.1. METHODS
Data sets in rats and mice exposed to nitrobenzene by inhalation for up to 2 years (Cattley
et al., 1994; CUT, 1993) showing at least a statistical trend for increased tumor incidence with
increasing exposure were fit using the multistage model in EPA's BMDS (version 1.3.2)
(U.S. EPA, 1999). The data modeled are shown in Table B-3.1. The exposure levels used were
those reported in the study, not adjusted for duration of exposure or converted to HECs prior to
modeling. Model parameters were restricted to be positive, consistent with a monotonic dose-
response relationship. A BMR of 10% increase in extra risk was used for kidney adenomas and
carcinomas, and a 5% increase in extra risk was used for liver and thyroid adenomas and
carcinomas.
Table B-3.1. Tumorigenic responses in experimental animals exposed to
nitrobenzene via inhalation for up to 2 years
Strain/species (sex)
Site
F-344 rats (male)
Liver: hepatocellular adenoma or carcinoma
Thyroid: follicular cell adenoma or adenocarcinoma
Kidney: tubular adenoma and carcinoma
B6C3F1 mice (male)
Lung: A/B adenoma or carcinoma
Thyroid: follicular cell adenoma
Incidence of neoplasms
Concentration of nitrobenzene (ppm)
0
l/43a
l/43a
0/43a
1
4/50
1/50
0/50
5
5/47
5/47
0/47
25
16/46
8/46
6/46
Concentration of nitrobenzene (ppm)
0
8/42a
0/41a
5
16/44
4/44
25
20/45
1/45
50
21/48
6/46
"Statistically significant positive exposure-related trend in incidence by Cochran-Armitage trend test (p < 0.05).
Sources: CUT (1993); Cattley et al. (1994).
B-3.2. RESULTS
The BMD modeling results are summarized in Table B-3.2. This table shows the BMDs
and BMDLs derived from each endpoint modeled. The BMDS outputs for all model runs are
presented below.
B-36
-------
Table B-3.2. Summary of BMD modeling results for nitrobenzene cancer
data
Tumor
Number of
stages
p Value for
model fit
BMD10
(ppm)
BMDL10
(ppm)
Male F344 rats
Liver: hepatocellular adenoma or carcinoma
Thyroid: follicular cell adenoma or adenocarcinoma
Kidney: tubular cell adenoma or carcinoma
1
1
la
0.63
0.38
0.99
6.76
13.64
22.81
4.46
7.78
16.78
Male B6C3F1 mice
Lung: A/B adenoma or carcinoma
Thyroid: follicular cell adenoma
1
la
0.18
0.056b
15.43
52.23
8.52
30.46
"Model was one-stage with a three-degree polynomial (coefficients for lower degrees were 0).
bWhile this data set had a statistically significant trend, the dose-response shape was not strictly monotonic (see
Output B-3.2. V). Within the context of applying one model for the purpose of consistent low-dose linear
extrapolation, the multistage model provides an adequate fit.
OUTPUT B-3.2.1. Male F344 rat liver tumors (hepatocellular adenoma or carcinoma)
BMDS MODEL RUN
The form of the probability function is:
The parameter betas are restricted to be positive
Dependent variable = liver
Independent variable = ppm
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-OOE:
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background = 0.0425205
Beta(l) = 0.0148284
Beta(2) = 0
Beta(3) = 8.47487e-007
B-37
-------
and do not appear in the correlation matrix
Background Beta(l)
Background 1 -0.57
Beta(l) -0.57 1
Std. Err.
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-64.3357 4
-64.7955 2 0.919545 2
-75.2506 1 21.8296 3
AIC:
Est. Prob.
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0. 95
6.76461
4.46073
18.5642
B-38
-------
Multistage Model with 0.95 Confidence Level
0.5
0.4
T3
-------
Default Initial Parameter Values
Background = 0.0358259
Beta(l) = 0.0064946
Beta(2) = 0
Beta(3) = 0
Asymptotic Correlation Matrix of Parameter Estimates
Background
Beta(l)
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -46.8328 4
Fitted model -47.7145 2 1.76351 2
Reduced model -52.1437 1 10.6218 3
Dose Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 13.6391
BMDL = 7.78329
BMDU = 34.2184
Taken together, (7.78329, 34.2184) is a 90 % two-sided confidence
interval for the BMD
B-40
-------
Multistage Model with 0.95 Confidence Level
u.oo
0.3
0.25
T3
CD
t> 0.2
.2 °'15
CO
£ 0.1
0.05
0
r
r
?
