ISSUE PAPER ON T?C
1>E hEALTH ASSESShCNT DOCUMENT
FOR
CARBON TETRACHLORIDE
Presented to:
The Science Advisory Board
Prepared by:
U.S. Environmental Protection Agency
Office of Research and Development
Environmental Criteria and Assessment Office
Cincinnati, OH 45268
Project Manager: Cynthia Sonich
November, 1982
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION 1-1
2. STATEMENT OF ISSUES 2-1
2.1. STRATOSPHERIC OZONE 2-1
2.2. SELECTION OF NOELs/LOAELs: RATIONALE 2-2
2.3. SELECTION OF CARCINOGENICITY STUDY USED IN THE
UNIT RISK ASSESSMENT: RATIONALE 2-3
2.4. USE OF MUTAGENICITY DATA AS AN INDICATOR OF
GENOTOXICITY 2-5
2.5. SELECTION OF ABSORPTION COEFFICIENTS: RATIONALE .... 2-6
2.6. CHANGES TO THE CARCINOGENICITY SECTIONS OF THE DOCUMENT. 2-8
2.6.1. Qualitative 2-8
2.6.2. Quantitative 2-8
3. RESPONSE TO PUBLIC COMMENTS 3-1
3.1. AIR PRODUCTS AND CHEMICALS, INC 3-1
3.1=1= Literature Survey 3-1
3.1.2. Carcinogenicity 3-2
3.1.3. Risk Assessment 3-3
3.1.4. Proposed Alternative Procedure 3-4
3.1.5. Other Comments 3-6
3.2. E.I. DU PONT DE NEMOURS AND COMPANY 3-9
3.2.1. Stratospheric Ozone 3-9
3.2.2. Toxicology 3-11
3.2.3. Carcinogenicity - Human 3-13
3.2.4. Unit Risk Estimate for Cancer 3-14
3.3. MUTAGENICITY COMMENTS BY AIR PRODUCTS AND DU PONT. ... 3-17
3.4. OFFICE OF DRINKING WATER, U.S. EPA 3-21
3.4.1. General Comments 3-21
3.4.2. Specific Comments 3-21
APPENDIX A. ADDITIONAL REFERENCES TO BE INCORPORATED OR CONSIDERED
FOR INCORPORATION INTO THE HEALTH ASSESSMENT DOCUMENT
FOR CARBON TETRACHLORIDE A-l
APPENDIX B. REVISION OF CHAPTER 10. MUTAGENICITY B-l
ii
-------
FOREWORD
The Health Assessment Document for Carbon Tetrachloride has been pre-
pared for the Office of Air Quality Planning and Standards by the Environ-
mental Criteria and Assessment Office, Cincinnati, Ohio. For additional
information or clarification contact Dr. Jerry F. Stara, Director, ECAO-
Cincinnati or Cynthia Sonich, Principle Author, ECAO-Cincinnati.
111
-------
1. INTRODUCTION
The purpose of this issue paper is two-fold. It is written in an effort
to focus upon the key scientific issues discussed within the document which
are subject to judgment mainly because of lack of data and to clarify any
judgments which involve policy decisions. Secondly, it will address the
public comments received to date. In this regard, this paper will evaluate
all of the major issues involved in assessing the potential health impact of
carbon tetrachloride. Hopefully, all questions regarding scientific con-
clusions and decisions will be clarified through the statement of issues and
response to public comments.
The Office of Health and Environmental Assessment, in consultation with
an Agency work group, has prepared this health assessment to serve as a
"source document" for EPA use. Originally the health assessment was devel-
oped for use by the Office of Air Quality Planning and Standards, however,
at the request of the Agency Work Group on Solvents, the assessment scope
was expanded to address multimedia aspects.
In the development of this assessment document, the scientific litera-
ture has been inventoried, key studies have been evaluated, and summaries
and conclusions have been prepared so that the chemical's toxicity and re-
lated characteristics are qualitatively identified. Observed effect levels
and dose-response relationships are discussed evaluating the potential toxi-
city of CC1.. Unit risk estimates for cancer are calculated to provide a
media-specific measure of carcinogenicity. This information can then be
placed in perspective with observed environmental levels.
1-1
-------
Writing of the document commenced June, 1981 and was completed December,
1981. A complete time table is provided in Table 1. Slight revisions were
made prior to its binding in March, 1982 and subsequent release to the pub-
lic. Since that time, the authors have remained current with regard to new
literature and have reviewed past literature searches in an effort to locate
new or previously overlooked references. Although additonal references were
uncovered, they do not change the final conclusions reached in the docu-
ment. Some of these findings will, however, be incorporated into the docu-
ment prior to final publication. Additional references were cited in the
public comments received and will also be considered. All such references
are listed in Appendix A.
It should be noted that this document has undergone extensive review
both within and outside of the Environmental Protection Agency prior to the
request for public comment and submission to the Science Advisory Board. A
list of reviewers appears in the front of the document. The reference to
the "authors" of the Health Assessment Document on Carbon Tetrachloride
(HAD-CCl^) throughout this issue paper includes those scientists partici-
pating in the peer review process. The participating offices are included
in Table 1.
1-2
-------
TABLE 1-1
Time Table for Completion of the Health
Assessment Document for Carbon Tetrachloride
Date
Activity
6/1/81
7/27/81
7/30/81
8/17/81
8/81
8/81
8/81
11/81
12/81
1/82
2/82
2/25/82
Begin literature search and writing
First Office of Health and Environmental
Assessment review begins
Toxic Substances Priority Committee
meeting
First Office of Health and Environmental
Assessment review completed
Environmental Criteria and Assessment
Office-Cincinnati internal and external
review
Office of Research and Development review
Second Internal Review Draft of the docu-
ment completed for the Toxic Substances
Priority Committee
Second and final Office of Health and
Environmental Assessment review completed
Final document completed incorporating
Office of Health and Environmental
Assessment, Office of Research and Devel-
opment, Toxic Substances Priority Commit-
tee, Environmental Criteria and Assess-
ment Office internal and Environmental
Criteria and Assessment Office external
comments
Office of Public Affairs approval and
clearance for printing as a draft for
public comment
Final changes made to the document
Final document submitted to Office of
Health and Environmental Assessment for
signoff
1-3
-------
TABLE 1-1 (Cont.)
Date Activity
3/82 Document printed by the Center for Envi-
ronmental Research Information
4/26/82 Federal Register notice of availability
and request for public comment
7/7/82 Public comment period ends
12/9/82 Science Advisory Board and public comment
review meeting
1-4
-------
2. STATEMENT OF ISSUES
2.1. STRATOSPHERIC OZONE
The main issues concerning the relationship between CCl^ emissions and
stratospheric ozone are:
1) does CC1. contribute to the depletion of the ozone layer?
2) is the ozone layer presently being depleted?
3) what are the actual health and environmental effects attributable to
depletion of the ozone layer?
Since none of these questions can be definitively answered, this topic be-
comes an issue.
In theory, laboratory evidence indicates that it is most probable that
CCl^ can contribute to ozone depletion. However, it is not known what the
net effect of COL is on stratospheric ozone nor of the combination of a
number of volatile pollutants such as methyl chloroform, CFC-11, CFC-12 and
CCl^ for example. Furthermore, the data are inadequate to conclusively
indicate that depletion of the ozone layer is presently occurring.
The authors of the HAD CCl^ recognize that ozone depletion cannot be
defined as a present problem. However, in keeping with the Agency's mandate
to protect the environment for present and future generations, the authors
are acting conservatively by recognizing a potential problem area if it is
allowed to go unchecked.
Thus, it is recommended that all factors contributing to ozone depletion
be examined. Once sufficient data are obtained, the issue can be re-exam-
ined ana the decision can be appropriately revised. However, it is pres-
ently maintained that it is not known if depletion of the ozone layer is
occurring but if it were to occur (as the laboratory data suggest), there
exists the potential for the following primary effects:
1) increased incidence of non-melanoma skin cancer in humans
2-1
-------
2) possible increased incidence of malignant melanoma skin cancer in
humans
3) possible crop yield reductions
4) possible killing of marine organisms.
This position requires a revision of the document as presently written.
The main reason for revision is the use of additional data published since
December, 1981 and reviewed more thoroughly since March, 1982. The appro-
priate sections of the document will be revised to indicate the theoretical
nature of the problem and its possible effects.
2.2. SELECTION OF NOELs/LOAELs: RATIONALE
The studies summarized in "Table 14-1. Dose-Related Toxic Effects of
Carbon Tetrachloride on Humans and Animals," includes all studies available
for review which provided information on dose, exposure duration, test ani-
mal and effect(s). The summary table: "Table 14=5. Reported No-Effect and
Low-Effect Levels for Toxicity of Carbon Tetrachloride" is extracted from
Table 14-1. Selections were made when possible, for each species and each
exposure duration (acute, sub-chronic and chronic) for those studies pro-
viding a NOEL (no-observed-effect level) or NOAEL (no-observed-adverseeffect
level) and corresponding LOAEL (lowest-observed-adverse-effect level). When
multiple studies were found on the same species for the same exposure dura-
tion, the study providing the highest NOEL (or NOAEL) or lowest LOAEL was
selected. For example, as noted in Table 14-1, Korsrud et al. (1972) re-
ported increased liver fat, increased liver weight and increased liver
enzyme levels in rats given a single dose of 40 mg/kg (CC1. per body
weight) by ingestion. Murphy and Malley (1969) reported similar effects:
increased liver weight and fatty infiltration and some liver necrosis in
rats given a single dose of 1600 mg/kg (CCl^ per body weight) by inges-
tion. The Korsrud et al. (1972) study is included in Table 14-5 rather than
2-2
-------
the Murphy and Malley (1969) study since effects were observed at a lower
oose, hence, it is the lowest-observed-adverse-effect level (LOAEL).
Only oral and inhalation exposures are considered because they are most
useful for risk assessment. One dermal exposure study was found in the
available literature providing the appropriate dose and exposure information
(Kronevi et al., 1979). However, there exists no approved methodology to
extrapolate from the dose and effects observed following a dermal exposure
in animals to corresponding dose and effects in humans. Studies adminis-
tering CCl^ via intraperitoneal, subcutaneous, intratracheal and other in-
jection routes though important in the qualitative determination of effects,
are not useful for quantitative hazard assessment.
2.3. SELECTION OF CARCINOGENICITY STUDY USED IN THE RISK ASSESSMENT:
RATIONALE
Six positive animal carcinogenicity studies were considered for use in
determining the unit risk estimate for the carcinogenicity of CCl^ because
they contained sufficient dose and exposure information on the test animals.
These were the mice studies by Edwards et al. (1942) and the NCI (1976), the
study on rats and mice by Weisburger (1977), the rat studies by Reuber and
Glover (1967) and the NCI (1976) and the hamster study by Delia Porta et al.
(1961).
Weisburger (1977) tested CCl^ by oral intubation in Osborne-Wendel
rats and B^C-F, mice of both sexes. However, the study was not used
since the author did not report the low dose but rather only the maximum
tolerated dose for each species and sex, and also did not report percent
survival. Delia Porta et al. (1961) orally administered CC1. to Syrian
golden hamsters. It is the only report found in the available literature of
the induction of tumors in hamsters by CCl^. It is not used for risk
assessment since the researchers provide no information on control animals
2-3
-------
and because of the small number of animals in the study (10 of each sex
originally, 5 of each sex surviving to the end of the experiment). Reuber
and Glover (1967) administered CCl^ by subcutaneous injections to inbred
Buffalo male and female rats. This study cannot be used for risk assessment
due to the inappropriate route of exposure since the mechanism of distribu-
tion is dissimilar to that of the oral route and subsequent distribution.
The NCI (1976) performed a bioassay of trichloroethylene in which CCl^ was
usea as the positive control. Osborne-Mendel rats of both sexes were admin-
istered CCl^ via oral gavage for 78 weeks. Although this study is used by
the National Research Council in determining the carcinogenic risk assess-
ment for CCl^ due to the dose levels used and the appropriate length of
the study, it is not recommended here to use this study because of the low
survival rates and the low absolute incidence of hepatocellular neoplasms
(< 535). This low incidence may possibly be the result of the resistance
of this rat strain to chlorinated hydrocarbons.
Fortunately, two carcinogenicity studies are useful for risk assessment
of CCi^. Edwards et al. (1942) administered CC1, via stomach tube to
inbred strain L mice. The mice were given 46 administrations of CCl^ over
a 4 month period (2 or 3 times per week) and found to be highly susceptible
to the induction of hepatomas. The NCI (1976) performed a bioassay on tri-
chloroethylene in which CC1. was used as the positive control. Male and
female B C,F, mice were given CCl^ by oral gavage for 78 weeks (5
times per week). Both sexes were found to be highly susceptible to hepato-
cellular carcinomas.
In the final selection of a study for use in the carcinogenic risk
assessment, all aspects of the studies under consideration are examined.
2-4
-------
Both the Edwards et al. and NCI studies orally administered CCl^. Both
used chronic exposures. However, the NCI study is chosen for use in the
risk assessment for CCl^ in this document due to the more appropriate
length of exposure (16 weeks vs. 78 weeks).
2.4. USE OF MUTAGENICITY DATA AS AN INDICATOR OF GENOTOXICITY
Included inthe original scope of this document was a discussion of the
hypothesis of some researchers, that CC1, causes cancer by a non-mutagenic
(possibly threshold*) mechanism although it has also been stated that avail-
able animal data do not conclusively define CC1. as being non-mutagenic.
Nor does the available literature conclusively define CC1. as genotoxic.
This has become a controversial issue primarily since the evidence of muta-
genicity is inconclusive with regard to a carcinogenic mechanism. However,
it was decided to review all evidence indicating that CCl^ may cause can-
cer through a mutagenic or non-mutagenic mechanism.
Under the non-mutation hypothesis, one could postulate that only those
levels of CC1. that induce organ toxicity are likely to pose a carcino-
genic risk to humans. However, with current data available to us as indi-
cated in the mutagenicity chapter, our present conclusion is that these
studies do not provide a conclusive judgment on the mutagenic potential of
carbon tetrachloride. Furthermore, some binding studies show that CCl^
metabolite(s) interact with macromolecules (e.g., DNA) in both in vitro and
in vivo systems which are suggestive that CC1. may be genotoxic. It
*The term "epigenetic" has been used to describe those chemicals that cause
cancer but do not cause genetic damage. The term has been used to also im-
ply a threshold. Some toxicologists believe that chemicals which only pro-
duce tumors at toxic dose levels are also epigenetic carcinogens. There is
considerable controversy over the use of the term epigenicity when des-
cribing carcinogenic mechanisms. Because of this confusion, the term "epi-
genetic" will either be deleted or better defined throughout the HAD-CC14.
