oEPA
EPA/63 5/R-02/00 IF
TOXICOLOGICAL REVIEW
OF
BENZENE
(NONCANCER EFFECTS)
(CAS No. 71-43-2)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
October 2002
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. Note: This document may undergo revisions in
the future. The most up-to-date version will be made available electronically via the IRIS Home
Page at http://www.epa.gov/iris.
11
-------
CONTENTS—TOXICOLOGICAL REVIEW FOR BENZENE (CAS No. 71-43-2)
FOREWORD vii
AUTHORS, CONTRIBUTORS, AND REVIEWERS viii
LIST OF ACRONYMS AND ABBREVIATIONS ix
EXECUTIVE SUMMARY xii
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS 2
3. TOXICOKINETICS RELEVANT TO ASSESSMENTS 3
3.1. ABSORPTION 3
3.1.1. Gastrointestinal Absorption 3
3.1.2. Respiratory Absorption 4
3.1.3. Dermal Absorption 5
3.2. DISTRIBUTION 8
3.2.1. Oral Exposure 8
3.2.2. Inhalation Exposure 8
3.2.3. Dermal Exposure 9
3.3. METABOLISM 9
3.3.1. Metabolic Pathways 10
3.3.2. Requirement for CYP2E1 10
3.3.3. Toxicity of Benzene Metabolites 12
3.3.3.1 Phenolic Products 12
3.3.3.2. Trans,trans-Muconaldehyde 13
3.3.3.3. Benzene Oxide 14
3.3.4. Species, Route, and Rate Differences 14
3.3.5. Induction of CYP2E1 15
3.3.6. Mechanism of Toxicity 16
3.3.6.1. Formation of Covalent Adducts 16
3.3.6.2 Genotoxicity 18
3.3.6.3 Oxidative Stress 18
3.3.6.4. Inhibition of Cytokine Formation 19
3.4. ELIMINATION AND EXCRETION 19
3.4.1. Oral Exposure 19
3.4.2. Inhalation Exposure 20
3.4.3. Dermal Exposure 21
3.4.4. Other Routes of Exposure 21
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS 22
3.6. TOXICOKINETICS SUMMARY 26
4. HAZARD IDENTIFICATION 28
iii
-------
CONTENTS-TOXICOLOGICAL REVIEW FOR BENZENE (CAS No. 71-43-2)
(continued)
4.1. STUDIES IN HUMANS 28
4.1.1. Oral Exposure 28
4.1.2. Inhalation Exposure 28
4.1.2.1. Hematotoxicity 28
4.1.2.2. Neurotoxicity 43
4.1.2.3. Reproductive Toxicity 44
4.1.2.4. Developmental Toxicity 47
4.2. ACUTE AND CHRONIC STUDIES IN EXPERIMENTAL ANIMALS 50
4.2.1. Hematotoxicity 51
4.2.1.1. Oral Exposure—Subchronic Studies 51
4.2.1.2. Oral Exposure—Chronic Studies 52
4.2.1.3. Inhalation Exposure—Subchronic Studies 53
4.2.1.4. Inhalation Exposure—Chronic Studies 60
4.2.1.5. Effects on Stem Cell Populations 63
4.2.1.6. Summary of Principal Hematotoxic Effects 77
4.2.2. Reproductive/Developmental Effects 79
4.2.2.1. Reproductive Toxicity 79
4.2.2.2. Developmental Toxicity 83
4.2.2.3. Summary of Principal Reproductive/Developmental
Toxicity Effects 95
4.2.2.4. Mechanisms of Developmental and Reproductive Toxicity ... 96
4.2.3. Neurotoxicity 97
4.2.3.1. Oral Exposure 97
4.2.3.2. Inhalation Exposure 98
4.2.3.3. Summary of Neurotoxic Effects 101
4.2.4. Immunotoxicity 101
4.2.4.1. Oral Exposure 101
4.2.4.2. Inhalation Exposure 103
4.2.4.3. Summary of Immunotoxic Effects 104
4.3. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS
AND MODE OF ACTION 104
4.4. SUSCEPTIBLE POPULATIONS 109
4.4.1. Childhood Susceptibility 109
4.4.2. Gender Differences 110
4.4.3. Genetically Susceptible Populations Ill
4.5. HAZARD IDENTIFICATION SUMMARY 113
5. DOSE-RESPONSE ASSESSMENTS 116
5.1. INHALATION REFERENCE CONCENTRATION (RfC) 117
5.1.1. Choice of Principal Study and Critical Effect 117
5.1.2. Benchmark Dose Modeling 119
5.1.3. RfC Derivation 121
5.1.4. Comparison Analysis Based on the LOAEL 123
5.1.5. Comparison Analysis Based on the Ward et al. (1985) Experimental
iv
-------
CONTENTS-TOXICOLOGICAL REVIEW FOR BENZENE (CAS No. 71-43-2)
(continued)
Animal Study 123
5.2. ORAL REFERENCE DOSE (RfD) 127
5.2.1. Choice of Principal Study and Critical Effect 127
5.2.2. Conversion of Inhalation Exposure to Equivalent Oral Dose Rate .... 127
5.2.3. RfD Derivation 129
5.2.4. Comparison Analysis Based on the LOAEL 130
5.2.5. Comparison Analysis Based on the NTP (1986) Experimental
Animal Study 130
5.3 DOSE-RESPONSE SUMMARY 135
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE-RESPONSE 141
6.1. HUMAN HAZARD POTENTIAL 141
6.2. DOSE-RESPONSE 144
6.2.1. Inhalation RfC 144
6.2.2. Oral RfD 144
7. REFERENCES 146
-------
LIST OF FIGURES
Figure 1. Metabolic pathways for benzene 11
Figure 2. Linear model of ALC data 121
Figure 3. Linear model of HCT data 125
LIST OF TABLES
Table 1. Hematotoxicity of benzene—occupational exposure 30
Table 2. Median benzene urine metabolites, by exposure category, in a study of
workers exposed to benzene in Shanghai, China, 1992 40
Table 3. Comparison of mean peripheral blood counts with standard deviations, by
exposure status, in a study of workers exposed to benzene in Shanghai, China, 1992 41
Table 4. Reproductive toxicity of inhaled benzene in humans 45
Table 5. Developmental toxicity of benzene—humans 48
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—
inhalation exposure 54
Table 7. Reproductive toxicity of inhaled benzene in test animals 80
Table 8. Developmental toxicity of inhaled benzene in test animals 84
Table 9. Joint effects of CYP2E1 activity and NQO1 genotype on benzene-induced
hematotoxicity in Chinese Workers 112
Table 10. Results of BMC modeling of Rothman et al. (1996a) data on benzene and ALC . . 122
Table 11. BMD modeling results of the NTP (1986) male mouse and male rat lymphocyte
counts, with untransformed data 133
Table 12. BMD modeling results of the NTP (1986) male mouse and male rat lymphocyte
counts, with transformed dose data 134
Table 13. Summary of uncertainty factors used for deriving the RfC and RfD 137
Table 14. Summary of RfC and RfD estimates using human and experimental animal data,
as well as benchmark dose modeling and LOAEL/NOAEL approaches 142
VI
-------
FOREWORD
The National Center for Environmental Assessment-Washington Division, Office of
Research and Development, has prepared this document on the Toxicological Review of Benzene
to serve as a source document for updating the noncancer health effects summary on benzene in
the Integrated Risk Information System (IRIS).
In the development of this document, the published relevant scientific literature has been
reviewed, key studies have been evaluated, and summary/conclusions have been prepared so that
the noncancer health effects from exposure to benzene are qualitatively and quantitatively
characterized and the derivation of the reference dose (RfD) and reference concentration (RfC) are
adequately described. The evaluation and review of the noncancer effects of exposure to benzene
have been conducted under the U.S. Environmental Protection Agency's standing guidance of
several relevant risk assessment guidelines dealing with reproductive, developmental, and
neurotoxic effects. This draft has undergone internal peer review and an expert external peer-
panel review in October 1998 as well as a 90-day public comment period.
The emphasis of this document is a discussion of the noncancer adverse health effects of
benzene, including the no-observed-adverse-effect levels, the lowest-observed-adverse-effect
levels, benchmark dose analysis, uncertainty factors, and any other considerations used to develop
the RfDs and RfCs for benzene.
vn
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
The National Center for Environmental Assessment-Washington Division (NCEA-W) of
the U.S. Environmental Protection Agency's (EPA's) Office of Research and Development (ORD)
was responsible for the preparation of this document. TN & Associates, Inc., under EPA Contract
No. 68-C6-0024, Work Assignment No. 1-09.1, conducted the literature search and prepared
preliminary drafts of the document.
AUTHORS
David Bayliss, NCEA-W
Jennifer Jinot, NCEA-W
Babasaheb Sonawane, NCEA-W
INTERNAL REVIEWERS
NCEA/ORD
Robert Bruce, NCEA-Cin
Margaret Chu, NCEA-W
Eric Clegg, NCEA-W
Anne Jarabek, NCEA-RTP
Carole Kimmel, NCEA-W
Robert McGaughy, NCEA-W
William Pepelko, formerly of NCEA-W
James Rowe, Office of Science Policy
Woody Setzer, National Health and Environmental Effects Research Laboratory-RTF
Other EPA Offices
Pam Brodowicz, Office of Mobile Sources, Office of Air and Radiation
Other Federal Offices
Dr. Nathaniel Rothman, National Cancer Institute, National Institutes of Health, Bethesda,
Maryland
Peer Review Panel
Dr. Donald Gardner, Chairperson, Inhalation Toxicology Associates, Raleigh, NC
Dr. Lynne Haber, Toxicology Excellence for Risk Assessment, Cincinnati, OH
Dr. John Keller, Consultant, Olney, MD
Dr. Michele Medinsky, Consultant, Durham, NC
Dr. Robert Snyder, Department of Pharmacology and Toxicology, Rutgers University,
Piscataway, NJ
Vlll
-------
LIST OF ACRONYMS AND ABBREVIATIONS
AChE
ALC
AML
ANOVA
ASA
AUC
BFU-E
BMC
BMCL
BMD
BMDL
BMR
CFU-C
CFU-C/tibia
CFU-E
CFU-HPP
CFU-S
CHO-M-OH
CI
Con A
COOH-M-CHO
COOH-M-OH
CYP2E1
EMG
EPA
F344
GD
acetyl cholinesterase
absolute lymphocyte count
acute myelogenous leukemia
analysis of variance
acetylsalicylic acid
area under the curve
burst-forming units-erythroid
benchmark concentration
95% lower confidence limit on the BMC
benchmark dose
95% lower confidence limit on the BMD
benchmark response
colony-forming units-culture
colony-forming granulopoietic stem cells
colony-forming units-erythroid
high proliferative potential colony-forming units
colony-forming units-spleen
6-hydroxy-^rami,^ram'-2,4-hexadienal
confidence interval
concanavalin A
6-oxo-trans, fra«5-2,4-hexadienoic acid
6-hydroxy-^ram1, ^rara'-2,4-hexadienoic acid
cytochrome P450 2E1
electromyographical
U.S. Environmental Protection Agency
Fischer 344
gestation day
IX
-------
LIST OF ACRONYMS AND ABBREVIATIONS (continued)
GM-CFU-C
GM-CFU-G
GPA
GSH
GTP
HCT
Hgb
HSDB
IC50
i.p.
IL-1
IL-laD
IL-lpD
IRIS
Kd
LOAEL
LPS
MA
MCH
MCV
MDS
MPO
MUC
NADPH
NCEA
NCEA-W
NK
granulocyte/macrophage colony-forming unit-culture
granulocyte/macrophage colony-forming unit
glycophorin A
glutathione
guanosine triphosphate
hematocrit
hemoglobin
Hazardous Substance Data Bank
Inhibition concentration 50%
intraperitoneal
interleukin-1
interleukin-1 alpha
interleukin-lbeta
Integrated Risk Information System
kilodalton
lowest-observed-adverse-effect level
lipopolysaccharide
trans,tmns-muconic acid
mean corpuscular hemoglobin
mean corpuscular volume
myelodysplastic syndrome
myeloperoxidase
muconaldehyde
nicotinamide adenine dinucleotide phosphate
National Center for Environmental Assessment
National Center for Environmental Assessment-Washington
Division
natural killer
-------
LIST OF ACRONYMS AND ABBREVIATIONS (continued)
NOAEL
NQO1 NAD(P)H
NTP
OR
PBPK
PCE
PC-PFC
PCV
PHA
RBC
RfC
RfD
SD
SE
SRBC
TWA
UF
U.S. EPA
WBC
YAC
no-observed-adverse-effect level
NAD(P)H quinone oxidoreductase
National Toxicology Program
odds ratio
physiologically based pharmacokinetic
polychromatic erythrocytes
polyclonal plaque-forming cells
packed cell volume
phytohemaglutinin
red blood cell
reference concentration
reference dose
standard deviation
standard error
sheep red blood cell
time-weighted average
uncertainty factor
U.S. Environmental Protection Agency
white blood cell
yeast artificial chromosome
XI
-------
EXECUTIVE SUMMARY
The U.S. Environmental Protection Agency's (EPA's, the Agency's) Office of Mobile
Sources, Office of Air and Radiation, requested that the National Center for Environmental
Assessment (NCEA) provide an updated characterization of human health risk from exposure to
benzene. Recently, EPA published Carcinogenic Effects of Benzene: An Update (EPA/600/P-
97/001F, April 1998 [U.S. EPA, 1998a]) and Extrapolation of the Benzene Inhalation Unit Risk
Estimate to the Oral Route of Exposure (NCEA-W-0517, April 1999). These documents serve
as a source of information regarding the carcinogenic effects of exposure to benzene.
The scope of the present report is limited to an assessment of the noncancer effects of
benzene under the Agency's standing guidance of several relevant risk assessment guidelines
dealing with reproductive, developmental, neurotoxic, and other noncancer effects, including
derivation of an oral reference dose (RfD) and inhalation reference concentration (RfC). This
toxicological review of benzene is to serve as a scientific source document for hazard
identification and dose-response assessment in updating the noncancer health effects summary
on benzene in the Integrated Risk Information System (IRIS).
Benzene, also known as benzol, is widely used as an industrial solvent, as an
intermediate in chemical syntheses, and as a component of gasoline; hence, the potential for
human exposure is great. Inhalation exposure is the major route of exposure to benzene,
although oral and dermal routes are also important. The toxicokinetics (absorption, distribution,
metabolism, and elimination) of benzene have been studied in humans and experimental animal
species. Benzene is readily absorbed by both test animals and humans and is distributed among
several body compartments. The parent compound is preferentially stored in fat, and the relative
uptake appears to be dependent on the blood perfusion rates of tissues. Metabolism of benzene
is required for expression of benzene toxicity.
Evidence indicates that following inhalation exposure to benzene, the major route of
elimination of unmetabolized benzene in humans is via exhalation. Absorbed benzene is
metabolized to phenol and muconic acid, followed by urinary excretion of conjugated sulfates
and glucuronides. Limited data exist on excretion of benzene in humans following dermal
exposure. Physiologically based pharmacokinetic models have been developed and are being
improved upon to better define interactions of benzene metabolism, toxicity, and dosimetry.
These interactions for humans exposed to low concentrations can be assessed only when the
mode of action is understood at a quantitative level and is incorporated within a physiological
modeling framework. However, the current models are insufficiently developed and validated to
allow them to predict with certainty the relationship between metabolism and toxicity of
benzene.
Benzene exposure results in adverse noncancer health effects by all routes of administration
to test animal species. Hematotoxicity has been consistently reported to be the most sensitive
indicator of noncancer toxicity both in limited studies in humans and experimental animals, with
bone marrow as the principal target organ. Chronic exposure to benzene results in progressive
deterioration of hematopoietic function. Whether the hematotoxic and carcinogenic effects of
benzene are due to a common mechanism has not been established. Although leukocytopenia has
been consistently shown to be a more sensitive indicator than anemia of benzene toxicity in
experimental animals, lymphocytopenia has been shown to be even more sensitive than
xii
-------
leukocytopenia. A decrease in absolute lymphocyte count (ALC) was observed to be the most
sensitive indicator of benzene exposure in several epidemiologic studies.
Benzene has also been shown to produce neurotoxic effects in test animals and humans after
short-term exposure, but at relatively high concentrations; however, long-term neurotoxicity
exposure studies are lacking. There is some evidence of adverse reproductive and developmental
effects due to benzene exposure from human epidemiologic studies, but data are not conclusive
enough to link low exposure levels to effects. No data from human studies were found to indicate
that children are more sensitive than adults to benzene toxicity, nor were any significant gender
differences found. The most frequently observed gender difference in test animals is a greater
sensitivity of male mice to benzene exposure; in rats, females appear to be more sensitive than
males. However, in the National Toxicology Program (NTP) study (NTP, 1986) the male was the
most sensitive sex in the rat.
Although a large number of epidemiologic and experimental animal studies are available for
evaluating noncancer health effects, there are few reports of human data with sufficiently reliable
estimates of exposure to benzene and few long-term, repeated-dose experiments in test animals.
Many of the epidemiologic studies are also complicated by exposure to other solvents, and some of
the longer-term animal studies have employed exposure levels that were too high to establish
reliable no-observed-adverse-effect level (NOAEL) values.
The epidemiologic occupational inhalation study by Rothman et al. (1996a) was chosen as
the principal critical study for deriving both an RfC and an RfD. This study showed significant
reductions in ALC, red blood cells (RBCs), and platelets in a subgroup of Chinese factory workers
exposed to a median 8-hour time-weighted average (TWA) concentration of 13.6 ppm (43.4 mg/m3).
In a subgroup of 11 workers exposed to a median 8-hr TWA concentration of 7.6 ppm (24 mg/m3),
only ALC was still significantly reduced. Thus this concentration is a lowest-observed-adverse-
effect level ( LOAEL) for benzene immunotoxicity in humans. The findings from this study also
indicate that white blood cells were significantly decreased and the mean corpuscular volume was
significantly increased in the total exposed group of 44 workers occupationally exposed to a median
8-hour TWA of 31 ppm (99 mg/m3) in comparison to an age- and sex-matched control group. These
dose-response effects are consistent with the well-known hematotoxic effects of benzene in humans
and experimental animals, as discussed in Sections 4.1.2.1 and 4.2.1 and summarized in Tables 1
and 6.
Reduction in ALC was the most sensitive endpoint in the Rothman et al. (1996a) study, and
the ALC exposure-response data were used in benchmark dose (BMD) modeling to obtain a point of
departure for the derivation of the RfC. It was necessary to transform the data to obtain an adequate
fit with the models for continuous data; then, the continuous linear model provided the best fit. In
the absence of a clear definition for an adverse effect for this continuous endpoint, a default
benchmark response of one standard deviation change from the control mean response was selected,
as suggested in EPA's draft Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b).
This default definition of a benchmark response for continuous endpoints corresponds to an excess
risk of approximately 10% for the proportion of individuals below the 2nd percentile (or above the
98th percentile) of the control distribution for normally distributed effects. The BMD modeling thus
resulted in a benchmark concentration (BMC) of 13.7 ppm (8-hour TWA) and a BMCL (95% lower
confidence unit on the BMC) of 7.2 ppm (8-hour TWA).
Xlll
-------
As suggested in EPA's draft Benchmark Dose Technical Guidance Document, the BMCL
was chosen as the point of departure for the RfC derivation. After converting to mg/m3 and
adjusting for continuous exposure, a BMCLADJ of 8.2 mg/m3 was obtained. Dividing this value by
an overall uncertainty factor (UF) of 300 yields a chronic inhalation RfC of 3 x 10"2 mg/m3. Because
the BMC is considered to be an adverse-effect level, an effect-level extrapolation factor analogous to
the LOAEL-to-NOAEL UF was used. EPA is planning to develop guidance for applying an effect-
level extrapolation factor to a BMD. In the interim, a factor of 3 was used in this analysis.
Additional factors of 10 for intraspecies variability, 3 for subchronic-to-chronic extrapolation, and 3
for database deficiencies—due to the absence of a two-generation reproductive/developmental
toxicity study for benzene—comprise the remainder of the 300 composite UF.
For comparison, an RfC was also calculated from the LOAEL of 7.6 ppm (8-hour TWA)
from Rothman et al. (1996a). After converting to mg/m3 and adjusting for continuous exposure, a
LOAEL AJJJ of 8.7 mg/m3 was obtained. Dividing this value by an overall UF of 1000 yields an RfC
of 9 x 10"3 mg/m3. The UF of 1000 was based on factors of 10 to account for the use of a LOAEL
because of the lack of an appropriate NOAEL, 10 for intraspecies variability, 3 for subchronic-to-
chronic extrapolation, and 3 for database deficiencies. This result of 9 x 10"3 mg/m3 based on the
LOAEL is in good agreement with the result of 3 x 10"2 mg/m3 based on the BMCL.
Support for this chronic inhalation RfC has been provided by the experimental animal study
of Ward et al. (1985). The subchronic inhalation study of Ward et al. (1985) was selected as a
supporting study because it was the experimental animal study with the longest inhalation exposure
duration and it provided good dose-response data. The exposure-response relationships for the
different hematologic endpoints in male mice (the most sensitive sex/species in this study) were
modeled using a BMD modeling approach, and decreased hematocrit (i.e., volume percentage of
erythrocytes in whole blood) was chosen as the critical effect. The continuous linear model was
selected because it provided the best fit. As above, a default level of one standard deviation change
from the control mean response was used as the benchmark response. A BMC of 100.7 ppm and a
BMCL of 85.0 ppm were obtained.
The BMCL was chosen as the point of departure for the RfC derivation. After converting to
mg/m3 and adjusting for continuous exposure, a BMCLADJ of 48.5 mg/m3 was obtained. As
discussed above, a UF of 3 is used as an effect-level extrapolation factor, analogous to a LOAEL-to-
NOAEL UF, because the BMC is considered an adverse-effect level. In addition, the standard UFs
of 3 for interspecies extrapolation for inhalation studies and 10 for intraspecies variability are
applied. A UF of 3 for database deficiencies is used, as above. Finally, a partial UF of 3 was used
to extrapolate from subchronic to chronic exposure. Dividing the BMCL^j by the overall
uncertainty factor of 1000 yields a chronic inhalation RfC of 5 x 10"2 mg/m3. This value is in good
agreement with the RfC of 3 x 10"2 mg/m3 based on BMD modeling of the ALC data from the
Rothman et al. (1996a) human study.
Similarly, for comparison purposes, a chronic inhalation RfC can be derived from the
NOAEL of 30 ppm observed for hematologic effects in the Ward et al. (1985) study. First, the
NOAEL is converted to mg/m3 and adjusted to equivalent continuous exposure, yielding 17.1
mg/m3. UFs are identical to those employed above, except that no NOAEL-to-LOAEL UF is used;
thus, the overall UF is 300. Dividing 17.1 by 300 results in an RfC of 6 x 10"2 mg/m3. This value is
also in good agreement with the RfC derived from the Rothman et al. (1996a) study.
xiv
-------
To derive the RfD, the same BMCL of 7.2 ppm (8-hour TWA) for the default benchmark
response of one standard deviation change from the control mean response for the critical effect of
reduced ALC in Rothman et al. (1996a) was used as the point of departure. After converting the
units, correcting for continuous exposure, and adjusting for the route-to-route extrapolation from
inhalation to oral exposure, a BMCLADJ-equivalent oral dose rate of 1.2 mg/kg/day was obtained.
Dividing the BMCL^j-oral of 1.2 mg/kg/day by an overall UF of 300 yields a chronic oral RfD of 4
x 10"3 mg/kg/day. As above, a UF of 3 is used as an effect-level extrapolation factor, analogous to
the LOAEL-to-NOAEL UF, because the BMC is considered to be an adverse-effect level.
Additional factors of 10 for intraspecies variability, 3 for subchronic-to-chronic extrapolation, and 3
for database deficiencies comprise the remainder of the 300 composite UF.
For comparison, an RfD was also calculated from the LOAEL of 7.6 ppm (8-hour TWA)
from Rothman et al. (1996a). After unit conversion, correction for continuous exposure, route-to-
route extrapolation, and division by a combined UF of 1000, an RfD of 1 x 10"3 mg/kg/day was
derived from the LOAEL. The combined UF of 1000 was based on factors of 10 to account for
using a LOAEL because of the lack of an appropriate NOAEL, 10 for intraspecies variability, 3 for
subchronic-to-chronic extrapolation, and 3 for database deficiencies. This RfD value of 1 x 10"3
mg/kg/day is in good agreement with the value of 4 x 10"3 mg/kg/day calculated from the BMCL.
For further comparisons, RfDs were calculated on the basis of the NTP's chronic
experimental animal study (NTP, 1986). In this study, Fischer 344 (F344) rats and B6C3F1 mice of
both sexes were administered benzene by gavage, and blood was drawn from subgroups of animals
at various time points. This study identified a LOAEL of 25 mg/kg for leukopenia and
lymphocytopenia in female F344 rats and male and female B6C3F1 mice; no NOAEL was
identified. Reduction in lymphocyte count was selected as the critical effect, and the dose-response
relationships at different time points were modeled using a BMD modeling approach. As above, a
default benchmark response of one standard deviation change from the control mean was used to
define the BMDs. The modeling results suggested that the male rat was the most sensitive
sex/species in this NTP gavage bioassay, and a BMDL (95% lower confidence limit on the BMD) of
1 mg/kg was selected as the point of departure for deriving the RfD.
The BMDL was first adjusted for exposure 7 days/week and then divided by a composite UF
of 1000. This composite UF reflects a UF of 3 used as an effect-level extrapolation factor,
analogous to LOAEL-to-NOAEL UF, because the BMD is considered to be an adverse effect; a UF
of 10 for interspecies extrapolation for oral studies; a UF of 10 for intraspecies variability; and a UF
of 3 for database deficiencies. The resulting RfD was 7 x 10"4 mg/kg/day, which is in reasonably
good agreement (within an order of magnitude) with the RfD of 4 x 10"3 mg/kg/day derived from the
Rothman et al. (1996a) human inhalation study.
An RfD was also derived from the LOAEL of 25 mg/kg observed in the NTP (1986) study.
The LOAEL was adjusted to a continuous exposure level of 17.9 mg/kg/day and then divided by a
UF of 3000 to derive an RfD of 6 x 10'3 mg/kg/day. The combined UF of 3000 is based on factors
of 10 for the absence of a NOAEL, 10 for interspecies extrapolation, 10 for intraspecies variability,
and 3 for database deficiencies. This value of 6 x 10"3 mg/kg/day is in good agreement with the
value 4 x 10"3 mg/kg/day derived from the Rothman et al. (1996a) human study.
xv
-------
1. INTRODUCTION
The overall purpose of this document is to provide a toxicological review of noncancer
health effects of exposure to benzene, including the no-observed-adverse-effect levels
(NOAELs), the lowest-observed-adverse-effect levels (LOAELs), benchmark dose (BMD)
modeling, and uncertainties in derivation of an inhalation reference concentration (RfC) and an
oral reference dose (RfD). This toxicological review document may serve as the scientific basis
for establishing a national air quality standard for ambient air exposure to benzene by the U.S.
Environmental Protection Agency's (EPA's) Office of Mobile Sources, Office of Air and
Radiation, under the Clean Air Act Amendments of 1990, Subchapter II (42 U.S.C. 7521-7590)
for hazardous air pollutants.
This document presents background and justification for the noncancer health hazard
and dose-response assessment summaries for benzene exposure in EPA's Integrated Risk
Information System (IRIS). IRIS summaries include an oral RfD, an inhalation RfC, and a
carcinogenicity assessment.
The RfD is based on the assumption that thresholds exist for certain toxic effects, such as
cellular necrosis, but may not exist for other toxic effects, such as some carcinogenic responses.
It is expressed in units of milligrams per kilograms per day (mg/kg/day). In general, the RfD is
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 noncancer effects during a lifetime. The inhalation RfC is analogous to the oral
RfD, but it provides a continuous inhalation exposure estimate. The inhalation RfC considers
toxic effects for the respiratory system (portal-of-entry effects) and for systems peripheral to the
respiratory system (extrarespiratory or systemic effects). It is generally expressed in units of
milligrams per cubic meter (mg/m3). The RfC and RfD do not represent a sharp dividing line
between safe and unsafe. In establishing the RfC and RfD, all relevant biologically significant
noncancer health effects in the published literature were reviewed and considered.
Development of these hazard identification and dose-response assessments for benzene
has followed the general framework for risk assessment, as set forth by the National Research
Council (1983). EPA guidelines that were used in the development of this assessment include
Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines
for Mutagenicity Risk Assessment (U.S. EPA, 1986b), Guidelines for Developmental Toxicity
Risk Assessment (U.S. EPA, 1991), Guidelines for Neurotoxicity Risk Assessment (U.S. EPA,
1998b), Proposed Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1996a), Draft Revised
Guidelines for Car'dnogen Risk Assessment (U.S. EPA, 1999b), Guidelines for Reproductive
Toxicity Risk Assessment (U.S. EPA, 1996b), Risk Assessment of Guidelines 1986 (U.S. EPA,
1987), Recommendations for and Documentation of Biological Values for Use in Risk
Assessment (U.S. EPA, 1988), Methods for Derivation of Inhalation Reference Concentrations
and Application of Inhalation Dosimetry (U.S. EPA, 1994), Use of the Benchmark Dose
Approach in Health Risk Assessment (U.S. EPA, 1995), and Benchmark Dose Technical
Guidance Document (external review draft; U.S. EPA, 2000b).
Literature search strategies employed for this compound were based on the Chemical
Abstracts Registry Number (CASRN) and at least one common name. At a minimum, the
1
-------
following databases were searched: RTECS, HSDB, TSCATS, CCRIS, GENETOX, EMIC,
EMICBACK, DART, ETICBACK, TOXLINE, CANCERLINE, MEDLINE, and MEDLINE
backfiles. Any pertinent scientific information submitted by the public to the IRIS Submission
Desk was also considered in the development of this document.
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS
Benzene also is known as benzol. Some relevant physical and chemical properties of
benzene are listed below.
CASRN: 71-43-2
• U.S. EPA Hazardous Waste No.: UO19 (in commercial product)
F005 (nonspecified source, spent solvent)
Hazardous Substance Data Bank (HSDB, 1997) No.: 2554
Synonyms: annulene, benzeen (Dutch), benzen (Polish), benzine, benzol, benzole,
benzolo (Italian), bicarburet of hydrogen, coal naphtha, cyclohjexatriene, fenzen (Czech),
mineral naphtha, motor benzol, NCI-C55276, phene, phenyl hydride, pyrobenzol,
pyrobenzole
• Registered trade name: Poly stream
Empirical formula: C6H6
Molecular weight: 78.11 (Budvari, 1989)
Vapor pressure: 75 mm Hg at 20°C (NFPA, 1994)
Water solubility: 1750 mg/L at 25°C (Banerjee etal., 1980). Miscible with ethanol,
ethyl ether, acetone, and chloroform.
Partition coefficients: LogKow: 2.13 (Hansch and Leo, 1985)
Log Koc: 1.8 to 1.9 (HSDB, 1997)
• Henry's law constant: 5.5 x 10"3 atm-m3/mol (Mackay and Leinonen, 1975) at 25°C
• Conversion factor: 1 ppm = 3.24 mg/m3 at 20°C; 1 mg/m3 = 0.31 ppm; 1 mg/L = 313 ppm
Melting point: 5.5°C (HSDB, 1997)
Boiling point: 80.1°C at 760 mm Hg (HSDB, 1997)
Odor: Aromatic (HSDB, 1997)
Lower 0.84 ppm (HSDB, 1997)
Upper 53 ppm (HSDB, 1997)
-------
Taste threshold: 0.5 to 4.5 mg/L (HSDB, 1997)
PEL (permissible exposure limit) TWA = 1 ppm (OSHA, 1987)
TLV (threshold limit value) TWA = 10 ppm (ACGIH, 1996)
3. TOXICOKINETICS RELEVANT TO ASSESSMENTS
The absorption, distribution, metabolism and excretion of benzene have been intensively
investigated in several experimental animal species and in humans. Benzene is readily absorbed
from oral and inhalation exposures. Dermal absorption is also rapid; however, quantitatively,
dermal absorption is very low due to rapid evaporation from skin. Benzene is rapidly distributed
throughout the body after exposure by all routes, and accumulation in fatty tissues is observed.
Metabolism of benzene is necessary for the expression of the characteristic hematotoxic and
carcinogenic effects of benzene. Despite extensive research, no single metabolite has been
identified as responsible for all the toxic effects of benzene, and the weight of evidence points
toward an interaction of several metabolites. At low doses, benzene is rapidly metabolized and
excretion occurs primarily as conjugated metabolites; however, at higher doses metabolic
pathways become saturated, and exhalation of unmetabolized benzene is observed to be the
primary route of excretion. Therefore, extrapolation from results observed at high doses
underestimates the potential toxic effects of low doses. The pathways of benzene metabolism
appear to be qualitatively similar across species. Quantitative differences in metabolism among
animal species are observed, however, and no good model for human metabolism has been
established. Thus, despite the development of several physiologically based pharmacokinetic
(PBPK) models, extrapolation of results from laboratory animal results to humans has proved
difficult.
3.1. ABSORPTION
Benzene is readily absorbed by both test animals and humans from inhalation, oral, and
dermal exposures (U.S. EPA, 2000a).
3.1.1. Gastrointestinal Absorption
Although only limited data are available on absorption of benzene in humans by the oral
route, accidental or intentional poisoning case studies indicate that benzene is readily absorbed
(Thienes and Haley, 1972).
Nearly complete absorption of orally administered benzene has been demonstrated in
laboratory animal studies. Parke and Williams (1953) performed a mass-balance study using
radiolabeled 14C-benzene with gavage exposures in rabbits at doses of 340-500 mg/kg.
Calculation of the mass balance 2-3 days after exposure showed that 84-89% of the
administered 14C could be accounted for as metabolites, CO2 and exhaled unchanged benzene.
Thus, approximately 90% of the dose was absorbed. Sabourin et al. (1987) similarly found that
gastrointestinal absorption of 14C-benzene administered to rats and mice at a dose range of 0.5 to
150 mg/kg was greater than 97%. Although this study used corn oil as a vehicle, it is reasonable
-------
to assume that oral absorption from water solutions would also be close to 100%. One could
expect that the presence of food in the stomach may delay absorption and an empty stomach may
enhance it.
The presence of soil had only minor effects on absorption of orally administered 14C-
benzene (Turkall et al., 1988). Male rats were gavaged with 5% aqueous solutions of gum
acacia with either benzene alone or with benzene preadsorbed onto a sandy soil or a clay soil.
Presence of soil had no significant effect on the absorption of benzene, as less than 2% of the
administered dose was excreted in the feces within 48 hours. Adsorption of benzene to soil
apparently increased the rate of gastrointestinal absorption, as the peak plasma benzene
concentrations were higher in both soil treatments, and the time to peak was reduced in the sandy
soil. The presence of both soils also increased the proportion of 14C excreted in the urine, as
opposed to excretion as unmetabolized benzene in expired air. Increased excretion in the urine
also indicated that the proportion of the benzene subject to metabolism was increased. Phenol
was the primary metabolite detected in the acid-hydrolyzed urine in all treatment groups,
followed by hydroquinone, catechol, and benzenetriol. No significant differences in the
distribution of metabolites among the treatments were found. Because the urine extracts were
acid-hydrolysed, conjugation products were not determined. Gastrointestinal absorption was
apparently rapid and efficient, even in the presence of soil.
3.1.2. Respiratory Absorption
There is a significant experimental database on the respiratory absorption of benzene in
humans (Nomiyama andNomiyama, 1974; Pekari et al., 1992; Srbova et al., 1950; Yu and
Weisel, 1998). Srbova et al. (1950) examined absorption of benzene by 23 human subjects
exposed to a range of concentrations, from 47 to 100 ppm (150 to 320 mg/m3), for 2-3 hours.
Absorption was greatest in the first 5 minutes of exposure (70-80%) but declined rapidly over
the next 15 minutes and varied between 20 and 60% after 1 hour and between 20 and 50% after 2
hours of exposure. Considerable variability between individuals was noted. Nomiyama and
Nomiyama (1974) determined both retention and uptake of benzene by three female and three
male subjects, 18-25 years of age, exposed to 52-62 ppm (166-198 mg/m3) for 4 hours.
Exhaled air was sampled every hour to determine respiratory excretion of benzene. Retention
declined from approximately 50% in the first hour and stabilized at 30% after 3 hours.
Respiratory uptake averaged 47%, with excretion of 17%. No significant differences between
males and females were reported; however, as noted, only three subjects of each sex were
examined.
Pekari et al. (1992) studied respiratory absorption of benzene in three males exposed to
1.7 and 10 ppm (5 and 39 mg/m3) benzene for 4 hours each. Absorption of benzene was
determined by measuring differences in concentration between inhaled and exhaled air. The
average absorption was 52% at the low concentration and 48% at the high concentration. In a
recent study, Yu and Weisel (1998) measured the uptake of benzene from sidestream tobacco
smoke by three female subjects. Benzene concentrations were measured in inhaled and exhaled
air. Smoke was generated by burning cigarettes, resulting in a variable benzene concentration
ranging from 32 to 69 ppm (102 to 220 mg/m3). Absorption in eight experiments averaged 64%,
with a range of 48-73%. The exposure periods were either 30 or 120 minutes, but no significant
decrease in absorption with the longer exposure period was observed.
-------
Considered together, these studies indicate that the respiratory absorption of benzene in
humans is approximately 50%. The observed decline in absorption with increasing exposure
time is apparently due to respiratory excretion of unmetabolized benzene. Animal studies
indicate that the metabolism of benzene is saturated at exposure concentrations in excess of 10
ppm (32 mg/m3). Thus, with the exception of the study by Pekari et al. (1992), benzene
metabolism may have been saturated in the human respiratory absorption studies. This would
have been expected to lead to greater respiratory excretion of unmetabolized benzene. Because
absorption at low dose levels is most relevant for derivation of an RfD, the estimates of 48-52%
retention derived by Pekari et al. (1992) are the most relevant values for evaluating absorption in
humans.
A number of test animal studies have also been conducted on the absorption of benzene
from inhalation exposure. Schrenk et al. (1941) noted a linear relationship between benzene
concentration (200-1300 ppm [639-4153 mg/m3]) and the equilibrium concentration of benzene
in the blood of dogs. A steady-state blood level was attained within 30 minutes at these
exposure concentrations. Exposure concentration appears to affect the retention of inhaled
radioactivity, as demonstrated by Sabourin et al. (1987) in rats and mice. The retention of
benzene by rats and mice during a 6-hour exposure decreased as exposure concentration
increased: 33 ± 6% to 15 ± 9% for rats, and 50 ± 1% to 10 ± 2% for mice, as exposure
concentration increased from 26 to 2600 mg/m3 (10 to 1000 ppm). This study also showed
species variability in the uptake and retention of inhaled benzene. At all exposure
concentrations, uptake was higher in mice. At exposure concentrations below 350 mg/m3, mice
retained approximately 50% more radioactivity per kilogram body weight than did rats, but there
was no significant difference at the highest (2500 mg/m3) concentration. In general, mice
inhaled greater amounts of benzene per kilogram body weight due to their higher relative minute
volume per kilogram than other species.
Henderson (1996) reviewed species differences in absorption and retention of benzene
given by either the oral or the inhalation route. Mice have both a higher respiratory rate and a
faster rate of metabolism for benzene than either rats or monkeys. After a 6-hour exposure to
7-10 ppm benzene, mice retained 20% of the dose, compared with 3-4% for rats and monkeys.
At higher exposure levels the metabolism becomes saturated and a sevenfold increase in
exposure concentration, between 130 and 925 ppm, results in only a threefold increase in
metabolism. When oral doses exceed the capacity for benzene metabolism, greater
concentrations are then exhaled.
3.1.3. Dermal Absorption
Studies of both humans and experimental animals indicate that benzene is rapidly
absorbed through the skin from both liquid and vapor phases. The percentage of absorption of
the applied doses is generally higher in experimental animals than in humans. The percutaneous
absorption of benzene has been studied in humans (Franz, 1984) and in laboratory animals
(Maibach and Anjo, 1981; Franz, 1983; Susten et al., 1985). Dermal absorption is minimal when
compared with inhalation or oral absorption; this is due in large part to benzene volatilizing
rapidly from the skin. If the absorption is based on the amount applied to the skin without
accounting for volatilization losses, then percentage absorption figures are low and usually less
-------
than 1%. Although this represents realistic exposure conditions for most situations, absorption
can be underestimated in situations where contact with the benzene source is maintained for a
prolonged period. A substantial portion of the absorbed benzene is excreted through the lungs.
Thus, experiments that measure excretion only in urine and feces substantially underestimate
true absorption.
Franz (1984) studied both in vivo and in vitro dermal absorption of benzene in humans.
A dose of 0.0026 mg/cm2 of 14C-benzene was applied to the ventral forearm skin of four
volunteers. Volatilization was rapid, as no free benzene on the skin was observed after 30
seconds. Absorption of the applied dose was estimated by urinary excretion of 14C for 36 hours.
No correction was made for volatilization, but a correction was made for the fact that not all
absorbed benzene is excreted in the urine. Correction was determined by the fraction of 14C from
a subcutaneous injection in rhesus monkeys that was not excreted in the urine. Excretion was
rapid, with 80% of the urinary excretion occurring by 8 hours after exposure. Total absorption
was estimated to be 0.05% of the applied dose. Absorption using in vitro techniques with human
skin was estimated to be 0.1% at a similar dose rate. Absorption in vitro increased linearly with
dose level and also as a function of exposure time.
Laitinen et al. (1994) studied occupational exposure to benzene in eight car mechanics in
Finland. Blood samples were taken 3-9 hours after exposure. The approximated benzene
concentration in blood corresponding to the time point of 16 hours after exposure showed much
higher levels of exposure than expected, based on corresponding air concentrations in the
workplace. Actual breathing zone concentrations of benzene varied from the Finish detection
limit of 0.2 cm3/m3 to 3.7 cm3/m3, depending on whether it was in unleaded or leaded gasoline.
Comparison of measured blood concentrations to predictions based on air measurements
suggested that dermal exposure could have accounted for 68% of exposure. These mechanics
had direct dermal contact with gasoline during the frequent changing of filters and fuel pumps.
This suggests far more dermal exposure to benzene than exposure via the inhalation route.
In a series of experiments conducted in a residence with a benzene-contaminated water
supply, Lindstrom et al. (1993) estimated, by modeling, that a dose of 281 |o,g benzene would be
absorbed during a 20-minute shower, with 60% derived from dermal absorption and 40% from
inhalation absorption, based on results from one individual only. The validity of this observation
is highly questionable.
Several dermal absorption studies have also been conducted with experimental animals.
In rhesus monkeys, minipigs, and hairless mice, dermal absorption was < 1% following a single
direct application of liquid benzene (Franz, 1984; Maibach and Anjo, 1981; Susten et al., 1985).
Absorption was rapid, with the highest urinary excretion observed in the first 8 hours following
exposure (Franz, 1984; Susten et al., 1985). Multiple applications, as well as application to
cellophane tape-stripped skin resulted in greater skin penetration (Maibach and Anjo, 1981). It
may be noted the percent absorption of the applied dose of benzene in each of these test animals
was approximately twofold to threefold higher than that of humans.
Dermal uptake of benzene from aqueous solutions is an important parameter for
evaluating the risk due to exposure to benzene-contaminated ground or surface waters, for
example, during showering, bathing, or swimming. Morgan et al. (1991) compared dermal
-------
uptake of undiluted benzene with one-third, two-thirds, and completely saturated aqueous
solutions of benzene. A capped skin depot (3.1 cm2 surface area) was glued to the shaved skin of
male Fischer 344 (F344) rats, and 2 mL of benzene or aqueous solutions were used. Blood
samples were collected via an implanted jugular catheter at 0, 0.5, 1, 2, 4, 8, 12, and 24 hours
after initiation of exposure. The blood concentration during exposure to undiluted benzene
continued to rise throughout the 24-hour exposure and reached a level of 24.2 mg/L. Dermal
absorption from aqueous solutions was rapid, and peak blood levels of benzene were reached
after 1 hour. Benzene could not be detected in blood from the one-third saturated solution; it
reached a peak level of approximately 0.18 mg/L in the two-thirds saturated treatment and 0.33
mg/L in the saturated treatment. Benzene was essentially completely absorbed, with less than
1% remaining in solution after 24 hours. Benzene was preferentially absorbed, because the
volume of aqueous benzene solution absorbed did not differ significantly from absorption of
distilled water. Benzene was rapidly and completely absorbed from aqueous solutions.
Tsuruta (1989) reported that dermal absorption of benzene increased linearly with dose in
hairless mice exposed to benzene vapors (the mice were attached to respirators to avoid
inhalation exposure). The dermal absorption rates at exposure concentrations of 200, 1000, and
3000 ppm (639, 3195, and 9584 mg/m3) were 4.11, 24.2, and 75.5 mmol/cnrVhour, respectively.
This is equivalent to an absorption rate of 0.31, 1.89, and 5.90 |ig/cm2/hour, respectively. The
skin absorption coefficient was 0.619 cm/hour. Using the mouse dermal absorption data and
human occupational exposure data, Tsuruta estimated that skin absorption of benzene by humans
would be 3.7% that of inhalation exposure at the same concentration.
Permeability constants for dermal absorption of benzene vapors were also estimated by
McDougal et al. (1990). In this study, rats were supplied air through latex masks and exposed to
benzene vapor at a concentration of 40,000 ppm (127,787 mg/m3) for 4 hours. Blood
concentration was monitored at 0.5, 1, 2, and 4 hours. A PBPK model was used to estimate the
permeability of the vapor in rat skin as well as human skin; the rat and human permeability
constants were estimated as 0.15 and 0.08 cm/hour, respectively. On the basis of these findings,
dermal exposure studies in rats probably provide a conservative estimate of the dermal
absorption of benzene by humans. In an in vitro experiment using F344 rat skin, Mattie et al.
(1994) determined a skin:air partition coefficient of 35 for benzene at 203 ppm (649 mg/m3),
with an equilibration time of 4 hours.
Adsorption of benzene to soils resulted in a small decrease in dermal absorption when
compared with free benzene (Skowronski et al., 1988). Radiolabeled 14C-benzene was applied to
the shaved skin of male Sprague-Dawley rats either as free benzene or mixed with a sandy soil
(4.4% organic matter) or a clay soil (1.6% organic matter). The application area was occluded
by gluing a shallow glass cover onto the skin prior to application. Statistically significant
decreases in the area under the plasma-concentration time curve were observed for both soils,
but the effect of the clay soil was greater. No statistically significant differences in the
absorption or elimination half-lives among treatments were observed. The greatest excretion of
14C was observed in the urine (45-86% of applied dose), with lesser amounts eliminated by
respiratory excretion (6-13%). Excretion in the feces was 0.2% or less. These values were
corrected for volatilization losses during administration of 67%, 59%, and 39% of applied
radioactivity for the free benzene, sandy soil, and clay soil treatments, respectively. Although
these experiments suggested that soil may reduce benzene absorption, the use of occlusion of the
-------
exposed skin is not realistic for the dermal exposure scenarios used in risk assessment. Thus, the
small decreases in uptake have little meaning for evaluation of risks from soil-absorbed benzene.
3.2. DISTRIBUTION
There were no studies reported regarding the distribution of benzene in humans after
either oral or dermal exposure. However, the animal data available provided some useful
information. Both human and animal data were available following inhalation exposure.
3.2.1. Oral Exposure
Low et al. (1989, 1995) followed the distribution of 14C-benzene administered by oral
gavage in female Sprague-Dawley rats. This study sought to determine whether the distribution
of benzene or benzene metabolites could explain the observations of tissue-specific induction of
solid tumors in the Zymbal gland, oral and nasal cavities, and mammary glands associated with
chronic oral exposure (NTP, 1986). Following oral doses of 0.15 or 1.5 mg/kg in olive oil, the
concentrations of 14C in various organs could be assigned to three groups, with the highest
concentrations in liver and kidneys, intermediate concentrations in the blood, and lowest
concentrations in the Zymbal gland, nasal cavity tissue, oral cavity tissue, mammary glands, and
bone marrow. At dose levels of 15 mg/kg or higher, disproportionate increases were found in
mammary glands and bone marrow. At the 0.15 mg/kg dose level, all of the 14C activity found in
the tissues or blood after 1 hour appeared as benzene metabolites, indicating that first-pass
metabolism in the liver was very efficient. A high proportion of the metabolites extracted by
ethyl acetate from the Zymbal gland and the nasal cavity tissue 1 hour after administration were
present as unidentified metabolites. Similarly, a high proportion of the water-soluble metabolites
from these tissues were also unidentified peaks that did not correspond to known metabolites of
benzene. Elimination of benzene-derived 14C in all organs was biphasic. Benzene did not
accumulate in the Zymbal gland; within 24 hours after administration, radiolabel derived from
14C-benzene in the Zymbal gland constituted less than 0.0001% of the administered dose. Thus,
preferential accumulation of benzene or metabolites in the Zymbal gland does not account for
tissue-specific tumor induction in this organ.
3.2.2. Inhalation Exposure
Results from test animal studies indicate that absorbed benzene is distributed throughout
several body compartments. The parent compound is preferentially stored in the fat, although
the relative uptake in tissues also appears to be dependent on the perfusion rate of tissues by
blood. Steady-state benzene concentrations in rats exposed via inhalation to 1600 mg/m3 (500
ppm) for 6 hours were blood, 1.2 mg%; bone marrow, 3.8 mg%; and fat, 16.4 mg% (Rickert et
al., 1979). Benzene was also found in the kidney, lung, liver, brain, and spleen. Levels of the
benzene metabolites phenol, catechol, and hydroquinone were higher in the bone marrow than in
blood, with phenol being eliminated more rapidly than catechol or hydroquinone after exposure.
Ghantous and Danielsson (1986) exposed pregnant mice to a benzene concentration of 6400
mg/m3 (2000 ppm) for 10 minutes and found benzene and its metabolites in lipid-rich tissues
such as brain and fat as well as in perfused tissues such as liver and kidney. Benzene was also
found in the placenta and fetuses immediately following exposure.
-------
Benzene was rapidly distributed throughout the bodies of dogs exposed via inhalation to
concentrations of 800 ppm (2556 mg/m3) for up to 8 hours per day for 8-22 days (Schrenk et al.,
1941). Fat, bone marrow, and urine contained about 20-fold the concentration of benzene in
blood; benzene levels in muscles and organs were onefold to threefold that in blood, and
erythrocytes contained about twice the amount of benzene found in plasma.
Studies in pregnant mice demonstrated that after inhalation exposure, 14C-benzene
crossed the placenta. Volatile radioactivity (unmetabolized benzene) was observed in the
placenta and fetuses immediately after and up to 1 hour after exposure (Ghantous and
Danielsson, 1986). Nonvolatile metabolites were also detected in the fetus, but at lower levels
than in maternal tissues. The label peaked in fetal tissues 30 minutes to 1 hour after inhalation,
similar to the peak in maternal tissues. No firmly tissue-bound metabolites of benzene were
detected in the fetal tissues in late gestation, indicating that the mouse fetus did not have the
ability to form the reactive metabolites.
Sato et al. (1975) compared elimination kinetics of benzene in men and women of similar
ages. Exposure was for 2 hours at 25 ppm. The level of benzene in the blood and the end-tidal
air was different for males and females. The shape of the decay curve was significantly steeper
in the males. The authors attributed these results to the higher fat content of females.
3.2.3. Dermal Exposure
No relevant studies were found regarding distribution in humans following dermal
exposure to benzene, and only one study in animals was identified, providing minimal
information on the distribution of benzene via the dermal route.
A study of male rats treated dermally with 0.004 mg/cm2 of 14C-benzene, with and
without 1 g of clay or sandy soil, revealed soil-related differences in tissue distribution following
treatment. The 14C activity (expressed as a percentage of initial dose per gram of tissue) 48
hours after treatment with soil-absorbed benzene was greatest in the treated skin
(0.059-0.119%), followed by the kidney (0.024%) and liver (0.013-0.015%), in both soil
groups. In the pure-benzene group, the kidney contained the largest amount of radioactivity
(0.026%), followed by the liver (0.013%) and treated skin (0.11%) (Skowronski et al., 1988). In
all three groups, less than 0.01% of the radioactivity was found in the following tissues:
duodenum, fat, bone marrow, esophagus, pancreas, lung, heart, spleen, blood, brain, thymus,
thyroid, adrenal, testes, untreated skin, and carcass.
3.3. METABOLISM
Despite extensive research, the metabolism of benzene is still not thoroughly understood.
It is generally accepted that benzene itself is not directly responsible for causing the toxic
effects; however, the metabolic product or products responsible for the noncancer and
carcinogenic effects of benzene exposure have not been clearly defined. The evidence suggests
that several metabolites, as well as interactions between these metabolites, may be necessary to
explain the toxic effects of benzene. A complete review of the metabolism of benzene is beyond
the scope of this toxicological review. The reader is referred to several excellent reviews for
more complete coverage (ATSDR, 1997; Snyder et al., 1993b; Snyder and Hedli, 1996; Ross,
-------
1996; Witz et al.,1996). This section provides an overview of benzene metabolism, focusing on
the issues that are relevant to the interpretation of the noncancer effects due to benzene exposure.
3.3.1. Metabolic Pathways
The first step in benzene metabolism is the formation of the epoxide, benzene oxide,
catalyzed by cytochrome P450 2E1 (CYP2E1) (Figure 1). After formation of benzene oxide, the
metabolism of benzene branches into several alternative metabolic pathways (Jerina et al., 1968;
Lovern et al., 1997). However, although Jerina et al. (1968) provided a kinetically plausible
argument for the existence of benzene oxide, they did not show that it was formed from benzene.
Benzene oxide rearranges nonenzymatically to form phenol, the major product of initial benzene
metabolism. Alternatively, benzene oxide may react with glutathione (GSH) to form phenyl-
mercapturic acid; undergo enzymatic conversion by epoxide hydrolase to benzene dihydrodiol
with subsequent formation of catechol; or undergo an iron-catalyzed, ring-opening reaction to
form ^ram^rara'-muconaldehyde (MUC) with subsequent metabolism to trans,trans-muconic
acid (MA). Phenol is further oxidized by CYP2E1 catalysis to hydroquinone. Further oxidation
of hydroquinone to/?-benzoquinone is catalyzed by myeloperoxidase (MPO) (Smith et al., 1989).
All of the phenolic products may be conjugated with sulfate or glucuronic acid, and the
conjugates of phenol and hydroquinone are the major benzene metabolites excreted in urine
(Sabourin et al., 1989; Wells and Nerland, 1991).
Mathews et al. (1998) studied the metabolism of 14C-benzene in male F344 rats over a
wide range of oral (gavage) doses, 0.02, 0.1, 0.5, or 100 mg benzene/kg body weight, in male
B6C3F1 mice at oral doses of 0.1 and 100 mg/kg and in male hamsters at 0.02, 0.1 and 100
mg/kg. In F344 rats, at lower doses (0.02, 0.1, and 0.5 mg/kg), greater than 95% of the dose was
recovered in the urine within 48 hours, and a small percentage (about 3%) was recovered in the
breath. At higher doses, 10 and 100 mg/kg, the percentage eliminated in the breath increased to
about 9 and 50%, respectively. Excretion in the feces was a minor route at all doses. A similar
pattern of disposition of the radiolabel dose was also observed in mice and hamsters. Mathews
et al. (1998) also examined the profile of urinary metabolites formed. Interestingly, the
percentage of prephenylmercapturic acid and phenylmercapturic acid, indicators of benzene
oxide production, was relatively constant across all doses for rats (-13%), mice (-5%), and
hamsters (-7%). However, the percentage of hydroquinone and related conjugates ranged from
about 3% at the highest dose to as much as 7% at the lowest doses. A higher percentage of
hydroquinone metabolites was seen in mice (-30%) and in hamsters (-30%), but it did not
appear to be dose dependent.
Figure 1. Metabolic pathways for benzene.
Source: Ross, 2000.
3.3.2. Requirement for CYP2E1
Oxidation of benzene by the CYP2E1 isoenzyme has been demonstrated to be required
for the expression of hematotoxicity and genotoxicity. Administration of toluene, a competitive
inhibitor of benzene metabolism, causes a decrease in benzene metabolite formation and a
reduction in toxicity (Andrews et al., 1977). The primary oxidation of benzene by CYP2E1
10
-------
I'!-. Kt JXID ASKS/
U"I'OXII>ATK)>
occurs in the liver, and further metabolism to the final toxic compound occurs in the target
tissues. Sammett et al. (1979) showed that partial hepatectomy of rats diminished both the rate
of metabolism of benzene and its toxicity, suggesting that a metabolite and/or metabolites
formed in the liver are necessary for toxicity. Immunoinhibition studies in rat and rabbit hepatic
microsomes also have implicated CYP2E1 as the major oxidative isoenzyme involved in
benzene metabolism (Johansson and Ingelman-Sundberg, 1988; Koop and Laethem, 1992).
Convincing evidence that CYP2E1 activity is required for expression of benzene toxicity was
provided using transgenic CYP2E1 knockout mice that do not express hepatic CYP2E1 activity
(Valentine et al., 1996). Benzene exposure at 200 ppm (640 mg/m3) for 6 hours/day for 5 days
resulted in severe genotoxicity and cytotoxicity in wild-type mice, but no toxicity was observed
in CYP2E1 knockout mice. Thus, the requirement for CYP2E1 metabolism has been clearly
established, but the identity of the toxic metabolites remains uncertain.
Recently, Bernauer et al. (2000) investigated a role of CYP2E1 expression in bone
marrow and its intra- and interspecies variability in rats, rabbits, and humans, because it is a
target organ for several chemicals, including benzene. The data demonstrated a presence of
CYP2E1 in the bone marrow of all species investigated, thus supporting the hypothesis of
CYP2E1-dependent local metabolism of several chemicals that includes benzene possibly
contributing to myelotoxicity and hematotoxicity.
11
-------
3.3.3. Toxicity of Benzene Metabolites
The products of the phenol pathway (catechol, hydroquinone, and/>-benzoquinone), MA,
and benzene epoxide have been proposed as benzene metabolites that may cause the toxic effects
of benzene exposure.
3.3.3.1. Phenolic Products
The production of the initial benzene metabolites occurs primarily in the liver; however,
the effects of benzene toxicity are primarily expressed as hematotoxicity and myelotoxicity in
the bone marrow (Snyder and Hedli, 1996). The evidence suggests that secondary metabolism in
bone marrow is required for expression of the toxicity of benzene (Schlosser and Kalf, 1989;
Subrahmanyam et al., 1990, 1991). An alternative hypothesis is that metabolism of benzene to
toxic metabolites can occur in the bone marrow itself. Metabolism of benzene to hydroquinone
in the bone marrow of rats has been demonstrated, but the amount of metabolites produced was
small (Irons et al., 1980). The presence of CYP2E1 has been detected in rabbit bone marrow
(Schnier et al., 1989); however, CYP2E1 could not be detected in the bone marrow of mice, the
species demonstrated to be most sensitive to benzene hematotoxicity (Genter and Reico, 1994).
Rickert et al. (1979) observed that catechol and hydroquinone concentrations persisted in the
bone marrow long after blood levels had declined following inhalation exposure.
Most research, therefore, has focused on the possibility that phenol, catechol, and
hydroquinone, generated by reactions in the liver, are transported to the bone marrow and other
target tissues and subsequently activated by the action of peroxidase (Smith et al., 1989;
Rushmore et al., 1984; Low et al., 1995). A potential role for conversion of hydroquinone top-
benzoquinone by the peroxidase component of prostaglandin H synthase has also been suggested
(Schlosser and Kalf, 1989). Subsequent research, however, suggested that the prostaglandin H
synthase is not involved in benzene toxicity (Ganousis et al., 1992).
A major problem with the hypothesis that phenolic metabolites are responsible for bone
marrow toxicity is that administration of phenol fails to duplicate the effects of benzene (Tunek
et al., 1981; NCI, 1980). Kenyon et al. (1995) suggested that the distribution of phenol-
conjugating enzymes and benzene-oxidizing enzymes within the liver might have accounted for
this result. The authors observed that phenol-conjugating enzymes were more concentrated in
the periportal area of the liver, the first region to absorb orally administered phenol, whereas
oxidizing enzymes were more concentrated in the pericentral region of the liver. This could
have lead to rapid excretion of orally administered phenol before it was further metabolized to
hydroquinone by CYP2E1 in the pericentral region of the liver. A combination of phenol and
hydroquinone is needed to cause bone marrow toxicity (Eastmond et al., 1987). Intraperitoneal
(i.p.) injection of either phenol or hydroquinone alone in B6C3F1 mice failed to cause significant
bone marrow toxicity; however, co-administration of these metabolites caused a reduction in
bone marrow cellularity with a clear dose-response curve. The presence of phenol apparently
stimulated peroxidase-dependent metabolism of hydroquinone to/>-benzoquinone. Increased
covalent binding of 14C-hydroquinone was observed in bone marrow when a combination of
phenol and hydroquinone was administered to mice (Subrahmanyam et al., 1990, 1991).
12
-------
In contrast to the observations of Eastmond et al. (1987), catechol was reported to
markedly stimulate peroxidase activation of hydroquinone in murine stroma (Ganousis et al.,
1992). In addition to the effects of phenol on the activity of MPO, co-administration of phenol
was shown to increase the concentration of phenol and hydroquinone in the blood in B6C3F1
mice. Legathe et al. (1994) measured the area under the curve (AUC) blood concentration-time
curve for phenol and hydroquinone administered alone or in combination at the same dose levels
used by Eastmond et al. (1987). Co-administration increased the phenol AUC by 1.4-fold and
the hydroquinone AUC by 2.6-fold in comparison to each compound administered alone.
Legathe et al. (1994) suggested that this resulted from saturation of the enzymes that form sulfate
and glucuronide conjugates of phenolics in the liver. These results suggest that interactions of
two or more phenolic metabolites of benzene may be necessary to cause the observed bone
marrow toxicity.
3.3.3.2. Trans, trans-Muconaldehyde
Muconaldehyde (MUC), a highly reactive six-carbon diene dialdehyde, has also been
proposed to be a benzene metabolite responsible for bone marrow toxicity. Evidence supporting
this hypothesis has been summarized by Witz et al. (1996). MUC has been shown to cause
hematotoxicity following short-term exposures. Administration of 2 mg/kg/day MUC to CD-I
mice for 16 days caused significant decreases in bone marrow cellularity, lymphocytes, red
blood cell (RBC) counts, hematocrit (HCT) (volume percentage of erythrocytes in whole blood),
and hemoglobin (Hgb) and significant increases in white blood cell (WBC) count and spleen
weight (Witz et al., 1985). Snyder et al. (1989) also found that administration of MUC caused
bone marrow toxicity in mice and that co-administration of MUC and hydroquinone resulted in a
dramatic decrease in 59Fe incorporation into red cell Hgb.
Production of MUC is expected as a step in the ring-opening pathway leading to MA.
Excretion of MA in the urine was demonstrated in rabbits in the early work of Parke and
Williams (1953), and excretion of MA was also demonstrated in mice, rats, cynomolgus
monkeys, chimpanzees, and humans (Gad-El Karim et al., 1985; Sabourin et al., 1988a, 1989,
1992). Metabolism of MUC to MA has been shown in vivo in mice (Witz et al., 1990a,b).
Urinary excretion of MA has been used as a sensitive biomarker of benzene exposure in humans.
At low doses, urinary MA concentration was found to be linearly correlated with time-weighted
average (TWA) benzene exposure concentrations (Bechtold and Henderson, 1993; Bechtold et
al., 1991). Thus, the ring-opening pathway is active in several animal species. There is no direct
evidence, however, for the in vivo formation of MUC from benzene.
Although MUC formation has not been demonstrated in animals in vivo, formation of
MUC from benzene has been demonstrated in a mouse hepatic microsomal system (Latriano et
al., 1986; Zhang et al., 1995a). MUC may be too reactive to be isolated from in vivo systems.
Using isolated rat livers perfused with 0.7 mM benzene solutions through the portal vein, Grotz
et al. (1994) demonstrated that the complete pathway for formation of MA from benzene was
active in the liver, but MUC was not detected in the perfusate collected via the hepatic vein.
When MUC was added to the perfusion solution, it was rapidly and efficiently metabolized to
MA, and only traces of MUC were detected in the perfusate collected after a single pass through
the isolated rat livers. Thus, it seems unlikely that sufficient quantities of MUC could reach the
target tissues by circulation in the blood. The acid-alcohol 6-hydroxy-^rami,^ram'-2,4-
13
-------
hexadienoic acid (COOH-M-OH) and MA were the major MUC metabolites detected in the rat
liver perfusate, but both the acid-aldehyde 6-oxo-^rarni,^rami-2,4-hexadienoic acid (COOH-M-
CHO) and the aldehyde-alcohol 6-hydroxy-^rami,^ram'-2,4-hexadienal (CHO-M-OH) were
detected as minor peaks (Grotz et al., 1994). The CHO-M-OH has been demonstrated to react
with GSH, to be cytotoxic to isolated rat hepatocytes, and to be hematotoxic in mice (Goon et
al., 1993; Zhang et al., 1995b). Thus, CHO-M-OH may be the ring-opened metabolite that
causes the hematoxic effects of administered MUC, and it may play a role in causing the toxic
effects of benzene exposure.