7
?
: <
Mi
r
k^^'
h
jltistage
t
^-T
t
BMDL
T
^
[ :
P :
. :
^
BMP ;
10
15
20
25
dose
10:3602/222008
OUTPUT B-3.2.3. Male F344 rat kidney tumors (tubular cell adenomas or carcinomas)
Multistage Model. $Revision: 2.1 $ $Date: 2000/08/21 03:38:21 $
Input Data File: C:\BMDS\UNSAVED1.(d)
Gnuplot Plotting File: C:\BMDS\UNSAVEDl.plt
Mon Jul 10 13:25:30 2006
The form of the probability function is:
Dependent variable = Incidence
Independent variable = Concentration
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
Default Initial Parameter Values
B-41
-------
Asymptotic Correlation Matrix of Parameter Estimates
Estimate
Analysis of Deviance Table
Log(likelihood) Deviance Test DF
-17.8118
-17.8645 0.105529 3
-26.5061 17.3886 3
i: 1
0.0000
i: 2
1.0000
i: 3
5.0000
i: 4
25.0000
Chi-square =
0 43
0 50
0 47
6 46
P-value = 0.996E
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 22.8198
BMDL = 16.7833
B-42
-------
Multistage Model with 0.95 Confidence Level
£=
O
"o
CO
0.3
0.25
0.2
0.15
0.1
0.05
Multistage
BMD Lower Bound
BMD
|BMD
10
15
dose
20
25
30
13:2507/102006
OUTPUT B-3.2.4. Male B6C3F1 mouse lung tumors (A/B adenomas or carcinomas)
Multistage Cancer Model. (Version: 1.5; Date: 02/20/2007)
Input Data File: C:\BMDS\NB_CIIT_MFEMALE_RFC.(d)
Gnuplot Plotting File: C:\BMDS\NB_CIIT_MFEMALE_RFC.plt
Fri Feb 22 13:49:01 2008
The form of the probability function is:
Dependent variable = thyroid
Independent variable = ppm
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
Default Initial Parameter Values
B-43
-------
Background = 0.286791
Beta(l) = 0.00593155
Beta(2) = 0
Beta(3) = 0
Asymptotic Correlation Matrix of Parameter Estimates
Beta(1)
Background 1 -0.71
Beta(l) -0.71 1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(1)
Beta(2)
Beta(3)
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f.
Full model -113.1 4
Fitted model -114.829 2 3.45823 2
Reduced model -117.28 1 8.36076 3
Dose Est. Prob.
Taken together, (8.523 , 61.7258) is a 90 % two-sided confidence
interval for the BMD
Multistage Cancer Slope Factor = 0.011733
B-44
-------
Multistage Cancer Model with 0.95 Confidence Level
0.6
0.5
Fraction Affected
0 0
co ji.
0.2
0.1
; Multistage Cance
r Linear extrapolatior
~-
~-
~-
- <
\
<
^
\
r
'"" T T;
r ^
>
h^
L
L
BMDL
^-^
4 — l
1 1;
-
BMD :
10
20 30
dose
40
50
13:4902/222008
OUTPUT B-3.2.5. Male B6C3F1 mouse thyroid follicular cell adenomas
Multistage Cancer Model. (Version: 1.4; Date: 01/18/2007)
Input Data File: G:\_BMDS\NB_CIIT_M_MICE.(d)
Gnuplot Plotting File: G:\_BMDS\NB_CIIT_M_MICE.plt
Thu Feb 14 14:48:52 2008
The form of the probability function is:
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
B-45
-------
Asymptotic Correlation Matrix of Parameter Estimates
Variable Estimate Std. Err.
Background
Beta(1)
Beta(2)
Beta(3)
Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Model
Full model
Fitted model
Reduced model
Log(likelihood) # Param's Deviance Test d.f.
-36.0112 4
-39.264 2 6.50555 2
-41.1473 1 10.2722 3
Est. Prob.