2-5
-------
is noted that although several researchers maintain that toxic effects are
concurrent with liver tumors, it has not been established that tissue damage
is a necessary precursor to CCl^ carcinogenesis. Thus, it is presently
recommended that "... further investigation, particularly in regard to geno-
toxic potential, is indicated to elucidate the carcinogenic mechanism of
action for carbon tetrachloride."
Thus, the position in the document following thorough review all of the
literature to date is that "... the evidence is inadequate to conclude that
carbon tetrachloride is not genotoxic." A discussion of the specific
studies leading to this conclusion is provided in section 3.3. of this issue
paper.
2.5. SELECTION OF ABSORPTION COEFFICIENTS: RATIONALE
The decision to use a 40% absorption coefficient for exposure via inha-
lation is based on two studies and one report:
30.4% (McCollister et al., 1951)
57-65% (Lehmann and Schmidt-Kehl, 1936)
30% (Stokinger and Woodward, 1958).
McCollister et al. (1951) investigated the absorption of CCl^ by in-
halation using Rhesus monkeys. Three female monkeys inhaled 99.9% C-
labelled CC1. vapor at an average concentration of 290 mg/m3 for 139,
300 or 344 minutes. By calculating the difference between inhaled and ex-
haled air, the researchers concluded that the monkeys absorbed an average of
30.4% of the total amount of CC14.
Lehman and Schmidt-Kehl (1936) studied humans exposed to CCl^ via in-
halation. In separate experiments, individuals (number unknown) were en-
closed in a room of C£L vapors for varying amounts of time. The percent
CC1. absorbed was calculated from the difference between the concentration
of CC1. in the inhaled air and the amount found in the exhaled breath.
The reported range was 57-65%.
2-6
-------
Stokinger and Woodward (1958) report an "accurate" absorption factor of
30% via inhalation. However, their conclusion is not referenced.
Although done on monkeys, the McCollister et al. study is the best
available. The Lehmann and Schmidt-Kehl data are excellent as work done
over 50 years ago. However, in the context of what we know today, there are
serious drawbacks both in the assumptions that were made in deriving the
57-65% uptake data and in the experimental procedures that were used.
Specifically, but briefly, 1) it is highly unlikely that an equilibrium
between alveolar air and blood will be established within a 30 minute expo-
sure period; 2) the 10 m3 chamber used for these experiments is rather
small; 3) only two subjects volunteered for these exposures; 4) the analyti-
cal methods used lacked the sensitivity of today's methods. The coefficient
reported by Stokinger and Woodward, by their own admission, was meant for
use in the emergency situation. Although they do not reference their con-
clusion, it does reflect the results of the McCollister et al. study.
Generally, when human data are available and good, they are used in risk
assessment, in this instance, however, problems have been associated with
the human data. The monkey data are better. However, in an effort to be
most protective yet realistic, a compromise was made to arrive at an absorp-
tion coefficient for exposure via inhalation of 40% (the tradeoff being
slightly more protective than the good animal experiment would indicate yet
not as strong"as the poorer human data would indicate). The Stokinger and
Woodward report is not overlooked but is used as supporting evidence of the
results reported by McCollister et al. and may in fact be reflective of it.
2-7
-------
The decision to use a 100% absorption coefficient for exposure via in-
gestion is based primarily on the lack of sufficient data. One study and
one report are cited:
80% (Marchland et al., 1970)
50% (Stokinger and Woodward, 1958)
Again, Stokinger and Woodward do not reference their conclusion and no paper
reporting a 50% absorption coefficient via ingestion could be found in the
available literature. Studying the distribution of CCl^ in rats,
Marchland et al. (1970) noted that at least 80% of the ingested CCl^ was
eliminated within 10 hours via the lungs in control animals. This indicates
that at least 80% of the CCl^ was absorbed. Thus, in an effort to be most
protective and in light of the available data, a 100% absorption coefficient
for exposure via ingestion is used.
It is obvious from this discussion that the supporting data are scanty.
Nevertheless, such interpretations and subsequent conclusions must be made
and the resulting assumptions used to perform present day risk calcula-
tions. It is hopefully understood that as additional data become available,
these coefficients should be improved.
2.6. CHANGES TO THE CARCINOGENICITY SECTIONS OF THE DOCUMENT
2.6.1. Qualitative. The only change necessary to the qualitative section
of the document with reference to the carcinogenicity of CCl^ is to incor-
porate the IARC classification scheme as related to CCl^. This will be
incluaea as a part of the summary section of Chapter 11.
2.6.2. Quantitative. It is recognized that the Appendix to the docu-
ment: "Unit Risk Estimate for Cancer" should provide additional explanation
ana rationale for the method employed. As suggested by the Science Advisory
Board through their discussions at the September, 1982 meeting reviewing the
"Quantitative Carcinogen Risk Assessment," the Carcinogen Assessment Group
2-8
-------
has prepared such an expansion for its "Evaluation of the Carcinogenicity of
Acrylonitrile." The changes so made to the Health Assessment Document for
Acrylonitrile will be incorporated into the CCl^ document as appropriate.
The information appropriate for inclusion to the CCl^ appendix is con-
tained within the Quantitative Estimation section of the Evaluation of the
Carcinogenicity of Acrylonitrile (November 2, 1982). This includes a dis-
cussion of the procedures for the determination of unit risk and a descrip-
tion of the low-dose extrapolation model. Methods of data selection as well
as the actual calculation of the unit risk estimates and their interpreta-
tion will also be included. Alternate approaches and their consequences
will be discussed. In order to provide perspective, a section on relative
potency will be included. The purpose of this section will be to compare
the potency of CCl^ with other carcinogens including the other carcino-
genic solvents. A histogram representing the frequency distribution of the
potency indices of 54 suspect carcinogens will be included and accompanied
by the data in tabular form.
The quantitative evaluation of the Carcinogenicity of CCl^ has been
incorporated into an appendix at the end of the HAD-CCl^. It is recog-
nized that the Science Advisory Board has suggested, in their review of the
other solvent documents, to incorporate the appendix into the document. The
revised CCl^ appendix can be incorporated into the Carcinogenicity chapter
of the document. However, it is left as an appendix for review by the SAB.
2-9
-------
3. RESPONSE TO PUBLIC COMMENTS
3.1. AIR PRODUCTS AND CHEMICALS, INC.
3.1.1. Literature Survey.
"The bibliography appears to be reasonably complete. Some additional
recent references which could be of interest are listed below ..."
The additional references provided on environmental measurements, envi-
ronmental fate and biochemical data are appreciated and will be incorporated
or considered for incorporation into the document as appropriate. Articles
disputed as inaccurate or obsolete may not be used although it is within the
scope to provide as complete a document as possible.
"The discussion on natural sources is incomplete. Bacteria and algae in
the sea, volcanoes, and other natural sources are known to generate
halocarbons." References were provided.
Some of the references cited on natural sources will not be included.
The report by Inn et al. (1981) discusses gaseous constituents in the plume
from eruptions of Mount St. Helens. However, CCl^ is not mentioned in the
report. Increased concentrations of methyl chloride were measured, but this
is hardly justification for inferring that volcanic eruptions are a natural
source for CCl^.
The report by Singh et al. (1975) proposes the formation of atmospheric
CCl^ from tetrachloroethylene. While this information should and will be
incorporated into the document it will not be used to identify a natural
source. Instead, it will be used to identify CCl^ as the researchers
state, a "secondary anthropogenic pollutant" since the tetrachloroethylene
is primarily anthropogenic in origin.
The study by Jacquez and Mukerman (1979) addresses the production of
haloform precursors, i.e., trihalomethanes, by water pipe bacterial films.
Carbon tetrachloride is not normally considered a trihalomethane and is not
a by-product of the chlorination process. The authors state that "— the
3-1
-------
major trihalomethane specie resulting from the chlorination of bacterial
films was the known carcinogen chloroform." The report is, thus, inappro-
priate for inclusion in the
"The statement . . . that the ubiquity of the substance is an indication
of the anthropogenic origin is the result of incorrect reasoning.
Natural sources will tend toward ubiquity more than will artifical
sources. The latter give more localized effects. See the data of
Bozelli ... Singh and others for confirmation of uneven distribution."
The statement relating ubiquity to the anthropogenic origin of CCl^ in
the document does need to be clarified. It was based on research articles
Discussing not only source, but also atmospheric half-life, persistence and
distribution. It is the reported levels of CC1, which are an indication
of its anthropogenic origin. The statement on page 1-1 will be expanded and
clarified to address this point.
3.1.2. Carcinogenic! ty.
"We agree with the disclaimer on pps. 14.18 and A. 6 (of the
that the upper bound risks ... are implausible. The standard bioassays
by NCI are not described by them as being suitable for risk assessment
purposes. Certainly, this positive control group dosed at deliberately
toxic levels, offers no hope of developing any dose response informa-
tion. "
The disclaimer is not that the upper bound risks are implausible but
rather that the upper bound of risk is presently regarded as having limited
plausibility. This is because of the inconclusive nature of the available
evidence for the mutagenicity of CCl^. The inappropriateness of the NCI
bioassays is not equivalent to "no hope." The information on cancer mechan-
isms and the dependence of such mechanisms on dose is admittedly incomplete
and poorly validated. As a consequence, it is not clear that high dose ex-
periments, with concomitant toxic effects, are truly unrelated to lower dose
carcinogenicity , as the commenter contends. Therefore, as long as low-dose
or moderate-dose studies are unavailable, the high-dose studies will be used
for risk assessment.
3-2
-------
"... carbon tetrachloride is a classic example of a nongenetic (some-
times termed epigenetic) carcinogen. This is substantiated amply by
oata presented in this report. The chronic effects are dependent on
dose rate (pps. 2.4, 2.6, 8.17); there are enormous sex and species dif-
ferences (8.2). There is a clear NOEL (8.23) and the chronic damage is
reversible upon cessation of dose (8.21). The substance is not muta-
genic (10.5) nor teratogenic (9.7-8), even though administered at doses
that produce maternal toxicity, and sometimes fetal toxicity. There-
fore, it is clear that tumors arise only after prolonged organ-damaging
treatment, and not as the result of genetic effects.
The nongenetic "proof" given by the commenter is only evidence of toxic-
ity. Since adequate long-term studies involving a large number of animals
exposed at low or medium doses do not exist, the argument that organ damage
is a precursor to cancer is not sustained. Furthermore, it has not been
concluded that CC1. not mutagenic but rather that the information is
insufficient to allow any conclusion to be made concerning the genotoxicity
of CC14.
3.1.3. Risk Assessment. Air Products provided a lengthy discussion on
risk assessment. Some of the points raised will be addressed here. Others
contained some misconceptions in concept, mathematics, chemistry and logic
and will be addressed in section 3.1.5. Other Comments.
"A major failing of most risk assessment efforts in the past has been
the attempt to use a mathematical exercise as the only component, ig-
noring the available scientific data which should modify the rote appli-
cation of a formula."
"... the risk assessment from the carbon tetrachloride data — used
only one data point from an NCI positive control rat feeding of that
material to develop a human risk assessment ... The Agency did attach
(a) cautionary statement ... We believe this to be a proper evaluation
of this attempt at risk assessment, and is appropriate for many other
substances as well."
Perhaps the most important statement made by the commenter is that a
systematic procedure must be developed for incorporating the many different
kinds of biological information into the quantitative risk assessment pro-
cess, especially since the treatment of carcinogens is so different from
that of toxicants (i.e., non-carcinogens, non-mutagens). Currently, several
3-3
-------
subjective judgments must be made during the risk assessment, not the least
of which is the decision that sufficient information exists to term the
chemical a potential carcinogen. Unfortunately, due to the poor under-
standing of cancer mechanisms, metabolic pathways, species differences and
due to the lack of adequate moderate dose cancer studies, these judgments
(which are only some of the sources of uncertainty), must be made. Other
sources of uncertainty are the quantitative procedures used, and the origi-
nal judgments or assumptions made by the researchers. Yet regulation is in
order if a substance has been shown to cause cancer. The Agency is continu-
ally working on new approaches and welcomes additional information on cancer
and toxicity risk assessment techniques. The commenter's only suggestion
was not to act at all.
One interpretation of the negative comments is that the many decisions
and judgments, along with the many types of information used in arriving at
the risk calculation are evidently obscured and need to be stated in some
explicit manner. An attempt was made, however, the extent of the questions
raised implies that still more detail is needed. As noted in section 2.6,
additional discussion and explanation will be incorporated into the document.
3.1.4. Proposed Alternative Procedure. The proposal by Air Products is
given in two steps. The first deals with how to decide if a chemical is a
potential human carcinogen. The second deals with the method of risk cal-
culation once it is decided that the chemical is a potential human carcino-
gen. The first step is essentially a recommendation for a sequential pro-
ceoure with all decision points and criteria clearly defined:
"It is necessary first to validate the experimental data, and to deter-
mine its relevance to humans ... It is then necessary to attempt to
understand the mechanism by which the putative carcinogen acted, geno-
toxic or non-genotoxic; whether the pharmacokinetics of humans resembles
that of the animal; whether the tumor was a new type or merely a (sig-
nificant?) increase in common types; and finally, whether epidemiology
is available to help set an upper risk limit or to modify the animal
data. A decision-tree approach is most helpful in integrating the
other necessary data with the bioassay."
3-4
-------
The proposal has merit. Much of this procedure is already followed by
EPA, although as noted above, it is not evident to the reader of the docu-
ment. This will be clarified in the revised version of the HAD-CCl^.
The second step is basically the "NOAEL-uncertainty factor" approach
used with systemic toxicants (defined as non-carcinogens, non-mutagens):
"If, after a decision-tree type evaluation of all the data, a decision
is reached that a substance is an animal carcinogen that also poses a
potential hazard to humans under reasonably expected circumstances, an
extrapolation of bioassay data should be made. This extrapolation
should utilize all available data and be carried out to some no-
observed-effect level not far from the experimental limits. Most good
bioassay data will allow reasonably accurate extrapolation to the 1%
incidence. There are, however, several NCI bioassays performed at
obviously highly chronic levels which are not suitable even for this
purpose... It is not important which "model" is used for this part;
most fit well and agree within the data limits."
"Once the 1% (10~2) risk point and its statistical error ranges have
been found, a safety factor of the desired magnitude is applied... This
is the allowable dose for the experimental animal at the desired risk."