3.3.3.3. Benzene Oxide
Benzene oxide was believed to be too reactive to escape the liver and to cause toxicity in
the bone marrow. However, Lovern et al. (1997) recently showed that benzene oxide constituted
7% of the benzene metabolites after 18 minutes' incubation with liver microsomes. Lindstrom et
al. (1997) demonstrated the presence of benzene oxide in the blood and estimated its half-life to
be about 8 minutes. Using a PBPK model, Lindstrom et al. (1997) predicted that the dose to the
body from benzene oxide would be about 22-fold greater than from 1,4-benzoquinone. Thus,
circulating benzene oxide can contribute to observed DNA and protein adduct formation.
Further research is needed to establish the role of benzene oxide in causing bone marrow
toxicity.
3.3.4. Species, Route, and Rate Differences
Differences in the rates of benzene metabolism and the metabolites formed correlate with
observed differences in sensitivity to benzene toxicity. Mice are more sensitive than rats to the
toxic effects of benzene (Huff et al., 1989; Snyder et al., 1978; Ward et al., 1985). Sabourin et
al. (1987) found that metabolism of benzene, determined by excretion of water-soluble
metabolites in the urine, became saturated at lower doses in B6C3F1 mice than in F344/N rats.
The amount of metabolites per kilogram body weight was similar in the two species at gavage
doses up to 50 mg/kg, but above this level, total metabolites in rats continued to increase but
total metabolites in mice did not increase further. Following inhalation exposures, total
metabolites excreted were higher for mice than for rats at all benzene concentrations due to a
higher amount inhaled by mice. Total metabolite excretion was exponentially related to benzene
concentration, but the concentration needed to reach half of the maximal metabolite formation
was lower in mice (220 mg/m3 [69 ppm]) than in rats (260 mg/m3 [81.5 ppm]). By both oral and
inhalation routes, the metabolism of benzene was saturated at lower concentrations in mice than
in rats. Thus, differences in total metabolite formation between rats and mice did not explain the
greater sensitivity of mice to benzene.
Sabourin et al. (1988a) compared the metabolites of benzene in the blood, liver, lung, and
bone marrow of male B6C3F1 mice and F344/N rats following 6 hours of inhalation exposure to
benzene at 50 ppm (160 mg/m3). Hydroquinone glucuronide, hydroquinone, and MA were
present in much higher concentrations in the mouse than in the rat tissues. Thus, metabolism in
mice leads to much greater exposures to potentially toxic benzene metabolites, which may
explain the greater sensitivity of mice to benzene toxicity.
14
-------
Exposure concentration and routes also affect the distribution of metabolites formed in
mice and rats (Sabourin et al., 1989). Male B6C3F1 mice and F344/N rats were given gavage
doses of 1, 10, or 200 mg/kg benzene or they were exposed to 5, 50, or 600 ppm (16, 160, or
1916 mg/m3) for 6 hours. Water-soluble metabolites were determined in blood, urine, liver,
lung, and bone marrow. A shift in the distribution from potentially toxic metabolites,
hydroquinone glucuronide, and MA to the detoxification metabolites, phenylglucuronide and
prephenylmercapturic acid, was observed from low concentrations to high concentrations
following both oral and inhalation exposures in mice. A similar shift was observed in rats,
except that hydroquinone glucuronide was a minor metabolite in rats at all concentrations. Thus,
extrapolation from toxicological studies conducted at high exposure concentrations may
significantly underestimate risks at lower exposure concentrations. There was also no simple
correlation between the dose of metabolites to tissues resulting from oral and inhalation
exposures. This may be due to saturation of benzene metabolism following bolus doses of
benzene administered by gavage.
Urinary excretion of benzene metabolites has also been examined in cynomolgus
monkeys and chimpanzees (Sabourin et al., 1992). The proportion of an oral gavage dose
excreted in the urine decreased from 50 to 15% as the dose rate increased from 5 to 500 mg/kg in
cynomolgus monkeys. Phenyl sulfate was the primary metabolite excreted, and the proportion
excreted as hydroquinone conjugates and MA decreased as the dose increased, as was observed
previously in mice and rats. The proportion excreted as MA decreased from 4.4% at the low
dose to 1.3% at the high dose. At all levels the proportion of MA excreted as MA was much
lower in the monkey than in either mice or rats. Chimpanzees were exposed at only one low
dose rate of 1 mg/kg by intravenous injection. Less than 15% of the administered dose was
recovered as hydroquinone conjugates and MA, and 79% was recovered as phenyl conjugates.
On the basis of the urinary metabolite profiles, the mouse appears to metabolize the largest
fraction of benzene via pathways leading to hydroquinone conjugates and MA (67% at low
doses) followed by monkeys (31%), rats (17%), and chimpanzees (14%). Thus, there are
apparently large quantitative differences among species in the metabolism of benzene to
potentially toxic metabolites. Because few data are available on the proportion of benzene
metabolized to potentially toxic metabolites in humans, considerable uncertainty exists in
determining which animal model best represents human metabolism.
Henderson (1996) reviewed the species differences in benzene metabolism. The
pathways of benzene metabolism appear to be similar in all species studied; however, there are
quantitative differences in the fraction of benzene metabolized by different pathways. Monkeys
and mice metabolize more of the benzene dose to hydroquinone metabolites than do rats or
chimpanzees, especially at low doses. Mice appear to have a greater overall capacity to
metabolize benzene than do rats and primates. This finding may explain why mice are more
sensitive than rats to benzene. In all species, a greater proportion of benzene is converted to
hydroquinone and ring-open metabolites at low doses than at high doses.
3.3.5. Induction of CYP2E1
Benzene exposure has been found to induce CYP2E1 activity, thereby increasing the rate
of toxic metabolite formation. Pretreatment of mice, rats, and rabbits subcutaneously with
benzene increased benzene metabolism in vitro without increasing total cytochrome P450
15
-------
concentrations (Arinc et al., 1991; Gonasun et al., 1973; Saito et al., 1973). In contrast, Sabourin
et al. (1990) found no significant effect on the metabolism of benzene when F344 rats and
B6C3F1 mice were pretreated by inhalation exposure to 600 ppm (316 mg/m3) of benzene. The
rate of benzene metabolism can be altered by pretreatment with various compounds. CYP2E1
also metabolizes alcohol and aniline, and CYP2E1 can be induced by these substrates (Chepiga
et al., 1990; Parke, 1989; Snyder et al., 1993a). Phenol, hydroquinone, benzoquinone, and
catechol have also been shown to induce P450 in human hematopoietic stem cells (Henschler
and Glatt, 1995). Therefore, exposure to chemicals that stimulate the activity of this enzyme
system prior to exposure to benzene could increase the rate of benzene metabolism.
Daiker et al. (1996) found that repeated oral benzene exposure of female B6C3F1 mice
for 3 weeks at 50 mg/kg/day decreased CYP2E1 activity by 34% and activated the detoxification
enzyme GSH transferase by 30% without affecting aldehyde dehydrogenase, another detoxifying
enzyme. The authors suggested that these changes in enzyme activity may serve a protective
role against repeated benzene exposure.
Exposure to benzene was found to interfere with the disappearance of ethanol from the
body in rats (Nakajima et al., 1985). The results of further studies showed that ethanol treatment
increased the production of hydroxylated benzene metabolites, phenol, and hydroquinone,
suggesting induction of benzene metabolism (Nakajima et al., 1987). The possibility therefore
exists of a synergism between alcohol and benzene.
3.3.6. Mechanism of Toxicity
Just as several benzene metabolites have been implicated in benzene toxicity, several
different mechanisms also may contribute to the overall toxic effects of benzene, as discussed
below.
3.3.6.1. Formation of Covalent Adducts
Benzene metabolites form covalent adducts with both cell proteins and DNA; however,
the role of adduct formation in toxicity is unclear. Treatments that reduce benzene toxicity also
reduce adduct formation. Partial hepatectomy of rats resulted in reduced hematotoxicity and also
in reduced covalent binding in bone marrow (Sammett et al., 1979). Levels of adduct formation
in the hematopoietic tissues of mice strains have been shown to correlate with the relative
sensitivity of the strains (Longacre et al., 1981).
Covalent binding of benzene metabolites to DNA was first demonstrated by Lutz and
Schlatter (1977). Mazzullo et al. (1989) found that formation of DNA adducts of benzene
metabolites was linear at low benzene concentrations, but saturation of adduct formation
occurred at high benzene concentrations. Hedli et al. (1996) investigated DNA adduct formation
from the benzene metabolites hydroquinone and 1,2,4-benzenetriol in combination with
investigations of the effects of these two metabolites on cell differentiation in a hematopoiesis
model system. Hydroquinone formed DNA adducts in human promyelocytic leukemia cells, but
1,2,4-benzenetriol did not. Both metabolites, however, inhibited retinoic acid-induced
maturation of human promyelocytic leukemia cells to granulocytes. Thus, DNA adduct
formation may be important in hydroquinone but not in 1,2,4-benzenetriol toxicity.
16
-------
Creek et al. (1997) showed that DNA and protein adduct formation following i.p.
administration of 14C-benzene was linear over a dose range spanning eight orders of magnitude
in B6C3F1 mice. At doses greater than 16 mg/kg body weight, however, adduct formation was
nonlinear. This corresponds very closely to the metabolic saturation level of 15 mg/kg body
weight observed by Sabourin et al. (1987). Benzene was administered to male B6C3F1 mice
over a dose range of 700 pg/kg to 500 mg/kg body weight. Liver DNA adduct levels peaked at
0.5 hours after exposure, but bone marrow DNA adduct levels peaked between 12 and 24 hours.
These results indicate that adduct formation is linear in the range of benzene concentrations
where benzene toxicity is first detected and becomes saturated at higher doses where most
toxicity experiments have been conducted.
Although the mechanism by which adduct formation results in bone marrow toxicity is
not well established, a few experiments have suggested that binding of hydroquinone orp-
benzoquinone to sulfhydryl groups at the active sites of proteins could be responsible for the
genotoxic effects of benzene. Hydroquinone and/>-benzoquinone interfere with binding of
guanosine triphosphate (GTP) to tubulin by alkylating nucleophilic sulfhydryl groups (Irons and
Neptun, 1980; Pfeifer and Irons, 1983). Binding of GTP is required for stabilization of tubulin
polymerization during microtubule formation and, therefore, interferes with spindle formation
during mitosis. Several metabolites of benzene have also been shown to inhibit the activity to
human topoisomerase II (Frantz et al., 1996). Topoisomerase II enzymes relieve torsional strain
on DNA during replication and transcription and are also believed to function during
recombination and chromosome condensation. Both/>-benzoquinone and MUC directly
inhibited topoisomerase at concentrations as low as 10 |j,M. With the addition of horseradish
peroxidase and peroxide to the assay mixture, the benzene metabolites phenol, 4,4'-biphenol,
2,2'-biphenol, hydroquinone, catechol, and 1,2,4-benzenetriol all inhibited topoisomerase II.
Topoisomerase inhibitors tend to be strong clastogens in mammalian cells. Thus, inhibition of
topoisomerase II could lead to the clastogenic effects observed following benzene exposure.
Addition of GSH to the assay mixture also protects against topoisomerase II inhibition, thus
suggesting that interaction of benzene metabolites with an essential sufhydryl residue occurs.
Hutt and Kalf (1996) also reported that/?-benzoquinone inhibits the activity of topoisomerase II.
Protein adduct formation by benzene metabolites differs greatly between mice and rats.
McDonald et al. (1994) used uniform 14C-benzene to determine total protein binding and C6-
13C-benzene to determine formation of adducts due to the reactions of benzene oxide, 1,4-
benzoquinone, or 1,2-benzoquinone. Total binding to protein and specific binding to cysteine
residues were estimated for blood Hgb and bone marrow total protein. Formation of adducts of
benzene oxide, 1,4-benzoquinone, and 1,2-benzoquinone accounted for 74% of the binding to
cysteine residues in rat Hgb but for only approximately 25% of the cysteine binding in mouse
Hgb or bone marrow proteins of either species. Thus, other benzene metabolites must also bind
to cysteine. Adducts of benzene oxide were highest in rat Hgb but accounted for only a small
proportion of adduct formation in mouse Hgb. Benzene oxide adducts accounted for less than
3% of total adduct formation in the bone marrow of either mice or rats, and benzoquinone adduct
formation accounted for a much greater proportion of total protein binding in bone marrow.
However, 1,2-benzoquinone adducts were more prevalent in rat bone marrow proteins and 1,4-
benzoquinone adducts were more prevalent in mouse bone marrow proteins. This suggests that
there are significant differences in benzene metabolism between mice and rats. A background
level of both 1,2- and 1,4-benzoquinone protein adducts that greatly exceeded the level of
17
-------
labeled adduct formation was also observed in bone marrow proteins of control mice and rats.
This result suggests that low doses of benzene may not cause toxicity through protein adduct
formation.
3.3.6.2. Genotoxicity
The association between benzene exposure and the appearance of structural and
numerical chromosomal aberrations in human lymphocytes suggests that benzene may be
considered as a human clastogen. In in vivo animal studies, benzene induced cytogenetic
effects, including chromosome and chromatid aberrations, sister chromatid exchanges, and
micronuclei (Anderson and Richardson, 1981; Au et al., 1991; Erexson et al., 1986; Fujie et al.,
1992; Kolachana et al., 1993; Siou et al., 1981; Toft et al., 1982; Ward et al., 1992). Several
lines of evidence also indicate that benzene is genotoxic in humans under occupational exposure
conditions (Ding et al., 1983; Sasiadek et al., 1989; Yardley-Jones et al., 1990; Major et al.,
1992; Eastmond, 1993; Tompa et al., 1994). However, these studies lacked good exposure
monitoring data, involved multiple chemical exposures, and were often poorly designed, with
inappropriate control groups.
Benzene has been shown to produce DNA breaks in Chinese hamster ovary cells
independent of metabolic activators; however, in vitro assays indicate that genotoxicity of
benzene is primarily due to its metabolites (Zhang et al., 1993; Eastmond et al., 1994). Benzene
is known to affect cell cycle progression, RNA and DNA synthesis, and DNA binding (Forni and
Moreo, 1967). Chen and Eastmond (1995) showed that benzene metabolites can adversely affect
human topoisomerases; however, some DNA repairs may occur in human cells (Chenna et al.,
1995).
Even with these studies, no data exist on the quantitative relationship between measured
benzene exposures and clastogenic effects.
Benzene itself is not mutagenic in short-term assays in either bacterial or animal systems
(Dean, 1985). Mutagenicity of benzene metabolites, however, is well established.
Hydroquinone, catechol, 1,2,4-trihydroxybenzene, and ^ram--l,2-dihydrodiol were active in
inducing elevated resistance to 6-thioguanine in Chinese hamster V79 cells. MUC is an active
mutagen in V79 cells, but it is only weakly active in Salmonella (Witz et al., 1990a).
3.3.6.3. Oxidative Stress
Benzene may also produce oxidative stress in the target tissues. The benzene metabolite
/>-benzoquinone is highly reactive and can deplete cellular levels of GSH (Brunmark and
Cadenas, 1988). Benzene metabolites can also be involved in redox cycling, resulting in the
production of reactive oxygen species that can also react with macromolecular components (Rao
and Snyder, 1995).
18
-------
3.3.6.4. Inhibition of Cytokine Formation
The stromal microenvironment of the bone marrow that normally modulates stem cell
proliferation and differentiation is a potential target for the hematotoxicity of benzene (Cox,
1991; Snyder et al., 1989; Kalf, 1987). The interaction of the stroma with the stem cells is
necessary for hematopoiesis. Furthermore, the stromal macrophage produces interleukin-1 (IL-
1), a cytokine also essential for hematopoiesis. Patients with aplastic anemia usually exhibit
monocyte dysfunction and decreased IL-1 production (Renz and Kalf, 1991).
Renz and Kalf (1991) demonstrated the disruption of IL-1 production by the stromal
macrophages of mice exposed to benzene. The mice were injected i.p. with 600 or 800
mg/kg/day benzene for 2 days. The stromal macrophages removed from these mice and cultured
with lipopolysaccharide produced the IL-1 precursor 34-kilodalton (Kd) pre-interleukin-1 alpha
(IL-1 a) but could not convert the precursor to the 17-Kd mature cytokine. Hydroquinone added
in vitro also inhibited the conversion of the pre-IL-lcc to the mature cytokine in mouse
macrophages. However, administration of recombinant mouse IL-1 a to mice before a bone-
marrow-suppressing dose of benzene ameliorated the bone marrow depression, probably by
circumventing the inability of the stromal macrophage in benzene-treated animals to process pre-
IL-lcc to the mature cytokine. Thus, Renz and Kalf (1991) suggested that benzene-induced
depression of bone marrow cellularity may result from the failure of the stromal macrophages to
process pre-IL-lcc to mature IL-1 a, which activates the stromal fibroblast production of the
colony-stimulating factor required for the differentiation of stem cells.
Niculescu et al. (1995, 1996) demonstrated that/>-benzoquinone, the oxidation product of
hydroquinone in the cell, causes a concentration-dependent inhibition of highly purified human
platelet calpain with an IC50 of 3 |j,M. Calpain is a calcium-activated, cysteine-dependent
protease that catalyzes the processing of pre-IL-lcc to the mature cytokine in vivo. The
investigators also demonstrated that/>-benzoquinone inhibits the processing of interleukin-1 P
(IL-1P), the product of a distinct second gene, to IL-1 by the sulfhydryl-dependent protease
referred to as IL-lp convertase.
3.4. ELIMINATION AND EXCRETION
3.4.1. Oral Exposure
The information base on the clearance of benzene from the body after oral exposure to
the compound is limited to reports of studies in experimental animals. For example, in one of a
series of toxicological reports on the metabolism and excretion of radiolabeled xenobiotics,
Parke and Williams (1953) recovered a substantial proportion of the administered dose on the
breath as unmetabolized product (43%). This amount compared with 33% urinary excretion in
which the label partitioned into such components as conjugated phenols (23% of the dose),
hydroquinone (4.8%), and catechol (2.2%), among others. Residual amounts of radioactivity
remained deposited in the tissues or were excreted in the feces.
In a study using male C57BL/6N mice given single oral doses of 14C-benzene (10 or 200
mg/kg), McMahon and Birnbaum (1991) reported the effects of age on benzene metabolism.
Radioactivity was monitored in urine, feces, and breath. Consistent with previous reports, the
19
-------
following urinary metabolites were detected: hydroquinone glucuronide, MA, phenyl
glucuronide, phenyl sulfate, catechol glucuronide, hydroquinone sulfate, and pre-
phenylmercapturic acid. At various time points up to 48 hours after dosing with 10 mg/kg, a
significant decrease in the urinary excretion of benzene-derived 14C was observed for 18- versus
3-month-old mice. When expressed as the percent of total benzene administered, the relative
amounts of some urinary metabolites varied between the 10 and 200 mg/kg dose groups, thereby
indicating dose-related quantitative changes in the urinary excretion of benzene metabolites. At
the low dose (10 mg/kg), urinary excretion was the major route of elimination. Hydroquinone
glucuronide, phenyl sulfate, and MA were the major metabolites at this dose, accounting for
40%, 28%, and 15% of the dose, respectively. At 200 mg/kg, urinary excretion decreased to
account for 42-47% of the administered dose, whereas respiratory excretion of volatile
components increased to 46-56% of the administered dose. Fecal elimination was minor and
relatively constant over both doses, accounting for 0.5-3% of the dose. Although age-related
differences in benzene disposition were observed, they could be attributed to alterations in
physiological function occurring with age; the significance of toxicity versus aging could not be
ascertained.
Fluctuation in the proportion of radiolabeled benzene excreted in urine was observed
with increasing dose by Sabourin et al. (1987), who exposed B6C3F1 mice, F344 rats, and
Sprague-Dawley rats via gavage to doses ranging from 0.5 to 300 mg/kg 14C-benzene. At the
lower doses (< 15 mg/kg), near quantitative amounts of label were recovered in various urinary
metabolites, whereas at the higher doses, unchanged 14C-benzene began to appear in expired air
in increasing proportions. For example, at doses of 150 mg/kg and above, greater than 50% of
the load was cleared on the breath as unchanged starting compound. These data point to the
saturation of the metabolic processes with increasing dose, although even at lower doses a
degree of fluctuation in the metabolic pattern was implied by dose-dependent changes in the
profile of metabolic products detected in the urine.
3.4.2. Inhalation Exposure
In a key report of studies that addressed the metabolic fate of inhaled benzene issue in
humans, Nomiyama and Nomiyama (1974) showed that at least a proportion of the absorbed
compound can be excreted in the urine as sulfate- or glucuronide-conjugated phenols or MAs.
At the benzene concentrations employed (52-62 ppm [166-198 mg/m3]), respiratory uptake was
considered to be in the region of 47% and respiratory excretion for the 4-hour exposure period
was approximately 17%. These values were in reasonably good agreement to those obtained in
an earlier study by Srbova et al. (1950), who showed that the respiratory excretion of retained
benzene could vary between 16.4 and 41.6% across a 7-hour exposure period.
In general, insufficient data exist to unequivocally assign one elimination route or
another as being of primary importance when human beings are exposed to benzene via
inhalation. For example, Sherwood (1988), using a complex experimental protocol, employed a
single human subject who was alternately exposed to either 6.4 ppm (20 mg/m3) benzene for 8
hours or 99 ppm (316 mg/m3) for 1 hour to monitor the kinetics of benzene elimination via the
various routes. Sherwood separated the release of benzene on the breath into several distinct
phases and was able to show that a greater proportion of the total dose was excreted in urine
rather than via expiration. These results also showed that urinary excretion of phenol conjugate
20
-------
was biphasic, with an initial rapid excretion phase followed by a slower excretion phase. The
importance of urinary excretion was also emphasized by the work of Inoue et al. (1986), who
showed a good correlation between urinary phenol levels and benzene exposure across a
relatively wide concentration range (1-200 ppm [3.2-640 mg/m3]).
Occupational exposure and cross-sectional studies have also pointed to the appearance of
benzene metabolites in urine as a consequence of exposure to such agents as sidestream tobacco
smoke (Bartczak et al., 1994) and gasoline vapors (Lagorio et al., 1994) or when the urine
content of smokers versus nonsmokers was compared (Kok and Ong, 1994; Melikian et al.,
1994). In general, one cigarette delivers an average of 55 jig of benzene, and an average smoker
takes in about 1.2-1.8 mg of benzene per day from smoking. Popp et al. (1994) detected
muconic acid and S-phenyl-N-acetyl cysteine levels in the urine of car mechanics at levels that
correlated with levels of the compound in the bloodstream and the breathing zone. These
observations support the earlier suggestion by Ghittori et al. (1993) that benzene and its
metabolites in the urine may be an important biomarker of occupational exposure to the
compound.
An accumulation of experimental data in laboratory animals has shown an essentially
similar pattern of benzene elimination and excretion as in human beings. In broad terms, this is
characterized by the release of unchanged compound on the breath and the appearance of
metabolites in the urine. For example, at high incidental concentrations of benzene (500 ppm
[1597 mg/m3] for 6 hours), an initial rapid phase of elimination was followed by a slower phase
of much longer duration (up to 13 hours or more) (Rickert et al., 1979).
Perhaps the most rigorous experiments that have examined the dose-response effects of
benzene via inhalation were those carried out by Sabourin et al. (1987) in parallel to those
already described for the oral route. The experimental protocol featured a 6-hour, nose-only
exposure to concentrations of 14C-benzene ranging from 10 to 1000 ppm (32 to 3195 mg/m3). At
the lower concentrations, less than 6% of the radioactivity was expired on the breath. However,
at the higher concentrations of benzene (> 850 ppm [2718 mg/m3]), both rats and mice exhaled
considerable proportions of the unchanged compound, amounting to 48% of the load in rats and
14% in mice. In a similar experiment, Dow Chemical Co. (1992a) reported that Duroc-Jersey
pigs exposed to 0, 20, 100, or 500 ppm (0, 64, 319, or 1597 mg/m3) benzene for 6 hours/day, 5
days/week for 3 weeks excreted levels of phenol in the urine that increased linearly according to
dose.
3.4.3. Dermal Exposure
Only limited data on excretion of benzene after dermal exposure in humans were found.
Franz (1984) reported that four volunteers exposed to 14C-benzene (0.0024 mg/cm2) on the skin
excreted trace amounts of label in the urine over a 36-hour period, suggesting the ability of the
compound to penetrate the dermal barrier. Similar results were obtained in monkeys and
minipigs, with excretion of the compound in monkeys approaching 0.065% (range
0.033-0.135%) of the applied dose, compared with 0.042% (range 0.030-0.054%) in minipigs
(Franz, 1984).
3.4.4. Other Routes of Exposure
21
-------
This section discusses very limited data available from three studies in which benzene
was either given intraperitoneally or subcutaneously. In these studies, the exposure was limited
to either a single injection or daily injections for up to 3 days.
Experiments by Cornish and Ryan (1965) used the i.p. route of administration to examine
the influence of the fed or fasted state on the rates of production of urinary metabolites such as
conjugated or unconjugated phenols, components that appeared to be increased markedly in the
fasted state versus fed. In line with their findings of the dose-response effects of benzene when
administered via gavage or inhalation, Sabourin et al. (1987) found that the proportion of
radioactivity (mainly as phenyl sulfate) excreted in the urine decreased with increasing dose
when monkeys were injected intraperitoneally with concentrations of 14C-benzene ranging from
5 to 500 mg/kg, again suggesting that benzene metabolism can be saturated at the higher dose
levels. Finally, Longacre et al. (1981) exposed C57BL/6N and DBA/2 mice to benzene
subcutaneously and monitored the appearance of conjugated phenol, catechol, and hydroquinone
in the urine, although with varying proportions when one strain was compared with another.
Amounts of phenyl mercapturic acid were similar, however.
3.5. PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS
PBPK models can be used to address the uncertainty of extrapolating from data of an
experimental study to a hypothetical administered dose for the same response in other animals or
humans. The approach seeks to predict the dose-response characteristics of a chemical's
potential toxicity through an understanding of the underlying mechanisms by which a substance
is absorbed, transported to, and metabolized in the primary target organs, on the assumption that
the biochemical processes involved in the toxicological response will be the same for both
humans and the animal model(s) in which the harmful effects of the compound had been
demonstrated. Necessary inputs to the model include a physiologically realistic discrimination
of a chemical's movement between metabolically and functionally linked compartments of the
body and the information necessary to describe this movement mathematically in terms of (1) the
kinetics of metabolism, (2) rates of movement to and from different target organ groupings, and
(3) the partitioning of a compound between physiological media. Thus, by modeling a target
organ-specific internal dose surrogate (parent compound or metabolite, as applicable) from the
NOAEL or LOAEL of an experimental study, an equivalent allometrically scaled internal dose
surrogate in human beings can be used to back-extrapolate to a hypothetical "effective" dose for
the same toxicological response in human beings.
Potential advantages of the PBPK approach reside in its ability to quantitatively address
interspecies differences and to take into account the nonlinearity of biological processes when
extrapolating outside the range of available experimental data. Limitations reside in the
accuracy/validity of the estimates that constitute the inputs to the model and in the
oversimplification that is inherent in seeking to define the movement and metabolism of a
xenobiotic or its metabolites in terms of a small number of functionally defined
subcompartments.
Several attempts have been made to develop PBPK models to define the understanding of
interactions between benzene metabolism and toxicity. The first model for benzene was
developed by Sato (Sato and Nakajima, 1979; Sato, 1988), who exposed three men to 25 and 100
22
-------
ppm (80 and 319 mg/m3) benzene vapor for 2 hours and then observed a triexponential decay of
benzene from their blood. The investigators constructed a three-compartment model consisting
of richly perfused tissues, poorly perfused tissues, and fat, which acted as a major sink for
benzene. Subsequently, PBPK models have been developed to take into account differences in
benzene metabolism between species and individuals using both experimental data and
simulations (Medinsky et al., 1989a; Travis et al., 1990a, b; Bois et al., 1991b; Cox, 1991, 1996).
A model (Medinsky et al., 1989a, b, c) was developed to describe and predict the fate of
benzene and to determine whether differences in the metabolic rates between rats and mice could
explain the differences in toxicity between these species. The model indicated that for inhalation
concentrations up to 1000 ppm (3195 mg/m3), mice metabolize at least two to three times as
much benzene as rats. However, following oral exposure, rats metabolized more benzene on a
body-weight basis than did mice at doses greater than 50 mg/kg. Patterns of metabolites also
differed between rats and mice. Mice produced primarily hydroquinone glucuronide and MA
metabolites linked to toxic effects; on the other hand, rats produced primarily phenyl sulfate, a
detoxification product. These simulated results agree with experimental data and provide a
framework for understanding the greater sensitivity of the mouse to benzene toxicity.
The Medinsky model was based on an earlier PBPK model developed by Ramsey and
Anderson (1984). The tissue compartments initially included in the model were (1) the liver,
presumed to be the only organ where benzene metabolism takes place; (2) a group of poorly
perfused tissues, including muscle and skin; (3) a group of richly perfused tissues, including
bone marrow, kidney, and intestines; and (4) a fat compartment. All metabolism of benzene,
consisting of initial metabolism to benzene oxide that is then further metabolized by one of four
pathways, was modeled by Michaelis-Menten kinetic parameters. Metabolic rate constants were
determined by fitting the results of model simulations to experimental data obtained by exposing
mice and rats to benzene orally and by inhalation (Medinsky et al., 1989b; Sabourin et al., 1987).
However, Bois et al. (1991b) found that the Medinsky model did not simulate the data of Rickert
et al. (1979) very well, thereby indicating a need for the model to be further refined.
Since the original Medinsky model was published, additional compartments have been
added to reflect an advancing understanding of benzene metabolism, and specific biochemical
and toxicokinetic parameters have been refined to reflect age, sex, and species-specific
differences (Schlosser et al., 1993; Seaton et al., 1994; McMahon et al., 1994; Kenyon et al.,
1995). Seaton et al. (1994) measured a 13-fold variability in CYP2E1 activity in human hepatic
microsomes and compared this to the activity in mouse and rat liver microsomes. The model
predicted the dependence of benzene metabolism on the measured CYP2E1 activity, and the
proportion of hydroquinone (the suspected toxic metabolite) produced in vitro was correlated
with the level of CYP2E1 activity. Seaton et al. (1995) measured the initial rates of the two
major conjugation reactions, phenol sulfonation and hydroquinone glucuronidation, in the
hepatic microsome preparations of humans, rats, and mice. This information was used in a
physiological compartment model to predict steady-state concentrations of phenol and
hydroquinone in blood.
Among humans, predicted steady-state concentration of phenol varied sixfold (0.38-2.17
nM), and predicted hydroquinone concentrations varied fivefold (6.66-31.44 nM). Predicted
steady-state concentrations of phenol were higher in mice than in rats (2.28 vs. 0.83 nM), and
23
-------
predicted hydroquinone concentrations were also higher in mice than in rats (42.44 vs. 17.99
nM). The predicted concentrations for mice were higher than the range for humans, but the rat
values were within the predicted concentrations for humans. On this basis, the authors suggested
that the rat may be a good model for humans with respect to tissue dosimetry for these benzene
metabolites. The authors also suggested that the mouse might be more sensitive than the human
and that in vitro metabolism data must always be placed within the context of the whole animal's
physiology.
The Medinsky PBPK model has served to organize the available information into a
coherent model that has helped to refine the specific experimental approaches used to fill the
gaps in the understanding of the mechanism of benzene toxicity. Although current PBPK
models may provide insights about putative toxic metabolites and potential biochemical
mechanisms, they are insufficiently developed to be able to reduce scientific uncertainty
(Medinsky et al., 1995, 1996; Medinsky, 1995).
Travis et al. (1990a, b) also developed a model to describe the pharmacokinetics of
benzene in rats, mice, and humans. The model contains five compartments, consisting of liver,
fat, bone marrow, muscle, and organs such as brain, heart, kidney, and viscera. The different
compartments are connected by the arterial and venous blood pathways. Metabolism of benzene
is assumed to follow Michaelis-Menten kinetics in all species and is assumed to occur primarily
in the liver and to a lesser extent in the bone marrow. Model simulations were compared with
experimental data from Sabourin et al. (1987, 1988a), Andrews et al. (1977, 1979), Nomiyama
and Nomiyama (1974), Snyder et al. (1981), Sato et al. (1975), and Rickert et al. (1979). The
Travis model successfully simulated uptake, metabolism, and excretion of benzene for mice,
rats, and humans using experimental data from the studies that were used to develop the model.
However, the model is of limited value because it does not predict the kinetics of benzene
metabolites (Bois et al., 1991b). In addition, the concentration of benzene in fat is poorly
predicted.
The model developed by Bois and Paxman (1992) provided evidence that exposure rate
had a strong influence on the rate of formation of several important metabolites of benzene. This
model has three components describing the pharmacokinetics of benzene and the formation of
metabolites in the rat. Distribution and elimination of benzene from a five-compartment system
composed of liver, bone marrow, fat, poorly perfused tissues, and well-perfused tissues make up
the first component of the model. The bone marrow compartment is included for its relevance to
human leukemia. Parameter values for this component were derived from the literature and from
the previously published work of Rickert et al. (1979) and Sabourin et al. (1987). The second
component describes the metabolic transformations of benzene and its by-products in the liver
and bone marrow. The reactions are assumed to follow Michaelis-Menten kinetics, except for
the transformation of benzene oxide into phenol, which occurs spontaneously and may be
described as a first-order reaction. The third component is the distribution of phenol. In addition
to the compartments described for benzene, phenol is also assumed to distribute to the lung and
gastrointestinal tract.
The model was validated against the data of Cassidy and Houston (1984), Sabourin et al.
(1987, 1988a, 1989), and Sawahata and Neal (1983). It was also used to predict metabolite
production for male rats exposed to benzene (nose only) at three different concentrations and for
24
-------
three different exposure durations in comparison with the experimental data of Sabourin et al.
(1989). The three exposure regimens were established to maintain a constant concentration/time
product. Simulation results indicated that the model may over- or underestimate the level of
urinary metabolites.
More recent efforts on development of the Bois and Paxman model have focused on
defining the physiological pharmacokinetic parameter distributions needed to develop models
useful in risk assessment (Spear and Bois 1994; Spear et al., 1991; Watanabe and Bois, 1996;
Bois et al., 1991a, 1996). Spear and Bois (1994) described the outcome of their modeling efforts
to explain the basis for the paradoxical observation that although phenol is a major initial
metabolite of benzene, a known carcinogen, a National Cancer Institute chronic study (NCI,
1980) did not demonstrate carcinogenic activity for phenol. The approach selected was to apply
Monte Carlo methods using parameter distributions coupled with a pass-fail fit criterion. The
advantage of this approach is that it acknowledges that in most biological applications, there is
no clear way to select a "best" set of fixed parameters. On the basis of the researchers' modeling
effort, hydroquinone was rejected as the ultimate toxic agent, and the pathway through benzene
glycol to catechol and MUC appeared to provide a better fit to the data.
Kenyon et al. (1995) investigated the metabolism of phenol. Even though phenol is
thought to be a key intermediate in benzene metabolism leading to toxicity, orally administered
phenol is neither carcinogenic or genotoxic (NCI, 1980). The authors found markedly higher
excretion of hydroquinone glucuronide after oral benzene exposure as compared with phenol.
Also, phenol sulfate and phenol glucuronide excretion was much lower following benzene
exposure than following phenol exposure. This could be explained by differences in the zonal
distribution of CYP2E1 and detoxification enzymes in the liver. Phenol initially entering the
liver had a relatively greater chance for conjugation (sulfonation or glucuronidation) in
periportal hepatocytes of zone 1 than of oxidation by CYP2E1 located in pericentral hepatocytes
in zone 3. Benzene, on the other hand, was more likely to pass through to zone 3 and be
oxidized to phenol.
Watanabe and Bois (1996) examined three methods (multiplicative, additive, and
allometric) to extrapolate physiological parameter distributions across species, specifically from
rats to humans. The results indicated that the multiplicative and allometric techniques were able
to extrapolate the model parameter distributions.
Bois et al. (1996) applied techniques from population pharmacokinetics, Bayesian
statistical inference, and physiological modeling to model distribution and metabolism in
humans. Statistical distributions for the parameters of a physiological model of benzene were
derived on the basis of existing data. The relationship between the fraction of benzene
metabolized in bone marrow and benzene exposure was linear up to 10 ppm (32 mg/m3). The
median population estimate of the fraction metabolized in bone marrow was 52% (90%
confidence interval [CI] 47-67%). At levels approaching occupational inhalation exposure
(continuous 1 ppm [3.2 mg/m3]), the estimated amount metabolized in bone marrow ranged from
2 to 40 mg/day. However, this model has not been tested for its ability to predict data from other
studies (Smith and Fanning, 1997).
25
-------
In summary, PBPK models continue to improve as additional data become available and
are incorporated and as additional techniques from other scientific fields are applied to modeling
benzene dosimetry. However, the current published models are insufficiently refined to allow
them to predict human metabolism accurately (Smith and Fanning, 1997). The key areas for
refinements appear to be the inclusion of the kinetics of the putative toxic metabolites of benzene
or their stable precursors. If benzene metabolites such as hydroquinone/ benzoquinone, MUC,
and/or benzene oxide are the toxic species, then PBPK models need to include descriptions of
their kinetics if they are to be useful in improving uncertainties in risk assessment.
3.6. TOXICOKINETICS SUMMARY
Benzene is almost completely absorbed from the gastrointestinal tract over a wide range
of dose levels in rats and mice. In contrast, several human studies have indicated that the
respiratory absorption of benzene is approximately 50%. At a concentration of 10 ppm (31.9
mg/m3), retention of benzene during a 6-hour exposure was 33% in rats and 50% in mice.
Retention decreased at higher concentrations (Sabourin et al., 1987). At high concentrations,
benzene metabolism is saturated, and respiratory excretion of unmetabolized benzene increases.
Dermal absorption is less than 1% of the applied dose due to rapid volatilization from
nonoccluded skin.
Benzene is rapidly distributed throughout the body regardless of the exposure route.
Following oral exposure, Low et al. (1989) found the highest concentrations in liver and kidneys,
intermediate concentrations in blood, and the lowest concentrations in the Zymbal gland, nasal
cavity, oral cavity, mammary gland, and bone marrow. Benzene is preferentially stored in fat
(Rickert et al., 1979). Metabolites of benzene are also found throughout the body, with phenol
being eliminated from bone marrow more rapidly than catechol or hydroquinone (Rickert et al.,
1979).
Despite extensive research, the metabolism of benzene to toxic metabolites is still not
completely understood. Metabolism of benzene by CYP2E1 is necessary for the expression of
hematotoxicity (Sammett et al., 1979; Valentine et al., 1996). Following initial oxidation by
CYP2E1 to benzene oxide, however, the benzene metabolic pathway branches to produce
several putative toxic metabolites. Most research has focused on the hypothesis that phenol,
catechol, and hydroquinone are produced in the liver and transported to the bone marrow, where
hydroquinone is activated to/>-benzoquinone by the action of MPO (Smith et al., 1989;
Rushmore et al., 1984; Low et al., 1995). However, the fact that administration of phenol does
not produce the same effects as benzene is a problem in the phenolic metabolite hypothesis. A
combination of phenol and hydroquinone is required to reproduce the hematotoxic effects of
benzene (Eastmond et al., 1987).
A ring-opening pathway of benzene metabolism leads to excretion of MA in the urine.
The production of the highly reactive ring-opened dialdehyde MUC as an intermediate in the
pathway to MA has been demonstrated in isolated hepatic microsomes; however, circulation of
MUC in the blood has not been demonstrated (Witz et al., 1996). Administration of MUC
reproduces the hematotoxic effects in mice; however, MUC is so reactive that it is not likely to
reach the target tissues. The less reactive ring-opened metabolite CHO-M-OH has been detected
in rat liver perfusate (Grotz et al., 1994) and has been shown to be hematotoxic in mice (Zhang
26
-------
et al., 1995b). Thus, CHO-M-OH is another potentially toxic benzene metabolite, but there has
been little research to define whether CHO-M-OH is quantitatively important.
Benzene oxide was believed to be too reactive to escape the liver, but it has recently been
measured in the bloodstream of mice, and its dose to the body may exceed that ofp-
benzoquinone (Lovern et al., 1997; Lindstrom et al., 1997). Thus, several metabolites of
benzene may play a role in inducing the toxic effects of benzene, and a combination of several
metabolites may be required to cause the full range of benzene-induced toxic responses. Large
quantitative differences in the production of putative toxic benzene metabolites have been
observed among different animal species (Sabourin et al., 1989, 1992). Because few data exist
on the proportion of different benzene metabolites produced in humans, there is considerable
uncertainty in selecting the most appropriate animal model for humans.
There is evidence to support several different mechanisms by which benzene metabolites
may cause hematotoxicity. Benzene metabolites form adducts with both proteins and DNA
(McDonald et al., 1994; Lutz and Schlatter, 1977). Adduct formation with protein sulfhydryl
groups inhibits tubulin polymerization during spindle formation and the activity of
topoisomerase II in DNA replication and transcription (Irons and Neptun, 1980; Pfeifer and
Irons, 1983; Frantz et al., 1996). These effects may explain the clastogenic activity of benzene.
Benzene itself is not mutagenic; however, mutagenicity of several of its metabolites is well
established (Dean, 1985). Benzene metabolites may also induce oxidative stress by depleting
levels of GSH and by production of reactive oxygen species that react with cellular
macromolecules (Snyder et al. 1987; Brunmark and Cadenas, 1988; Rao and Snyder, 1995).
Benzene also disrupts production of the cytokine IL-1, which is essential for hematopoiesis.
Niculescu et al. (1995, 1996) found that/7-benzoquinone inhibits the activity of calpain, an
enzyme that catalyzes the processing of pre-IL-lcc to the mature cytokine. All of these effects
may be involved in causing the toxic effects of benzene, but it is not possible to determine which
of these effects is the primary mechanism for benzene toxicity.
At low exposure levels, benzene is excreted primarily in the urine as sulfate or
glucuronide conjugatges of phenolic metabolites or as MA (McMahon and Birnbaum, 1991;
Sabourin et al., 1987). At oral doses of less than 15 mg/kg, 14C from benzene is nearly
quantitatively recovered as metabolites in urine in rats and mice, but at a dose of 150 mg/kg,
greater than 50% was cleared unmetabolized in the breath (Sabourin et al., 1987). Similarly, less
than 6% of 14C from benzene was exhaled in the breath from inhalation exposure to 10 ppm (32
mg/m3) in rats or mice, but at concentrations greater than 850 ppm (2718 mg/m3) rats exhaled
48% and mice exhaled 14%. Urinary phenol and MA concentrations are correlated with
benzene exposure level and can be used to monitor benzene occupational exposure (Ghittori et
al., 1993). However, the database for evaluating the relative importance of different excretion
pathways in humans is limited. Thus, there is a limited database for comparing the importance
of human excretion pathways at varying dose levels with the results of experimental animal
studies.
Several different PBPK models have been developed to mathematically describe the
uptake, distribution, metabolism, and excretion of benzene (Sato, 1988; Medinsky et al., 1989a,
b, c; Travis et al., 1990a, b; Bois and Paxman, 1992). Each of these PBPK models has
advantages, and some successfully simulate uptake, metabolism, and excretion in mice, rats, and
humans. However, the utility of PBPK modeling to predict the dose of toxic metabolites to
27
-------
target organs in humans is limited by the incomplete knowledge of the toxic metabolites and
difficulties in identifying a suitable experimental animal model for humans.
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS
4.1.1. Oral Exposure
Individual case reports of death from acute oral exposure to benzene have appeared in the
literature since the early 1900's. The benzene concentrations ingested by the victims often were
not known. However, lethal oral doses for humans have been estimated at approximately 125
mg/kg (Thienes and Haley, 1972). Accidental ingestion and attempted suicide with lethal oral
doses of benzene have produced the following signs and symptoms: staggering gait, vomiting,
shallow and rapid pulse, somnolence, and loss of consciousness followed by delirium,
pneumonitis, collapse, and then central nervous system depression, coma, and death (Thienes
and Haley, 1972). Lethality in humans has been attributed to respiratory arrest, central nervous
system depression, or cardiac collapse (Greenburg, 1926). Ingestion of lethal doses may also
result in visual disturbances or feelings of excitement and euphoria, which may suddenly change
to weariness, fatigue, sleepiness, convulsion, coma, and death (Von Oettingen, 1940).
4.1.2. Inhalation Exposure
4.1.2.1. Hematotoxicity
The recognition of benzene as etiologically significant in the development of aplastic
anemia, a potentially life-threatening suppression of bone marrow activities, and pancytopenia, a
reduction of the cellular elements of the peripheral blood, dates back to the 19th century
(Goldstein, 1988). Exposure to benzene has also been associated with acute myelogenous
leukemia (AML) and lymphoid malignancies in humans (Goldstein, 1988; Snyder, 1987),
although the precise relationship between frank pancytopenia and the benzene-induced
malignancies remains unclear (Snyder, 1987). That the hematopoietic abnormalities observed in
humans exposed to benzene in an occupational setting have been reproduced in numerous animal
experiments points to their universality. The comparatively low doses associated with their
onset suggest that they may represent a sensitive effect of benzene toxicity.
Human exposure to benzene occurs primarily via inhalation in the workplace, from
gasoline vapors, tobacco smoke, and automotive emissions. Individuals exposed to benzene
exhibit bone marrow depression, as evidenced by anemia (decreased RBC count), leukopenia
(decreased WBC count), and/or thrombocytopenia (decreased platelet count). A depression of
all three elements is called pancytopenia, and the simultaneous depression of RBCs, WBCs, and
platelets, accompanied by necrosis of the bone marrow, is diagnostic of aplastic anemia.
Patients with aplastic anemia also have exhibited mild bilirubinemia, changes in osmotic
fragility of erythrocytes, shortened erythrocyte survival time, increased fecal urobilinogen, and
mild reticulocytosis (Aksoy, 1991). The bone marrow profile may vary from aplasia to
hyperplasia, and these symptoms may vary in frequency and severity in different patients. Other
hematologic changes observed in humans chronically exposed to benzene include decreased
28
-------
osmotic resistance in leukocytes, decreased phagocytic function of neutrophils, reduced
glycogen content, and decreased activity of peroxidase of neutrophils and increased delta-
aminolevulinic acid activity in erythrocytes and increased coproporphyrins in the urine (Aksoy,
1991).
As summarized in Table 1, the large number of epidemiologic studies that have explored
the health consequences of benzene exposure in the workplace provide a considerable body of
inferential evidence for the compound-related onset of hematopoietic effects in chronically
exposed subjects. It should be recognized that for most of the studies there are uncertainties
and/or variabilities in precisely delineating the concentrations of benzene to which subjects were
exposed. For example, uncertainty surrounding the past or present workplace concentrations of
the chemical may not permit the dose to which subjects were exposed to be estimated accurately.
Even when an exposure term can be defined within an adequately cohesive range, its overall
contribution to the understanding of the compound's dose-response characteristics will probably
be limited to a single data point from any given study. Thus, the utility of such a study might
only arise through its consideration with others as a group, for example, in a "meta-analysis." In
such a survey, the establishment of a combined range for estimated exposures and their
toxicological consequences can permit (1) the emergence of a unifying concept of the
compound's toxicity to humans and (2) the delineation of a range of concentrations or doses
within which the compound's toxicological threshold may reside.
The paragraphs that follow describe the many epidemiologic studies that have helped to
establish the onset of benzene's hematologic effects.
In general, it is assumed that for the occupational exposure studies under consideration,
the levels of exposure to benzene were greater in earlier studies than in more recent studies, a
reflection of times when occupational exposure standards and guidelines were less rigorous. For
example, studies that have examined occupational exposure to benzene at relatively high
concentrations include several that evaluated hematologic surveillance records of 459 rubber
hydrochloride workers from the Pliofilm production departments of a rubber products
29
-------
Table 1. Hematotoxicity of benzene—occupational exposure
Population/Industry
Rotogravure printers
Rotogravure printers,
New York
Shoe factory, Finland
Shoe factory, Finland
Rubber factory, Ohio
Watch industry
Rubber-coating
industry
Shoe manufacturing,
Turkey
Controls (healthy
hospital workers)
Various industries
using benzene-
containing adhesives,
Turkey
Various industries
using benzene-
containing adhesives,
Turkey
No. of
Subjects
6
332
147
147
1104
216
217
100
32
diagnosed
with blood
dyscrasias
44
diagnosed
with blood
dyscrasias
Exposure
77-3386 mg/m3 benzene
35-3386 mg/m3 benzene
< 1278 mg/m3 benzene for -10 yr
Followup of Savilahti (1956) study, 9
years later
<1597 mg/m3 (mean, -319 mg/m3)
benzene
No exposure data; workers followed for
10 years after cessation of exposure to
benzene
< 80-399 mg/m3 benzene (prior to
installation of control measures)
96-671 mg/m3 benzene for 3 months-17
years
479-1597 mg/m3 benzene for 4
months-15 years
Follow-up study over 2-17 years after last
exposure to 479-2077 mg/m3 benzene for
4 months-15 years
Effects
Pancytopenia
23 cases of significant cytopenias
107 cases of hemato logic abnormalities
Persistent cytopenias; one death from acute
leukemia
83 cases of mild hematologic abnormalities, 25
cases of more severe pancytopenia (9 required
hospitalization, 3 of whom died)
Four cases of persistently decreased blood counts;
one death from aplastic anemia 9 years after
cessation of exposure
Follow-up suggested mild persistent anemia
51 cases of hematologic abnormalities, including 6
cases of pancytopenia
32 cases of significant aplastic anemia; eight deaths
from thrombocytopenic hemorrhage and infection
Complete remission in 23 (52%), death due to
complications of pancytopenia in 14 (32%),
leukemia in 6 (14%), myeloid metaplasia in 1 (2%)
Reference
Erf and Rhoads, 1939
Goldwater, 1941; Goldwater
and Tewksbury, 1941;
Greenburgetal., 1939
Savilahti, 1956
Hernberg et al, 1966
Wilson, 1942
Guberan and Kocher, 1971
Pagnotto et al., 1961
Aksoy etal., 1971
Aksoy etal., 1972
Aksoy andErdem, 1978
-------
Table 1. Hematotoxicity of benzene—occupational exposure (continued)
Population/Industry
No. of
Subjects
Exposure
Effects
Reference
Turkish, various
231
Exposure to solvents, thinners, and
similar materials in 40 small industries; in
one facility, benzene concentrations
averaged 351 mg/m3; in the remainder,
benzene levels < 1 ppm
351 mg/m3: two cases of acute myeloblastic
leukemia in facility where levels measured 351
mg/m3
< 1 ppm: mild abnormalities in 14 workers (6.1%);
differential diagnosis of 186 workers, monocytosis
in 8, eosinophilia in 5, and basophilia in 3; one case
each of acute myeloblastic leukemia, Hodgkin's
disease, and lymphoma
Aksoy etal., 1987
Chemical production
facility
282 men
Jobs assigned to exposure categories:
(!)<6mg/m3TWA
(2) 6-29 mg/m3 TWA
(3) 32-77 mg/m3 TWA
(4) 80- -96 ppm TWA
exposures ranged from < 1 years to > 20
years
Slight statistically significant (p<0.05) (but not
biologically significant) decreases in bilirubin and
RBC counts; no correlations between peripheral
blood counts and latency, duration, or intensity of
benzene exposure
Townsend et al., 1978
Refinery
303 men
Exposure to benzene levels of 1.70 mg/m3
(mean) for an average of 8 years for white
employees and 4.5 years for nonwhite
employees); average length of followup,
13 years
No significant changes in hematology of workers as
a group; one case of multiple myeloma; one death
from multiple myeloma and one from malignant
melanoma
Tsaietal., 1983
Chinese shoemaking
and printing
300
Mean TWA exposures:
Benzene—105 and 188 mg/m3 (55 and 65
months) for men and women,
respectively;
Toluene—173 and 154 mg/m3 (95 and 73
months);
Mixture—45 mg/m3 benzene + 68 mg/m3
toluene and 45 mg/m3 benzene + 79
mg/m3 toluene (159 and 120 months)
Changes in absolute and/or relative peripheral cell
counts (statistically significant at p<0. 01 or
Yinetal, 1987b
Benzene — decreased lymphocyte counts (women;
men and women, combined); eosinophilia (women;
men and women combined); decreased monocytes
(women)
Toluene — decreased lymphocyte counts (women;
men and women, combined); eosinophilia (all
groups)
Mixture — decreased lymphocyte counts (men;
women; men and women combined); significant
eosinophilia (men and women combined) _
-------
Table 1. Hematotoxicity of benzene—occupational exposure (continued)
Population/Industry
No. of
Subjects
Exposure
Effects
Reference
Rubber industry
459
Mean estimated exposure from 1948 to
1975, 48-64 mg/m3
Significant increases in WBCs, RBCs, and Hgb
during the 1940's (Pearson correlations: r = 0.50
for WBCs, r = 0.44 for RBCs, and r = 0.71 for
Hgb), but not during the ensuing 25 years; rapid
decline in exposure for the workers consistent with
increases in blood count values
Kipenetal., 1988, 1989
Workers
Unexposed controls
200
268
0.03^.5 mg/m3 8-hour TWA over a 10-
year period
No abnormal hematology
Collins etal, 1991
Workers in various
Chinese industries
Low-exposure group
Below median
Above median
Unexposed controls
11
22
22
44
Workers exposed to benzene, but not to
other solvents, at TWA of 24 (low
exposure), 44 (below median), or 294
(above median) mg/m3 for an average of
6.3 years; doses were measured by
personal dosimetry for 2 weeks
Significant increase in mean corpuscular volume
and significant decreases in ALC, WBCs, RBCs,
HCT, and platelets in above-median group;
significant reductions in ALC, RBCs, and platelet
count in below-median group; but only reduction in
ALC was significant in low-exposure group
Rothmanetal., 1996a
to
Rubber industry
657
Estimates of benzene exposure (<109
mg/m3) for last 30, 90, or 180 days before
blood tests were correlated with
hematologic data
Weak correlation between RBC and benzene
exposure, but strong correlation between low WBC
counts and benzene exposure, with no evidence of a
threshold
Ward etal., 1996
-------
manufacturer in Ohio (e.g., Kipen et al., 1988, 1989; Ward et al., 1996). The data indicate that
during the time period 1940-1948, significant increases in erythrocyte, Hgb, and leukocyte
levels were observed among the workers, with the responses displaying a positive time-related
trend (1940-1948). Thus, overall cell numbers and values appear to have been higher at the later
time points, the prevailing levels of exposure to benzene becoming reduced as industrial hygiene
practices improved. Kipen et al. (1988, 1989) compared the utility of two exposure assessment
models for correlating the observed hematologic consequences with plausible estimates of
benzene exposure in the absence of definitive monitoring data at the earlier time points. Thus,
between 1948 and 1975, when the workers were exposed to mean 8-hour TWA concentrations of
only 15-20 ppm (48-64 mg/m3), no significant relationships between benzene levels and any
hematologic parameters were observed. By contrast, significantly lower cell counts and
hemoglobin concentrations were observed in exposed subjects, where the levels of benzene
exposure were thought to be as high as 75 ppm (240 mg/m3) at the earlier time points (pre-1948)
(Kipen etal., 1989).
Cody et al. (1993) observed significant hematologic effects, including reduced RBC and
WBC counts in 161 male rubber workers exposed to median peak concentrations (i.e., only the
peak concentrations for any given exposure time were reported) of 30-54 ppm or more for a 12-
month period during 1948. The 30 ppm value was considered a 1-year LOAEL for hematologic
effects. In this rubber plant, workers who had blood dyscrasias were excluded from working in
the high-benzene units. Furthermore, individual workers who had more than a 25% decrease in
WBC counts from their pre-employment background count were removed from the high-benzene
units and placed in other units with lower benzene concentrations. Sensitive individuals
therefore could have been excluded from the analysis. The 30 ppm value is the low end of the
range of median values (30-54 ppm) reported by Crump and Allen (1984) and used in the Kipen
et al. (1988) and Cody et al. (1993) studies. These are the same Pliofilm workers studied by
Rinsky etal (1981, 1987).
The analysis of Ward et al. (1996) differs from the earlier investigations of the same
rubber worker cohort by Kipen et al. (1988, 1989) and Cody et al. (1993) in that a nested case-
control design was used. Incident cases were defined as the first occurrence of a low WBC or
RBC count, and matched controls were chosen from those tested within 6 months of the case's
blood test date. In contrast to the earlier analyses of hematologic screening data from the same
plant, the Ward et al. (1996) study used the entire data set, evaluated the exposure-response
relationship on the basis of individual dose metrics, and controlled for the temporal trends in pre-
employment blood count screening.
The Ward et al. (1996) analysis used hematologic screening data for 657 of 1037
individuals employed at the plant from 1939 through 1976. There was a total of 21,710 blood
test records. The study used a case-control design and estimated benzene exposures using the
job exposure matrix developed by Rinsky et al. (1981). The effects of benzene exposure in the
30, 90, and 180 days before the blood test date, as well as cumulative exposure until the blood
test date, were examined using conditional logistic regression. For WBC count there was a
strong exposure response, and all of the exposure metrics selected showed a significant
relationship with low blood count. For RBC count there was a weak positive exposure response
that was
33
-------
significant (p<0.03) for one of the dose metrics. The maximum daily benzene exposure estimate
in this study was 34 ppm (109 mg/m3). There was no evidence for a threshold for the
hematologic effects of benzene exposure, suggesting that even exposure to relatively low levels
(e.g., < 5 ppm [16 mg/m3]) could result in hematologic suppression. The results of this study are
consistent with test animal studies that have demonstrated a decrease in peripheral lymphocyte
counts at benzene exposures as low as 10 ppm (32 mg/m3) and a stronger effect of benzene
exposure on lymphocytes than on red cells; therefore, if one assumes that Ward's Pliofilm
workers exposure estimates are correct, they suggest that hematologic effects of benzene occur
even at 5 ppm. Because of controversy surrounding the exposure estimates used by Ward et al.,
this finding could be biased.
The nature of the problem is that the estimates of early exposure to benzene (prior to
1946) are not based on any measured ambient air inhalation data for any of the more than 1200
Pliofilm workers who begn work in the plants prior to 1946. Estimates of exposure to benzene
in the set of exposure indices chosen by Ward and her associates (1996) were based on the
exposure scenario described by Rinsky et al. (1981), which basically assumed that the exposures
were the same as the local and State standards established for benzene at the time (in the years
prior to 1946). This assumption has been challenged by others (Crump, 1992; Paustenbach et
al., 1992), who claim that the exposures during this time were much higher. Thus, the
assumption that the exposures could have been as low as < 5 ppm may not be correct. The
weaknesses in the Ward et al. study are discussed further in U.S. EPA (1998a). No LOAEL or
NOAEL could be established for this study.
In another occupational study that examined the health effects of benzene, Fishbeck et al.
(1978) studied the hematologic characteristics of 10 Dow Chemical employees who had been
exposed to high concentrations of the chemical in an ethyl cellulose sheeting operation for
various durations between 1937 and 1965. Duration of exposure was from fewer than 4 to nearly
30 years, with average 8-hour TWA benzene concentrations exceeding 25 ppm (80 mg/m3) for at
least a part of the period of employment. As described by Fishbeck et al. (1978), in a 1953
industrial hygiene survey, workers who had an "operator" job classification had the highest
exposure to the compound, with an 8-hour TWA of 30-35 ppm (96-112 mg/m3) and occasional
"spikes" of up to 937 ppm (2990 mg/m3). However, a 1963 study described by Fishbeck et al.
(1992) revealed 8-hour TWAs ranging from 37 to 132 ppm (118 to 422 mg/m3), with four
employees' exposure levels exceeding 100 ppm TWA (319 mg/m3). Concurrent hematologic
evaluations undergone by these employees included Hgb, HCT, RBC and WBC counts,
sedimentation rates, platelet counts, differential blood counts, clot retention determinations, and
blood indices.
Key findings included enlarged RBCs, transient anemia, reduced hemoglobin
concentrations, and mean corpuscular volumes (MCVs) exceeding medically accepted values.
However, the authors noted that since 1963, the employees' MCVs had shown a reduction from
the highest values obtained, such that 5 of the 10 employees had MCVs within the normal range.
From these data, the authors concluded that at least some of benzene's effects on the
hematopoietic system may be reversible. This conclusion may be premature. This small group
of 10 employees was a survivor population and therefore may have introduced a potential bias.
They were selected because they were listed on current employment rolls of the company when
the study was in progress. Employees who had left employment earlier, perhaps with a benzene-
34
-------
related nonreversible condition, could not be selected for inclusion in this study. Therefore,
persons with an irreversible condition due to exposure to benzene could not have been
represented in this group of 10. This study cannot be used to establish a NOAEL or a LOAEL.
Aksoy et al. (1971) assessed hematologic abnormalities in 217 apparently healthy men
exposed to benzene in the shoe manufacturing industry. The test battery included RBC, WBC,
platelet, and differential cell count; packed cell volume (PCV); bone marrow examination; and
Hgb analysis. The subjects, 12-58 years old, were compared with 100 unexposed, age-matched
hospital workers. Benzene levels ranged from 15 ppm (48 mg/m3) during nonworking hours to
210 ppm (671 mg/m3) during the use of adhesives containing benzene; exposure duration ranged
from 3 months to 17 years. A total of 217 exposed workers were involved in the study, with 51
(23.5%) displaying hematologic abnormalities. Of these, leukopenia was present in 9.7%,
thrombocytopenia in 1.8%, leukopenia associated with thrombocytopenia in 4.6%, pancytopenia
in 2.8%, acquired pseudo-Pelger Huet anomaly in 0.5%, eosinophilia in 2.4%, and basophilia in
0.5%.
Bone marrow examination of 11 of the 51 workers with abnormal peripheral blood cell
counts revealed (1) normocellularity with no abnormal histology (two patients), (2) slight
hypocellularity (three patients), (3) hypercellularity (one patient), or (4) normocellularity with
abnormal histology. Abnormal findings present in various combinations included maturation
arrest in the erythroid and myeloid series, maturation arrest in the granulocytic series, and
marked vacuolization in the myeloid series. One worker with normal hematology developed
acute erythroleukemia 4 years later. The investigators concluded, based on the incidences of the
various blood abnormalities, that benzene had a greater effect on leukocytes (with basophilia and
eosinophilia as inconsistent findings) than on platelets. In a follow-up study, Aksoy et al. (1972)
reported that 8 of 32 workers diagnosed with pancytopenia died, mainly from infection and
bleeding.
In an industrial hygiene survey of 282 Dow Chemical workers who were categorized by
job classification into four groups on the basis of their exposure to benzene, exposed individuals
were monitored for a number of clinical chemistry and hematologic parameters and compared
with an age- and sex-matched cohort of unexposed Dow employees serving as controls
(Townsend et al., 1978). Ranges of benzene concentrations within the groups were < 2, 2-9,
10-24, and > 24 ppm (< 6.4, 6.4-29, 32-77, and > 77 mg/m3). However, as noted by the
authors, neither the RBC counts nor any other hematologic parameter correlated with exposure
intensity, duration of exposure, or estimated career dosage (Townsend et al., 1978). In fact, no
consistent relationship between any monitored parameter and benzene exposure was apparent
from their data.
A report by Yardley-Jones et al. (1988, 1990) described a survey in which blood samples
were obtained from 66 refinery workers exposed to benzene at comparative low concentrations
and 33 unexposed employees serving as controls. Exposure levels of < 10 ppm (32 mg/m3) were
estimated through consideration of the exposed employee's workstation, supplemented by a
small amount of personal monitoring data. Exposed subjects were assessed for the appearance of
sister chromatid exchange in peripheral lymphocytes and the ability of a subject's
P-glucuronidase/sulfatase-treated urine to promote gene reversion in the Ames test. In addition,
some hematologic parameters were measured in whole blood, although the largely negative
35
-------
findings gave no indication of a cause-effect relationship for benzene within the estimated
exposure range.
Bogadi-Sare et al. (2000), in a medical surveilance study, examined 49 female workers in
the shoemaking industry who were exposed to solvent mixtures and 27 nonexposed controls.
Exposures were as high as 15 ppm benzene and 50 ppm toluene. Significant differences in the
levels of benzene and toluene in blood and phenols in postshift urine between the exposed and
the control group confirmed solvent exposure. The number of B-lymphocytes was lower in the
shoe workers than in the controls (p=0.01). Significant correlations were found between the
level of immunoglobulin G and benzene in the work atmosphere. Confounding factors appeared
to play no role in the immunological findings. The authors found that exposure to benzene
concentrations lower than 15 ppm can induce depression of the circulating B-lymphocytes.
Khuder et al. (1999) reported on a group of 105 petroleum workers exposed to 0.14-2.08
ppm (0.45-6.6 mg/m3) benzene who, over time, had small but statistically significant falls in
certain blood counts. However, there were problems with this study, including a decrease in the
red cell MCV, a finding contrary to what is observed in benzene toxicity (Goldstein and Cody,
2000).
Collins et al. (1991) used a cross-sectional study design to compare five hematologic
parameters in 200 benzene-exposed workers with those of 268 unexposed employees in the same
plant. The type of work is not identified in the study. Estimates of the 8-hour TWA benzene
concentration were obtained from monitoring information coupled with knowledge of the
applicable work assignment and professional judgment. Benzene concentrations ranged from
0.01 to 1.4 ppm (0.03-4.5 mg/m3) for exposed subjects. Parameters under investigation included
RBCs, WBCs, Hgb concentration, platelets, and MCV. The authors noted statistically
significant differences in some of the hematologic parameters, based on demographic and
lifestyle factors. Thus, when multiple regression analyses were applied using the confounding
factors and current exposure as independent variables, no significant correlations between
benzene exposure and any hematologic parameter was evident. Again, this was a study of
currently employed workers and, as such, represented a healthy group of individuals, who,
having survived potentially adverse effects from exposure to benzene and other environmental
hazards, were able to continue to work and be included in this cross-sectional study. Hence, this
study cannot be used to establish a LOAEL.
