SUMMARY OF THE U.S. EPA COLLOQUIUM
ON A
FRAMEWORK FOR HUMAN HEALTH RISK ASSESSMENT
Colloquium #2
Preparedfor:
U.S. Environmental Protection Agency
Risk Assessment Forum
401 M Street SW.
Washington, DC 20460
Contract No. 68-D5-0028
Work Assignment No. 3-98
Prepared by:
Eastern Research Group
110 Hartwell Avenue
Lexington, MA 02421 -313 6
September 1998
-------
SUMMARY OF THE U.S. EPA COLLOQUIUM
ON A
FRAMEWORK FOR HUMAN HEALTH RISK ASSESSMENT
Colloquium #2
Prepared for:
U.S. Environmental Protection Agency
Risk Assessment Forum
401M Street SW.
Washington, DC 20460
\
Contract No. 68-D5-0028
Work Assignment No. 3-98
Prepared by:
Eastern Research Group
110 Hartwell Avenue
Lexington, MA 02421-3136
September 1998
-------
NOTICE
This report was prepared by Eastern Research Group, Inc. (ERG), an EPA contractor, as a general
record of discussions during the U.S. EPA Colloquium on a Framework for Human Health Risk Assessment
(Colloquium #2). As requested by EPA, this report captures the main points and highlights of discussions
held during plenary sessions. The report is not a complete record of all details discussed nor does it
embellish, interpret, or enlarge upon matters that were incomplete or unclear.
-------
CONTENTS
Page
SECTION ONE BACKGROUND 1-1
Developing a Framework for Human Health Risk Assessment 1-1
The September 1997 Colloquium ' 1-2
The June 1998 Colloquium 1-2
SECTION TWO OPENING PLENARY SESSION 2-1
Welcoming Remarks 2-1
Goals of the Human Health Risk Assessment Framework 2-1
Introduction to Case Studies and Colloquium Issues and Charge to Breakout Groups 2-2
Questions/Comments 2-3
SECTION THREE BREAKOUT GROUP DISCUSSIONS
ON CASE-SPECIFIC QUESTIONS 3-1
Ethylene Thiourea 3-2
.Ethylene Oxide 3-5
Trichloroethylene 3-7
Vinyl Acetate 3-13
SECTION FOUR FINAL PLENARY SESSION 4-1
Lessons Learned and Their Applications to the Development of a Human Health
Risk Assessment Framework 4-1
Overview/Next Steps 4-4
-------
APPENDICES
APPENDIX A - White Paper
APPENDIX B - Participant and Observer Lists
APPENDIX C - Case Studies
APPENDIX D - Charge to the Participants
APPENDIX E - General Questions for Plenary Session
APPENDIX F - Breakout Group Assignments
APPENDIX G - Agenda
-------
SECTION ONE
BACKGROUND
Developing a Framework for Human Health Risk Assessment
The U.S. Environmental Protection Agency (EPA) has recognized the need to develop a framework
for human health risk assessment that puts a perspective on the approaches in practice throughout the
Agency. Current human health risk assessment approaches are largely endpoint driven. In its 1994 report
entitled Science and Judgment in Risk Assessment, the National Research Council (NRC) noted the
importance of an approach that is less fragmented, more consistent in application of similar concepts, and
more holistic than endpoint-specific guidelines. Both the NRC and EPA's Science Advisory Board have
raised a number of issues for both cancer and noncancer risk assessments that should be reconsidered in light
of recent scientific progress. EPA has recognized the need to develop a more integrated approach. In
response, the Agency's Risk Assessment Forum (RAF) has begun the long-term process of developing a
framework for human health risk assessment.
The framework will be a communication piece that will lay out the scientific basis, principles, and
policy choices underlying past and current risk assessment approaches and will provide recommendations for
integrating/harmonizing risk assessment methodologies for all human health endpoints.
As an initial step in this process, the RAF formed a technical panel in April 1996. An Issues Group
(Gar}' Kimmel and Vanessa Vu, co-chairs; Jane Caldwell; Richard Hill; and Ed Ohanian) was formed, and
this group developed a white paper, entitled Hitman Health Risk Assessment: Current Approaches and
Future Directions, to provide an overall perspective on the issue (see Appendix A). The RAF peer-reviewed
the white paper in February 1997. Its purpose is to serve as a basis for further discussion on current and
potential future risk assessment approaches. The paper highlights a number of issues regarding the Agency's
risk assessment approaches and their scientific basis, primarily with respect to dose-response and hazard
assessment. The paper discusses the scientific basis for cancer and noncancer risk assessment, including
differences and similarities. It also identifies knowledge/information gaps and areas where more work is
needed. /
As part of the continuing effort to develop a human health risk assessment framework, the RAF
organized a colloquium series, consisting of two internal colloquia. The colloquia brought together EPA
scientists for a dialogue on various scientific and policy issues pertaining to EPA's cancer and noncancer risk
assessment approaches. The first colloquium, held on September 28 and 29, 1997, in Arlington, Virginia,
focused on the role of mode of action information in re-examining and developing new risk assessment
approaches. The second colloquium, held on June 2 and 3, 1998, in Bethesda, Maryland, explored the more
quantitative aspects of mode of action, including dosimetry, dose-response relationships, and low-dose
extrapolation methods.
The overall goal of the first two colloquia was to provide Agency scientists an opportunity to share
perspectives on the role of mode of action in shaping future human health risk assessment approaches. The
RAF invited a cross-section of senior Agency scientists (from headquarters, Research Triangle Park,
Cincinnati. Las Vegas, and the regions) to participate in these discussions. As the Agency moves forward to
develop this framework, additional colloquia are anticipated, as well as workshops to gather input and
perspectives from scientists outside EPA.
1-1
-------
The September 1997 Colloquium
During the first colloquium, Agency scientists discussed the current standard default approach for
cancer and noncancer risk assessment, and the advantages and limitations of departing from this approach in
light of new information pertaining to chemical mode of action. The primary topics deliberated by the group
included defining mode of action, evaluating what events are critical, formulating dose metrics, determining
when enough information exists to support new risk assessment approaches, and strategizing on how mode of
action information can be effectively and systematically used in low-dose extrapolations. Group discussions
addressed general risk assessment issues and the overall use of mode of action in risk assessment. Case study
discussions followed. The colloquium's final session included discussions on "critical harmonization issues"
and quantitative dose-response issues to be covered at the second colloquium.
The "Summary of the U.S. EPA Colloquium on a Framework for Human Health Risk Assessment:
Colloquium #1," dated November 24, 1997, provides a detailed account of the outcome of the first
colloquium. A brief overview of the key results of the September 1997 colloquium was provided at the
opening session of the second colloquium (see Section Two).
The June 1998 Colloquium
Fifty EPA scientists and a small group of observers gathered for the second colloquium in June 1998
(see participant and observer lists in Appendix B). The 2-day colloquium focused on the role of mode of
action information in developing descriptive quantitative models, applicable to a variety of needs for carrying
t out a risk assessment. Mode of action and harmonization issues were discussed in the context of four
chemical-specific case studies: ethylene thiourea, ethylene oxide, trichloroethylene, and vinyl acetate.
Prior to the June colloquium each participant received one of the four case studies (Appendix C),
including case-specific questions; a "charge" (Appendix D); and a list of general questions developed to
guide colloquium discussions (Appendix E). During the colloquium, each participant was assigned to a
breakout group to discuss assigned case studies. Appendix F includes a list of breakout group assignments,
including the names of breakout group chairs and rapporteurs. As with the first colloquium, the RAF sought
to ensure a mix of expertise and Agency representation in making group assignments.
After opening remarks were made, the first day of the colloquium was devoted to breakout group
discussions on the case studies. During the second day, in plenary session, breakout group members
presented their key findings.. The closing plenary session involved an exchange of ideas on lessons learned
from the colloquia series. Participants discussed next steps in developing a risk assessment framework in
light of uncertainties and data gaps. The colloquium agenda is provided in Appendix G.
The following sections of this report highlight the outcome of the June 1998 colloquium. Section
Two presents opening statements. Section Three captures the breakout group discussions on the case studies
and Section Four presents highlights of the closing plenary session.
1-2
-------
SECTION TWO
OPENING PLENARY SESSION
Welcoming Remarks
William Wood, Risk Assessment Forum, EPA
Dr. Wood welcomed all participants, many of whom were at the first colloquium. He explained that
this RAF project was directed at developing a framework on integrating approaches for cancer and noncancer
risk assessment. Toward that end, the RAF workgroup's goal is to couple the outcome of the health effects
colloquia series with Agency work on the final cancer guidelines in setting the course for how EPA will
conduct future risk assessments. The outcome of the colloquium will also provide guidance for future
research. Dr. Wood encouraged the input and active participation of Agency scientists throughout the second
colloquium.
Dr. Wood acknowledged the hard work of the organizing committee whose members include Gary
Kimmel (co-chair), Vanessa Vu (co-chair), Kim Hoang, Annie Jarabek, Jennifer Seed, Gina Pastino, and
Wendy Yap. The colloquium participants then introduced themselves and their affiliations.
Goals of the Human Health Risk Assessment Framework
Vanessa Vu, National Center for Environmental Assessment, EPA
Dr. Vu reviewed the overall goals of the framework project, accomplishments to date, additional
short- and long-term plans, and the structure and charge of the second colloquium. She explained that the
Agency intends to develop a framework to accomplish the following:
Develop a conceptual piece to communicate a risk assessment approach (for the Agency and
public at large)
Layout past and current approaches.
Recommend approaches in integrating/harmonizing risk assessment approaches for all
endpoints
The major elements of the anticipated framework, she explained, include using mechanistic information to
enable integrating risk approaches for different endpoints, considering a range of default approaches, and
applying appropriate uncertainty factors.
2-1
-------
Obtaining buy-in and input from Agency scientists, Dr. Vu emphasized, is very important, especially
in the development stage of the framework. The RAF, therefore, has or plans to take the following steps:
1. Development of a white paper. The white paper, a "perspective" piece, was developed to
identify key issues related to current risk assessment approaches and harmonization. The
papers focuses on issues related to hazard and dose-response assessment and presents the
scientific basis for assessing cancer and noncancer risks. It identifies uncertainties in the
existing risk assessment process and areas requiring further guidance and research.
2. Organization of the colloquia series. The RAF organized the colloquia series to enable
Agency scientists to discuss white paper issues and to provide recommendations on the
approach of the framework. Agency scientists participating in the colloquia series were
charged with discussing scientific and policy issues associated with developing a more
consistent/holistic approach to risk assessment. During the first colloquium, discussions
centered on the significance of qualitative implications of mode of action for various risk
assessment endpoints. The second colloquium was designed to foster further qualitative
discussions and initiate discussions on quantitative issues associated with the application of
mode of action information (e.g., low dose extrapolation models).
3. Draft the framework. Based on the outcome of the colloquia series, the Agency anticipates
preparing a draft framework document. It is anticipated that the framework document will
undergo expert review, leading to future workshops and review by the Science Advisory
Board.
Dr. Vu briefly summarized the outcome of the first colloquium. During Colloquium #1, participants
developed a common appreciation for terminology and the role of mode of action in risk assessment. While
Colloquium #1 participants recognized that strictly defining mode of action was difficult, mode of action was
broadly defined as "knowledge of the series or sequence of biological events that influence the final toxic
outcome." The group agreed that the traditional use of threshold/nonthreshold approaches may no longer be
applicable in light of new scientific knowledge on mode of action. The group recommended greater use of
mode of action information when extrapolating from high to low dose, across species, and across routes of
exposure, as well as studying aggregate risk from chemicals that may have common mode of action.
Colloquium #1 case studies enabled participants to begin to explore new approaches to low-dose
extrapolation and evaluate commonalities across endpoints by reviewing toxicologic and mechanistic
information for five chemicals. Participants agreed that issues related to commonalities across toxicities
needed more emphasis. Continued development of the framework and future colloquia/workshops were
encouraged to pursue the complex issues associated with harmonization of risk assessment approaches.
Introduction to Case Studies and Colloquium #2 Issues and Charge to Breakout Groups
Dr. Vu explained that the purpose of the case study exercise at the second colloquium was to foster
more in depth discussions on critical issues related to mode of action and its role in harmonizing
cancer/noncancer risk assessment. Dr. Vu emphasized that the intent of the case studies was not to perform
chemical-specific risk assessments. Dose-response and mechanistic data were provided to help participants
2-2
-------
explore important factors related to developing descriptive quantitative models. Case-specific questions were
provided to guide discussions and to promote deliberations on harmonization issues.
Lastly, Dr. Vu acknowledged the efforts of the issues group, organizing committee, RAF (Bill Wood,
Jeanette Wilsey, and Carole Kimmel), and Eastern Research Group in helping to organize and coordinate the
activities of the workshop. Dr. Vu also thanked participants and observers for taking part in the colloquia
series.
Questions/Comments
The group briefly discussed possible limitations of the case studies. Points raised by participants
include the following:
Chemical-specific information presented in the case studies may not be 100 percent complete
or correct.' One participant questioned whether discussions should be limited to information
provided in the case studies or if new information could be introduced.
The group recognized that it would be impossible to present a complete data set for one- or
two-day discussions on a particular chemical. It was re-emphasized that participants were
not performing full-blown risk assessments on case-study chemicals, but rather raising and
evaluating case-specific issues related to more scientifically sound approaches to evaluating
human health risks. While it was agreed that scientists should introduce pertinent data during
the breakout sessions, it was also recognized that because of time constraints it is not
possible, nor necessary, to consider every chemical-specific detail. The ultimate purpose of
the case study exercise, the group was reminded, was to determine the best use of mode of
action information and how to generate the most credible risk assessment.
One participant questioned how the group should approach the issue of multiple modes of
action during case study deliberations, expressing concern that the group may try to "force
fit" a single mode of action for multiple endpoints.
Multiple modes of action should be considered in terms of their relative contribution to
pathogenesis. The intent of the case study exercise was to evaluate whether different
endpoints should be treated differently when a common mode of action has been identified,
not necessarily to identify' a single mode of action.
2-3
-------
SECTION THREE
BREAKOUT GROUP DISCUSSIONS ON CASE-SPECIFIC QUESTIONS
The first day of the colloquium was dedicated to breakout group discussions on the following four
case studies (See Appendix C).
Ethylene Thiourea (ETU)
Ethylene Oxide (EtO)
Trichoroethylehe (TCE)
Vinyl Acetate (VA)
The case studies include a summary of key human and animal studies and describe primary acute and chronic
effects. Depending on the chemical, the case study describes portal-of-entry effects; systemic toxicity;
reproductive and developmental toxicity; neurotoxicity; mutagenicity; and carcinogenicity. The case studies
also present pertinent dose-response, pharmacokinetic, and mode of action (MOA) information.
Each breakout group deliberated case-specific questions (included within each case study), but, in
general, the following questions capture the key issues discussed by each group.
1. Given what is known about MOA, are there commonalities among endpoints that would be
useful for quantitative analyses? For which endpoints should a common quantitative
analysis be conducted? For which endpoints should a separate analysis be conducted?
2. What additional information would be useful for quantitative analysis?
3. In the absence of this information, are any of the available data sets useful for quantitative
analysis?
4. Are dose and duration of exposure important considerations? If so, for which endpoints and
how should they be handled?
5. In the absence of case-specific physiologically-based pharmacokinetic (PBPK) models, how
should dose be adjusted for extrapolation to humans? Does choice of a specific endpoint
influence this decision? Why or why not?
If a PBPK model is available, which dose metrics should be considered for the dose-response
analysis?