"This allowable dose is then adjusted for humans by multiplication by
appropriate factors utilizing the data from the decision tree which
justified the conclusion that the substance posed a human risk. Factors
greater than one are used when the data support a larger acceptable dose
for humans than animals, factors lower than one for the opposite cases.
Factors to be considered include those mentioned above: type and hazard
of the tumor in question, exposure and absorption routes and times,
relative metabolism and pharmacokinetics, saturable repair or conversion
mechanism, and so on. The more data that are available, the more care-
fully the final dose can be set. This rewards good science and encour-
ages data gathering."
This approach is currently under investigation by the Agency. The
specific ideas in the American Industrial Health Council have not been con-
sidered since the paper was not available when this document was written.
The commenter does not go far enough, though, in this proposed approach.
Some procedure must exist for the usual situation where the ideal data do
not exist. When no information exists for the subject chemical concerning
human or animal metabolism, human or animal kinetics, human carcinogenicity,
or moderate- or low-dose-induced animal carcinogenicity, when no consensus
3-5
-------
exists on the mechanism of action of a carcinogen, and when the cellular
mechanisms for cancer initiation and growth are not well understood, some
approach must be in place to use in regulating a potential carcinogen.
There are other areas in this proposed procedure which should be ex-
panded, mainly involving the calculations. The mathematical areas of sta-
bility, reliability, robustness and accuracy all need to be explored. The
quantitative effect of certain decisions must also be investigated. For
example, how does the probit model differ from the multistage? How does the
maximum likelihooo estimate differ from a Bayesian approach? What is the
magnitude of change in "acceptable risk-specific dose" if a 90% or 99% con-
fidence limit is used instead of 95%? How should goodness-of-fit enter into
the risk calculation? All of these issues and others are currently being
investigated by the Agency. With some issues, the 1% response-uncertainty
factor approach is definitely inferior to the low-risk extrapolation
approach; with others, the proposed approach is a decided improvement.
Thus, this indicates that more investigation is needed.
In short, the proposed alternative approach has merit. It can be
discussed at the next methodology meeting but cannot be incorporated into
the present document.
3.1.5. Other Comments.
• "The statement in Sec. 2 that carbon tetrachloride will fail to volatize
if it separates as lower lower layer under water is incorrect. The rate
of volatilization will be slowed, because of the need for diffusion
through the water, but it will not be stopped. Carbon tetrachloride is
not insoluble in water, only slightly soluble."
This revision is valid and will be made.
"The effort to determine the annual intake of carbon tetrachloride from
realistic data in Sec. 4.4 is applauded. Too often these calculations
are based on unsupported assumptions. However, we believe that use of
U.S. data on food consumption from the FDA would have been more appro-
priate, and that the use of Bayonne, NJ air data as typical of the en-
tire U.S. is incorrect. See the Bozzelli data for other New Jersey
3-6
-------
sites, ana the Singh data for other (although probably not typical) U.S.
sites. Similarly, the water data probably are biased toward higher than
average levels. See the data made available by the EPA in conjunction
with the VOC in Drinking Water Proposal (47 FR 9350) for other (still
incomplete) data.
The "typical" air data were not those of New Jersey, rather only the
"maximum" level was from New Jersey. The statement on p. 4-21 will be
clarified and additional data will be incorporated as appropriate.
Several issues were raised with regard to risk assessment which con-
tained misconceptions and misinterpretations. These are discussed individu-
ally below.
• The commenter states several times that the risk assesment appears to be
a mathematical procedure which ignores biological theory and data.
The section on unit risk estimation is being expanded in part to explain
more fully the biological data considered prior to the risk estimation as
well as the information directly used in the risk calculation.
The commenter suggests that pharmacokinetic models are appropriate and
necessary for quantitative risk estimation. In particular, the comment-
er states that the tissue-carcinogen must be at least of second order
and that it implies a nonlinear dose-response curve.
EPA has found no information to date which shows the carcinogen-tissue
interaction to be a mass-action chemical reaction or which shows that that
order of a carcinogen reaction direclty indicates the shape of the popula-
tion dose-response curve. When such information becomes available pharmaco-
kinetic information will be used in the risk assessment approach.
The commenter states that specific information is not considered in the
risk assessment, namely background rate, exposure pattern, time-to-tumor
data, non-genetic mechanisms and toxicity-based NOELs.
The background rate is included directly in the risk calculation, as
described in the Federal Register (45 FR 79351). Several proposals are
being investigated by EPA for incorporating directly the exposure pattern
(including intermittent and partial lifetime exposure), time-to-tumor data
3-7
-------
and non-genetic or partially non-genetic mechamisms into the risk calcula-
tions. No consensus has been reached as of this date. Toxicity-based NOELs
and NOAELs are currently not used by EPA in cancer risk estimation because
the NOEL (or NOAEL) is not statistically based, and further, because the
connection between toxic effects and cancer development is not at all clear.
The commenter states that the EPA risk assessment has no biological
foundation.
In addition to devoting an entire chapter to the biological issues asso-
ciated with carcinogenicity, EPA also has demonstrated its concern for such
factors by choosing one of the few mathematical models with any supporting
biological theory. Biochemical evidence (interpreted as kinetic/metabolic
information) is not included because of the insufficient data available to
be of use in the estimation procedure.
The commenter implies that it is inappropriate to assume humans to be as
sensitive as the most sensitive animal.
When sufficient information on species metabolism and cancer response
exist, then such information has been used for study selection. Since human
studies on carbon tetrachloride carcinogenicity are virtually nonexistent,
the human-animal sensitivity question cannot be answered. The conservative
approach is chosen. It must be emphasized again that data are not yet suf-
ficient in scope to allow quantitative modification of the risk assessment
procedure. Desirable data include: rates of production of the active me-
tabolite for different exposure levels, relation between circulating levels
ana target organ concentration, relation between observable tissue damage
ana cancer risk, relation between intermittent or short-term exposure and
lifetime cancer risk, and the differences between species for each of these
factors.
3-8
-------
3.2. E.I. DU PONT DE NEMOURS AND COMPANY
3.2.1. Stratospheric Ozone
3.2.1.1. REVISIONS —
The commenter provided new information and additional references (some
of which were released after the completion of the document) for in-
clusion into the document.
The new information is appreciated. As stated in Section 2.1., informa-
tion contained within these new articles will be incorporated into the docu-
ment.
"The NAS in 1979 (NAS, 1979b) listed four categories of effects antici-
pated from increased UV-B, which, in turn, would result from depletion
of total ozone. They were:
(1) Increased incidence of nonmelanoma skin cancer in humans.
(2) Increased incidence of malignant melanoma skin cancer in humans.
(3) Significant crop yield reductions.
(4) Appreciable killing of marine organisms.
The assessment for three of these categories has changed substantial-
ly."
It is recognized that the NAS assessment for melanoma, crop damage and
mortality of lower marine organisms has changed. This change consists of
the removal of the quantitative estimates of damage contained within the NAS
1979 report. Although this change will be recognized in the document it
shoulo be noted that the removal of these quantitative estimates of damage
is not equivalent to belief in the relative safety of CCl^. The danger of
such effects occurring is still possible; however, the mechanisms are
fraught with uncertainties so as to make quantification difficult.
"Reaction rates and other basic data in the computer models have been
revised, with the result that calculated future depletion when CFCs are
considered alone has been sharply reduced from 16.5% to 5-7% for most
model calculations, or to about one-third the earlier values. A similar
proportinal decrease occurs for emission scencarios involving carbon
tetrachloride."
NAS (1982) estimates of ozone reduction are 5-9%, not 5-7% as stated.
Note that NAS (1982) also pointed out the sensitivity of the ozone reduction
to N02, namely, a doubling of atmospheric N02 leading to ozone reduction
3-9
-------
of 10-16%. These ranges are not exact nor are they statistical confidence
limits. Along with the observation that atmospheric NCL seems to be in-
creasing, this adds another source of uncertainty to the steady-state esti-
mate of ozone reduction. The MAS report is still based on one- and two-di-
mensional models. Since three-dimensional air pollution models exist, a
comment will also be added that future research should evaluate the uncer-
tainty introduced by restricting the models to two dimensions or less.
3.2.1.2. SKIN CANCER TRENDS —
"With recent revisions of calculated future ozone depletion, it is in-
correct to allude that carbon tetrachloride may increase the incidence
of certain forms of skin cancer. While epidemiological evidence indi-
cates that the incidence of skin cancer is increasing, this increase has
occurred in a period when there is no indication from actual measure-
ments that total ozone is being depleted, nor is there any indication of
an increase in UV-8 .... Thus there is no evidence or rational basis to
connect current epidemiological trends in skin cancer incidence with
ozone depletion or carbon tetrachloride emissions."
Although current epidemiological trends in skin cancer incidence are in-
conclusive with regard to ozone depletion or CCl^ emissions, safety has
not been demonstrated. Thus, if any mechanisms relating sunlight to melano-
ma are plausible then "may increase" (as opposed to "will increase") is cer-
tainly correct.
3.2.1.3. SPECIFIC LOCATIONS OF MISLEADING COMMENTS ON ENVIRONMENTAL
EFFECTS ON CARBON TETRACHLORIDE EMISSIONS — Specific comments were:
A. Carbon tetrachloride considered as the only emission affecting total
ozone, and absence of discussion of multiple perturbation calcula-
tions.
HAD-CC14 pages 2-2, 5-2, 6-5.
This comment is wrong. In each case, the phrase is "may contri-
bute," implying uncertainty and other causes.
B. Failure to distinguish adequately between the real atmosphere and
model calculation or attribution of predictive value to model calcu-
lations .
pages 2-2, 6-5, 6-6, 6-7.
3-10
-------
These pages will be revised to reflect more accurately, the model
uncertainties and to distinguish data from model predictions.
C. Use of early reports on effects of ozone depletion with inadequate
mention of subsequent technical revision.
HAD-CC14 pages 2-2, 4-3, 6-6, 6-7.
As stated in section 2.1. of this issue paper, these sections will
be so revised.
D. Misleading discussion on skin cancer.
HAD-CC14 pages 2-2, 4-3, 11-36.
These discussions are not misleading. The verbage used on pages
2-2 and 4-3 is such that it reflects the plausibility of an associ-
ation. The discussion on page 11-36 is abstracted from MAS (1982).
E. Lack of perspective on potential climate effects.
HAu-CCl4 page 4-3.
Again, the statement is "This might...bring about climatic
changes", i.e., it is plausible. It is irrelevant that in terms of
CFC emissions, the MAS (1982) report does not expect "important"
changes in climate.
3.2.2. Toxicology.
"The overall effect of this selection (of studies) is to present carbon
tetrachloride (0014) in a light so as to make more strict control
mandatory."
In preparing such documents, the authors critically review all available
literature, make scientific judgments about the specific studies and draw
scientifically defensible conclusions relating to the health hazard and risk
assessments for the chemical. The accusation of pre-selecting studies to
reach a biased conclusion is incorrect. The studies are clearly discussed
and criticized in chapter 8. The selection of studies for use in risk
assessment is further defended in chapter 14 and is included as an issue in
this paper (section 2.3.).
DuPont contends that "... the chronic inhalation study conducted by
Adams (1952) [sic] ... has not (been) given the appropriate balance ..."
and that "... although conducted in 1952, [sic] the Adams study is sound
and stands today as a good reference point in assessing the subchronic
toxicity of CC14."
3-11
-------
The Adams (1950 is the correct citation) study is not described in great
aetail for scientific reasons as noted in the document. The data presented
in the Adams study are incomplete, because numbers of animals tested, sur-
viving, or affected are not given and it is not possible to determine what
measurements were made at different exposures. Therefore, we are in dis-
agreement with DuPont that "... the Adams study is sound and stands today as
a good reference point in assessing the subchronic toxicity of
"... in controlling for human exposures (setting exposure limits), a
good portion of the data needs to be de-emphasized (rodent studies)
while another part (monkey studies) needs more careful analysis. Here
is a good example of the tenet that toxicological evaluations are most
relevant when the information obtained from animal surrogates is in a
species which handles the compound in a manner most like man. In this
case we have the data (granted the information in monkey and man is con-
siderably less complete than that in rodents) and should be willing to
use it."
The suggestion to use the monkey studies leads one to believe that the
commenter does not understand the quantitative issues involved in risk
assessment. The authors agree that "... the monkey may be the appropriate
animal for extrapolation to man." However, the data on monkeys exposed to
CC1 are insufficient in this case. Different monkeys were tested at 3
concentrations: 2 female monkeys tolerated 100 ppm CCl^ via inhalation
for 232 days; 2 male monkeys tolerated 50 ppm CC1. via inhalation for 277
days; 1 male monkey tolerated 25 ppm CC1. via inhalation for 212 days. It
is not clear if five monkeys were tested in total or if they are the only
monkeys tolerating the doses. Nevertheless, these data are hardly suffi-
cient for risk assessment.
Roaent data are used almost exclusively to predict the effects of CCl^
in man because they are the most appropriate. Rationale is provided in
chapter 14 of the HAD-CCl^ and sections 2.2. and 2.3. of this issue paper.
3-12
-------
Only rodent studies provided sufficient information for use in risk assess-
ment. Differences in species sensitivity as related to humans is considered
in the risk assessment methodology by the use of dose per body surface area
as the conversion between species. The authors of the HAD-CCl^ recognize
that the magnitude of the dose and length of exposure are related to the
effects produced by a given chemical. This is one reason why in the absence
of human data, animal data are carefully selected for use in extrapolation
to equivalent human dose risk assessment considering dose and exposure
information.
"The teratogenicity and other reproductive effects section does a good
job of reviewing the pertinent data. The review does not properly point
out that when reproductive changes (testicular histology, aspermatogene-
sis) have been produced in experimental animals, the dose used was
extremely high and the route of treatment sometimes not relevant (i.e.,
intraperitoneal). Changes seen regarding teratogenic, embryotoxic, or
reproductive effects have been seen only following higher doses."
This comment is well-taken. The high dose and inappropriate route of
exposure for use in risk assessment will be specifically mentioned. How-
ever, note that a general statement is made in chapter 14 addressing these
issues and rendering such studies inappropriate.
The comments specifically pertaining to the mutagenicity of CCl^ and
the mechanism of carcinogenicity of CC1, are addressed in section 3.3. of
this issue paper.
3.2.3. Carcinogenicity - Human.
"The study... (by Blair, 1979)... had several limitations that weaken
its results and make it difficult to interpret." DuPont lists such
weaknesses. "Thus, based on a closer look at the original article,
support for the statement on p. 11-37 that 'human data as reported by
Blair et al. (1979), also are consistent with this conclusion1 ['that
is a potential human carcinogen1] is weak."