Tsai et al. (1983) examined mortality patterns in a cohort of 454 former and current male
employees who were selected by industrial hygienists from a larger group of 20,000 present and
former refinery employees. The cohort members were chosen because they had worked directly
in the benzene, ethylene, aromatic distillate hydrogenation, or caiman units, the principal
petrochemical units in the refinery, for 1 to 21 years beginning in 1952. Average length of
"exposure" was 7.4 years. Maintenance workers who were assigned on a regular basis to these
same units were also included. Irregularly assigned workers to these units were not included,
although they may have been exposed to benzene. Personal exposures to benzene were
determined from 1979 to 1982. The median air concentration was 0.53 ppm in the "benzene-
related units," while in the refinery as a whole, the median was 0.14 ppm. In nonbenzene
exposure areas, the median was 0.07 ppm. This one-time "snapshot" of benzene exposure levels
says little about what the exposures may have been in the past or about study duration, nor do
36
-------
they tell us whether these personal samples relate to the 303 medically surveyed workers
discussed below, as they are a different group from the 454 who were part of the mortality
cohort. From this study it is impossible to determine the concentrations of benzene to which the
medically surveyed workers were exposed.
The analysis of overall mortality in the mortality cohort of 454 revealed no excesses. In
fact, mortality from all causes and from diseases of the circulatory system was significantly
below expected values, based on comparable groups of U.S. males. The authors concluded that
the presence of a healthy-worker effect produced significantly lower risk estimates, all causes
(standard mortality ratio = 0.53,/><0.05); only 34 deaths occurred, whereas 58.7 were expected.
The authors noted the relative youthfulness of the cohort. An internal but separate analysis of a
comparison group of 823 people consisting of 10% of the workers who were employed in the
same plant in operations not related to benzene produced higher relative risks of 0.90 and 1.31
for all causes and cancer at all sites, respectively (p<0.28 and/><0.23), higher than those of the
so-called "benzene-exposed" workers. Because of the appearance of significantly reduced risks
in this preselected cohort, the possibility of bias in the selection process should not be ruled out.
In addition, hematologic parameters were analyzed in a separate group of 303 male
"benzene workers" (according to the supervisors who selected them) beginning in 1959 and
ending in 1979. During this time, 11 deaths occurred in this group, whereas 21.5 were expected.
Up to four hematologic tests per year were conducted on just these workers. Total and
differential WBC counts, Hgb, HCT, RBC counts, platelets, and clotting times were calculated,
and a 95% range and standard deviation were derived for each hematologic test. Unfortunately,
as the authors pointed out, the data could not be compared with any "healthy population"
standard, because existing "normal" ranges are derived largely from hospital experiences. The
authors suggested that, based on the experience of this program, there is a need to establish
different interpretation criteria for screening for adverse effects of benzene exposure.
Thus, the Tsai et al. (1983) study is similar to the studies by Collins et al. (1991),
Fishbeck et al. (1978), and Yardley-Jones et al. (1988, 1990) in that the study population was a
currently employed group of workers (except for the 11 who died after the medical surveillance
was begun) and, as such, a survivor population in which few or no adverse health effects would
have been found. In addition, these medically surveyed employees were healthy and young.
They had been exposed for a short time to minimal levels of benzene that generally did not
exceed 0.53 ppm and probably were a lot less, based on survey results. Therefore, the selection
of a NOAEL is not considered on the basis of this study.
Considering data from the studies that have addressed the possible toxicological
consequences of exposure to benzene, a continuum of hematologic responses that are
etiologically linked to exposure to benzene is evident, although the precise threshold level for
these effects remains uncertain. However, some clarification of this issue has been provided by
an ongoing collaboration between the National Cancer Institute and the Chinese Academy of
Preventive Medicine that has generated a series of reports on both large and small cohorts of
subjects who were occupationally exposed to benzene.
Data from China describe clinical aplastic anemia in factories with exposure levels said
to range from 93 to 1156 mg/m3. Yin et al. (1987b) found 2676 cases of benzene
37
-------
poisoning—defined as a WBC count less than 4000 cells /mm3 and a history of benzene
exposure—from a review of more than 500,000 benzene-exposed workers in China. The
geometric mean concentration in 50,255 workplaces was 18.1 mg/m3. It was reported that 64.6%
of the workplaces had benzene exposure levels less than 40 mg/m3. From their review, the
authors concluded that cases of benzene poisoning can occur even in factories with less than 40
mg/m3 concentration of benzene.
Dosemeci et al. (1996) discussed the findings of a large-scale epidemiologic study
involving subjects employed between 1949 and 1987 in 672 factories in 12 cities in China. The
authors sought to establish plausible estimates of the levels of benzene exposure through an
industrial hygiene-based survey of all factory/work unit/job title/calender period data, providing
assessments that were subsequently correlated with one of six hypothetical exposure ranges
between < 1 and > 50 ppm. Thirty-eight percent of a total of 18,435 exposure estimates were
confirmed by workstation monitoring. The authors set hematologic criteria that defined benzene
poisoning to include a WBC count of < 4000 cells/mm3, platelets < 80,000/mm3, etc., thereby
allowing the identification of a case-defined subset of individuals who could be matched with
estimated levels of benzene exposure as described above. Relative risks of benzene poisoning
by intensity of exposure during the most recent 1.5-year exposure period compared with those
exposed to < 5 ppm (16 mg/m3) yielded estimated risks of 2.2 (95% CI = 1.7-2.9), 4.7 (95% CI =
3.4-6.5), and 7.2 (95% CI = 5.3-9.8) for exposure levels of 5-19 ppm (16-61 mg/m3), 20-39
ppm (64-125 mg/m3), and > 40 ppm (> 128 mg/m3), respectively. Clearly, if a significantly
elevated risk of benzene poisoning is an indication of hematotoxicity, then certainly exposures to
benzene at 5-19 ppm are hematotoxic. The decrease in ALC contributed substantially to the
decrease in WBC count, based on the results of the Rothman et al. (1996a) study discussed in the
next review. In general, recent exposure history appeared to be more closely related to elevated
risk than exposure duration, suggesting the comparatively rapid onset of benzene hematotoxicity.
These results support a threshold of benzene hematotoxicity in humans in the 5-19 ppm range, in
broad agreement with the emerging exposure-response range that is apparent from the
epidemiologic studies described in this section.
Some data that have shed further light on this issue have been provided by the same
group (Rothman et al., 1996a) that conducted a small cross-sectional study that compared 44
workers exposed to a range of benzene concentrations similar to those of Dosemeci et al. (1996)
with 44 age- and gender-matched unexposed controls. The purpose of this study was to show
that exposure to benzene affects all the major peripheral blood elements, with the ALC being the
most sensitive indicator of benzene-induced hematotoxicity. From public health district records,
workers from three workplaces in Shanghai, China, where benzene was used as a solvent and
unexposed workers from two workplaces in the same geographic area that did not use benzene
were selected as test subjects and controls, respectively. The three workplaces using benzene
included a factory that manufactured rubber padding for printing presses, a factory that
manufactured adhesive tape, and a factory that used benzene-based paint. The control
workplaces included a factory that manufactured sewing machines and an administrative facility.
Subjects who had a history of cancer, therapeutic radiation, chemotherapy, or current pregnancy
were excluded. Requirements for inclusion in the study were current employment for at least 6
months in a factory that used benzene, minimal exposure to other aromatic solvents, and no
exposure to other known bone marrow-toxic chemicals or ionizing radiation. Controls who had
38
-------
no history of occupational exposure to benzene or other bone marrow-toxic agents were
frequency-matched to the exposed subjects on age (5-year intervals) and gender.
Benzene exposure was monitored by organic vapor passive dosimetry badges worn by
each worker for a full workshift on 5 days within a 1- to 2-week period prior to collection of
blood samples. Benzene exposures of controls in the sewing machine factory were monitored
for 1 day, but no exposure monitoring was performed in the administrative facility. Compound
exposure was also evaluated by analyzing for benzene metabolites in urine samples collected at
the end of the benzene-exposure period for the exposed subjects. Median levels of the urine
metabolites phenol, catechol, MA, and hydroquinone correlated positively with the 8-hour TWA
personal air level measurements of benzene by exposure category (Table 2).
It is possible that air exposure to benzene may have been overestimated among workers
wearing marginally effective respirators. However, a comparison of urinary phenol levels to the
air benzene levels can be made. Urinary phenol is thought to increase in a linear fashion up to
100 ppm benzene. Roughly, 50 |ig/mg creatinine is thought to be equivalent to an 8-hour TWA
of 10 ppm benzene. For 19 workers exposed to more than 31 ppm who had an air and urine
sample taken on the same day, the median benzene exposure was 73 ppm and the median urine
phenol level was 351 ng/mg creatinine, about what one would predict (letter from Nathaniel
Rothman, National Institutes of Health, Bethesda, MD, to David Bayliss, NCEA-W, July 28,
1999). This is a test of the validity of the benzene air sampling effort. Dosemeci et al. (1996)
estimated that only about 5% of more than 1000 workers who were exposed to 100 ppm benzene
or higher in their factories were diagnosed with benzene poisoning. It is not known whether
severe hematotoxicity occurred to that 1 individual in 19 who was reported to have a TWA of
328.5 ppm. It is not always possible to obtain a perfect correlation between a surrogate measure
of exposure and an effect.
Historical benzene exposure of the subjects was evaluated by examining employment
histories. However, this information was not used to classify subjects by category of current
exposure. Data on age, gender, current and lifelong tobacco use, alcohol consumption, medical
history, and occupational history were collected by interview.
Six hematologic parameters were evaluated: total WBC count, ALC, HCT, RBC count,
platelet count, and MCV. Total WBC counts and ALCs were performed using a Coulter T540
blood counter. Abnormal counts were confirmed by hand. The Coulter company estimated that
the machine used in this study counts about 10,000 WBCs to generate a total WBC count and the
percent lymphocytes, which are multiplied to provide the ALC. The hand count involves
counting a standard 100 cells. The manual count also used to generate the percent lymphocytes
in this study was the standard 100 cells. Much higher-quality data would be expected for the
former and less precision for the latter, due to sampling variation (as well as human scoring
error). However, the Coulter counting machine actually generated larger and more significant
39
-------
Table 2. Median benzene urine metabolites, by exposure category, in a
study of workers exposed to benzene in Shanghai, China, 1992a
Exposure Category
Control (N = 44)
Exposed (< 3 1 ppm, N = 22)
Exposed (> 3 1 ppm, N = 22)
Phenol
17.3
38.9
348.7
Catechol
3.2
7.0
65.7
MA
0.18
8.15
46.8
Hydroquinone
1.6
12.8
64.3
aUnits ug/mg creatinine for all metabolites.
Source: Adapted from Rothman et al., 1996a.
differences in the ALC between the exposure groups than did manual counts. The authors
reported that previous studies that reported ALC and percent lymphocytes based on manual
counts may have underestimated group differences due to nondifferential measurement error.
Benzene metabolites in urine were measured by an isotope dilution gas chromatography/mass
spectrometry assay. Correlation analyses were completed using the Spearman rank order
correlation method. The Wilcoxon rank sum test was used to test for hematologic differences.
Twenty-one of the 44 subjects in both the exposed and control groups were female.
Mean standard deviation years of occupational exposure to benzene was 6.3 (standard
deviation = 4.4 years) with a range of 0.7-16 years. The median 8-hour TWA benzene exposure
concentration for the 44 exposed workers was 31 ppm (99 mg/m3). The 8-hour TWA for each
subject was derived from the geometric mean of the five air measurements that were gathered on
each of the exposed workers on the five different days during which the samples were taken.
Exposure to toluene and xylene was less than 0.2 ppm (0.6 mg/m3) in all exposure categories.
The exposed group was then subdivided into two equal groups of 22 workers consisting of those
exposed to greater than the median concentration of 31 ppm versus those exposed to less than the
median concentration according to the 8-hour TWA. The median (range) 8-hour TWA exposure
concentration was 13.6 (1.6-30.6) ppm (43.4 [5.1-97.8] mg/m3) for the low-exposure group and
91.9 (31.5-328.5) ppm (294 [101-1049] mg/m3) for the high-exposure group. A subgroup of the
low-exposure group composed of 11 individuals, none of whose five separate measurements
exceeded 31 ppm (100 mg/m3) at any time during the monitoring period, was also analyzed for
ALC. The median (range) 8-hour TWA exposure of these individuals was 7.6 (1-20) ppm (24
[3.2-64] mg/m3). All five air measurements were completed within 1 to 2 weeks of the
phlebotomies on all subjects.
All six blood elements measured were significantly different in the highest benzene
exposure group when compared with controls (Table 3). ALC, WBCs, RBCs, HCT, and
platelets were all significantly decreased, and MCV was significantly increased with increasing
benzene exposure. These effects are consistent with the hematotoxic effects of benzene shown
in Aksoy (1989) and Goldstein (1988). ALC was reduced from 1.9 x 103/|iL blood in controls to
40
-------
Table 3. Comparison of mean peripheral blood counts (with standard
deviations), by exposure status, in a study of workers exposed to benzene in
Shanghai, China, 1992
Exposure
Category
Control
(median =
0.02 ppm,
N = 44)
Exposed
(< 31 ppm,
N = 22)
Exposed
(>31ppm,
N = 22)
WBC Count
(xl03/nL blood)
6.8(1.7)
6.4(1.8)
5.6 (1.9)a
ALC
(xl03/nL blood)
1.9 (0.4)
1.6 (0.3)a
1.3 (0.3)c
RBC Count
(xlOVjiL blood)
4.7 (0.6)
4.6 (0.6)b
4.2 (0.6)c
HCT
(%)
42.0 (5.6)
41.2(5.7)
38.8 (5.3)b
MCV
(nm3)
88.9 (4.9)
89.8 (3.9)
92.9 (3.4)c
Platelets
(xl03/nL blood)
166 (59)
132 (45)b
121 (43)a
VO-05
Source: Adapted from Rothman et al., 1996a.
1.6 x 103/|iL (p<0.01) in the < 31 ppm (99 mg/m3) group and to 1.3 x 103/|iL (p<0.001) in the
group exposed to > 31 ppm (99 mg/m3) benzene. A dose-response relationship is evident in five
measures of hematotoxicity. The RBC and platelet counts were also significantly reduced in the
< 31 ppm (99 mg/m3) exposure group, but only ALC was significantly different in the lowest-
exposure subgroup (median = 7.6 ppm, 24 mg/m3). In this subgroup of 11 workers, whose
median 8-hour TWA exposure was 7.6 ppm (24 mg/m3) benzene, ALC (1.6 x 103/|iL) was
similar to that of the larger group of 22 who were exposed to < 31 ppm benzene. Thus, ALC is
the most sensitive measure of benzene hematotoxicity, representing a sentinel of a continuum of
the likely hematotoxic consequences of prolonged or enhanced exposure to benzene below 31
ppm.
One feature of the Rothman et al. (1996b) study was its attempt to probe the mechanisms
by which benzene may bring about the hematotoxicological symptoms that were evident in the
blood samples of exposed subjects. For example, although most peripheral blood cell levels
were decreased in exposed workers as compared with controls, peripheral cytokine levels (IL-3,
IL-6, erythropoietin, granulocyte colony-stimulating factor, and tissue necrosis factor-alpha)
were similar in exposed workers and controls, suggesting that benzene may not affect these
growth factor levels in peripheral blood. By contrast, when the rates of mutations were
measured at the glycophorin A (GPA) locus, an increased frequency of somatic cells with the
NN variant was interpreted as an indication of the compound's capacity to produce gene-
duplicating but not gene-inactivating mutations at the GPA locus in bone marrow cells (Rothman
etal., 1995).
41
-------
The study by Rothman et al. (1996a) is one of few occupational studies in which
individual- and workstation-monitored benzene concentrations provide a range of values that
encompass the exposure-response area into which, as judged by the data from other larger-scale
epidemiologic studies, the point-of-departure for the compound's hematologic effects appears
likely to fall. Further marking the study's relevance as a basis for developing quantitative
reference values for the compound is the exposure-response relationship that emerged between
ALCs and benzene concentrations. This observation is based on the inverse correlations
between these parameters that were reported for benzene-exposed workers and on the exposure
response that was observed when the data for exposed subjects were grouped on either side of
the overall median benzene concentration of 31 ppm (99 mg/m3).
Further assurance that the study met sufficient criteria for its use in developing an
inhalation RfC for benzene is provided by the facts that worker exposure to toluene and other
potentially harmful solvents was minimal in the exposed group and that the study protocol
fulfilled adequate criteria for age- and sex-matching of exposed and control subjects.
As discussed above, a number of subgroups of benzene-exposed workers have been
defined in the Rothman et al. (1996a) study, including (1) a high-dose subset of 22 workers, the
means of whose 5-day passive dosimetry measurements for benzene were greater than the
median (31 ppm [99 mg/m3]) of the 5-day mean benzene exposure concentration values for the
benzene-exposed group as a whole; (2) a low-dose group whose 5-day mean TWA benzene
exposure values were lower than 31 ppm (99 mg/m3) (median benzene concentration =13.6 ppm
[43.4 mg/m3]); and (3) a subset of the low-dose group whose exposure levels in all 5-day
consecutive breathing zone measurements were lower than 31 ppm (99 mg/m3) (median benzene
concentration = 7.6 ppm [24.3 mg/m3]). Because a reduction in ALC was the only significant
hematological effect obtained for all of the benzene groups, reduction in ALC was designated as
the critical effect, and the lowest exposure median of 7.6 ppm was defined as the LOAEL. There
is some concern in selecting the 7.6 ppm (24.3 mg/m3) concentration as the point of departure for
RfC development, because the subjects in question (N = 11) may provide an inadequate group-
wise comparison to the entire control group (N = 44). Nonetheless, a significant reduction in the
sensitive endpoint of ALC was observed in the subgroup exposed to a median benzene level of
7.6 ppm. Thus, this value is used as the LOAEL in Section 5.1.4. to derive a chronic inhalation
RfC for benzene and in Section 5.2.4. to derive a chronic oral RfD.
The primary quantitative analysis for developing the RfC and the RfD, however, uses
BMD modeling approaches to model the Rothman et al. (1996a) ALC data and derive a BMD at
the low end of the range of observation (Section 5.1.2). The BMDL is then used as the point of
departure for the application of UFs in the calculation of the RfC or RfD (Sections 5.1.3 and
5.2.3, respectively). Clearly, this is a preferable methodology because more of the exposure-
response data can be included in the analysis, and one is not restricted to selecting one of the
exposure group medians as a LOAEL. The LOAEL-based RfC and RfD estimates are provided
for comparison with the BMD-based estimates. RfC and RfD estimates are also derived from
applicable experimental animal studies for comparison with the human-based estimates (Sections
5.1.5 and 5.2.5, respectively).
42
-------
4.1.2.2. Neurotoxicity
Humans have displayed symptoms of neurotoxicity following acute inhalation of
relatively high concentrations of benzene (Snyder, 1987). Benzene produces generalized
symptoms such as dizziness, headache, and vertigo at levels of 250-3000 ppm (799-9584
mg/m3) (Brief et al., 1980), leading to drowsiness, tremor, delirium, and loss of consciousness at
700-3000 ppm (2236-9584 mg/m3) (ATSDR, 1997). Death may result from exposure to 20,000
ppm (63,894 mg/m3) benzene for 5-10 minutes (Sandmeyer, 1981). Neurological signs and
symptoms are similar for both lethal and nonlethal exposures to benzene and for exposure to
multiple solvents including benzene. These neurological symptoms are reversible upon removal
of the subject from exposure (Kraut et al., 1988).
Chronic exposure to benzene and toluene was studied in 121 workers exposed to benzene
for 2-9 years (Kahn and Muzyka, 1973). The air concentration of benzene between 1962 and
1965 was 6-15.6 ppm (20-50 mg/m3), whereas toluene vapors did not exceed the 5 mg/m3 level.
Subsequently, the air levels of both benzene and toluene did not exceed the 5 mg/m3 level.
Seventy-four of the examined workers complained of frequent headaches (usually at the end of
the workday), became tired easily, had difficulty sleeping, and complained of memory loss. The
limitations of this study are that workers were exposed to both benzene and toluene, and the dose
and duration of exposure were unknown.
Few studies have examined the neurological effects of chronic exposure to benzene.
Herregods et al. (1984) diagnosed transverse myelitis in a young man exposed daily to benzene
while working in the warehouse of a wholesale supplier of drugs and chemicals. The duration of
exposure was not given by the authors. The diagnosis of benzene poisoning was based on
differential diagnosis, ruling out other possible causes of poisoning. The clinical diagnosis was
based on the patient's occupation, high urinary phenol levels (28 mg/L in contrast to the normal
level, < 10 mg/L), and the coincidental decrease in urinary phenol and improvement of clinical
symptoms over about 6 months after cessation of exposure. The diagnosis of transverse myelitis
is consistent with an acute transection of the spinal cord affecting both the gray and white matter.
Baslo and Aksoy (1982) conducted neurological examinations on eight patients who had
a history of chronic exposure to benzene and who were diagnosed with aplastic anemia (cases
1-6) or preleukemia (cases 7 and 8). The seven males and one female ranged in age from 19 to
51 years (average age, 35.3 years). Cases 1, 2, 4, 5, 7, and 8 worked in the shoemaking industry,
case 3 was a leather worker, and case 6 was a whistle maker who dipped plastic material into an
open vessel of solution known to contain 88.42% benzene and 9.25% toluene. At times, the
concentrations of benzene in the working environments reached 210 ppm (671 mg/m3).
Exposure duration for the six patients with aplastic anemia ranged from 6 months to 8 years
(mean, 6 years), and the period between the cessation of exposure and the neurological
examinations ranged from 1 to 96 months. The two patients who had preleukemia (cases 7 and
8) were exposed for approximately 15 and 10 years, respectively, and were not exposed during
the 8 and 6 months preceding neurological examination. Neurological, electromyographical
(EMG), and motor conduction velocity examinations were performed on all patients (with the
exception of case 8, who did not have the EMG test). At least three different muscles were
tested in the EMG and motor conduction velocity examinations. Sensory conduction velocities
were measured in the upper and lower extremities of cases 1, 5, 6, 7, and 8.
43
-------
Four cases (1, 2, 3, and 8) had other neurological symptoms that included decreased
sensory vibration, atrophy of the leg, and exaggerated deep tendon reflexes; the other four (cases
4, 5, 6, and 7) had normal neurological test results. EMG tests revealed neurogenic involvement
in cases 1 and 2 but were normal in cases 3-7. Five of the cases (1, 2, 3, 7, and 8) had abnormal
motor conduction velocity tests, involving at least one of the nerves tested. These effects were
characterized by decreased motor conduction velocity and lengthening of latency in distal
nerves. Conduction velocity was decreased in the sensory nerves of cases 1, 6, and 7. In case 6,
the amplitudes of nerve action potentials were low. These effects reflected benzene-induced
effects on the axons of the peripheral nerves.
The investigators concluded that the neurological abnormalities in the four pancytopenic
individuals resulted from a direct effect of benzene or toluene (case 6) on peripheral nerves
and/or the spinal cord. They also concluded that the effects were related to the period of
nonexposure. For example, cases 3 and 4 had moderate to severe hematologic findings and
aplastic anemia, but no significant neurological findings. For these two cases, long periods of
nonexposure (53 and 96 months, respectively) preceded neurological testing.
The study is based on a small number of patients, lacks exposure and control data, and
does not rule out the possibility that the workers were exposed to chemicals other than benzene
and toluene; therefore, a reliable LOAEL or NOAEL could not be determined. The clinical
impression indicated a predominant involvement of the white matter in the peripheral nervous
system.
The reports summarized in this section have obvious deficiencies, such as small numbers
of patients under study and a lack of information regarding the intensity or duration of benzene
exposure. Thus, a reliable quantitative neurotoxic risk assessment is impossible at present.
Nevertheless, the occurrence of symptoms following exposure and the amelioration of effects
upon removal from exposure leave little doubt that benzene affects the central as well as the
peripheral nervous system.
4.1.2.3. Reproductive Toxicity
The available studies on the reproductive toxicity of benzene in humans are summarized
in Table 4.
Vara and Kinnunen (1946) examined 12 female workers who had gynecological
disorders attributed to benzene exposure: three were 25-28 years old, seven were 37-40 years
old, and two were 43-44 years old, and their exposure to benzene ranged from 1 to 10 years.
The investigators identified the third group as "nearing menopause" but found no evidence of
menopause; however, the symptoms of toxicity, consisting of hypermenorrhea and
hypomenorrhea, ovarian hypoplasia, sterility, degeneration of the ovary, and/or dysfunction of
the ovary were more severe in the two subjects nearing menopause. Peripheral blood counts
revealed distinct leukopenia in four subjects and decreased neutrophils and platelets in most
44
-------
Table 4. Reproductive toxicity of inhaled benzene in humans
Population
Female workers with
symptoms of benzene
poisoning
Female factory workers
with benzene-induced
ovarian hypofunction
Female gluing operators
Female solvent workers
Female workers in
shoemaking industry
No.
30
ND
360
174
223
Exposure
Overt symptoms suggested exposure to
levels greater than 3.2 mg/m3
ND
Exposure to gasoline (a major source of
benzene) and chlorinated hydrocarbons
via skin and inhalation; benzene levels,
< 5 mg/m3; 40% of chlorinated
hydrocarbon measurements exceeded
permissible limits by 1.2- to 2.4-fold;
controls had no chemical exposure
(1) < 131 mg/m3 benzene
(2) >131 mg/m3 benzene
(3) control
Subjects exposed to benzene and toluene
(compared with 327 controls)
Effects
12 subjects had menstrual disorders (e.g., hypermenorrhea)
and/or ovarian hypoplasia; women nearing menopause most
severely affected; 4 had leukopenia, and the majority had
reduced neutrophil and platelet counts
Workers exhibiting ovarian hypofunction had decreased
ascorbic acid concentrations in the blood, as compared with
clinically healthy female workers (p<0.001)
Functional disturbances of the menstrual cycle
(1) Hypermenorrhea in 4/40 (10%) (p<0.05)
(2) Hypermenorrhea in 4/47 (8.5%) (p<0.05)
(3) Hypermenorrhea in 1/87 (1.2%)
Increased menstrual disorders (exposed, 48.9%; control,
16.2%); spontaneous abortion (exposed, 5.7%; control, 2.4%);
gestosis (exposed 22.6%; control, 10.5%); all differences were
statistically significant
Reference
Vara and Kinnunen,
1946
Pushkinaetal., 1968
Mukhametova and
Vozovaya, 1972
Yinetal, 1987a
Huang., 1991
ND = no data.
-------
subjects. The investigators tentatively attributed the sparseness in menstruation to the ovarian
hypoplasia rather than to the benzene.
Pushkina et al. (1968) evaluated a group of workers who had ovarian hypofunction
related to exposure to benzene. Compared with clinically healthy female workers in the same
factory, the levels of ascorbic acid (changes in ascorbic acid metabolism were used as an index
of toxicity) were reduced in the blood of the workers who had ovarian dysfunction (subjects,
0.36 ± 0.02 mg%; controls, 0.49 ± 0.03 mg%;/X0.001). The details of this study were sparse,
but the results were supported by a study in rats that showed a similar effect. The justification for
the changes in the ascorbic acid as an index of toxicity was not well desribed.
A Russian study evaluated 360 women exposed to gasoline (a source of exposure to
benzene) and chlorinated hydrocarbons via inhalation and skin contact (Mukhametova and
Vozovaya, 1972). Benzene levels were < 1.6 ppm (5 mg/m3); chlorinated hydrocarbon levels
were 1.2-2.4 times higher than the permissible limits. Compared with workers with no exposure
to benzene, the subjects had increased incidences of menstrual disturbances. As exposure
duration increased, so did the number of premature interruptions of pregnancy, the percentage of
cases where theplacental membrane rupture during parturition was impeded, and the number of
cases of intrauterine asphyxia of the fetus. This study, although demonstrating exposure-related
reproductive effects and fetal toxicity, fails to distinguish between the effects of benzene and
chlorinated hydrocarbons.
Yin et al. (1987a) conducted a comprehensive study of the toxicity of benzene in which
300 workers exposed to benzene, toluene, or a mixture of the two were examined for subjective
symptoms and hematologic and biochemical effects. For 174 women, the mean TWA exposure
to benzene was 59 ppm (188 mg/m3) (exposure duration, 65 months). The women were divided
into low-exposure (1-40 ppm [3.2-128 mg/m3]) and high-exposure (41-210 ppm [131-671
mg/m3]) groups and were examined separately for subjective and pancytopenia-related subjective
symptoms. The benzene-exposed subjects reported an increase in the incidence of
hypermenorrhea, but the effect was not concentration-related (prevalence was 10% for lower
exposures and 8.5% for higher exposures; p<0.05 for both doses). Although the only
hematologic abnormality noted was a mild but statistically significant decrease in lymphocyte
count (p<0.05), this hypermenorrhea was tentatively considered to be related to pancytopenia.
These findings are strengthened by the large number of subjects, the measured exposure and
statistical evaluation, and similar observations in other studies; however, the significance of
subjective symptoms is debatable.
A Chinese report (Huang, 1991) described reproductive dysfunction in female workers
exposed to benzene and toluene in the leather shoemaking industry. The exposed group (223
women) exhibited increased incidence in the rate of menstrual disorders as compared with the
327 controls. The incidence rates of "mense-blood anomaly" and dysmenorrhea had a tendency
to increase with duration of employment. The incidences of spontaneous abortion and gestosis
(toxemia) were also increased. All increases were statistically significant. The investigators
concluded that both benzene and toluene had a deleterious effect on the reproductive function of
female workers.
46
-------
The data presented above provide limited evidence regarding reproductive effects of
benzene exposure in humans and involve concomitant exposure to other chemicals. These
studies did not provide good exposure monitoring data or quantitative dose-response
information.
4.1.2.4. Developmental Toxicity
The following discussion reviews the developmental toxicity of benzene in humans,
focusing on more recent reports. Available studies on the developmental toxicity of benzene in
humans are summarized in Table 5. There is no convincing evidence that benzene produces
malformations in humans or test animals; however, a few studies suggest that the chemical
induces adverse reproductive effects in humans and causes retarded fetal growth in test animals,
manifested mainly as decreased fetal weight and delayed ossification in the presence or absence
of maternal toxicity (Chatburn et al., 1981; ATSDR, 1997).
Benzene crosses the human placenta and is present in cord blood in quantities equal to or
greater than those in the maternal blood (Dowty et al., 1976). Summaries of developmental
toxicity of benzene in humans have been described in individual case reports (Forni et al., 1971;
Holmberg, 1979; Bordarier et al., 1991) and occupational studies (Vara and Kinnunen, 1946;
Pushkina et al., 1968; Mukhametova and Vozovaya, 1972; Funes-Cravioto et al., 1977; Axelsson
et al., 1984; Yin et al., 1987a; Savitz et al., 1989; Huang, 1991). Exposures were either to
benzene alone or, as is characteristic of occupational exposure, to multiple chemicals including
benzene. With only two exceptions (Bordarier et al., 1991; Mukhametova and Vozovaya, 1972),
all adult exposures in the studies summarized in these sections were via inhalation.
Only one study evaluated fetal effects following inhalation of benzene alone. A worker
who was exposed to benzene during her entire pregnancy suffered from severe pancytopenia and
exhibited increased chromosomal aberrations, but there were no effects on the fetus (Forni et al.,
1971).
Axelsson et al. (1984) evaluated by questionnaire the outcome of pregnancy among
personnel employed in laboratory work at the University of Gothenburg between 1968 and 1979.
Of 745 women who responded to the questionnaire, 556 had been pregnant (a total of 1160
pregnancies). The pregnant women were divided into two groups: those with exposure to
organic solvents during laboratory work and those without exposure. Responders to the
questionnaire reported exposure to at least 14 solvents; 41 workers remembered using benzene
during the first trimester of pregnancy and 5 used phenol. A slightly increased but not
significant difference in the miscarriage rate was found (relative risk = 1.31, 95% CI =
0.89-1.91). There were no differences in perinatal death rates or prevalence of malformations
between infants
whose mothers were exposed to solvents and those who were not exposed. The investigators
suggested that shift work, number of pregnancies, and age may have contributed to an increase in
miscarriage rate in the group of pregnant women not exposed to solvents, resulting in an
underestimation of the miscarriage rate of the exposed women. An additional concern was that
in spite of the fact that confounders were reduced by using as controls laboratory workers not
exposed to solvents, a mutagenic effect occurring during laboratory work (laboratory work the
woman had stopped working at the laboratory, thereby altering the miscarriage rate in the
47
-------
Table 5. Developmental toxicity of benzene—humans
Population
No.
Exposure
Effects
Reference
Pregnant worker
Exposure to benzene during entire
pregnancy
Maternal effects included severe pancytopenia and increased
chromosomal aberrations; no fetal effects
Fomietal, 1971
Employees in university
laboratory (-745
subjects, 1160
pregnancies), divided
into those with and
without exposure to
organic solvents
745
Responders to questionnaire reported
exposure to at least 14 solvents; 41
workers remembered using benzene
during the first trimester of pregnancy
and 5 used phenol
All women exposed to solvents had slight but not statistically
significant difference in miscarriage rate over those not exposed
(RR = 1.31, 95% CI = 0.89-1.91); 35 of 41 workers exposed to
benzene delivered, 1 had induced abortion, and 5 miscarried
(miscarriage rate for benzene-exposed subjects, 12.2%; miscarriage
rate for all responders to questionnaire, 11.1%; miscarriage rate for
unexposed responders, 10.1%); all 5 workers exposed to phenol
delivered; exposure to solvents did not affect perinatal death rates or
the incidence of malformations
Axelssonetal.,
1984
oo
Female gluing operators
360
Exposure to gasoline (a major source of
benzene) and chlorinated hydrocarbons
via skin and inhalation; benzene levels,
< 5 mg/m3; 40% of chlorinated
hydrocarbon measurements exceeded
permissible limits by 1.2- to 2.4-fold;
controls had no chemical exposure
Spontaneous abortions and premature births (17.2% vs. 4.9% in
controls), incidence of late membrane rupture, and intrauterine
asphyxia of the fetus increased with exposure duration
Mukhametova and
Vozovaya, 1972
Adult female workers
and 14 of their children
29
Adults exposed to benzene and other
organic solvents during pregnancy
(compared with 42 control adults and 7 of
their children)
Lymphocytes from adults exhibited approximately twofold increase
over controls in incidence of chromosomal aberrations and breaks;
their children exhibited increased frequency of chromatid breaks,
isochromatid breaks (p<0.0\, 14 children), and sister chromatid
exchanges (p<0.001, 4 children) in lymphocytes
Funes-Cravioto et
al., 1977
32-year-old pregnant
worker
Personal interview revealed exposure to
benzene and other solvents
(dichloromethane, methanol, and ether)
in laboratory during first trimester of
pregnancy; compared with matched
control
Stillborn anencephalic fetus
Holmberg, 1979
-------
Table 5. Developmental toxicity of benzene—humans (continued)
Population
No.
Exposure
Effects
Reference
23-year-old female
21 intramuscular injections of benzene in
an unsuccessful attempt to induce
abortion during first trimester of
pregnancy
Following normal delivery, infant exhibited slight dysmorphy
(hypotelorism and deep nasal bridge), moderate axial hypotonia and
abnormal ocular movements; at 1.5 months of age, child was
microcephalic, had severe axial hypotonia, severe peripheral
hypertonia, and bilateral optic atrophy, and CT scanning revealed
bilateral porencephalic cavities that created communication between
lateral ventricles and subarachnoid space; interventricular septum
lacking; child died from aspiration pneumonia at 2 months of age
Bordarier et al,
1991
RR = relative risk.
CT = computerized transverse tomography.
-------
control group. Another weakness in the study was that the subjects had worked in the laboratory
during 1968-1979, and some had to remember several years back.
Four other studies evaluated the developmental toxicity of multiple solvents. These
effects consisted of spontaneous abortions, premature births, and effects on "condition of the
fetus" (Mukhametova and Vozovaya, 1972), chromosomal abnormalities (Funes-Cravioto et al.,
1977), and stillbirth (Holmberg, 1979).
Savitz et al. (1989) examined the effects of parental (paternal as well as maternal)
occupational exposures on fetal development. The subjects, employed in various industries,
were exposed to a number of chemicals. Summarized here are the data pertaining only to
benzene exposures. The investigators accessed the National Natality and Fetal Mortality surveys
to obtain data on probability samples of live births and fetal deaths that occurred in the United
States in 1980 among married women. These data were merged with data provided by mothers
from questionnaires and from information from medical care providers for the public-use data
set. Omitted from the analysis were unmarried mothers, mothers who had not worked within 12
months of delivery, plural births, and births of unknown plurality.
The analysis, which was based on three case control studies of pregnancy outcome,
included case groups of stillbirths, preterm deliveries (birth before 37 weeks of gestation), and
small-for-gestational-age infants. The investigators used an exposure linkage system to
designate exposures. The exposure to a specific agent was assigned on the basis of the
occupation and industry of employment and on past studies and review of industrial processes.
Linkages of none, low, medium, high, or unknown were assigned to each agent, representing
probability (rather than intensity) of exposure. Unexposed working women served as controls.
The odds ratios (ORs) relating exposure to effect were adjusted for selected potential
confounders. Elevated ORs were found for maternal exposure to benzene and stillbirth (OR =
1.3, 95% CI = 1.0-1.8). Increased risks across low-, medium-, and high-linkage levels (crude
ORs of 0.9, 1.2, and 1.4, respectively) were observed. Exposure to benzene was also associated
with risk elevation for the fathers (OR =1.5, 95% CI = 1.1-2.3). The OR for low, medium, and
high linkages were 1.2, 1.5, and 2.0, respectively, with reasonably good precision for the two
highest levels (95% CI = 1.0-2.2 and 1.1-3.7, respectively). Benzene exposure linkage level
was unrelated to risk of preterm delivery for both parents. The investigators recognized that the
study had limitations. These included small population sizes, poor quality of exposure
information, and absence of statistical testing and the high rate of nonresponse to questionnaires
among women who were under 20 years old, 40 or more years old, and black and who had parity
four or greater, little or no prenatal care, and low education. However, in spite of these
limitations, the authors concluded that the results of the study encourage further evaluation of the
developmental effects of paternal exposure to benzene.
The data presented above provide inconclusive results regarding developmental toxicity
of benzene in humans. Most studies consisted of small numbers of subjects, lacked important
experimental details, and involved (in almost all cases) concomitant exposure to other chemicals,
and they did not provide monitoring data or quantitative dose-response information.
4.2. ACUTE AND CHRONIC STUDIES IN EXPERIMENTAL ANIMALS
50
-------
4.2.1. Hematotoxicity
4.2.1.1. Oral Exposure—Subchronic Studies
Subchronic oral studies have been conducted in F344 rats and B6C3F1 mice of both
sexes (NTP, 1986; Huff et al., 1989) and female Wistar rats (Wolf et al., 1956). Other
subchronic oral studies have been conducted to examine immunotoxicity, and these are
described in Section 4.2.4.1.
Hsieh et al. (1988b) exposed male CD-I mice to benzene in drinking water at
concentration levels 0, 31, 166, or 790 mg/L (0, 8, 40, or 180 mg/kg/day) for 28 days. Dose-
related hematologic effects (erythrocytopenia, leukocytopenia, lymphocytopenia, increased
MCV) were observed at all exposure levels. It can be estimated that these dose levels in
drinking water are roughly equivalent to inhaled doses of 6.4, 32, and 154 ppm. A major
drawback of this study is that there were only five animals per group. Details of the study are
discussed in Section 4.2.4.1.
In the National Toxicology Program (NTP) study (NTP, 1986), F344 rats and B6C3F1
mice (10/species/group/sex; 6-8 weeks of age) were treated with 0, 25, 50, 100, 200, 400, or 600
mg/kg benzene by gavage in corn oil 5 days/week for 17 weeks. The adjusted doses were 0, 18,
36, 71, 143, 286, or 429 mg/kg/day. An additional five animals/species/group/sex were tested at
the 0, 200, or 600 mg/kg dose levels and killed at 60 days of treatment. Hematologic analyses
were performed on all the animals killed at 60 days and on five animals/species/group/sex at the
end of the study. In addition, necropsies were performed on all animals, and the spleens of all
animals were examined histopathologically.
No compound-related deaths were observed for rats. Final body weight depression of
> 10% relative to controls was observed in male and female rats at dose levels of 200 mg/kg and
greater. Significant (p<0.05) leukopenia and lymphocytopenia were observed in male and
female rats after 60 days of treatment with 200 or 600 mg/kg (the only treatment groups tested
on day 60). On day 120 of treatment, significant leukopenia and lymphocytopenia were
observed in female rats at 25 mg/kg and higher, and significant lymphocytopenia was observed
in male rats at 400 mg/kg. Lymphoid depletion of B-cells in the spleen was observed in 100% of
male and female rats exposed to 600 mg/kg for 60 or 120 days. Increased extramedullary
hematopoiesis in the spleen was observed in four of five male and three of five female rats
treated with 600 mg/kg for 120 days. This study identified a LOAEL of 25 mg/kg (18
mg/kg/day) in female rats and a LOAEL of 200 mg/kg (143 mg/kg/day) in male rats for
hematologic effects following treatment by gavage for 17 weeks. The LOAEL for female rats
was at the lowest dose tested; thus, no NOAEL was established.
There were no compound-related deaths in the mice. A final body weight depression of
7% was seen in the 100 mg/kg dose group. Tremors were observed intermittently in male and
female mice treated with 400 or 600 mg/kg. No leukopenia or lymphocytopenia was observed in
male or female mice after 60 days of treatment with 200 or 600 mg/kg. At 120 days, significant
(p<0.05) leukopenia and lymphocytopenia were observed in male mice at dose levels of 50
mg/kg and greater and in female mice at 400 (only lymphocytopenia) and 600 mg/kg. A
51
-------
NOAEL of 25 mg/kg (18 mg/kg/day) and a LOAEL of 50 mg/kg (36 mg/kg/day) for
hematologic effects were identified in male mice treated by gavage for 17 weeks. A NOAEL of
200 mg/kg (143 mg/kg/day) and a LOAEL of 400 mg/kg (286 mg/kg/day) for hematologic
effects were identified in female mice treated by gavage for 17 weeks.
Female Wistar rats (10/group) were treated by gavage with benzene in olive oil 5
days/week for 6 months (Wolf et al., 1956). The reported doses were 0, 1, 10, 50, or 100 mg/kg
(0, 0.7, 7.1, 35.7, and 71.4 mg/kg/day). Parameters measured included mortality, clinical signs,
body and organ weights, hematology, blood biochemistry, bone marrow counts, and gross and
microscopic pathology of lungs, heart, liver, kidneys, spleen, testes, adrenals, and pancreas.
Leukopenia was reported for 10 mg/kg; at higher dose levels erythrocytopenia and leukopenia
were also observed. No quantitative data or statistical analysis were reported. The authors
reported that rats fed 1 mg/kg had "no evidence of ill effects." This study identified a NOAEL
of 0.7 mg/kg/day and a LOAEL of 7.1 mg/kg/day for hematologic effects in female rats treated
by gavage for 6 months.
Aoyama (1986) showed that a 14-day exposure of mice to 50 ppm (162 mg/m3) benzene
resulted in a significantly reduced blood leukocyte count.
4.2.1.2. Oral Exposure—Chronic Studies
In the NTP (1986) study, F344 rats and B6C3F1 mice of both sexes were treated by
gavage with benzene 5 days/week for 103 weeks. Results of this study have also been reported
by Huff et al. (1989). For rats, males (60/group) were administered doses of 0, 50, 100, or 200
mg/kg (0, 36, 71, or 143 mg/kg/day), and females (60/group) were administered doses of 0, 25,
50, or 100 mg/kg (0, 18, 36, or 71 mg/kg/day). Survival decreased with increasing dose in rats
of both sexes and was significantly decreased (p<0.05) at 200 mg/kg in males and at 50 and 100
mg/kg in females. Body weight depression of > 10% relative to controls was observed in male
rats treated with 100 mg/kg. Dose-related leukopenia was significant in female rats treated with
>25 mg/kg for 3, 6, 9, and 12 months; leukocyte levels were comparable with controls after 15,
18, 21, and 24 months of treatment. In male rats, dose-related leukopenia was significant at >50
mg/kg. Lymphoid depletion was observed in the thymus of 0/44, 4/42, 8/41, and 10/34 male rats
treated with 0, 50, 100, and 200 mg/kg benzene, respectively. In the spleen, lymphoid depletion
was observed in 0/49, 19/58, 8/47, and 23/47 male rats treated with 0, 50, 100, and 200 mg/kg,
respectively. This study identified a LOAEL of 25 mg/kg (18 mg/kg/day) for leukopenia and
lymphocytopenia in female F344 rats and 50 mg/kg (36 mg/kg/day) in male F344 rats. These
were the lowest doses tested, and thus no NOAEL was identified.
B6C3F1 mice (60/sex/group) were administered doses of 0, 25, 50, or 100 mg/kg (0, 18,
36, or 71 mg/kg/day). Survival decreased with increasing dose in mice of both sexes and was
significantly decreased (p<0.05) at 100 mg/kg. Body weight depression of > 10% relative to
controls was observed in mice of both sexes treated with 100 mg/kg. Significantly decreased
leukocyte counts were observed in males after 3, 6, 9, 12, 15, 18, and 21 months of treatment
with 50 or 100 mg/kg, but males treated with 25 mg/kg had significantly decreased leukocyte
counts only after 6 and 21 months. In female mice, leukopenia was observed only at 12 and 18
months, in both cases significant at all treatment levels. Significantly decreased lymphocyte
counts were observed in males after 3, 6, 9, 12, 15, 18, and 21 months with 50 or 100 mg/kg, but
52
-------
males treated with 25 mg/kg had significantly decreased lymphocyte counts only after 12
months. In female mice, significant lymphocytopenia was observed at >25 mg/kg at 12 and 18
months and at 100 mg/kg at 3 months. Hematopoietic hyperplasia of the bone marrow was
observed in 0/49, 11/48, 10/50, and 25/49 male mice and in 3/49, 14/45, 8/50, and 13/49 female
mice treated with 0, 25, 50, or 100 mg/kg, respectively. Increased splenic hematopoiesis was
observed in 5/49, 9/48, 19/49, and 24/47 male mice and in 9/49, 10/45, 6/50, and 14/49 female
mice treated with 0, 25, 50, or 100 mg/kg, respectively. In the female mice, increased incidences
of epithelial hyperplasia of the ovary occurred at all three doses, and increased incidence of
senile atrophy of the ovary occurred at the lower two doses compared with controls. This study
identified a LOAEL of 25 mg/kg (18 mg/kg/day) for leukopenia and lymphocytopenia in male
and female B6C3F1 mice. The observed LOAEL was at the lowest dose tested; thus, a NOAEL
was not identified.
Beginning in 1976, a series of carcinogenicity studies on oral treatment of rodents with
benzene were performed at the Bologna Institute of Oncology, including 52- and 104-week
studies on Sprague-Dawley and Wistar rats and Swiss and Rf/J mice (Maltoni et al., 1983, 1985).
Limited information regarding noncarcinogenic effects were reported for Sprague-Dawley rats,
but only carcinogenicity data were published for Wistar rats and mice. No statistical information
was included, making interpretation of the data difficult.
Maltoni et al. (1985) treated Sprague-Dawley rats (13 weeks of age, 30-35/sex/group) by
gavage with 0, 50, or 250 mg/kg benzene in oil 4-5 days/week for 52 weeks and then observed
them until death. Expanded doses were 0, 32, and 161 mg/kg/day. In addition, Sprague-Dawley
rats (7 weeks of age, 40-50/sex/group) were treated by gavage with 0 or 500 mg/kg benzene in
oil 4-5 days/week for 104 weeks and then observed until death. The expanded doses were 0 and
321 mg/kg/day. Mortality was higher in benzene-treated groups and appeared to be dose related;
body weights were not affected. Maltoni et al. (1983) stated that mortality in the first portion of
the study was due to direct toxic effects of treatment, and in the later portion it was partially due
to tumors. Mortality was similar to that of controls during treatments with 500 mg/kg for 92
weeks; body weight appeared to be somewhat depressed relative to controls. In Sprague-Dawley
rats exposed to 500 mg/kg for 84 or 92 weeks, decreased total RBCs (only at 92 weeks), WBCs,
and lymphocytes were observed. Insufficient information was provided to establish LOAEL or
NOAEL levels from these studies.
4.2.1.3. Inhalation Exposure—Subchronic Studies
The following discussion summarizes more recent studies of the effects of inhaled
benzene on the peripheral blood of experimental animals. The available studies on the
hematotoxicity of inhaled benzene to test animals are summarized in Table 6. These studies
form a database that supports the understanding that exposure to benzene at high concentrations
causes hematotoxic effects, including bone marrow suppression and the greater sensitivity of
mice over that of rats. Benzene-induced hematotoxicity has been observed from short-term as
well as long-term exposure to the chemical.
53
-------
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—inhalation exposure
Strain/Species
Endpoints
Exposure
Effects
LOAEL/NOAEL
(mg/m3)
Reference
CD-I mice
(11-12 males/
group)
Body weight, cells in
peripheral blood and bone
marrow, spleen weight, total
nucleated cells per spleen,
lymphocytes
0, 32 mg/m3, 6
hr/day, 5
days/wk for
10 weeks
Increased spleen weight, total nucleated
cells per spleen, and nucleated cells per
spleen
LOAEL: 32
Green et al.,
1981a, b
CD-I mice
(30 males and
30
females/group)
HCT; Hgb; WBC, RBC,
platelet, reticulocyte, and
differential counts; MCV;
mean corpuscular Hgb;
myeloid/erythroid ratio from
bone marrow; leukocyte
alkaline phosphatase; RBC
glycerol lysis time; and other
more general parameters of
subchronic toxicity
0, 3.2, 32, 96,
or 958 mg/m3
6 hr/day, 5
days/wk for
up to 91 days;
serial
sacrifices after
7, 14, 28, 56,
and 91 days of
exposure
At 958 mg/m3, decreased HCT, total
Hgb, RBC, leukocyte, and platelet
counts; decreased myeloid/erythroid
ratios and percentage of lymphocytes;
increased MCV, mean cell Hgb,
glycerol lysis time, and incidence and
severity of RBC morphological changes
(statistically significant [p<0.05] for
males and/or females); many effects
occurred and were statistically
significant by 14 days, persisting
throughout exposure; microscopic
changes in the thymus, bone marrow,
lymph nodes, spleen, ovaries, and testes;
changes in males occurred more often
and with greater severity than those in
the females
LOAEL: 958
NOAEL: 96
Wardetal.,
1985
AKR/Jmice
(60 males, test;
60 males,
control)
WBC and RBC counts
(Coulter), reticulocyte counts,
differential counts, Hgb, HCT,
lactic acid dehydrogenase,
RBC acetylcholinesterase,
reduced GSH, cytogenetic
analyses of bone marrow, and
microscopic examination of
lung, liver, spleen, and kidney
958 mg/m3
benzene, 6
hr/day, 5
days/wk for
life
Early mortality (mean survival time, 11
weeks for test animals, 39 weeks for
controls), severe lymphocytopenia, bone
marrow hypoplasia, granulocytosis, and
reticulocytosis
LOAEL: ND
NOAEL: ND
Snyderetal.,
1978
-------
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—inhalation exposure (continued)
Strain/Species
Endpoints
Exposure
Effects
LOAEL/NOAEL
(mg/m3)
Reference
AKR/Jmice
(50 males, test;
50 males,
control)
RBC and WBC counts
(Coulter) and differential
count; microscopic
examination of lung, liver,
spleen, kidney, and bone
marrow
319 mg/m3
benzene, 6
hr/day, 5
days/wk for
life
No statistically significant differences
between test and control mice in median
survival time, rate of weight gain or
lymphoma type or occurrence;
statistically significant lymphocytopenia
(cell counts were depressed to -65% of
normal throughout exposure); depressed
RBC counts (± 2 standard error of mean
at 9/19 monitoring periods); increased
neutrophils (± 2 standard error of mean
at 3/19 monitoring periods); bone
marrow hypoplasia in 20% of the
treated mice (p<0.05)
LOAEL: ND
NOAEL: ND
Snyderetal.,
1980
C57BL/6 mice
(40 males, test;
40 males,
control)
RBC and WBC counts
(Coulter) and differential
count; microscopic
examination of lung, liver,
spleen, kidney, and bone
marrow
958 mg/m3
benzene, 6
hr/day, 5
days/wk for
life
Decreased survival (median time 41
weeks, test animals; 75 weeks,
controls); lymphocytopenia (WBC
counts, ~3000/mm3 or 15% of normal);
anemia (RBC counts, 8 x lOVmm3 or
—80% of normal); neutrophilia
appearing at 17 weeks; abnormal blood
cell morphology; neutrophilic shift to
left; bone marrow hyperplasia in 33% of
test animals, none in controls;
hematopoietic neoplasms
LOAEL: ND
NOAEL: ND
Snyderetal.,
1980
-------
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—inhalation exposure (continued)
Strain/Species
Endpoints
Exposure
Effects
LOAEL/NOAEL
(mg/m3)
Reference
CD-I mice (40
males, test; 40
males, control)
RBC and WBC counts
(Coulter), and differential
count; microscopic
examination of lung, liver,
spleen, kidney, and bone
marrow
958 mg/m3
benzene, 6
hr/day, 5
days/wk for
life
Decreased median survival time (25.5
weeks for benzene-exposed animals,
52.7 weeks for controls); depressed
peripheral RBC and lymphocyte counts
(± 2 standard error of mean);
neutrophilia; abnormal blood cell
morphology; and a shift to immature
myeloid cells (at 217 days); tumor
incidence (5/40) not significant; acute
myeloblastic leukemia in one exposed
animal, chronic myelogenous leukemia
in one; bone marrow hyperplasia in 9/35
animals without neoplasia, bone marrow
hypoplasia in 11, splenic hemosiderin
pigments in 6, and splenic hyperplasia
in 19
LOAEL: ND
NOAEL: ND
Snyderetal.,
1982
Sprague-
Dawley rats
(10 males and
10
females/group)
HCT; Hgb; WBC, RBC,
platelet, reticulocyte, and
differential counts; MCV;
mean corpuscular hemoglobin;
myeloid/erythroid ratio from
bone marrow; leukocyte
alkaline phosphatase;
erythrocyte glycerol lysis time;
and other more general
parameters of subchronic
toxicity
0, 3.2, 32, 96,
or 958 mg/m3
6 hr/day, 5
days/wk for
up to 91 days;
serial
sacrifices after
7, 14, 28, 56,
and 91 days of
exposure
Statistically significant (p<0.05)
decrease in WBC counts (males on day
14 and females on days 14-91) and
slightly decreased femoral marrow
cellularity at 958 mg/m3 (300 ppm)
LOAEL: 958
NOAEL: 96
Wardetal.,
1985
-------
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—inhalation exposure (continued)
Strain/Species
Sprague-
Dawley rats
(45 males, test;
27 males,
control)
Female BDF1
mice (4/group)
Male
C57BL/6J
mice (5/group)
Male and
female C57/6
BNL mice
(5-10/group)
Male Sprague-
Dawley rats
Endpoints
WBC and RBC counts
(Coulter), reticulocyte counts,
differential counts, Hgb, HCT,
lactic acid dehydrogenase,
RBC acetylcholinesterase, and
reduced GSH, cytogenetic
analyses of bone marrow, and
microscopic examination of
lung, liver, spleen, and kidney
WBCs, reticuloctyes,
differential counts, and assays
for CFU-S, BFU-E, CFU-E,
and GM-CFU-C
Peripheral blood counts, spleen
and bone marrow BFU-E and
CFU-E
WBC, RBC, differential
counts, HCT, bone marrow
cellularity, and CFU-S in bone
marrow and spleen
Erythrocyte and lymphocyte
counts
Exposure
958 mg/m3
benzene, 6
hr/day, 5
days/wk for
life
0,319, 958, or
2875 mg/m3, 6
hr/day, 5
days/wk for
16 weeks
32 mg/m3, 6
hr/day, 5
days/wk for
up to 178 days
0, 32, 80, or
1278 mg/m3, 6
hr/day, 5
days/wk for 2
weeks
Oor319
mg/m3, 6
hr/day, 5
days/wk for
life
Effects
Mild weight depression,
lymphocytopenia, trend to anemia, fatty
changes in 77% of bone marrow
samples
Reduced lymphocyte count at > 3 19
mg/m3, decreased CFU-E, BFU-E,
CFU-S, and GM-CFU-C
Reduced circulating RBCs, lymphocyte
counts, bone marrow CFU-E and BFU-
E, and splenic CFU-E
Reduced lymphocyte counts after 2
weeks of exposure at 80 mg/m3 but no
effect at 32 mg/m3
No significant reductions in erythrocyte
or lymphocyte counts
LOAEL/NOAEL
(mg/m3)
LOAEL: ND
NOAEL: ND
LOAEL: 3 19
NOAEL: NA
LOAEL: 32
LOAEL: 80
NOAEL: 32
NOAEL: 3 19
Reference
Snyderetal.,
1978
Seidel et al.,
1989
Baarsonetal.,
1984
Cronkite etal.,
1985
American
Petroleum
Institute, 1983
-------
Table 6. Peripheral blood and hematopoietic effects of benzene in animals—inhalation exposure (continued)
Strain/Species
Male B6C3F1
mice (24/group
exposed, bone
marrow tests
on
3-16/group)
Male B6C3F1
mice
(3 or 9/group)
Male Sprague-
Dawley rats
(16/group)
Endpoints
Bone marrow; cell counts,
CFU-E, GM-CFU-C, andB-
and T-lymphocytes
Splenic, thymic, and femoral
lymphocytes and labeling
index of femoral lymphocytes
Thymus and spleen weight,
spleen and bone marrow
cellularity, spleen CD4+/CD5+,
CD8VCD5+ and Kappa+
lymphocytes
Exposure
0, 3.2, 32,
319, or 639
mg/m3, 6
hr/day, 5
days/wkfor 1,
2, 4, or 8
weeks
0, 3.2, 16, 32,
319, or 639
mg/m3, 6
hr/day, 5
days/wkfor 1,
2, 4, or 8
weeks
0, 96, 639, or
1278 mg/m3, 6
hr/day, 5
days/wk for 2
or 4 weeks
Effects
Reduced bone marrow cellularity,
progenitor cells, and differentiating
hematopoietic cells, at > 3 19 mg/m3, but
no effects observed at <32 mg/m3
Persistent and rapid reductions in
splenic and femoral lymphocytes at
>3 19 mg/m3, transient reduction of
thymic cell count and splenic B-
lymphocytes at 32 mg/m3 after 2 weeks,
but comparable with controls by 4
weeks
Reduced spleen B-CD4+/CD5+ and
CD5+ T-lymphocytes at 1278 mg/m3,
no effects observed at 639 mg/m3
LOAEL/NOAEL
(mg/m3)
LOAEL:319
NOAEL: 32
LOAEL:319
NOAEL: 32
LOAEL: 1278
NOAEL: 639
Reference
Farris et al.,
1996, 1997a
Farris et al.,
1997b
Robinson et
al., 1997
oo
ND = not determined because only one concentration was used.
NA = not applicable, adverse effects were observed at the lowest dose tested.
CFU-S = colony-forming unit-spleen.
BFU-E = burst-forming unit-erythroid.
CFU-E = colony-forming unit-erythroid.
GM-CFU-C = granulocyte/macrophage colony-forming unit-culture
-------
Early general toxicity studies reported leukopenia in dogs and fatal anemia in mice
exposed to 600 ppm (1917 mg/m3) benzene for 12-15 days (Hough and Freeman, 1944); changes
in bone marrow histopathology or leukopenia in rats, guinea pigs, and rabbits exposed to 80-85
ppm (256-272 mg/m3) benzene for 23-187 exposures (Wolf et al., 1956); and leukopenia in rats
exposed to 61 ppm (195 mg/m3) benzene for 2-4 weeks or to 44 ppm (141 mg/m3) for 5-8 weeks
(Deichmann et al., 1963).
Male CD-I mice (11-12/group) were exposed for 6 hours/day, 5 days/week to
concentrations of 0 or 10 ppm (0 or 32 mg/m3) benzene for 10 weeks or to 0 or 300 ppm (0 or
958 mg/m3) for 26 weeks (Green et al., 1981a,b). On the day of the last exposure, samples
(pooled from groups of 3-4 mice) were obtained from the peripheral blood, bone marrow, and
spleen to evaluate hematologic and hematopoietic cells. In mice exposed to 10 ppm (32 mg/m3),
no adverse effects were observed with respect to mortality, body weight, or cells in the
peripheral blood or bone marrow. Spleen weight, total nucleated cells per spleen, and nucleated
RBCs per spleen were significantly increased (p<0.05) in mice exposed to 10 ppm (32 mg/m3).
Mice exposed to 300 ppm (958 mg/m3) had the following significant (p<0.05) changes:
increased mortality rate; decreased numbers of lymphocytes and RBCs in peripheral blood;
decreased granulocyte/macrophage progenitor cells in bone marrow; decreased spleen weight
and numbers of lymphocytes; multipotential hematopoietic stem cells and committed
granulocyte/macrophage progenitor cells in the spleen; and increased incidence of atypical cell
morphology in the peripheral blood, bone marrow, and spleen. These studies identify a LOAEL
of 10 ppm (32 mg/m3) for slight hematopoietic effects in mice exposed to benzene for 10 weeks.
In a subchronic inhalation toxicity study, Ward et al. (1985) evaluated the peripheral
blood and bone marrow of benzene-exposed CD-I mice and Sprague-Dawley rats. Groups of 50
male and 50 female rats and 150 male and 150 female mice inhaled benzene concentrations of 0,
1, 10, 30, or 300 ppm (0, 3.2, 32, 96, or 958 mg/m3) 6 hours/day, 5 days/week for up to 13
weeks. Serial sacrifice of 30 mice and 10 rats/group took place after 7, 14, 28, 56, and 91 days
of exposure. The parameters of toxicity evaluated included behavior, body weight, organ
weights, clinical pathology, gross pathology, and histopathology. Chamber analyses of benzene
concentrations calibrated daily by gas chromatography and monitored continuously by infrared
analyzer were within 10% of the target concentrations.
Clinical observations and body weight data revealed no signs of exposure-related
toxicity. Hematologic effects were not observed in either species at 1, 10, or 30 ppm (3.2, 32, or
96 mg/m3). The mice exposed to 300 ppm (958 mg/m3) for 91 days, however, exhibited
decreased HCT and reductions in total Hgb concentration, RBC count, WBC count, platelet
count, myeloid/erythroid ratios, and percentage of lymphocytes; the RBCs of the exposed mice
also displayed increases in MCV, mean cell Hgb, and glycerol lysis time and in the incidence
and severity of red cell morphological changes. The values for RBC, motor conduction velocity,
mean corpuscular Hgb, and glycerol lysis time were statistically significant (p<0.05) for both
males and females, whereas the remainder of the values were statistically significant for males
only.
Many of the effects occurred (and were statistically significant) by 14 days of exposure
and most persisted throughout exposure, but the data provided indicated that the effects did not
increase in severity with duration of exposure. Microscopic examination of the high-
concentration group revealed changes in the thymus, bone marrow, lymph nodes, spleen, ovaries,
59
-------
and testes. The changes in the males occurred more often and with greater severity than those in
the females. The most common compound-related histopathologic findings included myeloid
hypoplasia of the bone marrow, lymphoid depletion of the periarteriolar sheath in the spleen and
mesenteric lymph node, increased extramedullary hematopoiesis in the spleen, and plasma cell
infiltration of the mandibular lymph node. The incidence and/or severity of thymic atrophy and
myeloid hypoplasia appeared to increase with duration of exposure. The rats, not as severely
affected as the mice, exhibited significantly (/X0.05) decreased WBC counts (males on day 14
and females on days 14-91) and slightly decreased femoral marrow cellularity at 300 ppm (958
mg/m3).
The investigators concluded that the hematologic effects observed in this study are
related to concentration and are similar to those reported by other investigators, but that they are
statistically significant only at 300 ppm (958 mg/m3). In addition, the mouse is more sensitive
than the rat to the effects of benzene, and male mice appear to be more sensitive than female
mice. The experiment provided a LOAEL of 300 ppm (958 mg/m3) and a NOAEL of 30 ppm
(96 mg/m3) for peripheral blood abnormalities in both rats and mice. The large numbers of
animals used in the study and the evaluation of several parameters of hematotoxicity strengthen
the conclusions of the investigators.
4.2.1.4. Inhalation Exposure—Chronic Studies
To further characterize the hematotoxicity of benzene, Snyder et al. (1978, 1980, 1982,
1984, 1988) conducted a series of studies using either 100 or 300 ppm (319 or 958 mg/m3)
benzene. These studies addressed mainly the influence of benzene concentration and animal
species on toxic responses. The 300 ppm (958 mg/m3) concentration was selected because it had
produced severe hematotoxicity in AKR mice exposed for life, but this concentration had
produced only lymphocytopenia in Sprague-Dawley rats exposed under the same conditions.