6. What response/endpoint(s) would be useful for dose-response modeling in the observable
range? Does MOA information influence this choice?
7. What quantitative method is recommended for low level exposures? Does this vary for
3-1
-------
different toxicities? Does MOA information influence the choice of models?
If a reference dose (RfD), reference concentration (RfC), or margin of exposure (MOE) were
to be calculated, does MOA information influence the choice of uncertainty factors or
influence uncertainties about data gaps?
The sections below summarize the main points discussed during breakout sessions, as captured by
the group rapporteurs and presented in plenary session. Vicki Dellarco, Kerry Deerfield, Vanessa Vu, and
Arnold Kuzmack presented the breakout group reports for ETU, EtO, TCE, and VA, respectively.
Ethylene Thiourea
In reviewing the ETU case study, the group considered the adverse health effects associated with
target organs/responses, common modes of actions across different responses, dose-related increases,
exposure duration issues, critical windows of exposure, and the relevancy of animal data to humans. The
group's responses to case-specific questions are provided below.
Given what is known about MOA, are there commonalities among endpoints?
The group identified the following ETU "targets:" thyroid, pituitary, liver, embryo/postnatal, and
central nervous system (CNS). The group described the following three potential modes of action likely to be
responsible for the effects in these target systems:
1. Thyroid/pituitary: In the rat, high concentrations of ETU result in decreased T3 and T4 and
increased TSH levels. The severity of hyperplasia increases with dose and possibly with
duration. These changes in T3/T4 and TSH levels are associated with thyroid hyperplasia
and tumor development in the thyroid (adenomas and carcinoma). These events can
eventually lead to pituitary tumors if substantial. Based on case study information,
mutagenicity or a direct DNA reactive mechanism does not seem to be a major influence on
tumor development. Perturbances of the pituitary-thyroid homeostasis is the essential event
leading to tumor development (i.e., an anti-thyroid MOA).
Some developmental effects (related to brain development in late gestational/postnatal
periods) are presumed to by thyroid-mediated.
2. Liver: A separate MOA appears responsible for liver effects. Effects appear to be
metabolite-dependent (FMO) and species-specific.
/
3. Non-thyroid developmental effects: Seen primarily in the rat, CNS malformations result
from necrosis of neuroblasts driven by ETU (parent compound). These effects are not
considered to be thyroid-mediated. Effects are species-specific.
3-2
-------
The group concluded that common modes of action for cancer and certain noncancer (e.g., CNS) endpoints
are associated with the disruption of the thryoid/pituitary homeostasis. This knowledge enables one to use
precursor events (e.g., changes in T3, T4, and TSH; increases in thyroid hyperplasia) instead of frank
toxicologic effects in protecting for different outcomes. No conclusions could be reached on the reversibility
of responses, however, because of the lack of data.
A question was raised following this discussion as to whether or not ETU exposures led to total
endocrine disruption and whether the pituitary should be considered separately. Another participant
questioned whether the group considered the relation of liver effects to thyroid/pituitary effects. It was noted
that data were not available to suggest any such relation.
What approaches should be considered for quantitative analysis?
Upon consideration of available dose-response data, the group suggested different approaches for the
quantitative analyses of the three identified MOAs. For thyroid/pituitary events and hyperplasia events,
effects on thyroid hormones should be used as indicators of both cancer arid noncancer endpoints. Given the
understanding of MO A in the th\Toid. the group suggested using a nonlinear approach for low-dose
extrapolation. For liver effects, the group noted that, in the absence of quantitative information and a full
understanding of MOA, the default linear approach should be used. The group commented, however, that
this approach might be overly conservative—the group emphasized the need to point out data set
uncertainties and the possibility 'that effects may be species-specific and not relevant to humans. For
developmental effects, the group suggested using the default nonlinear approach, but data were available to
also enable some benchmark modeling. .
What additional information \\ould be useful for quantitative analysis? What are the research
needs?
In general, the breakout group agreed that more comparative metabolism information (within and
across species) would be especially helpful in further evaluating MOA questions and the relevance of existing
data to humans. Response-specific information needs to include the following:
i
Thyroid: Because thyroid hormones are a good biomarker and evidence exists that there is age-
dependent susceptibility, it would be helpful to examine prenatal/early postnatal hormone levels. In
addition, obtaining more dose-duration information would be helpful in studying the issue of
reversibility. Comparative metabolism data (tissue distribution) between humans and rodents would
be helpful to better understand species differences.
Liver. More information is needed specific to mouse metabolism. Comparative metabolism studies
on FMO are needed.
NonthyroidMalformation: More comparative metabolism data are needed to study differences in
responses between humans and rats.
-------
Are dose and duration exposure important considerations?
The breakout group considered patterns of exposure and critical windows of susceptibility.
Responses in the thyroid/pituitary (severity of hyperplasia) appear to be dose limited and may be dependent
on duration. Not enough information is available to assess dose/duration considerations for liver and
developmental effects. Thyroid/pituitary and developmental effects were observed at similar ETU doses.
Dose was species-dependent for liver effects, which is an example of why more species-specific metabolism
data are needed.
In the absence of a PBPK-model, how should dose be adjusted for extrapolation to humans?
Does choice of a specific endpoint influence this decision? What quantitative method is recommended for
low level exposures? Does this vary for different toxicities? DoesMOA information influence the choice
of models?
Although no single extrapolation method was recommended (e.g., lack of an interspecies adjustment
versus using a scaling factor of body weight to the 3/4 power), the group strongly agreed that the approach
should be the same for cancer and noncancer endpoints in the thyroid/pituitary.
What endpoint(s) -would be useful for dose-response modeling in the observable range? Does MOA
information influence this choice?
The group agreed that MOA is relevant to thyroid/pituitary responses. It plays less of a role in
developing models for liver and developmental effects.
If an RfD were to be calculated, does MOA information influence choice of uncertainty factors or
influence uncertainties about data gaps?
Yes. The group reiterated, however, that more comparative data between rats and humans are needed
before fully answering this question. Qualitatively, the group agreed that uncertainty factors should be
applied in the same way for cancer and noncancer endpoints. In comparing RfD and margin of exposure
(MOE) approaches, the group agreed that, conceptually, the uncertainty factors applied are similar, hi
practice, however, they could be applied differently because the RfD approach is more compartmentalized
and the MOE approach involves more scientific judgment/interpretation. This issue, therefore, warrants
further studv and careful consideration.
3-4
-------
Ethylene Oxide
The breakout group initiated their evaluation of EtO by preparing a matrix of observed effects. EtO
induces a variety of effects including irritation, hematoxicity, neurotoxicity, reproductive and developmental
toxicity, and cancer. Group discussions focused primarily on the latter three. The group provided the
following responses to case-specific questions: .
Given what is known about MOA, are there commonalities among toxicities that would be useful for
quantitative analyses? Is there any reason to propose different mechanisms for the various endpoints?
Based on available data, two plausible MOAs exist for EtO: the formation of protein adducts and the
formation of DNA adducts. EtO distributes readily and is direct acting (no metabolite formation).
Distribution is even throughout the body. Although it is highly reactive (e.g., hemoglobin binding,
glutathione binding), free EtO distributes to target tissues. EtO binds to macromolecules (specific amino
acids in protein) and forms specific DNA adducts (e.g, 7-hydroxyethylquanine). These two mechanisms are
probably not mutually exclusive. The mechanisms related to neurotoxic outcomes are not completely
understood; these effects are not fully explained by DNA adduct formation, and may relate primarily to the
binding of EtO to protein.
The group categorized the endpoints and asked whether common MOAs exist.
Cancer: Tumors have been observed in multiple sites in animals (hematopoietic, brain, forestomach,
lung, ovary, lymph). In humans, epidemiologic studies suggest a link between EtO and hemopoietic
cancers. Because tumors appear in multiple locations, there is likely a common MOA for most of
these cancers and that is related to DNA binding mechanisms. Forestomach cancers, however,
appear to result from a local irritant effect, although this effect may be enhanced by the genotoxic
action of EtO.
Reproductive/Developmental Effects: Observed effects include spontaneous abortion, zygotic death,
lethality/viability, litter size, implant loss, and malformations. Dominant lethality appears to result
from the formation of DNA adducts. While insufficient data exist for all of these endpoints, the
group agreed that a common MOA probably exists for most reproductive/developmental endpoints. .
Data suggest that MOA is similar in animals and humans for tumors, but unknown for
developmental effects.
What additional information would be useful for quantitative analysis of the various toxicities? (For
example, is consideration of the entire spectrum ofmutational changes, such as the induction of gene
mutations, structural chromosome mutations, and numerical chromosome alterations important?)
Several data needs were identified.
• For mutagenic effects, existing information on point mutations needs to be considered. The
case study concentrated on chromosome breaks (translocation) data.
3-5
-------
Information on the shape of the curve at low doses. For example, is it linear or nonlinear?
Are DNA adducts formed at low levels? This is a research need.
Additional information on the causality of different endpoints.
Cell proliferation information at all dose levels.
Information on background rates. What is the background load (endogenous EtO)?
Information on exposures to other agents that may have the same MOA or make one more
susceptible to a MOA.
What quantitative, method is recommended for.low level exposures? Does this vary for different
responses? Does MOA information influence choice of models?
The group proposed the same approach for both cancer and developmental/reproductive effects
because the MOA suggests that both effects are related to the formation of DNA adducts. If one assumes
linear behavior, then a linear quantitative method is appropriate for low dose extrapolation because of the
mutagenic properties of EtO. The group, however, did discuss MOE and possible nonlinear approaches
because the data suggest that protein binding and DNA adduct formation may not be linear. One participant
noted that data on heritable effects versus dominant lethal effects suggest that a two-hit model and nonlinear
dose response may exist. The overall impression of the group was that MOE eliminates the theoretical
argument over linear versus nonlinear dose-response relationships and focuses on MOA. MOE would
therefore be a viable approach to bring to the risk manager. In general, the MOA for all effects is probably
related specifically to the electrophilic nature of EtO, and the ultimate action would be dependent on timing
and duration of exposure, where and to what it binds, etc.
The question on linear versus nonlinear dose response triggered a fairly lengthy discussion among the
plenary group. General and EtO-specific issues raised are highlighted below:
Because of the limited dose numbers in the NTP study, it is difficult to study linearity.
Adduct-formation is not the only factor to influence the shape of the dose-response curve.
Although adduct formation may be considered a linear response, a certain level may need to
be reached before a toxic outcome is observed. If adducts are easily repaired, a nonlinear
response may in fact be observed. What is happening beyond adduct formation needs to be
considered and is an argument for using the MOE approach.
"Toxicity" needs to be defined. Traditionally, toxicity was defined as an observable effect
(e.g., a tumor or malformation). Now with activities at the cellular level being considered
(e.g., biochemical changes or adduct formation), toxicologists need to agree on what the
"toxic endpoint" is.
One participant noted a definition of toxicity by Doull (of Cassarett and Doull): toxicity is
not achieved until the first "irreversible step" is observed. Several others disagreed citing
3-6
-------
ethanol exposure as an example where reversible effects still result in "toxicity."
Furthermore, RfDs have been developed based on nontoxic reversible effects. Doull's
definition, therefore, may not be relevant to these discussions.
It is important to study the nature of the lesions before deciding on a linear versus nonlinear
approach.
Are dose and duration of exposure important considerations? If so, which responses and how should they
be handled?
Very little dose rate information is available for most endpoints, but the group agreed that it is an
important consideration. For example, in a study of dominant lethality, dose and duration were found to be
extremely important when considering the effects of EtO.
In summary, it was agreed that EtO presents a good case for quantitatively treating different
endpoints similarly based on MOA. Although no specific approach was recommended, many felt that an
integrated MOE approach for each of the effects would provide risk managers with useful information.
Trichloroethylene
TCE. the group agreed, was one of the more complex case studies because of the variety of systems
affected and effects produced. It is further complicated because of the involvement of and uncertainties
associated with the metabolites. The group reviewed TCE effects and its MOA in several target systems, but
focused on effects in the liver, lung, and kidney.
Both the "minor" and "major" metabolic pathways for TCE were described (see case study figure in
Appendix C). The group identified the role of metabolites in mediated TCE-induced toxicities and
highlighted the relative species reactivity of the metabolites, as follows:
Effects
liver
lung •
kidney
Metabolites
TCA, DCA
Chloral
DCVC
Species reactivity
mouse>rat>humans
mouse>rat
rat>mouse>human
TCA = trichloroacetic acid
DCA = dichloroacetic acid
DCVC = s-1.2-dichlorovinyl cysteine
The breakout group summarized the effects of TCE in the liver, lung, and kidney, highlighting cross-
species and general dose duration differences. These discussions are summarized in Table 1.
3-7
-------
Table 1. Breakout Group Summary of TCE Effects
SPECIES
EXPOSURE
EFFECTS
Liver
human
rat
mouse
acute/high
occupational
acute/subchronic to high level
chronic/lower level
acute/subchronic to high level
chronic/relativelv lower level
liver failure/necrosis
impaired li ver function
some evidence of risk of cancer of the liver and
the biliary duct
enlarged liver, hypertrophy, necrosis
enlarged liver
enlarged liver, hypertrophy, necrosis
hepatomegaly, hypertrophy, tumors
Kidney
human
rat
mouse
occupational
acute exposure to high level
chronic to lower level
acute/chronic to high level
mild renal function changes
suggestive evidence of kidney cancer
nephropathy
increased kidney weight
mild karyomegaly
tumors
nephrotoxicity
no tumors
Lung
human
rat
mouse
acute/chronic
acute
chronic
no reported effects
ho effects
cytotoxicity to Clara cells
lung tumors
3-8
-------
The group briefly discussed lympho/hematopoietic, reproductive/developmental, and CNS effects. TCE-
related effects on the lympho/hematopoietic system include excess non-Hodgkin lymphoma in humans,
lymphoma in exposed mice (via inhalation), and effects on the spleen in rats and mice. The group noted
consistency across species. Inconclusive/conflicting evidence exists related to TCE-induced
reproductive/developmental effects in humans. Eye and cardiac malformations have been observed in rats
exposed in utero. Effects on sperm, implantations, and litter size have been observed in mouse reproductive
studies. CNS effects are reported in humans exposed to high levels of TCE (acutely) and in occupational
settings as well as in rats and mice exposed acutely, subchronically, and chronically.
Having highlighted key effects, the group then answered case-specific questions.
What seems to be the series of events leading to each observed toxic response? Are there any reversible
steps in the process? Can an irreversible step be identified in each process? Given that TCE-induced
toxicities are mediated through metabolites, are there common biological responses across toxicities that
would be useful for quantitative analyses?
The group developed schematics depicting key events in the liver, kidney, and lung (see Figures 1,2,
and 3). Discussions centered around whether common modes of action are present for different toxic
responses.
Figure 1. TCE MOA in the Liver
TO-
>- TPA
PPAR
V
\ (mouse) ^
[ROS]
1 MOA applies to human
tumors
hypertrophy
necrosis
hyperplasia
3-9
-------
Figure 2. TCE MOA in the Kidney (Rat)
TCE
GST
p-lyase
i
[RS = chloro thiokene]
lipid peroxidation
y
cytotoxicity
1
proximal tubular damage
Figure 3. TCE MOA in the Lung (mouse)
TCE
T
99
Chloral
mutagen
accumulation
in Clara cells
Cytotoxicity
vacuolization of Clara cells
perivascular inflammation
DNA damage
T
.cell proliferation
tumors
3-10
-------
The group also identified data gaps.
Liver: For the liver, the group emphasized that cellular proliferation appears to be the common event
leading to both tumors and liver toxicity. The MOA is relevant to humans based on available data.