The comments made on this section are basically correct and it is agreed
that the epidemiologic evidence for the carcinogenicity of CC1. is weak.
3-13
-------
The preliminary study by Blair et al. (1979) provides the best available
information. It is recognizably weak because it is a preliminary ecological
study using proportionate mortality ratios. This is stated in the document;
however, it will be modified to more strongly state these weaknesses. Since
it is preliminary and ecological, it is presented in such a manner indicat-
ing that it is not possible to identify associations between effects and
exposure but rather to merely defend the need for further more advanced
epidemiologic studies in this area. The comment that "... the study did not
take into account possible geographic or socioeconomic differences in
mortality rates ..." is inappropriate for this level of study. No control
is claimed. Thus, despite the fact that the study is preliminary and itself
cannot indicate the carcinogenicity of CC1. to humans, it does not contra-
dict the hypothesis generated by the animal studies that CCl^ is a poten-
tial human carcinogen. The statement on page 11-37 will be rewritten to
reflect a more cautious conclusion.
3.2.4. Unit Risk Estimate for Cancer.
The general criticism is to expand the discussion of uncertainties or
delete the risk assessment.
The EPA's decision has been to include a risk assessment. The best
possible estimate is presented and, therefore, the uncertainties must be
discussed. The necessary extent of such a discussion is difficult to judge.
Certain factors contributing to the uncertainty are quantifiable; others are
not. There is currently no accepted manner for combining subjective and
quantitative uncertainties into one range of values, although much research
has been performed in this area. Moreover, the commenter is correct in
warning that even the best discussion could be totally ignored, with the
risk estimate "accepted as fact." The discussion on uncertainties is being
expanded. Unfortunately, there is no way to completely prevent misinterpre-
tation of the risk estimates.
3-14
-------
"The linearized multistage model is only one of eight possible risk
extrapolation models (i.e., probit, Mantel-Bryan, logit, Weibull,
one-hit, multi-hit, linear, multistage) that could have been used in
this analysis. This narrow approach should be pointed out to the
reader... The reader should also be informed that the use of models, in
general, and the linearized multistage model, in particular, is not
widely accepted by the scientific community. For example, Squire (1981)
points out that 'no models can actually be based on biological events,
since these are not known for any carcinogens."1
Many models exist; the eight cited by no means form a complete listing.
Based on advice by internal scientists and expert scientists from industry,
acaoemics and other governmental groups, EPA maintains that the multistage
model is still preferable over the others as a general approach. Upper
limit risk estimates based on the multistage model have several mathematical
advantages and some supporting mechanistic theories which the others lack.
Thus, the choice of a "default" model is based on the consensus of scientif-
ic reviewers, and not prior policy. Squire's statement is wrong if strictly
interpreted since the probit is based on a distribution of discrete biologi-
cal events (instances of cancer). His intent, that biological mechanisms
are not known is, of course, true, although mechanistic theories do exist.
3.2.4.2. MORE DOCUMENTATION FOR "LIMITED PLAUSIBILITY" —
"The reasons for the unit risk estimates being 'presently regarded as
having limited plausibility' should be more fully documented. First, it
should be pointed out that the data base is of poor quality for risk
assessment because the doses were very high and produced essentially
100% response... We also emphasize that in the estimation of risk the
relevant dose is that seen by the target tissue rather than the adminis-
tered dose. This relevant dose is necessarily smaller than the adminis-
tered dose... It should also be pointed out that the linearized multi-
stage mooel is the most conservative risk assessment model and that the
estimates of potential risk are actually the upper 95% confidence limits
on potential risk. This approach, in effect, uses mathematical sophis-
tication to cover up two layers of conservatism. EPA has the right to
use such overly conservative procedures; however, EPA also has an
obligation to clearly describe their methodology to decision makers."
As stated under the previous comments, the discussion of uncertainties
and limited plausibility of risk assessment will be expanded. However, the
extent of the expanded documentation has not yet been decided. Some
3-15
-------
phrases, however, are misleading. First, since mechanisms are not known,
one cannot say that data based on high exposure levels are "... of poor
quality for risk assessment ...", although several theories exist as to why
high-exposure responses are poor indicators of low-exposure responses.
Second, the high "doses" as used by NCI are not only selected because of
"positive controls," but often are chosen to guarantee some difference in
response from that of the control group. Third, the linearized multistage
model is not the most conservative since the one-hit and linear models
produce higher risk estimates (assuming more than 2 points). Fourth, the
use of upper confidence limits is clearly stated in the document (p. A-l,
A-3 twice, A-A four times, A-5, A-6 twice). Fifth, since the true risk is
not known, the estimated risk could be an underestimate, so that the phrase
"overly conservative" is unjustified. Sixth, the methodology has been
clearly described and published in the Federal Register; it is not use of
"mathematical sophistication to cover up ... conservatism" as stated by the
reviewer. Nevertheless, additional explanation will be provided as indi-
cated in section 2.6. Note also that target tissue dose may not always be
lower than administered dose, depending on how the word "dose" is defined.
For example, if the commenter is talking about total dose, he is correct.
However, compounds exhibiting accumulation and nonuniform distribution may
result in higher target tissue concentrations than the exposure concentra-
tions. For example, DDT (or DDE) accumulates in adipose tissue to levels
(ppm of tissue) higher than the daily dose (ppm of diet). If adipose tissue
were a target organ in DDT toxicity, the commenter would be in error. Other
examples of such behavior are readily found.
3.2.4.3. UNIT RISK AND EXPOSURE —
"The unit risk estimate is meaningless without relating it to exposure.
A potential hazard isn't a risk until there is exposure. The quantita-
tive risk assessment is incomplete and of little value until exposure is
quantified and evaluated in the risk assessment."
3-16
-------
The definition of "risk assessment" is not clear nor agreed upon by most
scientists, not even by the Society for Risk Analysis. The phrase used in
this document describes the risk estimation process up to the point of
specific locations and exposure scenarios. Some suggestions for the next
step, application to a specific exposure, are given on pages A-4 and A-5 of
the document.
3.2.4.4. MORE THAN MODELING —
"Quantitative risk assessment involves more than the oose-response
modeling of tumor incidence data collected in animal studies. To be
useful and have scientific meaning, the results of the animal studies
must be related to the findings of mutagenicity, metabolism, and epide-
miologic studies and the presence of no-observed-effect-levels (NOEL)..."
As long as the mechanisms are unknown, as long as nongenetic carcino-
genicity remains a vague theory, until it can be shown that carcinogens act
wholly as "threshold" or wholly as "non-threshold" toxicants, any statements
relating mutagenicity and (non-cancer) toxicity to cancer risk are merely
conjecture. Quantitative documentation containing pharmacokinetics showing
that CCl^ must be metabolized to produce toxic effects is welcome. Once
such data are obtained, expansion of the risk assessment section is in
order. Any such modification would also need qualifiers, since metabolism
is not the only determinant of toxicity, others being distribution to
tissues and any detoxification or elimination mechanisms.
3.3. MUTAGENICITY COMMENTS BY AIR PRODUCTS AND DU PONT
Air Products and Chemicals, Inc. and E.I. DuPont DeNemours and Co. com-
mented that CC1. is not mutagenic. The position taken in HAD-CCl^ that
the availaole scientific data are inadequate to conclude that CC1. is not
a mutagen. The metabolism of CC1. to a reactive intermediate, such as the
trichloromethyl free radical (»CC13) (Reynolds and Moslen, 1980; Trudell
et al., 1982), results in covalent binding to nuclear protein, lipid, and
3-17
-------
DMA (Diaz Gomez and Castro, 1980a). This suggests that metabolic products
of CC1A have the potential to react with DMA and that the failure to ob-
serve a mutagenic effect may be due to inadequate testing protocols. Until
better data are available, it is premature to conclude that CC1, is car-
cinogenic by a nongenetic mechanism.
The negative mutagenicity studies mentioned in DuPont's comments on page
5 of part 2 were carried out with exogenous activation systems (S9 mix)
which may be inadequate for the metabolism of CCl^. Even if the S9 acti-
vation system were adequate for metabolism of CCl^, the primary reactive
metabolite of CC1. ('CC1,) is so reactive that it may have been
scavenged by microsomal protein and lipid before entering the bacterial
cell, thereby not having the opportunity to react with the DNA of the test
organism. Metabolically activated CC1A has been shown to bind to micro-
somal protein and lipid in vitro (Uehleke et al., 1977; Reynolds and Moslen,
1980) and to bind to and inactivate the protein responsible for the metabo-
lism of CC1.: the cytochrome P-450 component of the mixed function
oxidase system (DeGroot and Haas, 1980). Additional testing using adequate
protocols that take into account the above two factors (adequacy of the
exogenous metabolic system and ability of the reactive metabolite(s) to
reach the bacterial DNA) is necessary before a decision can be made re-
garding the mutagenic potential of CC1.. The negative mutagenicity data
reported by Uehleke et al. (1977) were inconclusive also because concentra-
tions of CC1. that resulted in toxicity greater than 10K were not tested
and it was not specifically stated in the paper whether rat, mouse, or
rabbit microsomes were used in the S9 activation system.
3-18
-------
Because of the problems described in the previous paragraph (extreme
reactivity of C£L metabolites and scavenging of the metabolites by
microsomal components), test systems utilizing endogenous activation for
investigating the mutagenic potential of CC14 are preferable to those
utilizing exogenous activation. Callen et al. (1980) carried out a study in
yeast, which possesses endogenous activation and is capable of assaying for
point mutations and DNA damage. Positive mutagenicity and DMA damage
results were reported, suggesting that C£L may have genotoxic potential.
DuPont commented (page 5, part 2) that no evidience of covalent binding
to nuclear DNA in mouse and rat liver tissue was found by Rocchi et al.
(1973). Rocchi et al. found evidence of covalent binding to DNA isolated
from mice pretreated with 3-methylcholanthrene. In this same study it was
also shown that rat or mouse microsomes, plus cofactors and activating
enzymes, metabolized CC1. to a substance that bound to calf thymus DNA _in
vitro. Furthermore, covalent binding of CCl^ metabolites to liver nuclear
DNA, protein, and lipid was also found by Diaz Gomez and Castro (1980a).
DuPont is evidently unaware of three recent publications by Diaz Gomez and
Castro (1980a,b, 1981) because only one study published in 1975 by this
group was mentioned in DuPont's comments on page 6 of part 2.
DuPont commented on page 6 of part 2 that Uehleke et al. (1977) measured
binding of ^CCl. to nucleic acid. This study contained no data on the
binding of ^CCl. to nucleic acid. Data on the binding of metabolically
activated CCl^ to microsomal protein and lipid and to added serum albumin
were presented. Significant binding to microsomal protein and lipid was
reported.
DuPont commented on page 2 of part 2 that two studies demonstrating that
CC1. does not induce unscheduled DNA syntheses (DOS) in hepatocytes of
3-19
-------
rats exposed in_ vivo were omitted in the mutagenicity chapter. The chapter
has been revised and includes a review of these studies. The revised chap-
ter is in Appendix B. These studies are specifically reviewed on pages 17
ana 18 of Appendix B.
Mirsalis and Butterworth (1980) tested two doses of CCl^ in an in vivo
UDS assay: 10 mg/kg and 100 mg/kg. The LD5Q for CC14 in rats is
2800 mg/kg. It was not determined in this study at what dose hepatic cell
toxicity would occur under the conditions used. Therefore, it is not clear
that adequate doses of CC14 were tested. Also, it is not clear whether
the 2-hour time period between exposure of the rates to CC1. and isolation
of the hepatocytes was sufficient for observation of UDS. If Mirsalis and
Butterworth had shown that the 2-hour exposure period was sufficient for
activation of the CC1A to a reactive intermediate, perhaps by demonstrat-
ing alkylation of protein by llf«CCl,, the negative results they obtained
would be more convincing. Thus, this negative result could be due to an
inadequate dose or to an inadequate exposure time period.
In a recent abstract, Mirsalis, Tyson, and Butterworth (1982) reported
negative results in the in_ vivo UDS assay after exposure of rats to CCl^
at 400 mg/kg, which is a larger dose than was used in the above study.
However, no details were reported in the abstract and it is not known
whether even this dose was adequate.
Craddock and Henderson (1978) carried out an _in vivo UDS study in which
hepatocyte nuclei were isolated and then assayed for radioactivity by
scintillation counting rather than by grain counting. This method may be
superior to the whole-cell grain counting procedure used by Mirsalis and
Butterworth (1980) as described above, particularly for chemicals that may
cause low UDS activity, because a small effect could be obscured when
3-20
-------
background cytoplasmic grain counts are subtracted from the nuclear grain
counts. In this study, Craddock and Henderson used a CC1. dose of
4000 mg/kg, which is significantly larger than the LD50 (2800mg/kg).
Negative results were obtained after a 2-hour exposure. The authors
suggested that this result may have been due to secondary effects, such as
lysosomal damage, which may result in release of DNA degradative enzymes.
In summary, the above in vivo UDS results suggest that CCl^ may not
damage DNA. However, additional in_ vivo UDS studies at doses just below
those resulting in hepatotoxicity and at exposure time periods longer than 2
hours are needed before a decision on the DNA damaging potential of CC1,
can be made. Also, use of a procedure for the UDS assay in which the
radioactivity of isolated nuclei is assayed, rather than one in which grain
counts of nuclei are corrected for grain counts in the cytoplasm, may better
allow for detection of low levels of UDS.
3.4. OFFICE OF DRINKING WATER, U.S. EPA
Although these comments were received and disposed of via internal re-
view procedures, response will be repeated here.
3.4.1. General Comments. It is recommended that all units be consistent
throughout the document. Conversions were made in the document as far as
scientifically defensible. Further conversions would require assumptions
which could lead to fallacious amounts. In such cases the reported doses
and units are preferred.
3.4.2. Specific Comments. The following specific comments were recieved.
"7. Compound Disposition and Relevant Pharmacokinetics." It will be
more appropriate to change this title to "Pharmacokinetic."
HAD-CC14 page 7-1
3-21
-------
This is not more appropriate since compound disposition is dis-
cussed in chapter 7 and it is not included in the general category
of pharmacokinetics. Pharmacokinetics refers to concentrations,
rates and products of the compound in the body. It does not refer
to disposition or distribution of the chemical, i.e., where it ends
up in the body.
Alumot et al. (1976) study shall be considered with caution throughout
this document since the authors admitted that chronic respiratory
disease was widespread in their animal colony.