76
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 52.2299
BMDL = 30.3586
BMDU = 334 . 893
Multistage Cancer Slope Factor =
B-46
-------
T3
€
0)
0.25
0.2
0.15
o
"C n 1
CO u- I
0.05
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
14:4802/142008
BMDL
BMD
10
20 30
dose
40
50
B-47
-------
B-3.3. Estimation of Overall Cancer Risk from Inhalation Exposure to Nitrobenzene by
Combining Risk Estimates Across Multiple Tumor Sites
In the CUT (1993) bioassay that was selected for use in the cancer dose-response
modeling of nitrobenzene, increased tumor incidences were observed at multiple sites in the male
rat following inhalation exposure to nitrobenzene (i.e., in the kidney, thyroid, and liver). Given
the multiplicity of tumor sites, basing the unit risk on one tumor site may underestimate the
carcinogenic potential of nitrobenzene. In addition, application of one model to a composite data
set does not accommodate biologically relevant information that may vary across sites or may
only be available for a subset of sites. Following the recommendations of NRC (1994) and the
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), a statistically appropriate upper
bound on total risk was estimated in order to gain some understanding of the total risk from
multiple tumor sites in the selected data set. Note that this estimate of overall risk describes the
risk of developing any combination of the tumor types considered, not just the risk of developing
all simultaneously.
NRC (1994) stated that an approach based on counts of animals with one or more tumors
(or tumor-bearing animals) would tend to underestimate overall risk when tumor types occur
independently and that an approach based on combining the risk estimates from each separate
tumor type should be used. This assessment considers both types of approach for estimating
total risk.
Separate dose-response fits for each site. Following the NRC (1994) recommendation,
there are several ways to address combining the risk estimates from each separate tumor type.
For individual tumor sites modeled using the multistage model,
P(d q) = 1 - exp[-(q0 + qid + q2d2 + ...
the model for the combined tumor risk would still be multistage, with a functional form that has
the sum of stage-specific multistage coefficients as the corresponding multistage coefficient
Pc(d q) = 1 - exp[-(Iq0l + Iqi,d + Zq2,d2 + ... + Iquff)]
The resulting equation for fixed extra risk (BMR) is polynomial in dose (when logarithms of
both sides are taken) and can be straightforwardly solved for combined BMD. But confidence
bounds (based on the profile likelihood method) for that BMD are not provided in available
software.
A Bayesian approach to finding a confidence bound on the combined BMD was
implemented by using WinBugs (Spiegelhalter et al., 2003). WinBugs software is freely
available and it implements Markov chain Monte Carlo computations. Use of WinBugs to
derive a distribution of BMD for a single multistage model (Kopylev et al., 2007) was
B-48
-------
straightforwardly generalized to derive a distribution of BMD for combined tumor risk,
according to the NRC (1994) methodology described above. The advantage of the Bayesian
approach is that it produces a distribution of the BMD, allowing better characterization of
uncertainty than would calculation of only the expected risk.
A diffuse (high variance or low tolerance) Gaussian prior restricted to be nonnegative
was used. The posterior distribution is based on three chains with 50,000 burn-ins (50,000 initial
simulations dropped) and thinning rate of 10, resulting in 150,000 simulations total. The mean
and 5th percentile of the posterior distribution provide estimates of the mean BMD ("central
estimate") and the BMDL for combined tumor risk, respectively. The WinBugs code for
generating these distributions is provided at the end of this section.
Table B-3.3 presents the results of the Bayesian analysis. The upper bound on the
combined risk estimate for nitrobenzene is 4 x 10~5 (jig/m3)"1, rounding the value derived above
to one significant figure. This analysis, using combined tumor sites, suggests a higher potency
than when considering only the most sensitive tumor site alone (i.e., liver tumors). In this case,
however, there is no appreciable difference between the two approaches when the potencies are
rounded to one significant figure. Regardless, neither of these risk estimates should be used with
continuous lifetime nitrobenzene exposures greater than about 3 x 103 |ig/m3, the human
equivalent of the POD defined for the male rat liver tumors, because the observed dose-response
curve is not likely to be linear at higher doses.