Also of interest was whether benzene would affect the incidence of lymphoma in the different
strains of rodent. The AKR strain carries a virus that produced a high incidence of spontaneous
lymphoma, killing the mice within 1 year, and the C57BL strain carries a virus that yields a high
incidence of lymphoma following exposure to radiation, carcinogens, and immunosuppressive
agents (Snyder et al., 1980).
Exposures were similar in all the studies. Essentially, test animals and control animals
were placed in identical stainless steel inhalation chambers and exposed to benzene or filtered
air, respectively, 6 hours/day, 5 days/week for life. Benzene concentrations, monitored every 30
minutes by ultraviolet absorbance at 255 nm, were read from a calibration curve. Tail bleeding
of test and control animals took place at least every 3-5 weeks during exposure and at about the
same time of day for analysis of the following endpoints: WBC and RBC counts (Coulter),
reticulocyte counts, differential counts, Hgb, and HCT. Evaluations of lactic acid dehydrogenase
levels, RBC acetylcholinesterase levels, and reduced GSH levels were conducted less frequently.
The evaluation also included cytogenetic analyses of bone marrow. Other tissues taken for
microscopic examination included lung, liver, spleen, and kidney.
In the first study of the series, Snyder et al. (1978) exposed AKR/J mice (60 test and 60
control) and Sprague-Dawley rats (45 test and 27 control) to 300 ppm (958 mg/m3) benzene.
The rats had mild body weight depression that started at 30 weeks and persisted throughout life.
60
-------
The only significant hematologic findings in the rats (± 2 standard error [SE]) were
lymphocytopenia and a trend to anemia. Lymphocyte counts for the exposed animals were
approximately 75% of normal at 2 weeks and 53% of normal at 20 weeks. Because of increased
mortality, blood analyses were discontinued after 1 year of exposure. The investigators also
noted fatty changes in 77% of the bone marrow samples taken from the benzene-exposed
animals compared with 42% of the controls. The last mouse in the study died after 28 weeks of
exposure, probably of aplastic anemia (mean survival time, 11 weeks for test animals, 39 weeks
for control animals). The treated animals suffered severe weight loss (59% weight change at 16
weeks), marked lymphocytopenia (e.g., lymphocyte values that dropped to < 1000/mm3 were
-3% of control levels after 36 days of exposure), bone marrow hypoplasia (81% of exposed mice
and 6% of controls,/?<0.001 by chi-square), granulocytosis (fivefold increase after 9 weeks of
exposure), and reticulocytosis. The blood analyses were suspended after 92 days of exposure
due to early mortality. Neither species exhibited any signs of leukemia or preleukemia (the
absence of overt signs of leukemia, characterized by bone marrow dysfunction, usually low
numbers of certain cell types). The investigators suggested that this might be related to early
mortality from aplastic anemia in the mice and a lack of opportunity for recovery of the bone
marrow (a possible factor in leukemogenesis) in the rats.
The investigators concluded from this study that inhalation of 300 ppm (958 mg/m3)
benzene causes significantly decreased survival, severe lymphocytopenia, and anemia
accompanied by reticulocytosis and granulosis in AKR/J mice and lymphocytopenia, mild
anemia, and moderately decreased survival in Sprague-Dawley rats. They further concluded that
AKR/J mice are more susceptible to benzene-induced hematotoxicity than are Sprague-Dawley
rats.
The severity of hematologic effects induced by 300 ppm (958 mg/m3) benzene in AKR/J
mice prompted Snyder et al. (1980) to conduct a similar study with 50 AKR/J mice exposed to a
lower concentration (100 ppm [319 mg/m3]) of benzene. The investigators also exposed 40
C57BL/6 mice to 300 ppm (958 mg/m3) benzene to compare the susceptibility of a different
strain of mice with that of rats. Hematologic parameters were measured in both strains at regular
intervals during exposure.
No statistically significant differences were seen between the AKR/J test and control
mice in median survival (39 and 41 weeks, respectively), rate of weight gain, or in type and
occurrence of lymphoma. Effects that did occur were not as severe as those observed at 300
ppm (958 mg/m3). But even at 100 ppm (319 mg/m3), the mice exhibited statistically
significantly reduced lymphocyte counts after 1 week of exposure (cell counts were depressed to
65% of control levels), and the depression persisted throughout exposure. RBC counts were also
consistently depressed relative to controls, and the depression was statistically significant (± 2
SE) at 9/19 monitoring periods. Neutrophils were elevated (± 2 SE) in 3/19 monitoring periods.
Ten (20%) of the treated mice, but only one control, had bone marrow hypoplasia (p<0.05).
C57BL/6 mice exposed to 300 ppm (958 mg/m3) benzene had a median survival time of
41 weeks, compared with 75 weeks for the controls. Starting 1 week after the beginning of
exposure, mice exposed to benzene exhibited a statistically significant increase in incidence of
lymphocytopenia at 29 of 30 monitoring periods (WBC counts were about 3000/mm3 or -15% of
control values), and of anemia in 30 of 30 monitoring periods (RBC counts were -80% of
61
-------
control values). After 17 weeks, the incidence of neutrophilia was significantly increased in 15
of the 22 remaining monitoring periods. Throughout exposure, the peripheral RBCs exhibited
anisocytosis and poikilocytosis. A neutrophilic shift to the left, characterized by the appearance
of metamyelocytes, indicated immaturity. At the end of the first year, myelocytes,
promyelocytes, and giant platelets began to appear. Blood analyses were discontinued after 61
weeks because of decreased survival. Examination of the bone marrow revealed hyperplasia,
mainly of granulopoietic elements, in 13 (33%) of the exposed mice; hyperplasia was not present
in the controls. There was a significant increase in the incidence of hematopoietic neoplasms,
including six cases (15%) of thymic lymphoma. Two control mice died with nonthymic
lymphoma.
The investigators concluded from this study that the two strains of mice are more
sensitive than the Sprague-Dawley rat to the hematologic effects of benzene and that the AKR/J
mouse is more sensitive than the C57BL/6 mouse. The response of the C57BL/6 mouse
producing a proliferative effect on the myeloid cell line is significant, because this is the cell line
that undergoes leukemic transformation in humans exposed to benzene.
In the next study, Snyder et al. (1982) examined the effects of 300 ppm (958 mg/m3)
benzene on CD-I mice, a strain not known to harbor any endogenous lymphoma virus. The
effects were similar to those observed in the AKR/J and C57BL/6 strains and included decreased
median survival time (25.5 weeks for benzene-exposed animals, 52.7 weeks for controls),
statistically significantly depressed peripheral RBC and lymphocyte counts after the first week of
exposure (± 2 SE), and statistically significant neutrophilia (± 2 SE) after 29 weeks (-209 days)
of exposure and the appearance of Howell-Jolly bodies (remnants of nuclear chromatin seen in
erythrocytes in certain anemic conditions) at 7 days, anisocytosis (erythrocytes showing
excessive variation in size) at 22 days, and poikilocytosis (erythrocytes showing excessive
variation in shape) at 92 days and a shift to immature myeloid cells at 217 days.
Tumor incidence was not significant. One benzene-exposed animal developed acute
myeloblastic leukemia, one had chronic myelogenous leukemia, one had a benign lung adenoma,
and two had malignant lymphoma with thymic involvement. Of the 35 benzene-exposed animals
that did not have neoplasia, 9 had bone marrow hyperplasia, 11 had bone marrow hypoplasia, 6
had splenic hemosiderin pigments (indicating hemolysis or ineffective erythropoiesis), and 19
had splenic hyperplasia. The investigators noted that "exposed mice dying with bone marrow
hypoplasia survived, on the average, 75 fewer days than mice dying with bone marrow
hyperplasia. This may indicate that bone marrow hypoplasia is an early response to exposure
followed by bone marrow hyperplasia. It may also indicate two different responses to the
exposure."
Snyder et al. (1988) examined the influence of modifications in inhalation protocol on
benzene-induced hematotoxicity and tumor development in C57BL/6 and CD-I mice. Exposure
was by two protocols, one representing intermittent occupational exposures (1 week of exposure
to 300 ppm [958 mg/m3] benzene followed by 2 weeks of nonexposure alternately for life), and
the other representing intense but short-term exposures (1200 ppm [3834 mg/m3] benzene for 10
weeks); all exposures were 6 hours/day, 5 days/week. The long-term intermittent exposures
produced earlier mortality in both strains of mice than did the short-term, 10-week exposures.
Both types of exposure produced severe lymphocytopenia and moderate anemia in both strains
62
-------
of mice, but peripheral blood values in the 1200 ppm (3834 mg/m3)/ 10-week exposure groups
returned to normal after termination of exposure. The 300 ppm (958 mg/m3) exposure induced
lymphocytopenia and anemia throughout the course of exposures. Tumor incidences were
significantly increased (p<0.05-0.001) with both protocols in CD-I mice and with the
intermittent, long-term exposure in the C57BL/6 strain. A puzzling finding was that neither
leukemia nor lymphoma occurred in either strain by either protocol. The investigators concluded
that benzene was cytotoxic by either protocol, especially to circulating lymphocytes (but
recovery took place when exposure to the high concentration ceased) and that both protocols did
induce tumors. The Snyder et al. (1978, 1980, 1982, 1984, 1988) studies were designed to
answer specific questions and used only one concentration at a time (except for the study on
exposure protocol), and therefore they were not sufficient for the derivation of LOAEL or
NOAEL.
4.2.1.5. Effects on Stem Cell Populations
The benzene-induced peripheral blood abnormalities reflect a disruption of
hematopoiesis in the bone marrow; thus many experimental studies have focused on this process,
analyzing the total cellularity (the total number of bone marrow cells obtained from a femur), as
well as the individual cellular components of the bone marrow. The pluripotent hematopoietic
stem cells are capable of self-renewal and can differentiate along at least three lineages:
erythrocytic, granulocytic/macrophagic, and thrombocytic (Snyder, 1987). These cells are
quantitated by the spleen colony assay in which the colonies formed by the cells in the spleens of
lethally irradiated host animals are counted; these are called colony-forming units-spleen (CFU-
S). The specific progenitor cells, in contrast to the pluripotential stem cells, are committed to
differentiate, and they also have some capacity for self-renewal (cell division). These cells are
enumerated by counting the colonies formed in cell cultures grown in the presence of specific
growth-factor stimuli (Snyder, 1987). The two types of erythroid progenitor cells are the less-
differentiated burst-forming units-erythroid (BFU-E), which responds to a factor called "burst-
promoting activity" or to IL-2, and the more differentiated erythroid progenitor, the colony-
forming units-erythroid (CFU-E), which responds to erythropoietin. The
granulocyte/macrophage colony-forming units-culture (GM-CFU-C) respond to conditioned
media from human leukemia cells or from organs of animals treated with plant lectins. The T-
and B-lymphocytes arise from the pluripotential hematopoietic stem cells in the bone marrow.
The cells progress to the recognizable blast cells and their progeny and further mature to the
functional forms normally observed in peripheral blood (Snyder, 1987).
Seidel et al. (1989) examined the effects of benzene on peripheral blood cell counts and
the hematopoietic stem cell compartments of the bone marrow. Female BDF1 mice were
exposed to 100, 300, or 900 ppm (319, 958, or 2875 mg/m3) benzene 6 hours/day, 5 days/week
for 16 weeks. Benzene concentrations were monitored twice a day. Control mice, placed in
chambers, were exposed to air. The evaluation of hematopoietic effects included WBC
(Coulter), reticulocyte, and differential counts and assays for CF-U-S, BF-U-E, CFU-E, and
GM/(CFU-C) using standard procedures. The assays, conducted about every 2 weeks, usually
took place about 6 hours after the end of the exposure period of the preceding week.
There was a dose-dependent reduction in lymphocyte count after both 4 and 8 weeks of
exposure. Lymphocyte counts of the animals exposed to 300 and 900 ppm (958 and 2875
63
-------
mg/m3) returned to control values 1-2 weeks after cessation of exposure, whereas the
lymphocytes in the 100 ppm (319 mg/m3) group remained low during that time. The animals
exposed to 300 and 900 ppm (958 and 2875 mg/m3) developed a slight anemia that did not
worsen with longer exposure. Three-fourths of the animals with elevated CFU-E in the bone
marrow also showed reticulocytosis at 13 weeks. The number of granulocytes was apparently
not affected. The investigators noted that control values for the 300 ppm (958 mg/m3) groups
were approximately twice those for the 100 and 900 ppm (319 and 2875 mg/m3) groups, possibly
because the 300 ppm (958 mg/m3) experiment had been performed 1 year earlier.
Assays performed after 4 and 8 weeks of exposure demonstrated decreases in stem cell
numbers that were dose related, particularly at 8 weeks. At 4 weeks, the CFU-E count was
depressed more than other stem cell counts. At 100 ppm (319 mg/m3), the CFU-E were
depressed to 48% of control, whereas the BFU-E, CFU-S, and CFU-C (colony-forming
unit-culture) were depressed to 88%, 92%, and 94% of control, respectively. At 8 weeks, the
values for CFU-E in animals exposed to 100, 300, and 900 ppm (319, 958, and 2875 mg/m3)
were 99%, 73%, and 35% of controls, respectively. Similar patterns of depression were
observed for BFU-E and CFU-C numbers of animals exposed to benzene for 8 weeks, and CFU-
S numbers were depressed to 65% of control at 300 ppm (958 mg/m3), the only dose tested after
8 weeks of exposure. Recovery studies demonstrated a gradual regeneration of progenitor cell
numbers, and animals exposed to 300 ppm (958 mg/m3) for 16 weeks had completely recovered
by 73-185 days postexposure. The investigators concluded that the erythroid system is more
sensitive than the myeloid cell system to the effects of benzene, with the CFU-E showing more
sensitivity than other components of the erythroid system under the conditions of this study.
Two studies were designed to examine the effects of low concentrations of benzene on
the erythroid and myeloid progenitor cells (Baarson et al., 1984; Baarson and Snyder, 1991).
Baarson et al. (1984) examined the effects of subchronic exposure to a low concentration of
benzene on the erythroid progenitor cells in the bone marrow and spleen of mice. Male
C57BL/6J mice inhaled 10 ppm (32 mg/m3) benzene vapor 6 hours/day, 5 days/week for as long
as 178 days. Peripheral blood counts and bone marrow and spleen BFU-E and CFU-E assays
were performed on days 32, 66, and 178 of exposure. The numbers of circulating RBCs in the
benzene-treated groups were significantly decreased (p<0.05) at both 66 and 178 days.
Lymphocyte values were decreased (p<0.05) at all three time points. Benzene did not affect
levels of circulating neutrophils, bone marrow cellularity, and numbers of nucleated red cells in
the bone marrow.
The numbers of bone marrow CFU-E gradually declined during exposure to only 5% of
control values after 178 days. The decline, statistically significant (p<0.01) at all three time
points, was apparently exponential with a half-life of about 2 months. Bone marrow BFU-E
colonies were significantly depressed in benzene-exposed animals (to about 55% of control;
/X0.01) at 66 days, but recovered to control values by 178 days. Depressions in splenic
nucleated red cell numbers (to about 15% of control values; p<0.05) and in splenic nucleated
cellularity (p<0.05) occurred at day 178. Splenic CFU-E colonies in benzene-exposed mice
were decreased to 10% of control values at 178 days (p<0.05). In contrast, splenic BFU-E
colonies were increased (but not significantly) at all time points. The spleen is a site of
extramedullary erythropoiesis in the rodent, particularly under stressful conditions. The
reduction in the numbers of nucleated red cells and CFU-E in the spleen following exposure to
benzene indicated that the compensatory mechanism of the spleen was also affected. The
64
-------
investigators concluded that low concentrations of benzene are hematotoxic and that further
studies to examine this effect would be warranted.
In the later study, Baarson and Snyder (1991) reported that the ingestion of ethanol in
combination with exposure to benzene may enhance benzene-induced toxicity of the erythroid
progenitor cells in C57BL/6J mice. These studies identify a LOAEL of 10 ppm (32 mg/m3) for
depressed hematopoiesis in mice.
Vacha and coworkers (1990) exposed female C57BL/6xDBA/2Fl hybrid mice to 0 or
300 ppm (0 or 958 mg/m3) benzene 6 hours/day, 5 days/week for 6-7 weeks. Indices of
hematopoiesis were measured in peripheral blood (RBC and WBC count, Hgb, HCT,
reticulocyte, and leukocyte count) in addition to 59Fe accumulation in the erythropoiesis organs
(spleen and bone marrow) and in the peripheral RBCs. The distribution of developmental
classes of erythroblasts was also determined. This study found that animals became anemic after
6-7 weeks of benzene exposure. The number of erythroblasts in the bone marrow was not
different; however, exposure to benzene shifted the population to a less mature class of cells.
The number of colonies derived from BFU-E and CFU-E were decreased to 70 and 34% of
controls, respectively. A LOAEL of 300 ppm (958 mg/m3) was established for hematotoxic
effects.
Toft et al. (1982) evaluated the adverse effects of occupationally relevant levels of
benzene on the bone marrow of mice. Male NMRI mice inhaled concentrations of benzene that
ranged from 1 to 200 ppm (3.2 to 639 mg/m3). Exposures were either continuous (24 hours/day
for 4 to 10 days) or intermittent (8 hours/day, 5 days/week for 2 weeks) using five mice per
group. Endpoints included the number of nucleated cells per tibia, the number of colony-
forming granulopoietic stem cells (CFU-C/tibia), and frequency of micronuclei in polychromatic
erythrocytes (PCE). Mice exposed continuously to >21 ppm (67 mg/m3) benzene exhibited
significant and concentration-dependent alterations in all three parameters. The values after 4
days of exposure to 21 ppm (67 mg/m3) of benzene (estimated from a graph) were cells/tibia,
-24% of control; CFU-C/tibia, -32% of control; micronuclei/500 PCE, 6 (control value,
0.41/500).
Furthermore, intermittent inhalation exposure to 1.0, 10.5, 21, 50, 95, and 107 ppm (3.2,
33, 67, 160, 303, and 342 mg/m3) for 2 weeks produced concentration-dependent decreases in
CFU-C/tibia and increases in the frequency of micronuclei, significant at >21 ppm (67 mg/m3),
and concentration-dependent increases in cellularity, significant at >50 ppm (160 mg/m3).
Animals exposed intermittently to 95 or 201 ppm (303 or 642 mg/m3) benzene 2-8 hours/day, 5
days/week for 2 weeks had significant decreases in cellularity and CFU-C/tibia (95 ppm [303
mg/m3] for 6 and 8 hours/day) and increases in the frequency of micronuclei (95 ppm [303
mg/m3] for 4 hours). At 201 ppm (642 mg/m3), all exposure protocols, except for the 2
hours/day exposure, produced significantly adverse effects on all parameters. Statistical
significance was determined by the Student's t test for cellularity and CFU-C and by the Wilcox-
Whitney method for micronuclei, which was set at/><0.05. Statistical values were not given for
the individual data points.
The investigators concluded that the CFU-C/tibia was suppressed to a greater extent than
was the overall cellularity for most exposures and that CFU-C are more sensitive to prolonged
65
-------
exposure to low levels of benzene than are the majority of bone marrow cells. However, higher
exposures for short durations produced the reverse, suppressing cellularity but not CFU-C. This
could indicate that short exposures cause rapid injury to the bone marrow but that the injury
could be offset by a compensatory increase in the proliferation rate of the bone marrow. The
investigators further concluded that intermittent exposure is less effective than continuous
exposure, that the induction of micronuclei is important because somatic mutations precede most
chemically induced cancers, and that the cytotoxicity observed in the study could lead to aplastic
anemia.
Because of the relationship between exposure to benzene and the development of
myeloblastic leukemia, Dempster and Snyder (1990) evaluated the possibility that exposure
provides a growth advantage for granulocytic progenitor cells over erythroid progenitor cells.
Basically, DBA/2 male mice inhaled benzene concentrations of 0, 10, 30, or 100 ppm (0, 32, 96,
or 319 mg/m3) 6 hours/day, 5 days/week for 5 days and were evaluated for BFU-E, CFU-E, and
GM-CFU-C on days 1 and 5 after exposure. To determine the effects of benzene on stem cells
during increased erythropoiesis, subgroups of five exposed mice received injections of hemolytic
doses of phenylhydrazine during or after exposure. Controls were exposed to air and/or injected
with saline. Brief descriptions of treatment and the effects, significant at/><0.05, are as follows.
Benzene only. One day after exposure, there was a dose-dependent depression in bone
marrow BFU-E and CFU-E, but GM-CFU-C were not affected (except for a spurious increase at
30 ppm [96 mg/m3]). Splenic CFU-E and BFU-E were increased at 10 ppm (32 mg/m3) and
CFU-E were depressed at 100 ppm (319 mg/m3). Splenic granulocytic cells were unaffected.
Five days after exposure, BFU-E and CFU-F were the same as in controls, but GM-CFU-C were
decreased at 10 and 100 ppm (32 and 319 mg/m3); splenic CFU-E exhibited a concentration-
related increase, significant at 30 and 100 ppm (96 and 319 mg/m3), whereas the numbers of the
other two progenitor cell types were normal.
Benzene + phenylhydrazine. One day after exposure, bone marrow BFU-E were
depressed at 30 and 100 ppm (96 and 319 mg/m3), and CFU-E were depressed only at 30 ppm
(96 mg/m3); there was no effect on the number of GM-CFU-C. Splenic BFU-E and GM-CFU-C
showed concentration-dependent depressions, and CFU-E were depressed, but only at 100 ppm
(319 mg/m3). Five days after exposure, bone marrow CFU-E were elevated in mice exposed to
10 ppm (32 mg/m3), and GM-CFU-C exhibited a concentration-dependent depression; bone
marrow BFU-E counts were similar to control counts. Splenic GM-CFU-C were elevated at all
concentrations, CFU-E were elevated at 10 and 30 ppm (32 and 96 mg/m3) and depressed at 100
ppm (319 mg/m3), and BFU-E were elevated only at 10 ppm (32 mg/m3).
The investigators concluded that acute exposure to benzene has different effects on
erythroid and granulocytic progenitor cell populations, which resulted in a growth advantage for
granulocytic cells in both the bone marrow and the spleen of exposed mice. This observed shift
toward granulopoiesis occurred even in mice treated with the erythropoiesis stimulus
phenylhydrazine, but the effects were short-lived. The bone marrow erythroid progenitor cells
had recovered from their depression 5 days after exposure, and at that time the granulocytic
progenitor cells had become depressed. The dose-dependent increase in splenic CFU-E 5 days
after exposure probably reflects the spleen's attempt to repopulate the erythron.
66
-------
In a series of studies, Luke et al. (1988a, b) and Tice et al. (1989) examined the effect of
sex, route, schedule, and duration of exposure on benzene-induced cytotoxicity and genotoxicity
of the bone marrow in mice. The effects on percentage of PCE in peripheral blood, PCV, and
bone marrow cellularity (all considered by the investigators to be measures of bone marrow
cytotoxicity) are reviewed below.
Groups of 6-10 male and female DBA/2 mice, male B6C3F1 mice, and male C57BL/6J
mice inhaled either 300 ppm (958 mg/m3) benzene or ambient air 6 hours/day for 13 weeks.
Exposures took place on either 3 consecutive days/week or 5 consecutive days/week. For
comparison with the oral route of exposure, one group of B6C3F1 males received doses of 400
mg/kg by gavage on 5 consecutive days/week for 14 weeks. The oral dose of 400 mg/kg was
estimated to exceed the total amount of benzene absorbed by a mouse during a 6-hour exposure
to 300 ppm (958 mg/m3). Peripheral blood smears, prepared weekly, were used to examine PCE.
PCVs and bone marrow cellularity were determined at the end of the study. The data for the
three parameters were statistically analyzed using temporal averages and then evaluated for
differences related to sex, strain, regimen, and route of exposure using a two-way Brown
Forsythe analysis of variance (ANOVA). Group mean data were compared using Student's t
test.
Benzene initially induced a significant depression in the number of PCE in the peripheral
blood of all three strains exposed by either route and both inhalation regimens. However, the
extent and duration of the depression varied by sex, strain, and exposure regimen. Female
DBA/2 mice exhibited the smallest initial decrease in PCE, and values returned to control levels
by week 3 of exposure. The extent of depression was not dependent on exposure regimen.
Among the males of the different strains, the DBA/2 mouse was the most severely affected,
particularly when exposed to benzene 3 days/week. ANOVA revealed significant differences in
the ability of benzene to suppress erythropoiesis that were related to both strain (for C57BL/6
and DBA/2 mice) and exposure regimen. Oral treatment produced an initial suppression in PCE
levels in the male B6C3F1 mice that persisted for only 2 weeks and was not revealed in the
temporal average.
The study also demonstrated a significant depression in the PCV and bone marrow
cellularity (determined at the end of the study) of the exposed mice. With regard to the PCV
values, the depression was independent of the inhalation regimen in B6C3F1 and C57BL/6 mice.
After correcting for differences in PCV levels in controls, the data showed that inhalation
produced a greater effect than oral exposure to benzene, the extent of which was strain-
dependent for all three strains (B6C3F1 ~ C57BL/6 > DBA/2; p<0.0163). In DBA/2 mice the
extent of depression was sex dependent (male > female; /X0.025). With regard to bone marrow
cellularity, the depression observed after inhalation exposure depended on sex (male > female;
/X0.0002) and strain (DBA/2 < B6C3F1 ~ C57BL/6;/X0.0012). Among the male mice, the
extent of depression was dependent on route (inhalation > oral;/><0.012) and was not dependent
on regimen.
These studies used rather small numbers of animals but otherwise appear to have been
thoroughly conducted and evaluated; however, the information provided is for exposure to
benzene that is significantly above the most recent occupational limits. The investigators
67
-------
suggest that further studies are needed to evaluate the influences on benzene toxicity relative to
regimen and route of exposure at lower doses.
In a series of experiments, Cronkite et al. (1985, 1989) examined the hematotoxicity of
inhaled benzene in mice. The study examined the effects on and recovery of peripheral blood,
bone marrow, and progenitor cells, as well as the development of neoplasia and the influence of
exposure regimen on toxicity. In the first study, male and female C57BL/6 BNL mice were
exposed for 6 hours/day, 5 days/week to benzene concentrations of 10, 25, 100, or 400 ppm (32,
80, 319, or 1278 mg/m3) for 2 weeks or to 300 ppm (958 mg/m3) for durations of exposure
ranging from 2 to 16 weeks. The endpoints of toxicity included WBC, RBC (Coulter), and
differential counts; HCT; bone marrow cellularity; CFU-S in the bone marrow and spleen; and
tumor development. Assays were performed 18-22 hours after termination of exposure or after
various periods of recovery up to 16 weeks.
Inhalation of 25, 100, or 400 ppm (80, 319, or 1278 mg/m3) benzene for 2 weeks did not
affect granulocyte counts but did result in a concentration-related depression in the number of
lymphocytes in the peripheral blood; there was no effect on lymphocytes at 10 ppm (32 mg/m3).
HCT values were decreased at 100 and 400 ppm (319 and 1278 mg/m3); RBC counts were also
decreased, but no data were given. At the end of exposure, lymphocyte depression ranged from
about 20 to 65% of control for the three highest exposure groups. For mice exposed to 300 ppm
for 2 weeks, lymphocyte levels were similar to control levels by 4 weeks after the end of
exposure, and recovery of the remaining exposure duration groups was complete by 8 weeks
after exposure. This experiment provided an estimated LOAEL of 25 ppm (80 mg/m3) and a
NOAEL of 10 ppm (32 mg/m3) for lymphocyte depression in mice exposed to benzene for 2
weeks (Cronkite et al., 1985). These results differed from those of Ward et al. (1985), which
suggested a LOAEL of 300 ppm (958 mg/m3) for all hematologic parameters tested in both mice
and rats. The main difference in the two studies was the use of different strains of mouse.
Cronkite et al. (1985) also determined bone marrow cellularity, the percentage of bone
marrow stem cells synthesizing DNA, numbers of pluripotent stem cells in bone marrow, and
CFU-S. Cellularity was determined in animals inhaling 0, 10, 25, or 100 ppm (0, 32, 128, or 319
mg/m3) for 2 weeks (the 400 ppm [1278 mg/m3] concentration was not tested as it was for the
other parameters of the study). Benzene levels of 100 ppm (319 mg/m3) significantly depressed
bone marrow cellularity and stem cell CFU-S numbers (p<0.003 and/><0.001, respectively) in
exposed mice. In contrast, 10 and 25 ppm (32 and 128 mg/m3) had no effect on these
parameters. A single assay demonstrated a depression in the fraction of stem cells in DNA
synthesis in mice exposed to 10 ppm (32 mg/m3). There was no effect at 25 ppm (128 mg/m3),
and slight increases in the fraction in DNA synthesis at 100 and 400 ppm (319 and 1278 mg/m3)
(p<0.17 and/><0.08, respectively). These results for DNA synthesis are of uncertain relevance.
In another experiment, the recovery of bone marrow CFU-S population was tracked after
exposure of the animals to 300 ppm (958 mg/m3) for 2, 4, 8, or 16 weeks. In animals exposed
for 2 and 4 weeks, bone marrow CFU-S numbers were depressed to 90% of control values at the
end of exposure but had recovered to >100 and >95% of control values, respectively, 4 weeks
later. In animals exposed for 8 weeks, CFU-S levels were decreased to 50% of control values 4
weeks after the last exposure, but they recovered 8 weeks later (after 16 weeks, the values were
>100% of control). In animals exposed for 16 weeks, CFU-S values were depressed to 27% of
68
-------
control, and although recovery was occurring by week 2 after the last exposure, it was still not
complete by 16 weeks after termination of exposure. The investigators concluded that exposure
to 300 ppm (958 mg/m3) benzene for 16 weeks significantly depressed the hemopoietic stem
cells and that recovery was incomplete 16 weeks after termination of exposure.
In reviewing the effects of benzene on the CFU-S compartment, Snyder (1987) pointed
out that under normal circumstances the CFU-S cells are nondividing and in that state are
probably resistant to benzene toxicity; however, in response to benzene-induced injury of more
differentiated cell types, the CFU-S may be drawn into the cell cycle and thereby become
susceptible to benzene.
Cronkite et al. (1985) also noted that animals exposed to 300 ppm (958 mg/m3) benzene
for 16 weeks began to die from thymic and nonthymic lymphomata and solid tumors, beginning
at 330 days of age. In a continuation of this study, Cronkite et al. (1989) (1) evaluated the
effects of benzene on erythrocyte progenitors (BFU-E, CFU-E) and on neutrophil and
macrophage progenitors (GM-CFU-C), (2) compared the effects of two different exposure
regimens on peripheral blood cell counts and progenitor cell numbers, (3) evaluated mortality
and neoplasia, and (4) tested the functional capacity of the stem cells using a "rescue assay."
Hale-Stoner mice, aged 12-14 weeks, were used in the progenitor cell assays. Bone marrow cell
suspensions harvested from mice exposed to 400 ppm (1278 mg/m3) benzene 6 hours/day, 5
days/week for 9.5 weeks and from their corresponding controls were counted and incubated to
form plasma clots. The cells used in the CFU-E assay were cultured in 2% fetal calf serum; the
cells used in the BFU-E assay were cultured in erythropoietin and pokeweed-mitogen
conditioned medium. The clots were scored for erythrocytic colonies (CFU-E) or
megakaryocytes (BFU-E). The cells used in the CFU-G and GM-CFU assays were cultured in
agar for subsequent scoring. Progenitor cell assays resulted in the following: CFU-E numbers
did not change significantly after 4 days of exposures, but the numbers had decreased
significantly (to approximately 45-70% of control values) by assay days 29, 48, and 65. Twelve
days after exposure ended, the cell numbers had recovered to 200% of control values. The BFU-
E values also decreased significantly (to just a few percentage points of control values) at 29, 48,
and 65 days, but they had recovered to only about 40% of control values on day 12 after
exposure. During the time of this assay, the peripheral RBC counts were decreased to about
50-75% of control values.
Granulocyte/macrophage aggregates, also observed in the clots, were decreased to about
10-15% of control values by day 29 and had recovered to about 85% of control values by 12
days after exposure. The GM-CFU cells in an agar assay decreased to approximately 40% of
control values after 30 days of exposure, increased to about 75% by termination of exposure at
day 65, decreased to approximately 60% by 72 days, and then recovered to almost 80% by 80
days. The numbers of granulocytes in the blood also fluctuated during this time.
Influence of exposure protocols. CBA/Ca BNL male mice were exposed to either 316
ppm (1010 mg/m3) benzene administered 6 hours/day, 5 days/week for 19 exposures or to 3000
ppm (9584 mg/m3) benzene administered 6 hours/day for 2 successive days (Cronkite et al.,
1989). After 19 exposures, 316 ppm (1010 mg/m3) benzene had reduced lymphocyte counts
from 7500/|J,L to 300/|J,L, and 3000 ppm (9584 mg/m3) for 2 days had resulted in a reduction
from 6700/|a,L to 3300/|o,L 1 day after the end of the exposure. The differences in the effects of
69
-------
the two regimens still persisted at 214 days after exposure. Neutrophil counts of the group
exposed to 316 ppm (1010 mg/m3) were also depressed more than those exposed to 3000 ppm
(9584 mg/m3). By 214 days after exposure, the neutrophil counts in the group treated with 3000
ppm (9584 mg/m3) had recovered, whereas those in the group treated with 300 ppm (958 mg/m3)
had not. All leukocyte cell types were decreased in number as compared with those in sham-
exposed mice on day 1, and all absolute counts except for the large unstained cells (mostly large
lymphocytes) were depressed on 32, 67, and 214 days after the last exposure. These data were
not given for the animals exposed to 3000 ppm (9584 mg/m3).
On day 1 after exposure, both treatments had induced a reduction in bone marrow
cellularity; this was significantly greater in the animals exposed to 316 ppm (1010 mg/m3) for 19
exposures (depression to -33% of control value) than in those exposed to 3000 ppm (9584
mg/m3) for 2 days (depression to -59% of control value). Cellularity had returned to control
values by 32, 67, and 214 days after exposure. The CFU-S values followed the same response
pattern except that by day 214 after exposure recovery was incomplete for the group exposed to
316 ppm (958 mg/m3).
In another experiment, exposure to 300 ppm (958 mg/m3) benzene for 80 weeks produced
a severe depression in CFU-S (to 38% of control) that had not completely recovered 178 days
after exposure ended. Exposure to 3000 ppm (9584 mg/m3) for 8 weeks produced a milder
depression in CFU-S (to 62% of control) that had recovered by 30 days. Decreases in bone
marrow cellularity were nearly the same for the two exposure scenarios.
Rescue assay. The injection of normal bone marrow cells into lethally irradiated rodents
enhances their survival (Cronkite et al., 1989). The serial rescue assay tested the effect of
benzene on this particular function of the bone marrow. C57BL/6BNL male and female mice
were irradiated with 850 rad, then injected with bone marrow cells from mice that had been
exposed to 300 ppm (958 mg/m3) benzene 6 hours/day, 5 days/week for 16 weeks 752 days
earlier. Thirty days after the initial transplants, some of the survivors' marrow was injected into
another group of irradiated mice for a secondary transplant, and 30 days later those recipients
served as donors in the third transplant to still another group of irradiated mice. The recipients
of bone marrow in each transplantation of the series were observed for up to 594 days. In the
first rescue, the 30-day survival was 100%. At 290 days, survival of the mice that received cells
from control mice was 93%, and survival of the mice that received cells from benzene-exposed
mice was 64%.
In the secondary transplant, no mice died by 30 days, but the mice that received control
cells started dying at day 171 and mice that received cells from donors exposed to benzene
started to die at day 70. In the tertiary rescue, 50% of the recipients of bone marrow from
benzene-exposed donors were dead by day 30 and 95% were dead by day 245. Among the
animals that received cells from the sham-exposed mice, only one died by day 30 and 50% died
by day 375. The last animal died on day 594. The mechanism for the stem cell malfunction that
led to increased mortality is not clear, but the investigators suggest that the observation could be
associated either with genetic injury to the stem cells, exhaustion of the G0 hematopoietic stem
cells, or a radiation effect on the stroma of the recipient.
70
-------
BDF1 mice were exposed to 0, 100, 300, or 900 ppm (0, 319, 958, or 2875 mg/m3)
benzene for up to 4 weeks (Seidel et al., 1990). The numbers of hematopoietic progenitor cells,
early and late progenitors (BFU-E, CFU-E), and granuloid progenitors (CFU-C) were
determined. A group was generated to establish the effect of ethanol (drinking water) on these
effects. This study demonstrated that benzene decreased the number of CFU-E per femur in a
concentration-dependent manner. This effect was evident at 300 and 900 ppm (958 and 2875
mg/m3) concentration; however, the effect in the 100 ppm (319 mg/m3) exposure group was
uncertain, as the study focused on the effects of ethanol on benzene toxicity. The
LOAEL/NOAEL was thus difficult to determine.
Male Sprague-Dawley rats (40/group) were exposed to vapor concentrations of 0 or 100
ppm (0 or 319 mg/m3) benzene 6 hours/day, 5 days/week for life (American Petroleum Institute,
1983). Blood samples were obtained at 2- to 4-week intervals throughout the treatment period.
The erythrocyte and lymphocyte counts were depressed at nearly every sampling time in treated
rats, but the extent of decrease was rarely statistically significant. Nonetheless, the frequencies
of depression were highly significant for each cell type (p<0.002, Wilcoxon Rank Sum Test),
and the overall results were interpreted as evidence of hematotoxicity. Significantly increased
incidences of splenic hyperplasia (p<0.005) and hemosiderin pigments (p<0.001) were observed
in benzene-exposed rats. The incidences of normally rare tumors in treated rats were 4/40 in
liver, 2/40 in Zymbal gland, and 1/40 chronic myelogenous leukemias. The authors considered
these tumors to be related to the benzene exposure. This study identifies a LOAEL of 100 ppm
(319 mg/m3) for hematologic effects in rats.
Decreases in bone marrow cellularity in the femur, HCT, and leukocytes were seen in
DBA/2 mice (20/group) exposed to 300 ppm (958 mg/m3) benzene in air 6 hours/day, 5
days/week for 2 weeks (Chertkov et al., 1992). In most cases no erythroid or myeloid
clonogenic cells could be recovered in bone marrow cultures started after the last benzene
exposure. After 2 weeks of recovery, however, body weight, HCT, bone marrow cellularity, and
committed hematopoietic progenitor cells had recovered to near normal values.
Neun et al. (1992, 1994) compared the in vivo sensitivity of Swiss-Webster and
C57B1/6J mice to inhaled benzene with the in vitro sensitivity of culture bone marrow cells to
benzene metabolites. Mice were exposed to 300 ppm (958 mg/m3) benzene 6 hours/day, 4
days/week for 2 weeks. Swiss-Webster mice were more sensitive to benzene exposure, as
indicated by much greater reductions in femoral bone marrow cellularity and in the number of
CFU-E per femur after in vivo exposure. Neither phenol nor muconic acid were toxic to cultured
CFU-E from either mouse strain. CFU-E from Swiss-Webster mice were more sensitive than
CFU-E from C57B1/6J mice to 1,4-benzoquinone or hydroquinone. Thus, both in vitro and in
vivo data indicated that Swiss-Webster mice were more sensitive to benzene toxicity than were
C57Bl/6Jmice.
Female BDF1 mice (C57BL/6 x DBA/2F1 hybrids) were exposed to 0, 100, 300, or 900
ppm (0, 319, 958, or 2875 mg/m3) benzene 6 hours/day, 5 days/week for up to 8 weeks (Plappert
et al., 1994). Hematologic studies included peripheral blood data, T4 and T8 lymphocyte counts
in the blood and spleen, and hematopoietic stem and progenitor cell assays in the marrow CFU-
S, CFU-C, and BFU-E, CFU-E. No significant changes were observed in the peripheral blood
data of mice exposed to 900 ppm (2875 mg/m3) benzene. Some perturbation of the reticulocyte
71
-------
numbers was observed, but values at days 3 and 5 did not differ from controls. Absolute
numbers of lymphocytes and neutrophils did not differ from controls (no data shown). Slight
anemia was observed at 4 and 8 weeks of treatment with 300 and 900 ppm (958 and 2875
mg/m3) benzene. Minor changes occurred in the stem and progenitor cells. CFU-E depression
after 4 days of exposure was significant. A dose-dependent depression of colony-forming cell
number appeared at 4 and 8 weeks of exposure, with maximal effect at the level of CFU-E.
Farris et al. (1997a) exposed male B6C3F1 mice to 0, 1, 10, 100, or 200 ppm (0, 3.2, 32,
319, or 639 mg/m3) benzene 6 hours/day, 5 days/week for 1, 2, 4, or 8 weeks. A separate 4-week
experiment at benzene concentrations of 0, 1, 5, and 10 ppm (0, 3.2, 16, and 32 mg/m3) was also
conducted. Another group of animals was exposed to benzene for 4 weeks and then allowed to
recover for up to 25 days. Bone marrow cell counts, high proliferative potential colony-forming
unit (CFU-HPP), percentage of CFU-HPP cells in S-phase, bone marrow CFU-E, bone marrow
GM-CFU, labeling index, B and T lymphocytes, and hematology parameters were determined.
Data sets for each parameter were individually evaluated with a univariate two-way ANOVA
using contrasts to determine treatment effect. Statistical significance was determined at a level
of/X0.05.
There were no significant effects on hematopoietic parameters from exposure to 10 ppm
(32 mg/m3) benzene or less. Exposure of mice to 100 or 200 ppm (320 or 640 mg/m3) reduced
the number of total bone marrow cells, progenitor cells, differentiating hematopoietic cells, and
most blood parameters, with a concentration-related effect. Replication of primitive progenitor
cells in the bone marrow was increased during the exposure period, possibly in compensation for
benzene cytotoxicity. Recovery after benzene exposure was rapid. Most hematopoietic
parameters returned to control levels within 4 days following the 100 ppm (320 mg/m3) exposure
and within 11 days following the 200 ppm (640 mg/m3) exposure. The percentage of CFU-HPP
in S-phase was elevated throughout the 25-day recovery period evaluated. Farris et al. (1996)
reported that the frequency of micronucleated erythrocytes was increased in B6C3F1 male mice
exposed to 100 and 200 ppm (320 and 640 mg/m3) benzene for 8 weeks. The authors suggested
that the increased proliferation of primitive progenitor cells, in concert with genetic damage,
provides the components for producing an increased incidence of lymphoma in mice. The
effects on hematopoietic parameters indicate a LOAEL of 100 ppm (320 mg/m3) and a NOAEL
of 10 ppm (32 mg/m3).
Farris et al. (1997b) exposed male B6C3F1 mice to 0, 1, 10, 100, or 200 ppm (0, 3.2, 32,
319, or 639 mg/m3) benzene 6 hours/day, 5 days/week for 1, 2, 4, or 8 weeks. A separate 4-week
experiment at benzene concentrations of 0, 1, 5, and 10 ppm (0, 3.2, 16, and 32 mg/m3) was also
conducted. Spleen and thymic lymphocyte counts were determined for three mice in each
treatment. Femoral B lymphocyte counts were evaluated in 10 mice per group. Labeling index
of bone marrow B lymphocytes was determined by BrdU incorporation. To minimize the
potential for recovery, mice were sacrificed and sampled within 2 hours of the end of the last
exposure. The spleen, thymus, and bone marrow data sets were individually evaluated with a
univariate two-way ANOVA using contrasts to determine treatment effect within each time
point. Statistical significance was determined at a level ofp<0.05.
Exposure to 100 or 200 ppm (319 or 639 mg/m3) benzene induced rapid and persistent
reductions in femoral B, splenic T and B, and thymic T lymphocytes; total nucleated thymus
72
-------
cells count; and total nucleated spleen cells. There was a significant decrease in thymic cell
count and B lymphocytes in the spleen at week 2 in mice exposed to 10 ppm (32 mg/m3) in the
8-week study and no effects at 1 ppm (3.2 mg/m3). These parameters, however, were
comparable with the controls by week 4 of the 8-week exposure experiment, and in the separate
4-week experiment there were no significant decreases in these parameters at any time in the
mice exposed to 10 ppm (32 mg/m3) benzene or less. Thus, the 100 ppm (319 mg/m3) level was
considered a LOAEL and 10 ppm (32 mg/m3) a NOAEL for these effects. The percentage of
femoral B lymphocytes and thymic T lymphocytes in apoptosis was increased 6- to 15-fold by
200 ppm (639 mg/m3). Replication of femoral B lymphocytes was increased during the exposure
period in the bone marrow, apparently in compensation for the lymphocyte loss induced by 100
or 200 ppm (319 or 639 mg/m3) benzene exposure.
Robinson and coworkers (1997) exposed male Sprague-Dawley rats (16/group) to 0, 30,
200, or 400 ppm (0, 96, 639, or 1278 mg/m3) benzene 6 hours/day, 5 days/week for either 2 or 4
weeks. Of the 16 animals in each group, 8 were challenged intravenously with sheep red blood
cells (SRBCs) at 4 days before the end of the exposure to allow an assessment of humoral
immunocompetence using the enzyme-linked immunosorbent assay technique. Thymus and
spleen weights were determined, as were total cell counts for the spleen and femur. Spleen
CD4+/CD5+, CD8VCD5+, and Kappa+ lymphocytes were determined by immunostaining.
Statistical evaluations were performed using Dunnett's test.
Total spleen cell counts were significantly reduced (29%) in animals exposed to 400 ppm
(1278 mg/m3) after 4 weeks, and thymus weight was also significantly reduced (28%). The
spleen weight and cellularity were comparable with the controls at both 2 and 4 weeks in the 200
ppm (639 mg/m3) exposure group. After 4 weeks at 400 ppm (1278 mg/m3) there was a
significant reduction in spleen B, CD4+/CD5", and CD5+ T lymphocytes. Rats exposed to 30,
200, or 400 ppm (99, 639, or 1278 mg/m3) benzene for 2 or 4 weeks and then challenged with
SRBCs developed a humoral response comparable to the controls', and only rats exposed to 400
ppm (1278 mg/m3) for 2 weeks showed a significant reduction in spleen B lymphocytes
(Robinson et al., 1997). These results indicate that 400 ppm (1278 mg/m3) is a LOAEL and 200
ppm (639 mg/m3) is a NOAEL for immunotoxicity in rats.
Taken as a group, the results demonstrate that the reduction in bone marrow function
following benzene exposure is related to decreases in the growth and maturation of pluripotential
stem cells (CFU-S), lineage-restricted stem cells (e.g., GM-CFU), progenitor cells in various
stages of maturation, and the stromal cells that provide growth factors necessary for bone
marrow function. These studies tend to demonstrate the sensitivity of erythropoiesis and
lymphopoiesis to benzene and the relative resistance of granulopoiesis to benzene.
Snyder and Kalf (1994) reviewed the effects of benzene on bone marrow cells. Bone
marrow stromal cells, which include leukocytes, erythrocytes, endothelial cells, reticular cells,
and fat cells, constitute the hematopoietic connective tissue in which stem cells undergo
maturation and amplification to form the cells of the circulation. Garnett et al. (1983) reported
that the marrow adherent layer from benzene-treated animals had an altered morphology and
failed to support the differentiation of stem cells. In a related study, Kalf et al. (1996)
demonstrated that/?-benzoquinone, an important reactive metabolite of benzene, prevents the
conversion of pre-interleukins-lcc and -ip to the active cytokines by inhibiting calpain and the
73
-------
interleukin-lp converting enzyme, respectively. Thus, a key benzene metabolite inhibits the
formation and release of an important cytokine, IL-1, from stromal macrophages.
Irons et al. (1992) isolated human CD34+ human bone marrow myeloid progenitor cells,
which proliferate in response to IL-3 and GM-CSF. Addition of hydroquinone, within a limited
range of concentrations, resulted in synergistic enhancement of colony formation. Hazel et al.
(1996) made similar observations using mouse bone marrow CD32 cells. The mechanism is
thought to include the up-regulation of the 5-lipoxygenase pathway in these cells, with increases
in the formation of LTD4, which in turn reacts with the LTD4 receptor to increase cell
replication. Hydroquinone independently reacts with the LTD4 receptor to also enhance cell
replication. Morphological evaluation suggests that maturation of the myeloblast stage to the
myelocyte stage has been enhanced. These cells remain in their immature state in the presence
of hydroquinone and do not advance to mature neutrophils, but proliferation continues.
The implications of these observations are significant in terms of a mechanism by which
benzene can cause expansion of the pool of myelocytes, which do not then undergo either further
differentiation to neutrophils or apoptosis, which is the usual method used by the bone marrow to
control the size of the myelocyte pool. This might provide a pathway for a benzene-induced
leukemogenic response. It may be hypothesized that if benzene, acting via DNA adduct
formation or oxidative damage to DNA, causes cell transformation at an earlier stage of
differentiation, the creation of an expanding pool of myelocytes by benzene and hydroquinone
provides a mechanism for promotion. Thus, the criteria for a two-stage mechanism of
carcinogenesis (i.e., initiation followed by promotion) are satisfied. However, the specific issue
of progression has not been addressed by studies of this type.
The studies by Tennant et al. (1995) have demonstrated the production of a granulocytic
leukemia in TgAC mice when treated with benzene dermally but not when given benzene by
gavage. This is the first animal model that we have had for a benzene-induced granulocytic
leukemia. The TgAC mouse is derived from the FVB mouse in which the viral Ha-ras gene has
been inserted into chromosome 8. The granulocytic leukemia develops within 6 months. TgAC
mice may represent initiated animals in which benzene and hydroquinone act as promoting
agents leading to proliferation of the leukemia cells, and thus could support the argument for the
two-stage model for benzene-induced leukemogenesis. It is significant that in the course of
these studies bone marrow depression was observed, indicating the close relationship between
noncancer and cancer endpoints in benzene exposure.
The following discussion offers some thoughts about the attempt to understand the
mechanism by which DNA damage occurs as a result of benzene exposure. DNA damage is not
extensively dealt with here because this document concentrates on noncancer endpoints. It is not
clear, however, that there is no impact on noncancer endpoints by DNA damage. Some
discussion of DNA damage leading to translocation and gross breaks as well as a discussion of
DNA repair and the impact of benzene metabolites is warranted. The study by Snyder and Kalf
(1994) and the more recent study by Smith et al. (1998) relate to benzene-induced chromosome
damage. On the issue of DNA damage induced by benzene, there is ample evidence to show that
benzene metabolites covalently bind to protein and DNA. Also, the structures of putative DNA
adducts are known. An alternative mechanism for DNA damage is via the generation of reactive
oxygen species. For example, /?-benzoquinone and hydroquinone increase superoxide, nitric
74
-------
oxide, and hydrogen peroxide in HL-60 cells activated with phorbol ester. Boersma et al. (1994)
suggested that hydroquinone itself is unlikely to generate reactive oxygen species.
An interesting complimentary alternative mechanism, proposed by Singh et al. (1994),
suggests that benzene- induced DNA damage is mediated by (1) release of free iron in the bone
marrow of benzene-treated animals (probably by polyphenolic metabolites of benzene), followed
by (2) the chelation of iron by hydroquinone or benzenetriol to yield (3) a reactive oxygen-
generating species such as superoxide, which in turn (4) causes oxidative damage to DNA.
Their suggestion that glutathionyl-hydroquinone may be a key intermediate resonates well with
the suggestion of Brunmark and Cadenas (1988). In a radical-rich system, glutathionyl-
hydroquinone may well be converted to glutathionyl-benzenetriol. Singh et al. (1994) suggest a
chemical structure for the iron-benzenetriol chelate in which the iron is bound between the
hydroxyl group at positions 1 and 2 on 1,2,4-benzenetriol with the release of two protons.
Indeed, it may well be that the auto-oxidation of glutathionyl-benzenetriol is enhanced with an
iron chelated between positions 1 and 2. These hypotheses can be reconciled by assuming that
the superoxide produced by the auto-oxidation of gluthionyl-benzenetriol is converted to
hydroxyl radical and singlet oxygen by the benzenetriol-iron chelate. The resulting reactive
oxygen species can hydroxylate guanine and other DNA bases, resulting in a mutagenic effect.
The discussion of the mechanisms of benzene toxicity provides a framework for
understanding the total disease process. It has become clear that benzene
toxicity/leukemogenesis is a continuity of effects. Although this is not intended to be a
discussion of the carcinogenic effects of benzene, all of the effects on the bone marrow caused
by benzene appear to be time- and dose-related in a sequential manner. The literature suggests
that chronic low doses lead to cytopenias, which may be reversible if exposure ceases. At
intermediate chronic doses cytopenias persist and may eventually lead to myelodysplastic
syndrome (MDS), which is a preleukemic state terminating in AML. If the cytopenias are
sufficiently severe the patient may succumb to an infection due to failure of the immune system.
Higher doses may lead directly to bone marrow aplasia and death without the characteristic
appearance of MDS. Death is usually the result of immune deficiency. Thus, the cytopenias (the
ultimate form of which is aplastic anemia) leading to decreases in circulating cell numbers and
immune impairment are not separable from MDS and, eventually, AML.
The following studies of exposure to benzene and its metabolites are considered briefly
because of their value in the study of the mechanisms of benzene toxicity.
Tunek et al. (1982) examined the hematotoxicity of benzene and two of its phenolic
metabolites, hydroquinone and catechol, in male NMRI mice. Daily doses of 440 mg/kg
benzene to body weight were injected subcutaneously for 6 days, and CFU-C/tibia, cells/tibia,
and micronuclei/2000 PCE were assayed on the day after the last injection. Micronuclei
appeared after one injection, peaked after three injections, then decreased with injections 4, 5,
and 6. The investigators presumed that the numbers of micronuclei decreased as a result of
toxicity and cell death before the cells reached the PCE stage. By day 4, the CFU-C/tibia and
number of cells/tibia were reduced to about 5-10% of control values.
In another experiment, the animals received doses ranging from 0.7 to 440 mg/kg/day
benzene on 6 consecutive days and were sacrificed on day 7. Increased levels of micronuclei
75
-------
were observed at doses above 28 mg/kg/day, and their frequency increased dramatically at doses
above 55 mg/kg/day. At -100 mg/kg/day, micronuclei levels were about 50/2000 PCE, whereas
at the highest dose of 440 mg/kg/day, they were 218/2000 PCE (vs. 3 micronuclei/2000 PCE for
controls). CFU-C/tibia and bone marrow cellularity were suppressed by about 40% at doses as
low as 3.5 mg/kg/day. At 100-440 mg/kg/day, both values dropped to 5-10% of control values.
Six injected doses of hydroquinone ranging from 20 to 100 mg/kg produced micronuclei,
with a sharp increase occurring at 6 x 80 mg/kg. Bone marrow cellularity was slightly elevated
at low doses and was suppressed at higher doses, with the number of CFU-C following a similar
pattern. Catechol, injected at doses ranging from 5 to 42 mg/kg for 6 consecutive days, had no
effect on any of the three parameters examined. The toxic effects of benzene were alleviated by
-30-40% when toluene (which competitively inhibits benzene metabolism) was injected
simultaneously. The toxic effects of hydroquinone were somewhat reduced by simultaneous
administration of toluene, but not nearly to the extent observed with benzene. The investigators
calculated that a dose of 440 mg/kg benzene would yield an excretion of -20 mg/kg
hydroquinone. However, 20 mg/kg hydroquinone was virtually nontoxic, whereas 440 mg/kg
benzene produced severe hematologic effects. On the basis of these differences and the
differences in the sequence of the responses to the two chemicals, Tunek et al. (1981) speculated
that the hematotoxicity of benzene at low doses is due to agents other than hydroquinone, but
that effects at and above the threshold dose result from the metabolic formation and
accumulation of hydroquinone.
Using an iron-uptake method, Snyder et al. (1989) demonstrated that i.p. administration
of benzene, hydroquinone, />-benzoquinone, and MUC, singly and in combination with each
other, inhibited erythropoiesis in female mice. The combination of hydroquinone plus MUC was
most effective in decreasing iron uptake. Toluene alleviated the effects of benzene but not the
effects of hydroquinone or/?-benzoquinone.
MacEachern et al. (1992) evaluated the effects of benzene on the morphology and
function of bone marrow phagocytes. Male Balb/c mice received i.p. injections of either
benzene (660 mg/kg) or a combination of hydroquinone and phenol (50 mg/kg each) once per
day for 3 days. Control animals received corn oil or phosphate-buffered saline. The animals
were sacrificed after the last injection and the following assays were performed: characterization
and quantitation of subpopulations of bone marrow cells using monoclonal antibody techniques,
measurement of chemotaxis (the process by which activated phagocytes migrate to an injured
site in a tissue), and measurement of the oxidative metabolism of phagocytes (to evaluate
maturation and activation of phagocytes). The bone marrow contained three distinct populations
of cells: population 1—a larger more dense population (41%); population 2—a population of
intermediate size and density (23%); and population 3—a smaller, less dense population (33%).
Population 1 consisted of 85-90% granulocytes, including neutrophils, basophils, and
eosinophils; population 2 contained a mixture of mononuclear phagocytes (35-40%) and
immature precursor cells (55-65%); and population 3 contained lymphocytes (86%) and
immature precursor cells (14%). This pattern of distribution is similar to that observed for
human peripheral blood leukocytes and bone marrow cells (Landay and Bauer, 1988; Lund-
Johansen et al., 1990, both cited in MacEachern et al., 1992). Differential staining of sorted
mononuclear phagocytes revealed an increase in the number of mature, morphologically
activated macrophages in the bone marrow of benzene-treated mice.
76
-------
Granulocytes and mature macrophages were the only cells to migrate toward
chemoattractants C5a or TPA, respectively, and cells from benzene-treated mice exhibited
increased chemotactic activity when compared with controls (p<0.02). The migration of bone
marrow granulocytes from mice treated with hydroquinone and phenol toward C5a was
depressed (p<0.02), whereas the migration of the macrophages from these mice toward TPA was
enhanced (p<0.02).
Benzene treatment also induced a significant increase in basal oxidative metabolism in
bone marrow granulocytes (population 1), but treatment with phenol and hydroquinone did not.
The investigators concluded that phagocytes and granulocytes from the bone marrow of
benzene-treated mice are activated and stimulated to differentiate. Further studies are in
progress to determine whether these phagocytes or their mediators contribute to the development
of benzene hematotoxicity.
Gaido and Wierda (1984) demonstrated that in vitro treatment with hydroquinone andp-
benzoquinone decreased the ability of stromal cells to support granulocyte/macrophage colony
formation. Catechol and 1,2,4-benzotriol inhibited colony formation, but only at very high
concentrations.
Eastmond et al. (1987) demonstrated that coadministration of 75 mg/kg phenol and
25-75 mg/kg hydroquinone 2 times/day for 12 days i.p. to mice resulted in a significant, dose-
related depression in bone marrow cellularity. The compounds alone produced neither a
significant nor a dose-related response. Catechol had no effect, either alone or in combination.
Hydroquinone (10 mg/kg/day), benzoquinone (2 mg/kg/day), or benzenetriol (6.25
mg/kg/day) administered i.p. to rats produced significant decreases in bone marrow cell counts,
RBCs, and Hgb (Rao et al., 1988). Thms^ram'-muconaldehyde administered to CD-I mice i.p.
daily for 10 and 16 days produced hematotoxicity similar to that of benzene (Witz et al., 1985).
Benzene metabolites added to macrophage cultures significantly and selectively inhibited
macrophage function (Lewis et al., 1988).
Irons et al. (1992) observed an enhanced colony-forming response by mouse bone
marrow cells treated in vitro, first with hydroquinone and then with recombinant
granulocyte/macrophage colony-stimulating factor. However, treatment with phenol, catechol,
or MUC plus the colony-stimulating factor did not enhance colony formation. The combination
of hydroquinone and the stimulating factor also appeared to recruit a segment of the myeloid
progenitor cell population that was not normally responsive to the stimulating factor. These
alterations in the differentiation of the myeloid progenitor cell population may be relevant in the
pathogenesis of chemically induced AML.
4.2.1.6. Summary of Principal Hematotoxic Effects
A considerable number of studies in experimental animals have pointed to the changes in
peripheral blood and bone marrow induced by benzene as being among the most sensitive effects
of the compound's toxicity. That these effects were seen in most, if not all, of the species tested,
as well as in humans occupationally exposed to benzene (see Section 4.1.2.1.), suggests that their
threshold dosimetry may be important in identifying a point of departure for the derivation of
77
-------
quantitative risk estimates. Principal among the adverse effects described are leukopenia,
lymphocytopenia, granulocytosis, anemia, and reticulocytosis. Comparing the incidence of one
response with another, lymphocyte counts appear to be depressed sooner and more severely than
other cell types, and granulocytes may be the most resistant (Snyder et al., 1978, 1980).
Differences in responsiveness to benzene have been observed among species, with mice being
more sensitive than rats to the hematotoxicity of benzene (Ward et al., 1985; Snyder et al., 1978,
1980). Intraspecies variability was demonstrated by Snyder et al. (1978, 1980), who reported
that AKR/J mice were more sensitive than C57BL/6 mice to the compound. Ward et al. (1985)
showed that in CD-I mice, males were more sensitive than females.
A striking feature of the studies of Ward et al. (1985) and Cronkite et al. (1985) was that
hematologic abnormalities were evident after only 2 weeks of exposure, with comparatively little
progression during the rest of the exposure period. However, the results of Cronkite et al. (1985,
1989) point to the possibility that peripheral blood elements may have the capacity to recover
from benzene-induced toxicity, depending on the duration of exposure and concentration. For
example, lymphocyte counts, which were depressed in C57BL/6 mice exposed to benzene at 300
ppm (32-1279 mg/m3) for 2-16 weeks, tended toward normal values during a 4- to 8-week
recovery period.
Dose-response studies such as that of Ward et al. (1985) have served as a basis for the
identification of NOAELs and LOAELs. For example, a NOAEL of 30 ppm (96 mg/m3) and a
LOAEL of 300 ppm (958 mg/m3) have emerged for CD-I mice from the Ward et al. study. The
data from this study have also served as a basis for BMD modeling. Using the BMDL derived
from this modeling as a point of departure for RfC determination yields a value that is in good
agreement with the RfC obtained from the hematotoxicity data from the occupational exposure
study of Rothman et al. (1996a), as described in Section 5.1.
Benzene-induced peripheral blood abnormalities reflect a disruption of all levels of
hematopoiesis in the bone marrow. For example, several studies have demonstrated
concentration-dependent depression of bone marrow cellularity in mice treated with benzene, the
lowest concentration producing this effect being 21 ppm administered for 2 weeks (Toft et al.,
1982). Similarly, the stem cell compartment (CFU-S) appears to be sensitive to the adverse
effects of benzene, particularly when drawn into the cell cycle in response to benzene-induced
damage to other cells (Snyder, 1987). Thus, concentrations of 100 ppm (320 mg/m3) benzene
have produced significant depressions in the CFU-S population in the bone marrow of mice
exposed for 2-4 weeks (Cronkite et al., 1985; Seidel et al., 1989).
The granulocytic and erythropoiesis progenitor cells are also sensitive to the effects of
the compound, exhibiting significant depression at 21 ppm (67 mg/m3) benzene for 2 weeks and
10 ppm (32 mg/m3) for 4 weeks, respectively (Toft et al., 1982; Seidel et al., 1989). In addition,
both the stem cells and the progenitor cells can recover from benzene-induced depression, at
least partially, depending on the concentration and duration of exposure. Cronkite et al. (1989)
observed that depression of CFU-E and BFU-E were paralleled by depressions in peripheral
RBC counts.
Bone marrow macrophages also appear to be a target of benzene toxicity (MacEachern et
al., 1992). Thus, stromal macrophages from benzene- or hydroquinone-treated mice failed to
78
-------
convert IL-1 to the mature cytokine 17-Kd, a component essential for hematopoiesis (Renz and
Kalf, 1991). This allows the suggestion that the inhibitory effect of benzene on lymphocyte
proliferation can be mediated through the suppression of cytokine production.
Another important consideration is whether the benzene-induced perturbation of the
hematopoietic system is mediated through the effects of the parent compound or via one or more
metabolites. The fact that both in vivo and in vitro studies have demonstrated the ability of some
metabolites to induce qualitatively similar toxicological effects to those of benzene provides
evidence consistent with the concept that the hematotoxicity of the compound may be mediated
through its metabolites.
4.2.2. Reproductive/Developmental Effects
4.2.2.1. Reproductive Toxicity
Multiple-generation reproductive toxicity studies of benzene were not found in the
literature. Relevant information pertinent to the reproductive toxicity of benzene in animals is
described below.
4.2.2.1.1. Oral exposure. NTP conducted the 90-day subchronic toxicity and 2-year bioassay
studies of benzene in mice and rats (NTP, 1986). Other than ovarian effects, no signs of
reproductive toxicity were reported. Benzene was administered by gavage to B6C3F1 mice and
F344/N rats. In the 90-day study, the doses of benzene, administered in corn oil 5 days/week,
ranged from 25 to 600 mg/kg; in the 2-year study, doses were 50, 100, or 200 mg/kg for male
rats and 25, 50, or 100 mg/kg for female rats and all mice. There were no ovarian effects in
either the rats or the mice treated for 90 days or in the rats treated for 2 years. Mice treated for 2
years exhibited ovarian lesions ranging from atrophy to neoplasia. However, the incidence of
nonneoplastic lesions was not dose related. The nonneoplastic lesions and their incidence, based
on the total number of animals examined for each group, for mice administered 0, 50, 100, or
200 mg/kg benzene were ovarian atrophy (32, 79, 65, and 46%, respectively) and epithelial
hyperplasia (26, 89, 63, and 60%, respectively).