HAD-CC14 page 8-18
It is true that Alumot et al. (1976) reported widespread respira-
tory disease in their animal colony. However, the disease began at
14 months. The study summarized on page 8-18 is on the weanling
rats studied by Alumot et al. They were not affected by the respi-
ratory disease.
Since we don't have Mallinckrodt ' s report on the possible impurities
present in the CCl^ sample, it will be inappropriate to refute Callen
et al. (1980) observations which illustrate the genotoxic potential of
CC14.
page 10-2
The lack of the Mallinckrodt data is mentioned on page 10-4. The
revised version of chapter 10 (Appendix B) contains the additional
information.
The correct reference for NCI's bioassay for trichloroethylene is
"Carcinogenesis Bioassay of Trichloroethylene." CAS No. 79-01-6,
NCI-CG-TR-2. U.S. Dept. of Health, Education, and Welfare.
HAD-CC14 page 11-7
The citation remains correct on page 11-7 as NCI, 1976. However,
the citation is in error in the reference section (chapter 15) and
will be corrected.
"Health Effects of Major Concern and Hazard Assessment" seems to be a
more appropriate title for this section.
This is not a more appropriate title since hazard assessment can
mean many things. This chapter deals only with the hazard assess-
ment relating to health. Thus, the title remains: "Health Effects
of Major Concern and Health Hazard Assessment".
HAD-CC14 page 14-1
This subsection should also list pertinent animal ingestion studies such
as Korsrud et al. (1972), Alumot et al. (1976), etc.
HAD-CC14 page 14-18
It does. See Table 14-5.
3-22
-------
APPEfCIX A
Additional References to be Incorporated or Considered for Incorporation
into the Health Assessment Document for Carbon Tetrachloride
Anderson, G.T. et al. 1980. Human Exposure to Atmospheric Concentrations
of Selected Chemicals. Reported to EPA by Systems Applications, Inc. Con-
tract #68-02-3066. May.
Asenov, A. et al. 1980. Khim Ind. (Sofia). 2: 82.
Barber, E.D., W.H. Donish and K.R. Mueller. 1981. A procedure for the
quantitative measurement of the mutagenicity of volatile liquids in the Ames
Salmonella/microsome assay. Mutat. Res. 90: 31-48.
Berger, D.S. and F. Urbach. 1982. Photochem. and Photobiol. 35: 187-192.
Bozelli, J.W. and B.B. Kerbekus. 1980. Analysis of Selected Volatile Or-
ganic Substances in Ambient Air. NTIS, Springfield, VA. PB80-144694.
Bozelli, J.W. et al. 1980. Analysis of Selected Carcinogenic Substances in
Ambient Air in New Jersey. New Jersey Dept. of Environmental Protection.
May.
Callen, D.F., C.R. Wolf and R.M. Philpot. 1980. Cytochrome P-450 mediated
genetic activity and cytotoxicity of seven halogenated aliphatic hydrocar-
bons in Saccharomyces cervisiae. Mutation Res. 90: 31-48.
Cessi, C., C. Colombini and L. Mameli. 1966. The reaction of liver pro-
teins with a metabolite of carbon tetrachloride. Biochem. J. 101: 46c-47c.
A-l
-------
Clark, C.S. et al. 1982. An Environmental Health Survey of Drinking Water
Contamination of Leachate from a Pesticide Waste Dump in Hardeman County,
Tennessee. Arch. Environ. Hlth. 37(1): 9.
Cooper, K. and C. Witmer. 1982. Low oxygen tension and its effects on
CCl^ toxicity and mutagenicity in the Ames test (Abstract Bh-6). Environ..
mental Mutagen Society 13th Annual Meeting, Feb. 26-March 1. 13: 99.
Craddock, V.M. and A.R. Henderson. 1978. De_ novo and repair replication of
DMA in liver of carcinogen-treated animals. Cancer Res. 38: 2135-2143.
Dean, B.J. and G. Hodson-Walker. 1979. An in vitro chromosome assay using
cultured rat liver cells. Mutat. Res. 64: 329-337.
De Groot, H. and W. Haas. 1980. 02-Independent damage of cytochrome
P-450 by CCl^ metabolites in hepatic microsomes. FEBS Lett. 115: 253-256.
Diaz Gomez, M.I. and J.A. Castro. 1980a. Covalent binding of carbon tetra-
chloride metabolites to liver nuclear DMA, proteins, and lipids. Toxicol.
Appl. Pharmacol. 56: 199-206.
Diaz Gomez, M.I. and J.A. Castro. 1980b. Nuclear activation of carbon tet-
rachloride and chloroform. Res. Commun. Chem. Pathol. Pharmacol.
27: 191-194.
Diaz Gomez, M.I. and J.A. Castro. 1981. Reaction of trichloromethyl free
radicals with deoxyribonucleic acid bases. Res. Commun. Chem. Pathiol.
Phannacol. 32: 147-153.
A-2
-------
Dolar, P. et al. 1981. Exposure to Carcinogenic Chemicals and Smoking In-
creases Urinary Excretion of Mutagens in Humans. J. Toxicol. Environ.
Hlth. .8(1-2).
Fishbein, L. 1976. Industrial mutagens and potential mutagens. I. Halo-
genated aliphatic derivatives. Mutat. Res. 32: 267-308.
Harris, R.N. and M.W. Anders. 1980. Effect of Fasting, Diethylmaleate, and
Alcohols on Carbon Tetrachloride-Induced Hepatotoxicity. Toxicol. Appl.
Pharmacol. 56(2): 191.
Harris, R.N. et al. 1982. Interaction Hepatotoxicity of Chloroform and
Carbon Tetrachloride. Toxicol. Appl. Pharmacol. 63(2): 281.
Henry, E.C. 1980. Renal Effects of Inhaled Carbon Tetrachloride and the
Influence of Inhaled Isopropanol: A Comparison with the Hepatic Effects.
NTIS, Springfield, VA. VR-3490-1776.
Hjelle, J.J. et al. 1982. Carbon tetrachloride-ethanol Interactions in
Mice. Toxicol. Lett. 10(1): 17.
Jakobson, I. et al. 1982. Uptake Via the Blood and Elimination of 10 Or-
ganic Solvents Following Epicutaneous Exposure of Anesthetized Guinea Pigs.
Toxicol. Appl. Pharmacol. 63(2): 181.
Jones, B.V. and M. Shah. 1982. Clinico-chemical Changes in Goats Given
Carbon Tetrachloride. Nord. Vet. Med. 34(1-2).
A-3
-------
Keller, E.B. and P.C. Zamecnik. 1956. The effect of guanosine diphosphate
and triphosphate on the incorporation of labeled amino acids into proteins.
J. Biol. Chem. 221: 45-59.
Kraybill, H.F. 1978. Carcinogenesis Induced by Trace Contaminants in
Potable Water. Bull. NY Acad. Med. 54(4): 413.
Kraybill, H.F. 1980. Evaluation of Public Health. Aspects of Carcinogenic/
Mutagenic Biorefractories in Drinking Water. Prev. Med. 9(2): 212.
Kubic, V.L. and M.W. Anders. 1980. Metabolism of carbon tetrachloride to
phosgene. Life Sci. 26: 2151-2155.
Laseter, J.L. and B.J. Dowty. 1977. Association of Biorefractories in
Drinking Water and Body Burden in People. Ann. NY Acad. Sci. 298: 547.
Lee, P.Y. et al. 1982. Evidence for Carbon Tetrachloride-Induced Lipid
Peroxidation in Mouse Liver. Biochem. Pharmacol. 31(3): 405.
Liebovitz, B. and B.V. Siegel. 1980. Antioxidants and Carbon Tetrachloride
Toxicity. Fed. Proc. (3 part 1): 511.
Louria, D.B. and J.D. Bogden. 1980. The Dangers of Limited Exposure to
Carbon Tetrachloride. Crit. Rev. In Toxicology, CRC Press. 7(2): 177.
Lovelock, J.E. 1974. Nature. 252: 292.
A-4
-------
Lovelock, J.E. et al. 1973. Nature. 241: 194.
Lovelock, J.E. 1977. Halogenated Hydrocarbons in the Atmosphere. Ecotox-
icology and Environmental Safety. 1: 399.
Masuda, Y. and T. Murano. 1977. Carbon tetrachloide-induced lipid peroxi-
dation of rat liver microsomes in vitro. Biochem. Pharmacol. 26: 2275-2282.
McCann and Ames. 1976. Proc. Natl. Acad. Sci. 73: 950.
McCann, J., and E. Choi, E. Yamasaki, and B.N. Ames. 1975. Detection of
carcinogens as mutagens in the Salmonella/microsome test: Assay of 300
chemicals. Proc. Natl. Acad. Sci. 72: 5135-5139.
Mirsalis, J.C. and B.E. Butterworth. 1980. Detection of unscheduled DMA
synthesis in hepatocytes isolated from rats treated with genotoxic agents:
an in vivo - in vitro assay for potential carcinogens and mutgens. Carcino-
genesis. 1: 621-625.
Mirsalis, J.C., C.K. Tyson and B.E. Butterworth. 1982. Induction of DNA
repair in hepatocytes from rats treated in_ vivo with genotoxic agents (Ab-
stract Bi-5). Environmental Mutagen Society 13th Annual Meeting, February
26 - March 1. 13: 102.
Nakajima, T. et al. 1982. Dietary Modification of Metabolism and Toxicity
of Chemical Substances - With Special Reference to Carbohydrate. Biochem.
Pharmacol. 31(6): 1005.
A-5
-------
NAS (National Academy of Sciences). 1978. Chloroform, Carbon Tetrachlor-
ide, and Halomethanes: An Environmental Assessment.
NAS. 1979a. Stratospheric Ozone Depletion by Halocarbons: Chemistry and
Transport.
NAS. 1979b. Protection Against Depletion of Stratospheric Ozone by Chloro-
fluorocarbons.
NAS. 1982. Causes and Effects of Stratospheric Ozone Reduction: An Update.
Noguchi, T. et al. 1982. Selective Early Loss of Polypeptides in Liver
Microsomes of CCl^-treated Rats. Relationship to Cytochrome P-450 Con-
tent. Biochem. Pharmacol. 31(5).
O1Donovan, D.J. and T.M. Dunne. 1977. Structural and Metabolic Effects of
Carbon Tetrachloride Administration in the Rat. J. Anat. 124(2): 490.
Patterson, J.W. and P.S. Kodukala. 1981. C.E.P. 77(4): 48.
Pohl, L.R., R.V. Branchflower, R.J. Highet, J.L. Martin, D.S. Nunn, T.J.
Monks, J.W. George and J.A. Hinson. 1981. The formation of diglutathionyl
dithiocarbonate as a metabolite of chloroform, bromotrichloromethane, and
carbon tetrachloride. Drug Metab. Disposit. 9: 334-339.
A-6
-------
Popp, J.A., S. Hisashi and E. Farber. 1978. The protective effects of di-
ethyldithiocarbamate and cycloheximide on the mutliple hepatic lesions in-
duced by carbon tetrachloride in the rat. Toxicol. Appl. Pharmacol.
45: 549-564.
Powell-Jackson, P. and M. Davis. 1979. Occupational Liver Disease. Prac-
titioner. 223(1333): 67.
Reynolds, E.S. and M.T. Moslen. 1980. Free-radical damage in liver. In:
Free Radicals in Biology, Vol. IV, pp. 49-94, W.A. Pryor (ed.). Academic
Press, New York.
Rocchi, P., G. Prodi, S. Gilli and A.M. Ferreri. 1973. _In vivo and in
vitro binding of carbon tetrachloride with nucleic acids and proteins in rat
and mouse liver. Int. J. Cancer. 11: 419-425.
Saleh, M.A. 1980. Mutagenic and Carcinogenic Effects of Pesticides. J.
Env. Sci. Hlth. (B) 15(6): 907.
Sawicki, E. 1977. Chemical Composition and Potential 'Genotoxic1 Aspects
of Polluted Atmospheres. IARC Sci. Pub. (16): 127.
Shah, H., S.P. Hartman and S. Weinhouse. 1979. Formation of carbonyl
chloride in carbon tetrachloride metabolism by rat liver in_ vitro. Cancer
Res. 39: 3942-3947.
A-7
-------
Simmon, V.F., K. Kauhanen and R.G. Tardiff. 1977. Mutagenic activity of
chemicals identified in drinking water. In: Progress in Genetic Toxicology,
D. Scott, B.A. Bridges, and F.H. Sobels (eds.), pp. 249-258. Elsevier/
North-Holland Biomedical Press, New York.
Singh, H.F. et al. 1975. Atmospheric Formation of Carbon Tetrachloride
from Tetrachloroethylene. Env. Letters. 10(3): 253.
Sipes, G.I., G. Krishna and J.R. Gillette. 1977. Bioactivation of carbon
tetrachloride, chloroform and bromotrichloromethane: Role of cytochrome
P-450. Life Sci. 20: 1541-1548.
Tabak, H.H. et al. 1981. Biodegradability Studies with Organic Priority
Pollutant Compounds. J. Water Poll. Control Fed. 53(10): 1503.
Takizawa, S. 1975. Preparative Action of Carbon Tetrachloride in Liver
Tumorigenesis by a Single Application of N-Butylnitrosourea in Male ICR/ICL
Strain Mice. Gann. 66(6): 603.
Tanabe, S. et al. 1981. Distribution and Total Burdens of Chlorinated
Hydrocarbons in Bodies of Striped Dolphins (Stenella coeruleoalba). Agric.
Biol. Chem. 45(11).
Trudell, J.R., B. Bosterling and A.J. Trevor. 1982. Reductive metabolism
of carbon tetrachloride by human cytochromes P-450 reconstituted in phos-
pholipid vesicles: Mass spectral identification of trichloromethyl radical
bound to dioleoyl phosphatidylcholine. Proc. Natl. Acad. Sci. 79: 2678-2682.
A-8
-------
Uehleke, H., H. Greim, M. Kramer, and T. Werner. 1976. Covelant binding of
haloalkanes to liver constitutents, but absence of mutagenicity on bacteria
in a metabolizing test system. Mutat. Res. 38: 114.
Uehleke, H., H. Greim, M. Kramer and T. Werner. 1977. Metabolic activation
of haloalkanes and tests in vitro for mutagenicity. Xenobiotica. 7: 393-400.
Ugazio, G. et al. 1972. Mechanism of Protection Against Carbon Tetrachlor-
ide by Prior Carbon Tetrachloride Administration. Exptl. Mol. Pathol.