Table B-3.3. Overall risk based on kidney, thyroid, or liver
adenomas/carcinomas in F344 rats, using simulated data derived from
observed incidences (Bayesian approach)
Tumor site and type
Kidney, adenoma or
carcinoma
Thyroid, follicular cell
adenoma or carcinoma
Liver, hepatocellular
adenoma or carcinoma
Combined tumors
Bioassay concentrations
Median BMC
(BMC10), ppm
19.4C
13.6
6.7
4.1
BMCL10
(ppm)
14.2C
7.8
4.4
2.9
Human equivalents
BMC 10/HEC
(Hg/m3)a
17,500
12,200
6,030
3,650
BMCL10/HEc
(Hg/m3)
12,800
7,020
4,000
2,610
Slope from
BMC10
frig/m3)-1"
5.7 x KT6
8.2 x KT6
1.7 x 10~5
2.7 x 1(T5
Upper bound
risk estimate,
(Hg/m3)-lb
7.8 x 10~6
1.4 x 10~5
2.5 x 10~5
3.8 x 1(T5
aBMC(L)HEc = BMC(L) x 5.04 mg/m3 x 5/7 x 6/24, assuming ratio of animal to human airblood partition
coefficients is 1.
bThe slope to background from the BMCio/HEc = 0.1/BMCio/HEc, and the upper bound risk estimate =
0.1/BMCL10/HEc, using the models developed and reported in section B-3.3.
°Note that the BMCio and BMCL10 for kidney tumors are not the same values estimated by using BMDS (which are
22.8 and 16.8 ppm, respectively; see Output, B-3.2.III). Alternate methodology does not generally provide
identical results. In this case, the Bayesian results are expected to be more accurate. However, this tumor site is
the least sensitive and with very little impact on the composite estimated risk.
B-49
-------
Tumor-bearing animals. For comparison, dose-response analysis was carried out on the
incidence of tumor-bearing animals. Tabulation of tumor-bearing animals is best carried out
with histopathology results reported for individual animals. These results were not available in
the CUT (1993) report. Under the assumption that the three tumor types are independent of each
other, the incidences of tumor-bearing animals in each group can be estimated using the
following probability relationship:
P(A, B, or C) = P(A) + P(B) + P(C) - P(A)P(B) - P(A)P(C) - P(B)P(C) + P(A)P(B)P(C)
where A, B, and C represent the three tumor types, and P( ) denotes the proportion responding.
This relationship sums the proportions responding across tumor types and adjusts for double
counting. A summary of the application of this relationship to the male rat tumor incidence (in
Table B-3.1) is provided in Table B-3.4. The estimated incidences in the control and low-dose
groups are the same as just adding the raw incidences, while those in the mid- and high-dose
groups are approximately 1 and 16% less than the sums of the raw incidences for those groups.
A one-stage multistage model provided an adequate fit to the estimated incidences of
tumor-bearing animals (see Output, B-3.4.1), with a BMDio of 3.8 ppm and a BMDLio of
2.7 ppm, very similar to the BMDio and BMDLio for overall risk derived above using the
Bayesian approach. The BMCHEc was calculated as BMC x 5.04 mg/m3 x 5/7 x 6/24 =
3,420 |ig/m3, assuming the ratio of animal to human airblood partition coefficients is 1. The
slope to background from the BMCio/HEc = 0.1/BMCio/HEc or 3 x 10~5 (jig/m3)"1. The BMCLHEc
was calculated as BMCL x 5.04 mg/m3 x5/7 x 6/24 = 2,430 |ig/m3. The upper bound risk
estimate = 0. l/BMCLi0/HEc = 4 x 10~5 (jig/m3)"1.
Table B-3.4. Estimated incidence of male F344 rats with liver, thyroid, or
kidney tumors from separately tabulated site-specific incidences following
nitrobenzene exposure for 2 years
Exposure
(ppm)
0
1
5
25
Incidence of neoplasms
and (%) responding
Liver
1/43 (2.3%)
4/50 (8.0%)
5/47 (10.6%)
16/46 (35%)
Thyroid
1/43 (2.3%)
1/50 (2.0%)
5/47 (10.6%)
8/46 (17%)
Kidney
0/43 (0%)
0/50 (0%)
0/47 (0%)
6/46 (17%)
Estimated
probability of an animal
having at least one of the
tumor types
0.046
0.098
0.201
0.532
Estimated
incidence of
tumor-bearing
animals
2/43
5/50
9/47
24/46
Although both approaches in this case lead to a similar result, it is not clear that the two
approaches will generally agree. Further examination of the operating characteristics of the
respective approaches is necessary for any such generalizations.