Because quantitative information on cellular proliferation is lacking, it is not known whether
reversible steps exist. The specific steps leading to tumors and liver toxicity are not clear. One group
member noted that speculation exists as to whether DCA is a promoter or an initiator.
Kidney. While the metabolite DCVC is common to the two endpoints (i.e., tumors and proximal
tubular damage), a common MOA is not observed for these endpoints.
Lung: TCE action in the lung of mice was described. Both cytotoxicity and DNA damage appear to
be the result of the accumulation of chloral in the Clara cells. Because of many unknowns, no specific
common biological events could be identified to account for either TCE-induced tumors or toxicity in
the lung.
Which of the above-selected responses is most relevant to humans regarding specificity (response
concordance) and sensitivity (dose range of response)?
Liver and kidney MOA and responses in test animals are relevant to humans. Lung responses,
however, are not. Data are not sufficient to judge sensitivity of response. Epidemiologic data provide good
qualitative information but do not enable quantification. Animal studies show more tumors in the liver versus
the kidney following TCE exposure.
What additional information would be useful for quantitative analysis? .
The group stressed that obtaining more dose-response information on cell proliferation was critical.
No dose-response curve is available. Cell proliferation data are needed for initiated versus noninitiated cells.
A labeling index study for age range is also needed.
Are dose and duration of exposure important considerations? If so, for which toxicity and how should
they be handled?
Dose and duration appear to be important in the liver and the kidney. In animal studies, liver tumor
response depends on dose, but not enough is known to specifically answer the dose/duration question. Not
enough data are available to answer this question for the kidney. In addition, more information is needed on
dose/duration issues in humans.
5-11
-------
What response(s) 'would be useful for dose-response modeling in the observable range for each toxicity?
How does MO A information influence this choice? Given the availability of the PBPK models, what would
be the appropriate dosimeters for the toxicity observed in the liver, lung, and kidney? Which quantitative
models should be used for the observed data?
Dose-response modeling could be considered for liver and kidney responses. Cell proliferation in the
liver is the preferable response choice, but because of high background and species variability, coupled with
the lack of quantitative data, it may be problematic. PBPK models could be used to estimate internal TCE
dose. More information is needed, however, relating TCE to its metabolites so that an internal dose of
metabolites can be obtained.
Given what is known about theMOAfor each toxicity, what quantitative approach would be recommended
for characterizing risk associated with low level exposures (i.e., beyond the observable range} for each
toxicity?
The group focused on the liver response for this question. Opinions varied regarding the best
quantitative approach to take in light of available data. Although no one approach was recommended, it was
agreed that applying a biologically-based dose response (BBDR) model would be the ideal choice. The group
considered two scenarios: (1) assume quantitative cell proliferation data are available, and (2) assume
quantitative cell proliferation data are not available.
Assuming quantitative cell proliferation data were available, the group considered linear and MOE
approaches. Half of the breakout group felt an MOE approach was preferable because it gives more
consideration to science and nearly an equal number felt it is really a polity choice. One individual preferred
a linear approach because it is more conservative and because the threshold for lifetime exposure is not
known.
In the absence of cell proliferation data or a BBDR model (where tumor and liver toxicity would be
considered as the responses), the group was again divided as to what approach is most appropriate. The
following quantitative approaches were proposed, with the group divided equally on each of the three options.
1. Status quo. Several individuals supported using default approaches (i.e., linear for tumor
and an RfD/RfC for noncancer effects). These individuals felt resorting to the existing
models was more conservative in light of data gaps.
2. Same approach for both responses. Because of common MOA, others felt it was more
appropriate to use the same approach for both cancer and noncancer outcomes. Both linear
and MOE approaches were considered. The overall preference of the group was an MOE
approach because of observed receptor-threshold effects. One member noted that, in the
absence of data, no compelling reason exists to assume a linear curve at low doses; he
emphasized, however, that all endpoints should be considered and the most sensitive should
be used to select the RfD/benchmark dose.
3. Policy choice.
5-12
-------
Presentation of these choices resulted in lively discussions both in breakout and plenary sessions.
The group conveyed the following general points about choosing an appropriate quantitative approach.
How much information is enough to support a decision to choose a nondefault approach?
Because of the uncertainties in most data sets, opinion will vary widely.
In the case of TCE, one participant questioned how one could conclude simply from the
evidence of cell proliferation whether a threshold or nonthreshold response existed. He
provided the dioxin example where several factors led to identifying a threshold. He could
not accept the threshold concept for the complicated TCE story.
The group did not discuss other sensitive noncancer effects of TCE (e.g., neurotoxic effects).
In focusing on the noncancer effects in the liver (cell proliferation), a potentially more
sensitive outcome in another system (neurotoxic) may be overlooked.
If an RfD orMOE were to be developed, which factors should be considered to account for uncertainties
in risk assessment?
The group agreed that the following uncertainty factors should be considered as common to both RfD
and MOE approaches:
— intraspecies differences
— interspecies differences
— nature of response
— steepness of the dose-response curve at point of departure region
— lack of understanding
Further discussion on uncertainty factors was held in the final plenary session and is summarized in Section
Four of this report.
Vinyl Acetate
It was noted that the action of VA is unique from the chemicals evaluated in the other case studies in
that it exhibits effects at the portal-of-entry (upper respiratory tract). There is a spatial specificity of lesion
location, with most effects concentrated in the olfactory region of the rat. Li mice, the location of the lesions
is consistent with air-flow patterns and tissue-specific enzymes. Case-specific questions varied slightly,
therefore, to foster discussions on this unique aspect of VA.
The group reviewed the established metabolic pathway for VA. Carboxylesterase catalyzes the
initial hydrolysis of VA to vinyl alcohol and acetic acid (AA). Vinyl alcohol rearranges to acetaldehyde
(AAld) which aldehyde dehydrogenase subsequently metabolizes to additional AA. These enzymes have
been localized histochemically and are found in discrete cell types in the respiratory and olfactory mucosae.
The metabolism scheme (as presented in the case study) is depicted in Figure 4:
5-13
-------
Figure 4. Metabolic Pathways for VA
Aldehyde Dehydrogauue
Two mechanisms of action were identified: (1) AA causes cytotoxicity which may progress to cell
proliferation, (2) AAld, which is a known clastogen and sister chromatid exchange initiator, leads to multi-hit
genetic damage. Tumors are seen only in male rats at the highest concentration tested, 600 parts per million
(ppm), and only at the terminal sacrifice of a 2-year b'ioassay; no effects are observed at concentrations below
50 ppm. It was hypothesized that because mice can restrict respiration (reflex apnea), less of an effect is
observed. This species difference was shown to be the case with formaldehyde, another upper respiratory
tract (URT) irritant.
Does the existing database support the URT lesions as the sentinel toxicityfor inhalation exposures to
VA?
The group agreed the database clearly supports URT lesions as the sentinel toxicity. The proximal to
distal pattern and the concentration response are both important to the argument.
Can the cytotoxic changes caused by I'A exposure be considered as sequentially linked to the observed
rumor outcome? What are the key considerations to characterize the conditions of hazard (e.g., high dose
versus low dose)? How do the genotoxic data factor in this characterization?
• Cytotoxic changes caused by VA are linked to tumors.
• AAld are linked with different tumor types. Responses in both pathways appear to be at
high doses only. The group noted that the spatial distribution of tumors was consistent.
• A "good" PBPK model exists that relates metabolism, physical layout, and fluid mechanics
in human and rodents. The PBPK model accounts for the observed species and gender
differences.
• Knowledge of cytotoxicity, cell proliferation and temporal aspects, and localization of
enzymes is helpful.
5-14
-------
Cytotoxicity may cause death but some cells will survive and those will have an increased
probability for genotoxic effects, especially at high concentrations.
What mechanistic data are most relevant to characterizing tumor outcome? Which would be useful for
dose-response modeling in the observable range? What are the implications oftheMOA information for
extrapolation of risk to low dose?
The data most relevant to tumor outcome include: cytotoxicity, cell proliferation, genotoxicity, site
specificity (localization of effect), and metabolism. Dose-response modeling based on tumor outcome is not
possible, however, because only two non-zero points (the second lowest with a response of 1) exist in the
observable range. Because effects are seen only at the highest exposure concentration and only at the last
sacrifice, the group overall felt this suggests that a nonlinear approach is appropriate for low-dose
extrapolation. This was supported by clear relationships of genotoxicity, cytotoxicity, and cell proliferation
only with high concentrations.
One breakout group member, however, disagreed that all effects are only at high concentrations. He
noted that AA leads to cytotoxicity as a result of changes of pH, which may ultimately lead to cellular
changes in the URT and to cancer. He agreed that the effect of AAld is significant only at high doses.
Evidence includes the fact that cross links are only significant at high doses and that there are no long-lived
DNA adducts. He noted, however, that large-scale changes in DNA have been observed that may have
required multiple events. He noted that these large-scale changes are important to humans and should be
examined closely. Dose-response data are lacking for observed DNA damage. In addition, there is a lack of
mechanistic understanding of the process. A low dose linear situation may, therefore, exist.
Given the availability of the PBPK model, which dose metrics should be considered for the dose-response
analysis? Does this choice of dose metric address consideration of the role of exposure duration?
Limited time was spent discussing the PBPK model although its usefulness in addressing the
toxicokinetic issue of species to species extrapolation was recognized. The dose metrics (about seven
tabulated) need to be further explored for implications to quantitative dose-response assessment. At 50 ppm
VA, the model predicts the same decrement in pH projected in animals and humans. The group concluded
that, at lower doses, animal and human responses would be quantitatively the same, but that the case study
did not present the model in sufficient detail to quantitatively explore the interspecies differences in dosimeny
(e.g.. airflow).
What are the uncertainties in using these data to characterize human risk?
The group identified several uncertainties and data gaps that, if filled, would enable further
consideration of the mechanistic actions and commonalities across endpoints.
• Reflex apnea in mice.
• Description of lesions (coverage in case study was brief).
3-15
-------
Effects of lowered pH in the respiratory tract on cancer.
Effects of acetic acid and other aldehydes.'
Gender differences.
Differences in deposition patterns in the respirator}' tract of humans versus rodents.
Dose-response data for DNA effects.
Human metabolism data (qualitatively metabolism between rodents and humans appear
similar, but rates may be different).
Should an RfC be developed separately? If an RfC orMOE were to be developed, which factors should be
considered to account for uncertainties in the extrapolations applied?
The group agreed that developing a separate RfC is justified. The potential role of lesions such as
atrophy and hyperplasia would have to be considered in the context of later tumor outcome. Uncertainty
factors would include one to account for animal to human extrapolation (based on further study of the PBPK
model) and one for intrahuman variability.
What mechanistic data would be useful for development of risk estimates of exposures via the oral route?
The group did not evaluate the oral exposure route but agreed that more than site-specific (i.e., URT)
effects need to be examined. More data are needed to learn whether using site of toxicity dose metrics is
protective of other effects.
3-16'
-------
SECTION FOUR
FINAL PLENARY SESSION
Lessons Learned and Their Applications to the Development of a Human Health Risk Assessment
Framework
In efforts to integrate information deliberated throughout the two colloquia and to assist in the
development of the framework, the group broadly discussed the questions listed below.
• Should a common quantitative analysis be conducted when there are commonalities among
toxicities?
• In the absence of case-specific PBPK models, is there a common approach for dose
adjustment for interspecies extrapolation for all responses? Does this differ for different
routes of exposure?
In the presence of PBPK models, how does MOA information influence the dose surrogate in
characterizing toxicity? Can it be different for different responses? '
• In the absence of BBDR models, how does MOA information influence the default
approach(es) to characterize in quantitative terms the potential risk of toxicities at low levels
of exposure (i.e., beyond the range of observation)? Are there common default approaches?
• The 1996 "Proposed Guidelines for Carcinogen Risk Assessment" have recommended that
five factors be considered w:hen determining the margin of exposure. These included
intraspecies variation, interspecies variation, nature of the response, steepness of the dose-
response curve, and biopersistence.
The current quantitative approach for noncancer effects generally involves development of a
single RfD/RfC for a "critical effect." Factors used include intraspecies variation,
interspecies variation, subchronic to chronic extrapolation, LOAEL to NOAEL
extrapolation, and completeness of the data base. An additional factor may be applied to
account for scientific uncertainties in the study selected for derivation of the RfD/RfC.
If the goal is to harmonize across toxicities, can a consistent set of factors be identified?
How does MOA information influence the choice of these factors?
Discussions focused on criteria and factors one should consider when evaluating integrated risk
assessment approaches. In addition, factors relevant to MOE application and appropriate "uncertainty"
factors were detailed. Prior to these discussions, the group clarified terminology related to dose response:
4-1
-------
Linear: When assuming a linear dose response, the ED10 (or point of departure) assumes
that from the point of departure (POD) there is a linear extrapolation down to zero.
Nonlinear: For a nonlinear dose response, the ED10 (or "benchmark dose") is divided by
uncertainty factors to develop an RfD.
MOE: The MOE is the ED,0-divided by the human exposure estimate of interest. It can be
applied to linear or nonlinear dose-response curves and for any endpoint
The group agreed upon this definition of MOE but noted that the description of
MOE in EPA's cancer guidelines is somewhat confusing and, therefore, needs to be
clarified.
Some participants preferred the term "margin of protection;" however, it was
pointed out that the term MOE was developed and used purposely so not to imply
"safety" or "protection."
The group considered how adequate and useful MOE is to the risk management decision and
discussed the possible basis on which an MOE should be set. The group agreed that regulators need these
"numbers" for compliance purposes. Like RfDs, MOEs need to represent exposures "without appreciable
risk." One participant noted that there are social, political, and legal issues as well as the science driving the
decision. Another participant noted that it is ultimately a risk management decision—is the MOE acceptable
given a certain set of conditions? It was noted that an MOE can be more powerful than an RfD because, in
evaluating an acceptable MOE. the entire toxicity database is examined. It is the scientist/risk assessor's
responsibility to bring the relevant information to the risk manager so that he/she can understand the
significance of a given MOE.
Colloquium participants agreed on the following points or questions regarding the application of
MOEs:
IRJS needs to include additional risk characterization information. One participant
commented that it could be included in Section 6.
A criteria list is needed to guide risk assessors and managers in applying the MOE concept
(a consistent series of questions). The list should include uncertainty issues for cancer and ,
noncancer effects.
One participant noted that a consistent approach may be difficult (across programs and the
different regions).
Both the numerator (ED10) and denominator (human exposure of interest) values need to be
clearly explained to the risk manager, including the confidence in each value.
Adequacy of the MOE will be based largely on experience.
4-2
-------
Factors considered when deriving an RED and when deciding on an MOE are similar, but not
identical. While both consider toxicity and dose-response, one important distinction is that
application of an MOE also considers the magnitude and uncertainty in the exposure
estimate. Furthermore, as mentioned previously, the entire toxicity database is considered
when deciding on an MOE.
Mode of action needs to be carefully examined when deciding if MOE is the most
scientifically viable approach for assessing risks.
The group listed the following key "uncertainty" factors for consideration when integrated
approaches are applied. No "values" were assigned.
Intraspecies differences: Differences in toxicokinetics and toxicodynamics within species.
Interspecies differences: Differences in toxicokinetics and toxicodynamics across species.
Quantitative linkages between toxicokinetics and toxicodynamics.
Severity of endpoint/effects.
Structure activity relationship information.
Human exposure scenario information (e.g., frequency, pattern, etc.).
Confidence limits on ED10 (experimental variability).