Spano et al. (1989) examined the cytotoxic action of benzene on mouse germ cells using
flow cytometric DNA content measurements. Testicular monocellular suspensions were
obtained from (C57BL/Cne x C3H/Cne) Fl male mice receiving single doses by gavage of 0, 1,
2, 4, 6, or 7 mL/kg benzene to body weight. The effects were measured in three animals per
group 7, 14, 21, 28, and 70 days after treatment. Testicular cells were classified as mature
haploid, immature haploid (haploid is split into two parts because of different staining on the
elongated and round spermatids), diploid, or tetraploid, depending on DNA content.
Benzene treatment did not affect body weight or testes weight, but it did alter the ratio of
testicular cell types. DNA histograms of mouse testis cells obtained at different times after
benzene exposures showed a dose-related decrease in the tetraploid cell fraction (mainly primary
spermatocytes); 7 days after treatment the tetraploid cell number in animals exposed to 6 and 7
mL/kg benzene was depressed to -80% of control values. The percentage of round spermatids
in animals exposed to 7 mL/kg benzene was also decreased to -80% of control values; however,
in this case the dose relationship was not distinct. These effects indicate cytotoxicity of
79
-------
differentiating spermatogonia. Dose-dependent recovery processes for the tetraploid cells had
begun 21 days after treatment, simultaneously with reduction of other cell subpopulations, but
they were still incomplete by 70 days. Recovery of the round spermatids began by day 28
posttreatment and was complete by day 70. There was also a time- and exposure-dependent
reduction (to 60% of control values for the 4, 6, and 7 mL/kg doses) in the percentage of
elongated spermatids at 28 days after treatment; recovery was complete by 70 days. The
investigators concluded from this study that benzene can induce acute cytotoxicity in mouse
germ cells. Both the NTP (1986) and the Spano et al. (1989) studies provide valuable
descriptive information regarding the effects of benzene on the reproductive system. It should
be noted that longer term and more relevant studies are needed to confirm the conclusion of
Spano etal. (1989).
4.2.2.1.2. Inhalation exposure. The available studies on the reproductive toxicity of inhaled
benzene to test animal species is summarized in Table 7.
Subchronic toxicity studies have provided information relevant to the reproductive
toxicity of benzene. In an early inhalation study, Wolf et al. (1956) demonstrated moderately
increased testicular weights in groups of 10-25 rats exposed to 6600 ppm (21,084 mg/m3)
benzene for 13 weeks but not in rats exposed to 88 ppm (281 mg/m3) for 30 weeks, to 2200 ppm
(7028 mg/m3) for 30 weeks, to 4400 ppm (14,056 mg/m3) for 5 weeks, or to 9400 ppm (30,030
mg/m3) for more than 1-19 days. Groups of 5-10 male guinea pigs exposed to 88 ppm (281
mg/m3) benzene for -9.6 months had a slight increase in testicular weights, but guinea pigs
Table 7. Reproductive toxicity of inhaled benzene in test animals
Species
Male rat
Rabbit
Male
guinea pig
Male and
female
Sprague-
Dawley
rats
No./
group
10-25
1-2
5-10
50/sex
Treatment
2 1,084 mg/m3, 7
hrs/day, 5 days/wk
for 13 weeks
256 mg/m3, 7
hrs/day, 5 day/wk
for —8.5 months
281 mg/m3, 7
hrs/day, 5 days/wk
for— 9.6 months
0, 3.2, 32, 96, 958
mg/m3 6 hrs/day, 5
days/wk for 13
weeks;
10/sex/group
sacrificed after 7,
14, 28, 56, and 91
days of treatment
Reproductive effects
Moderate changes in testes
weight (no quantitative
data)
Slight change in
histopathology of the testes
(no details)
Slight change in weight of
testes (no quantitative data)
No microscopic lesions of
ovaries and testes
LOAEL/
NOAEL
(mg/m3)
LOAEL: NA
NOAEL: NA
LOAEL: NA
NOAEL: NA
LOAEL: NA
NOAEL: NA
LOAEL: 958a
NOAEL: 96
Reference
Wolf etal.,
1956
Wolf etal.,
1956
Wolf etal.,
1956
Ward etal.,
1985
80
-------
Species
CD-I
mouse
Female
Sprague-
Dawley rat
No./
group
150/sex
26
Treatment
0, 3.2, 32, 96, 958
mg/m3 6 hrs/day, 5
days/wk for 13
weeks;
30/sex/group
sacrificed after 7,
14, 28, 56, and 91
days of treatment
0, 3.2, 32, 96, 958
mg/m3 6 hrs/day, 5
days/wk for a 10-
week period before
and during mating,
during gestation to
GD 20, and from
day 5 to day 21 of
lactation
Reproductive effects
Cystic ovaries (4/10),
testicular atrophy (7/10),
decreased sperm count
(6/10), and increased
abnormal sperm forms
(9/10) only at 958 mg/m3
for 91 days
Reduction in body and liver
weight in the female pups at
958 mg/m3 (p<0.05); no
effect on maternal
mortality, body weight,
physical parameters,
pregnancy rate, length of
gestation, number of live
and dead pups at birth, and
sex distribution data; no
effect on pup survival and
growth, and gross
postmortem manifestations
LOAEL/
NOAEL
(mg/m3)
LOAEL: 958
NOAEL: 96
LOAEL: 958
NOAEL: 96
Reference
Wardetal.,
1985
Kuna et al.,
1992
aEffects on hematologic parameters place the LOAEL for the overall study in rats at 958 mg/m3 and the NOAEL at
96 mg/m3.
GD = gestation day.
NA = not applicable (only one concentration tested).
81
-------
exposed to 88 ppm (281 mg/m3) for 4 weeks did not. Groups of 1-2 rabbits exposed to benzene
concentrations of 88 ppm (281 mg/m3) for -8.5 months had slight histopathologic changes in the
testes, described by the authors as degeneration of the seminiferous tubules (no other
concentrations were tested in rabbits). Exposures were routinely 7 to 8 hours/day, 5 days/week.
Air-exposed or unexposed animals served as controls. The vapor concentrations were
maintained within 10% of the desired concentration. No further details were given for the results
of this study, and quantitation was not possible. Although the study was not specifically
designed to detect reproductive effects, the gonadal effects noted stimulated further interest in
the potential reproductive effects of benzene.
Ward et al. (1985) exposed male and female CD-I mice and Sprague-Dawley rats to
benzene concentrations of 1, 10, 30, or 300 ppm (3.2, 32, 96, or 958 mg/m3) 6 hours/day, 5
days/week for 13 weeks. Animals exposed to filtered air served as controls. Groups of 10 males
and 10 females per species were sacrificed after 7, 14, 28, 56, and 91 days of treatment.
Endpoints of the investigation that were relevant to reproductive toxicity included body and
organ weights and gross and microscopic pathology. The mice exposed to 300 ppm (958 mg/m3)
benzene exhibited statistically significant decreases in testes weights at days 28, 56, and 91 and
decreased testes/body weight ratios on days 56 and 91 (data not given). Testicular weight for the
rats was comparable with control values. The mice exposed to 300 ppm (958 mg/m3) for 91 days
had testicular lesions that included minimal to moderately severe bilateral atrophy and
degeneration (7/10 mice), moderate to moderately severe decreases in spermatozoa in the
epididymal ducts (6/10 mice), and minimal to moderate increases in abnormal sperm forms
(9/10). Four of 10 female mice had bilateral ovarian cysts after 91 days of exposure to 300 ppm
(958 mg/m3) benzene. These testicular and ovarian lesions were not observed at earlier
sacrifices, but similar lesions did appear at lower concentrations; however, biological
significance is questionable. Microscopic findings in the control animals were not mentioned.
The rats had no signs of reproductive toxicity; however, hematotoxicity occurred at 300
ppm (958 mg/m3), giving the overall study a LOAEL of 300 ppm (958 mg/m3) and a NOAEL of
30 ppm (96 mg/m3). The experiment appeared to be a carefully performed subchronic toxicity
study, and the appearance of microscopic changes in the testes and ovaries of the 300 ppm (958
mg/m3) group and similar—but only occasional—findings at 10 and 30 ppm groups suggest a
concentration response.
Kuna et al. (1992) assessed the effects of benzene on female fertility in Sprague-Dawley
rats. The study tested occupational exposure levels as well as higher exposures that previously
demonstrated developmental toxicity. Groups of 26 Sprague-Dawley female rats inhaled
benzene vapor concentrations of 0, 1, 30, and 300 ppm (0, 3.2, 96, and 958 mg/m3) benzene
(purity, 99.96%) 6 hours/day, 5 days/week for a 10-week premating and mating period, and daily
from gestation days (GDs) 0 to 20 and lactation days 5 to 20. Daily vaginal smears were
examined to determine whether estrus was affected by treatment.
Neither strain had any treatment-related effects, as evidenced by data on maternal
mortality, body weight, physical parameters, pregnancy rate, length of gestation, number of live
and dead pups at birth, and sex distribution. No effect was observed on pup survival and growth
or on gross postmortem findings. In the Sprague-Dawley rats there was a trend toward reduced
body and organ weights in the 21-day-old pups exposed to 30 and 300 ppm (96 and 958 mg/m3)
82
-------
levels; however, the differences were statistically significant (p<0.05) only for reduced body
weight (89% of control value) and liver weight (80% of control value) in the female pups at 300
ppm (958 mg/m3). Increases in relative kidney weight for the female Sprague-Dawley pups were
attributed to differences in body weight. The CD rats exhibited no such changes. The
investigators concluded that benzene at concentrations as high as 300 ppm (958 mg/m3) did not
induce reproductive toxicity in CD rats. The study in CD rats does not identify a LOAEL or a
NOAEL. The study in Sprague-Dawley rats identifies a LOAEL of 300 ppm (958 mg/m3) and a
NOAEL of 30 ppm (96 mg/m3), based on the statistically significant reduction in body and liver
weights, and based on past studies showing similar effects at similar concentrations of benzene.
4.2.2.2. Developmental Toxicity
4.2.2.2.1. Oral exposure. Seidenberg et al. (1986) evaluated benzene, along with 54 other
chemicals of known and unknown developmental toxicity potential, to validate the
Chernoff/Kavlock assay as a screen for developmental effects. Benzene doses of 1300
mg/kg/day were administered in cottonseed oil by gavage to pregnant ICR/SEVI mice on GDs 8
through 12, and the dams were allowed to deliver. Maternal toxicity was assessed using body
weights, mortality, or other clinical signs of overt toxicity as endpoints. Benzene had no effect
on maternal body weight, but it did produce significantly lower neonatal body weight (to -95%
of control value, p<0.05), as measured on days 1 and 2 after birth. This was a screening study
that used only one dose, and therefore it does not provide dose-response information. However,
the study does demonstrate evidence for fetal toxicity in the absence of observable maternal
toxicity.
In the Exxon Chemical Company (1986) study, pregnant female Sprague-Dawley rats
(20-22/group) were treated by gavage with 0, 50, 250, 500, or 1000 mg/kg/day benzene on GDs
6-15. No dose-related mortality was observed. Significant findings in the treated dams as
compared with controls were decreased food consumption at >250 mg/kg, decreased body
weight, body weight gains at >500 mg/kg, and increased incidence of alopecia at 1000 mg/kg
(p<0.05). Developmental toxicity was limited to decreased fetal body weights at >500 mg/kg
(p<0.05). Fetuses were examined only for external malformations, not for skeletal and visceral
malformations. This study identified a NOAEL of 50 mg/kg/day and a LOAEL of 250
mg/kg/day for maternal toxicity and a NOAEL of 250 mg/kg/day and a LOAEL of 500
mg/kg/day for developmental toxicity in Sprague-Dawley rats.
4.2.2.2.2. Inhalation exposure. This section summarizes the data for the developmental toxicity
of inhaled benzene in animals. The studies included experiments with three species (rats, mice,
and rabbits) and intermittent (6 or 7 hours/day) and continuous (24 hours/day) exposures. Some
studies tested more than one concentration of benzene and demonstrated a concentration
response and a LOAEL and/or a NOAEL. Some studies used only one concentration and were
not useful for determining dose-response relationships, but they did provide supporting evidence
for fetal toxicity. Three studies evaluated fetal hematologic parameters, and one study examined
the possible synergistic effects of benzene with other chemicals. The available studies on the
developmental toxicity of inhaled benzene are summarized in Table 8.
83
-------
Table 8. Developmental toxicity of inhaled benzene in test animals
Strain/specie
$
Sprague-
Dawley rat
Sprague-
Dawley rat
Sprague-
Dawley rat
Rat
No. of
dams/group
14-15
exposed,
11 controls
35-37
exposed,
32-34
controls
14-18
5-12
Exposure3
0, 32, 160, or 1597
mg/m3, 7 hrs/day on
GDs 6-15; sacrificed
on day 20
0, 3.2, 32, 128, or
319 mg/m3, 6 hrs/day
on GDs 6-1 5;
sacrificed on day 20
0,319, 958, or 7028
mg/m3, 6 hrs/day on
GDs 6-15; sacrificed
on day 21
1.0-670 mg/m3 24
hrs/day, 10-15 days
before mating and
throughout
pregnancy
Effects
Maternal
Decreased body
weight and body
weight gain at 160
and 1597 mg/m3,
days 5-15; dose-
related15
None in any group
Decreased body
weight gain at 7028
mg/m3 (p< 0.01)
No data
Developmental
Decreased mean live body
weight at 160 and 1597
mg/m3, day 20°; decreased
crown-rump distance, 1597
mg/m3c; skeletal and
visceral (brain)
abnormalities at 160 and
1597 mg/m3; increased
incidence of malformations
at 1597 mg/m3
Decreased body weight at
319mg/m3(^<0.05);
variants in all but one group
(including controls), not
dose related; no increase in
incidence of malformations
Decreased body weight and
length at 7028 mg/m3 (p<
0.05); increased skeletal
variants all exposure groups
(/?<0.05at319and7028
mg/m3; females more
sensitive than males); no
increase in incidence of
malformations
Tendency toward decreased
litter sizes at 64 mg/m3;
complete absence of litters
at 670 mg/m3
Effect levels
maternal/developmental
(mg/m3)
LOAEL: 160/160
NOAEL: 32/32
LOAEL:NA/319
NOAEL: NA/128
LOAEL: 7028/3 19
NOAEL: 958/NA
LOAEL: ND
NOAEL: ND
Reference
Kuna and
Kapp, 1981
Coate et
al., 1984
Green et
al., 1978
Gofmekler,
1968
oo
-------
Table 8. Developmental toxicity of inhaled benzene in test animals (continued)
Strain/specie
s
CFY rat
CFY rat
CFY rat
No. of
dams/group
19 exposed,
28 controls
20-22
exposed,
48 controls
17
Exposure3
0 or 1000 mg/m3, 24
hrs/day on days
9-14 of pregnancy
0, 150, 450, 1500, or
3000 mg/m3, 24
hrs/day on GDs
7-14; sacrificed on
day 21
400 mg/m3, 24
hrs/day on GDs 7-15
Effects
Maternal
Decreased body
weight gain (p<0.0l)
Decreased body
weight gain at >150
mg/m3 (p<0.001),
somewhat dose
related; liver/body
weight increased
(^<0.05or0.01)
Decreased body
weight gain
(£><0.001); increased
relative liver weight
Developmental
Decreased body weight
(p<0.01) and growth
retardation (p<0.05);
retarded skeletal
development and increased
incidence of extra ribs and
fused sternebrae (p<0.05
for both); no increase in
incidence of malformations
Decreased body weight at
> 150 mg/m3 (p<0.001),
increased resorptions and
skeletal and weight
retardation (p<0.0 1-0.05),
effects not dose related; no
increase in incidence of
malformations
Retarded weight gain
(p<0.01); skeletal growth
retardation (p<0. 001)
Effect levels
maternal/developmental
(mg/m3)
LOAEL: NA
NOAEL: NA
LOAEL: 150/150
NOAEL: NA
LOAEL: NA
NOAEL: NA
Reference
Hudak and
Ungvary,
1978
Tatrai et
al., 1980
Ungvary,
1985
oo
-------
Table 8. Developmental toxicity of inhaled benzene in test animals (continued)
Strain/species
Swiss-Webster
mouse
Swiss-Webster
mouse
No. of
dams/group
5
8
Exposure3
16, 32, or 64 mg/m3,
6 hrs/day on GDs
6-15
(a) 32 mg/m3
benzene GDs 6-15
(b) 5% ethanol in
drinking water ad lib
(c) 32 mg/m3
benzene + 5%
ethanol
(d) air + distilled
water
Effects
Maternal
None observed
No data
Developmental
16-day -old fetus: no effect
on hematologic parameters
2-day-old neonates:
reduced circulating
erythroid precursor cells
(all concentrations) (p<0.05
at 64 mg/m3); increased
hepatic hemato-poietic blast
cells, lympho-cytes, and
granulopoietic precursor
cells and decreased hepatic
erythropoiesis precursor
cells(all/?<0.05at64
mg/m3)
6-week-old adult: similar
pattern of enhanced
granulopoiesis (64 mg/m3)
Bone marrow samples from
6-week-old offspring:
protocols (a) and (b) caused
changes in CFU-E counts,
males only; protocol (c)
caused changes in CFU-E
counts, females onlyd
Effect levels
maternal/developmental
(mg/m3)
LOAEL: NA/64
NOAEL: 64/32
LOAEL: NA
NOAEL: NA
Reference
Keller and
Snyder,
1988
Corti and
Snyder,
1996
oo
-------
Table 8. Developmental toxicity of inhaled benzene in test animals (continued)
Strain/species
CF-1 mouse
CFLP mouse
New Zealand
rabbit
New Zealand
rabbit
No. of
dams/group
35-37
15 exposed,
115 controls
20
11 or 15
exposed,
60 controls
Exposure3
0 or 1597 mg/m3
ppm, 7 hr/day on
GDs 6-15; sacrificed
on day 18
0, 500 or 1000
mg/m3, 24 hrs/day on
GDs 6-15
0 or 1597 mg/m3, 7
hr/day on GDs 6-18;
sacrificed on day 29
0, 500 or 1000
mg/m3, 24 hrs/day on
GDs 7-20
Effects
Maternal
None
Not mentioned
None
Decreased weight
gain and increased
relative liver weight
at 1000 mg/m3 (p<
0.05)
Developmental
Decreased body weight
(/?<0.05), "significantly"
increased skeletal variants
of fetuses; no increase in
incidence of malformations
but was toxic to fetuses
Weight and skeletal
retardation, both
concentrations (p< 0.05),
somewhat dose-related
Statistically significant
decrease in minor skeletal
variants, lumbar spurs and
proportion with 13 ribs, in
exposed fetuses
Decreased body weight
and increased abortions and
skeletal variants at 3 13 ppm
(p<0.05 for all effects)
Effect levels
maternal/developmental
(mg/m3)
LOAEL: NA
NOAEL: NA
LOAEL: ND/500
NOAEL: NA/NA
LOAEL: NA
NOAEL: NA
LOAEL: 1,000/1,000
NOAEL: 500/500
Reference
Murray et
al., 1979
Ungvary
and Tatrai,
1985
Murray et
al., 1979
Ungvary
and Tatrai,
1985
oo
Conversion factors, 1 ppm = 3.26 mg/m3; 1 mg/m3 = 0.31 ppm
bStatistically different from control as determined by pairwise multiple comparison procedures
'Statistically different from control as determined by Cochran's approximation tot ft')
*Data from abstract; no other details available
GD = gestation day
NA = not applicable
ND = not determined
-------
In a review of some of the earlier, mostly unpublished developmental toxicity studies on
benzene, Brief et al. (1980) observed that the developmental effects in animals were
characterized mainly by fetal toxicity in rats exposed to 40, 50, 500, and 2200 ppm (128, 159,
1597, and 7028 mg/m3) (Dow Chemical Co., 1992b; Green et al., 1978). Maternal toxicity and
some fetal malformations occurred at 500 ppm (1597 mg/m3); concentrations of 10 ppm (32
mg/m3) produced conflicting results (Brief et al., 1980).
Kuna and Kapp (1981) exposed pregnant Sprague-Dawley rats (14-15/group) to benzene
concentrations of 0, 10, 50, or 500 ppm (0, 32, 160, or 1597 mg/m3) 7 hours/day on GDs 6-15
and sacrificed the dams on day 20 to evaluate maternal and developmental effects. Hematologic
evaluation of the dams, performed on GDs 5 and 20, included RBC, WBC, and differential
counts.
There were no deaths, observable illness, or hematologic changes for dams in any
exposure group. During GDs 5 through 15, maternal body weight gains were significantly
decreased to 66% of control values in the 50 ppm (160 mg/m3) group and to 63% of control
values in the 500 ppm (1597 mg/m3) group during GDs 5-15 (p<0.05 for both groups), whereas
during GDs 15-20, body weights in the dams exposed to 10 ppm (32 mg/m3) and weight gain in
the dams exposed to 10 and 500 ppm (32 and 1597 mg/m3) were increased. Body weight
corrected for gravid uterine weight was not determined. No differences were observed between
exposed and control groups in the number of implantation sites per number of ovarian corpora
lutea (implantation efficiency); the incidences of resorbed, dead, or live fetuses; and the sex
distribution.
Fetal-crown rump lengths were decreased in the 500 ppm (1597 mg/m3) group, and mean
body weight of live fetuses were decreased in both the 50 and 500 ppm (160 and 1597 mg/m3)
groups (control, 4.4 ± 0.6 g; 50 ppm [160 mg/m3], 3.8 ± 0.7 g; 500 ppm [1597 mg/m3], 3.6 ± 0.8
g; all statistically significant, p<0.05). One fetus from each of four litters from dams treated with
500 ppm (1597 mg/m3) benzene displayed exencephaly (one fetus), angulated ribs (one fetus), or
ossification of the forefeet out of sequence (two fetuses); these abnormalities were not observed
in control fetuses. This group also exhibited delayed ossification in the skull, vertebral column,
rib cage, pelvic girdle, and extremities and significantly fewer tail bones than the controls
(p<0.05). There was evidence of a dose-related (but not statistically significant) decrease in the
mean number of phalanges and metacarpals.
In summary, 13 litters and 142 fetuses were examined from the group exposed to 500
ppm (1597 mg/m3); 30 fetuses from six litters had delayed ossification (p<0.05), and four fetuses
from four litters had skeletal variants and abnormalities. In the group exposed to 50 ppm (160
mg/m3), 125 fetuses from 15 litters were examined; 23 fetuses from six litters had variants
(p<0.05). The investigators concluded that the effects observed in this study were benzene
induced at concentrations of 50 and 500 ppm (160 and 1597 mg/m3). The study appears to have
been conducted according to standard protocols and recommendations (U.S. EPA, 1991) for
developmental toxicity. The study establishes a LOAEL for the developmental toxicity of
inhaled benzene of 50 ppm (160 mg/m3) and a NOAEL of 10 ppm (32 mg/m3).
Coate et al. (1984) exposed pregnant Sprague-Dawley rats by inhalation to 0, 1, 10, 40,
or 100 ppm (0, 3.2, 32, 128, or 319 mg/m3) benzene 7 hours/day on GDs 6-15. Benzene levels
88
-------
in the chambers, monitored at least once per day using both infrared and gas chromatography
methods, were within 10% of target concentrations by both analyses, except for the gas
chromatography analysis of the 10 ppm (32 mg/m3) concentration, which was within 17% of
target.
Maternal body weight and mortality rate did not differ between treated and control
animals (other than a slight but insignificant decrease in body weight of the 100 ppm [319
mg/m3] group), and there were no clinical signs or gross pathology suggestive of maternal
toxicity during gestation. Body weight corrected for gravid uterine weight was not determined,
and hematologic assays were not performed. The average number of implantations, number of
resorptions, resorption incidence, number of live fetuses, and the incidence of dams with one or
more resorbed implantations were comparable in all control and exposed groups. The fetuses in
the 100 ppm (319 mg/m3) group exhibited a significant decrease (p<0.05) in average male and
female fetal body weights and a statistically insignificant decrease in fetal length. These values
were within 10% of control values. There were no significant differences in the percentage of
fetuses per litter with one or more variants, although slight increases in variants (dilation of the
renal pelvis and ureters) occurred in the 1 and 100 ppm (3.2 and 319 mg/m3) groups. Soft tissue
anomalies occurred in all exposed groups, but they were not dose related. Delays in ossification
of the skull, vertebral centra, and extremities occurred at 100 ppm (319 mg/m3). The
investigators concluded that benzene is weakly toxic to the fetus at 100 ppm (319 mg/m3) (a
concentration that was not toxic to the dams) and is not toxic to the fetus at 40 ppm (128 mg/m3).
This study establishes a LOAEL of 100 ppm (319 mg/m3) and a NOAEL of 40 ppm (128 mg/m3)
for the fetal toxicity of benzene.
Green et al. (1978) exposed groups of pregnant Sprague-Dawley rats to benzene
concentrations of 100, 300, and 2200 ppm (319, 958, and 7028 mg/m3) 6 hours/day on GDs
6-15. Daily body weights were measured as an indicator of maternal toxicity; peripheral blood
cell counts were not performed. The dams were sacrificed on day 21, and developmental
toxicity was evaluated. Maternal weight gain was similar to control values for the rats exposed
to 100 and 300 ppm (319 and 958 mg/m3); however, body weight for the females exposed to
2200 ppm (7028 mg/m3) was significantly reduced from GDs 8 to 20 to ~44%-83% (p<0.01 for
all days) of control values. There were no differences between all exposure groups and controls
with regard to implantation sites/litter, live fetuses/litter, percentage resorption/implantation site,
percentage litters with resorption, litters totally resorbed, and resorption/litter with resorption.
At concentrations of 100 and 300 ppm (319 and 958 mg/m3), fetal sex ratio, mean fetal body
weight, and mean fetal crown-rump length were comparable with control values; at 2200 ppm
(7028 mg/m3), mean fetal weight and mean crown-rump length were significantly lower than
controls (p<0.05).
The incidence of delayed ossification was similar for the exposed groups and controls;
however, among the fetuses exposed to 300 and 2200 ppm (958 and 7028 mg/m3) and exhibiting
delayed ossification, the incidence was increased significantly for females over males (p<0.05).
The number of litters displaying an increased incidence of unossified sternebrae was increased at
100 ppm (319 mg/m3) (9/18 litters) and 2200 ppm (7028 mg/m3) (11/15 litters), and at 2200 ppm
(7028 mg/m3) the number of females with the abnormality was significantly increased over
males (p<0.05). The concentration of 2200 ppm (7028 mg/m3) benzene was maternally toxic;
however, fetal toxicity manifested as skeletal abnormalities was observed at concentrations that
89
-------
were not maternally toxic (and were statistically significant). The apparently increased
sensitivity of the female fetus to the effects of benzene is supported by other observations that
female rabbits, mice, and rats are more sensitive to these effects of benzene (Desoille et al.,
1961; Ito, 1962a-d; Sato et al., 1975). The investigators tentatively suggested that the
differences in the responses of the males and females may be related to hormonal differences.
In an early study, Gofmekler (1968) exposed female rats to benzene concentrations
ranging from 0 to 210 ppm (670 mg/m3) 24 hours/day for 10-15 days prior to mating and all
during pregnancy. The rats exhibited a dose-related decrease in litter size and produced no
litters at the highest concentration of 210 ppm (670 mg/m3). This effect was not observed in the
experiments of Green et al. (1978) at higher benzene concentrations administered over a
different exposure period (6 hours/day on GDs 6-15). This suggests that the rat embryo is more
sensitive to continuous exposure than to intermittent exposure only during organogenesis.
However, females in that study were exposed before mating, and ovulation rate and fertility may
have been affected independently of developmental effects. The Gofmekler (1968) study
demonstrated significant differences in the weight of liver, lung, spleen, kidney, and thymus
from fetuses of benzene-treated dams. The weights were mostly decreased, but the effect was
not dose related. Gofmekler (1968) concluded that exposure of pregnant rats to relatively low
concentrations of benzene can disturb normal fetal development. However, the lack of a
concentration response in the organ weight data for fetal rats precludes the determination of a
LOAEL or a NOAEL for the study.
Hudak and Ungvary (1978) conducted an experiment with CFY rats exposed to 313 ppm
(1000 mg/m3) benzene 24 hours/day on GDs 9 through 14. Exposure to benzene resulted in
statistically significant decreases in maternal body weight gain (p<0.01), fetal weights (p<0.01)
and percentage of weight-retarded fetuses (p<0.05) and increased incidences of extra ribs and
fused sternebrae (p<0.05 for both). The use of only one dose precluded the derivation of a
LOAEL and/or a NOAEL for the study.
A later study (Tatrai et al., 1980) examined the developmental toxicity of benzene using
several doses. Pregnant CFY rats inhaled 0, 47, 140, 465, or 930 ppm (0, 150, 450, 1500, or
3000 mg/m3) 24 hours/day on GDs 7-14. Animals were sacrificed on day 21 to evaluate
developmental effects. The dams exhibited significantly decreased body weight gain at all
concentrations (/X0.001, ANOVA). The effect was dose related at 47, 140, and 465 ppm
(150,450, and 1500 mg/m3), but not at 930 ppm (3000 mg/m3). For example, the body weight
gain for dams exposed to 465 ppm (1500 mg/m3) benzene was 28.5% of the starting body weight
compared with the control value of 62.8%, whereas the body weight gain for dams exposed to
930 ppm (3000 mg/m3) benzene was 37.0%. Liver/body weight ratios were significantly
increased in the dams exposed to 140, 465, or 930 ppm (450, 1500, or 3000 mg/m3), with a
response pattern similar to the body weights (p<0.05 or/><0.01, ANOVA).
Fetal body weights were decreased at >47 ppm (150 mg/m3) (p<0.001, all
concentrations); fetal loss in the percent of total implantation sites was increased 4-7 times
control values at >140 ppm (450 mg/m3) (p<0.05 or/><0.01, Mann Whitney U test); and skeletal
and weight retardation were observed in the groups exposed to the three highest concentrations
(p<0.01 or/><0.05). For these effects, there was no distinct relationship between concentration
and response, and there was no increase in the incidence of malformations, even at
90
-------
concentrations >47 ppm (150 mg/m3) that were maternally toxic. The main difference observed
for the days 7-14 exposure regimen (Tatrai et al., 1980) study versus the days 9-14 regimen
(Hudak and Ungvary, 1978) was the increase in fetal mortality (30-40% of the implants) that
occurred when exposure was started on GD 7.
To test the hypothesis that the congenital effects of industrial solvents may be attributable
to simultaneous exposure to a combination of two or more chemicals, Ungvary and Tatrai (1985)
conducted a study using combinations of benzene and its methyl derivatives, toluene and xylene,
with each other or with acetylsalicylic acid (ASA), a known developmental toxicant. Groups of
pregnant CFY rats inhaled combinations of 123 ppm (400 mg/m3) benzene and 307 ppm (1000
mg/m3) toluene or 189 ppm (600 mg/m3) xylene 24 hours/day on GDs 7-15. Additional groups
of 7-14 animals were exposed to 798 ppm (2600 mg/m3) benzene on GDs 10-12 and given oral
doses of either 250 or 500 mg/kg of ASA on GD 12.
At 400 mg/m3, benzene produced maternal toxicity, as evidenced by decreased weight
gain (47% gain in the benzene-treated dams vs. 69% gain in the controls [p<0.001]); benzene
alone induced retardation of fetal and skeletal growth but did not increase the incidence of
skeletal anomalies or skeletal, internal, or external malformations or the percentage of abnormal
survivors. The combined developmental effects of the solvents were not additive. However,
when each solvent was administered by inhalation in combination with orally administered ASA,
it enhanced maternal as well as fetal toxicity. In addition, benzene and its methyl derivatives
significantly increased the frequency of AS A-induced malformations. The induction of most
malformations was not solvent dependent; cleft lip and palate and abnormalities of the spinal
column occurred only in response to the combined treatment with solvents and ASA. The
investigators concluded that under the conditions of this experiment, neither benzene nor its
alkyl derivatives induced malformations, alone or in combination, and that the fetal toxicity
observed with all three solvents was not additive when the solvents were combined; however, all
three solvents potentiated the toxic effects of ASA, including the induction of malformations.
Pregnant Swiss-Webster mice (five per exposure level per progeny age group; initial age
8-12 weeks) were exposed via inhalation to nominal vapor concentrations of 0, 5, 10, or 20 ppm
(0, 16, 32, or 64 mg/m3) benzene 6 hours/day on GDs 6-15 (Keller and Snyder, 1986). Two
fetuses/litter/sex were sacrificed on GD 16, two neonates/litter/sex were sacrificed 2 days after
birth, and one adult/litter/sex was sacrificed 6 weeks after birth to measure hematopoietic
progenitor cells (CFU-E, BFU-E, and GM-CFU-C) from the liver (fetuses and neonates) and
bone marrow and spleen (adults). In addition, 10-week-old progeny from litters in the control
and mid-exposure group were exposed for 2 weeks to 10 ppm (32 mg/m3) benzene, then
sacrificed to measure hematopoietic progenitor cells from the bone marrow and spleen.
There was no evidence of maternal or nonhematopoietic developmental toxicity in
benzene-exposed mice. There was a significant increase (p<0.05) in the numbers of BFU-E
from livers of male and female fetuses exposed to the low- and mid-exposure levels. The
following significant changes (p<0.05) were observed with respect to CFU-E: in fetuses, there
were increases in liver CFU-E at the low- and mid-exposure levels and decreases at the high-
exposure level; in male neonates, there were increases and decreases in liver CFU-E at the mid-
exposure level and increases at the high-exposure level; and in adult mice, there were decreases
in bone marrow CFU-E and increases in spleen CFU-E in males exposed to 10 ppm (32 mg/m3)
91
-------
in utero. Liver GM-CFU-C in neonates was significantly decreased (p<0.05) at the mid-
exposure level (males only) and increased at the high-exposure level. Mice exposed to 10 ppm
(32 mg/m3) benzene in utero and for 2 weeks as adults had significantly (p<0.05) decreased bone
marrow CFU-E (males only) and splenic GM-CFU-C; mice exposed to air in utero and 10 ppm
(32 mg/m3) benzene for 2 weeks as adults had no changes in bone marrow or splenic CFU-E, but
they had a significant (p<0.05) decrease in splenic GM-CFU-C (females only). The authors
concluded that benzene treatment in utero induced hematopoietic alterations in fetuses that
persisted until at least 10 weeks after birth.
This study could be interpreted as identifying a LOAEL of 5 ppm (16 mg/m3) for
developmental hematopoietic effects in mice, because statistically significant changes in BFU-E
and CFU-E were observed in the 16-day fetuses from dams exposed to 5 ppm (16 mg/m3).
However, the responses observed were typically biphasic in nature, showing increases in BFU-E
and CFU-E at 5 and 10 ppm (16 and 32 mg/m3) followed by decreases at 20 ppm (64 mg/m3).
Only five pregnant animals were used in the study per exposure dose level, and limited numbers
of animals (two fetuses, two neonates, or one adult/litter/sex) were used for the evaluation of
hematotoxic effects. Also, the responses did not establish a consistent pattern in the different
ages of progeny examined. Therefore, there is a high degree of uncertainty associated with
determining whether the effects observed at 5 ppm (16 mg/m3) are truly adverse effects.
Keller and Snyder (1988) further examined the effects of in utero benzene on the
developing recognizable hematopoietic precursor cells and peripheral blood cells and Hgb
production in Swiss-Webster mice. Three separate exposure experiments were performed in
which pregnant mice (five per group) were exposed to benzene concentrations of 0, 5, 10, or 20
ppm (0, 16, 32, or 64 mg/m3) in chambers 6 hours/day on GDs 6-15. Maternal toxicity was
evaluated on the basis of morbidity, mortality, or weight loss, but maternal peripheral blood
count was not determined. In experiment 1, the fetuses of the benzene-exposed dams were
assayed for signs of hematotoxicity on GD 16 (two fetuses/sex/litter). In experiment 2, the
offspring (two neonates/sex/litter) were assayed at 2 days. In experiment 3, progeny
(one/sex/litter) were examined at 6 weeks of age. Parameters of toxicity for the fetuses and
2-day-old neonates included the number of live, dead, and resorbed fetuses; body weights; gross
abnormalities; RBC, WBC, and blood cell differentials; Hgb levels; and the number of liver cells
in the hematopoietic differentiating, proliferating pool. The responses of 6-week-old progeny
were evaluated on the basis of peripheral WBC and RBC counts, Hgb, and smears made from
femur bone marrow or spleen to determine the number of cells in the differentiating,
proliferating pool. Although not stated, the experimental animals may have been the same for
both studies (Keller and Snyder, 1986, 1988), as all the details of the exposures were identical.
Benzene exposures monitored hourly were well within target range (Keller and Snyder,
1988). There was no evidence of maternal toxicity at any exposure level. For the fetuses, there
were no effects on litter size, male/female ratio, or body weight or the numbers of dead,
resorbed, or malformed. Significant findings (p<0.05) relevant to hematopoietic toxicity in the
three groups of progeny included the following.
Peripheral blood. Hgb levels of exposed and control mice were similar for all groups.
Peripheral blood cell and differential cell counts of the 16-day fetuses and differential cell counts
of the 6-week-old progeny were comparable with those of the controls. The 2-day-old neonates
92
-------
exposed in utero to 5 ppm (16 mg/m3) benzene exhibited significant increases in RBC counts and
decreases in mean corpuscular Hgb, and the 6-week-old adults exposed in utero to 5 and 20 ppm
(64 mg/m3) benzene had increased RBC counts. However, these effects did not occur with a
distinct pattern, and the investigators concluded that they probably had no toxicological
significance. Peripheral blood cell differentials from the 2-day-old neonates revealed
significantly decreased numbers of early nucleated red cells (statistically significant at all
concentrations); the effect was concentration dependent. The 2-day-old neonates exposed in
utero to 20 ppm (64 mg/m3) also had decreased numbers of late nucleated red cells and increased
numbers of nondividing granulocytes; these effects were not concentration dependent.
Hematopoietic organs. The 16-day fetuses from dams exposed to benzene exhibited no
changes in hematopoietic parameters of the liver at any benzene concentration. The 2-day-old
neonates exposed in utero to 20 ppm (64 mg/m3) benzene had changes in the liver that were
significant at 20 ppm (64 mg/m3). These included increase in the number of blast cells to 3.2
times the control value, depression in the number of late nucleated red cells (polychromatic
normoblasts and their nucleated progeny) to 70% of the control value, an increase in the number
of lymphocytes to 1.4 times the control value, an increase in the number of nondividing
granulocytes to 1.9 times the control value, and an increase in the numbers of dividing
granulocytes to 2.5 times the control value. These effects were not statistically significant at
lower concentrations. The effects on the early red cells and granulocytes were concentration
related; the others did not show a distinct dose response. The 6-week-old adults exposed in utero
to 20 ppm (64 mg/m3) benzene also exhibited statistically significant changes in the cells of
hemopoietic organs. These included fewer early nucleated red cells in the liver (to 55% of the
control value), more splenic blast cells (to 6.5 times the control value), and increased dividing
and nondividing splenic granulocytes (to 2.9 and 3.2 times the control value, respectively). The
effects were not dose related and, with the exception of one cell count at 5 ppm (16 mg/m3), were
not statistically significant at lower concentrations.
The authors concluded that benzene induces hematotoxicity in the offspring of mice
exposed during pregnancy, as evidenced by reductions in the numbers of early nucleated red
cells in the peripheral blood of 2-day-old neonates at all exposure levels and by the increase in
the number of dividing granulocytes in the liver of 2-day-old neonates and in the spleen of 6-
week-old adults. However, the toxicological significance of these results is unclear. The only
clearly dose-related response was the decrease in early nucleated red cells in the peripheral blood
of 2-day-old neonates. Early nucleated red cells, however, were present in the peripheral blood
only in very young animals, as clearly indicated by the near absence of such cells in the 6-week-
old adult mice. A limited number of pregnant animals (five per group) and progeny
(two/litter/sex) were tested. Therefore, biological significance and confidence in the data are
questionable. Thus, a NOAEL or a LOAEL cannot be derived from the results of the Keller and
Snyder (1988) report.
Corti and Snyder (1996) examined the influence of gender, development, pregnancy, and
ethanol consumption on the hematotoxicity of inhaled benzene. They exposed age-matched
male, virgin female, and pregnant Swiss-Webster mice (number of animals in each group was
not given, but there were 12 or 13 dams) to 10 ppm (32 mg/m3) for 6 hours/day for 10
consecutive days (GDs 6-15 for the pregnant females). One-half of the animals also received
5% ethanol in the drinking water during the exposure period. On day 11, bone marrow cells
93
-------
from the adults and liver cells from the fetuses were assayed for the numbers of CFU-E. CFU-E
assays were also performed on bone marrow cells isolated from 6-week-old males and females
exposed in utero. Depressions in CFU-E numbers were seen only in males. CFU-E in adult
males was depressed to -70% of control values by exposure to benzene, ethanol, or ethanol plus
benzene. The action of the two agents was neither additive nor synergistic. CFU-E from fetal
livers was significantly decreased in males exposed to benzene or benzene plus ethanol but not
for ethanol alone. The CFU-E value for males exposed to both benzene and ethanol, however,
was reduced by only 5% in comparison with the control value as compared with 20% for ethanol
alone. There were no significant differences for female fetuses. Assays of CFU-E in 6-week-old
mice exposed to benzene in utero showed reductions of as much as 50% in comparison with
controls for males, but there were no significant reductions for females. In fact females exposed
to benzene and ethanol in utero showed a 40% elevation of CFU-E in comparison with
unexposed controls. For both ethanol and benzene, only a single concentration was used in these
experiments. Thus, it is not possible to evaluate whether the results are part of a dose-response
relationship.
Murray et al. (1979) conducted developmental toxicity studies in mice that also included
the evaluation of hematologic parameters in the offspring. CF-1 mice were exposed to 500 ppm
(1597 mg/m3) benzene 7 hours/day on GDs 6-15 and were sacrificed on GD 18. There were no
significant signs of maternal toxicity, and values for PCV, RBCs, Hgb, and WBCs were
comparable with control values for the adult mice. There were no adverse effects on
developmental parameters such as number of live fetuses or number of resorptions/litter;
however, body weight of the mouse fetuses was slightly but significantly decreased (to 94% of
control values; p<0.05). The investigators described a "significant" increase in the incidence of
skeletal variations of the mouse fetuses, as evidenced by delayed ossification of sternebrae and
skull bones and unfused occipital bones of the skull, but there was no significant increase in the
incidence of malformations. The investigators concluded that benzene did not induce
malformations in mice at 500 ppm (1597 mg/m3), but it was toxic to the fetuses. This study used
only one dose, and statistical analysis for the variations were not clearly stated; however, the
study did demonstrate fetal toxicity at a dose that was not maternally toxic.
Ungvary and Tatrai (1985) examined the developmental toxicity of benzene in CFLP
mice. Groups of 11 to 15 pregnant animals inhaled 156 or 313 ppm (500 or 1000 mg/m3)
benzene 24 hours/day during GDs 6-15. Unexposed and air-exposed (chamber) animals served
as controls. The mice were sacrificed on GD 18. The authors mentioned in the abstract of their
paper that all solvents caused moderate and concentration-dependent maternal toxicity. This was
the only information provided for maternal toxicity among the mice. Of the mouse fetuses from
dams exposed to 156 and 313 ppm (498 and 1000 mg/m3), 25 and 27%, respectively, exhibited
weight retardation, and 10 and 11%, respectively, had retarded skeletal development. All of
these effects were statistically significant (p<0.05). There were no differences between control
and exposed groups in the incidence of minor anomalies or malformations. The investigators
concluded that under the conditions of this study, benzene does not induce malformations, but it
does induce fetal toxicity in mice exposed to 156 and 313 ppm (498 and 1000 mg/m3), a
maternally toxic dose. The mouse data suggest a LOAEL of 156 ppm (498 mg/m3).
Murray et al. (1979) conducted developmental toxicity studies in rabbits that also
included the evaluation of hematologic parameters in the offspring. New Zealand rabbits were
94
-------
exposed to 500 ppm (1597 mg/m3) benzene 7 hours/day on GDs 6-18. The rabbits were
sacrificed on GD 29. The rabbits exhibited no significant signs of maternal toxicity, and there
were no adverse effects on developmental parameters such as number of live fetuses or number
of resorptions/litter. Body weight and length of the rabbit fetuses were comparable with control
values. Two minor skeletal variants, lumbar spurs and the proportion of fetuses with 13 ribs (the
normal number is 12 or 13), occurred significantly less often among the exposed litters.
Increases in the incidence of malformations were not significant. Values for PCV, RBCs, Hgb,
and WBCs were comparable to control values for adult and fetal rabbits. The investigators
concluded that benzene did not induce malformations in rabbits at 500 ppm (1597 mg/m3) and
was only weakly toxic to the fetuses. This study used only one dose, and statistical analyses for
the variations observed were not clearly stated.
Ungvary and Tatrai (1985) also examined the developmental toxicity of benzene and
other solvents in New Zealand rabbits. Groups of 11-15 pregnant animals inhaled 156 or 313
ppm (500 or 1000 mg/m3) benzene 24 hours/day on GDs 7-20. Unexposed and air-exposed
(chamber) animals served as controls. The rabbits were sacrificed on GD 30. The maternal
rabbits exhibited concentration-related decreases in weight gain (to 37% of control values at 313
ppm [1000 mg/m3]) and increases in relative liver weights (1.2 times control values at 313 ppm
[1000 mg/m3]); compared with control values, the effects were statistically significant at 313
ppm (1000 mg/m3) (p<0.05). Two dams died and six aborted at the higher concentration. The
fetuses exhibited concentration-related decreases in body weight (to 83% of control values,
/?<0.05 at 313 ppm [1000 mg/m3]), concentration-related increases in the percent of dead or
resorbed fetuses (3.1 times control values,p<0.05 at 313 ppm [1000 mg/m3]), concentration-
related increases in skeletal retardation (not statistically significant) and minor anomalies (2.5
times control values,/><0.05 at 313 ppm [1000 mg/m3]), and concentration-related decreases in
the percent of malformations. The investigators concluded that under the conditions of this
study, benzene does not cause malformations, but it induces fetal toxicity in rabbits at 313 ppm
(1000 mg/m3), a maternally toxic concentration. The evidence supporting the LOAEL and the
NOAEL in rabbits is weakened by a lack of experimental details in the report, the small numbers
of pregnant rabbits employed, and the use of the fetus rather than the litter as the experimental
unit.
4.2.2.3. Summary of Principal Reproductive/Developmental Toxicity Effects
As shown in the studies summarized in the previous subsections and in Tables 7 and 8,
administration of benzene has been associated with the onset of reproductive/developmental
effects in a wide range of experimental animals.
The susceptibility of reproductive organs to the toxic effects of benzene was shown in
studies on the longer-term administration of benzene in which, among other organs, the rat testes
(Wolf et al., 1956) and the ovaries of female CD-I mice (Ward et al., 1985; NTP, 1986) were
reduced in weight compared with controls and histopathologically altered as a result of benzene
administration. The potential for these changes to affect reproductive function in males is
suggested by reduced sperm count and increased numbers of abnormal sperm when CD-I mice
inhaled up to 300 ppm (958 mg/m3) benzene for 13 weeks.
95
-------
The experiments of Kuna et al. (1992) provided evidence that the inhalation of benzene
by female Sprague-Dawley rats before and during pregnancy resulted in reduced body and liver
weights in weanlings, a valid reproductive/developmental response because it occurred in the
absence of any overt effects of the compound on maternal toxicity. In addition, a considerable
number of experiments have addressed the potential of inhaled benzene to induce reproductive,
developmental, or teratogenic effects in the offspring of pregnant rodents exposed to the
compound during the period of principal organogenesis. For example, the studies by Kuna and
Kapp (1981) of Sprague-Dawley rats point to the possibility of benzene-induced skeletal
variations and abnormalities in the fetuses of exposed dams, although the quantitative extent of
the effect was marginal at best. However, the weight of evidence for benzene as a
developmental toxicant arises from the extensiveness of the overall information base. Several
studies that employed essentially similar experimental protocols (see, e.g., Green et al., 1978)
contribute to an overall picture of the compound as inducing reduced fetal growth and delayed
ossification. In addition, other studies reported reductions in fetal body weights, for example, in
Sprague-Dawley rats (Coate et al., 1984) and New Zealand white rabbits (Ungvary and Tatrai,
1985).
The developing hematopoietic system in mice is affected by maternal exposure to
benzene (Keller and Snyder, 1986, 1988; Corti and Snyder, 1996), which is consistent with
benzene-induced hematologic abnormalities in humans and adult animals. The offspring of
dams exposed during gestation exhibited increased granulopoiesis and changes in the numbers of
hematopoietic progenitor and precursor cells, particularly during postnatal development.
Multigenerational animal studies on the reproductive/developmental toxicity of benzene
were not found in the literature, an information gap that contributes to uncertainty about the
overall context of the compound's reproductive and developmental toxicity. As discussed in
Sections 4.1.2.3 and 4.1.2.4, the evidence regarding reproductive and developmental effects from
human studies is limited. The human studies suffer from major limitations, including small
numbers of subjects, exposures to a mixture of chemicals, and poorly defined benzene exposure
levels and durations of exposure. In many cases data on controls are lacking.
4.2.2.4. Mechanisms of Developmental and Reproductive Toxicity
The mechanisms for the developmental and reproductive toxicity of benzene are not well
understood. Summarized below are a few of the suggested mechanisms that pertain specifically
to the developmental and reproductive toxicity of benzene.
• Following administration of benzene to pregnant rats, Pushkina et al. (1968) observed
decreased ascorbic acid content in the whole fetus and in maternal organs as the
concentration of benzene increased, first in the maternal liver and later in the placenta
and fetal liver. Benzene also increased the DNA content and decreased the RNA content
in the placenta, fetal liver, and fetal brain and it decreased DNA content in the maternal
liver. The authors suggested that these alterations in ascorbic acid, RNA, and DNA
content are possible mechanisms for fetal toxicity.
• Ungvary and Donath (1984) suggested that damage to the peripheral noradrenergic fibers
observed in their study in pregnant rats may result in a disturbed control of ovarian and
96
-------
uterine blood flow and steroid production and may thus be instrumental in the
embryotoxic action of organic solvents.
• Tatrai et al. (1980) suggested that several factors may be responsible for the
embryotoxicity of benzene. First, because of its lipophilicity, benzene can pass the
placental barrier and affect the embryonal cells directly. Second, phenol, a major
metabolite of benzene shown to inhibit DNA synthesis in bone marrow in vivo, can also
pass the placental barrier. Third, benzene can damage maternal circulation and cause
bone marrow depression, resulting in adverse nutritional conditions for the fetus.
Benzene oxide was suggested as the toxic metabolite of benzene, and it can be
translocated from liver to blood and bone marrow. The enzymes responsible for its
production appear in the rat fetus late in pregnancy. However, in the human fetus, the
enzymes are present during weeks 9-13 of pregnancy. The authors suggested that a
transplacental effect is more plausible than any other mechanism.
• Reports indicating that paternal (as well as maternal) exposure to benzene is associated
with increased risk for stillbirths and the findings of increased incidences of testicular
lesions in benzene-exposed animals suggest that exposure of males may be important in
the reproductive toxicity of benzene (Savitz et al., 1989; Spano et al., 1989; Ward et al.,
1985).
4.2.3. Neurotoxicity
4.2.3.1. Oral Exposure
Sprague-Dawley rats given a single dose of 1870 mg/kg benzene exhibited tremors and
tonic-clonic convulsions and died within minutes. A dose of 352 mg/kg produced slight
nervous system depression (Cornish and Ryan, 1965). The LD50 for benzene in nonfasting rats is
0.81 g/kg.
Hsieh et al. (1988a) evaluated the effects of benzene on neurotransmitters of the brain.
CD-I adult male mice (five/group) received benzene in the drinking water for 4 weeks ad lib.
Controls received tap water. The nominal concentrations were 0, 40, 200, and 1000 mg/L.
However, based on water consumption, the actual daily estimated doses of benzene were 0, 8,
40, and 180 mg/kg/day.
The benzene concentrations used in this study did not significantly alter behavior, body
weights, or food and water consumption. Generally, the increases in levels of monoamine
neurotransmitters were dose related at 8 and 40 mg/kg/day, but in several cases there were no
further increases at 180 mg/kg/day. The increases were greatest for norepinephrine in the
hypothalamus (increased over control by approximately 38, 55, and 58% at 8, 40, and 180
mg/kg/day, respectively) and medulla oblongata (16, 42, and 20%). For serotonin, a similar
pattern of increase was observed in the higher association center with increased exposure to
benzene. In the hypothalamus, the serotonin levels were 21, 86, and 93% above the control
values. A generalized increase was also observed in the medulla oblongata (5, 25 and 19%) and
the midbrain (8, 46, and 23%). The increases in the parent compounds were associated with
increases in their corresponding metabolites, reflecting increased turnover of the amines. The
97
-------
data from this study did not establish a no-effect level of benzene in drinking water. In fact,
although the assay has potential as a biomarker for exposure, the biological significance of these
findings is questionable. The findings, however, suggest that neurotoxicological effects in the
hypothalamus and the peripheral nervous system could be of particular concern for the
developing central nervous system, resulting in permanent effects.
4.2.3.2. Inhalation Exposure
Ten rabbits exposed to ~ 45,000 ppm (143,760 mg/m3) benzene exhibited light narcosis
after 3.7 minutes of exposure, followed by tremors, chewing, excitement, and running
movements after 5 minutes (Carpenter et al., 1944). There was a loss of pupillary reflex to
strong light after 6.5 minutes, loss of blink reflex to tactile stimulus after 11.4 minutes, pupillary
contraction after 12 minutes, involuntary blinking after 15.6 minutes, and death after 36.2
minutes (Carpenter et al., 1944).
Andersson et al. (1983) examined the effects of high concentrations of benzene on
dopamine and noradrenaline turnover within various parts of the hypothalamus that are involved
in neuroendocrine regulation. Turnover of the catecholamines was assessed by measuring
changes in the degree of catecholamine stores following tyrosine hydroxylase inhibition.
Sprague-Dawley rats exposed to 1500 ppm (4792 mg/m3) benzene 6 hours/day for 3 days were
divided into two groups of four or six animals. In the first group (-inhibitor), rats that served as
controls were sacrificed immediately following exposure. In the second group (+inhibitor), the
injection of tyrosine hydroxylase inhibitor immediately followed exposure, and the animals were
sacrificed 2 hours later. The extent of depletion of the catecholamines was determined by
calculating the percentage depletion in the test group, based on the levels of catecholamine
present in the control group at the time of the injection of the hydroxylase inhibitor.
Quantitative microfluorimetry demonstrated that benzene induced statistically significant
alterations in catecholamine content and turnover in various sections of the hypothalamus.
Benzene (-inhibitor) produced increases in the catecholamine fluorescence in the median and
lateral palisade zones of the median eminence (p<0.05) and within the posterior periventricular
hypothalamic region (p<0.05). Benzene (+inhibitor) enhanced the disappearance of
catecholamine in the median palisade zone (p<0.002), within the posterior periventricular
hypothalamic region (p<0.05), within the parvocellular part of the paraventricular hypothalamic
nucleus (p<0.01), and within the dorsomedial hypothalamic nucleus (p<0.01).
The investigators concluded that benzene produced a pattern of discrete changes in
noradrenalin and dopamine turnover in certain areas of the hypothalamus. The study included
only one concentration of benzene and a small number of animals. No LOAEL or NOAEL was
established. Tyrosine hydroxylase is the key enzyme for biosynthesis of catecholamine, and the
hypothalamus is one of the major association centers in the central nervous system. It is clear
that benzene will affect the functions of all these centers at an inhalation dose of 500 ppm (1597
mg/m3), as demonstrated in the changes in dopamine levels of various areas, probably including
centers controlling respiration, hunger, and thirst.
Ungvary and Donath (1984) evaluated the effects of benzene on the noradrenergic
innervation of reproductive organs. CFY rats were exposed to 465 (1500 mg/m3) benzene 8
98
-------
hours/day on GDs 8-10 ppm. The abundance of fluorescent noradrenergic fibers normally found
in the ovaries and uterus of the pregnant rats decreased, whereas background fluorescence
increased (interpreted by the investigators to indicate an increased release of noradrenalin). To
confirm that the effect was selective to nerve fiber, the investigators injected benzene into the
anterior chamber of the eye. The benzene-induced damage to the sympathetic nerve-plexus of
the iris was similar to that in the uterus and ovary. The density of fluorescent fibers was
decreased (dose related), and the iris exhibited substantial hyperemia and increased background
fluorescence 72 hours after injection. The investigators concluded that benzene has a selective
and differential toxic effect on postganglionic neurons, with potential embryotoxicity.
Dempster et al. (1984) examined the temporal relationship between the behavioral and
hematologic effects of inhaled benzene. Mice (30-45/group) were exposed to benzene
concentrations of 100-3000 ppm (319-9584 mg/m3) 6 hours/day for the number of days
necessary to achieve a concentration x time product of 3000 ppm-days. The controls were age-
matched and exposed to air. The behavioral parameters evaluated included milk licking (an
observation of the mice licking a spout protruding from a wall in order to obtain milk), hindlimb
grip strength (a mouse was held by the tail and steadily pulled backward through a trough until
both hind paws grasped a wire triangle and then it was pulled until its grip was broken), and
home cage food and water intake (measured by weighing water bottles and feeders daily).
Lymphocyte numbers were reduced to 68% of control values after five exposures to 100 ppm
(319 mg/m3) benzene, to 50% after two exposures to 300 ppm (958 mg/m3), and to 50% after one
exposure to 1000 and 3000 ppm (3194 and 9584 mg/m3). Maximal depression in the lymphocyte
counts occurred after 10 days of exposure to 300 ppm (958 mg/m3). The lymphocyte counts
remained depressed throughout each exposure regimen.
RBC counts were not depressed as rapidly or to the same extent as were the lymphocytes.
At the end of exposure to 300 ppm (958 mg/m3), for example, the RBC counts were reduced to
70% of control values. Increased milk licking, the most sensitive of the behavioral parameters,
was statistically significant following 1 or 2 days of exposure to 100 ppm (319 mg/m3) and after
4 or 5 days of exposure to 300 ppm (958 mg/m3). The maximal increase in this behavior pattern
occurred after 7-8 days of exposure to 300 ppm (958 mg/m3), following the same time course as
the hematologic effects. Short-term exposures to high concentrations did not increase milk
licking, but they did increase food consumption. Thus the increased milk licking at 100 ppm
(319 mg/m3) was apparently not due to hunger. Exposure to 1000 ppm (3195 mg/m3) for one
exposure reduced hindlimb grip strength. These effects disappeared following termination of
exposure.
In a study designed to reflect occupational exposure, male CD-I and C57BL/6 mice were
exposed to 300 or 900 ppm (958 or 2875 mg/m3) benzene 6 hours/day for 5 days followed by 2
weeks of no exposure, after which the exposure regimen was repeated for an unspecified amount
of time (Evans et al., 1981). Seven categories of behavioral activities were monitored in exposed
and control animals: stereotypic behavior, sleeping, resting, grooming, eating, locomotion, and
fighting. Only minimal and insignificant differences were observed between the two strains of
mice. Increased behavioral activity in both strains was observed after exposure to benzene.
Mice exposed to 300 ppm (958 mg/m3) benzene had a greater increase than those exposed to 900
ppm (2875 mg/m3), probably because of narcosis-like effects induced at the higher exposure
99
-------
level. It is not known whether benzene induces behavioral changes by direct action on the
central nervous system.
Frantik et al. (1994) found that male albino SPF rats from a Wistar-derived strain
exposed to benzene for 4 hours in glass chambers (dose not specified) exhibited depression of
evoked electrical activity in the brain; the authors calculated the 30% effect level (depressed
activity) as 929 ppm (2968 mg/m3). When female H strain mice were exposed to benzene for 2
hours, the 30% effect level for depression of evoked electrical activity in the brain was 856 ppm
(2735 mg/m3).
Adult male Kunming mice (five/group) were exposed to 0, 0.78, 3.13, or 12.52 ppm (0,
2.5, 10, 40 mg/m3) benzene 2 hours/day, 6 days/week for 30 days (Li et al., 1992). Exposures
were conducted under static conditions in 300 m3 plexiglass chambers. Benzene was monitored
by gas chromatography every 30 minutes for 3 days at the beginning of the experiment, but
apparently it was not monitored for the rest of the experiment. The animals were monitored for
Y maze performance (rapid response); locomotor activity; forelimb grip strength; acetyl
cholinesterase (AChE) activity in the blood and brain; brain, liver, spleen, and kidney weights;
and bone marrow cellularity. Statistical significance was evaluated by t value for forelimb grip
strength and locomotor activity and by U value for rapid response frequency.
Significantly increased grip strength was observed at 0.78 ppm (2.5 mg/m3), but at higher
doses, grip strength was significantly decreased. The frequency of rapid response (running a Y
maze in less than 3 seconds) followed a pattern similar to forelimb grip strength. The frequency
increased significantly (p<0.05) from 33.7% in controls to 43.8% in the 0.78 ppm group but
declined significantly to 29.4% and 25.5% in the 3.13 and 12.52 ppm (10 and 40 mg/m3) groups,
respectively. No statistically significant differences were observed in AChE activity in either
blood or brain or in locomotor activity at any dose level. Relative liver weight was significantly
increased and relative spleen weight was significantly (p<0.05) decreased in the high-dose
group. No statistical analysis of the bone marrow histologic investigation was presented. There
were no apparent responses in the 0.78 or 3.13 ppm (2.5 or 10 mg/m3) groups. In the high-dose
12.52 ppm (40 mg/m3) group, however, there were reductions of 91% in myeloblasts, 64% in
premyeloblasts, 77% in metamyelocytes, 100% in reticulum, and 100% in erythroblasts.
The neurological effects reported in this study—forelimb grip strength and frequency of
rapid response—are unique and interesting. However, several limitations in this study prevent
its use in establishing an RfC. The number of animals used in each group (five) was low,
exposures were performed under static conditions, and benzene concentrations were monitored
on only the first 3 of 30 exposure days. The very large decreases in several blood parameters are
in contrast with the findings of most other studies, which found minimal or no response in bone
marrow parameters at similar exposure concentrations. Most other studies on the hematotoxicity
of inhaled benzene used exposure durations of 6 hours/day, in contrast to the 2 hours/day in this
study. Thus, the dramatic effects on bone marrow parameters in this study suggest that actual
benzene exposure may have been higher than reported.
100
-------
4.2.3.3. Summary ofNeurotoxic Effects
The database for establishing threshold dose levels for neurotoxicity effects is limited.
Neurological effects of benzene have been observed in animals and humans, and no clear
NOAEL could be identified in any species. No complete neurological testing has been
conducted in animals or humans. The exposure levels used in the neurotoxicity studies
summarized here were typically high and the exposure durations were short. The longest
exposure period was 4 weeks (Hsieh et al., 1988a). It should be noted that neurological concerns
have not been raised in occupational studies where exposure was at lower levels. Kahn and
Muzyka (1973) reported neurological complaints in workers exposed to 6-16 ppm
(20-52mg/m3), but no objective evaluation was conducted, exposure was poorly characterized,
and there was confounding exposure to low levels of toluene (1.6 ppm [5 mg/m3]). Other human
neurotoxicity studies reported much higher exposure levels. No complete functional
observational battery has been conducted in animals. Dempster et al. (1984) reported reduced
milk licking in mice exposed to 100 ppm (319 mg/m3 for 1 or 2 days, but grip strength was not
affected until exposure reached 1000 ppm (3195 mg/m3). Li et al. (1992) reported effects on grip
strength, locomotor activity, and Y-maze performance (rapid response) at 0.78 ppm (2.5 mg/m3),
but poor exposure characterization and marked inconsistencies between this and other studies'
hematologic effects mean that this value is unreliable.
4.2.4. Immunotoxicity
4.2.4.1. Oral Exposure
Male Charles River CD-I mice (five/group, 6-7 weeks of age) were exposed to 0, 31,
166, or 790 mg/L (0, 8, 40, or 180 mg/kg/day) benzene in drinking water for 28 days (Hsieh et
al., 1988b). The treatment had no adverse effects with respect to mortality, clinical signs, body
weight change, liver weight, or gross necropsy. A dose-related decrease in relative spleen
weight was observed, significant (p<0.05) at the high-exposure level. In one test, spleen
cellularity was reported to be significantly decreased at all exposure levels, and in a separate test
only at the high-exposure level. Although relative thymus weights were decreased at all
exposure levels, the values were not statistically significantly different from control values.
Dose-related hematologic effects (erythrocytopenia, leukocytopenia, lymphocytopenia, increased
MCV) were observed at all exposure levels. The authors indicated that the increased MCV and
decreased HCT and numbers of RBCs were indicative of severe macrocytic anemia.
Biphasic responses were observed in immunological tests, including mitogen-stimulated
(lipopolysaccharide [LPS], pokeweed mitogen, concanavalin A [Con A], phytohemaglutinin
[PHA]) splenic lymphocyte proliferation; mixed splenic lymphocyte culture response to
allogenic yeast artificial chromosome [YAC]-1 cells; cytotoxic splenic T lymphocyte response to
allogenic YAC-1 cells with a significantly increased response at the low-exposure level; and
significantly decreased responses at the mid- and/or high-exposure level. Using several methods
to determine primary antibody response to SRBC, significantly decreased responsiveness was
observed at the mid- and/or high-exposure levels. This response was either significantly
increased or not different from controls in mice at the low-exposure level. This study identified
a LOAEL of 8 mg/kg/day (the lowest dose tested) for hematologic and immunological effects in
male mice exposed to benzene in drinking water for 30 days. No NOAEL was established.