16: 281.
Vainio, H., M.G. Parkki, and J. Marniemi. 1976. Effects of aliphatic
chlorohydrocarbons on drug-metabolizing enzymes in rat liver in_ vivo. Xeno-
biotica. 6: 599-604.
Wang, W.C. et al. Greenhouse Effects due to Antropogenic Perturbations of
Trace Constituents in the Earth's Atmosphere. Goddard Institute for Space
Studies.
Whittle, K.J. et al. 1977. Occurrence and Fate of Organic and Inorganic
Contaminants in Marine Animals. Ann. NY Acad. Sci. 298: 47.
World Meteorological Organization. Issued in 1982. The Stratosphere 1981:
Theory and Measurements. WMO Global Ozone Research and Monitoring Project
Report No. 11, May 1981.
A-9
-------
Wuebbles, D.J., F.M. Luther, and J.E. Penner. 1982. Effect of coupled
anthropogenic perturbations on stratopsheric ozone. 183rd ACS National
Meeting, Las Vegas, NV. April 1, 1982. (Submitted to the Journal of Geo-
physical Research for publication.)
A-10
-------
APPENDIX B
Rewrite of Chapter 10. Mutagenicity
-------
THE .REPRODUCTIVE EFFECTS ASSESSMENT GROUP'S
DRAFT REPORT
ON THE MUTAGENICITY OF CARBON TETRACHLORIDE
Peter E. Voytek, W.0.
Di rector
Prepared by
Sheila L. Rosenthal, Ph.D.
DRAFT
DO NOT QUOTE OR CITE
This document is a preliminary draft. It has not been formally released by
the EPA and should not at this time be construed to represent Agency policy.
It is being circulated for comment on its technical accuracy and policy
implication.
-------
TABLE OF CONTENTS
Introduction 1
Metabolism and Covalent Binding to Macromolecules 1
Mutagenldty Studies In Bacterial Test Systems . 7
Studies in Eucaryotic Test Systems 12
Studies Indicative of Primary DNA Damage 15
Suggested Additional Testing. 17
Summary and Conclusions 18
References 21
DRAPT
DO NOT QUOTE OR
CITE
MOV 1 2 1982
-------
INTRODUCTION
The mutagenic potential of carbon tetrachloride (€014) has b
DRAFT
00 NOT QUOTE OR
CITS
NOV1 2 1982
sen
assessed by evaluation of the results from seven bacterial studies, one yeast
study, one 1n vitro mammalian chromosome damage study and two in vivo DNA
damage studies in rodents. The majority of these studies were negative.
Information relating to the metabolism of CC14 and to covalent binding of
the metabolites to cellular macromolecu!es (including DNA) is presented before
the sections assessing the genotoxicity of €014. This was done to set the
stage for the discussion of the largely negative results obtained in the
mutagenicity studies described below and for the suggestion that CC14 may be
a weak mutagen. In addition, suggestions for additional testing are
presented.
METABOLISM AND COVALENT BINDING TO MACROMOLECULES
The evidence described in the section suggests that CC14 is metabolized
in the liver to.highly reactive intermediates (the trichloromethyl free
radical and phosgene). The evidence also indicates that metabolically
activated CC14 covalently binds to protein, lipid, and DNA, suggesting that
CC14 mav nave genotoxic potential.
Metaboli sm
CC14 is metabolized in the liver endoplasmic reticulum by the
cytochrome P-450 component of the mixed function oxidase system (Reynolds and
Moslen 1980). The available evidence indicates that metabolism of CC14
results in the generation of the trichloromethyl free radical -CC^
(Reynolds and Moslen 1980, Trudell et al. 1982) and phosgene (Shah et al.
1979, Kubic and Anders 1980, Pohl et al. 1981). These two substances are the
most likely of the metabolites to exhibit reaction with tissue macromolecules
-------
because of their high reactivity. By using human cytochrome P-450
reconstituted in phospholipid vesicles, Trudell et al. (1982) have
demonstrated that *CCl3 is the major product of the reductive metabolism
of CC14 as determined by mass spectral identification of the adducts formed
between *CCl3 and the phospholipid dioleoyl phosphatidylcholine.
Phosgene results from further metabolism of »CCl3 (Shah et al . 1979)
(see Fig. 7-1).
CC1 4
Pohl et al . (1981) measured the amount of phosgene (as diglutathionyl
dithiocarbonate) produced from the aerobic metabolism of 0014, CHC13, and
CBrCl3 by liver microsomes (from phenobarbital -treated rats) plus cofactors.
The results indicate that phosgene production from CC14 is only 4% of that
produced from CHC13. Thus, the level of phosgene production from aerobic
metabolism of CC14 is small. The card nogeni city and mutagenicity of
phosgene are currently under "investigation in Dr. B.L. Van Duuren's laboratory
at New York University Medical Center (Dr. Sipra Banerjee, personal
communication). Metabolites of CC14 are so reactive that they bind to and
inactivate the cytochrome P-450 enzymes that were responsible for their
generation (suicide mechanism) (Vainio et al . 1976, Sipes e£ £]_! 1977, De
Groot and Haas 1980, Cooper and Witmer 1982).
Covalent Binding to Macromol ecu! es
Metabolically activated CC14 has been found to bind covalently* to
*Amount of label bound was determined after washing and extraction
indicating that the binding was covalent.
DRAFT
DO NOT QUOTE OR
CITE
NOV 1 2 1982
-------
lipid and protein both in vivo (Rocchl et al. 1973, Diaz Gomez and Castro
1980) and j£ vitro (Rocchi et al. 1973, Uehleke et al. 1977). Uehleke et al.
(1977) measured covalent binding of 1 mM 14CC14 (0.25 Ci/mol) to
microsomal protein and lipid in liver microsome suspensions (5 mg protein/ml
plus cofactors) from phenobarbital-treated rabbits. About 10% of the 14C
label was covalently bound to endoplasmlc reticulum protein and greater than
i
30% was bound to microsomal lipid. Extramicrosomal binding was evaluated by
the addition of 5 mg of bovine serum albumin per ml to the CCl4/m1crosome
mixture. The binding of metabolically activated ^CCl4 to added bovine
serum albumin (1.4 nmol/mg in 60 min) was about 1.5% of that bound to
microsomal protein (20.0 nmol/mg) plus lipid (76.0 nmol/mg). Thus, it appears
that binding of metabolically activated 14CC14 to extramicrosomal
macromolecules is negligible compared to binding to microsomal constituents.
(Mutagenicity studies were also described in this paper and are discussed in
the next section.)
Evidence that the CC14 metabolite, phosgene, reacts with proteins was
obtained by Cessi et £j_. (1966) when they measured the in vivo binding of
CC14 to rat liver proteins and compared it to the in vitro acylation of
poly-L-lysine and serum albumin by phosgene. Similar reaction products were
obtained in both systems, suggesting that phosgene reacts with the lysine
£-amino groups in proteins, leading to cross-linked carbonyl derivatives:
C1-C-C1 + 2 lysTne ^ lysine-fl-C-ft-lysine
phosgene proteins cross-linked proteins
or
Q NH2 NH2 NH-C-NH^
C1-C-C1 + ....lysine....lysine... > lysine lysine
phosgene protein cross-linked protein
DRAFT
DO NOT QUOTE OR
CUE
MOV
-------
Such cross-linked proteins would undoubtedly exhibit Impaired biological
activity. It is also possible that similar cross-linking reactions of
phosgene can occur with ami no groups in DNA, resulting in alterations in DMA
structure and function.
Binding of metabolically activated CC14 to DNA was found by two groups.
Rocchi et aj_. (1973) have studied the binding of CC14 to nucleic acids and
protein. 14C-Labeled CC14 (367 umol/kg) was injected into rats and mice
and the amount of metabolite(s) of CC14 that covalently bound to liver DNA,
RNA, nuclear proteins, and cytoplasmic proteins was measured. They reported
that significant amounts of labeled material were found associated with rRNA,
nuclear proteins, and cytoplasmic proteins in rats. When the rats were
pretreated with 3-methylcholanthrene (5 mg, 24 hr before treatment with
0014), the amount of label associated with the macromolecules increased. No
label was associated with DNA in the rat studies. Similar studies in mice
indicated that DNA binding occurred (108 umol CCl4/mol DNA), but only after
pretreatment with 3-methylcholanthrene (1 mg, 24 hr before CC14 dosing).
In an in vitro experiment, Rocchi et £l_. (1973) used rat or mouse liver
microsomes to activate labeled CC14 in the presence of calf thymus DNA.
They found that pretreatment of animals with 3-methylcholanthrene enhanced the
amount of label associated with DNA. Furthermore, pH 5 enzyme preparations
which contain activating enzymes (Keller and Zamecnik 1956), were also found
to increase the amount of label bound to DNA. Therefore, from these results,
it appears that metabolites of CC14 can interact with DNA; however, for
optimal binding conditions, microsomal enzymes had to be activated with
3-methylcholanthrene and the binding assay had to be carried out in the
presence of pH 5 enzymes.
CO SOI QUOTE OR
CITE
NOV 1 2 1982
-------
Diaz Gomez and Castro (1980a) have also measured binding of metabolically
activated CC14 to cellular macromolecules. 14C from 14CCl4 (specific
activity, 27 C1/mol) irreversibly bound in vivo to liver nuclear DNA, protein,
and lipids in strain A/J mice and Sprague-Dawl ey male rats. Mouse and rat
liver nuclear DNA Isolated from animals given 14CCl4 16 hr before they
were killed exhibited a small but significant labeling (mice, 0.72 pmol/mg;
rats 0.52, pmol/mg). The count from the assay done in the presence of
unlabeled DNA was subtracted from the experimental counts before binding was
calculated. In contrast to the results of Rocchi et aU (1973), induction of
liver enzymes by 3-methylcholanthrene was not required for binding of ^C
from 14CCl4 to DNA. It should be mentioned that the purified DNA samples
contained 0.2% protein. However, contamination by protein at this low level
could not account for all the covalent binding measured in the DNA sample.
Significant in vitro binding of metabolically activated CC14 to
isolated mouse liver DNA (1.81 pmol/mg) was observed by Diaz Gomez and Castro
(1980a) in anaerobic Incubation mixtures containing microsomes and NADPH. It
was also found that *CCl3 chemically produced by reaction of CC14 with
benzoyl peroxide interacted with DNA (826 pmol/mg). This result suggests that
•CC13 may be the species involved in the binding to DNA.
In addition to the above observations that metabolites of CC14 bind to
DNA, Diaz Gomez and Castro (1980a) also observed significant binding in vivo
of metabolically activated CC14 to rat liver nuclear protein and lipid. The
label bound to nuclear protein was 47.7 pmol/mg and that bound to nuclear
lipid was 113.5 pmol/mg. The authors suggested that the binding of
metabolically activated CC14 to nuclear lipids is significant in terms of
the carcinogenicity of CC14, because nuclear lipid is derived from the
nuclear membrane which contains the P-450 necessary for activation of CCl
DO NOT QUOTE OR
CITE
NOV 1 2 1982
-------
to the probable reactive metabolites *CCl3 and phosgene. Since these
metabolites are highly unstable and therefore are not likely to exist long
enough to travel from the endoplasmlc reticulum P-450 enzymes to nuclear
components, activation by nuclear membrane P-450 enzymes Is more likely to
yield Interactions of the metabolites with nuclear components (e.g., DNA) than
will activation in the endoplasmlc reticulum.
01az Gomez and Castro (1980b) have assessed the CC14 activation
potential of purified rat liver nuclei by measuring covalent binding of
nuclear activated CC14 to nuclear protein and lipid. Binding to DNA was not
measured 1n this study. The results were compared to results obtained from
similar incubation mixtures containing microsomes instead of purlfTed nuclei.
The incubation mixtures containing either nuclei (1.3 mg protein/ml) or
microsomes (1.56 mg protein/ml) were incubated for 30 min in 37.6 nM
^CCl4 (6.94 Ci/mol) and an NADPH generating system in an 02-free N2
atmosphere.
The authors observed that the extent of binding to proteins in the nuclear
preparations was 43.5% of that observed for microsomes (nuclear suspensions,
21.9 pmol/mg; microsomes, 50.3 pmol/mg). Binding to nuclear lipids was 77.3%
of that observed for microsomes (nuclear suspension, 147 pmol/mg; microsomes,
190 pmol/mg). Thus, isolated nuclei were less efficient than microsomes in
metabolizing (£14 but the results were within the same order of magnitude.
This study suggests that metabolism of CC14 to reactive intermediates
can occur in nuclear membranes and indicates that the in vivo binding observed
in the previous study (01az Gomez and Castro 1980a) may have been due to
nuclear rather than microsomal activation of CC14- It should be
mentioned,however, that the nuclear preparations were contaminated with trace
amounts of endoplasmic reticulum, which may have been sufficient to result in
at least part of the nuclear activation observed.
CITE
6 NOV 1 2 1982
DRAFT
DO NOI QUOTE OR
-------
Diaz Gomez and Castro (1981) have published preliminary evidence that
•CCIs, chemically generated from the benzoyl peroxide-catalyzed
decomposition of CC14, reacts primarily with guanine and adenine and to a
lesser extent with cytosine and thymine. This result suggests that 'CC^
may bind to DNA in vivo by interaction with the deoxyribonucleic acid bases.
In summary, it has been shown that CC14 is metabolized to reactive
intermediates (*CCl3 and phosgene). It has also been shown that
metabolically activated CC14 binds to DNA, protein, and lipid. These
results suggest that CC14 has genotoxic potential. The negative results in
six of the seven bacterial mutagenicity studies described in the next section
may be due to inadequate metabolic activation in the test systems used or may
result from the scavenging by protein or lipid of any very reactive metabolic
intermediates formed (e.g., *CCl3 and phosgene) under conditions of
exogenous activation, thus limiting their availabliity for reaction with DNA.
Thus, test systems used to assess the genotoxicity of CC14 should
incorporate adequate metabolic activation and an indication that any highly
reactive metabolites formed were not scavenged by microsomal protein or lipid
before reaching the DNA of the test organism.
MUTAGENICITY STUDIES IN BACTERIAL TEST SYSTEMS
Studies to determine the mutagenic activity of CC14 in the Salmonella
typhimurium revertant system have been primarily negative. A review article
written by McCann et jil_. (1975) stated that an assay using Aroclor-induced S9
activation and strains TA 100 and TA 1535 was negative. The McCann et al.
article contained no details of the procedure used to generate this negative
result. Another review article (Fishbein 1976) in which no data were
••——
DO NOT* QUOTS OR
CITE
\ 2 >982
-------
presented contained a statement that CC14 was not mutagenic when assayed in
a spot test with the TA 1950 strain.