Shape and steepness of dose-response curve.
Integration of multiple factors.
Species specificity/sensitivity.
Quality of database.
Quality of individual studies.
Knowledge of MOA.
Reversibility/irreversibility of effects.
Biopersistence (e.g., is it sequestered in fat?) (toxicokinetics).
Bioavailability (toxicokinetics).
Particularly susceptible population (e.g., children, genetic susceptibility, pre-existing
4-3
-------
disease).
Route of exposure.
Route to route extrapolation.
Relationship between MOA and human exposure scenarios.
Confidence in PBPK models.
Biopersistence in the environment.
Biomarkers of effect/exposure.
Overview/Next Steps
Both colloquia were instrumental in soliciting expert opinion on evolving issues related to MOA and
integrated risk assessment approaches. Participants offered their impressions on the current state of scientific'
knowledge and on the next steps in developing a human health risk assessment framework. Having worked
through the case studies, the group agreed that, in light of available knowledge, new more scientifically-based
approaches can and should be applied. The group clearly recognized, however, that many uncertainties exist.
The following ideas were communicated by participants and reiterated throughout the colloquium.
• As was evidenced through case study discussions, a range of opinions still exist on the best
approach (e.g., shape of the dose response curve, common MOAs. etc.).
• Before integrated risk assessment approaches can fully evolve, more quantitative
information is needed.
• Risk assessors will inevitably be faced with limited data sets. The general scheme of
toxicologic events may be known, but specific mechanisms may not be fully understood.
What dp we do if only limited MOA information is available? Do we fall back on current
default approaches0 Scientists will need to evaluate when "enough" data are available.
• The process requires a good deal of data interpretation. Developing a system to aid in this
process will be challenging. Others agreed, asking "Can we come up with an approach that
is scientifically viable and useful from a regulatory perspective?"
• As integrated approaches are explored further, a case study(ies) that would use an MOE
approach needs to be developed. A set of key factors related to cancer and noncancer effects
also should be formally developed.
• The overall goal of the nsk assessment framework is to consider how to practice and
communicate the "best science" in predicting risks.
• The best available science should be used to generate the most credible risk assessment, but
4-4
-------
presented in a way that is useful to the risk manager.
Scientists need to know when not to harmonize, even when similar MOAs exist.
In closing, members of the health effects framework planning committee provided a brief overview of
next steps in the framework development. The input from agency experts during this colloquia series will be
reviewed. Numerous questions and issues were raised that will need to be re-examined and/or further
explored. The planning committee would like to see discussions from this colloquia series expanded. A
collaborative workshop, including EPA and outside groups (e.g., SOT and SRA) is being contemplated.
Participants noted that additional forums would be helpful in offering additional insight. The group
also expressed interest in future colloquia to discuss topics such as exposure and health outcome data and
PBPK models.
4-5
-------
APPENDIX A
White Paper
-------
FINAL DRAFT- Do Not Cite or Quote
Human Health Risk Assessment:
Current Approaches & Future Directions
September 1997
Risk Assessment Forum
U.S. Environmental Protection Agency
Technical Panel
Co-Chairs
Gary Kimmel, Office of Research and Development
Vanessa Vu, Office of Prevention, Pesticides and Toxic Substances
Members
Jane Caldwell, Office of Air and Radiation
Richard Hill, Office of Prevention, Pesticides and Toxic Substances
Edward Ohanian, Office of Water
-------
FINAL DRAFT- Do Not Cite or Quote
Contents
1. INTRODUCTION 1
2. MODE OF ACTION/DOSE-RESPONSE CONSIDERATIONS:
CANCER VERSUS NONCANCER EFFECTS - 3
2.1. Cancer Risk Assessment Approach 3
2.1.1. Overview of 1986 Cancer Risk Assessment Guidelines 3
2.1.2. Rationale for 1986 Cancer Risk Assessment Guidelines 4
2.1.3. New Directions for Cancer Risk Assessment 5
2.2 Noncancer Risk Assessment Approach 6
2.2.1. Overview of Current Approach 6
2.2.2. Rationale for Current Approach 7
2.2.3. New Directions for Noncancer Risk Assessment 7
2.3. Summary 8
3. POINT OF DEPARTURE FOR CANCER AND NONCANCER DOSE-RESPONSE
EXTRAPOLATION: CENTRAL TENDENCY OR LOWER BOUND ESTIMATE 9
3.1. Proposed Departure Dose Point (Benchmark Dose) for Noncancer Assessment 9
3.2. Proposed Departure Dose Point for Cancer Risk Extrapolations 10
3.3. Summary 12
4. INTERSPECIES ADJUSTMENTS FOR DOSE 12
4.1. Default Procedure for Dose Extrapolation for Noncarcinogens 13
4.1.1. ' Oral Exposure 13
4.1.2. Inhalation Exposure 14
4.1.3. Dermal Exposure : 15
4.2. Default Procedure for Dose Extrapolation for Carcinogens 16
4.2.1. Oral Exposure -•. 16
4.2.2. Inhalation Exposure 17
4.2.3. Dermal Exposure 17
. 4.3. Summary 17
-------
FINAL DRAFT- Do Not Cite or Quote
Contents (Continued)
5. APPROPRIATENESS OF UNCERTAINTY FACTORS 18
5.1. Noncancer 18
5.2. Cancer 21
5.3. Summary 22
6. NONCANCER RISK ASSESSMENT 22
6.1. Critical Health Endpoint Versus Entire Spectrum of Adverse Effects 22
6.2. Exposure-Duration Relationships 24
6.3. Dose-Response Assessment for Contaminants with Beneficial Effects
At Low Doses 25
7. . REFERENCES 28
in
-------
FINAL DRAFT- Do Not Cite or Quote
1. INTRODUCTION
Human health risk assessment entails the evaluation of available scientific information on the
biological and toxicological properties of an agent to make an informed judgment about the potential
toxicity in humans as a consequence of environmental exposure to the agent. The National Research
Council (NRC), in its report entitled Risk Assessment in the Federal Government: Managing the Process
(NRC, 1983) defined risk assessment as including some or all of the following components: hazard
identification, dose-response assessment, exposure assessment, and risk characterization. This has been
supported more recently in Science and Judgment in Risk Assessment (NRC, 1994). As recommended by
the NRC, EPA has developed health risk assessment approaches, modified them over time and
incorporated them into endpoint-specific guidelines for the evaluation of mutagenicity (USEPA, 1986),
carcinogenicity (USEPA, 1986, 1996a), developmental toxicity (USEPA, 1986, 1991), reproductive
toxicity (USEPA 1988a, 1988b, 1996b), and neurotoxicity (USEPA, 1995a). Guidelines on exposure
(USEPA 1986, 1992a) and chemical mixtures (USEPA 1986) have also been developed.
The NRC, in Science arid Judgment in Risk Assessment (NRC, 1994), noted the importance of an
approach that is less fragmented, more consistent in application of similar concepts, and more holistic
than endpoint-specific guidelines. The report also points out a number of issues in EPA's current risk
assessment approaches that need to be reexamined in light of the current scientific knowledge. For
example, the report questions the application of a non-threshold quantitative approach as a default in all
cancer risk assessments. Conversely. the use of a threshold concept as a default for agents that cause
neuro-, reproductive and developmental toxicity or that act on various systems through receptor-
mediated events is also questioned. The need for explicit accounting of variability in sensitivity among
individuals due either to inherent susceptibility or differentia^exposure was also a major point of
discussion of the NRC report. EPA's Science Advisory Board, in its review of the Draft Reproductive
Toxicity Risk Assessment Guidelines, raised similar concerns over the appropriateness of current default
approaches, that include the assumption of a threshold (USEPA, 1995b). Finally, scientists are
encouraging the use of mechanistic data in risk assessment (e.g. Butterworth et al., 1995, Purchase and
Auton, 1995). Thus, there is a recognized need for the development of a framework for human health
risk assessment which includes all of these perspectives.
-------
FINAL DRAFT- Do Not Cite or Quote
In response, the Agency's Risk Assessment Forum is beginning the development of a human
health risk assessment framework as a communication piece for risk assessors and risk managers, as well
as members of the public who are interested in health risk assessment issues. The primary purpose of the
framework document is to discuss the scientific bases and policy choices behind EPA's current risk
assessment approaches and to lay out recommended future directions for health risk assessment in the
Agency. The framework will emphasize the need for problem formulation at the beginning of the risk
assessment process and for integration and harmonization of risk assessment methodologies and
procedures of all health endpoints.
The present paper serves as the initial step in the development of a framework for a more
integrated approach to human health risk assessment. This paper discusses a number of issues regarding
the Agency's risk assessment approaches and their scientific bases to begin to examine their
compatibility with current scientific developments. Several variations in health risk assessment
approaches for carcinogenicity and for toxicological endpoints other than cancer and heritable mutations
(hereafter "noncarcinogenic" or "noncancer" effects) are examined. These include several of the default
assumptions and methodologic procedures used in the hazard and dose-response evaluations of cancer
and noncancer effects, and in accounting for potential beneficial effects at low doses. This paper is
intended as a perspectives piece and serves as a basis for further discussion of the scientific basis for
current and future risk assessment approaches.
-------
FINAL DRAFT- Do Not Cite or Quote
2. MODE OF ACTION / DOSE-RESPONSE CONSIDERATIONS: CANCER VERSUS
NONCANCER EFFECTS
Assessment of risk from exposures to environmental agents has traditionally been performed
differently, depending on whether the response is cancer or a noncancer health effect. This is because
different modes of action were thought to be involved in the two cases. Cancer has been thought to
largely be the consequence of chemically induced DNA mutations which unleash processes leading to
tumor formation. Since a single chemical-DNA interaction may lead to a mutation and since cancer is
thought to arise from single cells, it follows that any dose of an agent that produces mutations may be
associated with some finite risk. This has led the Agency to employ a science policy that cancer risk
should be estimated by a linear, nonthreshold dose-response method. On the other hand, noncancer
effects have been thought to result from multiple chemical reactions within multiple cells of an anlage,
tissue, organ or system. The Agency's science policy has been that threshold effects would pertain to
noncancer risk assessment dose-response analyses.
2.1. Cancer Risk Assessment Approach
2.1.1.. Overview of 1986 Cancer Risk Assessment Guidelines
In the Agency's 1986 cancer guidelines, observation of tumors in animals and humans are the
primary determinants of carcinogenic hazard to humans (USEPA, 1986). Other toxicologic and
mechanistic information only play a modulating role. Cancer risk estimations use dose-response models
to extrapolate tumor incidence observed in an epidemiologic or experimental study at high doses to the
much lower doses typical of human environmental exposures. Since mode of action information is
generally not available, the linearized multistage (LMS) procedure is employed as the default. An
important feature of the LMS procedure is that it assumes increased risk is proportional to dose at low
doses, even if it displays nonlinear behavior in the region of observation. A statistical confidence-limit
procedure is incorporated in the LMS to generate what is known as an upper bound on excess lifetime
cancer risk per unit of dose.
-------
FINAL DRAFT- Do Not Cite or Quote
2.1.2. Rationale for 1986 Cancer Risk Assessment Guidelines
Since the inception of EPA's cancer policy in 1976 (USEPA, 1976), the Agency has taken risk
averse positions on the identification of carcinogenic hazards and the estimation of risks. The Agency
recognized a range of evidence bearing on carcinogenesis but relied primarily on human and especially
chronic animal studies, in keeping with current scientific guidance at the time (NCAB, 1976). A single
positive animal study was generally sufficient to identify potential carcinogens, and mutagenicity and
other information played only supporting roles. A linear extrapolation of risk was assumed, based on
experience with ionizing radiation, lung cancer from smoking and the induction of genetic mutations
(Albert et al., 1977; Anderson et al., 1983; Albert, 1994). The Millers at the McArdle Institute developed
the thesis that carcinogens were electrophiles (or were metabolized to them) which interacted with
nucleophilic sites in cells, namely the DNA, to induce mutations and commence carcinogenesis (Miller
& Miller, 1976). These positions were adopted broadly among Federal agencies (IRLG, 1979).
With time it was recognized that not all carcinogens seem to be mutagens. Some researchers
suggested that mode of action could in some way be incorporated into the risk assessment process by
dividing agents into genotoxic and epigenetic categories (Weisburger & Williams, 1981). Various
groups, including EPA, considered the potential of using mode of action information, but given the
paucity of chemical-specific information, thought that such actions were largely premature (USEPA,
1982a, 1982b; IARC, 1983; Upton etal., 1984).
By 1985, it was generally accepted that mode of action may play a part in cancer risk
assessments, but there was still a significant emphasis on health-conservative default positions (OSTP,
1985; USEPA, 1986). In addition, arguments for linear dose-response relationships had centered upon
the concept of additiviry to background. This position asserts that if a chemical has a mode of action
similar to any ongoing, background process (i.e., mutations), then the risk from the chemical will simply
add to that of the background, resulting in no threshold of response and being consistent with low-dose
linearity (Crump et al., 1976).
-------
FINAL DRAFT- Do Not Cite or Quote
2.1.3. New Directions for Cancer Risk Assessment
Within the last decade, it has become generally held by various groups that mode of action can
influence significantly the conduct of risk assessments (IARC, 1991; Vainio et al., 1992; NRC, 1994;
Strauss et al., 1994). Carcinogenesis is recognized to embody changes in key genes that regulate the cell
replication cycle and can be influenced by mutagenic and non-mutagenic modes of action. Non-
mutagenic events include mitogenic and cytotoxic events that result in an increase in cellular
proliferation, immunotoxic events and modulation of key cellular control phenomena [e.g., hormonal,
receptor-mediated processes (Purchase et al., 1995)]. These concepts have been incorporated into the
EPA's 1996 Proposed Cancer Risk Assessment Guidelines (USEPA, 1996a).
Today, direct-acting mutagenic agents are assumed, as a science policy default, to influence the
potential for cancer hazard and risk at any dose (e.g., linear, non-threshold), using the same rationale as
the original 1976 EPA cancer policy. Linearity in the dose-response is also supported when anticipated'
human exposures are already in the part of the dose-response curve where effects are observed.
However, when direct mutagenic events do not pertain and other mode of action considerations apply,
the likelihood exists that cancer would be secondary to other events (e.g., stimulation of cell division).
Under such conditions a potential for cancer would exist only at doses of an agent that are sufficient to
produce the events. Such events can be anticipated to demonstrate significant nonlinearities in the slope
of the dose-response curve. In some cases thresholds may apply. Accordingly, for secondary
carcinogenic processes, a margin of exposure (MOE) analysis is proposed as the science policy default in
the proposed revisions to the 1986 Cancer Risk Assessment Guidelines (USEPA, 1996a), similar to the
approach that has been taken for non-cancer health effects (see below). Finally, in the absence of
information on mode of action, the science policy position is to assume that a linear default will apply.
2.2. Noncancer Risk Assessment Approach
2.2.1. Overview of Current Approach
The Agency treats chemicals exerting noncancer health effects as if there is a dose below which
there is no potential for risk and above which the potential for risk is undefined. Accordingly, it is
-------
FINAL DRAFT- Do Not Cite or Quote
assumed as a matter of science policy is^that thresholds apply for the risks of health effect from exposure
to such pollutants.