101
-------
In a subchronic study, groups of male C57BL/6 mice were exposed via drinking water to
0, 152, or 853 mg/L benzene for 7-28 days. Using estimated daily water intakes, the authors
calculated benzene dosages of 27 and 154 mg/kg/day (Fan, 1992). Five mice per group were
sacrificed after 7, 14, 21, or 28 days of exposure. An unspecified number of mice were exposed
to 152 mg/L for 28 days and sacrificed 7, 14, or 21 days after the last dosage. The focus of this
study was to determine the toxicity of benzene on natural killer (NK) cells involved in
nonspecific host resistance and on IL-2, which is the primary growth factor of T cells, a growth
factor for B cells and NK cells, and is involved in the regulation of granulocyte and eosinophil
production that occurs in response to NK cell activity and IL-2 production. No overt signs of
toxicity were observed in the benzene-exposed mice. Significant decreases in the number of
spleen cells were observed in both groups of benzene-exposed mice. This effect was observed
after 21 days of exposure in the 152 mg/kg/day group and after 14 days in the 852 mg/kg/day
group. After 21 days, a significant increase in splenic NK cell activity was observed in both
groups; however, after 28 days, the activity was not significantly different from that of controls.
Splenic IL-2 production was significantly depressed after 28 days in both groups. Spleen cell
numbers and IL-2 production were also depressed in the mice exposed to 152 mg/kg/day for 28
days and sacrificed 7 and 14 days (IL-2 levels only) after the end of the exposure. This study
identified a LOAEL of 152 mg/kg/day (the lowest dose tested) for effects on the immune system
in male mice. A NOAEL was not identified.
Female B6C3F1 mice (12/group, 6-7 weeks of age) were exposed to benzene in drinking
water (containing emulphor to increase solubility of benzene) at levels of 0, 50, 1000, or 2000
mg/L (0, 12, 195, or 350 mg/kg/day, as calculated by the authors) for 30 days (White et al.,
1984). Body weight was significantly decreased at the high-exposure level (p<0.05). A dose-
related decrease in absolute and relative spleen weight was observed (/X0.01). In one test,
spleen cellularity was reported to be significantly decreased at all exposure levels (p<0.05), and
in a separate test at only the mid- and high-exposure levels. Dose-related leukopenia and
lymphocytopenia were observed (p<0.05). A dose-related decrease in eosinophils was observed
(p<0.01). At the high-exposure level, significant decreases in levels of erythrocytes and Hgb
were observed (p<0.05). No exposure-related effects were observed for levels of blood urea
nitrogen, serum creatinine, serum glutamic oxaloacetic transaminase, or serum glutamic pyruvic
transaminase, indicators of renal and hepatic damage.
Dose-related changes were observed in immunological tests on spleen cells and in assays
of bone marrow (p<0.05); decreases were observed with respect to IgM antibody forming
cells/spleen in response to SRBC, lymphocyte proliferation response to the T cell mitogen Con
A and the B cell mitogen LPS, number of T lymphocytes, and femoral GM-CFU; and an increase
was observed in bone marrow cell DNA synthesis. These effects were not significant at 12
mg/kg/day but were dose related (p<0.05). Of all the immunological indices tested, only one
endpoint (stimulation index for lymphocyte proliferation of spleen cells in response to medium
containing 0.5 |o,g/mL Con A) was significantly decreased at 12 mg/kg/day (p<0.05). The
number of B lymphocytes was not affected, but the investigators commented that the number of
B lymphocytes in the controls was lower than for historical controls for their laboratory. This
study identifies a NOAEL of 12 mg/kg/day and a LOAEL of 195 mg/kg/day for hematologic
effects in mice exposed to benzene in drinking water for 30 days, and a LOAEL of 12 mg/kg/day
for immunological effects.
102
-------
4.2.4.2. Inhalation Exposure
Male C57BL/6J mice (7-8/group) were exposed to benzene at a concentration of 0, 10.2,
31, 100, or 301 ppm (0, 32.6, 99, 319, or 962 mg/m3) in whole-body dynamic inhalation
chambers 6 hours/day for 6 days (Rozen et al., 1984). Mice were bled within 30-90 minutes of
the last exposure to determine peripheral blood counts. Five animals with blood counts closest
to the group mean were selected for lymphocyte assays. Single-cell suspensions of bone marrow
and spleen cells from these animals were used for T and B cell enumeration and mitogen-
induced proliferative assays. Statistical significance (/X0.05) was determined by one-way
ANOVA.
Lymphocyte counts were depressed at all exposure levels. Lymphocyte counts were
significantly reduced to approximately 65% of controls in the 10.2 ppm (32.6 mg/m3) exposure
group and to 35% of controls at 100 or 300 ppm (319 and 962 mg/m3). Erythrocyte counts were
significantly stimulated (115% of control levels) at 10.2 ppm (32.6 mg/m3) and were depressed
only at 100 and 300 ppm (319 and 962 mg/m3). At exposures of 10.2 ppm (32.6 mg/m3), the
frequency of femoral B lymphocyte colony-forming cells was reduced to approximately 30% of
the control value. In contrast, the number of femoral B lymphocytes was not significantly
reduced in the low-exposure group, but the number was reduced to less than 10% of the control
value at 100 or 300 ppm (319 and 962 mg/m3). Similarly, splenic PHA-induced blastogenesis
was significantly depressed at 31 ppm (99 mg/m3) without a concomitant depression in numbers
of T lymphocytes. The numbers of T lymphocytes were reduced to less than 50% of controls in
the 100 and 300 ppm (319 and 962 mg/m3) exposure groups. These results demonstrate that
short-term exposure to inhaled benzene even at low exposure concentrations can cause
reductions in immune-associated processes. This study identifies a LOAEL for depression of
lymphocytes of 10.2 ppm (32.6 mg/m3), the lowest dose tested.
Rosenthal and Snyder (1985) investigated the effects of exposure to benzene at 0, 10, 30,
100 (320 mg/m3) or 300 ppm (960 mg/m3) on the immune response of male C57BL/6J mice
(5-7/group) to challenge with the facultative intracellular pathogen Listeria monocytogenes. All
mice (5-7/group) were exposed to benzene for 5 days (6 hours/day) before infection with L.
monocytogenes. At this point benzene exposure was stopped for half of the groups, and the other
half continued to be exposed for 7 days after infection, for a total of 12 days' exposure.
Bacterial proliferation in the spleen was measured at 1, 4, and 7 days after infection as an index
of the resistance of the mice to infection. Body and spleen weights were determined, T and B
lymphocytes were enumerated, spleen cellularity was determined, and spleen
monocyte/macrophage, polymorphs, lymphocytes, and nucleated RBCs were scored.
None of the benzene exposure treatments affected L. monocytogenes counts in the spleen
1 day after infection. Preexposure to benzene at 300 ppm resulted in a sevenfold increase in
spleen bacterial counts at 4 days after infection, but lower concentration had no significant
effect. However, with continued benzene exposure after infection, concentration of 30 ppm (96
mg/m3) or more resulted in dose-dependent increases in spleen bacterial counts at 4 days after
infection. By 7 days after exposure, spleen bacterial counts had returned to control levels in all
treatments. The authors suggested that benzene exposure caused a delay in the cell-mediated
immune response, as there was a temporary increase in spleen bacterial counts. Both T and B
103
-------
lymphocytes were particularly sensitive to benzene exposure. Lymphocyte counts were
depressed at benzene exposure concentrations of >30 ppm, and counts did not return to control
levels even after cessation of benzene exposure. This study identified a LOAEL of 30
(96 mg/m3) ppm and a NOAEL of 10 (32 mg/m3) ppm for effects on the immune system.
4.2.4.3. Summary of Immunotoxic Effects
As detailed above, dose-related adverse effects on spleen weight and cellularity and on
various measures of immune function have been observed following both oral and inhalation
exposures to benzene. Dose-related decreases in spleen weight and cellularity were consistently
observed following oral benzene exposure (Hsieh et al., 1988b: Fan, 1992; White et al., 1984).
Dose-related adverse effects on several measures of immune function have been observed,
including lymphocyte count, mitogen-stimulated lymphocyte proliferation, primary antibody
response to SRBCs, IL-2 production, femoral B lymphocyte colony-forming cells, number of T
lymphocytes, and proliferation of L. monocytogenes bacteria in the spleen (Hsieh et al., 1988b;
Fan, 1992; White et al., 1984; Rozen et al., 1984). The results indicate that exposure to benzene,
whether oral or inhaled, adversely affects the immune response.
4.3. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS AND
MODE OF ACTION
Benzene exposure results in adverse noncancer effects by all routes of administration.
Hematotoxicity and immunotoxicity have been consistently reported to be the most sensitive
endpoints of noncancer toxicity in limited studies in humans and experimental animals, and these
effects have been the subject of several reviews (Aksoy, 1989; Goldstein, 1988; Snyder et al.,
1993b; Ross, 1996). The bone marrow is the target organ for the expression of benzene
hematotoxicity and immunotoxicity.
Chronic exposure to benzene results in progressive deterioration in hematopoietic
function. Anemia, leukopenia, lymphocytopenia, thrombocytopenia, pancytopenia, and aplastic
anemia have been reported after chronic benzene exposure (see detailed discussions and
summaries in Sections 4.1.2.1 and 4.2.1 and Tables 1 and 6). In contrast to these blood
cellularity depression effects, benzene is also known to induce bone marrow hyperplasia. AML
has been frequently observed in studies of human cohorts exposed to benzene, and there is
evidence linking benzene exposure to several other forms of leukemia (U.S. EPA, 1998a).
Whether the hematotoxic/immunotoxic effects of benzene and its carcinogenic effects are caused
by a common mechanism is not yet known, due in part to the fact that although the bone marrow
depressive effects of benzene in humans can be readily duplicated in several experimental animal
model systems, a suitable experimental animal system for the induction of leukemia has not yet
been developed. In addition, the hematotoxicity/immunotoxicity of benzene leads to significant
health effects apart from potential induction of leukemia.
Although the decreased ALC and leukemias observed from benzene exposure both result
from bone marrow toxicity, they do not necessarily result from the same mechanisms, and
decreased ALC may not be a necessary precursor for leukemia. For example, decreased ALC
may be the result of cytotoxicity independent of the interaction of benzene metabolites with
DNA and/or DNA-associated proteins and tumor formation. Thus, integration of the cancer and
104
-------
noncancer health assessments of benzene is not considered possible given the current state of
knowledge regarding mode of action.
Leukocytopenia has been consistently shown to be a more sensitive indicator of benzene
toxicity than anemia in experimental animal systems, and lymphocytopenia has been shown to
be an even more sensitive indicator of benzene toxicity than is overall leukocytopenia (Snyder et
al., 1980; Ward et al., 1985; Baarson et al., 1984). Rothman et al. (1996a) also found that a
decrease in absolute lymphocyte count was the most sensitive indicator of benzene exposure in a
group of workers. Similarly, Ward et al. (1996) demonstrated a strong relationship between
benzene exposure and decreased WBC counts in a rubber worker cohort, but no significant
relationship with RBC counts was found. A common observation in experimental animal and
human studies on benzene and lymphocytopenia—except for the occupational study of Collins et
al. (1991)—has been the absence of a clear threshold.
Human and experimental animal studies indicate that benzene affects both humoral and
cellular immunity. Dose-related adverse effects on spleen weight and cellularity and various
parameters of immune function have been observed following both oral and inhalation exposure
to benzene.
Benzene toxicity to humans exposed in the workplace has been characterized as having
either early hematotoxicity or, with prolonged exposure to high doses, irreversible bone marrow
damage. Studies of worker populations in factories in which benzene was used as a solvent
(Snyder and Kocsis, 1975) showed a range of hematotoxic effects, including anemia, leukopenia,
and thrombocytopenia. In some cases, more than one cell type was decreased. A decrease in the
levels of all the classes of blood cell types in circulation is termed pancytopenia and is usually
associated with irreversible bone marrow aplasia. Aplastic anemia is fatal in most cases. In
those who survive aplasia, the marrow appears to be dysplastic. Myelodysplastic syndrome,
which has been called preleukemia, is likely an early stage of AML.
The evidence is strong that benzene metabolism plays a critical role in toxicity (Snyder
and Hedli, 1996). Hepatic metabolism plays an important role in toxicity (Andrews et al., 1977).
In addition to hepatic metabolism, it appears that secondary metabolism of benzene metabolites
in bone marrow contributes to toxicity (Irons et al., 1980; Schlosser and Kalf, 1989). Thus,
elucidation of the metabolic pathway for benzene biotransformation is essential for a full
understanding of the mechanism of toxicity.
There are alternative routes by which the first step of benzene metabolism, namely,
phenol formation, can occur. CYP2E1, and perhaps other cytochromes P450, can generate H2O2
when acting as oxidases of nicotinamide adenine dinucleotide phosphate (NADPH). The
hydroxyl radical formed from H2O2 can hydroxylate benzene to yield phenol. An alternative
mechanism for phenol formation reflects on the fate of the benzene oxide-oxepin system (see
Figure 1). When benzene oxide is the first product, it can rearrange nonenzymatically to form
phenol. Alternatively, benzene oxide can be further metabolized by epoxide hydrolase to yield
1,2-benzene dihydrodiol, which can in turn be oxidized via dihydrodiol dehydrogenase to form
catechol. The reaction of benzene oxide with GSH catalyzed by GSHS transferase leads to the
formation of the premercapturic acid. It is likely that benzene oxide or its oxepin (intermediate)
are precursors to ring opening (Witz et al., 1996). Phenol can be further hydroxylated to form
105
-------
hydroquinone or catechol. In theory, 1,2,4-benzenetriol may be formed by the hydroxylation of
either hydroquinone or catechol, but Inoue et al. (1989) suggested that catechol is not a precursor
of 1,2,4-benzenetriol in humans.
Any of the phenolic metabolites may be conjugated with either sulfate or glucuronide. In
addition to L-phenylmercapturic acid reported by Parke and Williams (1953), other
mercapturates include 6-N-acetylcysteinyl-S-2,3-cyclohexadienol (Sabourin et al., 1988a, b) and
2,5-dihydroxy-phenyl-mercapturic acid (Nerland and Pierce, 1990). The urine also contains two
ring-opening products, MA (Drummond and Finar, 1938) and 6-hydroxy-/rami, trans-2,4-
hexadienoic acid (Kline et al., 1993), and the residue of a covalently bound DNA adduct, i.e.,
N7-phenylguanine (Norpoth et al., 1988).
The production of benzene metabolites, largely in the liver, is followed by their transport
to the bone marrow and other organs. There are many possible causes of bone marrow toxicity.
Pfeifer and Irons (1983) suggested that covalent binding of hydroquinone to spindle fiber protein
could explain inhibition of cell replication by benzene. Damage to DNA could result in bone
marrow depression, leading to aplastic anemia, which in survivors leads to marrow dysplasia and
ultimately to AML (Snyder and Kalf, 1994). Benzene metabolites could cause damage to DNA
by two possible mechanisms. One pathway focuses on the metabolic activation of benzene to
species that covalently bind to DNA to produce mutagenic events that are expressed as leukemia.
The second mechanism involves the production of metabolites that cause oxidative stress,
subsequent oxidative damage to DNA, and a mutagenic effect that has the same consequences.
Within the bone marrow, both hematopoietic progenitor cells and bone marrow stromal
cells are potential targets of benzene toxicity (Aksoy, 1988; Gaido and Wierda, 1985).
Progenitor cells are thought to be the cells of origin for leukemias (Greaves, 1993) and are
attractive as potential targets of toxins such as benzene, whose toxic effects are not restricted to a
single hematopoietic lineage. Stromal cells are critical in the regulation of normal hemopoiesis
(Dorshkind, 1990) and have been considered as important targets of benzene toxicity (Gaido and
Wierda, 1985). Stromal cells are intimately associated with developing blood cells and regulate
hemopoiesis via direct cell-to-cell interactions, the production of extracellular matrix
components, and the secretion of soluble mediators such as cytokines and eicosanoids (Billips
etal., 1991).
Although metabolism is central to benzene toxicity, studies of the metabolic capability of
human bone marrow are scarce, particularly in human stromal and CD34+ cells.
Benzene hematotoxicity is a complex process that most likely involves interaction among
several metabolites (Billips et al., 1991). Possible intermediates considered at present to be
important in benzene hematotoxicity are polyhydroxylated benzene metabolites, for example,
hydroquinone and 1,2,4-benzenetriol; their quinone oxidation products, for example, p-
benzoquinone, formed via oxidation of hydroquinone, and semi quinone free radical
intermediates formed during the oxidation of polyhydroxylated metabolites to quinones; reactive
oxygen species formed during the oxidation of the polyhydroxylated metabolites; and ring-
opened aldehydic benzene metabolites, for example, MUC and 6-hydroxy-trans, trans-2.,4-
hexadienal.
106
-------
MUC and its metabolite 6-hydroxy-trans, ^ram'-2,4-hexadienal are reactive ring-opened
hematotoxic compounds. They exhibit a host of biological activities that could potentially be
important in their mechanisms of toxicity in relation to benzene.
Benzene hematotoxicity occurs when its hepatic metabolites (Sammett et al.,
1979)—phenol, catechol, and hydroquinone—are transported to the bone marrow (Greenlee et
al., 1981) and further oxidized in a peroxidase-mediated (Smith et al., 1989) reaction to
biologically reactive intermediates that can potentially affect hematopoiesis.
The ability to alter cytokine-dependent growth and differentiation in hematopoietic
progenitor cells appears to be a property of many agents, with leukemogenic potential for
humans (Irons and Stillman, 1993). Several studies have reported significant effects of benzene
on hematopoietic stem and progenitor cells (Cronkite et al., 1989). In one study (Seidel et al.,
1989), a dose-dependent depression of all stem cell compartments was observed in BDF1 mice
exposed for 16 weeks to airborne concentrations of benzene as high as 99 ppm (317 mg/m3) for 6
hours/day, 5 days/week. The results of this study indicated that the GM-CFU was much less
sensitive than the erythroid CPUs at higher doses of benzene.
Another study (Dempster and Snyder, 1990) reported that short-term exposure of mice to
benzene induced a shift toward granulocytic differentiation, a growth advantage for granulocytic
progenitor cells in the bone marrow and spleen, and to a resultant increase in the total number of
granulocytes. These results suggest that benzene, or hydroquinone, is acting on the myeloid
stem or progenitor cells. The administration of benzene, or hydroquinone, specifically
stimulates granulopoiesis in mice and induces granulocytic differentiation in myeloblasts of the
human promyelocytic leukemia cell line, HL-60, as well as the normal murine IL-3 dependent
myeloblastic cell line, 32D.3 (G). Benzene and hydroquinone do this by replacing the
requirement for granylocyte colony-stimulating factor and leukotriene D4, respectively, for
induction of differentiation.
In conclusion, benzene-induced mechanisms of hematotoxicity are poorly understood.
Most proposed mechanisms of benzene-induced effects such as cytotoxicity, apoptosis,
mutagenesis, and cell replication can be demonstrated in vitro at relatively high concentrations of
benzene metabolites (Snyder et al., 1993b). Benzene effects in vitro do not necessarily correlate
with benzene-induced hematotoxicity following in vivo exposures to low concentrations, as in
vitro assays of hematopoiesis lack the complex hematopoietic cell interrelationships, fine-tuned
regulation, and compensation mechanisms that are essential in normal hematopoiesis. The most
important aspects of benzene-induced hematotoxicity are the possible toxic effects caused by
low concentrations of benzene. Concurrent assessment of toxicity elicited by high
concentrations of benzene (100 [320 mg/m3] and 200 [640 mg/m3] ppm) established a reference
for comparison.
In common with many other organic solvents, benzene has been shown to produce
neurotoxic effects in test animals and humans after short-term exposures to relatively high
concentrations. The neurotoxicity of benzene, however, has not been extensively studied, and no
systematic studies of the neurotoxic effects of long-term exposure were located. Benzene
produces generalized symptoms such as dizziness, headache, and vertigo at levels of 250-3000
ppm (799-9584 mg/m3) (Brief et al., 1980), leading to drowsiness, tremor, delirium, and loss of
107
-------
consciousness at 700-3000 ppm (2236-9584 mg/m3). These neurological symptoms are
reversible upon removal of the subject from exposure. Kahn and Muzyka (1973) reported that
workers exposed to benzene for 2-9 years at 6-15.6 ppm (20-50 mg/m3) complained of frequent
headaches, became tired easily, had difficulties sleeping, and complained of memory loss. The
limitations of this study were that the workers were exposed to both benzene and toluene and the
dose and duration of exposure were unknown.
Few studies in test animals have examined the neurotoxicity of benzene, and those that
have used short-term exposures, usually to concentrations that have been shown to induce
significant hematotoxicity. One exception is the study by Li et al. (1992), which observed
biphasic responses in forelimb grip strength and frequency of rapid response in running a Y
maze following inhalation exposure to 0, 0.78, 3.13, or 12.52 ppm (0, 2.5, 10, or 40 mg/m3)
benzene 2 hours/day, 6 days per week for 30 days. Both responses increased at the lower
concentration and declined at the intermediate and high concentrations. There were, however, a
number of uncertainties in the experimental protocol that need to be resolved before these
observations can be used in establishing NOAEL or LOAEL values.
No complete neurological testing has been conducted in animals and humans. The
toxicological significance of these responses is also not clear. A detailed discussion of
neurotoxicity studies is presented in Section 4.2.3.
There is some evidence of reproductive and developmental benzene toxicity from human
epidemiology studies, but the data did not provide conclusive evidence of a link between
exposure and effects. In most cases, there was exposure to other chemicals as well, and the
quantitative data were not sufficient to determine a NOAEL or a LOAEL. Some test animal
studies provide limited evidence that benzene affects reproductive organs; however, these effects
were limited to high-exposure concentrations that exceeded the maximum tolerated dose.
Fertility studies that have shown adverse effects on the number of live fetuses used benzene
concentrations that caused severe maternal toxicity, as indicated by large reductions in body
weight gain. Studies that used lower benzene concentrations showed no reduction in fertility
(Coate et al., 1984; Green et al., 1978; Kuna et al., 1992; Murray et al., 1979).
Results of inhalation studies conducted in test animals are fairly consistent across species
and demonstrate that benzene is fetotoxic and causes decreased fetal weight and/or minor
skeletal variants at concentrations of greater than 47 ppm (150 mg/m3) (Coate et al., 1984; Green
et al., 1978; Kuna and Kapp, 1981; Murray et al., 1979). Exposure of mice to benzene in utero
during development has been shown to cause changes in the hematogenic progenitor cells in
fetuses, 2-day-old neonates, and 6-week-old adults (Keller and Snyder, 1986, 1988). However,
the biological significance of these effects is questionable because of the experimental design
limitations (see detailed discussion of studies in Section 4.2.2).
Although benzene exposure has been shown to result in structural and numerical
chromosomal abberations in human lymphocytes, the quantitative relationship between
measured benzene exposures and clastogenic effects in humans is unknown. Associations
between benzene exposure and genotoxic effects in humans under occupational conditions have
been demonstrated. In animal studies benzene has been shown to induce cytogenic effects,
including chromosome and chromatid aberrations, sister chromatid exchanges, and micronuclei
108
-------
in vivo and in vitro. Mutagenicity of benzene metabolites is well established. However,
benzene itself is not mutagenic in bacterial and animal systems (for a detailed discussion on
benzene genotoxicity, see Section 3.3.6.2 and U.S. EPA, 1998a).
4.4. SUSCEPTIBLE POPULATIONS
4.4.1. Childhood Susceptibility
This section reviews information regarding whether infants and children may be more
susceptible than adults to the hematotoxic effects of benzene. Developmental toxicity was
described in Section 4.2.2. For example, inhalation exposure of benzene in pregnant rats on GDs
6-15 resulted in fetotoxicity at 100 ppm (319 mg/m3). The study established a LOAEL of 100
ppm (319 mg/m3) and a NOAEL of 40 ppm (128 mg/m3) for the fetal toxicity of benzene.
Benzene could be a potential risk factor for the development of childhood leukemia (see reviews,
OEHHA, 1997; Smith and Zhang, 1998; U.S. EPA, 1998a). Evidence from human and animal
studies (McKinney et al., 1991; Shaw et al., 1984; Shu et al., 1988; Buckley et al., 1989)
suggests that increases in childhood leukemia may be associated with in utero exposures and
maternal and paternal exposure prior to conception.
Evidence in animals suggests that exposure to benzene in utero alters maturation of
lymphocytes, erythrocytes, and granulocytes (OEHHA, 1997). The consequences of in utero
exposure to benzene can be detected as alterations in cell population numbers and functional
properties into adulthood. Damage during the initial in utero stages of hematopoiesis could have
lasting effects, as has been demonstrated for a number of other toxicants (OEHHA, 1997).
Exposure of mice to benzene in utero during development has been shown to cause
changes in hematogenic progenitor cells in fetuses, 2-day-old neonates, and 6-week-old adults
(Keller and Snyder, 1986, 1988; Corti and Snyder, 1996). These results indicate that lasting
damage to the hematopoietic system can occur during development. It is not known whether the
effect is the same as in adults or is unique to developmental exposures. Dose-related
hematotoxicity was observed in the absence of any apparent maternal effects in Swiss-Webster
mice (Keller and Snyder, 1986, 1988). At high doses, significantly increased skeletal variations
in CF-1 mouse fetuses (Murray et al., 1979) and in rabbits (Ungvary and Tatrai, 1985) were
observed. Corti and Snyder (1996) found effects on precursor cells but provided no information
on peripheral blood effects. The absence of a clear NOAEL for fetal/neonatal exposure makes it
difficult to directly compare fetal/neonatal and adult toxicity.
There are few data on the effects of direct exposure of children to benzene. However,
some indirect evidence suggests that children may be susceptible to benzene-induced
hematotoxicity. There is mounting evidence that key changes related to the development of
childhood leukemia occur in the developing fetus. Several studies have reported that genetic
changes related to eventual leukemia development occur before birth. For example, there is one
study of genetic changes in sets of twins who developed T cell leukemia at 9 years of age
(reviewed in Smith and Zhang, 1998). Because of their small size, increased activity, and
increased ventilation rates, as compared with those of adults, children may have greater exposure
to benzene in the air on a unit-body-weight basis (U.S. EPA, 1998a). Infants and children also
109
-------
may be vulnerable to leukemia induction from benzene because their hematopoietic cell
populations are undergoing maturation and differentiation (U.S. EPA, 1998a).
4.4.2. Gender Differences
Human exposure data regarding differential effects of benzene exposure on males and
females are very limited. Some studies indicate that females may be at greater risk than males.
Sato et al. (1975) exposed male and female workers to 25 ppm (80 mg/m3) benzene for 2 hours
and showed that benzene was retained longer in the female subjects. During 5 hours of
exposure, the blood concentrations were higher in men, whereas at the end, exhalation breath
levels of both men and women were essentially equivalent. However, 4 hours after the exposure
had stopped, the female subjects' blood and exhaled air benzene levels were higher than those of
male subjects. The shape of the decay curve was significantly steeper in the males. Recently,
Brown et al. (1998), using the Sato et al. (1975) data, developed a PBPK model simulation for
adult men and women. Results demonstrated that physiochemical gender differences result in
women metabolizing 23-26% more benzene than do men when subjected to the same exposure
conditions. Benzene blood levels were generally higher in men. These results suggest that
women may be at significantly higher risk for certain effects of benzene exposure. In the
occupational benzene exposure study by Rothman et al. (1996a), 21 of the 44 workers in both the
exposed and control groups were female. The study, however, did not indicate that either gender
was more affected.
There are limited indications of gender differences in studies with test animals. The most
frequently observed gender difference is a greater sensitivity of male mice to benzene. Ward et
al. (1985) reported that microscopic examination revealed that changes in the thymus, bone
marrow, lymph nodes, spleen, and reproductive organs occurred more often and with greater
severity in males than in females following inhalation exposure at 300 ppm (958 mg/m3) of CD-I
mice for 13 weeks. Male DBA/2 mice showed greater depression in bone marrow cellularity
than did female DBA/2 mice in response to inhalation exposure to 300 ppm (958 mg/m3) for 13
weeks (Luke et al., 1988a).
Male mice were also more sensitive to benzene by oral exposure (NTP, 1986). In a 17-
week oral gavage study, the LOAEL for hematologic effects in male B6C3F1 mice was 50
mg/kg, compared with 400 mg/kg for females. Female rats, in contrast, appear to be more
sensitive than males to benzene toxicity. In the NTP (1986) 17-week oral gavage study, the
LOAEL for hematologic effects for female F344/N rats was 25 mg/kg, compared with 100
mg/kg for males.
Male mice are twofold to threefold more sensitive to the genotoxic effects of benzene, as
measured by micronuclei induction and sister chromatid exchanges (Luke et al., 1988a).
Available studies indicate that the differential susceptibility to benzene is subject to hormonal
regulation of the CYP2E1. Renal levels of CYP2E1 in males are 20-fold higher than in females,
but studies have indicated that renal levels of CYP2E1 can be induced in females by testosterone
treatment (Hu et al., 1993; Pan et al., 1992). Kenyon et al. (1996) found that male B6C3F1 mice
have an almost twofold faster rate of benzene oxidation than do females. Phenol disappearance
from the blood was also faster in male mice, suggesting that phenol metabolism is faster in
males. These differences in benzene metabolism correlated with the sensitivity to genotoxicity.
110
-------
Corti and Snyder (1996) found that CFU-E levels were depressed in livers of male
fetuses of Swiss-Webster mice exposed in utero to 10 ppm (32 mg/m3) benzene in comparison
with controls. Similarly, bone marrow CFU-E was also depressed in 6-week-old adult male
mice that had been exposed in utero. No depression in CFU-E was observed in female fetuses or
adult females. This suggests that differences in the sensitivity of male mice can be expressed in
utero and persist into adulthood. This result is in contrast to the results observed by Siou et al.
(1981), who found that immature male and female mice did not differ in sensitivity to benzene.
Green et al. (1978) reported that exposure of pregnant mice to 2200 ppm (7028 mg/m3)
benzene caused a greater increase in missing sternebrae in female fetuses than in male fetuses,
but this was not observed at concentrations of 300 ppm (958 mg/m3) or less. The dose causing
this effect induced extreme maternal toxicity, complicating interpretation of this effect. The
apparently increased sensitivity of the female fetus to the effects of benzene is supported by
other observations that female rabbits, mice, and rats are more sensitive to these effects of
benzene (Desoille et al., 1961; Ito, 1962a-d; Sato et al., 1975). The investigators tentatively
suggested that the differences in the responses of the males and females may be related to
hormonal differences.
4.4.3. Genetically Susceptible Populations
Significant sources of variability in the population stem from genetic polymorphisms in
key enzymes involved in the metabolism of benzene, namely, CYP2E1, NADPH-dependent
quinone oxidoreductase, MPO, GSH transferase, and others. Dietary and endogenous sources of
phenol, hydroquinone, and other primary metabolites of benzene confer potentially large
differences in susceptibility to benzene toxicity. These polymorphisms may increase or decrease
an individual's susceptibility to the toxic effects of benzene, as described below.
Smart and Zannoni (1984, 1985) demonstrated the importance of reductase activity in
bone marrow as a protective mechanism against benzene toxicity. Recent studies have also
shown the importance of polymorphic forms of GSH transferase as a protective characteristic.
By the same token it appears that the reason that the liver is not a target for benzene toxicity is
its high level of reductase activity, compared with the relatively low level in the bone marrow,
where oxidative reactions predominate. Furthermore, London et al. (1997) described a
polymorphism in MPO, an enzyme important in activation of polyhydroxylated metabolites of
benzene to quinones, suggesting that people who genetically display lower levels of this enzyme
may display a lower risk to carcinogenesis. Although biomarkers for susceptibility to benzene
have not been validated, it might be hypothesized that an individual who displays high CYP2E1
activity and low GSH transferases in liver and/or bone marrow and low bone marrow reductase
and high MPO could be highly susceptible to benzene toxicity.
Ross et al. (1996) have suggested that NAD(P)H:quinone oxidoreductase (NQO1
NAD(P)H) may play a critical protective role in benzene toxicity. This enzyme detoxifies the
reactive 1,4-benzoquinone generated by MPO by reducing it back to hydroquinone. Ross (1996)
summarized evidence suggesting that the target cells of benzene toxicity have a high ratio of
MPO:NQO1 activities in vivo. Traver et al. (1992) characterized a point mutation in the NQO1
gene that leads to a total loss of NQO1 activity. This appears to be a true polymorphism in the
NQO1 gene, because Rosvold et al. (1995) found that the frequency of this mutant allele in a
111
-------
reference population was 13% and Edwards et al. (1980) reported thatNQOl was absent in 4%
of samples taken from a British population. If the postulated role of NQO1 in detoxification of
benzene metabolites is correct, then individuals who lack this enzyme activity could be
especially susceptible to benzene toxicity. Rothman et al. (1997) carried out a case-control study
in a population of benzene-exposed workers in Shanghai, China. CYP2E1 and NQO1 genotypes
were determined by polymerase chain reaction-restriction fragment length polymorphism
analysis, and CYP2E1 activity was estimated by the fractional excretion of chlorzoxanone.
Investigators found that subjects who had both rapid chlorzoxanone excretion and two copies of
the NQO1 609OT mutation had a 7.6-fold increased risk of hematotoxicity when compared with
subjects who had slow chlorzoxanone excretion and who carried one or two wild-type NQO1
alleles.
The joint effects of CYP2E1 and NQO1 status, shown in Table 9, indicate that rapid
CYP2E1 activity or a nonfunctional NQO1 increase an individual's risk of benzene
hematotoxicity. Those individuals who had rapid CYP2E1 activity and heterozygous or
homozygous variant NQO1 genotypes were at greatest risk of developing benzene-induced
hematotoxicty.
Seaton et al. (1994) reported differences of up to 13-fold in liver microsome CYP2E1
among individual humans. Hepatic metabolism of benzene by the CYP2E1 enzyme has been
shown to be the first step in generation of reactive benzene metabolites that are responsible for
the toxicity of benzene. Thus, differences in CYP2E1 between individual humans could indicate
potential differential susceptibility to benzene toxicity.
Rossi et al. (1999) investigated the role of genetic polymorphisms in modulating urinary
excretion of two benzene metabolites—MA and S-phenylmercapturic acid—in 59 nonsmoking
city bus drivers, who were professionally exposed to benzene via vehicle exhausts. Genetic
Table 9. Joint effects of CYP2E1 activity and NQO1 genotype on benzene-induced
hematotoxicity in Chinese Workers
CYP2E1 activity
Slow
Slow
Rapid
Rapid
NQO1 genotype
Wild type
Variant
Wild type
Variant
Cases
8
6
21
13
Odds ratio (95% CI)
benzene hematotoxicity"
1.0
2.4 (0.6-9.7)
2.9(1.0-8.2)
7.6(1.8-31.2)
aAdjusted for the matching variables age and sex.
Source: Rothman etal., 1997.
polymorphisms at six loci encoding cytochrome P450-dependent monooxygenases (CYP2E1 and
CYP2D6), GSH transferases (GSTT1, GSTP1, and GSTM1), and NQOR were determined by
112
-------
polymerase chain reaction based methods. Metabolic variability was observed for CYP2D6,
GSTT1, and NQOR, potentially contributing polymorphism as one of the risk factors for
benzene-induced adverse health effects. However, no evidence emerged for a possible role of
CYP2E1, GSTM1, and GSTP1 polymorphisms in determining the wide differences observed in
the biotransformation.
4.5. HAZARD IDENTIFICATION SUMMARY
Studies in experimental animals and human occupational studies have pointed to the
changes in peripheral blood and bone-marrow-induced exposure to benzene as being among the
most sensitive indicators of the compound's toxicity. Principal among the adverse effects
described are leukopenia, lymphocytopenia, granulocytosis, anemia, and reticulocytosis.
Human occupational exposure studies have revealed clear evidence of benzene-induced
hematotoxicity (Table 1). Difficulties arise in directly applying most of these findings to
quantitative estimation of potential human health impacts. This is due to uncertainty in defining
the level of benzene exposure and to potential exposure to mixtures of other potentially harmful
substances. One human occupational study, however, reported evidence that can be used to
support the derivation of RfC and RfD values for benzene. In a small cross-sectional study of 44
age- and gender-matched controls, Rothman et al. (1996a) observed a dose-response relationship
and inverse correlation between hematologic responses and exposure level. Six blood
parameters (ALC, WBC count, RBC count, HCT, platelets, and MCV) were significantly
different in the high-benzene exposure group (8-hour TWA of 91.9 ppm [294 mg/m3]) in
comparison with controls. However, in the lower-exposure group (median 8-hour TWA of 13.6
ppm [43.4 mg/m3], only ALC, RBCs, and platelets were significantly different.
In a second case-control study of a rubber worker cohort employed between 1939 and
1976 that relied on controversial exposure estimates (Ward et al., 1996), a strong exposure-
response relationship between WBC count and estimated benzene exposure concentration for 30,
90 and 180 days before the blood test date was evident; however, there was only a weak positive
exposure response for RBC count. The maximum daily exposure estimate was 34 ppm (109
mg/m3). There was no evidence for a threshold, suggesting that exposure to relatively low levels
of benzene (e.g., < 5 ppm [16 mg/m3]) could result in hematologic suppression, if the exposure
estimates are correct.
Further support for the hypothesis that low exposures to benzene can reduce lymphocyte
counts comes from a recent study (Bogadi-Sare et al., 2000) of female employees in the
shoemaking industry. This study found that the number of circulating B lymphocytes was lower
in 49 shoe workers exposed to benzene concentrations lower than 15 ppm (48 mg/m3) as
compared with 27 nonexposed controls. The authors concluded that benzene concentrations
lower than 15 ppm (48 mg/m3) can induce depression of circulating B lymphocytes. Additional
evidence that low exposures to benzene induce reductions of blood parameters comes from a
study of workers employed in factories in China (Dosemeci et al., 1996). Significant relative
risks of "benzene poisoning" were recorded in workers exposed to as little as 5-19 ppm (16-61
mg/m3) during the most recent 1.5 years as compared with workers exposed to < 5 ppm (16
mg/m3). The authors defined benzene poisoning by two criteria: a WBC count of < 4000
cells/mm3 and a platelet count of < 80,000/mm3. These data suggest that a threshold of
113
-------
hematotoxicity in humans is observable in the 5-19 ppm (16-48 mg/m3) range of exposure to
benzene.
Hematotoxicity has been observed in experimental animal studies following both oral and
inhalation exposure (Table 6). Hematologic abnormalities developed rapidly and were evident
after only 2 weeks of exposure, but there was comparatively little progression with longer
exposure periods (Ward et al., 1985; Cronkite et al., 1985). The dose-response studies of Ward
et al. (1985) identified a NOAEL of 30 ppm (96 mg/m3) and a LOAEL of 300 ppm (958 mg/m3)
in CD-I mice. The results of Cronkite et al. (1985, 1989) suggest that peripheral blood elements
may have the capacity to recover from benzene-induced toxicity, depending on the duration of
exposure and concentration. For example, lymphocyte counts, which were depressed in
C57BL/6 mice exposed to benzene at 300 ppm (958 mg/m3) for 2-16 weeks, tended towards
normal values during a 4- to 8-week recovery period.
Benzene-induced peripheral blood abnormalities reflect a disruption at all levels of
hematopoiesis in the bone marrow (Toft et al., 1982; Snyder, 1987; Cronkite et al., 1985; Seidel
et al., 1989; MacEachern et al., 1992). Bone marrow cellularity, stem cell compartments,
granulocytic and erythropoietic progenitor cells, and bone marrow macrophages have all been
observed to be adversely affected following benzene exposure. Recent evidence has suggested
that the inhibitory effect of benzene on lymphocyte proliferation can be mediated through the
suppression of cytokine production (Renz and Kalf, 1991). The fact that both in vivo and in
vitro studies have demonstrated the ability of some metabolites to induce toxicological effects
qualitatively similar to those of benzene provides evidence consistent with the concept that the
hematotoxicity of benzene may be mediated through its metabolites.
There is no convincing evidence that benzene produces either reproductive or
developmental toxicity in humans. The available data are summarized in Tables 4 and 5. Most
studies consisted of small numbers of subjects, lacked important experimental details, involved
(in almost all cases) concomitant exposure to other chemicals, and did not provide monitoring
data or quantitative dose-response information.
A number of reproductive and developmental studies have been conducted in a wide
range of experimental animals (Tables 7 and 8). Changes in testicular weight in guinea pigs
(with slight histopathologic alterations) and rats were observed by Wolf et al. (1956), and
reductions in ovarian and testicular weights with histopathologic alterations were observed in
mice by Ward et al. (1985) after prolonged exposure to benzene. However, the concentrations
were higher than those observed to cause hematotoxicity. Several studies have suggested that
benzene can cause developmental toxicity in the absence of maternal toxicity in rats. Effects
observed include reduced body and liver weights in weanlings (Kuna et al., 1992), skeletal
variations in fetuses (Kuna and Kapp, 1981), and reduced fetal weight and delayed ossification
in fetuses (Green et al., 1978; Coate et al., 1984; Ungvary and Tatrai, 1985). Although benzene
can cause reproductive and developmental toxicity, the LOAEL and NOAEL values associated
with reproductive/developmental toxicity are higher than for hematotoxicity.
The hematologic effects in offspring of dams exposed to benzene during gestation were
observed at exposure concentrations similar to those in adult animals (Keller and Snyder, 1986,
1988; Corti and Snyder, 1996). Changes in granulopoiesis and hematopoiesis in offspring at
114
-------
some growth stages were observed. Thus, the results were not consistent. The results of these
studies cannot be used for quantitative evaluation of human health risk because of the limited
number of replicate animals.
Symptoms of neurotoxicity have been observed in humans following acute exposure to
relatively high concentrations of benzene (Brief et al., 1980; Kraut et al., 1988). Kahn and
Muzyka (1973) reported subjective symptoms—headache, fatigue, difficulty sleeping, and
memory loss—in factory workers exposed to benzene. Baslo and Aksoy (1982) reported
neurological abnormalities in patients diagnosed with aplastic anemia after prolonged exposure
to benzene. These reports, however, have obvious deficiencies: lack of exposure data, small
numbers of subjects, and unknown exposure to other chemicals. Thus, no reliable quantitative
evaluation of neurotoxicity in humans is possible.
Neurotoxicity studies with experimental animals were also very limited. The exposure
levels were generally high and exposure durations short. The longest exposure period was 4
weeks (Hsieh et al., 1988a). These investigators observed significant increases in monoamine
neurotransmitters in the brain of rats at doses of 8 and 40 mg/kg/day, but further increases were
not observed at 180 mg/kg/day. Benzene-induced behavioral and learning disorders in mice
following inhalation exposures of 100 ppm (319 mg/m3) or greater for 5 days (Evans et al., 1981;
Dempster et al., 1984). Li et al. (1992) observed increased grip strength and enhanced rapid-
response maze performance after exposure to 0.78 ppm (2.5 mg/m3) for 30 days and decreases in
these parameters at higher concentrations. There were, however, several limitations in the
experimental procedures used in these experiments. There is a limited body of evidence
indicating that benzene is neurotoxic; however, there are no human or animal studies that could
be used for quantitative evaluation of potential human health risks.
No comprehensive immunotoxicity studies in human populations have been reported, but
the study by Rothman et al. (1996a) indicates decreased ALC as the most sensitive indicator of
benzene toxicity. This could be interpreted as an effect on immune function.
Immunotoxicity studies in experimental animals have demonstrated dose-related adverse
effects on spleen weight and cellularity and on several measures of immune function following
both oral and inhalation exposure to benzene (Hsieh et al., 1988b; Fan, 1992; White et al., 1984;
Rosenthal and Snyder, 1985; Rozen et al., 1984). NOAEL and LOAEL values for adverse
effects on immune function were similar to those established for hematotoxic effects. These
studies indicate that even short-term exposures to benzene adversely affect the immune response
in experimental animals.
There is no convincing evidence to indicate that children are more susceptible to the toxic
effects of benzene; however, there is evidence that differences in gender and subpopulation
susceptibility may exist. Differences in responsiveness to benzene have been observed among
species. Mice were found to be more sensitive than rats (Ward et al., 1985; Snyder et al., 1978,
1980). Intraspecies variability has also been demonstrated. AKR/J mice were found to be more
sensitive than C57BL/6 mice (Snyder et al., 1978, 1980). Absorption studies with humans
suggest that absorption by females is higher than by males, and modeling results indicate that
females metabolize 23-26% more benzene than men under the same exposure conditions (Sato
et al., 1975; Brown et al., 1998). Differences in benzene metabolism could result in differences
115
-------
in susceptibility, because conversion of benzene to metabolites is necessary for expression of
toxicity. Several studies have indicated that male mice are more sensitive than females (Ward et
al., 1985; NTP, 1986; Siou et al., 1981; Kenyon et al., 1996; Corti and Snyder, 1996). In
contrast, other studies using mice, rats, and rabbits have indicated that females are more sensitive
than males (Desoille et al., 1961; Ito, 1962a-d; Sato et al., 1975).
The balance of NQO1 NAD(P)H activity to peroxidase activity in cells may be important
in determining benzene toxicity by modulating the concentration of 1,4-benzoquinone, a
suspected toxic metabolite. A mutant allele for lack of NQO1 NAD(P)H activity has been
reported in a human population at a frequency of 13% (Rosvold et al., 1995). Edwards et al.
(1980) reported that 4% of the British population lack NQO1 NAD(P)H activity. Rothman et al.
(1997) also observed that Chinese workers homozygous for the mutant allele and with high
chlorzoxazone excretion (a measure of CYP2E1 activity) had a 7.6-fold higher risk of
developing benzene toxicity. Thus, there is good experimental evidence to indicate that
benzene-sensitive human subpopulations may exist.
Several experimental animal studies have observed that ethanol and benzene have an
interactive effect on the production of hydroxylated benzene metabolites (Nakajima et al., 1987).
When administered at the same time, ethanol and benzene enhanced the number of erythroid
progenitor cells in C57BL/6J mice (Baarson and Snyder, 1991), decreased the number of CFU-E
cells per femur in BDF1 mice (Seidel et al., 1990), and produced a 70% reduction of CFU-E in
male Swiss-Webster mice livers (Corti and Snyder, 1996). On the other hand, a 40% elevation
of CFU-E occurred in female Swiss-Webster mice exposed to benzene and ethanol compared to
unexposed controls (Corti and Snyder, 1996). There are no relevant published data on the
interactive effects of benzene exposure and ethanol use in the human population.
5. DOSE-RESPONSE ASSESSMENTS
The primary target organs for benzene toxicity are the hematopoietic, immune, and
nervous systems. There are several reports of short-term exposure in humans and animals. Data
on the effects of benzene on the human hematopoietic system following occupational inhalation
exposure are scant, but they indicate effects such as leukopenia, anemia, and thrombocytopenia.
Data on hematologic effects in experimental animals from short-term inhalation/oral exposures
are extensive, indicating changes in peripheral erythrocytes (e.g., Cronkite et al., 1985; Rozen et
al., 1984), in peripheral leukocytes (e.g., Aoyama, 1986; Li et al., 1986; Green et al., 1981a), and
in bone marrow cells (e.g., Cronkite et al., 1989; Dempster and Snyder, 1991). However, these
studies were generally determined to be unsuitable for derivation of the RfD/RfC because they
involved short-term exposures and small numbers of experimental animals; in addition, exposure
levels were often too high to establish meaningful LOAELs or NOAELs.
Data on adverse hematologic effects in humans following intermediate and chronic
duration exposures to benzene in occupationally exposed individuals are available (Aksoy et al.,
1971, 1972, 1974, 1987; Cody et al., 1993; Greenberg et al., 1939; Kipen et al., 1989; Li et al.,
1994; Townsend et al., 1978; Yin et al., 1987b). However, in the majority of these studies the
exposure level estimates and duration of exposure are poorly defined and/or characterized, and
potential confounding is problematic; therefore, these studies could not be used in determinating
116
-------
the RfC and RfD, with the exception of the Rothman et al. (1996a) study, which is used as the
key study. The study of occupational inhalation exposure to benzene by Tsai et al. (1983) seems
to suggest a NOAEL for hematologic effects; however, the study had no appropriate screening
guidelines or referral criteria with which to contrast results, no follow-up, a very pronounced
healthy-worker effect, and a very young employed survivor population. No relevant human data
are available to evaluate hematologic effects following oral exposure.
Several subchronic and chronic studies in experimental animals following inhalation
exposure (see Table 6) and oral exposure (NTP, 1986; Huff et al., 1989; Wolf et al., 1956; see
also Section 4.4.2.1) identified NOAELs and LOAELs for hematologic effects and were
considered in the derivation of the RfC and RfD. Ultimately, the results of the Ward et al.
(1985) subchronic inhalation study and the NTP (1986) chronic gavage study were selected as
the most appropriate animal data for deriving an RfC and an RfD. These RfC and RfD values
are calculated for comparison with the values derived from the Rothman et al. (1996a) human
study.
Whether the hematotoxic/immunotoxic effects of benzene and its carcinogenic effects are
due to a common mechanism is not yet known. This is in part due to the fact that although the
bone marrow depressive effects of benzene in humans can be readily duplicated in several
experimental animal model systems, a suitable experimental animal system for the induction of
leukemia has not yet been developed. In addition, the hematotoxicity/immunotoxicity of
benzene leads to significant health effects apart from potential induction of leukemia. Although
the decreased ALC and leukemias observed from benzene exposure both result from bone
marrow toxicity, they do not necessarily result from the same mechanisms, and decreased ALC
may not be a necessary precursor for leukemia. For example, decreased ALC may be the result
of cytotoxicity independent of the interaction of benzene metabolites with DNA and/or DNA-
associated proteins and tumor formation. Thus, integration of the cancer and noncancer health
assessments of benzene is not considered possible given the current state of knowledge regarding
mode of action.
5.1. INHALATION REFERENCE CONCENTRATION (RfC)
5.1.1. Choice of Principal Study and Critical Effect
The human occupational inhalation study by Rothman et al. (1996a) was selected as the
principal study for the derivation of the RfC because it is a well-conducted human exposure
study that demonstrates a dose-response relationship for hematologic effects, responses that are
considered to be among the more sensitive indices of benzene toxicity. The California EPA
(2001) selected the Tsai et al. (1983) study to develop a drinking water noncancer protective
concentration of 0.026 mg/L, because no increased hematological changes were observed among
U.S. refinery workers chronically exposed to an average 0.53 ppm of benzene. This value was
assumed to be a NOAEL by the California EPA. The U.S. EPA did not choose this study to
develop an RfC or an RfD because of deficiencies in the study design. However, the California
EPA's numbers were similar to those of the U.S. EPA, although different studies were used. The
cross-sectional Rothman et al. (1996a) study provides exposure-response data for some of the
lowest exposure concentrations at which effects have been observed, as discussed in Section
4.1.2.1. These exposure-response data are suitable for BMD modeling to derive a point of
117
-------
departure for the calculation of the RfC. Reduced ALC is the most sensitive of the hematologic
effects reported in this study, and it was selected as the critical effect. For comparison purposes,
a chronic inhalation RfC is also derived from the subchronic experimental animal study of Ward
et al. (1985), based on the critical effect of decreased HCT (see Section 5.1.5).
Rothman et al. (1996a) compared the hematologic outcomes of 44 workers
occupationally exposed to benzene with those of 44 age- and gender-matched unexposed
controls, all based in Shanghai, China. ALC, WBCs, RBCs, and platelets were all significantly
decreased and MCV was significantly increased in the exposed group (median 8-hour TWA of
31 ppm [99 mg/m3]) as compared to those in the control group. These effects are consistent with
the hematotoxic effects of benzene shown in Aksoy (1989), Goldstein (1988), and Dosemeci et
al. (1997). In the low-dose subjects (< 31 ppm [99 mg/m3], median 8-hour TWA of 13.6 ppm
[43.4 mg/m3]), the ALC, RBCs, and platelet count were reduced compared with controls.
Similarly, in a selected subgroup exposed to a median 8-hour TWA of 7.6 ppm (24 mg/m3)
benzene, a statistically significant difference in ALC versus controls was observed (p<0.03).
The study by Rothman et al. (1996a) is notable among epidemiology studies because
benzene exposures were monitored; exposure to other chemicals, including toluene, was
minimal; and subjects were compared with matched controls. Furthermore, a dose-response
relationship was established between ALC and benzene exposure, as monitored by organic vapor
passive dosimetry and the level of benzene metabolites in the urine. As discussed in Section
4.1.2.1, the median 8-hour TWA of 7.6 ppm (24 mg/m3) was designated the LOAEL for these
effects and is used herein to calculate a chronic inhalation RfC for benzene hematotoxicity in
humans for comparison with the RfC derived by BMD modeling of the ALC exposure-response
data.
The choice of a reduction in ALC as the primary effect in an RfD/RfC derivation is
partially based on the response's potential role as a "sentinel" effect for a cascade of early
hematologic and related biological changes that might be expected to result in the more profound
examples of benzene poisoning observed in other cohorts of the National Cancer
Institute/Chinese Academy of Preventive Medicine study, as described by Dosemeci et al.
(1996). Considered together, the statistically significant inverse correlation between ALC and
the level of benzene metabolites in the urine of exposed subjects in the Rothman et al. (1996a)
study and the strong relationship between exposure to benzene in persons with benzene
poisoning observed by Dosemeci et al. (1996) point clearly to the utility of fluctuations in ALC
as a marker for the onset of potentially harmful hematologic changes. That ALC depletion is
accompanied by gene-duplicating mutations in somatic cells under the same range of exposure
conditions suggests that benzene can cause repeated damage to longer-lived stem cells in human
bone marrow, further implicating the compound as etiologically important in the onset of
benzene-associated leukemia. These findings underline the importance of basing public health
concern for benzene on a toxicological effect that is representative of the earliest biological
changes induced by the compound.
In summary, there is overwhelming evidence, in both experimental animals and humans,
in the published literature cited in this report and in the cancer update documents (U.S. EPA,
1998a, 1999a), that chronic exposure to benzene leads to the onset of irreversible bone marrow
118
-------
depression, which is characterized by anemia, leukocytopenia with emphasis on
lymphocytopenia, and/or thrombocytopenia.
It is a well-known fact that benzene is a leukemogen. The mechanistic steps that lead to
that outcome involve decreases in circulationg blood cells, as characterized by abnormal marrow
architecture, inadequate hematopoiesis, and demonstration of chromosomal damage in many
cells. Benzene toxicity is a continuum of events, leading from decreases in circulating blood
cells to pancytopenia and aplastic anemia or to MDS and acute nonlymphocytic leukemia.
Although the role of several metabolites of benzene and the "dose x times" relationship are only
partially understood, the ultimate result as a function of marrow dysplasia or aplasia or leukemia
is undisputable. In the Rothman et al. (1996a) study, benzene air levels were inversely
correlated with the absolute lymphocyte counts among exposed workers. This has been the
subject of more detailed discussion in several sections of this document as well as in the cancer
update documents (U.S. EPA, 1998a, 1999a). The reader's attention is drawn to sections 4.1,
4.2, 4.5, as well as to some selected references: Rothman et al. (1996a, b), Dosemeci et al.
(1997), Bogardi-Sari et al. (2000), Hsieh et al. (1988b), NTP (1986), and Ward et al. (1985).
5.1.2. Benchmark Dose Modeling
The exposure-response data from Rothman et al. (1996a, Table IV) that were used for the
BMD modeling are reproduced below.
Median exposure ALC Transformed
(ppm: 8-hr TWA) Number of subjects (mean ± SD x 103/|iL blood) exposures
0.02 (control) 44 1.9 ±0.4 0.0198
13.6 22 1.6 ±0.3 2.68
91.9 22 1.3 ±0.3 4.53
The modeling was done using EPA's Benchmark Dose Modeling Software (version
1.20). The data are fairly supralinear, that is, the change in ALC per unit change in exposure
decreases with increasing exposure; therefore, in order to fit the data with one of the available
continuous models, the exposure levels were first transformed according to the equation
d' = ln(d + 1). Then the exposure-response data were fit using the continuous linear model,
which has the form
Y[dose] = po + p^dose,
where, in this case, Y[dose] is the mean ALC and dose is the transformed exposure, d'. The
parameters were estimated using the method of maximum likelihood. A constant variance model
was used. The resulting parameter estimates are:
Variable Estimate SE
119
-------
P0 1.91029 26.8966
-0.129822 70.7846
The model fit was good, as is apparent from the/>-value of 0.5443 and the graphical display in
Figure 2.
In the absence of a clear definition for an adverse effect for this continuous endpoint, a
default benchmark response (BMR) of one standard deviation change from the control mean was
selected, as suggested in EPA's draft Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b). This default definition of a benchmark response for continuous endpoints
corresponds to an excess risk of approximately 10% for the proportion of individuals below the
2nd percentile (or above the 98th percentile) of the control distribution for normally distributed
effects (see U.S. EPA, 2000b). The software uses the estimated standard deviation. For the
resulting BMC, a 95% lower confidence limit (BMCL) was calculated using the likelihood
profile method. A BMC of 2.69 and a BMCL of 2.10 were obtained with the transformed
exposures. Transforming these values back to the original exposure scale yields the following
values: BMC = 13.7 ppm, 8-hour TWA; BMCL = 7.2 ppm, 8-hour TWA.
A two-degree restricted polynomial model also adequately fits the data, yielding a BMC
of 3.00 and a BMCL of 2.14 with the transformed doses (i.e., 19.1 and 7.5 ppm 8-hour TWA,
respectively); however, the linear model was selected because it is the most parsimonious (i.e., it
120
-------
Linear Model with 0.95 Confidence Level
c
3
o
o
o
_c
Q.
£
1.8
CD
-t— '
1 1.6
Ifl
-Q
TO
CD
to
c
o
Q.
w
CD
cr
CO
CD
1.4
1.2
Linear
BMDL
BMD
01234
transformed exposure level
Figure 2. Linear model of ALC data from Rothman et al. (1996a).
is the model with the fewest parameters that fits the data) and 2 x (the absolute difference in the
log-likelihood values between the two models) < Xi2, that is, the parameter P2 for the coefficient
of dose2 in the two-degree polynomial model is not statistically different from 0. Results using a
restricted power model were identical to those using the two-degree polynomial.
Note that the 7.6 ppm exposure subgroup (11 subjects) was not included explicitly in the
exposure-response modeling because it was a subset of the 13.6 ppm group (22 subjects) and the
data were not available for the remaining 11 subjects in the 13.6 ppm group (nor were the
individual data available). Thus, using the 7.6 ppm group rather than the 13.6 ppm group would
have meant that some of the data, for one-fourth of the exposed subjects, would not have been
used at all. Comparison analyses were conducted to examine the effects of using the 7.6 ppm
group and either omitting the subjects in the 2nd exposure quartile or estimating the exposure
level of the 2nd quartile as the midpoint between 13.6 and 31 ppm. The BMCLs were similar to
those obtained when the 7.6 ppm group was not included separately (see Table 10).
5.1.3. RfC Derivation
As suggested in the draft BMD technical guidance document (U.S. EPA, 2000b), the
BMCL is chosen as the point of departure for the RfC derivation. An adjusted BMCL is
calculated by converting ppm to mg/m3 and adjusting the 8-hour TWA occupational exposure to
Table 10. Results of BMC modeling of Rothman et al. (1996a) data on benzene and
ALCa
121
-------
Dataset
without 7.6 ppm subgroup
with 7.6 ppm subgroup,
omitting 2nd quartile
with 7.6 ppm subgroup,
estimating 2nd quartile
BMC (ppm)
13.7
14.7
17.1
BMCL (ppm)
7.2
7.4
8.4
aUsing log-transformed exposures, a linear model, and a BMR level of one standard deviation change from the
control mean.
an equivalent continuous environmental exposure. The BMCL is first converted to mg/m3 using
the molecular weight of 78.11 for benzene and assuming 25°C and 760 mm Hg:
BMCL (mg/m3) = 7.2 ppm x 78.11/24.45 = 23.0 mg/m3
The converted value is then adjusted from the 8-hour TWA to an exposure concentration
adjusted for continuous exposure using the default occupational minute volume (U.S. EPA,
1994).
j = BMCL (mg/m3) x (VEho/VEh) x 5 days/7 days
j = 23.0 mg/m3 x (10 m3/20 m3) x 5 days/7 days = 8.2 mg/m3
where:
= the BMCL dosimetrically adjusted to account for continuous exposure
BMCL = occupational exposure level (8-hour TWA)
VEho = human occupational default minute volume (10 m3/8 hours)
VEh
= human ambient default minute volume (20 m3/24 hours).
To calculate an RfC using the BMCLADJ value of 8.2 mg/m3, several UFs were applied.
First, because the BMC is considered to be an adverse-effect level, an effect-level
extrapolationfactor analogous to the LOAEL-to-NOAEL UF is used. EPA is planning to
develop guidance for applying an effect-level extrapolation factor to a BMD. In the interim, a
factor of 3 is used in this analysis, based on the professional judgement that, while the BMC
corresponds to an adverse-effect level at the low end of the observable range, the endpoint is not
very serious in and of itself. Decreased ALC is a very sensitive sentinel effect that can be
measured in the blood, but it is not a frank effect, and there is no evidence that it is related to any
functional impairment at levels of decrement near the BMR. For a more serious effect, a larger
factor, such as 10, might be selected.
122
-------
Second, a factor of 10 was used for intraspecies differences in response (human
variability) as a means of protecting potentially sensitive human subpopulations. Third, a
subchronic-to-chronic extrapolation factor was applied because the mean exposure duration for
the subjects in the principal study was 6.3 years, which is less than the exposure duration of 7
years (one tenth of the assumed human life span of 70 years) which has been used by the
Superfund program as a cut-off for deriving a subchronic human reference dose (U.S. EPA,
1989). However, because the mean exposure duration (range 0.7 to 16 years) was near the
borderline of what would be considered chronic (i.e., 6.3 years vs. 7 years), a partial value of 3
(vs. 10) was felt to be appropriate for the UF.
Finally, a UF of 3 was chosen to account for database deficiencies because, despite the
extensive nature of the overall toxicological database for benzene, no two-generation
reproductive/developmental toxicity study is available. Therefore, an overall UFof3xiOx3x
3 = 300 is used to calculate the chronic inhalation RfC, as follows:
RfC = BMCLADJ H- UF = 8.2 mg/m3 - 300 = 3 x 102 mg/m3
5.1.4. Comparison Analysis Based on the LOAEL
A median 8-hour TWA concentration of 7.6 ppm was designated the LOAEL for the
Rothman et al. (1996a) study for a comparison calculation of a chronic inhalation RfC for the
critical effect of reduced ALC. A LOAEE^ can be calculated, as described for the BMCLADJ in
Section 5.1.3, by converting to mg/m3 and adjusting from the occupational ventilation rate and
intermittent work-week schedule to a continuous 24-hour exposure, 7 days/week:
LOAEL (mg/m3) = 7.6 ppm x 78.11/24.45 = 24.3 mg/m3
LOAELADj = 24.3 mg/m3 x (10 m3/20 m3) x 5 days/7 days = 8.7 mg/m3
To calculate an RfC using the LOAELADJ value of 8.7 mg/m3, the following UFs were
selected. A factor of 10 was applied to account for using a LOAEL because of the lack of an
appropriate NOAEL. A factor of 10 was used for intraspecies differences in response (human
variability). A partial UF of 3 for subchronic-to-chronic extrapolation was applied and a UF of 3
was chosen to account for database deficiencies, as discussed in Section 5.1.3 above. Therefore,
an overall UF of IQx 10x3x3 = 1000 is used in the calculation of the chronic inhalation RfC
from the LOAEL, as follows:
RfC = LOAELADj H- UF = 8.7 mg/m3 - 1000 = 9 x 10'3 mg/m3
This value is in good agreement with the RfC of 3 x 10"2 mg/m3 calculated from the BMC.
5.1.5. Comparison Analysis Based on the Ward et al. (1985) Experimental Animal Study
A chronic inhalation RfC was also derived from the subchronic experimental animal
study of Ward et al. (1985) for comparison with the RfC of 3 x 10"2 mg/m3 based on the
Rothman et al. (1996a) human study. The Ward et al. (1985) study was selected because it used
a relatively long inhalation exposure duration and an adequate number of animals, and it
123
-------
provided dose-response data. The study identified both a LOAEL of 300 ppm and a NOAEL of
30 ppm. The investigators exposed male and female CD-I mice and Sprague-Dawley rats to 0,
1, 10, 30, or 300 ppm (0, 3.2, 32, 96, or 960 mg/m3) benzene 6 hours/day, 5 days/week for 91
days and measured various hematologic endpoints. The male mouse appears to be the most
sensitive sex/species in this study. The exposure-response relationships for the different
hematologic endpoints were modeled using a BMD modeling approach, and decreased HCT was
chosen as the critical effect. The exposure-response data from Ward et al. (1985, Table II, 91
days) for the male mouse are reproduced below.
Exposure Number of % HCT
(ppm) subjects (mean ± SD)
0 20 41.1 ±1.61
1 20 38.4 ±3.93
10 20 40.8 ±3.22
30 20 38.4 ±5.65
300 20 27.9 ±4.75
EPA's Benchmark Dose Modeling Software (version 1.20) was used for the modeling.