B-50
-------
References
Kopylev, L., Chen, C., White, P. (2007). Towards quantitative uncertainty assessment for cancer risks: central
estimates and probability distributions of risk in dose-response modeling. Regul Tox Pharmacol 49:203-207.
NRC (1994). Science and Judgment. National Academy Press. Washington, DC.
Spiegelhalter, D., Thomas, A., Best, N., (2003). WinBugs Version 1.4 User Manual, http://www.mrc-
bsu.cam.ac.uk/bugs/winbugs/manuall4.pdf
B-51
-------
Computer Code for Bayesian Approach
WinBugs computer code that was used to derive posterior distribution of multistage
parameters for individual tumors is provided below. The resulting distributions can be used to
solve for distribution of individual and combined BMD as described above.
To run the following programs, one needs to have R and WinBugs (freeware) installed,
and the R package RtoWinbugs should be installed and loaded in R. The next step is to create
two text files in the directory that R uses:
sum.txt is a text file containing the following text:
model {
for( i in 1 : doses ) {
<- -bO-bl*x[i]
y2[i]~dbin(g2[i],n2[i])
Iog(g2[i])<--c0-cl*x[i]-c2*pow(x[i],2)-c3*pow(x[i],3)
y3[i]~dbin(g3[i],n3[i])
Iog(g3[i])<- -dO-dl*x[i]
}
bO~dnorm(0,.0001)I(0,)
bl~dnorm(0,.0001)I(0,)
cO~dnorm(0,.0001)I(0,)
cl~dnorm(0,.0001)I(0,)
c2~dnorm(0,.0001)I(0,)
c3~dnorm(0,.0001)I(0,)
dO~dnorm(0,.0001)I(0,)
dl~dnorm(0,.0001)I(0,)
data.txt is a text file containing the following text:
list(doses=4,
nl=c(43, 50, 47, 46),
n2=c(43, 50, 47, 46),
n3=c(43, 50, 47, 46),
x=c(0, 1,5,25),
yl=c(42, 46, 42, 30),
y2=c(43, 50, 47, 40),
y3=c(42, 49, 42, 38)
The next step is to run three R commands:
inits<-function(){list(bO=ranif(l,0,.l),bl=ranif(l,0,.l),cO=ranif(l,0,.l),cl=ranif(l,0,.l),c2=ranif(l,0,.l),c3=ranif(l,0,.l),
dO=ranif(l,0,.l),dl=ranif(l,0,.l))}
parameters<-c("b 1 ","c 1 ","c2","c3,"dl ")
NB.sim<-bugs("data.txt",inits,parameters, "sum.txt" ,n.chains=3,n.iter=550000,n.burnin=50000,n.thin=10)
The last step is to run the R command
attach(NB.sim)
that creates posterior distributions of the individual multistage parameters (here bl for liver, dl for thyroid and cl,c2,c3 for thyroid).
B-52
-------
Output B-3.4.1.
Multistage Model. (Version: 2.7; Date: 01/18/2007)
Input Data File: M:\_BMDS\NB_TBA_RM.(d)
Gnuplot Plotting File: M:\_BMDS\NB_TBA_RM.plt
Wed Mar 26 13:50:47 2008
The form of the probability function is:
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
Default Initial Parameter Values
Background = 0.0649281
Beta(l) = 0.0269233
Beta(2) = 0
Beta(3) = 0
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
1 -0.54
-0.54 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(1)
Beta(2)
Beta(3)
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -79.1379 4
Fitted model -79.2961 2 0.316284 2 0.8537
Reduced model -96.8271 1 35.3784 3 <.0001
B-53
-------
Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 3.78878
BMDL = 2.70515
BMDU = 10.7878
0.7
0.6
0.5
T3
-§0.4
£
cO.3
o
=8
20.2
LJ_
0.1
Multistage Model with 0.95 Confidence Level
Multistage
BMPL
iBMP
10 15
dose
20
25
B-54
------- |