Evaluating human risks for non-cancer effects has generally proceeded along two lines within
the Agency. The first is derivation of the oral Reference Dose (RiD) or the inhalation Reference
Concentration (RfC). The RfC is derived for continuous airborne exposures and includes adjustments
based on respiratory physiology for animal to human extrapolation. The RfD/RfC is defined as an
"estimate with uncertainty spanning perhaps an order of magnitude of a daily exposure to the human
population, including sensitive subgroups, that is likely to be without an appreciable risk of deleterious
effects during a lifetime" (Barnes & Dourson, 1988; USEPA, 1994a). The RfD/RfC is a dose
operationally calculated from a human or animal study by dividing the no-observed-adverse-effect level
(NOAEL) for a critical effect by various (usually 3-1 OX) Uncertainty Factors (UFs) and a Modifying
Factor (MF) that reflect the various types of data used. UFs are applied on a case-by-case basis to
compensate for application of a study that identifies a Lowest-Observed-Adverse-Effect-Level (LOAEL)
instead of a NOAEL, subchronic instead of chronic study, within human variability, animal to human
•
extrapolation, and an incomplete data base. The MF also varies by up to a factor of 10 and depends upon
the uncertainties of the study and data base not explicitly treated above (Dourson and Stara, 1983; Barnes
and Dourson, 1988; USEPA, 1994a; Ohanian, 1995). A more complete discussion of uncertainty factors
is provided in section 5.0.
The second way of expressing noncancer risks is to calculate a Margin of Exposure (MOE),
which is the ratio of the critical NOAEL to the expected human exposure level. The larger the ratio, the
less likely an agent poses a risk to humans; the smaller the ratio, the greater the chance of some risk.
Part of the evaluation of the adequacy of the MOE may include the UFs and MF that might have been
applied for the case under investigation had an RfD/RfC been calculated. '
2.2.2. Rationale for Current Approach
Studies on many compounds show that before toxicity occurs, an agent must deplete physiologic
reserves or overcome repair capacity. For instance, toxicity may occur within a cell when there has been
-------
FINAL DRAFT- Do Not Cite or Quote
sufficient lipid peroxidation or when levels of glutathione have been depleted and the chemical then has
the ability to affect the cell. Likewise, toxicity is seen to occur when not just one cell is affected, but
when multiple cells in an embryonic anlage, tissue, organ or system have been perturbed. Thus as
science policy, it is assumed that toxic effects occur only after homeostatic, compensating, repair, and
adaptive mechanisms fail. Accordingly, if exposure is below that required to cause such failures,, the
noncancer effect should not be manifest.
2.2.3. New Directions for Noncancer Risk Assessment
Over time it has been recognized that threshold considerations may not be applicable to all
noncancer effects cases. Sometimes, effects are manifest at existing environmental exposure levels so
that no apparent NOAEL exists, as is the case with exposure to lead (Markowitz et al., 1996). As studies
on lead exposure in humans have been refined and conducted at lower and lower exposure levels, effects
continue to be manifest. Thus, responses within the human population is already on the observed part of
the dose-response curve, and obviously a threshold has not been defined for lead. The same seems to
apply to certain receptor-mediated effects, like those associated with 2,3,7,8-TCDD and some hormones
(e.g., estrogens).
Application of mode of action information, toxicokinetics and biologically based dose-response
models may also play a role in the evolution of assumptions concerning dose-response relationships for
noncancer effects. For instance, exposure to various mutagenic agents (e.g., ethylene oxide, ethylene
nitrosourea) of pregnant mice carrying zygotes or two-celled embryos, leads to malformations and death
later in embryonic and fetal stages (Generoso et al., 1987; Rutledge et al., 1992). Certainly these effects
arise from single exposures at the 1- and 2-cell stages, but the mechanisms leading to them have not been
determined. Maternal toxicity has been ruled out as an etiological agent, as have structural chromosome
aberrations (Katoh et al., 1989). Gene mutations are a potential cause of the effects, but they have not
been directly investigated. Likewise, it is possible that the compounds are not working via mutagenesis
but by changes in gene expression. Therefore, it is possible that thresholds would not apply in such cases.
-------
FINAL DRAFT- Do Not Cite or Quote
In addition, it is not usually feasible to distinguish empirically between a threshold and a
nonlinear dose response relationship. This has led the EPA Science Advisory Board, when deliberating
the draft risk assessment guidelines for reproduction (USEPA, 1996b) and neurotoxicity (USEPA,
1995a), to recommend a shift in the assumption about dose-response relationships from threshold to
nonlinear. However, this recommendation does not fundamentally change the ways RfDs/RfCs are
derived and interpreted.
2.3. SUMMARY
The current scientific data base indicates that automatic application of traditional approaches of
separating dose-response relationships for cancer and noncancer risk assessment, may no longer be
justified. Given mode of action information available today, the Agency is proposing to depart from the
assumption that all cancer effects show linear dose-response relationships (USEPA, 1996a). Likewise, it
\
may not be reasonable to assume that all noncancer effects show threshold dose-response relationships.
In addition, focus on mechansisms of carcinogenesis directs attention away from tumors per se toward
earlier biological and toxicological responses that are critical in the carcinogenic process. Such
responses are relevant .to both noncancer effects and cancer and serve as a bridge to link their risk
assessments.
3. POINT OF DEPARTURE FOR CANCER AND NONCANCER DOSE-RESPONSE
EXTRAPOLATION: CENTRAL TENDENCY OR LOWER BOUND ESTIMATE
The point of departure refers to that estimate of dose-response information in the observable
range from.which low-dose extrapolation occurs. Historically, EPA has used no observed adverse effect
levels (NOAELs) as the point of departure for calculation of RfDs/RfCs or margins of exposure. Cancer
risks were estimated using the linearized multistage procedure which incorporates all dose-response
information for tumor incidence in projecting risks at any finite exposure level. In recent years, the
Agency has been developing the benchmark dose (BMD) approach as an alternative for noncancer risk
assessment (USEPA, 1995c). Using this method, uncertainty factors are applied to a BMD rather than a
NOAEL. An approach similar to that of the BMD has recently been proposed for cancer risk assessment
-------
FINAL DRAFT- Do Not Cite or Quote
(USEPA, 1996). Comment is divided whether the lower bound on extrapolated dose should be used or
the point estimate of extrapolated dose should be employed for the point of departure in cancer and
noncancer dose-response assessments.
3.1. Proposed Departure Dose Point (Benchmark Dose) for Noncancer Assessment
The historical approach to defining a NOAEL and calculating a RfD/RfC has a number of
limitations. For example, this type of method does not specifically take into account both the slope of
the dose-response curve and the baseline variability in the end point in question. The resulting NOAEL
from a study using a small number of experimental animals may be significantly higher than the one
identified from a study with a larger number of animals. Finally, the NOAEL is generally limited to one
of the doses in a study and is contingent upon the dose spacing.
In response to these limitations, the Risk Assessment Forum has developed guidance on Agency
use of an alternative approach, the BMD approach (USEPA, 1996c). The BMD is defined as a statistical
lower confidence limit on the dose producing a predetermined level of change in adverse response
compared with the background response. A BMD is derived by fitting a mathematical model to the dose-
response data. In addition to the BMD approach, categorical regression analysis has been proposed to
evaluate health effects sorted into categories of progressively greater severity (e.g., no adverse effect,
mild-to-moderate effect, and severe effect) (Hertzberg, 1989; Dourson, 1994; Rees and Hattis, 1994).
With respect to the dose point of departure, participants at a workshop on the benchmark dose
recommended the use of the lower confidence limit on the 10% incidence (or some other incidence level)
of effect as the point of departure (Barnes et al., 1995). The lower confidence limit provides a means of
including the variability of the data in the analysis, and addresses one of the limitations of the current
RfD/RfC approach.
-------
FINAL DRAFT- Do Not Cite or Quote
3.2. Proposed Departure Dose Point for Cancer Risk Extrapolation
The proposed revisions to the cancer risk assessment guidelines (USEPA, 1996a), like the BMD
approach, divide dose-response assessment in two parts. The first is assessment of the data in the range
of empirical observation. This is followed by low-dose extrapolations either by modeling, if there are
sufficient information to support the use of case-specific model, or by a default procedure if there is not.
The default procedure may utilize a linear or nonlinear approach, or both, based on information of the
agent's likely mode of action. For those agents producing cancer that 1) lack mutagenic activity and 2)
have sufficient evidence of a nonlinear dose response relationship, an analysis of margin of exposure
(MOE) is conducted to provide perspective on how much risk reduction is associated with reduction in
dose. The MOE is the ratio of the dose point of departure to the human exposure level. The point of
departure can be obtained in several ways for cancer dose-response assessment. To be consistent with
the process for the BMD for noncancer endpoints, the current proposal is to calculate either (1) the lower
95% confidence limit on dose for the observed or calculated 10% tumor incidence level, or (2) the lower
95% confidence limit on dose for the observed or calculated 10% incidence'of some tumor precursor
(e.g., hyperplasia, hormone levels) (USEPA 1996A).
At a workshop in the fall of 1994 (USEPA, 1994b) that evaluated an early draft of the cancer risk
assessment guidelines, there was a strong recommendation that the Agency use dose associated with a
particular tumor or tumor precursor response (e.g., 10%) instead of the lower confidence limit as is done
for non-cancer health endpoints in the benchmark dose procedure as the point of departure. The
importance of calculating the upper and lower 95% confidence limits on the 10% tumor incidence and
conveying that information to risk managers as part of the risk characterization was recognized and
recommended. It was thought that using the lower 95% confidence limit alone resulted in introducing a
level of exactitude and public health conservatism that was unnecessary as a part of the analysis of
observed data and given the uncertainties inherent in later extrapolation to lower doses outside the
observed data range. However, in order to be consistent with the proposed noncancer BMD procedure,
the Agency proposed in the 1996 cancer guidelines that the lower confidence limits on the 10%
incidence dose be used. In the Federal Register notice of the proposed guidelines, the Agency
specifically requested comments on how to proceed with defining the point of departure (USEPA,
10
-------
FINAL DRAFT- Do Not Cite or Quote
1996a). At a more recent workshop on the BMD approach (USEPA, 1996d), in which there had been
adequate time for reflection on the proposals for the cancer risk assessment guidelines, participants were
divided as whether to use the lower confidence limit (BMD) or the point estimate (e.g., 10% response) as
the departure point.
3.3. Summary
The Agency is interested in developing consistent principles both for analysis of observed data
and extrapolation below the observed range of exposures. However, a number of issues have been raised
with the revision of the cancer risk assessment guidelines and the development of the BMD approach for
noncancer risk assessment. There is still debate over the use of lower confidence limit on the dose or the
point estimate as the proposed departure pint for low-dose extrapolation. Is there a reason to apply.
different approaches to cancer or other health effects? Cancer testing in animals regularly uses 50 or
more animals per dose group, a number greater than in most testing of noncancer endpoints. Would it be
preferable to use a point of departure that is based on the power of the study, yet may differ for different
endpoints? There are numerous options to consider.
4. EVTERSPECIES ADJUSTMENTS FOR DOSE
There are a number of uncertainties in the extrapolation of dose-response data from animals to
humans. EPA's risk assessment guidelines and procedures provide specific guidance for the application
.of default approaches and procedures to compare dose between species and to account for potential
species differences in the carcinogenic and noncarcinogenic responses to environmental agents. One of
the critical steps in risk assessment is the selection of the measure of exposure for definition of the
exposure-dose-response relationship. EPA's exposure guidelines (USEPA 1992a) describes several types
of exposure measures for such definition. Administered dose is the amount of chemical ingested, inhaled.
or applied to the skin. Internal dose is the amount of a chemical that has been absorbed across the
applicable barriers (i.e., the gut wall, the skin, or the lung lining) and is available for biological
interactions. Delivered dose is the amount transported to an individual organ, tissue, or fluid of interest.
Biologically effective dose is the amount of the chemical that actually reaches cells, sites, or membranes
11
-------
FINAL DRAFT- Do Not Cite or Quote
where adverse effects occur. Ideally, the biologically effective dose is used as the basis for defining the
dose-response relationship and for assessing risk.
EPA has recommended the use of physiologically-based pharmacokinetics (PBPK) models as the
procedure of choice to account for metabolism and pharmacokinetics processes and, thereby, improve
confidence in dose estimation (USEPA, 1986, 1994). This approach for dose extrapolation between
species, however, is not possible for most compounds since the use of PBPK models requires extensive
comparative metabolism and pharmacokinetics.data for use in the modeling process, as well as a good
understanding of the agent's mode(s) of action. These data are generally not available for most
compounds. As a result, EPA has developed default procedures to compare dose between species in the
absence of sufficient pharmacokinetics information. The default assumption is that the administered
dose and biologically effective dose are directly proportional.
4.1. Default Procedure for Dose Extrapolation for Noncarcinbgens
The RfD/RfC methodologies represent quantitative approaches to estimate levels of exposure
with little appreciable risk of adverse effects for noncancer endpoints. A major difference between the
two approaches is that the RfC methodology includes dosimetric adjustments to account for the
relationship between exposure concentrations with that of deposited or delivered doses, whereas the RfD
does not. ,
4.1.1. Oral Exposure
In the derivation of a RfD, it is assumed that the dose administered orally is proporti6nal to the
delivered dose as well as the biologically effective dose, and is equivalent across species on a body
weight basis (BW). The underlying scientific bases for this assumption are not provided in the guidance
describing the methodology. However, such procedures are common among other agencies as well as
internationallv.
12
-------
FINAL DRAFT- Do Not Cite or Quote
4.1.2 Inhalation Exposure
In the RfC methodology, the disposition of inhaled toxicants is determined by several factors.
EPA has established standard methods for derivation of the human equivalent concentration (HEC)
estimates from animal exposure data. Disposition is defined for inhalation exposure as encompassing the
processes of deposition, absorption, distribution, metabolism, and elimination. Major factors include the
respiratory tract anatomy and physiology, as well as the physicochemical characteristics of the inhaled
toxicant. In addition, the relative contribution of these factors is also influenced by exposure conditions
such as concentration and duration. Finally, default adjustment factors are used which are based on
default dosimetry models for relatively insoluble and non-hygroscopic particles and three categories of
gases (USEPA, 1994).
The default deposition model for particles provides estimates of regional deposition of the major
respiratory tract regions [i.e., extrathoracic (ET), tracheobronchial (TB), and pulmonary (PU) regions].
The model, however, does not take into account the clearance and distribution of the deposited dose
which would allow for a more accurate estimation of the retained dose and would be a better measure of
chronic dose for the derivation of a RfC. For particles, a multiplicative factor ( RDDDr or regional
deposited dose ratio), is used to adjust an observed inhalation particulate exposure concentration of an
animal to that of a human that would be associated with the same dose delivered to a specific regional (r)
tissue. Depending on whether the observed toxicity is in the respiratory tract or at distal
(extrarespiratory) sites, RDDR, is used in conjunction with default normalizing factors for the
physiological parameter of interest. Because insoluble particles deposit and clear along the surface of
the respiratory tract, dose per unit surface area is the recommended normalizing factor for respiratory
effects due to particulate deposition. Body weight is often used to normalize dose to distal target tissues.
13
-------
FINAL DRAFT- Do Not Cite or Quote
For gases, the dosimetric adjustments are dependent on the type of gas as well as the effect to be
assessed, i.e., respiratory effects or extrarespiratory toxicity. The two categories of gases with the
greatest potential for respiratory effects are those that are highly water soluble and/or rapidly irreversibly
reactive in the respiratory tract (Category 1), and those that are water soluble and rapidly reversibly
reactive, or moderately to slowly irreversibly metabolized in respiratory tract tissue (Category 2).
Because they are not as reactive in the respiratory tract tissue as Category 1 gases, gases in Category 2
also have the potential for significant accumulation in the blood and, therefore, have a higher potential
for both respiratory and distal toxicity. Gases in Category 3 are relatively water insoluble and unreactive
and their uptake is predominantly in the pulmonary region. The site of toxicity of these gases is
generally at sites remote to the respiratory tract.