An assumption of constant variance was used, although the test for homogeneity of the variances
failed. The continuous linear, polynomial, and power models all resulted in the same BMC and
BMCL estimates; however, the linear model had better results for the fit statistics. The linear
model had a/>-value of 0.09, which is of borderline adequacy (EPA's draft Benchmark Dose
Technical Guidance Document [U.S. EPA, 2000b] recommends a/?-value of > 0.1), whereas the
other models had ^-values of 0.04. Thus the continuous linear model was selected. This model
has the form
Y[dose] = po + pj * dose
where, in this case, Y[dose] is the mean HCT level and dose is the experiment exposure
concentration. The parameters were estimated using the method of maximum likelihood. The
resulting parameter estimates are
Variable Estimate SE
P0 40.0962 2.44042
p! -0.0407075 329.232
See Figure 3 for a graphical display of the data and model.
In the absence of a clear definition for an adverse effect for this continuous endpoint, a
default BMR of one standard deviation from the control mean was selected, as suggested in the
EPA's draft BMD technical guidance document (U.S. EPA, 2000b). The software uses the
estimated standard deviation. A 95% lower confidence limit (BMCL) on the resulting BMC was
calculated using the likelihood profile method. A BMC of 100.7 ppm and a BMCL of 85.0 ppm
124
-------
Linear Model with 0.95 Confidence Level
42
gS 40
| 38
"5
CD 36
CD
CO
c.
CD
cr
c
cc
32
30
28
26
24
Linear
BMCL BMC
50 100 150 200
exposure concentration (ppm)
250
300
Figure 3. Linear model of HCT data from Ward et al. (1985)
were obtained.
It should be noted that the dose spacing in this study was less than ideal. Responses in the
three lower-exposure groups for all the hematologic endpoints tended to clump near control
group levels, and significant deviations in response were generally seen only in the 300 ppm
group, with a large exposure range in between, including where the BMC is located, for which
there are no response data (see Figure 3). Therefore, there is some uncertainty about the actual
shape of the dose-response curve in the region of the BMR and, thus, some corresponding
uncertainty about the values of the BMC and BMCL estimates.
ALCs were not reported in Ward et al. (1985), so this endpoint could not be compared
with the human ALC results. Total WBC counts were reported, and they exhibited the largest
percent change in response between the control and the 300 ppm group. However, the data for
this endpoint also had substantial variance, and because the BMR used for this analysis is a
function of the standard deviation, WBC counts did not yield the lowest BMC estimate. The
actual lowest BMC estimates were obtained for increased mean cell Hgb (78 ppm; BMCL = 67
ppm) and increased mean cell volume (79 ppm; BMCL = 68 ppm); however, these endpoints are
probably not adverse per se. On the other hand, they are likely to be compensatory effects—and
thus markers of toxicity—and one could probably justify using them as the critical effects. In any
event, the BMC estimates are not much different from the BMC of 100 ppm obtained for
decreased HCT. The results are also similar for total blood Hgb (BMC =104 ppm; BMCL = 88
125
-------
ppm). RBC results were in between those for MCV and mean corpuscular Hgb (MCH) and
those for HCT and total Hgb; however, the model fits were not adequate for the RBC data; thus,
the RBC results have more uncertainty.
To derive the RfC, the BMCL is used as the point of departure, as suggested in the draft
BMD technical guidance document (U.S. EPA, 2000b). An adjusted BMCL is calculated by
converting ppm to mg/m3 and adjusting from animal experiment exposure to equivalent
continuous exposure. (For conversion of the inhalation exposures across species, ppm
equivalence was assumed; this is identical to using EPA's inhalation dosimetry methodology
with the regional gas dose ratio for the respiratory tract region (RGDR,.) = 1 [U.S. EPA, 1994].)
The BMCL is first converted to mg/m3 using the molecular weight of 78.11 for benzene and
assuming 25°C and 760 mm Hg:
BMCL (mg/m3) = 85.0 ppm x 78.11/24.45 = 272 mg/m3
The converted value is then adjusted to an equivalent continuous exposure as follows:
j = 272 mg/m3 x 6 hours/24 hours x 5 days/7 days = 48.5 mg/m3
To derive the RfC, several UFs are applied to the BMCLADJ. As discussed in Section
5.1.3, a UF of 3 is used as an effect-level extrapolation factor, analogous to a LOAEL-to-
NOAEL UF, because the BMC is considered an adverse-effect level. In addition, the standard
UFs of 3 for interspecies extrapolation for inhalation studies and of 10 for intraspecies variability
are used. Also as discussed in Section 5.1.3, a UF of 3 for database deficiencies is used due to
the absence of a two-generation reproductive/developmental toxicity study for benzene. Finally,
a partial UF of 3 was used to extrapolate from subchronic to chronic exposure. This partial value
was selected on the basis of the observation that hematologic fluctuations such as reductions in
RBCs and WBCs in the high-dose mice were noted at interim sacrifice (14 days) as well as at
termination (91 days), suggesting that the responses occurred early in the exposure cycle and
then remained comparatively unchanged. Therefore, an overall UF of 3x3x10x3x3 = 1000
is used in the calculation of the chronic inhalation RfC from the BMCL^j, as follows:
RfC = BMCLADJ-UF = 48.5 mg/m3 - 1000 = 5 x 10-2mg/m3
This value is in good agreement with the RfC of 3 x 10"2 mg/m3 calculated from the BMC from
the Rothman et al. (1996a) human study.
Similarly, for comparison purposes, a chronic inhalation RfC can be derived from the
NOAEL of 30 ppm observed for hematologic effects in the Ward et al. (1985) experimental
animal study. First, the NOAEL is converted to mg/m3 and adjusted to equivalent continuous
exposure, as above:
30 ppm x 78.11/24.45 = 95.8 mg/m3
95.8 mg/m3 x (6 hours/24 hours) x 5 days/7 days =17.1 mg/m3
126
-------
(ppm equivalence across species was assumed; this is identical to using EPA's inhalation
dosimetry methodology with RGDR^l [U.S. EPA, 1994]). UFs are identical to those employed
above, except that no NOAEL-to-LOAEL UF is used, thus the overall UF is 300, resulting in an
RfCof
NOAELADJ -UF = 17.1 -300 = 6 x 10-2mg/m3
This value is also in good agreement with the RfC of 3 x 10"2 mg/m3 derived from the Rothman
etal. (1996a) study.
It should be noted, however, that other experimental animal studies have reported
significant hematologic effects at benzene exposures of 10-25 ppm, which are lower than the
NOAEL of 30 ppm from the Ward et al. (1985) study. These studies have insufficient data for
dose-response modeling, and they used shorter exposure durations and/or fewer experimental
animals than did the Ward et al. study; nonetheless, they observed statistically significant
hematologic effects at 10-25 ppm. Baarson et al. (1984), for example, exposed male C57BL/6J
mice (five/group) to 10 ppm benzene 6 hours/day, 5 days/week for 178 days and observed
statistically significant reductions in blood lymphocytes at each of the three monitoring time
points (32, 66, and 178 days) as compared with controls. The magnitude of the reduction in
lymphocytes ranged from about 53% at 32 days to about 68% at 178 days. Cronkite et al. (1985)
exposed male and female C57BL/6 BNL mice to various concentrations of benzene 6 hours/day,
5 days/week for 2 weeks and observed no decrease in blood lymphocytes at 10 ppm; however,
the investigators did observe a statistically significant reduction of about 21% at 25 ppm as
compared with controls (5-10 mice/group). Thus, lower RfCs than those calculated above for
the Ward et al. (1985) study are possible, based on other experimental animal results. In the
most extreme case, using a LOAEL of 10 ppm and an overall UF of 3000 yields a LOAELADJ of
5.7 mg/m3 and an RfC of 2 x 10'3 mg/m3.
5.2. ORAL REFERENCE DOSE (RfD)
5.2.1. Choice of Principal Study and Critical Effect
As with the inhalation RfC (Section 5.1), the human occupational inhalation study of
Rothman et al. (1996a) was selected as the principal study, and reduced ALC was selected as the
critical effect, for the derivation of the chronic oral RfD. This study was selected because it was
a well-conducted human exposure study that demonstrated a dose-response relationship for
hematologic effects, responses that are considered to be among the more sensitive indices of
benzene toxicity (see Section 5.1.1 for more details). Furthermore, no relevant human data are
available to evaluate hematologic effects following oral exposure. For comparison purposes, a
chronic oral RfD is also derived from the chronic experimental animal gavage study of the NTP
(1986) on the basis of the same critical effect of reductions in lymphocyte count (see Section
5.2.5).
5.2.2. Conversion of Inhalation Exposure to Equivalent Oral Dose Rate
As discussed in Section 5.1.2, BMD modeling of the ALC data of Rothman et al. (1996a)
yielded a BMC of 13.7 ppm (8-hour TWA) and a BMCL of 7.2 ppm (8-hour TWA) for the
127
-------
default BMR of one standard deviation change from the control mean response. This default
definition of a BMR for continuous endpoints corresponds to an excess risk of approximately
10% for the proportion of individuals below the 2nd percentile (or above the 98th percentile) of
the control distribution for normally distributed effects (see U.S. EPA, 2000b). As suggested in
the draft Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b), the BMCL is
chosen as the point of departure for the RfD derivation. Calculation of an equivalent oral dose
rate from the inhalation BMCL of 7.2 ppm (8-hour TWA) is shown below.
The BMCL is first converted to mg/m3 using the molecular weight of 78. 1 1 for benzene
and assuming 25°C and 760 mm Hg:
BMCL (mg/m3) = 7.2 ppm x 78.11/24.45 = 23.0 mg/m3
The converted value is then adjusted from the 8-hour TWA to an exposure concentration
adjusted for continuous exposure using the default occupational minute volume (U.S. EPA,
1994).
j = BMCL (mg/m3) x (VEho/VEh) x 5 days/7 days
j = 23.0 mg/m3 x (10 m3/20 m3) x 5 days/7 days = 8.2 mg/m3
where:
BMCL^j = the BMCL dosimetrically adjusted to account for continuous exposure
BMCL = occupational exposure level (8-hour TWA)
VEho = human occupational default minute volume (10 m3/8 hours)
VEh = human ambient default minute volume (20 m3/24 hours)
In the support document for the benzene cancer assessment on IRIS (U.S. EPA, 1999a),
EPA provided a simple method for extrapolating benzene-induced cancer risk from the
inhalation to oral route. The same method is applied here for noncancer (hematopoietic) effects.
The method is based on the relative efficiency of benzene absorption across routes of exposure,
especially pulmonary and gastrointestinal barriers. An inhalation absorption rate of 50% and an
oral absorption rate of 100% were used to calculate the absorbed benzene dose. These values are
based on human inhalation absorption studies (Nomiyama and Nomiyama, 1974; Pekari et al.,
1992; Srbova et al., 1950) and the study of Sabourin et al. (1987), which compared inhalation
and oral absorption in rats and mice.
Sabourin et al. (1987) found that the retention of 14C-benzene by rats and mice during a
6-hour exposure decreased as exposure concentration increased. Retention decreased from 33 ±
6% to 15 ± 9% for rats and from 50 ± 1% to 10 ± 2% for mice as exposure concentration
increased from 32 to 3,200 mg/m3 (10 to 1,000 ppm). In the same study, gastrointestinal
absorption of benzene administered by gavage was > 97% for doses between 0.5 and 150 mg/kg
body weight. At oral doses below 15 mg/kg, > 90% of the 14C excreted was in the urine as
128
-------
nonethyl acetate-extractable material. At higher doses, an increasing percentage of the orally
administered benzene was exhaled unmetabolized. Thus, in the dose range represented by the
BMCL from the study by Rothman et al. (a, b), absorption of a comparable oral dose was
assumed to be 100%. See also U.S. EPA (1999a) for more details about the route-to-route
extrapolation of benzene inhalation results to oral exposures.
To calculate an equivalent oral dose rate, the BMCLADJ is multiplied by the default
inhalation rate, multiplied by 0.5 to correct for the higher oral absorption, and divided by the
standard default human body weight of 70 kg (U.S. EPA, 1988):
8.2 mg/m3 x 20 nrVday x 0.5 H- 70 kg = 1.2 mg/kg/day
5.2.3. RfD Derivation
To calculate an RfD using the BMCLADJ-equivalent oral dose rate value of 1.2
mg/kg/day, several UFs were applied. First, because the BMC is considered to be an adverse
effect level, an effect-level extrapolation factor analogous to the LOAEL-to-NOAEL UF is used.
EPA is planning to develop guidance for applying an effect level extrapolation factor to a BMD.
A factor of 3 will be used in this analysis, as in Section 5.1.3. For a more serious effect, a larger
factor, such as 10, might be selected. Second, a factor of 10 was used for intraspecies
differences in response (human variability) as a means of protecting potentially sensitive human
subpopulations. Third, a partial UF of 3 was used for subchronic-to-chronic extrapolation, as
discussed in Section 5.1.3. Finally, a UF of 3 was chosen to account for database deficiencies
due to the lack of a two-generation reproductive/developmental toxicity study for benzene.
Therefore, an overall UF of 3x10x3x3 = 300 is used to calculate the chronic oral RfD, as
follows:
RfD = equivalent oral dose rate ^ UF = 1.2 mg/kg/day ^ 300 = 4 x 103 mg/kg/day.
Use of a modifying factor of 3 to account for uncertainty in the route-to-route extrapolation was
considered; however, it was deemed unnecessary. The RfD is based on human data for a
sensitive endpoint; thus, it was felt that the composite UF of 300 provides sufficient protection.
129
-------
5.2.4. Comparison Analysis Based on the LOAEL
A median 8-hour TWA concentration of 7.6 ppm was identified from the Rothman et al.
(1996a) study as a LOAEL for the critical effect of reduced lymphocyte counts. An equivalent
oral dose rate can be calculated from the LOAEL, as shown in Section 5.2.2 above and in
abbreviated form below. First, a LOAEL ^ is calculated by converting to mg/m3 and adjusting
from the occupational ventilation rate and intermittent work-week schedule to a continuous 24-
hour exposure, 7 days/week:
LOAEL (mg/m3) = 7.6 ppm x 78.11/24.45 = 24.3 mg/m3
j = 24.3 mg/m3 x (10 m3/20 m3) x 5 days/7 days = 8.7 mg/m3
Next, an inhalation absorption rate of 50% was used to calculate the absorbed benzene
dose, and oral absorption was assumed to be 100%, as discussed in Section 5.2.2. To calculate
an equivalent oral dose rate, the LOAEL ADJ is multiplied by the default inhalation rate,
multiplied by 0.5 to correct for the higher oral absorption, and divided by the standard default
human body weight of 70 kg (U.S. EPA, 1988):
8.7 mg/m3 x 20 m3/day x 0.5 H- 70 kg = 1.2 mg/kg/day
To calculate an RfD using the LOAEL ADJ-equivalent oral dose rate value of 1 .2
mg/kg/day, the following UFs were selected: a factor of 10 to account for using a LOAEL due
to the lack of an appropriate NOAEL, a factor of 10 for intraspecies differences in response
(human variability), a factor of 3 for subchronic-to-chronic extrapolation, and a factor of 3 for
database deficiencies, as above. Therefore, an overall UFoflQx 10x3 x3 = IQOO has been
used to calculate the chronic oral RfD, as follows:
RfD = equivalent oral dose rate - UF = 1.2 mg/kg/day - 1000 = 1 x 10'3 mg/kg/day.
This value is in good agreement with the RfD of 4 x 10"3 mg/kg/day calculated from the BMC.
5.2.5. Comparison Analysis Based on the NTP (1986) Experimental Animal Study
The proposed RfD is based on the Rothman et al. (1996a) study because it was decided
that it was preferable to use the human data rather than the experimental animal data. The NTP
(1986) study rather than the Hsieh et al. (1988b) study was used for the comparison analysis
because it was much larger (50 vs. 5 animals/dose group), it was conducted for 2 years (vs. 28
days) with blood measurements made at multiple time points, and it examined both male and
female rats and mice (vs. just male mice). The Hsieh et al. (1988b) study provided a lower
LOAEL (8 mg/kg/day vs. 18 mg/kg/day adjusted to 7 days/week); however, the lower LOAEL is
not crucial, because the main analysis currently uses a BMD approach. Both studies yielded
similar BMDLs for decreased ALC (0.7 mg/kg/day for the male rat NTP data vs. 1 .4 mg/kg/day
for the Hsieh et al. data; both based on a linear model with transformed doses and a BMR of one
standard deviation change from the control mean).
130
-------
A chronic oral RfD was also derived from the NTP (1986) experimental animal study for
comparison with the RfD of 1 x 10"3 mg/kg/day based on the Rothman et al. (1996a) human
study. In the NTP (1986) study, F344 rats and B6C3F1 mice of both sexes were administered
benzene by gavage 5 days/week for 103 weeks (see also Section 4.2.1.2). Male rats (50/group)
were administered doses of 0, 50, 100, or 200 mg/kg and females (50/group) were administered
doses of 0, 25, 50, or 100 mg/kg. B6C3F1 mice (50/sex/group) were administered doses of 0,
25, 50, or 100 mg/kg. Blood was drawn from 10 randomly preselected animals per
species/sex/dose group at 12, 15, 18, and 21 months and from all animals at the terminal kill at
24 months. Additional groups of 10 animals of each sex and species were administered benzene
for 51 weeks at the same doses of the 2-year study, and blood was drawn at 0, 3, 6, 9, and 12
months. This study identified a LOAEL of 25 mg/kg for leukopenia and lymphocytopenia in
female F344 rats and male and female B6C3F1 mice and of 50 mg/kg in male F344 rats. These
were the lowest doses tested and, thus, no NOAEL was identified.
Reduction in lymphocyte count was selected as the critical effect, and attempts were
made to model the dose-response relationships using a BMD modeling approach. However, the
dose-response analysis of the NTP data was problematic for a number of reasons. First, only the
grouped data, not the individual animal data, were provided. Second, as discussed in Appendix
N of the NTP (1986) report, not all of the groups were comparable for a variety of reasons,
including experiment design, differences in bleeding methods and time of bleeding, and
differences in measurement instrumentation. Third, the grouped data exhibited substantial
variability, and baseline measurements (vehicle control groups) were not constant over time.
Furthermore, for the rats, the strong dose effect in lymphocyte count was accompanied by
significant temporal variability, which appears to be largely due to temporal variability in the
vehicle control group means. For these reasons, it was not possible to do a pooled analysis
across time points. Instead, modeling was performed for each dataset in two data groupings
within which the datasets are comparable (months 6 and 9; and months 12, 15, 18, and 21), and
ranges of results are presented. Each of these datasets had at most 10 animals/dose, so the dose-
response results are not very robust. The males of both species exhibited more dramatic and
consistent reductions in lymphocyte count than did the females, but it was not clear a priori
which species was more sensitive; therefore, dose-response analyses were performed for both the
male mouse (NTP Table N8) and the male rat (NTP Table N4).
The continuous linear, polynomial, and power models in EPA's Benchmark Dose
Modeling Software (version 1.20) were used for the modeling. The software estimates the
parameters using the method of maximum likelihood. Most of the data were supralinear (i.e., the
magnitude of the reductions in lymphocyte count decreased with increasing unit dose) and not
amenable to modeling with restricted models (in this case, with parameters constrained to be
nonpositive), as suggested in EPA's draft Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b). Therefore, two different strategies were attempted. The first was to use
unrestricted polynomial models and exclude the high-dose group when necessary to fit the
models, and the second was to transform the dose data before modeling. Dropping the high-dose
groups alone did not resolve the problem because there was still too much supralinear curvature
in the two lower dose groups to fit restricted models. Results are summarized below in Tables
11 and 12, for untransformed and transformed data, respectively.
131
-------
For each dataset, the selected model was chosen on the basis of the lowest Akaike's
Information Criterion value with consideration of the graphical display, as suggested in the draft
BMD technical guidance document. For selecting between models within a family of models
(e.g., between a linear and a two-degree polynomial model), consideration was given to the log-
likelihood values to evaluate the statistical significance of adding an extra parameter. In some
cases there was no model that adequately fit the data. The draft technical guidance document
suggests a/>-value of 0.1 for model fit. In addition, there was substantial variability in these
data, but the variability appeared to be random and not amenable to modeling. Therefore,
constant variance was assumed for all the models, although in some cases the variances failed
the test for homogeneity.
In the absence of a clear definition for an adverse effect for this endpoint, a default BMR
of one standard deviation change from the control mean response was selected as suggested in
the draft BMD technical guidance document. This definition of the BMR is highly sensitive to
the substantial variability in data such as these; thus, the BMR itself is not very robust. The
usefulness of this default definition would be strengthened by the use of a larger dataset of
historical control data, but such data were not located. The software uses the estimated
"constant" standard deviation as the standard deviation for all the group means. The 95% lower
confidence limit (BMDL) on the resulting BMD was then calculated using the likelihood profile
method.
Because many of the datasets could not be modeled using the available models without
dropping the parameter constraints and sometimes even excluding the highest dose group (Table
11), a second modeling approach was tried in which the dose levels were transformed before the
dose-response modeling was conducted, as was done for the human data modeled in Section
5.1.2. The dose levels were transformed according to the formula: transformed dose =
ln(dose +1). Results of the dose-response modeling using the transformed doses are presented
in Table 12.
The results in Table 12 suggest that the male rat is more sensitive than the male mouse to
lymphocyte count reductions from exposure to benzene in this NTP (1986) gavage bioassay,
because the ranges of BMDs/BMDLs are substantially lower for the male rat, especially for year
2. The ranges for the male rat are also fairly tight, and the models selected provide good fits to
all the male rat datasets. However, all but one of the calculated BMDs for the male rat are over
an order of magnitude below the lowest exposure dose of 50 mg/kg. Ideally, BMDs should be
closer to the low end of the range of observation (i.e., the range of the actual exposure doses) to
reduce the impacts of model selection and the uncertainties inherent in extrapolating to lower
doses.
Nevertheless, data from two drinking water studies (Hsieh et al., 1988b; White et al.,
1984) provide support for selecting a BMD in this range. These two studies were of shorter
duration and used fewer experimental animals than did the NTP (1986) study; however, they do
provide dose-response data for BMD modeling and they also have the advantage of being
drinking water studies; thus the benzene exposure scenario is more relevant to human oral
benzene exposures. In one study, Hsieh et al. (1988b) exposed male CD-I mice (five/group) to
0, 8, 40, or 180 mg/kg/day benzene in drinking water for 28 days. Hematologic effects were
132
-------
Table 11. BMD modeling results of the NTP (1986) male mouse and male rat
lymphocyte counts, with untransformed data
Dataset
Model
Variance
homogeneity
Fit
BMDa
(mg/kg)
BMDLa
(mg/kg)
Male mouse
6-month
9-month
unrestricted 2-degree
polynomial
unrestricted 2-degree
polynomial
OK
no
yes
p=Q.l3
yes
p=Q.12
year 1 range
12-month
15 -month
18-month
21 -month
unrestricted 2-degree
polynomial
linear
linear
linear
OK
no
no
no
nob
^=0.005
yes, ^=0.98
yes,/>=0.16
yes, ^=0.25
year 2 range
20.52
23.06
20.52-23.06
16.68b
51.94
35.50
42.91
35.50-51.94
12.96
13.70
12.96-13.70
11.41b
36.80
26.92
30.44
26.92-36.80
Male rat
6-month
9-month
unrestricted 2-degree
polynomial
unrestricted 2-degree
polynomial
OK
no
yes
^=0.56
yes
^=0.37
year 1 range
12-month; w/o
high-dose
group
15 -mo nth; w/o
high-dose
group
18-month; w/o
high-dose
group
21 -month; w/o
high-dose
group
unrestricted 2-degree
polynomial
unrestricted 2-degree
polynomial
unrestricted 2-degree
polynomial
unrestricted 2-degree
polynomial
OK
OK
no
OK
yes0
yes0
yes0
yes0
year 2 range
15.70
16.88
15.70-16.88
10.51
7.56
13.84
7.21
7.21-13.84
11.79
12.72
11.79-1272
7.3
5.56
9.07
5.03
5.03-9.07
"Unadjusted animal doses in mg/kg.
bThese data could not be fit, and the BMD and BMDL are excluded from the reported ranges.
°Models were judged to fit on the basis of the graphical displays and x2-residuals; however, there were insufficient
degrees of freedom for calculation of a^-value.
133
-------
Table 12. BMD modeling results of the NTP (1986) male mouse and male rat
lymphocyte counts, with transformed dose data
Dataset
Model
Variance
homogeneity
Fit
BMDa
(mg/kg)
BMDLa
(mg/kg)
Male Mouse
6-month
9-month
2-degree polynomial
linear
OK
no
borderline ^=0.047
yes,/>=0.35
year 1 range
12-month
15 -month
18-month
21 -month
linear
power
power
power
OK
no
no
no
yes, ^=0.30
yes,/?=0.31
borderline ^=0.09
yes, p=Q.l5
year 2 range
19.68
9.07
9.07-19.68
3.74
47.46
28.93
23.34
3.74^7.46
6.57
4.05
4.05-6.57
2.32
18.55
13.99
5.80
2.32-18.55
Male rat
6-month
9-month
power
linear
OK
no
yes, ^=0.30
yes,/?=0.11
year 1 range
12-month
15 -mo nth
18-month
21 -month
linear
linear
linear
linear
no
OK
no
OK
yes, ^=0.22
yes, ^=0.93
yes, ^=0.22
yes, ^=0.54
year 2 range
9.92
3.71
3.71-9.92
1.34
1.34
2.73
1.69
1.34-2.73
4.52
2.30
2.30-4.52
0.95
0.95
1.74
1.10
0.95-1.74
aUnadjusted animal dose in mg/kg, after transforming the results back according to the formula: dose =
exp(transformed dose) - 1.
observed at all exposure levels. BMD modeling of the ALC yielded a BMD of 2.2 mg/kg/day
and a BMDL of 1.4 mg/kg/day, based on a linear model with transformed doses and a BMR of
one standard deviation change from the control mean, as above. In the second study, White et al.
(1984) exposed female B6C3F1 mice to 0, 12, 195, or 350 mg/kg/day benzene in drinking water
for 30 days. BMD modeling of the ALC (five to six mice/group) resulted in a BMD of 11.6
mg/kg/day and a BMDL of 5.3 mg/kg/day (also based on a linear model with transformed doses
and a BMR of one standard deviation change from the control mean, as above).
The results in Table 12 from BMD modeling of the male rat ALC data from the NTP
(1986) study show the lowest BMDL of about 1 mg/kg at three time points in the second year;
therefore, this was selected as the point of departure for an RfD calculation. Adjusting for
134
-------
exposure 7 days/week yields a BMDL^j of 0.7 mg/kg/day. As discussed in previous
subsections, a UF of 3 is used as an effect-level extrapolation factor, analogous to a LOAEL-to-
NOAEL UF, because the BMD is considered to be an adverse-effect level. In addition, the
standard UFs of 10 for interspecies extrapolation for oral studies and for intraspecies variability
are used. Finally, a UF of 3 for database deficiencies is used, as in the previous derivations.
These values yield an overall UF of 1000 and an RfD as follows:
RfD = BMDLADJ - UF = 0.7 mg/kg/day - 1000 = 7 x 10'4 mg/kg/day
This RfD value is in reasonably good agreement (within an order of magnitude) with the RfD
value of 4 x 10"3 mg/kg/day derived from the Rothman et al. (1996a) human inhalation study.
For comparison purposes, a chronic oral RfD can be derived from the LOAEL of 25
mg/kg identified for hematologic effects in the NTP (1986) study (there was no NOAEL).
Adjusting from 5-day to 7-day exposure yields a LOAEL^ of 18 mg/kg/day, which can be used
to calculate an RfD for benzene as follows:
RfD = LOAELADJ - UF = 18 mg/kg/day - 3000 = 6 x 10'3 mg/kg/day,
where the combined UF of 3000 is made up of component factors of 10 for LOAEL-to-NOAEL
extrapolation, 10 for interspecies extrapolation, 10 for intraspecies variability, and 3 for database
deficiencies. This value is in good agreement with the RfD of 4 x 10"3 mg/kg/day calculated
from the BMD analysis of the Rothman et al. (1996a) human data.
5.3. DOSE-RESPONSE SUMMARY
For the derivation of a chronic inhalation RfC, the human occupational inhalation study
of Rothman et al. (1996a) was selected as the principal study, and the inhalation study of Ward
et al. (1985) was selected as a supporting experimental animal study. These two studies
provided the best dose-response data for quantitatively evaluating the potential human health
risks due to inhalation of benzene. The cross-sectional study of Rothman et al. (1996a) was
designated as the principal study because it is a human exposure study with well-quantified
benzene exposure estimates and data on sensitive hematological endpoints. The study compared
the hematologic evaluations of 44 workers occupationally exposed to benzene with those of 44
age- and gender-matched unexposed controls, all based in Shanghai, China. ALC, WBCs,
RBCs, and platelets were all significantly decreased, whereas MCV was significantly increased
in the group of exposed workers (median 8-hour TWA of 31 ppm [99 mg/m3]), compared with
the age- and sex-matched control group. In the low-dose subjects (< 31 ppm, median 8-hour
TWA of 13.6 ppm [43.4 mg/m3]), the ALC, RBCs, and platelet count were reduced as compared
with controls. Similarly, in a selected subgroup exposed to a median 8-hour TWA of 7.6 ppm
(23 mg/m3) benzene, a statistically significant difference in ALC versus controls was observed
(p<0.03). Furthermore, a dose-response relationship was established between the ALC and
benzene exposure, as monitored by organic vapor passive dosimetry and the level of benzene
metabolites in the urine. The median 8-hour TWA of 7.6 ppm was designated the LOAEL for
these effects and was used to calculate a chronic inhalation RfC for benzene hematotoxicity in
humans for comparison with the RfC derived in the primary analysis using BMD modeling of
the ALC exposure-response data.
135
-------
BMD modeling of the ALC exposure-response data was conducted using the continuous
models from EPA's Benchmark Dose Modeling Software (version 1.20). In order to fit the data
with one of the available models, the data were first transformed according to the equation
d' = ln(d + 1). The exposure-response data were then fit using the continuous linear model,
which provided a good fit. The two-degree polynomial model and the power model also
provided adequate fits to the data; however, the linear model was selected because it was more
parsimonious and the additional parameters were not statistically significant. In the absence of a
clear definition for an adverse effect for this continuous endpoint, a default BMR of one standard
deviation change from the control mean response was selected, as suggested in EPA's draft
Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b). This default definition of a
BMR for continuous endpoints corresponds to an excess risk of approximately 10% for the
proportion of individuals below the 2nd percentile (or above the 98th percentile) of the control
distribution for normally distributed effects. Transforming the resulting BMC and BMCL back
to the original exposure scale yielded a BMC of 13.7 ppm (8-hour TWA) and a BMCL of 7.2
ppm (8-hour TWA).
As suggested in the draft BMD technical guidance document, the BMCL was chosen as
the point of departure for the RfC derivation. After converting to mg/m3 and adjusting for
continuous exposure, a BMCLADJ of 8.2 mg/m3 was obtained. Dividing this value by an overall
UF of 300 yields a chronic inhalation RfC of 3 x 10~2 mg/m3, based on BMD modeling of the
ALC data from the Rothman et al. (1996a) human study. Because the BMC is considered to be
an adverse-effect level, an effect-level extrapolation factor analogous to the LOAEL-to-NOAEL
UF was used. EPA is planning to develop guidance for applying an effect-level extrapolation
factor to a BMD. In the interim, a factor of 3 was used in this analysis. For a more serious
effect, a larger factor, such as 10, might be selected. Additional factors of 10 for intraspecies
variability, 3 for subchronic-to-chronic extrapolation (exposure range 0.7 to 16 years), and 3 for
database deficiencies, due to the absence of a two-generation reproductive/developmental
toxicity study for benzene, comprise the remainder of the 300 composite UF. These UFs, as well
as the UFs for the comparison analyses and the RfD calculations, are summarized in Table 13.
For comparison, an RfC was also calculated from the LOAEL of 7.6 ppm (8-hour TWA)
from Rothman et al. (1996a). After converting to mg/m3 and adjusting for continuous exposure,
a LOAELADJ of 8.7 mg/m3 was obtained. Dividing this value by an overall UF of 1000 yields an
RfC of 9 x 10"3 mg/m3. The UF of 1000 was based on a factor of 10 to account for the use of a
LOAEL because of the lack of an appropriate NOAEL, a factor of 10 for intraspecies differences
in response (human variability), a partial UF of 3 for subchronic-to-chronic extrapolation, and a
factor of 3 for database deficiencies in the absence of a two-generation reproductive/
developmental toxicity study for benzene. This result of 9 x 10"3 mg/m3 based on the LOAEL is
in good agreement with the result of 3 x 10"2 mg/m3 based on the BMCL. The BMD modeling
approach is chosen as the primary analysis in this document because BMD modeling is a
generally superior methodology that addresses some of the limitations of the LOAEL/NOAEL
approach. For example, BMD modeling makes use of all the dose-response data, and the BMD
is not restricted to being one of the doses used in the study. The BMD also provides a more
136
-------
Table 13. Summary of uncertainty factors used for deriving the RfC and RfDa
Basis
RfC based on
Rothman et
al. (1996a)
(human)
Comparison
RfC based on
Ward et al.
(1985)
(rodent)
RfD based on
Rothman et
al. (1996a)
(human)
Comparison
RfD based on
NTP (1986)
(rodent)
BMCL or
BMDL
(NOAEL or
LOAEL)"
8.2 mg/m3
(8.7 LOAEL)
48.5 mg/m3
(17 NOAEL)
1.2
mg/kg/dayc
(1.2 LOAEL)
0.7
mg/kg/day
(18 LOAEL)
Effect level
extrapolation
factor
(LOAEL-to-
NOAEL UF)
3
(10)
3
(NA)
o
5
(10)
o
5
(10)
Interspecies
UF
NA
3
NA
10
Intraspecies
UF
10
10
10
10
Subchronic-
to-chronic UF
3
3
o
J
NA
Database
deficiencies
UF
3
3
o
J
o
J
Composite
UF
300
(1000)
1000
(300)
300
(1000)
1000
(3000)
RfD or RfC
0.03 mg/m3
(0.009)
0.05 mg/m3
(0.06)
4 x lO'3
mg/kg/day
(1 x icr3)
7 x ICr4
mg/kg/day
(6 x lO'3)
00
Trimary results are based on BMD analysis; comparison results using NOAEL/LOAEL approach are presented in parentheses where different.
bThe BMC was based on a BMR of one standard deviation change from the control mean. All values are adjusted for continuous exposure.
°Oral BMDL (and LOAEL) for Rothman et al. (1996a) inhalation study was derived by route-to-route extrapolation with the assumptions that inhalation
absorption was 50% and oral absorption was 100% in the dose range near the BMC.
NA = not applicable.
-------
standardized point of comparison across endpoints and studies because it corresponds to a
specific response level rather than being merely one of the doses used in a particular study.
Support for this chronic inhalation RfC was provided by the experimental animal study of
Ward et al. (1985). The subchronic inhalation study of Ward et al. (1985) was selected as a
supporting study because it used a relatively long inhalation exposure duration and an adequate
number of animals; it also provided dose-response data. The study identified both a LOAEL of
300 ppm and a NOAEL of 30 ppm. Ward et al. exposed male and female CD-I mice and
Sprague-Dawley rats to 0, 1, 10, 30 or 300 ppm (0, 3.2, 32, 96 or 960 mg/m3) benzene 6
hours/day, 5 days/week for 91 days and measured various hematologic endpoints. The male
mouse appears to be the most sensitive sex/species in this study. The exposure-response
relationships for the different hematologic endpoints in male mice were modeled using a BMD
modeling approach, and decreased HCT was chosen as the critical effect.
BMD modeling was conducted using the continuous models from EPA's Benchmark
Dose Modeling Software (version 1.20). The linear, polynomial, and power models all resulted
in the same BMC and BMCL estimates; however, the linear model had better results for the fit
statistics. Thus, the continuous linear model was selected. In the absence of a clear definition
for an adverse effect for this continuous endpoint, a default BMR of one standard deviation for
an adverse effect for this continuous endpoint, a default BMR of one standard deviation change
from the control mean response was selected, as suggested in EPA's draft Benchmark Dose
Technical Guidance Document (U.S. EPA, 2000b). A BMC of 100.7 ppm and a BMCL of 85.0
ppm were obtained. It should be noted that because of the large dose spacing between the 30
ppm and the 300 ppm groups, there is some uncertainty about the actual shape of the dose-
response curve in the region of the BMC and, thus, some corresponding uncertainty about the
values of the BMC and BMCL estimates.
ALCs were not reported in Ward et al. (1985), so this endpoint could not be compared
with the human ALC results. Total WBC counts were reported and exhibited the largest percent
change in response between the control and the 300 ppm group; however, the data for this
endpoint also had substantial variance. In addition, because the BMR used for this analysis is a
function of the standard deviation, WBC counts did not yield the lowest BMC estimate. The
actual lowest BMC estimates were obtained for increased mean cell Hgb (78 ppm; BMCL = 67
ppm) and increased mean cell volume (79 ppm; BMCL = 68 ppm); however, these endpoints are
probably not adverse per se. On the other hand, they are likely to be compensatory effects—and
thus markers of toxicity—and one could probably justify using them as the critical effects. In
any event, the BMC estimates are not much different from the BMC of 100 ppm obtained for
decreased HCT. The results are also similar for total blood Hgb (BMC =104 ppm; BMCL = 88
ppm). RBC results were in between those for MCV and MCH and those for HCT and total Hgb;
however, the model fits were not adequate for the RBC data. Thus, the RBC results have more
uncertainty.
As suggested in the draft BMD technical guidance document, the BMCL was chosen as
the point of departure for the RfC derivation. After converting to mg/m3 and adjusting for
continuous exposure, a BMCLADJ of 48.5 mg/m3 was obtained for the critical effect of decreased
HCT. To derive the RfC, several UFs are applied to the BMCL^j. As discussed above, a UF of
3 is used as an effect-level extrapolation factor, analogous to a LOAEL-to-NOAEL UF, because
138
-------
the BMC is considered an adverse-effect level. In addition, the standard UFs of 3 for
interspecies extrapolation for inhalation studies and of 10 for intraspecies variability are applied.
A UF of 3 for database deficiencies is used due to the absence of a two-generation
reproductive/developmental toxicity study for benzene. Finally, a partial UF of 3 was used to
extrapolate from subchronic to chronic exposure. This partial value was selected on the basis of
the observation that hematologic fluctuations such as reductions in RBCs and WBCs in the high-
dose mice were noted at interim sacrifice (14 days) as well as at termination (91 days),
suggesting that the responses occurred early in the exposure cycle and then remained
comparatively unchanged. Dividing the BMCL^j by the overall UF of 1000 yields a chronic
inhalation RfC of 5 x 10"2 mg/m3. This value is in good agreement with the RfC of 3 x 10"2
mg/m3, based on BMD modeling of the ALC data from the Rothman et al. (1996a) human study.
Similarly, for comparison purposes, a chronic inhalation RfC can be derived from the
NOAEL of 30 ppm observed for hematologic effects in the Ward et al. (1985) experimental
animal study. First, the NOAEL is converted to mg/m3 and adjusted to equivalent continuous
exposure, yielding 17.1 mg/m3. UFs are identical to those employed above, except that no
NOAEL-to-LOAEL UF is used; thus, the overall UF is 300. Dividing 17.1 by 300 results in an
RfC of 6 x 10"2 mg/m3. This value is also in good agreement with the RfC derived from the
Rothman et al. (1996a) study.
For derivation of an RfD, the human occupational inhalation study of Rothman et al.
(1996a) was again selected as the principal study, and the 103-week gavage study conducted by
the NTP (1986) in F344 rats and B6C3F1 mice was selected as a supporting experimental animal
study. These studies provided the best dose-response data for quantitatively evaluating the
potential human health risks due to oral benzene exposure. The cross-sectional study of
Rothman et al. (1996a) is designated as the principal study because it is a human exposure study,
with well-quantified exposure estimates and data on sensitive hematological endpoints.
As with the RfC, the BMCL of 7.2 ppm (8-hour TWA) for the default BMR of one
standard deviation change from the control mean response for the critical effect of reduced ALC
in Rothman et al. (1996a) was used as the point of departure for the derivation of the RfD. After
converting the units, correcting for continuous exposure, and adjusting for the route-to-route
extrapolation from inhalation to oral exposure, a BMCLADJ-equivalent oral dose rate of 1.2
mg/kg/day was obtained. The route-to-route extrapolation assumes 50% absorption of inhaled
doses and 100% absorption of oral doses. These assumptions were based on experimental data
(Nomiyama and Nomiyama, 1974; Pekari et al., 1992; Srbova et al., 1950; Sabourin et al., 1987)
and are the same as those used by EPA for route-to-route extrapolation of the inhalation cancer
risk estimates to obtain cancer risk estimates for oral exposures (U.S. EPA, 1999a). Dividing the
BMCL^j-oral of 1.2 mg/kg/day by an overall UF of 300 yields a chronic oral RfD of 4 x 103
mg/kg/day, based on BMD modeling of the ALC data from the Rothman et al. (1996a) human
study. Because the BMC is considered to be an adverse-effect level, an effect-level
extrapolation factor analogous to the LOAEL-to-NOAEL UF was used. EPA is planning to
develop guidance for applying an effect-level extrapolation factor to a BMD. In the interim, a
factor of 3 was used in this analysis. Additional factors of 10 for intraspecies variability, 3 for
subchronic-to-chronic extrapolation (exposure range 0.7 to 16 years), and 3 for database
deficiencies due to the absence of a two-generation reproductive/developmental toxicity study
for benzene comprise the remainder of the 300 composite UF.
139
-------
For comparison, an RfD was also calculated from the LOAEL of 7.6 ppm (8-hour TWA)
from Rothman et al. (1996a). After unit conversion, correction for continuous exposure, route-
to-route extrapolation, and division by a combined UF of 1000, an RfD of 1 x 10"3 mg/kg/day
was derived from the LOAEL. The route-to-route extrapolation was based on an assumption of
50% absorption of inhaled doses and 100% absorption of oral doses, as above. The combined
UF of 1000 was based on a factor of 10 to account for using a LOAEL because of the lack of an
appropriate NOAEL, a factor of 10 for intraspecies variability, a factor of 3 for subchronic-to-
chronic extrapolation, and a factor of 3 for database deficiencies. This RfD value of 1 x 10"3
mg/kg/day is in good agreement with the value of 4 x 10"3 mg/kg/day calculated from the BMC.
Support for this chronic oral RfD was provided by the experimental animal study of the
NTP (1986). In this study, F344 rats and B6C3F1 mice of both sexes were administered benzene
by gavage 5 days/week for 103 weeks (see also Section 4.2.1.2). For rats, males (50/group) were
administered doses of 0, 50, 100, or 200 mg/kg, and females (50/group) were administered doses
of 0, 25, 50, or 100 mg/kg. B6C3F1 mice (50/sex/group) were administered doses of 0, 25, 50,
or 100 mg/kg. Blood was drawn from 10 randomly preselected animals per species/sex/dose
group at 12, 15, 18, and 21 months and from all animals at the terminal kill at 24 months.
Additional groups of 10 animals of each sex and species were administered benzene for 51
weeks at the same doses of the 2-year study, and blood was drawn at 0, 3, 6, 9, and 12 months.
This study identified a LOAEL of 25 mg/kg for leukopenia and lymphocytopenia in female F344
rats and male and female B6C3F1 mice and 50 mg/kg in male F344 rats. These were the lowest
doses tested and, thus, no NOAEL was identified.
Reduction in lymphocyte count was selected as the critical effect, and attempts were
made to model the dose-response relationships using a BMD modeling approach. The males of
both species exhibited more dramatic and consistent reductions in lymphocyte count than did the
females, but it was not clear a priori which species was more sensitive; therefore, dose-response
analyses were performed on datasets for various time points for both the male mouse (NTP Table
N8) and the male rat (NTP Table N4). Various continuous models were used to fit the different
datasets, as appropriate, and one standard deviation change from the control mean was used as
the BMR, as discussed above. The modeling results suggested that the male rat is more sensitive
than the male mouse to lymphocyte count reductions from exposure to benzene in this NTP
gavage bioassay. However, all but one of the calculated BMDs for the male rat were over an
order of magnitude below the lowest exposure dose of 50 mg/kg. Ideally, BMDs should be
closer to the low end of the range of observation (i.e., the range of the actual exposure doses) to
reduce the impacts of model selection and the uncertainties inherent in extrapolating to lower
doses. Nonetheless, BMD modeling of two subchronic drinking water studies (Hsieh et al.,
1988b; White et al., 1984) supported the selection of a BMD in the range of BMDs estimated
from the NTP (1986) rat data, and a BMDL of 1 mg/kg was selected as the point of departure.
Adjusting the BMDL of 1 mg/kg for exposure 7 days/week yielded a BMDLADJ of 0.7
mg/kg/day. A UF of 3 was used as an effect-level extrapolation factor, analogous to a LOAEL-
to-NOAEL UF, because the BMD is considered to be an adverse-effect level. Also, UFs of 10
for interspecies extrapolation for oral studies, 10 for intraspecies variability, and 3 for database
deficiencies due to the absence of a two-generation reproductive/developmental study were
applied, resulting in a composite UF of 1000. Dividing the BMDLADJ by 1000 yields an RfD of
7 x 10"4 mg/kg/day. This RfD value is in reasonably good agreement (within an order of
140
-------
magnitude) with the RfD value of 4 x 10"3 mg/kg/day derived from the Rothman et al. (1996a)
human inhalation study.
For comparison purposes, a chronic oral RfD was also calculated using the LOAEL from
the NTP (1986) study. The LOAEL of 25 mg/kg was adjusted to a continuous exposure level of
17.9 mg/kg/day and then divided by a UF of 3000 to derive an RfD of 6 x 10'3 mg/kg/day. The
combined UF of 3000 is based on a factor of 10 for the absence of a NOAEL, a factor of 10 for
interspecies extrapolation, a factor of 10 for intraspecies differences in response (human
variability), and a factor of 3 for database deficiencies, as above. This value of 6 x 10"3
mg/kg/day is in good agreement with the value 4 x 10~3 mg/kg/day derived from the Rothman et
al. (1996a) human study.
In summary, the chronic inhalation RfC values calculated on the basis of the human and
the experimental animal data and using BMD modeling and NOAEL/LOAEL approaches are in
good agreement, yielding values that range from 9 x 10"3 to 6 x 10"2 mg/m3 (Table 14). This
consistency in results provides increased confidence in the selected chronic inhalation RfC of 3
x 10"2 mg/m3, which is based on BMD modeling of the Rothman et al. (1996a) human data.
Similarly, the chronic oral RfD values calculated on the basis of the human and the
experimental animal data and using BMD modeling and NOAEL/LOAEL approaches are in
generally good agreement, yielding values that range from 7 x 10"4 to 6 x 10"3 mg/kg/day. This
consistency in results provides increased confidence in the selected chronic oral RfD of 4 x 10"3
mg/kg/day, which is based on BMD modeling of the Rothman et al. (1996a) human data.
The RfC and RfD values based on the Rothman et al. (1996a) human study were selected
over the values derived from experimental animal studies because they are based on good-
quality human data and therefore are not subject to the uncertainties inherent in interspecies
extrapolation. Reference values based on BMD modeling were selected over those calculated
using the NOAEL/LOAEL approach because BMD modeling is a superior methodology that
makes better use of the exposure-response data. In any event, especially in the case of the RfC,
all the estimates are in sufficiently good agreement as to be effectively indistinguishable. The
overall confidence in this RfC and RfD assessment is medium.
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF
HAZARD AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
Benzene is widely used as an industrial solvent, as an intermediate in chemical syntheses,
and as a gasoline additive (NTP, 1994). Because of its widespread use, the potential for human
exposure is great. The toxicity of benzene has been recognized for more than a century, and the
141
-------
Table 14. Summary of RfC and RfD estimates using human and experimental
animal data, as well as BMD modeling and LOAEL/NOAEL approaches"
Approach
BMD modeling
LOAEL/NOAE
L
RfC (mg/m3)
Human
3 x 10 2
9 x 1C'3
Rodent
5 x 1C'2
6 x 1C'2
RfD (mg/kg/day)
Human
4 x 10 3
1 x 1C'3
Rodent
7 x 1C'4
6 x 1C'3
Selected values are in bold.
biological impacts of benzene exposure have been extensively studied in humans and in
experimental animal models. Metabolism of benzene is necessary for the compound's toxic
effects to develop. Evidence has accumulated indicating that oxidation of benzene by CYP2E1
in the liver is the first step in initiation of benzene toxicity. Convincing evidence of the
importance of CYP2E1 was provided by Valentine et al. (1996), who showed that genetic
knockout mice lacking expression of the CYP2E1 protein produced much lower levels of
benzene metabolites and failed to develop signs of genotoxicity and hematotoxicity following
acute benzene exposure at dose levels that resulted in severe genotoxicity and cytotoxicity in
both wild-type and B6C3F1 mice.
The majority of benzene metabolism occurs in the liver, but the bone marrow is the target
organ where its toxicity is expressed with the greatest sensitivity. The major hepatic metabolites
of benzene are phenol, catechol, and hydroquinone. Catechol and hydroquinone have been
shown to accumulate in bone marrow after benzene exposure (Rickert et al., 1979). The bone
marrow has high peroxidase activity, which results in oxidation of the phenolic metabolites
produced in the liver to the highly reactive 1,4-benzoquinone (Smith et al., 1989). Other target
tissues are also characterized by high peroxidase activity (Low et al., 1995). The metabolic basis
for the toxicity of benzene has been extensively studied (Snyder and Hedli, 1996).
The most frequently observed toxic effect of benzene, both in humans and test animal
models, is bone marrow depression, which leads to lymphocytopenia, leukocytopenia,
thrombocytopenia, anemia, and aplastic anemia (Aksoy, 1991; Goldstein, 1988; Dosemeci et al.,
1996). The most sensitive effect observed in humans is the depression of ALC in peripheral
blood (Rothman et al., 1996a). In test animal studies, the most sensitive effects observed are
depressions of the colony-forming ability of bone marrow progenitor cells. These cells are
responsible for producing the blood cells needed to replace the aging blood cells in the
circulatory system. The regulation of hematopoiesis is a dynamic process in which stem and
progenitor cells, in conjuction with bone marrow stroma, give rise to mature blood cells. The
survival and proliferation of these cells are controlled by multiple growth factors or cytokines
that regulate hematopoiesis. Several studies in mice and human bone marrow cultures have
shown that benzene alters cytokine production or response to cytokines (Dempster and Snyder,
1990; Farris et al., 1993; Cronkite et al., 1989; Irons et al., 1992; Irons and Stillman, 1993, 1996;
Rothman et al., 1996b)
142
-------
Although a large number of human and experimental animal studies have been
conducted, there are few human studies with reliable estimates of exposure to benzene and few
long-term, repeated-dose experiments in test animals. Human studies frequently are also
complicated by exposure to other solvents. The long-term test animal studies have used
exposure levels that were too high to establish a reliable NOAEL, as significant adverse effects
were observed even in the lowest dose tested in all the long-term studies examined. Thus, the
lack of reliable NOAEL values from either human or test animal studies is an area of uncertainty
in establishing RfD and RfC values that are protective of human health. The use of BMD
modeling obviates the need for NOAELs; however, good-quality exposure-response data from
the low end of the observable response range are still required to reliably estimate BMDs.
Another area of scientific uncertainty in this assessment concerns the neurotoxic effects
of benzene. As is the case with many other organic solvents, benzene has been shown to
produce neurotoxic effects in experimental animals and humans after short-term exposures to
relatively high concentrations of the compound. Benzene produces generalized symptoms such
as dizziness, headache, and vertigo, leading to drowsiness, tremor, delirium, and loss of
consciousness. In an occupational study, Kahn and Muzyka (1973) reported that workers
complained of frequent headaches (usually at the end of the workday), tired easily, had
difficulties sleeping, and complained of memory loss. Overall, there is a lack of reliable
information on dose-related neurotoxic effects under low-dose chronic exposure conditions in
either humans or experimental animal model systems. Li et al. (1992) reported that forelimb grip
strength and the frequency of rapid response in Y-maze running in Kunming mice was increased
following brief inhalation exposure to 0.78 ppm (2.5 mg/m3) benzene, but both responses were
decreased at concentrations of 3.13 ppm (10 mg/m3) or higher. Several experimental
deficiencies in this study prevent its use for calculating the risk to humans, but further
investigation of these effects could reveal neurotoxic effects of concern to human health.
There have been a number of developmental and reproductive studies in humans and in
test animal model systems. Several test animal studies have shown developmental effects
exhibiting manifestations such as reductions in numbers of live fetuses, reductions in live
weight, and minor skeletal variants, but the benzene concentrations used also caused severe
maternal toxicity. The studies by Keller and Snyder (1986, 1988), however, demonstrated that
exposure to low concentrations of 5, 10, or 20 ppm (16, 32, or 64 mg/m3) benzene in utero
during development caused changes in colony-forming hematopoietic cells. These studies were
considered supporting, because the LOAEL of 5 ppm is below the LOAEL of 7.6 ppm
established for hematotoxic and immunotoxic effects observed in humans (Rothman et al.,
1996a). However, the confidence in the human data is much higher than that in the limited
hematotoxic endpoints measured in mice by Keller and Snyder (1986, 1988). Furthermore, the
responses did not establish consistent patterns in different ages of the progeny. Also, a limited
number of animals were examined. These limitations make this study less useful because of the
high degree of uncertainty that these are truly adverse effects. (See section 4.2.2.2.)
143
-------
6.2. DOSE RESPONSE
6.2.1. Inhalation RfC
Quantitative estimates of human health risk as a result of low-level chronic exposure to
benzene via the inhalation route are based on data from human occupational inhalation exposure
and from a subchronic inhalation study in experimental animals. Hematotoxicity and
immunotoxicity are the critical effects observed in both humans and test animals.
The air concentration of benzene considered to be without any appreciable risk with
lifetime chronic human exposure (the RfC) is 3 x 10"2 mg/m3. This value was obtained by
applying a UF of 300 to the BMCLADJ obtained from BMR modeling of the exposure-response
data for reductions in ALC from the Rothman et al. (1996a) occupational study, using a BMR
level of one standard deviation change from the control mean response. This RfC is in good
agreement with the value of 9 x 10"3 mg/m3 based on the LOAEL in the Rothman et al. (1996a)
study, with a UF of 1000.
The RfC of 3 x 10"2 mg/m3 is similarly in good agreement with the value of 5 x 10"2
mg/m3 obtained by applying a UF of 1000 to the BMCLADJ estimated from BMD modeling of the
exposure-response data for the hematologic effect of decreased HCT in male mice from the
Ward et al. (1985) subchronic inhalation study, using a BMR level of one standard deviation
change from the control mean response. Finally, both of these values are in good agreement
with the value of 6 x 10'2 mg/m3 based on the NOAEL in the Ward et al. (1985) study, with a UF
of 300.
The overall confidence in this RfC assessment is medium. The Rothman et al. (1996a)
and Ward et al. (1985) studies were both well conducted, and various methodologies for deriving
the RfC yielded similar results for the two studies. Furthermore, the availability of good-quality
human data for a sensitive endpoint eliminates the uncertainty associated with basing the RfC on
experimental animal data. However, with continuous endpoints such as the hematologic
parameters measured in these studies, there is uncertainty about when a change in a parameter
that has inherent variability becomes an adverse effect. Other uncertainties explicitly recognized
in the quantitative derivation include intraspecies variability, that is, the need to accommodate
sensitive human subgroups; the extrapolation of subchronic results to a lifetime exposure
scenario; and database deficiencies.
6.2.2. OralRfD
Quantitative estimates of human health risk as a result of low-level chronic exposure to
benzene via the oral route are based on data from human occupational inhalation exposure and
from a chronic gavage study in experimental animals. Once again, hematotoxic responses are
the critical effects observed.
The human chronic dose of ingested benzene considered to be without any appreciable
risk (the RfD) is 4 x 10"3 mg/kg/day. This value was obtained by applying a UF of 300 to the
oral equivalent dose extrapolated from the RMCL^ obtained from BMD modeling of the ALC
data from the human occupational study of Rothman et al. (1996a) using a BMR level of one
144
-------
standard deviation change from the control mean response, as for the RfC discussed above. This
RfD is in good agreement with the value of 1 x 10"3 mg/kg/day based on the oral equivalent
LOAEL from the Rothman et al. (1996a) study, with a UF of 1000.
This RfD of 4 x 10"3 mg/kg/day is also in good agreement with the values of 7 x 10"4
mg/kg/day and 6 x 10"3 mg/kg/day, which are 1/1000 of the BMDL^j derived from the male rat
ALC data in the NTP (1986) chronic gavage study and 1/3000 of the LOAEL from the NTP
(1986) study, respectively.
The overall confidence in this RfD assessment is medium. The Rothman et al. (1996a)
study was well conducted, and the availability of good-quality human data for a sensitive
endpoint eliminates the uncertainty associated with basing the RfD on experimental animal data.
However, with continuous endpoints such as hematologic parameters, there is uncertainty about
when a change in a parameter that has inherent variability becomes an adverse effect. Other
uncertainties explicitly recognized in the quantitative derivation include intraspecies variability,
to accommodate sensitive human subgroups; the extrapolation of subchronic results to a lifetime
exposure scenario; and database deficiencies.
A further uncertainty in the RfD ensues from the use of route-to-route extrapolation to
estimate oral equivalent doses from inhalation exposures resulting from analysis of the Rothman
et al. (1996a) occupational data. In experiments conducted to compare the metabolite doses to
the target organ following oral or inhalation exposure, Sabourin et al. (1987, 1989) found that
there was no simple relationship between the two routes of exposure. Oral doses and inhalation
exposures that produced similar concentrations of one metabolite in the blood produced very
different doses of another metabolite. The target specificity of benzene toxicity for the bone
marrow progenitor cells irrespective of route of administration, however, is well documented
both in humans and experimental animal models. Thus, route-to-route extrapolation is justified
and introduces a lower degree of uncertainty than extrapolating from test animals to humans
(U.S. EPA, 1999a). Use of a modifying factor of 3 to account for uncertainty in the route-to-
route extrapolation was considered; however, it was deemed unnecessary. The RfD is based on
human data for a sensitive endpoint; thus, it was felt that the composite UF of 300 provides
sufficient protection.
145
-------
7. REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). (1996) Threshold limit values for chemical
substances and physical agents and biological exposure indices. American Conference of Governmental Industrial
Hygienists. Cincinnati, OH.
Aksoy, M. (1988) Benzene hematotoxicity. In: Benzene carcinogenicity. Aksoy, M, ed. Boca Raton, FL: CRC
Press, pp. 59-112.
Aksoy, M. (1989) Hematotoxicity and carcinogenicity of benzene. Environ Health Perspect 82:193-197.
Aksoy, M. (1991) Hematotoxicity, leukemogenicity and carcinogenicity of chronic exposure to benzene. In:
Molecular aspects of monooxygenases and bioactivation of toxic compounds. Arinc, E; Schenkman, JB; Hodgson,
E, eds. New York: Plenum Press, pp. 415-434.
Aksoy, M; Erdem, K. (1978) A follow-up study on the mortality and the development of leukemia in 44
pancytopenic patients associated with long-term exposure to benzene. Blood 52:285-292.
Aksoy, M; Dincol, K; Akgun, T; et al. (1971) Hematological effects of chronic benzene poisoning in 217 workers.
BrJIndMed 28:296-302.
Aksoy, M; Dincol, K; Erdem, T; et al. (1972) Details of blood changes in 32 patients with pancytopenia associated
with long-term exposure to benzene. Br J Ind Med 29:56-64.
Aksoy, M; Erdem, S; Dincol, G. (1974) Leukemia in shoe-workers exposed chronically to benzene. Blood. 44:837-
841.
Aksoy, M; Ozemis, S; Sabuncu, H; et al. (1987) Exposure to benzene in Turkey between 1983 and 1985: A
hematological study on 231 workers. Br J Ind Med 44:785-787.
American Petroleum Institute. (1983) Evidence for hematotoxicity and tumorigenesis in rats exposed to 100 ppm
benzene. TSCA FYI submission. OTS fiche #OTS0000241-0.
Anderson, D; Richardson, CR. (1981) Issues relevant to the assessment of chemically induced chromosome damage
in vivo and their relationship to chemical mutagenesis. Mutat Res 90:261-272.
Andersson, K; Nilsen, OG; Toftgard, R; et al. (1983) Increased amine turnover in several hypothalamic
noradrenaline nerve terminal systems and changes in prolactin secretion in the male rat by exposure to various
concentrations of toluene. Neurotoxicology 4:43-55.
Andrews, LS; Lee, EW; Witmer, CM; et al. (1977) Effects of toluene on the metabolism, disposition and
hematopoietic toxicity of [3H]benzene. Biochem Pharmacol 26:293-300.
Andrews, LS; Sasame, HA; Gillette, JR. (1979) 3H-Benzene metabolism in rabbit bone marrow. Life Sci 25:567-
572.
Aoyama, K. (1986) Effects of benzene inhalation on lymphocyte subpopulations and immune response in mice.
Toxicol Appl Pharmacol 85(1):92-101.
Arinc, E; Adali, O; Iscan, M; et al. (1991) Stimulatory effects of benzene on rabbit liver and kidney microsomal
cytochrome P-450 dependent drug metabolizing enzymes. Arch Toxicol 65:186-190.
ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for benzene (update).
Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
Au, WW; Ramanujam, VM; Ward Jr., JB. (1991) Chromosome aberrations in lymphocytes of mice after sub-acute
low-level inhalation exposure to benzene. Mutat Res 260(2):219-224.
146
-------
Axelsson, G; Luetz, C; Rylander, R. (1984) Exposure to solvents and outcome of pregnancy in university
laboratory employees. BrJIndMed 41:305-312.
Baarson, KA; Snyder, CA. (1991) Evidence for the disruption of the bone marrow microenvironment by combined
exposures to inhaled benzene and ingested ethanol. Arch Toxicol 65:414-420.
Baarson, KA; Snyder, CA; Albert, RE. (1984) Repeated exposure of C57B1 mice to inhaled benzene at 10 ppm
markedly depressed erythropoietic colony formation. Toxicol Lett 20:337-342.
Banerjee, S; Yalkowsky, SH; Valvani, SC. (1980) Water solubility and octanol/water partition coefficients of
organics. Limitations of the solubility-partition coefficient correlation. Environ Sci Technol 14:1227-1229.
Bartczak, A; Kline, SA; Uy, R; et al. (1994) Evaluation of assays for the identification and quantitation of muconic
acid, a benzene metabolite in human urine. J Toxicol Environ Health 42:245-258.
Baslo, A; Aksoy, M. (1982) Neurological abnormalities in chronic benzene poisoning: A study of 6 patients with
aplastic anemia and 2 with preleukemia. Environ Res 27:457-465.
Bechtold, WE; Henderson, RF. (1993) Biomarkers of human exposure to benzene. J Toxicol Environ Health
40:377-386.
Bechtold, WE; Lucier, G; Birnbaum, LS; et al. (1991) Muconic acid determinations in urine as a biological
exposure index for workers occupationally exposed to benzene. Am Ind Hyg Assoc J 52:473-478.
Bernauer, U; Vieth, B; Ellrich, R; et al. (2000) CYP2E1 expression in bone marrow and its intra- and interspecies
variability: approaches for a more reliable extrapolation from one species to another in the risk assessment of
chemicals. Arch Toxicol 73(12):618-624.
Billips, LG; Petitte, D; Hostutler, R; et al. (1991) Suppression of bone marrow stromal cell fucntion. Ann NY Acad
Sci 628:313-322.
Boersma, GM; Balvers, WG; Boeren, S; et al. (1994) NADPH-cytochrome reductase catalyzed redox cycling of
1,4-benzoquinone; hampered at physiological conditions, initiated at increased pH values. Biochem Pharmacol
47:1954-1955.
Bogadi-Sare, A; Zavalic, M; Trosic, I; et al. (2000) Study of some immunological parameters in workers
occupationally exposed to benzene. Int Arch Occup Environ Health 73:397-400.
Bois, FY; Paxman; DG. (1992) An analysis of exposure rate effects for benzene using a physiologically based
pharmacokinetic model. Regul Toxicol Pharmacol 15:122-136.
Bois, FY; Smith, MT; Spear, RC. (1991a) Mechanisms of benzene carcinogenesis: Application of a physiological
model of benzene pharmacokinetics and metabolism. Toxicol Lett 56:283-298.
Bois, FY; Woodruff, TJ; Spear, RC. (199 Ib) Comparison of three physiologically based pharmacokinetic models of
benzene disposition. Toxicol Appl Pharmacol 110:79-88.
Bois, FY; Jackson, ET; Pekari, K; et al. (1996) Population toxicokinetics of benzene. Environ Health Perspect 104
(supp!6):1405-1412.
Bordarier, C; Robain, O; Ponsot, G. (1991) Bilateral porencephalic defect in a newborn after injection of benzol
during pregnancy. Brain Dev 13:126-129.
Brief, RS; Lynch, J; Bernath, T; et al. (1980) Benzene in the workplace. Am Ind Hyg Assoc J 41:616-623.
Brown, EA; Shelley, ML; Fisher, JW. (1998) A pharmacokinetic study of occupational and environmental benzene
exposure with regard to gender. Risk Anal 18:205-213.
147
-------
Brunmark, A; Cadenas, E. (1988) Reductive addition of glutathione to/>-benzoquinone, 2-hydroxy-p-
benzoquinone, and/>-benzoquinone epoxides. Effect of hydroxy- and glutathionyl substituents on/>-benzoquinone
autooxidation. Chem Biol Interact 86:273-298.