In-an abstract, Uehleke et al. (1976) reported that CC14 was not
mutagenic 1n liquid Salmonella typhimurium strains TA 1535 (test for base-pair
substitution) or TA 1538 (test for frame-shift mutation) and Escherichia coli
K12 Incubated with microsomes. However, these results cannot be evaluated
because no data were presented.
In a more detailed study, Uehleke et al. (1977) tested the mutagenicity
of CC14 in suspension assays with S^. typhimurium strains TA 1535 and TA
1538. No mutagenic activity was observed. About 6-9 x 10s bacteria were
incubated for 60 min under N2 in tightly closed test tubes with 8 mM CC14
and microsomes (5 mg protein) plus cofactors. The mutation frequencies
(his"1" colony forming units/10^ his" colony forming units) were less than
10 for both strains and the spontaneous mutation frequencies were 3.9 +_ 3.7
for strain TA 1535 and 4.4 +_ 3.5 for strain TA 1538. At this concentration of
CC14, survival of the bacteria was at least 90%. Thus, the negative result
for 0014 cannot be interpreted since concentrations of CC14 resulting in
less than 90% survival should also have been tested. Dimethylnitrosamine,
cyclophosphamide, 3-methylcholanthrene, and benzo[a]pyrene were the positive
controls used in this study. Although these chemicals were mutagenic in the
presence of the S9 activation system, they may not be appropriate controls
because they in no way resemble halogenated alkanes.
Because metabolically activated CC14 did not significantly bind to
exogenously added bovine serum albumin (see binding studies above), Uehleke e_t_
al. (1977) concluded that any reactive species generated by the microsomes may
not have distributed into the incubation medium and, thus, may have been
inaccessible to the test bacteria, resulting in the negative mutagerH-ert-y- ,„
NOV 1 2 1982
-------
response observed. However, It is not clear from the description provided in
this report whether rat, mouse, or rabbit microsomes were used in the
mutagenicity studies. It is clearly stated that rabbit microsomes were used
for the binding studies described above. If mouse or rat microsomes rather
than rabbit microsomes were used for the mutagenicity experiments, it cannot
be assumed that CC14 was sufficiently activated, since activation sufficient
for binding of 14CC14 to macromolecules was shown 1n this paper only with
rabbit microsomes. Another deficiency in this study is that the Ames strains
TA 98 and TA 100 were not used. These strains contain an R factor plasmid
that increases the sensitivity of the tester strains to certain mutagens.
Because of these deficiencies, the negative mutagenicity results in this paper
are inconclusive.
The mutagenicity of CC14 was also tested in a study designed to
evaluate the mutagenic potential of chemicals identified in drinking water
(Simmon et al. 1977). No mutagenic activity was detected with CC14. The
authors tested 71 of the 300 chemicals that had been identified in public
water supplies. 0014 was tested in this study in a desiccator to assess
mutagenicity due to vapor exposure and to avoid excessive loss of CC14 to
the atmosphere. The desiccator contained a magnetic stirrer which acted as a
fan to aid in evaporation of the measured amount of CC14 and to maintain an
even distribution of the vapors. Plates were exposed to the vapors for 7-10
hr and then removed from the desiccators, covered, and incubated approximately
40 hr before scoring. Mutagenic activity was not observed and no information
on toxicity was provided.
This study by Simmon et al. (1977), although lacking some specific details
of the CC14 assay, clearly identifies certain trihalomethanes (CHBT3,
, CHBrClg) as mutagens in the vapor assay in desiccators. Mettw*
DRAPT
DO HOT QUOTE OR
CITE
1 2 1982
-------
bromide, methyl chloride, methyl iodide, and methylene chloride were also
found to be mutagenic in the desiccator assay. However, these seven
halogenated compounds did not require metabolic activation to exhibit
mutagenic activity. It may be that CC14 itself is not mutagenic and the rat
liver S9 does not effectively metabolize CC14 to a potential mutagenic
reactive Intermediate (? •CCl3)r even though the demonstration of
mutagenicity oiMthree oHthe chemicals- tested [bis(2-chloroisopropyl )ether,
vinyl chloridev and vinyTidene chloride] required or was enchanced by this S9
mix. It may also be that a reactive intermediate was formed, but it was too
reactive or short-lived to be detected in a test system that uses exogenous
metabolic activation.
Another negative result for the mutagenicity of CC14 was obtained in a
recent study using the Salmonella/microsome assay in which escape of the
volatile compounds was prevented by the use of a specially designed, closed,
inert incubation system (Barber e_t £l_. 1981). Seven of the ten halogenated
alkane solvents tested gave positive mutagenicity results when the Ames assays
were carried out in the closed incubation system. Under standard conditions
(in which volatilization was not prevented), only 2 of the 10 solvents tested
gave a positive result. Thus, the closed incubation system allowed for the
detection of five more mutagens than could be detected under standard
conditions. CC14 was one of the three solvents tested that gave a negative
result in both the standard and closed incubation systems. The investigators
indicated that CC14 was tested at concentrations high enough to produce
observable toxicity (determined by absence of background lawn). The
Salmonella strains used were TA 1535, TA 1537, TA 1538, TA 98, and TA 100.
Levels of CC14 tested were 4.7, 5.7, 10.2, 12.3, and 18.4 umol per plate.
No dose-related response was observed. The seven solvents that tested
DRAFT'
DO NOT QUOTE OR
CITE
10 NOV 1 2 1982
-------
positive in this closed system did not require metabolic activation by S9 mix
prepared from Aroclor-induced rats (the S9 mix did activate the positive
control 2-aminoanthracene). It 1s possible that the S9 mix used, although
adequate for activation of 2-aminoanthracene, was not adequate to metabolize
the other three solvents (Including CC14). It is also possible that active
metabolites, If formed, reacted with microsomal components (see below) before
reaching the bacterial DNA.
In a recent abstract, Cooper and Witmer (1982) reported that exposure of
Salmonella strain TA 100 to CC14 for 20 minutes in a 1-ml liquid suspension
at low oxygen tension (3-ml vacutainer tubes) before preparation of plates
resulted in a twofold increase in revertants above background (background, 72
+_ 9 revertants; 5 umol CC14, 141 +_ 20 revertants). Evidence for a
dose-response relationship was not reported in this abstract. The weak
mutagenic result was observed in two separate experiments and only when fresh
rabbit liver S9 was used as the activation system. It is unlikely that the
rabbit liver S9 alone was responsible for the mutagenic activity observed
because plate assays with rabbit S9 exhibited no mutagenic activity. The
CC14 was spectrophotometric grade (Spectrar) purchased from Mallinkrodt (Dr.
Charlotte Witmer, personal communication). CCl^ was toxic in this
suspension assay; 2 umol/ml caused 80% toxicity with microsomal activation and
20% without activation. Addition of 0.1 mM EDTA to inhibit microsomal lipid
peroxidation decreased bacterial toxicity.
The authors concluded that the suspension assay in the vacutainer tubes
may be a suitable system for testing volatile compounds that undergo a
reductive metabolism. This weak positive result constitutes preliminary
evidence that CC14 may be mutagenic in the Ames assay, but only under
certain assay conditions (low oxygen tension, exposure of bacteria in
DRAFT
DO NOT QUOTE OR
CITE
11 NOV 1 2 1982
-------
suspension In the presence of EDTA, and use of fresh rabbit liver S9).
In summary the results from the above bacterial studies are Inconclusive,
since false negative results could have been obtained due to a number of
factors, including:
1. The activation systems used may have been inadequate for metabolism
of CC14.
2. *CCl3 and phosgene, the primary reactive metabolites of CC14,
are unstable and highly reactive. Because exogenous activation
systems were used in many of these studies, any *CCl3 or phosgene
generated (assuming an adequate activation system) may have been
scavenged by microsomal protein or lipid before reaching the DNA.
3. Adequate exposure to CC14 may not have occurred if appropriate
precautions were not taken to prevent the evaporation of 0014.
STUDIES IN EUCARYOTIC TEST SYSTEMS
Call en e££]_. (1980) carried out a study on the mutagenicity of CC14 in
the D7 strain of Saccharomyces cerevislae, which contains an endogenous
cytochrome P-450 dependent monooxygenase activation system. By using this
strain of yeast, Call en and his coworkers eliminated the need for the
exogenous type of metabolic activation system used in the above bacterial
studies. Three different genetic endpoints can be examined with this system:
gene conversion at the trpS locus, mitotic crossing over at the ade2 locus,
and gene reversion at the ilvl locus. The effect of CC14 (Mallinckrodt,
99+% pure) on these endpoints was measured by exposing cells in suspension
to 3.23, 4.31, and 5.13 g of CC14 per liter of buffer (21 mM, 28 mM, and 34
mM, respectively), well above the solubility level of CC14 in water (0.8 g/1
at 25°C). Therefore, a dose-response relationship could not be obtained
because the dose was essentially the same in all cases -the solubility level
of CC14 in water at 37°C. Escape of volatilized CC14 is not expected to
have occurred to any significant extent, because the incubations were carried
out in screw-capped glass tubes. Although the dose is essentially constant, DRAFT
DO NOT QUOTE OR
CITE
MOV 1 2 1982
-------
amounts In suspension will vary. Extracellular or membrane effects may result
in the high toxidty observed at 5.13 g/1.
Results of the Callen et £l_. study are presented in Table 1. A 1-hr
treatment of cells with CC14 at the highest amount tested (5.13 g/1)
resulted 1n significant increases 1n gene conversion (31-fold) and mitotic
crossing over (25-fold). Gene reversion was also increased, but to a lesser
extent (3-fold increase). Survival was only 10% at this concentration of
CC14. At the intermediate level (4.31 g/1) of CCl*;, much smaller effects
in the three genetic systems were observed (2- to 3-fold increases). Survival
at this level was 77%. Thus, the data in the Callen et al. (1980) study are
suggestive of a weak positive response, but additional studies are needed
before it can be conclusively stated that CC14 causes genetic effects in
yeast.
Negative results have been obtained in a recently developed in vitro
chromosome assay that utilized an epithelial-type cell line derived from rat
liver (Dean and Hodson-Walker 1979). This cell line has sufficient
metabolizing activity to activate various chemical mutagens and carcinogens
without the need for an exogenous activating system. Sealed-flask cultures
were treated for 22 hr with CC14 dissolved in growth medium at 0.005, 0.010,
and 0.020 mg/1. CC14 at these low concentrations did not induce any
chromatid or chromosomal aberrations, whereas a number of direct-acting
mutagens and several requiring metabolic activation produced chromatid
deletions, gaps, and exchanges. However, none of the other substances tested
was a heavily chlorinated compound. In addition, one of these substances was
assayed at levels comparable to that used for CC14 and the remaining 10
compounds were assayed at doses several orders of magnitude larger. The doses
chosen for each substance assayed were determined from cytotoxiclt^ assays.
13
DO «OT QUOTE OR
CUE
MOV 1 2 1982
-------
iiUiii.cMi.1 ai. tun, y/
0
3.23
4.31
5.13
Survival
Total colonies
% of control
trpS locus (gene conversion)
Total convertants
Convertants/10^ survivors
ade2 locus (mitotlc crossing over)
Total twin spots
Mitotlc cross-overs/10* survivors
Total genetically altered colonies
Total genetically altered colonies/
survivors
1 1 v 1 locus (gene reversion)
Total revertants
Revertants/lO** survivors
1454 1252 1120 152
100 86 77 10
285 331 350 506
2.0 2.6 3.1 61.7
1 3 3
1.6 5.3 5.8
11 19 16
1.7
3.4
3.1
38 41 57
2.6 3.3 5.1
10
40.1
65
33.3
11
7.2
aSource: Adapted from Call en et aK. 1980
bThe total number of colonies In the different classes represent total counts of colonies
from five plates in the case of survival, conversion and revertant-frequency estimations. Mltotl
crossing over was estimated from counts of colonies growing on a total of 30 plates, 20 plates
containing medium on which all surviving cells grew and 10 plates containing medium on which only
trpS convertants grew.
14
DRAFT
DO HOI QUOTE OR
CITE
NOV1 2 1982
-------
Apparently, CCl^ was very toxic to rat liver cells. EDTA (0.1 mM) has been
found to decrease the cytotoxicity of CC14, without affecting mutagenicity in
bacteria (Cooper and Witmer 1982), apparently by decreasing membrane lipid
peroxidation (Masuda and Murano 1977). Perhaps addition of EDTA to a
mammalian in vitro chromosome assay, such as in the1study carried out by Dean
and Hodson-Walker (1979), would allow for the use of larger concentrations of
CC14.
STUDIES INDICATIVE OF PRIMARY DNA DAMAGE
Mirsalis and Butterworth (1980) measured unscheduled DNA synthesis (UDS)
in primary rat hepatocyte cultures following in vivo treatment of adult male
Fischer-344 rats (200-250 g) with CCU (certified ACS grade, Fischer
Scientific Co.) by oral gavage. Control rats received corn oil by gavage.
Acetylaminofluorine and dimethylnitrosamine were also tested. At 2 hr after
treatment, the livers were perfused in situ and hepatocytes were isolated.
Approximately 6 x 105 viable cells were seeded in culture dishes containing
coverslips and allowed to attach to the coverslips for 90 min. After the
coverslip cultures were washed, they were incubated in a medium containing 10
uCi [3H]thymidine (42 Ci/mmol) per ml for 4 hr. The cultures were washed
again and incubated in medium containing 0.5 mM cold thymidine for 14-16 hr.
The extent of UDS was assessed by autoradiography. Net grains/nucleus were
calculated as the silver grains over the nucleus minus the highest grain
count of three adjacent nuclear-sized areas over the cytoplasm. The area of
the silver grains was counted rather than the grain number, so that densely
labeled cells where silver grains were touching could be accurately measured.
Cells from control animals ranged from -2 to -6 net grains/nucleus.
Treatment of rats with dimethylnitrosamine (i.p.) at 10, 1, or 0.1 m
15
3 DRAFT
DO NOT QUOTE OR
CITE
NCV 1 2 1982
-------
yielded 36.6, 6.4, and -0.9 net grains/nucleus, respectively;
dimethylnitrosamine at 10 mg/kg (p.o.) produced 22.2 net grains/nucleus. Oral
doses of acetylaminofluorine at 50 or 5 mg/kg yielded 14.0 and 6.4 net
grains/nucleus, respectively. CC14 at 100 or 10 mg/kg (p.o.) yielded -3.2
and -5.1 net grains/nucleus, respectively. Thus, dose-related increases in
DOS were observed for dimethylnitrosamine and acetylaminofluorine, whereas
CC14 produced no response.