\
s
I
For gases, a ratio of regional dose of a gas in the laboratory animal species to that of humans for
region (r) of interest for the toxic effect (RGDR,) is used to dosimetrically adjust the experimental
NOAEL to an HEC. The default equations to calculate the RGDR, for the different gas categories are
dependent on the types of effects - respiratory effects versus effects at remote sites. For respiratory
effects, the default RGDR, is based on species differences of ventilatory parameters and regional
respiratory surface areas (i.e., ET, TB, PU) of concern. For extrarespiratory effects, the default approach
assumes that the toxic effects observed are related to the arterial blood concentration of the inhaled
agent, and that the animal alveolar blood concentrations are periodic with respect to time for the majority
of the experiment duration. Thus, the NOAEL[HEC] is dependent on the ratio of the blood to gas (air)
partition coefficient of the gas for the animal species to the human value. For the situation in which
blood to gas (air) partition coefficients are unknown the default value of 1 is recommended.
4.1.3. Dermal Exposure
No official Agency guidance has been developed for evaluating health risks from dermal
exposure to chemicals. However, EPA's Office of Research and Development (ORD) has developed
interim methods and procedures for estimating dermally absorbed dose resulting from direct contact
with environmental contaminants in soil, water, and contact with vapors (USEPA, 1992c). The
guidance document provides a range of default values to be used in situations where exposure
14
-------
FINAL DRAFT- Do Not Cite or Quote
information and chemical-specific data (e.g. permeability coefficient) are not available.
Due to the paucity of dose-response data from dermal exposure to chemicals, the default
practice for characterizing noncancer risks from dermal contact with contaminants in soil and water is to
utilize chemical-specific oral RfD, with some adjustment for dermal bioavailability when feasible.
4.2. Default Procedure for Dose Extrapolation for Carcinogens
4.2.1. Oral Exposure
To derive a human equivalent oral dose from animal data, the default procedure as
recommended in the 1986 Cancer Risk Assessment Guidelines was to scale the lifetime average daily
dose by 2/3 power of body weight as a measure of differences in body surface area. Dose extrapolation
on the basis of body surface area was thought to be appropriate because certain pharmacological effects
commonly scale according to surface area (USEPA, 1986). Recently, the Agency has adopted the
recommendation made by an interagency workgroup that interspecies scaling be based on 3/4 power by
body weight (USEPA, 1996a). The underlying assumption is that lifetime cancer risks are equal in
animals and humans when average daily administered dose are proportional to each species' body
weight. This default procedure is based on empirical observation that rates of physiological processes
consistently tend to maintain proportionality with 3/4 power by body weight (USEPA 1992b).
15
-------
FINAL DRAFT- Do Not Cite or Quote
4.2.2. Inhalation Exposure
The default procedure to derive a human equivalent concentration of inhaled particles, gases, and
vapors is that for estimating inhaled dose in the derivation of RfC (see discussion above).
4.2.3. Dermal Exposure
As discussed in section 4.1.3, interim guidance is available for the estimation of dermally
absorbed dose resulting from direct contact with environmental contaminants in soil, water, and contact
with vapors (USEPA, 1992c). Potential cancer risk from dermal exposure to systemic carcinogens for
which dose-response information by the oral route is available can be estimated with some adjustment
for dermal bioavailability. This default procedure is only applicable for chemicals that are expected to
be readily absorbed via animal and human skin.
4.3. Summary
As illustrated from the discussion above, different default assumptions and methodologies are
being utilized to account for interspecies differences for dose in the assessment of cancer and noncancer
risks. There are also differences in the methods applied to different routes of exposures. The underlying
scientific bases for these default assumptions need to be re-examined in light of the need to better
harmonize and integrate the assessment for potential human cancer and non-cancer health effects. A
number of questions have been raised: (1) Should EPA's science policy for dosimetric adjustments be
.the same for cancer and noncancer assessments from lifetime oral exposure, as it has now been
recommended for inhalation exposure? (2) What would they be? (3) What are the interagency and
international implications of adopting similar default procedures? In addition, more guidance is needed
for the evaluation of potential cancer and noncancer risks from dermal exposures. Current EPA risk
assessment guidelines primarily focus on oral and inhalation pathways.
16
-------
FINAL DRAFT- Do Not Cite or Quote
5.0 APPROPRIATENESS OF UNCERTAINTY FACTORS
Efforts have been made to account for major sources of variation in responses when estimating
levels of human exposure that may not be attended with significant risk for noncancer and, more
recently, for certain cancer risk assessments. Uncertainty factors (UFs) have been used to account for
response differences of various types. They have often been used, along with a modifying factor (MF)
which is dependent on the completeness of the data, for calculation of an RfD/RfC or evaluation of the
significance of a margin of exposure (MOE) (NOAEL/estimated human exposure). Questions have
arisen concerning the magnitude of individual uncertainty factors and the appropriateness of
compounding a number of such factors together for evaluation of potential risk.
5.1. Noncancer
Traditionally, UFs of up to 1 OX have been used to adjust for differences in variability of
response following oral exposures for differences: (a) within species, (b) between species, (c) when-
using less than chronic data, (d) when using a lowest observed adverse effect level (LOAEL) instead of a
NOAEL, and (e) incompleteness of the data base (Barnes & Dourson, 1988; USEPA, 1994).
The initial choice of 10X for these UFs was somewhat arbitrary (Lehman and Fitzhugh, 1954).
Empirical analyses presented in Table 1 (see page 20 ) indicate that these values are usually conservative
estimates of the underlying variability (Dourson & Stara, 1983; Calabrese, 1985; Lewis et al., 1990). For
instance^ '
a. Nair et al. (1995) investigated NOAELS for a large number of subchronic and chronic studies in
rats, mice and dogs that were investigated by FAO/WHO and a smaller number of studies
conducted by Monsanto. Interspecies comparisons could be made for 7 to 73 studies. Of these
cases, 80-100% of interspecies comparisons are covered by a 1 OX factor, and the median is
usually less than a factor of 3X, although there is one exception.
17
-------
FINAL DRAFT- Do Not Cite or Quote
b. Human variability can be quite marked for certain inherited conditions, but about 80 to 95% of
cases people are covered by a 10-fold factor (Calabrese, 1985). This is also born out when
comparisons are made for various pharmacokinetic factors as well as for the elimination half life
or the therapeutic dose of Pharmaceuticals (Naumann, 1995).
c. Variability in extrapolating from subchronic to chronic studies ranges from 9 to over 40 study
comparisons (Weil & McCollister, 1963; McNamara, 1976; Abdel-Rahman, 12995; Nair et al.,
1995; Nessel et al., 1995). Median differences are 4 fold or less; the 90th percentile is usually
about 5 fold; and essentially 100% of cases are within a factor of 10 fold.
d. In comparisons of the LOAEL vs. a NOAEL in a study, investigators have noted median
differences of less than 4 fold and 90th percentile fold differences of about 5, with almost all
cases being covered by a factor of 10 fold (Weil & McCollister, 1963;.Abdel-Rahman, 1995;
. Kadryetal., 1995).
These data indicate that uncertainty factors of 10 are generally inclusive pf the variation that
exists for the various factors, often with the median significantly less than 1 OX. Even the 90th percentile
for a number of the factors mav onlv be about a factor of 5X.
18
-------
FINAL DRAFT- Do Not Cite or Quote
Table 1. Observed Variability of Responses
Factor
Interspecies
Intraspecies
Subchronic
to chronic
LOAEL to
NOAEL
Nairetal., 1995 rat/mouse (N=31)
(N= 7)
rat/dog (N=73)
(N= 7)
mouse/dog (N=30)
Calabrese, 1985
Hattis et a!., 1987 p'kinetic factors
Naumann, 1995 elimination tlc
therapeutic dose
Weil & McCollister, 1963 (N=33)
McNamara, 1976 (N=41)
Abdel-Rahman, 1995 (N= 3)
Nesseletal., 1995 oral (N=22)
inhalation (N=|9)
Nairetal., 1995 (N=22)
Weil & McCollister, 1963 (N=33)
Kadryetal., 1995 (N= 9)
Abdel-Rahman, 1995 (N=24)
Fold level at
named %
50th
3.0
5.3
2.0
1.8
2.9
-
<2.0
2.0
4.0
3.3
<3.0
2.0
<3.5 .
90th
-
<5.0
<5.0
<5.0
3.5
7.6
<5.0
5.0
Proportion
of cases
below 10-
fold level
80%
85%
92%
100%
83%
80-95%
100%
100%
88%
97%
100%
100%
96%
100%
68%
100%
100%
96%
19
-------
FINAL DRAFT- Do A/of Cite or Quote
Given the inclusive nature of individual 10X UFs, compounding of multiple factors all with this
magnitude could result in a significant overestimation of the inherent total variability. For instance, the
combination of five factors of 10X to calculate an RfD is 100,000. If the individual UFs were actually
3X each instead of 10X, the overall estimate of variability would be 27, a value nearly 4000 times
smaller than the default value. Partially in recognition of this problem, the Agency limits the maximum
product of the UFs and MF for RfD/RfC calculation to 3000. If factors in a given case are in excess of
3000, then an RfD is not calculated. An empirical analysis of the influence of compounding UFs on 231
RfDs found that none of the calculated values was greater than the 30th percentile of the distribution of
potential human threshold doses and over half were below the 5% level (Baird et al., 1996).
In addition, for calculation of some RfDs EPA has deviated from using the default 10X factors:
(a) when human variability is less than the default, (b) when the database is partially complete, (c) for
essential nutrients when default factors would result in exposures below maintenance levels, (d) when
the LOAEL is a minimal effect, and (e) when animal studies warrant reduction, as when they share a
common target toxicity with humans (Cicmanec & Poirier, 1995).
5.2. Cancer
In the 1996 proposed cancer risk assessment guidelines, an MOE approach is used when there is
sufficient information to conclude the agent is not mutagenic and mode of action findings support a non-
linear dose-response relationship. In evaluating MOEs, default factors of not less than 10X are
suggested to account for differences in sensitivity (a) within species and (b) between species. If humans
are less sensitive than animals, the default value is 0.1. Basically all hazard and dose response
information are to be considered in evaluating the adequacy of the MOE. Other factors should be
evaluated include things like (c) slope of and uncertainties about the dose response curve at the point of
departure, (d) nature of the endpoint used for dose response assessment, and (e) persistence of the agent
in the body. Only qualitative guidance is given as to how to use this information.
20
-------
FINAL DRAFT- Do Not Cite or Quote
5.3 Summary
Traditional use of 1 OX uncertainty factors seems to account for the variability in responses of a
number of factors and may overestimate it in most cases. Exceptions do exist, however. Compounding
multiple UFs may only propagate either over or underestimates in calculating RfDs/RfCs and in
evaluating MOEs.
4-
Several issues deserve consideration such as the following. Should default UFs remain the same
as in the past or be changed? Should assessments include the use of central tendency values for UFs or
continue with default 1 OX positions? How should the employment of multiple UFs be presented and
characterized in risk assessments?
6. NONCANCER RISK ASSESSMENT
6.1. Critical Health Endpoints Versus Entire Spectrum of Adverse Effects
As discussed in the introduction section, the Agency has published several guidelines for
assessing specific non-cancer, non-mutagenic endpoints, such as developmental toxicity (USEPA, 1986,
1991); reproductive toxicity (USEPA 1988a, 1988b, 1996b), and the proposed neurotoxicity (USEPA,
1995a). These guidelines set forth principles and procedures to guide EPA scientists in the interpretation
of studies that follow EPA's testing guidelines and other toxicologic and epidemiologic information to
make inferences about the potential hazard to specific health endpoints and identification of data and
.knowledge gaps. In practice, EPA risk assessments do not routinely make a full evaluation and
characterization of various potential health effects. Rather, most EPA non-cancer assessments focus on
the "critical effect" of an agent (i.e., the adverse effect or its known precursor which occurs at the lowest
dose) to derive an RfD or RfC for oral and inhalation exposures, respectively. The RfD/RfC approach
assumes that if exposure can be limited so that such a critical effect does not occur, then no other effects
of concern will occur. Consequently, this approach fulfills the regulatory needs in various EPA programs
for defining an exposure level(s) below which there is negligible risk of adverse non-cancer and non-
mutagenic effects from exposure to a given agent.
21
-------
FINAL DRAFT- Do Not Cite or Quote
EPA also conducts endpoint specific assessments for identification of potential hazards for
priority setting or hazard ranking, for making decisions whether to invest resources in collecting data for
a full assessment, or for determination of whether there is scientific basis for listing an agent on the
Agency's regulatory lists of hazardous substances of.concern. These hazard assessments can be of
screening or comprehensive level depending mainly on the regulatory need. Accordingly, the scope and
depth of a given EPA assessment for noncarcinogenic effects vary depending on its intended purpose, the
available data and resources, and other factors including the nature of risk management needs. Critical to
the process is communication between risk assessors and risk managers to insure that scientific
information is best analyzed and used.
Risk assessments that focus only on the critical health endpoint, in effect, minimize
characterization of other adverse effects the chemical may cause and the doses where they are found. As
such, the full spectrum of potential effects are not characterized. In trying to identify potential health
effects in humans from studies of an agent in experimental animals, the assessor seldom knows which
•
effects are predictive of those which may occur in humans. Therefore, there is merit in presenting the
myriad of effects in experimental animals at differing dose levels. As a result, risk managers may have a
better appreciation of the potential effects in humans and can better evaluate risk reduction options. In
addition, performing non-cancer effects in this way would have several advantages: 1) a better
appreciation of possible hazards at various exposures is developed with little more investment of time
and effort, 2) because it is not known whether sensitivity to different effects is the same for humans as
that of the test animals, a more full consideration of effects that may be closely spaced in appearance
with increasing exposure could be realized; and 3) non-cancer effects that may underlie potential
carcinogenic endpoints could be discerned and examined. A presentation of a spectrum of effects is
/
currently being accomplished in the ATSDR toxicological profiles which feature graphic means to
summarize observed effects.
22
-------
FINAL DRAFT- Do Not Cite or Quote
6.2. Exposure-Duration Relationships
Historically, the risk assessment of noncancer effects has placed emphasis on the potential health
effects from continuous lifetime exposures. However, there is an increasing recognition that other
exposure scenarios such as intermittent occupational and consumer exposures, as well as accidental
exposures are also of regulatory concern. As a result, various EPA regulatory program offices have
developed or are developing exposure guidelines or advisories for acute, short-, or intermediate-term
exposures. For example, the Office of Water has developed health advisories for 1-day and 10-day
consumption levels, which consider exposures to both adults and children. The Office of Pollution
Prevention and Toxics is leading an Agency effort, in collaboration with other federal and state agencies,
to develop acute exposure guideline levels (AEGL) for the general public from emergency or accidental
exposures to hazardous chemicals. The risk evaluation method for AEGL is based on the methodology
developed by the National Academy of Sciences (NAS, 1993). The Office of Pesticides Program has
recently completed its effort in the development of risk assessment methods for less-than-lifetime
exposures to pesticides.
However, all of the available approaches, described above to estimate short-duration exposure
limits, assume a constant relationship between level of an exposure and its duration with respect to the
expected response. Specifically, the exposure basis used in risk assessment calculations is a "daily
exposure", regardless of the actual timing, duration, or frequency of exposure. Even in the derivation of
a reference dose or reference concentration for developmental toxicity (RfDDT, RfCDT), the risk
assessment is based on the overall daily exposure.