Buckley, JD; Robison, LL; Swotinsky, L; et al. (1989) Occupational exposure of parents of children with acute
nonlymphocytic leukemia: a report from the Children's Cancer Study Group. Cancer Res 49:4030-4037.
Budvari, S (ed.). (1989) The Merck Index: An encyclopedia of chemicals, drugs, and biologicals. 11th ed.
Rathway, NJ: Merck & Co.
California EPA. (2001) Public Health Goal for Benzene in Drinking Water. Office of Environmental Health Hazard
Assessment, California Environmental Protection Agency, Sacramento.
Carpenter, CP; Shaffer, CB; Weil, CS; et al. (1944) Studies on the inhalation of 1:3 butadiene; with a comparison of
its narcotic effect with benzol, toluol, and styrene, and a note on the elimination of styrene by the human. J Ind Hyg
Toxicol 26:69-78.
Cassidy, MK; Houston, JB. (1984) In vivo capacity of hepatic and extra hepatic enzymes to conjugate phenol.
Drug Metab Dispos 12:619-624.
Chatburn, G; Sharratt, M; Wickramaratne, GA. (1981) Chemical Industries Association/Institute of Petroleum Joint
Committee on Benzene: reports of task forces on toxicology and teratology of benzene. II. Reproductive effects,
embryotoxicity. Regul Toxicol Pharmacol 1:205-210.
Chen, H; Eastmond, DA. (1995) Topoisomerase inhibition by phenolic metabolites: A potential mechanism for
benzene's clastogenic effects. Carcinogenesis 16(10):2301-2307.
Chenna, A; Hang, B; Rydberg, B; et al. (1995) The benzene metabolite p-benzoquinone forms adducts with DNA
bases that are excised by a repair activity from human cells that differs from an ethenoadenine glycosylase. Proc
Natl Acad Sci USA 92(13):5890-5894.
Chepiga, TA; Yang, CS; Snyder, R. (1990) Benzene metabolism by two purified, reconstituted rat hepatic mixed
function oxidase systems. Adv Exp Med Biol 283:261-265.
Chertkov, JL, Lutton, JD; Jiang, S; et al. (1992) Hematopoietic effects of benzene inhalation assessed by murine
long-term bone marrow culture. J Lab ClinMed 119:412-419.
Coate, WB; Hoberman, AM; Durloo, RS. (1984) Inhalation teratology study of benzene in rats. Adv Mod Environ
Toxicol 6:187-198.
Cody, RR; Strawderman, WW; Kipen, HM. (1993) Hematologic effects of benzene. J Occup Med 35(8):776-782.
Collins, JJ; Connor, P; Friedlander, BR; et al. (1991) A study of the hematologic effects of chronic low-level
exposure to benzene. J Occup Med 33:619-626.
Cornish, HH; Ryan, RC. (1965) Metabolism of benzene in nonfasting, fasted and aryl-hydroxylase inhibited rats.
Toxicol Appl Pharmacol 7:767-771.
Cord, M; Snyder, CA. (1996) Influences of gender, development, pregnancy and ethanol consumption on the
hematotoxicity of inhaled 10 ppm benzene. Arch Toxicol 70:209-217.
Cox, LA. (1991) Biological basis of chemical carcinogenesis: insights from benzene. Risk Anal 11:453-464.
Cox, LA. (1996) Reassessing benzene risks using internal doses and Monte-Carlo uncertainty analysis. Environ
Health Perspect 104 (suppl 6): 1413-1429.
148
-------
Creek, RM; Mani, C; Vogel, JS; et al. (1997) Tissue distribution and macromolecular binding of extremely low
doses of [14C]-benzene inB6C3Fl mice. Carcinogenesis 18:2421-2427.
Cronkite, EP; Drew, RT; Inoue, T; et al. (1985) Benzene hematotoxicity and leukemogenesis. Am J Ind Med 7:447-
456.
Cronkite, EP, Drew, RT; Inoue, T; et al. (1989) Hematotoxicity and carcinogenicity of inhaled benzene. Environ
Health Perspect 82:97-108.
Crump, KS. (1992) Exposure-response analyses of Pliofilm cohort. Work supported by Western States Petroleum
Association, Fairfax, VA. Draft.
Crump, KS; Allen, BC. (1984) Quantitative estimates of risk of leukemia from occupational exposure to benzene.
Prepared for the Occupational Safety and Health Administration by Science Research Systems, Inc., Ruston, LA.
Daiker, DH; Moslen, MT; Carr, JB; et al. (1996) Repeated oral benzene exposure alters enzymes involved in
benzene metabolism. J Toxicol Environ Health 48:43 9-451.
Dean, BJ. (1985) Recent findings on the genetic toxicology of benzene, toluene, xylenes, and phenols. Mutat Res
154:153-181.
Deichmann, WB; MacDonald, WE; Bernal, E. (1963) The hemopoietic tissue toxicity of benzene vapors. Toxicol
Appl Pharmacol 5:201-224.
Dempster, AM; Snyder, CA. (1990) Short term benzene exposure provides a growth advantage for granulopoietic
progenitor cells over erythroid progenitor cells. Arch Toxicol 64:539-544.
Dempster, AM; Snyder, CA. (1991) Kinetics of granulocytic and erythroid progenitor cells are affected differently
by short-term, low-level benzene exposure. Arch Toxicol 64(7):556-561.
Dempster, AM; Evans, HL; Snyder, CA. (1984) The temporal relationship between behavioral and hematological
effects of inhaled benzene. Toxicol Appl Pharmacol 76:195-203.
Desoille, H., Philbert, M; Albahary, C. (1961) The effect of hormones on chronic benzene poisoning in rabbits. I.
The influence of natural oestrogens on the white and red corpuscles. Arch Mai Prof 12:681-693.
Ding, X-J; Li, Y; Ding Y; et al. (1983) Chromosome changes in patients with chronic benzene poisoning. Chin
Med J (Peking Engl. Ed.) 96:681-685.
Dorshkind, K. (1990) Regulation of hemopoiesis by bone marrow stromal cells and their products. Annu Rev
Immunol8:lll-137.
Dosemeci, M; Yin, S-N; Linet, M; et al. (1996) Indirect validation of benzene exposure assessment by association
with benzene poisoning. Environ Health Perspect 104 (suppl 6): 1343-1347.
Dosemeci, M; Rothman, N; Yin, SN; et al. (1997) Validation of benzene exposure assessment. Ann NY Acad Sci
837:114-121.
Dow Chemical Co. (1992a) Initial submission: Effects of benzene vapor in the pig and rat pertaining to hematology
and immunology with cover letter dated 05/14/92. EPA/OTS Doc # 88-920003196.
Dow Chemical Co. (1992b) Initial submission: Embryotoxicity study with inhaled benzene in mice and rabbits with
cover letter dated 08/20/92. EPA/OTS Doc # 88-920006580.
Dowty, BJ; Laseter, JL; Storer, J. (1976) The transplacental migration and accumulation in blood of volatile
organic constituents. PediatrRes 10:696-701.
149
-------
Drummond, JC; Finar, IL. (1938) Muconic acid as a metabolic product of benzene. Biochemistry 32:79-84.
Eastmond, DA. (1993) Induction of micronuclei and aneuploidy by the quinone-forming agents benzene and o-
phenylphenol. Toxicol Lett 67(1-3): 105-118.
Eastmond, DA; Smith, MT; Irons, RD. (1987) An interaction of benzene metabolites reproduces the myelotoxicity
observed with benzene exposure. Toxicol Appl Pharmacol 91(7) 85-95.
Eastmond, DA; Rupa, DS; Hasegawa, LS. (1994) Detection of hyperdiploidy and chromosome breakage in
interphase human lymphocytes following exposure to the benzene metabolite hydroquinone using multicolor
fluorescence in situ hybridization with DNA probes. Mutat Res 322(1):9-20.
Edwards, YH; Potter, J; Hopkinson, DA. (1980) Human FAD-dependent NAD(P)H diaphorase. BiochemJ
187:429-436.
Erexson, GL; Wilmer, JL; Steinhagen, WH; et al. (1986) Induction of cytogenetic damage in rodents after short-
term inhalation of benzene. Environ Mutagen 8:29-40.
Erf, LA; Rhoads, CP. (1939) The hematological effects of benzene (benzol) poisoning. J Ind Hyg Toxicol 20:421-
435.
Evans, HL; Dempster, AM; Snyder, CA. (1981) Behavioral changes in mice following benzene inhalation.
Neurobehav Toxicol Teratol 3:481-485.
Exxon Chemical Company. (1986) Determination of maternal toxicity and fetal toxicity of benzene in rats
following oral exposure. TSCA 8E submission. OTS Fiche # OTS0536017.
Fan, X-H. (1992) Effect of exposure to benzene on natural killer (NK) cell activity and interleukin-2 (IL-2)
production in C57BL/6 mice. J Nippon Med Sch 59:393-399.
Farris, GM; Everitt, JI; Irons, RD; et al. (1993) Carcinogenicity of inhaled benzene in CBA mice. Fundam Appl
Toxicol 20(4):503-507.
Farris, GM; Robinson, SN; Gaido, KW; et al. (1996) Effects of low concentrations of benzene on mouse
hematopoietic cells in vivo. A preliminary report. Environ Health Perspect 104 (suppl 6): 1275-1276.
Farris, GM, Robinson, SN; Gaido, KW; et al. (1997a) Benzene-induced hematotoxicity and bone marrow
compensation in B6C3F1 mice. Fundam Appl Toxicol 36:119-129.
Farris, GM; Robinson, SN; Wong, BA; et al. (1997b) Effects of benzene on splenic, thymic, and femoral
lymphocytes in mice. Toxicology 118:137-148.
Fishbeck WA, Townsend, JC; Swank, MG. (1978) Effects of chronic occupational exposure to measured
concentrations of benzene. J Occup Med 20:539-542.
Forni, A; Moreo, L. (1967) Chromosome studies in a case of benzene-induced erythroleukaemia. Eur J Cancer
5:459-463.
Forni A, Pacifico, E; Limonta, A. (1971) Chromosome studies in workers exposed to benzene or toluene or both.
Arch Environ Health 22:373-378.
Frantik, E; Hornychova, M; Horvath, M. (1994) Relative acute neurotoxicity of solvents: Isoeffective air
concentrations of 48 compounds evaluated in rats and mice. Environ Res 66:173-185.
Frantz, CE; Chen, H; Eastmond, DA. (1996) Inhibition of human topoisomerase II in vitro by bioactive benzene
metabolites. Environ Health Perspect 104 (suppl 6):1319-1323.
150
-------
Franz, TJ. (1983) Percutaneous absorption. On the relevance of in vitro data. J Invest Dermatol 64:190-195.
Franz, TJ. (1984) Percutaneous absorption of benzene. In: Advances in modern environmental toxicology. VolVI.
Applied toxicology of petroleum hydrocarbons. MacFarland, HN; Holdsworth, CE; MacGregor, JA; et al., eds.
Princeton, NJ: Princeton Scientific Publishers, pp. 61-70.
Fujie, K; Ito, Y; Maeda, S. (1992) Acute cytogenetic effect of benzene on rat bone marrow cells in vivo and the
effect of inducers or inhibitors of drug-metabolizing enzymes. MutatRes 298(2):81-90.
Funes-Cravioto, F; Zapata-Gayon, C; Kolmodin-Hedman, B; et al. (1977) Chromosome aberrations and sister
chromatid exchange in workers in chemical laboratories and a rotoprinting factory and in children of women
laboratory workers. Lancet 2:322-325.
Gad-El Karim, MM; Ramanujam, VMS; Legator, MS. (1985) trans,trans-Mucovic acid, an open chain urinary
metabolite of benzene in mice: Quantification by high-pressure liquid chromatography. Xenobiotica 15:211-220.
Gaido, K; Wierda, D. (1984) In vitro effects of benzene metabolites on mouse bone marrow stromal cells. Toxicol
Appl Pharmacol 76:45-55.
Gaido, K; Wierda, D. (1985) Modulation of stromal cell function in DBA/2J and B6C3F1 mice exposed to benzene
or phenol. Toxicol Appl Pharmacol 81:469-475.
Ganousis, LG; Goon, D; Zyglewska, T; et al. (1992) Cell-specific metabolism in mouse bone marrow stroma:
studies of activation and detoxification of benzene metabolites. Mol Pharmacol 42:1118-1125.
Garnett, H; Cronkhite, EP; Drew, RT. (1983) Effect of in vivo exposure to benzene on the characteristics of bone
marrow adherent cells. Leuk Res 7:803-810.
Center, MB; Reico, L. (1994) Absence of detectable P450 2E1 in bone marrow of B6C3F1 mice: relevance to
butadiene-induced bone marrow toxicity. Fundam Appl Toxicol 22:469-473.
Ghantous, H; Danielsson, BRG. (1986) Placenta! transfer and distribution of toluene, xylene, and benzene, and
their metabolites during gestation in mice. Biol Res Pregnancy 7:98-105.
Ghittori, S; M.L. Fiorentino, ML; L. Maestri, L; et al. (1993) Urinary excretion of unmetabolized benzene as an
indicator of benzene exposure. J Toxicol Environ Health 38:233-243.
Gofmekler, VA. (1968) Effect of embryonic development of benzene and formaldehyde in inhalation treatments.
HygSanit 33:327-332.
Goldstein, BD. (1988) Benzene toxicity. State of the art reviews. OccupMed 3:541-554.
Goldstein, BD; Cody, R. (2000) Assessment of complete blood count variations among workers exposed to low
levels of benzene [letter]. J Occup Environ Med 42(2): 113-114.
Goldwater, LJ. (1941) Disturbances in the blood following exposure to benzene (benzol). J Lab Clin Med 26:957-
973.
Goldwater, LJ; Tewksbury, MP. (1941) Recovery following exposure to benzene (benzol). JIndHyg23:217.
Gonasun, LM; Witmer, C; Kocsis, J; et al. (1973) Benzene metabolism in mouse liver microsomes. Toxicol Appl
Pharmacol 26:398-406.
Goon, D; Matsura, J; Ross, D. (1993) Metabolism and cytotoxicity of/raw5,/ra«5-muconaldehyde and its
derivatives: potential markers of benzene ring cleavage reactions. Chem Biol Interact 88:37-53.
Greaves, MF. (1993) Stem cell origins of leukemia and curability. Br J Cancer 67:413-423.
151
-------
Green, JK; Leong, BKJ; Laskin S. (1978) Inhaled benzene fetotoxicity in rats. Toxicol Appl Pharmacol 46:9-18.
Green, JD; Snyder, CA; LoBue, J; et al. (1981a) Acute and chronic dose/response effect of benzene inhalation on
the peripheral blood, bone marrow, and spleen cells of CD-I mice. Toxicol Appl Pharmacol 59:204-214.
Green, JD; Snyder, CA; LoBue, J; et al. (1981b) Acute and chronic dose/response effect of inhaled benzene on the
multipotential hematopoietic stem (CFU-S) and granulocyte/macrophage progenitor (GM-CFU-C) cells in CD-I
mice. Toxicol Appl Pharmacol 58:492-503.
Greenburg, L. (1926) Benzol poisoning as an industrial hazard. Public Health Rep 41:1357-1375.
Greenburg, L; Mayers, MR; Goldwater, L; et al. (1939) Benzene poisoning in rotogravure printing. JIndHyg
Toxicol 21:395-420.
Greenlee, WF; Gross, EA; Irons, RD. (1981) Relationship between benzene toxicity and the disposition of 14C-
labeled benzene metabolites in the rat. Chem Biol Interact 33:285-299.
Grotz, VL; Ji, S; Kline, SA; et al. (1994) Metabolism of benzene and trans,trans-muconaldehyde in the isolated
perfused liver. Toxicol Lett 70:281-290.
Guberan, E; Kocher, P. (1971) Pronostic lointain de l'intoxication benzolique chronique: controle d'une population
10 ans apres 1'exposition. Schweiz Med WochnschrlO 1:1789.
Hansch, C; Leo, AJ. (1985) Medchem project. Issue No. 26. Pomona College, Claremont, CA.
Hazel, BA; O'Connor, A; Niculescu, R; et al. (1996) Induction of granulocytic differentiation in a mouse model by
benzene and hydroquinone. Environ Health Perspect 104 (suppl 6): 1257-1264.
Hedli, CC; Rao, NR; Reuhl, KR; et al. (1996) Effects of benzene metabolite treatment on granulocytic
differentiation and DNA adduct formation in HL-60 cells. Arch Toxicol 70:135-144.
Henderson, R.F. (1996) Species differences in the metabolism of benzene. Environ Health Perspect 104 (suppl
6):1173-1175.
Henschler, R; Glatt, HR. (1995) Induction of cytochrome P4501al in haemopoietic stem cells by hydroxylated
metabolites of benzene. Toxicol in vitro 9:453-457.
Hernberg, S; Savilahti, M; Ahlman, K; et al. (1966) Prognostic aspects of benzene poisoning. Br J Ind Med
23:204.
Herregods, P; Chappel, R; Mortier, G. (1984) Benzene poisoning as a possible cause of transverse myelitis.
Paraplegia 22:305-310.
Holmberg, PC. (1979) Central nervous system defects in children born to mothers exposed to organic solvents
during pregnancy. Lancet 2:177-179.
Hough, H; Freeman, S. (1944) Relative toxicity of commercial benzene and a mixture of benzene, toluene and
xylene. Fed Proc Fed Am Soc Exp Biol 3:20.
HSDB (Hazardous Substances Data Bank). (1997) National Library of Medicine, National Toxicology Program
(Via Toxnet), Bethesda, MD.
Hsieh, GC; Parker, RD; Sharma, RP. (1988a) Subclinical effects of groundwater contaminants. II. Alteration of
regional brain monoamine neurotransmitters by benzene in CD-I mice. Arch Environ Contam Toxicol 17:799-805.
Hsieh, GC; Sharma, RP; Parker, RDR. (1988b) Subclinical effects of groundwater contaminants. I. Alteration of
humoral and cellular immunity by benzene in CD-I mice. Arch Environ Contam Toxicol 17:151-158.
152
-------
Hu, JJ; Lee, MJ; Vapiwala, M; et al. (1993) Sex-related differences in mouse renal metabolism and toxicity of
acetaminophen. Toxicol Appl Pharmacol 122:16-22.
Huang, X-Y. (1991) Influence on benzene and toluene to reproductive function of female workers in leather shoe-
making industry. Chin J Prev Med 25:89-91. (In Chinese; Eng. abstr. TOXLINE).
Hudak, A; Ungvary, G. (1978) Embryotoxic effects of benzene and its methyl derivatives: Toluene and xylene.
Toxicology 11:55-63.
Huff, JE; Haseman, JK; DeMarini, DM; et al. (1989) Multiple-site carcinogenicity of benzene in Fischer 344 rats
and B6C3F1 mice. Environ Health Perspect 82:125-163.
Hutt, AM; Kalf, GF. (1996) Inhibition of human topoisomerase II by hydroquinone and^-benzoquinone, reactive
metabolites of benzene. Environ Health Perspect 104 (suppl 6): 1265-1269.
Inoue, O; Seiji, K; Kasahara, M; et al. (1986) Quantitative relation of urinary phenol levels to breathzone benzene
concentrations: a factory survey. Br J Ind Med 43:692-697.
Inoue, O; Seiji, K; Nakatsuka, H; et al. (1989) Excretion of 1,2,4-benzenetriol in the urine of workers exposed to
benzene. BrJ Ind Med 46:559-565.
Irons, RD; Neptun, DA. (1980) Effects of principle hydroxy-metabolites of benzene on microtubule
polymerization. Arch Toxicol 45:297-305.
Irons, RD; Stillman, WS. (1993) Cell proliferation and differentiation in chemical leukemogenesis. Stem Cells
11:235-242.
Irons, RD; Stillman, WS. (1996) Impact of benzene metbolites on differentiation of bone marrow progenitor cell.
Environ Health Perspec 6:1247-1250.
Irons, RD; Dent, JG; Baker, TS; et al. (1980) Benzene is metabolized and covalently bound in bone marrow in situ.
ChemBiol Interact 30:241-245.
Irons, RD; Stillman, WS; Colagiovanni, DB; et al. (1992) Synergistic action of the benzene metabolite
hydroquinone on myelopoietic activity of granulocyte/macrophage colony-stimulating factor in vitro. Proc Natl
Acad Sci USA 89:3691-3695.
Ito, T. (1962a) Study on the sex difference in benzene poisoning. Report 1. On the obstacles in benzene workers.
Showa Igakukai Zasshi 22:268-272.
Ito, T. (1962b) Study on the sex difference in benzene poisoning. Report 2. Animal experiment on the sex
difference in benzene inhalation. Showa Igakukai Zasshi 22:273-277.
Ito, T. (1962c) Study on the sex difference in benzene poisoning. Report 3. The influence of femaleness upon the
benzene poisoning. Showa Igakukai Zasshi 22:278-284.
Ito, T. (1962d) Study on the sex difference in benzene poisoning. Report 4. The influence of maleness upon the
benzene poisoning. Showa Igakukai Zasshi 22:285-290.
Jerina, D; Daly, J; Witkop, B; et al. (1968) Role of arene oxide-oxepin system in the metabolism of aromatic
substances. I. In vitro conversion of benzene oxide to a premercapturic acid and a dihydrodiol. Arch Biochem
Biophys 128:176-183.
Johansson, I; Ingelman-Sundberg, M. (1988) Benzene metabolism by ethanol-, acetone-, and benzene-inducible
cytochrome P-450(IIE1) in rat and rabbit liver microsomes. Cancer Res 39:5387-5390.
153
-------
Kahn, H; Muzyka, V. (1973) The chronic effect of benzene on porphyrin metabolism. Work Environ Health
10:140-143.
Kalf, GF. (1987) Recent advances in the metabolism and to xicity of benzene. Crit Rev Toxicol 18:141-159.
Kalf, GF; Renz, JF; Niculescu, R. (1996) />-Benzoquinone, a reactive metabolite of benzene, prevents the
processing of pre-interleukins-lcc and-lp to active cytokines by inhibition of the processing enzymes, calpain, and
interleukin-lp converting enzyme. Environ Health Persp 104 (supp!6):1251-1256.
Keller, KA; Snyder, CA. (1986) Mice exposed in utero to low concentrations of benzene exhibit enduring changes
in their colony forming hematopoietic cells. Toxicology 42:171-181.
Keller, KA; Snyder, CA. (1988) Mice exposed in utero to 20 ppm benzene exhibit altered numbers of recognizable
hematopoietic cells up to seven weeks after exposure. Fundam Appl Toxicol 10:224-232.
Kenyon, EM; Seeley, ME; Janzen, D; et al. (1995) Dose-, route-, and sex-dependent urinary excretion of phenol
metabolites in B6C3F1 mice. J Toxicol Environ Health 44:219-233.
Kenyon, EM; Kraichely, RE; Hudson, KT; et al. (1996) Differences in rates of benzene metabolism correlate with
observed genotoxicity. Toxicol Appl Pharmacol 136:49-56.
Khuder, SA; Youngdale, MC; Bisesi, MS; et al. (1999) Assessment of complete blood count variations among
workers exposed to low levels of benzene. J Occup Environ Med 41:821-826.
Kipen, HM; Cody, RP; Goldstein, BD. (1988) Hematologic effects of benzene: A thirty-five year longitudinal study
of rubber workers. Toxicol Ind Health 4:411-430.
Kipen, HM; Cody, RP; Goldstein, BD. (1989) Use of longitudinal analysis of peripheral blood counts to validate
historical reconstructions of benzene exposure. Environ Health Perspect 82:199-206.
Kline, SA; Robertson, JF; Grotz, VL; et al. (1993) Identification of 6-hydroxy-/raw5,/ra«5-hexadienoic acid, a novel
ring-opened urinary metabolite of benzene. Environ Health Perspect 101:310-312.
Kok, PW; Ong, CN. (1994) Blood and urinary benzene determined by headspace gas chromatography with
photoionization detection: application in biological monitoring of low-level nonoccupational exposure. Int Arch
Occup Environ Health 66:195-201.
Kolachana, P; Subrahmanyam, W; Meyer, KB; et al. (1993) Benzene and its phenolic metabolites produce
oxidative DNA damage in HL60 cells in vivo and in the bone marrow in vivo. Cancer Res 53(5): 1023-1026.
Koop, DR; Laethem, CL. (1992) Inhibition of rabbit microsomal cytochrome P-450 2E1 -dependent/>-nitrophenol
hydroxylation by substituted benzene derivatives. Drug Metab Dispos 20:775-777.
Kraut, A; Lilis, R; Marcus, M; et al. (1988) Neurotoxic effects of solvent exposure on sewage treatment workers.
Arch Environ Health 43:263-268.
Kuna, RA; Kapp, RW. (1981) The embryotoxic/teratogenic potential of benzene vapor in rats. Toxicol Appl
Pharmacol 57:1-7.
Kuna, RA; Nicolich, MJ; Schroeder, RE; et al. (1992) A female rat fertility study with inhaled benzene. J Am Coll
Toxicol 11:275-282.
Lagorio, S; Forastiere, F; lavarone, I; et al. (1994) Exposure to benzene and urinary concentrations of 8-
hydroxydeoxyguanosine, a biological marker of oxidative damage to DNA. Occup Environ Med 51:73 9-743.
Laitinen, J; Kangas, J; Pekari, K; et al. (1994) Short time exposure to benzene and gasoline at garages.
Chemosphere 28:197-205.
154
-------
Landay, A; Bauer, K. (1988) Flow cytometric analysis of cells of the immune system. EOSJImmunol
Immunopharmacol 8:166-171.
Latriano, L; Goldstein, BD; Witz, G. (1986) Formation of muconaldehyde, an open-ring metabolite of benzene, in
mouse liver microsomes: An additional pathway for toxic metabolites. Proc Natl Acad Sci USA 83:8356-8360.
Legathe, A; Hoener, BA; Tozer, TN. (1994) Pharmacokinetic interaction between benzene metabolites, phenol and
hydroquinone, in B6C3F! mice. Toxicol Appl Pharmacol 124:131-138.
Lewis, JG; Odom, B; Adams, DO. (1988) Toxic effects of benzene and benzene metabolites on mononuclear
phagocytes. Toxicol Appl Pharmacol 92:246-254.
Li, G-L; Yin, S-N; Watanabe, T; et al. (1986) Benzene-specific increase in leukocyte alkaline phosphatase activity
in rats exposed to vapors of various organic solvents. J Toxicol Environ Health 19:581-589.
Li, L; Sun, W; Gong, Z; et al. (1992) Effect of low benzene exposure on neurobehavioral function, AChE in blood
and brain and bone marrow picture in mice. Biomed Environ Sci 5:349-354.
Li, G-L; Linet, MS; Hayes, RB; et al. (1994) Gender differences in hematopoietic and lymphoproliferative
disorders and other cancer risks by major occupational group among workers exposed to benzene in China. J Occup
Med36(8):875-881.
Lindstrom, AB; Highsmith, VR; Buckley, TJ; et al. (1993) Gasoline-contaminated ground water as a source of
residential benzene exposure: a case study. J Exposure Anal Environ Epidemiol 4:183-195.
Lindstrom, AB; Yeowell-O'Connell, K; Waidyanatha, S; et al. (1997) Measurement of benzene oxide in the blood
of rats following administration of benzene. Carcinogenesis 18:1637-1641.
London, SJ; Lehman, TA; Taylor, JA. (1997) Myeloperoxidase genetic polymorphism and long cancer risk. Cancer
Res 57:5001-5003.
Longacre, SL; Kocsis, JJ; Snyder, R. (1981) Influence of strain differences in mice on the metabolism and toxicity
of benzene. Toxicol Appl Pharmacol 60:397-409.
Lovern, MR; Turner, MJ; Meyer, M; et al. (1997) Identification of benzene oxide as a product of benzene
metabolism by mouse, rat and human liver microsomes. Carcinogenesis 18:1695-1700.
Low, LK; Meeks, JR; Norris, KJ; et al. (1989) Pharmacokinetics and metabolism of benzene in Zymbal gland and
other key target tissues after oral administration in rats. Environ Health Perspect 8:2:215-222.
Low, LK; Lambert, CD; Meeks, JR. (1995) Tissue-specific metabolism of benzene in Zymbal gland and other solid
tumor target tissues in rats. J Am Coll Toxicol 14:40-60.
Luke, CA; Tice, RR; Drew, RT. (1988a) The effect of exposure regimen and duration on benzene-induced bone-
marrow damage in mice. I. Sex comparison in DBA/2 mice. Mutat Res 203:251-271.
Luke, CA; Tice, RR; Drew, RT. (1988b) The effect of exposure regimen and duration on benzene-induced bone-
marrow damage in mice. II. Strain comparisons involving B6C3F1, C57BL/6 and DBA/2 male mice. Mutat Res
203:273-295.
Lund-Johansen, F; Bjerknes, R; Laerum, OD. (1990) Flow cytometric assay for the measurement of human bone
marrow phenotype, function and cell cycle. Cytometry 11:610-616.
Lutz, WK; Schlatter, CH. (1977) Mechanism of the carcinogenic action of benzene: Irreversible binding to rat liver
DNA. ChemBiol Interact 18:241-245.
155
-------
MacEachern, L; Snyder, R; Laskin, DL. (1992) Alterations in the morphology and functional activity of bone
marrow phagocytes following benzene treatment of mice. Toxicol Appl Pharmacol 117:147-154.
Mackay, D; Leinonen, PJ. (1975) Rate of evaporation of low-solubility contaminants from water bodies to
atmosphere. Environ Sci Technol 9:1178-1180.
Maibach, HI; Anjo, DM. (1981) Percutaneous penetration of benzene and benzene contained in solvents used in the
rubber industry. Arch Environ Health 36:256-260.
Major, J; Kemeny, G; Tompa, A. (1992) Genotoxic effects of occupational exposure in the peripheral blood
lymphocytes of pesticide preparing workers in Hungary. Acta Med Hung 49(l-2):79-90.
Maltoni, C; Conti, B; Cotti, G. (1983) Benzene: a multipotential carcinogen: results of long-term bioassays
performed at the Bologna Institute of Oncology. Am J Ind Med 4:589-630.
Maltoni, C; Conti, B; Cotti, G; et al. (1985) Experimental studies on benzene carcinogenicity at the Bologna
Institute of Oncology: current results and ongoing research. Am J Ind Med 7:415-446.
Mathews, JM; Etheridge, AS; Mathews, HB. (1998) Dose-dependent metabolism of benzene in hamsters, rats, and
mice. Toxicol Sci 44:14-21.
Mattie, DR; Bates, Jr., GD; Jepson, GW; et al. (1994) Determination of skin:air partition coefficients for volatile
chemical: experimental method and applications. Fundam Appl Toxicol 22:51-57.
Mazzullo, M; Bartoli, S; Bonora, B; et al. (1989) Benzene adducts with rat nucleic acids and proteins: dose-
response relationship after treatment in vivo. Environ Health Perspect 82:259-266.
McDonald, TA; Yeowell-O'Connell, K; Rappaport, SM. (1994) Comparison of protein adducts of benzene oxide
and benzoquinone in the blood and bone marrow of rats and mice exposed to [14C/13C6] benzene. Cancer Res
54:4907-4914.
McDougal, JN; Jepson, GW; Clewell III, HJ; et al. (1990) Dermal absorption of organic chemical vapors in rats
and humans. Fundam Appl Toxicol 14:299-308.
McKinney, PA; Alexandria, FE; Cartwright, RA; et al. (1991) Parental occupations of children with leukaemia in
West Cumbria, North Humberside, and Gateshead. Br Med J 302(6778):681-687.
McMahon, TF; Birnbaum, LS. (1991) Age-related changes in disposition and metabolism of benzene in male
C57BL/6N mice. Drug Metab Dispos 19:1052-1057.
McMahon, TF; Medinsky, MA; Birnbaum, LS. (1994) Age-related changes in benzene disposition in male
C57BL/6N mice described by a physiologically based pharmacokinetic model. Toxicol Lett 74:241-253.
Medinsky, MA. (1995) The application of physiologically-based pharmacokinetic/ pharmacodynamic modeling to
understanding the mechanism of action of hazardous substances. Toxicol Lett 79:185-191.
Medinsky, MA; Sabourin, PJ; Lucier, G; et al. (1989a) A physiological model for simulation of benzene
metabolism by rats and mice. Toxicol Appl Pharmacol 99:193-206.
Medinsky, MA; Sabourin, PJ; Henderson, RF; et al. (1989b) Differences in the pathways for metabolism of
benzene in rats and mice simulated by a physiological model. Environ Health Perspect 82:43-49.
Medinsky, MA; Sabourin, PJ; Lucier, G; et al. (1989c) A toxicokinetic model for simulation of benzene metabolic.
ExpPathol 37:150-154.
Medinsky, MA; Kenyon, EM; Schlosser, PM. (1995) Benzene: A case study in parent chemical and metabolite
interactions. Toxicology 105:225-233.
156
-------
Medinsky, MA; Kenyon, EM; Seaton, MJ; et al. (1996) Mechanistic considerations in benzene physiological model
development. Environ Health Perspect 104 (suppl 6): 1399-1404.
Melikian, AA; Prahalad, AK; Seeker-Waker, RH. (1994) Comparison of the levels of the urinary benzene
metabolite trans,trans,-muconic acid in smokers and nonsmokers, and the effects of pregnancy. Cancer Epidemiol
3:239-244.
Morgan, DL; Cooper, SW; Carlock DL; et al. (1991) Dermal absorption of neat and aqueous volatile organic
chemicals in the Fischer 344 rat. Environ Res 55:51-63.
Mukhametova, DVI; Vozovaya, MA. (1972) Reproductive power and the incidence of gynecological disorders in
female workers exposed to the combined effect of benzene and chlorinated hydrocarbons. Gig Tr Prof Zabol 16:6-9
(Russian).
Murray, FJ; John, JA; Rampy, LW; et al. (1979) Embryotoxicity of inhaled benzene in mice and rabbits. Am Ind
HygAssoc 140:993-998.
Nakajima, T; Okuyama, S; Yonekura, I; et al. (1985) Effects of ethanol and phenobarbital administration on the
metabolism and toxicity of benzene. Chem Biol Interact 55:23-38.
Nakajima, T; Okino, T; Sato, A. (1987) Kinetic studies on benzene metabolism in rat liver - possible presence of
three forms of benzene metabolizing enzymes in the liver. Biochem Pharmacol 36:2799-2804.
National Research Council. (1983) Risk assessment in the federal government: Managing the process. Committee
on the Institutional Means for Assessment of Risks to Public Health, Commission on Life Sciences, NRC.
Washington, DC: National Academy Press.
NCI (National Cancer Institute). (1980) Bioassay of phenol for possible carcinogenicity. Technical Report Series
NCI-CG-TR-203. Bethesda, MD.
NFPA (National Fire Protection Association). (1994) Benzene. Fire protection guide to hazardous materials, llth
edition. National Fire Protection Association. Quincy, MA.
Nerland, DE; Pierce, WM. (1990) Identification of A^-acetyl-5'-(2,5-dihydroxyphenyl)-/-cysteine as a urinary
metabolite of benzene, phenol and hydroquinone. Drug Metab Dispos 18(6):958-961.
Neun, DJ; Perm, A; Snyder, CA. (1992) Evidence for strain-specific differences in benzene toxicity as a function of
host target cell susceptibility. Arch Toxicol 66:11-17.
Neun, DJ; Perm, A; Snyder, CA. (1994) Erythroid progenitor cells that survive benzene exposure exhibit greater
resistance to the toxic benzene metabolites benzoquinone and hydroquinone. Arch Toxicol 68:535-540.
Niculescu, R; Bradford, HN; Colman, RW; et al. (1995) Inhibition of the conversion of pre-interleukins-lcc and ip
to mature cytokines by ^-benzoquinone, ametabolite of benzene. ChemBiol Interact 98:211-222.
Niculescu, R; Renz, RF; Kalf, GF. (1996) Benzene-induced bone marrow cell depression caused by inhibition of
the conversion of pre-interleukins-lcc and -ip to active cytokines by hydroquinone, a biological reactive metabolite
of benzene. In: Biological reactive intermediates V. Snyder, R, ed. New York: Plenum Press.
Nomiyama, K; Nomiyama, H. (1974) Respiratory retention, uptake, and excretion of organic solvents in man:
Benzene, toluene, n-hexane, trichloroethylene, acetone, ethyl acetate and ethyl alcohol. Int Arch Arbeitsmed 32:75-
83.
Norpoth, K; Stucker, W; Krewet, E; et al. (1988) Biomonitorintg of benzene exposure by trace analysis of
phenylguanine. Int Arch Occup Enviorn Health 60:163-168.
157
-------
NTP (National Toxicology Program). (1994) Seventh annual report on carcinogens: 1994 Summary: National
Toxicology Program: Benzene. Public Health Service, U.S. Department of Health and Human Services,
Washington, DC.
NTP. (1986) Toxicology and carcinogenesis studies of benzene (CAS No. 71-43-2) in F344/N rats and B6C3F1
mice (gavage studies). National Toxicology Program. Research Triangle Park, NC.
OEHHA (Office of Environmental Health Hazard Assessment). (1997) Hazard identification of the developmental
and reproductive toxic effects of benzene. California Environmental Protection Agency. Sacramento, California.
Draft, September 1997.
OSHA (Occupational Safety and Health Administration). (1987) Benzene. U.S. Department of Labor,
Occupational Safety and Health Administration. Code of Federal Regulations. 29 CFR 1910.1028.
Pagnotto, LD; Elkins, HB; Brugsch, HG; et al. (1961) Industrial benzene exposure from petroleum naphtha. I.
Rubber coating industry. Am Ind Hyg Assoc J 22:417-421.
Pan, J; Hong JY; Yang, CS. (1992) Post-transcriptional regulation of mouse renal cytochrome P450 2E1 by
testosterone. Arch Biochem Biophys 299:110-115.
Parke, DV. (1989) Introduction: session on metabolism. Environ Health Perspect 82:7-8.
Parke, DV; Williams, RT. (1953) Studies in detoxication. 49. The metabolism of benzene containing ["CJ.
Biochem 154:231-238.
Paustenbach, DJ; Price, PS; Ollison, W; et al. (1992) Reevaluation of benzene exposure for the Pliofilm
(rubberworker) cohort (1936-1976). J Toxicol Environ Health 36:177-231.
Pekari, K; Vainiotalo, S; Heikkila, P; et al. (1992) Biological monitoring of occupational exposure to low levels of
benzene. Scand J Work Environ Health 18:317-322.
Pfeifer, RW; Irons, RD. (1983) Alteration of lymphocyte function by quinones through sulfhydryl-dependent
disruption of microtubule assembly. Int J Immunopharmacol 5:463-470.
Plappert, U; Barthel, E; Raddatz, K; et al. (1994) Early effects of benzene exposure in mice. Hematological versus
genotoxic effects. Arch Toxicol 68:284-290.
Popp, W; Rauscher, D; Muller, G; et al. (1994) Concentrations of benzene in blood and s-phenylmercapturic and
t,t-muconic acid in urine in car mechanics. Int Arch Occup Environ Health 66:1-6.
Pushkina, NN; Gofmekler, VA; Klevtsova, GN. (1968) Changes in content of ascorbic acid and nucleic acids
produced by benzene and formaldehyde. Bull Exp Biol Med 66:51-53.
Ramsey, JC; Anderson, ME. (1984) A physiologically based description of the inhalation pharmacokinetics of
styrene in rats and humans. Toxicol Appl Pharmacol 73:159-175.
Rao, NR; Snyder, R. (1995) Oxidative modifications produced in HL-60 cells on exposure to benzene metabolites.
J Appl Toxicol 15:403-409.
Rao, GS; Siddiqui, SM; Pandya, KP; et al. (1988) Relative toxicity of metabolites of benzene in mice. Vet Human
Toxicol 30:517-520.
Renz, JF; Kalf, GF. (1991) Role forinterleukin-1 (IL-1) in benzene-induced hematotoxicity: inhibition of
conversion of pre-IL-1 alpha to mature cytokine in murine macrophages by hydroquinone and prevention of benzene-
induced hematotoxicity in mice by IL-lalpha. Blood 78:938-944.
158
-------
Rickert, DE; Baker, TS; Bus, JS; et al. (1979) Benzene disposition in the rat after exposure by inhalation. Toxicol
Appl Pharmacol 49:417-423.
Rinsky, RA; Young, RJ; Smith, AB; et al. (1981) Leukemia in benzene workers. Am J Ind Med 2:217-245.
Rinsky, RA; Smith, AB; Hornung, R; et al. (1987) Benzene and leukemia: an epidemiological risk assessment. N
EngJMed 316:1044-1050.
Robinson, SN; Shah, R; Wong, BA; et al. (1997) Immunotoxicological effects of benzene inhalation in male
Sprague-Dawley rats. Toxicology 119:227-237.
Rosenthal, GJ; Snyder, CA. (1985) Modulation of the immune response to Listeria monocytogenes by benzene
inhalation. Toxicol Appl Pharmacol 80:502-510.
Ross, D. (1996) Metabolic basis of benzene to xicity. EurJHaematol 57:111-118.
Ross, D. (2000) The role of metabolism and specific metabolites in benzene-induced toxicity: evidence and issues.
J. Toxicol and Environ Health 61,5-6:357-372.
Ross, D; Seigel, D; Schattenberg, DG; et al. (1996) Cell-specific activation and detoxification of benzene
metabolites in mouse and human bone marrow: identification of target cells and a potential role for modulation of
apoptosis in benzene toxicity. Environ Health Perspect 104 (suppl 6): 1177-1182.
Rossi, AM; Guarnieri, C; Rovesti, S; et al (1999) Genetic polymorphisms influence variability in benzene
metabolism in humans. Pharmacogenetics 9(4):445-451.
Rosvold, EA; Mcglynn, KA; Lustbader, ED; et al. (1995) Identification of an NAD(P)H:quinone oxidoreductase
polymorphism and its association with lung cancer and smoking. Pharmacogenetics 5:199-206.
Rothman, N; Haas, R; Hayes RB; et al. (1995) Benzene induces recombination-type mutations at the glycophorin
A locus in the bone marrow of exposed workers. Proc Natl Acad Sci, USA 92:4069-4073.
Rothman, N; Li, GL; Dosemeci, M; et al. (1996a) Hematotoxicity among Chinese workers heavily exposed to
benzene. Am J Ind Med 29:236-246.
Rothman, N; Smyth, MT; Hayes, RB; et al. (1996b) An epidemiological study of early biologic effects of benzene
in Chinese workers. Environ Health Perspect 104 (suppl 6): 1365-1370.
Rothman, N; Smith, MT; Hayes, RB; et al. (1997) Benzene poisoning, a risk factor for hematological malignancy,
is associated with the NQO1609C forward T mutation and rapid fractional excretion of chlorzoxazone. Cancer Res
57(14):2839-2842.
Rozen, MG; Snyder, CA; Albert, RE. (1984) Depressions in B- and T-lymphocyte mitogen induced blastogenesis
in mice exposed to low concentrations of benzene. Toxicol Lett 20:343-349.
Rushmore, T; Kalf, G; Snyder, R. (1984) Covalent binding of benzene and its metabolites to DNA in rabbit bone
marrow mitochondria in vitro. Chem Biol Interact 49:133-154.
Sabourin, PJ; Chen, BT; Lucier, G; et al. (1987) Effect of dose on the absorption and excretion of [C14]benzene
administered orally or by inhalation in rats and mice. Toxicol Appl Pharmacol 87:325-336.
Sabourin, PJ; Bechtold, WE; Birnbaum, LS; et al. (1988a) Differences in the metabolism and disposition of inhaled
[3H]benzene by F344/N rats and B6C3F1 mice. Toxicol Appl Pharmacol 94:128-140.
Sabourin, PJ; Bechtold, WE; Henderson, RF. (1988b) A high pressure liguid chromatographic method for the
separation and quantitation of water-soluble radiolabeled benzene metabolites. Anal Biochem 170:316-327.
159
-------
Sabourin, PJ; Bechtold, WE; Griffith, W; et al. (1989) Effect of exposure concentration, exposure rate, and route of
administration on metabolism of benzene by F344 rats and B6C3F! mice. Toxicol Appl Pharmacol 99:421-444.
Sabourin, PJ; Sun, JD; MacGregor, JT; et al. (1990) Effect of repeated benzene inhalation exposures on benzene
metabolism, binding to hemoglobin, and induction of micronuclei. Toxicol Appl Pharmacol 103:452-462.
Sabourin, PJ; Muggenburg, BA; Couch, RC; et al. (1992) Metabolism of 14C benzene by cynomolgus monkeys and
chimpanzees. Toxicol Appl Pharmacol 114:277-284.
Saito, FU; Kocsis JJ; Snyder, R. (1973) Effect of benzene on hepatic drug metabolism and ultrastructure. Toxicol
Appl Pharmacol 26:209-217.
Sammett, D; Lee, EW; Koscis, JJ; et al. (1979) Partial hepatectomy reduces both the metabolism and toxicity of
benzene. J Toxicol Environ Health 5:785-792.
Sandmeyer, EE. (1981) Aromatic hydrocarbons: benzene. Patty's Ind Hyg Toxicol 3:3253-3283.
Sasiadek, M; Jagielski, J; Smolik, R. (1989) Localization of breakpoints in the karyotypes of workers professionally
exposed to benzene. Mutat Res 224:235-240.
Sato, A. (1988) Toxicokinetics of benzene, toluene and xylenes. In: environmental carcinogens methods of
analysis and exposure measurement. Vol. 10, Benzenes and Alkylated Benzenes. Fishbein, L; O'Neil, IK, eds.,
Lyon, France, IARC Scientific Publ. No. 85, pp. 47-64.
Sato, A; Nakajima, T. (1979) Dose-dependent metabolic interaction between benzene and toluene in vivo and in
vitro. Toxicol Appl Pharmacol 48:249-256.
Sato, A., Nakajima, T; Fujiwara, Y; et al. (1975) Kinetic studies on sex difference in susceptibility to chronic
benzene intoxication - with special reference to body fat content. Br J Ind Med 32:321-328.
Savilahti, M. (1956) Over 100 cases of benzene poisoning in a shoe factory. Arch Gewerbepathol Gewerbehyg
15:147-157.
Savitz, DA; Whelan, EA; Kleckner, RC. (1989) Effect of parents' occupational exposures on risk of still birth,
preterm delivery, and small-for-gestational-age infants. Am J Epidemiol 129:1201-1218.
Sawahata, T; Neal, RA. (1983) Biotransformation of phenol to hydroquinone and catechol by rat liver microsomes.
Mol Pharmacol 23:453-460.
Schlosser, MJ; Kalf, GF. (1989) Metabolic activation of hydroquinone by macrophage peroxidase. ChemBiol
Interact 72:191-207.
Schlosser, PM; Bond, JA; Medinsky, MA. (1993) Benzene and phenol metabolism by mouse and rat liver
microsomes. Carcinogenesis 14:2477-2486.
Schnier, GC; Laethem, CL; Koop, DR. (1989) Identification and induction of cytochromes P450, P450IIE1 and
P450IA1 in rabbit bone marrow. J Pharmacol Exp Ther 251:790-796.
Schrenk, HH; Yant, WP; Pearce, SJ; et al. (1941) Absorption, distribution, and elimination of benzene by body
tissues and fluids of dogs exposed to benzene vapor. J Ind Hyg Toxicol 23:20-34.
Seaton, MJ; Schlosser, PM; Bond JA; et al. (1994) Benzene metabolism by human liver microsomes in relation to
cytochrome P450 2E1 activity. Carcinogenesis 15:1799-1806.
Seaton, MJ; Schlosser, PM; Medinsky, MA. (1995) In vitro conjugation of benzene metabolites by human liver:
potential influence of interindividual variability on benzene toxicity. Carcinogenesis 16:1519-1527.
160
-------
Seidel, HJ; Barthel, E; Zinser, D. (1989) The hematopoietic stem cell compartments in mice during and after long-
term inhalation of three doses of benzene. Exp Hematol 17:300-303.
Seidel, HJ; Bader, R; Weber, L; et al. (1990) The influence of ethanol on the stem cell toxicity of benzene in mice.
Toxicol Appl Pharmacol 105:13-18.
Seidenberg, JM; Anderson, DG; Becker, RA. (1986) Validation of an in vivo developmental toxicity screen in the
mouse. Teratog Care Mutagen 5:361-374.
Shaw, G; Lavey, R; Jackson, R; et al. (1984) Association of childhood leukemia with maternal age, birth order, and
paternal occupation. A case-control study. Am J Epidemiol 119(5):788-795.
Sherwood, R.J. (1988) Pharmacokinetics of benzene in a human after exposure at about the permissible limit. Ann
NY Acad Sci 534:635-647.
Shu, XO; Gao, YT; Brinton, LA; et al. (1988) A population based case-control study of childhood leukemia in
Shanghai. Cancer 62:635-644.
Singh, V; Ahmad, S; Rao, GS. (1994) Prooxicant and antioxidant properties of iron-hydroquinone and iron-1,2,4-
benzenetriol complex. Implications for benzene toxicity. Toxicology 89(l):25-33.
Siou, G; Conan, L; Haitem, M. (1981) Evaluation of the clastogenic action of benzene by oral administration with 2
cytogenetic techniques in mouse and Chinese hamster. Mutat Res 90:273-278.
Skowronski, GA; Turkall, RM; Abdell-Rahman, MS. (1988) Soil adsorption alters bioavailability of benzene in
dermally exposed male rats. Am Ind Hyg Assoc J 49:506-511.
Smart, RC; Zannoni, VG. (1984) DT-Diaphorase and peroxidase influence the covalent binding of the metabolites
of phenol, the major metabolite of benzene. Mol Pharmacol 26:105-111.
Smart, RC; Zannoni, VG. (1985) Effect of ascorbate on covalent binding of benzene and phenol metabolites to
isolated tissue preparations. Toxicol Appl Pharmacol 77:334-343.
Smith, MT; Fanning, EW. (1997) Report on the workshop entitled: "Modeling chemically induced
leukemia-implications for benzene risk assessment." Leuk Res 21(5):361-374.
Smith, MT; Zhang, L. (1998) Biomarkers of leukemia risk: benzene as a model. Environ Health Perspect 106
(suppl 4):937-946.
Smith, MT; Yager, JW; Steinmetz, K; et al. (1989) Peroxidase-dependent metabolism of benzene's phenolic
metabolites and its potential role in benzene toxicity and carcinogenicity. Environ Health Perspect 82:23-29.
Smith, MT; Zhang, L; Wang, J; et al. (1998) Increased translocations and aneusomy in chromosomes 8 and 21
among workers exposed to benzene. Cancer Res 58:2176-2181.
Snyder, CA. (1987) Benzene. Toxic Metab Ind Solv 2:3-37.
Snyder, R; Hedli, CC. (1996) An overview of benzene metabolism. Environ Health Perspect 104 (suppl 6): 1165-
1171.
Snyder, R; Kalf, GF. (1994) A perspective on benzene leukemogenesis. Crit Rev Toxicol 24(3): 177-209.
Snyder R; Kocsis, JJ. (1975) Current concepts of benzene toxicity. Crit Rev Toxicol 3:265-288.
Snyder, CA, Goldstein, BD; Sellakumar, A. (1978) Hematotoxicity of inhaled benzene to Sprague-Dawley rats and
AKR mice at 300 ppm. J Toxicol Environ Health 4:605-618.
161
-------
Snyder, CA; Goldstein, BD; Sellakumar, AR; et al. (1980) The inhalation toxicity of benzene: Incidence of
hematopoietic neoplasms and hematoxicity in AKR/J and C57BL/6J mice. Toxicol Appl Pharmacol 54:323-331.
Snyder, CA; Erlichman, MN; Laskin, S; et al. (1981) The pharmacokinetics of repetitive benzene exposures at 300
and 100 ppm in AKR mice and Sprague-Dawley rats. Toxicol Appl Pharmacol 57:164-171.
Snyder, CA; Goldstein, BD; Sellakumar, AR; et al. (1982) Toxicity of chronic benzene inhalation: CD-I mice
exposed to 300 ppm. Bull Environ Contam Toxicol 29:385-391.
Snyder, CA; Goldstein, BD; Sellakumar, AR; et al. (1984) Evidence for hematotoxicity and tumorigenesis in rats
exposed to 100 ppm benzene. Am J Ind Med 5:429-434.
Snyder R; Jowa, L; Witz, G; et al. (1987) Formation of reactive metabolites from benzene. Arch Toxicol 60:61-64.
Snyder, CA; Goldstein, BD; James, DJ; et al. (1988) The carcinogenicity of discontinuous inhaled benzene
exposures in CD-I and C57BL/6 mice. Arch Toxicol 62:331-335.
Snyder, R; Dimitriadis, E; Guy, R; et al. (1989) Studies on the mechanism of benzene toxicity. Environ Health
Perspect 82:31-35.
Snyder, R; Chepiga, T; Yang, CS. (1993a) Benzene metabolism by reconstituted cytochromes P450, 2B1, and 2E1
and its modulation by cytochrome b5, microsomal epoxide hydrolase, and glutathione transferases: Evidence for an
important role of microsomal epoxide hydrolase in the formation of hydroquinone. Toxicol Appl Pharmacol
122:172-181.
Snyder, R; Witz, G; Goldstein, BD. (1993b) The toxicology of benzene. Environ Health Perspect 100:293-306.
Spano, M; Pacchierotti, F; Uccelli, R; et al. (1989) Cytotoxic effects of benzene on mouse germ cells determined by
flow cytometry. J Toxicol Environ Health 26:361-372.
Spear, RC; Bois, FY. (1994) Parameter variability and the interpretation of physiologically based pharmacokinetic
modeling results. Environ Health Perspect 102:61-66.
Spear, RC; Bois, FY; Woodruff, T; et al. (1991) Modeling benzene pharmacokinetics across three sets of animal
data: Parametric sensitivity and risk implications. Risk Anal 11:641-654.
Srbova, J; Teisniger, J; Skramovsky, S. (1950) Absorption and elimination of inhaled benzene in man. Arch Ind
HygOccupMed2:l-8.
Subrahmanyam, W; Doane-Setzer, P; Steinmetz, K; et al. (1990) Phenol-induced stimulation of hydroquinone
bioactivation in mouse bone marrow in vivo: possible implications in benzene myelotoxicity. Toxicology 62:107-
116.
Subrahmanyam, W; Kolanchana, P; Smith, MT. (1991) Hydroxylation of phenol to hydroquinone catalyzed by a
human myeloperoxidase-superoxide complex: possible implications in benzene myelotoxicity. Free Radic Res
Comm 15:285-296.
Susten, AS; Dames, BL; Burg, JR; et al. (1985) Percutaneous penetration of benzene in hairless mice: An estimate
of dermal absorption during tire-building operations. Am J Ind Med 7:323-335.
Tatrai, E; Ungvary, GY; Hudak, A; et al. (1980) Concentration dependence of the embryotoxic effects of benzene
inhalation in CFY rats. J Hyg Epidemiol Microbiol Immunol 24:363-371.
Tennant, RW; French, JE; Spalding, JW. (1995) Identifying chemical carcinogens and assessing potential risk in
short-term bioassays using transgenic mouse models. Environ Health Perspect 103(10):942-950.
Thienes, H; Haley, TJ. (1972) Clinical toxicology, 5th ed. Philadelphia: Lea & Febiger, pp. 124-127.
162
-------
Tice, RR; Luke, CA; Drew, RT. (1989) Effect of exposure route, regimen, and duration on benzene-induced
genotoxic and cytotoxic bone marrow damage in mice. Environ Health Perspect 82:65-74.
Toft, K; Olofsson, T; Tunek, A; et al. (1982) Toxic effects on mouse bone marrow caused by inhalation of benzene.
Arch Toxicol 51:295-302.
Tompa, A; Major, J; Jakob, MG. (1994) Monitoring of benzene-exposed workers for genotoxic effects of benzene;
improved-working-condition-related decrease in the frequencies of chromosomal aberrations in peripheral blood
lymphocytes. MutatRes 304(2):159-165.
Townsend, JC; Ott, MG; Fishbeck, WA. (1978) Health exam findings among individuals occupationally exposed to
benzene. JOccupMed 20:543-548.
Traver, RD; Horikoshi, T; Danenberg, KD; et al. (1992) NAD(P)H:quinone oxidoredactase gene expression in
human colon carcinoma cells: Characterization of a mutation which modulates DT-diaphorase activity and
mitomycin sensitivity. Cancer Res 52:797-802.
Travis, CC; Quillen, JL; Arms, AD. (1990a) Pharmacokinetics of benzene. Toxicol Appl Phamacol 102:400-420.
Travis, CC; White, RK; Wards, RC. (1990b) Interspecies extrapolation of pharmacokinetics. J Theoret Biol
142:285-304.
Tsai, SP; Wen, CP; Weiss, NS; et al. (1983) Retrospective mortality and medical surveillance studies of workers in
benzene areas of refineries. J Occup Med 25:685-692.
Tsuruta, H. (1989) Skin absorption of organic solvent vapors in nude mice in vivo. Ind Health 27:37-47.
Tunek, A; Olofsson, T; Berlin, M. (1981) Toxic effects of benzene and benzene metabolites on granulopoietic stem
cells and bone marrow cellularity in mice. Toxicol Appl Pharmacol 59:149-156.
Tunek, A; Hogstedt, B; Olofsson, T. (1982) Mechanism of benzene toxicity. Effects of benzene and benzene
metabolites on bone marrow cellularity, number of granulopoietic stem cells and frequency of micronuclei in mice.
ChemBiol Interact 15:129-138.
Turkall, RM; Skowronski, G; Gerges, S; et al. (1988) Soil absorption alters kinetics and bioavailability of benzene
in orally exposed male rats. Arch Environ Contam Toxicol 17:159-164.
Ungvary, G. (1985) The possible contribution of industrial chemicals (organic solvents) to the incidence of
congenital defects caused by teratogenic drugs and consumer goods: An experimental study. Prog Clin Biol Res
1636:295-300.
Ungvary, G; Donath, T. (1984) Effect of benzene and its methyl-derivatives (toluene, p-xylene) on post ganglionic
noradrenergic nerves. Z Mikrosk-Anat Forsch 98:755-763.
Ungvary, G; Tatrai, E. (1985) On the embryotoxic effects of benzene and its alkyl derivatives in mice, rats, and
rabbits. Arch Toxicol 8:425-430.
U.S. EPA (U.S. Environmental Protection Agency). (1985) final draft for drinking water criteria document on
benzene. Washington, DC: Office of Drinking Water, U.S. EPA. PB86-118122.
U.S. EPA. (1986a) Guidelines for the health risk assessment of chemical mixtures. Federal Register
51(185):34014-34025.
U.S. EPA. (1986b) Guidelines for mutagenicity risk assessment. Federal Register 51(185):34006-34012.
U.S. EPA. (1987) Risk assessment guidelines of 1986. Prepared by Office of Health and Environmental
Assessment, Office of Research and Development, Environmental Protection Agency, Washington, DC.
163
-------
EPA/600/8-87/045, August 1987.
U.S. EPA. (1988) Recommendations for and documentation of biological values for use in risk assessment.
EPA/600/6-87/008, NTIS PB88-179874/AS, February 1988.
U.S. EPA. (1989) Risk assessment guidance for superfund. Volume 1. Human health evaluation manual. Part A.
Washington, DC: U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response; EPA
report no. EPA/540/1-89/002.
U.S. EPA. (1991) Guidelines for developmental toxicity risk assessment, dated December 5, 1991. Federal
Register 56( 234):63798-63826.
U.S. EPA. (1994) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry, EPA/600/8-90/066F, October 1994.
U.S. EPA. (1995) Use of the benchmark dose approach in health risk assessment. EPA/630/R-94/007.
U.S. EPA. (1996a) Proposed guidelines for carcinogen risk assessment. FederalRegister61(79):17960-18011.
U.S. EPA. (1996b) Guidelines for reproductive toxicity risk assessment. Federal Register 61(212):56274-56322.
U.S. EPA. (1998a) Carcinogenic effects of benzene: An update. Prepared by Office of Research and Development,
Washington, DC. EPA/600/P-97/001F.
U.S. EPA. (1998b) Guidelines for neurotoxicity risk assessment. Federal Register 63(93):26926-26954.
U.S. EPA. (1999a) Extrapolation of the Benzene Inhalation Unit Risk Estimate to the Oral Route of Exposure.
National Center for Environmental Assessment, Office of Research and Development, Washington, DC. NCEA-W-
0517.
U.S. EPA. (1999b) Draft revised guidelines for carcinogen risk assessment. Federal Register 66(230):59593-59594.
U.S. EPA. (2000a) Integrated Risk Information System (IRIS). Substance file - benzene. CASRN 71-43-2. Online.
National Center for Environmental Assessment, Cincinnati, OH.
U.S. EPA. (2000b) Benchmark dose technical guidance document (external review draft). EPA/630/R-00/001.
Vacha, J; Znojil, V; Seidel, HJ; et al. (1990) Ferrokinetics and erythropoiesis in mice after long-term inhalation of
benzene. Blut 60:41-47.
Valentine, RL; Lee, SS; Seaton, MJ; et al. (1996) Reduction of benzene metabolism and toxicity in mice that lack
CYP2E1 expression. Toxicol Appl Pharmacol 141:205-213.
Vara, P; Kinnunen, O. (1946) Benzene poisoning as a gynecological problem. ActaObstet Gynecol Scand 26:433-
452. (OTS 8EHQ-0279-0244)
Von Oettingen, WF. (1940) Toxicity and potential dangers of aliphatic and aromatic hydrocarbons: A critical
review of the literature. Washington, DC: Division of Industrial Hygiene, National Institutes of Health, U.S. Public
Health Service, Washington, DC. Public Health Bull 255:66-97, 125-133.
Ward, CO; Kuna, RA; Snyder, NK; et al. (1985) Subchronic inhalation toxicity of benzene in rats and mice. Am J
IndMed 7:457-473.
Ward, JB, Jr; Ammenheuser, MM; Ramanujam, VMS; et al. (1992) The mutagenic effects of low level sub-acute
inhalation exposure to benzene in CD-I mice. Mutat Res 268(l):49-57.
164
-------
Ward, E; Hornung, R; Morris, J; et al. (1996) Risk of low red or white blood cell count related to estimated benzene
exposure in a rubberworker cohort (1940-1975). Am J Ind Med 29:247-257.
Watanabe, KH; Bois, FY. (1996) Interspecies extrapolation of physiological pharmacokinetic parameter
distributions. Risk Anal 16:741-754.
Wells, MS; Nerland, DE. (1991) Hematotoxicity and concentration-dependent conjugation of phenol in mice
following inhalation exposure to benzene. Toxicol Lett 56:159-166.
White, KL, Jr; Lysy, HH; Munson, JA; et al. (1984) Immunosuppression of B6C3F1 female mice following
subchronic exposure to benzene from drinking water. TSCA 8E Submission. OTS fiche #OTS0536214.
Wilson, RH. (1942) Benzene poisoning in industry. JLabClinMed 27:1517.
Witz, G; G.S. Rao, GS; Goldstein, BD. (1985) Short-term toxicity of /raws-./raws-muconaldehyde. Toxicol Appl
Pharmacol 80:511-516.
Witz, G; Gad, SC; Tice RR; et al. (1990a) Genetic toxicity of the benzene metabolite /ra«5-/ra«5-muconaldehyde in
mammalian and bacterial cells. Mutat Res 240:295-306.
Witz, G; Kirley, TA; Maniara, WM; et al. (1990b) The metabolism of benzene to muconic acid, a potential
biological marker of benzene exposure. Biol React Inter IV 283:613-618.
Witz, G; Z. Zhang, Z; Goldstein, BD. (1996) Reactive ring-opened aldehyde metabolites in benzene
hematotoxicity. Environ Health Perspect 104 (suppl 6): 1195-1199.
Wolf, MA; Rowe, VK; McCollister, DD; et al. (1956) Toxicological studies of certain alkylated benzenes and
benzene. AMA Arch Ind Health 14:387-398.
Yardley-Jones, A; Anderson, D; Jenkinson, PC; et al. (1988) Genotoxic effects in peripheral blood and urine of
workers exposed to low level benzene. BrJIndMed 45:694-700.
Yardley-Jones, A; Anderson, D; Lovell, DP; et al. (1990) Analysis of chromosomal aberrations in workers exposed
to low level benzene. Br J Ind Med 47:48-51.
Yin, S; Li, G; Hu, Y; et al. (1987a) Symptoms and signs of workers exposed to benzene, toluene or the
combination. Ind Health 25:113-130.
Yin, SN; Li, Q; Tian, F; et al. (1987b) Occupational exposure to benzene in China. Br J Ind Med 44:192-195.
Yu, R; Weisel, CP. (1998) Measurement of benzene in human breath associated with an environmental exposure. J
Expos Anal Environ Epidemiol 6:261-277.
Zhang, L; Robertson, ML; Kolachana, P; et al. (1993) Benzene metabolite, 1,2,4-benzenetriol, induces micronuclei
and oxidative DNA damage in human lymphocytes and HL60 cells. Environ Mol Mutagen 21:339-348.
Zhang, Z; B.D. Goldstein, BD; Witz, G. (1995a) Iron-stimulated ring-opening of benzene in a mouse liver
microsomal system. Mechanistic studies and formation of a new metabolite. Biochem Pharmacol 50:1607-1617.
Zhang, Z; Schafer, F; Schoenfeld, H; et al. (1995b) Hematotoxicological studies of 6-hydmxy-trans-trans-2,4-
hexadienal in CD-I mice. Toxicol Appl Pharmacol 132:213-219.
165
------- |