In this study two doses of CC14 were tested: 10 mg/kg and 100 mg/kg.
The LDso for CC14 in rats is 2800 mg/kg. It was not determined in this
study at what dose hepatic cell toxicity would occur under the conditions
used. Therefore, it is not clear that adequate doses of CC14 were tested.
Also, it is not clear whether the 2-hr time period between exposure of the
rats to 0014 and isolation of the hepatocytes was sufficient for observation
of UDS. In a study by Popp et_ £l_. (1978), in which CCl4-induced
hepatocellular changes were noted, the shortest exposure period studied was 6
hr. If Mirsalis and Butterworth had shown that the 2-hr exposure period was
sufficient for activation of the CC14 to a reactive intermediate, perhaps by
demonstrating alkylation of protein by 14-CCl3, the negative results
they obtained would be more convincing. Thus, this negative result could be
due to an inadequate dose or to an inadequate exposure time period.
In a recent abstract, Mirsalis, Tyson, and Butterworth (1982) reported
negative results in the in vivo UDS assay after exposure of rats to CC14 at
400 mg/kg, which is a larger dose than was used in the above study. However,
no details were reported in the abstract and it is not known whether even this
dose was adequate.
Craddock and Henderson (1978) carried out an in^ vivo UDS study in which
hepatocyte nuclei were isolated and then assayed for radioactivity by
16
DRAFT
DO NOT QUOTE OR
CITS
"0V 1 2 1982
-------
scintillation counting rather than by grain counting. This method may be
superior to the whole-cell grain counting procedure used by Mlrsalls and
Butter-worth (1980) as described above, particularly for chemicals that may
cause low UDS activity, because a small effect could be obscured when
background cytoplasmic grain counts are subtracted from the nuclear grain
counts. In this study, Craddock and Henderson used a CC14 dose of 4000
mg/kg, which is significantly larger than the LD$Q (2800 mg/kg). Negative
results were obtained after a 2-hr exposure. A positive effect was observed
after a 17-hr exposure. The authors suggested that this result may have been
due to secondary effects such as lysosomal damage, which may result in release
of DMA degradative enzymes.
In summary, the above in vivo UDS results suggest that CC14 may not
damage DMA. However, additional in vivo UOS studies at doses just below those
resulting in hepatotoxicity and at exposure time periods longer than 2 hr are
needed before a decision on the DMA damaging potential of CC14 can be made.
Also, use of a procedure for the UDS assay in which the radioactivity of
isolated nuclei Is assayed, rather than one in which grain counts of nuclei
are corrected for grain counts in the cytoplasm, may better allow for
detection of low levels of UDS.
SUGGESTED ADDITIONAL TESTING
Suggested additional testing falls into five categories:
1. The DNA damage studies reported by Craddock and Henderson (1978) and
Mirsalis and Butterworth (1980), which suggest that CC14 does not damage DNA
following in vivo treatment of rats, should be corroborated. Testing of
CC14 doses lower than the LD$Q (2800 mg/kg) but higher that those tested
by Mirsalis and Butterworth and for exposure time periods longer than 2 hr is
needed.
DRAFT
DO NOT QUOTE 07
CITE
NOV1 2 1982
-------
2. Additional data is needed to corroborate the Callen e_t jil_. (1980) study
in the yeast system, which utilizes an endogenous activation system and is
capable of assaying for point mutations, mitotic crossing over, and gene
conversion. The Callen et^^. study is the only yeast study found so far.
3. Additional cytogenetic testing for chromosome effects in mammalian
systems is needed before CC14 can be considered to be adequately tested for
chromosome damage. Because EDTA has been reported to decrease cytotoxicity
due to CCl4.in bacteria (Cooper and Witmer 1982), perhaps in vitro mammalian
liver cell cytogenetic assays should be carried out in the presence of EDTA so
that higher levels of CC14 could be assayed than were used in the Dean and
Hodson-Walker (1979) study. In addition, in vivo cytogenetic studies such as
the bone marrow micronucleus test, are needed.
4. Corroboration of the preliminary evidence for the weakly mutagenic
response in the Ames test reported by Cooper and Witmer (1982) is needed. The
same experimental conditions (fresh rabbit liver 39 and exposure of the
bacteria in suspension to CC14 under reduced oxygen tension in the presence
of EDTA) should be used and several concentrations of CC14 should be tested
to establish whether or not a dose-response relationship exists.
5. Studies on the ability of CC14 to reach reproductive organs and
cause germ cell mutation were not available. If results from the tests
suggested above for somatic cell mutation are positive, then studies assessing
heritable risk are needed. See Mutagenicity Risk Assessments: Proposed
Guidelines published in the Federal Register in 1980 for guidance on such
tests.
SUMMARY AND CONCLUSIONS
CC14 has been tested for its mutagenic potential in bacteria, yeast,
20 30T ^UurS OR
CITS
18
NOV1 2 1982
-------
and a mammalian cell line and for its DNA damaging potential In rat
hepatocytes when administered In vivo. Six of the seven point mutation
studies were negative. The remaining study (Cooper and Witmer 1982) was
preliminary and only suggestive of a weak mutagenic response.
In none of the negative studies was it shown that CC14 was activated or
metabolized by the exogenous S9 activation system used. Metabolism of
2-am1noanthracene or vinyl compounds (used as positive controls) is probably
an inadequate indication that the activation system can metabolize 0014.
These substances are very different from CC14. A better indication that the
activation system is sufficient for metabolism of CC14 may be to show that
it metabolizes 14CCl4 to intermediates that bind to macromolecules. It is
also conceivable that potentially mutagenic reactive intermediates of CCI^
(such as the free radical -CC^ and phosgene) are generated in the
presence of an S9 activation system but that they are too short-lived to
interact with DNA in in vitro test systems.
The Call en et aU (1980) study was designed to overcome this problem by
the use of an in vivo activation system in yeast. Positive mutagenicity and
DNA damage results were reported. However, because of the lack of
corroborative studies, the evidence is not adequate to conclude whether or not
CC14 is genotoxic. Binding studies by Rocchi jit aH_. (1973) and by Diaz
Gomez and Castro (1980a, 1980b, and 1981) indicate that metabolically
activated CC14 can interact with DNA.
Therefore, the negative genotoxicity test results that have been reported
may be due to any of three (or more) factors: (i) excessive volatilization
and escape of CC14 if appropriate precautions are not taken, (ii) inadequate
activation of CCI^ by the S9 system used to a metabolite capable of causing
mutations (e.g., 'CCH or phosgene), or (iii) inability of any reactive— >
j KIAFT j
20 SOT -riUOTS OR !
CISE I
19 NOV1 2 1982 I
i
-------
Intermediates formed to reach DNA before being scavenged by lipid and protein
(particularly under conditions of exogenous activation, such as in the Ames
test).
Additional tests should be conducted in which appropriate measures are
taken to ensure that (1_) volatilization and escape of CC14 does not decrease
exposure of the test organism or cell to levels of CC14 that are too low to
be effective, (11) metabolic activation 1s occurring, and (111) DNA 1s exposed
to the activated chemical species.
In summary, evidence described in this report is insufficient to allow any
conclusion to be made concerning the genotoxicity of CC14.. The borderline
results for binding of reactive intermediates to DNA (Rocchi et £l_. 1973, and
Diaz Gomez and Castro 1980a and 1981), the study of Call en et al_. (1980), and
the highly preliminary evidence for mutagenicity in the Ames assay (Cooper and
Witmer 1982) are insufficient evidence for genotoxicity, but are suggestive of
weak genotoxic effects and are an indication that further studies should be
done.
20
DHA7T
DO SOT QUOTE OR
CITB
NOV1 2 1982
-------
REFERENCES
Barber, E.O., W.H. Donlsh and K.R. Mueller. 1981. A procedure for the
quantitative measurement of the mutagenicity of volatile liquids in the
Ames Salmonel1 a/microsome assay. Mutat. Res. 90:31-48.
Call en, D.F., C.R. Wolf, and R.M. Phil pot. 1980. Cytochrome P-450 mediated
genetic activity and cytotoxicity of seven halogenated aliphatic
hydrocarbons in Saccharomyces cerevisiae. Mutat. Res. 77:55-63.
Cessi. C., C. Colombini, and L. Mameli. 1966. The reaction of liver proteins
with a metabolite of carbon tetrachloride. Biochem. J. 101:46c-47c.
Cooper, K., and C. Witmer. 1982. Low oxygen tension and its effects on
CC14 toxicity and mutagenicity in the Ames test (Abstract Bh-6).
Envrionmental Mutagen Society 13th Annual Meeting, Feb. 26-March 1.
13:99.
Craddock, V.M. and A.R. Henderson. 1978. De novo and repair replication of
ONA in liver of carcinogen-treated animaTs. Cancer Res. 38:2135-2143.
Dean, B.J., and G. Hodson-Walker. 1979. An in vitro chromosome assay using
cultured rat liver cells. Mutat. Res. 647329-337.
De Groot, H., and W. Haas. 1980. O?-Independent damage of cytochrome P-450
by CC14 metabolites in hepatic microsomes. FEBS Lett. 115:253-256.
Diaz Gomez, M.I., and J.A. Castro. 1980a. Covalent binding of carbon
tetrachloride metabolites to liver nuclear DNA, proteins, and lipids.
Toxicol. Appl. Pharmacol. 56:199-206.
Diaz Gomez, M.I., and J.A. Castro. 1980b. Nuclear activation of carbon
tetrachloride and chloroform. Res. Commun. Chem. Pathol. Pharmacol.
27:191-194.
Diaz Gomez, M.I., and J.A. Castro. 1981. Reaction of trichloromethyl free
radicals with deoxyribonucleic acid bases. Res. Commun. Chem. Pathol.
Pharmacol. 32:147-153.
Fishbein, L. 1976. Industrial mutagens and potential mutagens. I.
Halogenated aliphatic derivatives. Mutat. Res. 32:267-308.
Keller, E.B. and P.C. Zamecnik. 1956. The effect of guanosine diphosphate
and triphosphate on the incorporation of labeled ami no acids into
proteins. J. Biol. Chem. 221:45-59.
Kubic, V.L., and M.W. Anders. 1980. Metabolism of carbon tetrachloride
to phosgene. Life Sci. 26:2151-2155.
Masuda, Y. and T. Murano. 1977. Carbon tetrachloride-induced lipid
peroxidation of rat liver microsomes in vitro. Biochem. Pharmacol.
26:2275-2282.
21
DRAFT
DO NOT QUOTE OR
CITE
NOV 1 2 1982
-------
McCann, J., E. Choi, E. Yamasaki, and B.N. Ames. 1975. Detection of
carcinogens as mutagens in the Salmonella/microsome test: Assay of 300
chemicals. Proc. Nat!. Acad. Sci. 72:5135-5139.
Mirsalis, J.C. and B.E. Butterworth. 1980. Detection of unscheduled DNA
synthesis in hepatocytes isolated from rats treated with genotoxic agents:
an in vivo - in vitro assay for potential carcinogens and mutagens.
Carcfhogenesis 1:621-625.
Mirsalis, J.C., C.K. Tyson, and B.E. Butterworth. 1982. Induction of DNA
repair in hepatocytes from rats treated in vivo with genotoxic agents
(Abstract B1-5). Envrionmental Mutagen "Society 13th Annual Meeting,
February 26 - March 1. 13:102.
Pohl, L.R., R.V. Branchflower, R.J. Highet, J.L. Martin, D.S. Nunn, T.J.
Monks, J.W. George, and J.A. Hinson. 1981. The formation of
diglutathionyl dithiocarbonate as a metabolite of chloroform,
bromotrichloromethane, and carbon tetrachloride. Drug Metab. Disposit.
9:334-339.
Popp, J.A., S. Hisashi, and E. Farber. 1978. The protective effects of
diethyldithiocarbamate and cycloheximide on the multiple hepatic lesions
induced by carbon tetrachloride in the rat. Toxicol. Appl. Pharmacol.
45:549-564.
Reynolds, E.S. and M.T. Moslen. 1980. Free-radical damage in liver. In:
W.A. Pryor (ed.), Free Radicals in-Biology, Vol. IV, pp. 49-94. New York:
Academic Press.
Rocchi, P., G. Prodi, S. Grilli, and A.M. Ferreri. 1973. In vivo and in
vitro binding of carbon tetrachloride with nucleic acids and proteins in
rat and mouse liver. Int. J. Cancer. 11:419-425.
Shah, H., S.P. Hartman and S. Weinhouse. 1979. Formation of carbonyl
chloride in carbon tetrachloride metabolism by rat liver in vitro.
Cancer Res. 39:3942-3947.
Simmon, V.F., K. Kauhanen, and R.G. Tardiff. 1977. Mutagenic activity of
chemicals identified in drinking water. In: D. Scott, B.A. Bridges, and
F.H. Sobels (eds.), Progress in Genetic Toxicology, pp. 249-258. New
York: Elsevier/North-Holland Biomedical Press.
Sipes, G.I., G. Krishna, and J.R. Gillette. 1977. Bioactivation of carbon
tetrachloride, chloroform and bromotrichloromethane: Role of cytochrome
P-450. Life Sci. 20:1541-1548.
Trudell, J.R., B. Bosterling, and A.J. Trevor. 1982. Reductive metabolism of
carbon tetrachloride by human cytochromes P-450 reconstituted in
phospholiped vesicles: Mass spectral identification of trichloromethyl
radical bound to dioleoyl phosphatidylcholine. Proc. Natl. Acad. Sci.
79:2678-2682.
Uehleke, H., H. Greim, M. Kramer, and T. Werner. 1976. Covelant binding of
haloalkanes to liver constituents, but absence of mutagenicity qr> bacteria. •
in a metabolizing test system. Mutat. Res. 38:114. * DQ I;C°™TE OR
22
CITS
MOV 1 2 1932
-------
Uehleke, H., H. Greim, M. Kramer, and T. Werner. 1977. Metabolic
activation of haloalkanes and tests in vitro for mutagenicity.
Xenobiotica. 7:393-400.
Vainio, H., M.G. Parkki, and J. Marniemi. 1976. Effects of aliphatic
chlorohydrocarbons on drug-metabolizing enzymes in rat liver in vivo.
Xenobiotica. 6:599-604.
23
CO SOT Q3JSX2 OR
CITS
NOV 1 2 1982
------- |