Consequently, \vhiie approaches for incorporating less-than-lifetime exposures in-the risk
assessment process have been developed, our understanding of the influence of the timing, duration, and
frequency of exposure on chemical toxicity is limited at best. There is a need for the development of an
Agency risk assessment guidelines for the evaluation of "less-than-lifetime exposures". These guidelines
should set forth the general principles and approaches, and the underlying assumptions of available
methodologies for various exposure scenarios other than continuous lifetime exposures and stress the use
of toxicokinetic data where possible. These guidelines should also be useful in identifying major gaps in
our scientific knowledge.
23
-------
FINAL DRAFT- Do Not Cite or Quote
63 Dose-Response Assessment for Contaminants with Beneficial Effects at Low Doses
Essential elements are those elements that must be present in small quantities in the human diet
to maintain normal physiological, and biochemical functions. The 10th edition of the NRC's
Recommended Dietary Allowances (NRC, 1989) identifies nine essential elements. For four of these
(iodine, iron, selenium, and zinc), the database was considered acceptable to set a Recommended Dietary
Allowance (RDA), and for the other five (chromium, copper, fluoride, manganese, and molybdenum), a
range of estimated safe and adequate daily dietary intakes (ESADDIs) was generated. The NRC also
addressed several other trace elements (e.g., arsenic, boron, nickel and silicon), for which there is some
evidence of essentiality but where physiological/biochemical requirements and functions in humans have
not been proven.
For each essential element, there are two ranges of exposure or intake associated with adverse
health effects; intakes that are too low and result in nutritional deficiency, and intakes that are too high
and cause toxicity. The general dose-response for adverse effects for these elements thus has been
visualized as U-shaped, composed of overlapping curves for deficiency and toxicity (ILSI, 1994).
Ideally, the "trough" of the U-shaped curve would define the region of acceptable (safe and adequate)
intakes. In practice, the available data are seldom adequate to clearly describe the shape of the curve,
and values such as the RDA are established with a margin of safety based on the best scientific evidence
available.
On the toxicity side of the U-shaped relationship, EPA establishes oral RfDs. Because human
data on the toxicity of these elements are limited, RfDs often must be based to a considerable extent on
experimental data from animal studies, and in most cases, there is a large uncertainty factor associated
with such RfDs. In fact, in one case, zinc, the RDA and RfD were found to be almost identical, and for
other cases the values were within an order of magnitude or less. This apparent convergence of values
associated with beneficial effects on one hand and minimal risk of toxicity on the other suggests the need
for a closer look at the Agency's risk assessment methodology for contaminants with beneficial effects at
low doses (Calabrese, 1995). The following examples illustrate this point of view (ILSI, 1994).
24
-------
FINAL DRAFT- Do Not Cite or Quote
1. The RDA for zinc (15 mg/day for males, 12 mg/day for females) and the RfD (0.3 mg/kg/day, or
21 mg/day for a reference 70-kg adult) represent somewhat convergent doses.' Furthermore, the
RfD for this element is below the RDA for infants, children, adolescents, and (possibly) pregnant
or lactating women, an overlap that is acknowledged in IRIS.
2. Selenium has an RDA of 70 ug/day for males and 55 ug/day for females, compared with an RfD
of 5 ug/kg/day (350 ug/day). Both the RDA and RfD for selenium are based on studies in China.
The actual estimated dietary selenium intakes of Americans vary, ranging from 60 to 234
ug/Se/day. For some apparently healthy individuals, however, selenium intakes appear to be
greater than the RfD, with no apparent adverse effects.
Based on the above discussion, it is quite timely that the Agency evaluates its existing risk
assessment methodologies to apply "common sense" while attempting to maximizing beneficial effects
at low doses and minimizing toxic effects at high doses.
25
-------
FINAL DRAFT- Do Not Cite or Quote
1. REFERENCES
Abdel-Rahman, M.S. & Kadry, A.M. 1995. Studies on the use of uncertainty factors in deriving RfDs.
Hum. Ecol. Risk Assess. 1:614-624.
Albert, R.E. 1994. Carcinogen risk assessment in the U.S. Environmental Protection Agency. Crit. Rev.
Toxicol. 24: 75-85.
Albert, R.E., Train, R.E. & Anderson, E. 1977. Rationale developed by the Environmental Protection
Agency for the assessment of carcinogenic risks. J. Natl. Cancer Inst. 58:1537-1541.
Anderson, E.L. and the Carcinogen Assessment Group of the U.S. Environmental Protection Agency.
1983. Quantitative approaches in use to assess cancer risk. Risk Anal. 3:277-295.
Baird, S.J.S., Cohen, J.T., Graham, J.D., Shlyakhter, A.I. & Evans, J.S. (1996) Non cancer risk
assessment: a probabilistic alternative to current practice. Hum. Ecol. Risk Assess. 2:79-102.
Barnes, D.G. & Dourson, M. 1988. Reference dose (RfD): Description and use in health risk
assessments. Regul. Toxicol. Pharmacol. 8:471-486.
Barnes, D.G.k, Daston, G.P.. Evans. J.S., Jarabek, A.M., Kavlock, R.J., Kimmel, C.A., Park, C. &
Spitzer, H.L. (1995) Benchmark dose workshop: Criteria for use of a benchmark dose to estimate a
reference dose. Regul. Toxicol. Pharmacol. 21:296-306.
Calabrese, E.J. 1985. Uncertainty factors and interindividual variation. Regul. Toxicol. Pharmacol. 5:
190-196.
Calabrese, E. J. 1995. -Expanding the RfD concept to incorporate and optimize beneficial effects while
preventing toxic responses from non-essential toxicants. BELLE Newsletter 4: 1-10.
Cicmanec, J.L. &. Poirier. K..A. 1995. Selected applications of reduced uncertainty factors in noncancer
risk assessment. Hum. Ecol. Risk Assess. 1: 637-640.
Crump, K.S., Hoel. D.G.. Langley. C.H. & Peto, R. 1976. Fundamental carcinogenic processes and their
implications for low dose risk assessment. Cancer Res. 36: 2973-2979.
26
-------
FINAL DRAFT- Do Not Cite or Quote
Dourson, M.L. and Stara, J.F. 1983. Regulatory history and experimental support of uncertainty (safety)
factors. Regul. Toxicol. Ind. Health. 3:224-238.
Generoso, W.M., Rutledge, J.C., Cain, K.T., Hughes, L.A. & Braden, P.W. 1987. Exposure to female
mice to ethylene oxide within hours after mating leads to fetal malformation and death. Mutat. Res.
176:269-274.
International Agency for Research on Cancer (IARC) 1983. Approaches to classifying chemical
carcinogens according to mechanism of action. IARC internal tech. rept. no. 83/001. Lyon:
International Agency for Research on Cancer.
International Agency for Research on Cancer (IARC) 1991. A consensus report of and IARC
monographs working group on the use of mechanisms of carcinogenesis in risk identification. IARC
internal tech. rept. no. 91-001. Lyon: International Agency for Research on Cancer.
International Life Sciences Institute (ILSI) 1994. Risk assessment of essential elements, ILSI Press,
Washington, D.C.
Interagency Regulatory Liaison Group (IRLG) 1979. Scientific bases for identification of potential
carcinogens and estimation of risks. Fed. Regist. 44: 39858-39879. Also published in J. Natl. Cancer
Inst. 63:241-268.
Kadry, A.M., Skowronski, G.A., Khodair, A.I., & Abdel-Rahman, M.S. 1995 Determining "safe" levels
of exposure: The validity of the use of 10X safety factors. Hum. Ecol. Risk Assess. 1: 565-575.
Katoh, M., Cacheiro, N.L.A., Cornett, C.V., Cain, K.T., Rutledge, J.C. & Generoso, W.M. 1989. Fetal
anomalies produced subsequent to treatment of zygotes with ethylene oxide or ethyl methanesulfonate
are not likely due to the usual genetic causes. Mutat. Res. 210: 337-344.
Lehman, A.J. & Fitzhugh, O.G. 1954 100-fold margin of safety. Assoc. Fd. Drug Offic. U.S. Quart.
Bull. 18:33-35.
Lewis, S.C., Lynch, J.R. & Nikiforov, A.I. (1990) A new approach to deriving community exposure
guidelines from "no-observed-adverse-effect levels." Regul. Toxicol. Pharmacol. 11: 314-330.
Markowitz, M.E., Bijur, P.E., Ruff, H.A., Balbi, K., an d Rosen, J.F. 1996. Moderate lead poisoning:
Trends in blood lead levels in unchelated children. Environ. Health Perspect. 104:968-972.
27
-------
FINAL DRAFT- Do Not Cite or Quote
McNamara, B.P. 1976 Concepts in health evaluation of commercial and industrial chemicals. In
Mehlman, M.A., Shapiro, R.E. & Blumenthal, H., eds. New Concepts in Safety Evaluation. Washington,
DC: Hemisphere, pp. 61-115.
Miller, B.C. & Miller, J.A. 1976. The metabolism of chemical carcinogens to reactive electrophiles and
their possible mechanisms of action in carcinogenesis. In Searle, C.E., ed. Chemical carcinogens. ACS
monogr. 173. Washington, DC: American Chemical Society, pp. 737-762.
Nair, R.S., Sherman, J.H., Stevens, M.W. & Johannsen, F.R. 1995 Selecting a more realistic uncertainty
factor: Reducing compounding effects of multiple uncertainties. Hum. Ecol. Risk Assess. 1:576-589.
National Academy of Sciences (NAS). 1993. Guidelines for developing community emergency exposure
levels for hazardous substances. Committee on Toxicology. Washington DC: National Academy Press.
National Cancer Advisory Board (NCAB). 1976. General Criteria for assessing the evidence for
carcinogencity of chemical substances: Report of the Subcommittee on Environmental Carcinogenesis,
National Cancer Advisory Board. J. Natl. Cancer Inst. 58:461-465.
Naumann, B.D. & Weideman, P.A. 1995 Scientific basis for uncertainty factors used to establish
occupational exposure limits for phamaceutical active ingredients. Hum. Ecol. Risk Asssess. 1: 590-
613.
Nessel, C.S., Lewis, S.C., Stauber. K.L., & Adgate, J.L. 1995 Subchronic to chronic exposure
extrapolation: Toxicologic evidence for a reduced uncertainty factor. Hum. Ecol. Risk-Assess. 1:516-
526.
National Research Council (NRC). 1983. Risk Assessment in the Federal Government: Managing the
Process. Washington D.C.: National Academy Press.
National Research Council (NRC). 1989. Recommended dietary allowances, 10th ed. National Academy
Press, Washington, D.C.
»
National Research Council (NRC). 1994. Science and judgment in risk assessment. Washington, DC:
National Academy Press.
Ohanian, E.V. 1995 Use of the reference dose in risk characterization of drinking water contaminants.
Hum. Ecol. Risk Assess. J. 1: 625-631.
28
-------
FINAL DRAFT- Do Not Cite or Quote
Office of Science and Technology Policy (OSTP). 1985. Chemical carcinogens: A review of the
science and its associated principles. Fed. Regist. 50:10371-10442. Also published Environ. Health
Perspect. 67: 201-282. 1986.
Purchase, I.H.F. & Auton, 1995. Thresholds in chemical carcinogenesis. Regulatory Toxicol.
Pharmacol. 22: 199-205.
Rutledge, J.C., Generoso, W.M., Shourbaji, A., Cain, K.T., Gans, M. & Oliva, J. 1992 Developmental
anomalies derived from exposure of zygotes and first-cleavage embryos to mutagens. Mutat. Res. 296:
167-177.
Strauss, B., Hanawalt, P., Swenberg, J. 1994 Risk assessment in environmental carcinogenesis. An
American Association for Cancer Research special conference in cancer research cosponsored by the
Environmental Mutagen Society. Cancer Res. 54: 5493-5496.
Upton, A.C., Clayson, D.B., Jansen, J.D., Rosenkranz, H. & Williams, G. 1984 Report of ICPEMC task
group on the differentiation between genotoxic and non-genotoxic carcinogens. Task group 5. Mutat.
Res. 133: 1-49.
U.S. Environmental Protection Agency. (USEPA). 1976. Interim procedures and guidelines for health
risk and economic impact assessments of suspected carcinogens. Fed. Regist. 41: 21402-21405.
U.S.Environmental Protection Agency (USEPA). 1982a. Report on workshop on estimating ambient
water quality criteria for epigenetic carcinogens. Cincinnati, OH: Environmental Criteria Assessments
Office. (February 17).
U.S. Environmental Protection Agency (USEPA). 1982b. Report on workshop for the review of the use
of uncertainty factors in developing ambient water quality criteria for epigenetic carcinogens.
Cincinnati, OH: Environmental Criteria Assessments Office. (March 12).
U.S. Environmental Protection Agency (USEPA). 1986. The risk assessment guidelines of 1986.
EPA/600/8-87/045.
U.S. Environmental Protection Agency (USEPA). 1988a.. Proposed guidelines for assessing female
reproductive risk. Notice. Fed register 53:24834-24847.
U.S. Environmental Protection Agency (USEPA). 1988b.. Proposed guidelines for assessing male
reproductive risk and request for comments.. Notice. Fed register 53:24850-24869.
U.S. Environmental Protection Agency (USEPA). 1991. Guidelines for Developmental Toxicity Risk
Assessment. Fed Register 58(234):63798-63826.
U.S. Environmental Protection Agency (USEPA). 1992a. Guidelines for exposure assessment. Fed.
Register 57(104):22888-22938.
29
-------
FINAL DRAFT- Do Not Cite or Quote
U.S. Environmental Protection Agency (USEPA). 1992h. Draft report: A cross-species scaling factor for
carcinogen risk assessment based on equivalence of mg/kg3/4/day. Fed. Register 57(109):24152-24173.
U.S. Environmental Protection Agency (USEPA). 1992c. Dermal Exposure Assessment: Principles and
Application. EPA/600/8-91/01 IB. Office of Health and Environmental Assessment, Washington D.C.
U.S. Environmental Protection Agency (USEPA). 1994a. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. EPA/600/8-90/066F.
U.S. Environmental protection Agency (USEPA). 1994b. Report on the workshop on cancer risk
assessment guidelines issues. Risk Assessment Forum. USEPA: Washington, DC
U.S. Environmental Protection Agency (USEPA). 1995a. Proposed guidelines for neurotoxicity risk .
assessment. Fed. Register 60(192):52032-52056. -
U.S. Environmental Protection Agency (USEPA ).1995b. Report of the SAB on the draft Reproductive
Guidelines.
U.S. Environmental protection Agency (USEPA). 1995c. The Use of the Benchmark Dose Approach in
Health Risk Assessment. EPA Document EPA/63 O/R-94/007.
U.S. Environmental protection Agency (USEPA). 1996a. Proposed guidelines for carcinogen risk
assessment. Fed. Register 61(79): 17960-18011.
U.S. Environemntal protection Agency (USEPA). 1996b. Reproductive Toxicity Risk Assessment
Guidelines. Fed. Register 61(212):56274-56322. .
U.S. Environmental Protection Agency (USEPA).1996c. Draft benchmark dose technical guidance
document. EPA/600/P-96/002A.
U.S. Environmental Protection Agency (USEPA) .1996d. Benchmark dose peer consultation workshop.
Holiday Inn Bethesda, Bethesda, MD, September 10-11, 1996. Washington, DC: U.S. Environmental
Protection Agency.
Vainio, H., Magee, P.N., McGregor, D.B. & McMichael, A.J. 1992 Mechanisms of carcinogenesis in
risk identification. IARC scientif. publ. no. 116. Lyon: International Agency for Research on Cancer.
Weil, C.S. & McCollister, D.D. 1963 Relationship between short-and long-term feeding studies in
designing an effective toxicity test. Agric. Fd. Chem. 11: 486-491.
Weisburger, J.H. & Williams, G.M. 1981 Carcinogen testing: Current problems and new approaches.
Science. 214:401-407.
30
-------
APPENDIX B
Participant and Observer Lists
-------
United States Environmental Protection Agency
Office of Research and Development
Risk Assessment Forum
Framework for Human Health
Risk Assessment Colloquia Series
Colloquium #2
Holiday Inn Bethesda
Bethesda, MD
June 3-4, 1998
Participant List
Barbara Abbott
Research Toxicologist
Developmental Biology Branch
Reproductive Toxicology Division
U.S. Environmental Protection Agency
(MD67)
Research Triangle Park, NC 27711
919-541-2753
Fax: 919-541-4017
E-mail: abbott.barbara@.epa.gov
Charles Abernathy
Health and Ecological Criteria Division
Office of Water
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-5374
Fax: 202-260-1036
E-mail: abernathy.charles@.epa.gov
Karl Baetcke
Senior Scientist
Health Effects Division
Office of Pesticide Programs
U.S. Environmental Protection Agency
401 M Street, SW (7509-C)
Washington, DC 20460
703-305-7397
Fax: 703-305-5147
E-mail: baetcke.karl@epa.gov
Donald Barnes
Staff Director, Science Advisory Board
U.S. Environmental Protection Agency
401 M Street, SW (1400)
Washington, DC 20460
202-260-4126
Fax: 202-260-9232
E-mail: barnes.don@epa.gov
Bob Benson
Toxicologist
Drinking Water
U.S. Environmental Protection Agency
999 18th Street-Suite'500
Denver, CO 80202-2466
303-312-7070
Fax: 303-312-6131
E-mail: benson.bob@epa.gov
Ethel Brandt
Biologist
Existing Chemical Assessment Branch
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-2953
Fax: 202-260-1216
E-mail: brandt.ethel@epa.gov
Ann-Marie Burke
Toxicologist
Technical Support Branch
Office of Site Remediation
and Restoration
U.S. Environmental Protection Agenq
JFK Federal Building (HBS)
Boston, MA 02203-0001
617-223-5528
Fax: 617-573-9662
E-mail: burke.ann-marie@epa.gov
Chao Chen
Statistician
Quantitative Risk Method Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8602)
Washington, DC 20460
202-564-3244
Fax: 202-565-0079
E-mail: chen.chao@epa.gov
UX' Printed on Recycled Paper
-------
Eic Clegg
Reproductive lexicologist
Effects Identification and
Characterization Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-W)
Washington, DC 20460
202-564-3297
. Fax: 202-565-0078
E-mail: clegg.eric@epa.gov
Jim Cogliano
Chief, Quantitative Risk Methods
Office of Reasearch and Development
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3269
Fax: 202-565-0079
E-mail: cogliano.jim@epa.gov
Marion Copley
Health Effects Division
Office of Pesticide Programs
U.S. Environmental Protection Agency
401 M Street, SW (7509-C)
Washington, DC 204600
703-305-7434
Fax: 703-305-5147
E-mail: copley.marion@,epa.gov
Kevin Crofton
Neurotoxicologist
Neurotoxicology Division
National Health and Environmental
Effects Research Laboratory
U.S. Environmental Protection Agency
(MD 74B)
Research Triangle Park, NC 27711
919-541-2672
Fax: 919-541-4849
E-mail: crofcon.kevin@epa.gov
Kerry Dearfield
Science Administrator
Office of Science Policy
U.S. Environmental Protection Agency
401 M Street, SW (8103-R)
Washington, DC 20460
202-564-6486
Fax: 202-565-2925
E-mail: dearfield.kerry@.epa.gov
Lois Dicker
Existing Chemical Assessment Branch
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-3387
Fax: 202-260-1216
E-mail: ldicker@epa.gov
Janine Dinan
Environmental Health Scientist
Office of Emergency and
Remedial Response
U.S. Environmental Protection Agency
401 M Street, SW (5202-G)
Washington, DC 20460
703-603-8824
Fax: 703-603-9133
E-mail: dinan.janine@epa.gov
Vicki Dellarco
Senior Geneticist
Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-7336
Fax: 202-260-1036
E-mail: dellarco.vicki@epa.gov
Julie Du
Toxicologist
Office of Water
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-7583
Fax: 202-260-1036
E-mail: du.julie@epa.gov
Penelope Fenner-Crisp
Special Assistant to the
Assistant Administrator
Office of Prevention, Pesticides,
and Toxics Substances
U.S. Environmental Protection Agency
401 M Street, SW (7101)
Washington, DC 20460 "
202-260-0947
Fax: 202-260-1847
E-mail: fenner-crisp.penelope@epa.gov
Terry Harvey
Director
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
26 West Martin Luther King
Drive (MS-117)
Cincinnati, OH 45268
513-569-7531
Fax: 513-569-7475
E-mail: harvey.terry@epa.gov
Oscar Hernandez
Chemical Screening and Risk
Assessment Division
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1835
Fax:202-260-1283
E-mail: hernandez.oscar@epa.gov
Richard Hill
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7101)
Washington, DC 20460
202-260-2894
Fax: 202-260-1847
E-mail: hill.richard@epa.gov
Kim Hoang
Environmental Engineer
Quantitative Risk Methods Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
202-564-3303
Fax: 202-565-0079
E-mail: hoang.kim@epa.gov
Lee Hofmann
Environmental Health Scientist
Office of Emergency and
Remedial Response
U.S. Environmental Protection Agency
401 M Street, SW (5202-G)
Washington, DC 20460
703-603-8874
Fax: 703-603-9133
E-mail: hofmann.lee@epa.gov
-------
Annie Jarabek
lexicologist
Hazardous Pollutant
Assessment Branch
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
(MD-52)
Research Triangle Park, NC 27711
919-541-4847
Fax: 919-541-1818
E-mail: jarabek.annie@epa.gov
Jennifer Jinot
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3281
Fax: 202-565-0079
E-mail: jinot.jennifer@epa.gov
Carole Kimmel
Senior Scientist
Office of Research and Development
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3307
Fax: 202-565-0078
E-mail: kimmel.carole@epa.gov
Gary Kimmel
Effects Identification and
Characterization Division
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
202-564-3308
Fax: 202-565-0078
E-mail: kimmel.gary@epa.gov
Aparna Koppikar
Epidemiologist
Exposure Assessment and Risk
Characterization Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460 .
202-564-3242
Fax: 202-565-0076
E-mail: koppikar.aparna@epa.gov
Arnold Kuzmack
Senior Science Advisor
Office of Water
U.S. Environmental Protection Agency
401 M Street, SW (4301)
Washington, DC 20460
202-260-5821
Fax: 202-260-5394
E-mail: kuzmack.arnold@epa.gov
David Lai
Senior Toxicologist
Existing Chemical Assessment Branch
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-6222
Fax: 202-260-1279
E-mail: lai.david@epa.gov
Christopher Lau
Pharmacologist
Reproductive Toxicology
National Health and Environmental
Effects Research Laboratory
U.S. Environmental .Protection Agency
(MD67)
Research Triangle Park, NC 27711
919-541-5097
Fax:919-541-4017
E-mail: lau.christopher@epa.gov
Bob Luebke
Research Biologist
Immunotoxicology Branch
Experimental Toxicology Branch
U.S. Environmental Protection Agency
(MD92)
Research Triangle Park, NC 27711
919-541-3672
Fax: 919-541-4284
E-mail: luebke.robert@epa.gov
Amal Mahfouz
Senior Toxicologist
Office of Water
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-9568
Fax: 202-260-1036
E-mail: mahfouz.amal@epa.gov
Elizabeth Margosches
Statistician
Risk Assessment Division
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1511
E-mail: margosches.elizabeth@epa.gov
Marc Mass
Research Biologist
Biochemistry and Pathology Branch
Environmental Carcinogenesis Division
U.S. Environmental Protection Agency
(MD68)
Research Triangle Park, NC 27711
919-541-3514
Fax: 919-541-0694
E-mail: mass.marc@epa.gov
Robert McGaughy
Senior Scientist
Exposure Analysis and Risk
Characterization Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3244
Fax: 202-565-0079
E-mail: mcgaughy.robert@epa.gov
Edward Ohanian
Senior Science Advisor
Health and Ecological Criteria Division
Office of Science and Technology
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-7574
Fax:202-260-1036
E-mail: ohanian.edward@epa.gov
Gina Pastino
AAAS Risk Assessment Fellow
Office of Research and Development
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3372
Fax: 202-565-0059
E-mail: pastino.gina@epa.gov
-------
William Pepelko
lexicologist'
Effects Identification and
Characterization Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3309
Fax: 202-565-0078
E-mail: pepelko.william@epa.gov
James Rowe
Science Administrator
Office of Research and Development
Office of Science Policy
U.S. Environmental Protection Agency
401 M Street, SW (8103-R)
Washington, DC 20460
202-564-6488
Fax: 202-565-2925
E-mail: rowe.james@epa.gov
Daljit Sawhney
U.S. Environmental Protection Agency
4212 St. Jerome Drive
Annandale, VA 22003
202-260-0289
Louis Scarano
lexicologist
High Production Volume
Chemicals Branch
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1259
Fax: 202-260-1279
E-mail: scarano.louis@epa.gov
Rita Schoeny
Associate Director'
Health and Ecological Criteria Division
Office of Science and Technology
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-3445
Fax: 202-260-1036
E-mail: schoeny.rita@epa.gov
Cheryl Scott
Quantitative Risk Method Group
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8602)
Washington, DC 20460
202-564-3286
Fax: 202-565-0079
E-mail: scott.cheryl@epa.gov
Jennifer Seed
Branch Chief
Existing Chemicals Assessment Branch
Risk, Assessment Division
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1301
Fax: 202-260-1279
E-mail: seed.jennifer@epa.gov
William Sette
Senior Toxicologist
Science Analysis Branch
Health Effects Division .
Office of Pesticide Programs
U.S. Environmental Protection Agency
401 M Street, SW (7509-C)
Washington, DC 20460
703-305-6375
Fax: 703-305-5147
E-mail: sette.william@epa.gov
R. Woodrow Setzer
Methematical Statistician
Office of Research and Development
National Health and Environmental
Effects Research Laboratory
U.S. Environmental Protection Agency
(MD-55)
Research Triangle Park, NC 27711
919-541-0128
Fax: 919-541-4002
E-mail: setzer.woodrow@epa.gov
Mark Stanton .
Research Environmental
Health Scientist
Neurobehavioral Toxicology Branch
National Health and Environmental
Effects Research Laboratory
U.S. Environmental Protection Agency
(MD-74B)
Research Triangle Park, NC 27711
919-541-7783
Fax: 919-541-4849
E-mail: stanton.mark@epa.gov
Letty Tahan
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1301
E-mail: tahan.letty@epa.gov
Vanessa Vu
Associate Director of Health
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (8623-D)
Washington, DC 20460
202-564-3282
Fax: 202-565-0066
E-mail: vu.vanessa@epa.gov
John Whalan
Toxicologist
Registration Action Branch
Health Effects Division
U.S. Environmental Protection Agency
8650 Chase Glen Circle
Fairfax Station, VA 22039
703-305-6511
Fax: 703-503-5147
E-mail: whalan.john@epa.gov
Paul White
Exposure Assessment Group
Office of Research and Development
U.S. Environmental Protection Agency
401 M Street, SW (8603)
Washington, DC 20460
202-260-2589
E-mail: white.paul@epa.gov
Yin-Tak Woo
Senior Toxicologist
Risk Assessment Division
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Strteet, SW (7403)
Washington, DC 20460
202-260-0291
Fax: 202-260-1279
E-mail: woo.yin-tak@epa.gov
-------
Bill Wood
Executive Director
Risk Assessment Forum
. Office of Research and Development
National Center for
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (86601-D)
Washington, DC 20460
202-564-3358
Fax: 202-565-0062
E-mail: wood.bill@epa.gov
-------
AEPA
United States Environmental Protection Agency
Office of Research and Development
Risk Assessment Forum
Framework for Human Health
Risk Assessment Colloquia Series
Colloquium #2
Holiday Inn Bethesda
Bethesda, MD
June 3-4, 1998
s
Observer List
Ethel Brandt
Biologist
Existing Chemical Assessment Branch
Office of Pollution Prevention and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-2953
Fax: 202-260-1216
E-mail: brandt.ethel@epa.gov
Lois Dicker
Existing Chemical Assessment Branch
Office of Pollution Prevention and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-3387
Fax: 202-260-1216
E-mail: ldicker@epa.gov
Amal Mahfouz
Senior Toxicologist
Office of Water-
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
401 M Street, SW (4304)
Washington, DC 20460
202-260-9568
Fax: 202-260-1036
E-mail: mahfouz.amal@epa.gov
Daljit Sawhney
U.S. Environmental Protection Agency
4212 St. Jerome Drive
Annandale, VA 22003
202-260-0289
Letty Tahan
Office of Pollution Prevention
and Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW (7403)
Washington, DC 20460
202-260-1301
E-mail: tahan.letty@epa.gov
David Bennett
U.S. Environmental Protection Agency
40.1 M Street, SW (5202G)
Washington, DC 20460
703-603-8759
Fax: 703-603-9133
Printed on Recycled Paper
-------
APPENDIX C
Case Studies
-------
FRAMEWORK FOR HUMAN HEALTH RISK ASSESSMENT
Colloquium #1
Case Study: Ethylene Thiourea
Executive Summary
Ethylene thiourea (ETU) is a decomposition and metabolic product of the
ethylenebisdithiocarbamate fungicides. ETU is primarily used as an accelerator for volcanizing
elastomers. It is also used as a clearing agent in metallic electroplating baths, and as an
intermediate in the production of dyes, synthetic resins, antioxidants, and Pharmaceuticals.
Animal studies of ETU have demonstrated that the thyroid, pituitary, liver, and the developing
organism are major targets of ETU. The effects of ETU on the thyroid and pituitary have a
common mode of action. In addition, some of the anticipated effects of ETU on the developing
CNS also share aspects of this same mode of action. ETU has been found to be very weakly
genotoxic at best. In the thyroid, ETU inhibits the enzyme thyroid peroxidase in the presence of
iodide ion with concomitant oxidative metabolism to imidazoline and bisulfite ion. The
inhibition ceases upon consumption of ETU with no loss of enzymatic activity and negligible
covalent binding to thyroid peroxidase. Enzyme inhibition leads to a decrease in serum T3 and
T4 levels which leads to an increased secretion of TSH from the pituitary gland. This results in a
•hyperplastic, highly vascularized thyroid gland which if continued may lead to tumor
development. Animal studies have demonstrated the relationship between the dose and duration
of exposure to*ETU and the subsequent series of events that occur, namely decreased serum T4
levels and increased TSH levels, thyroid follicular cell hyperplasia, and subsequent tumor
formation. A similar sequence of events occurs in the pituitary leading to tumor formation.
When exposure to ETU is discontinued, this sequence of events is reversible up to a point, but
becomes irreversible after exposure to ETU at the appropriate dose and duration. It is also
apparent from these studies that while hyperplasia is a requisite step in the process of tumor
formation, the presence of hyperplasia does not necessarily mean that rumors will develop.
There are species differences in the administered dose required for this sequence of events; these
differences are due at least in part to species differences in the metabolism of ETU.
A similar mode of action is presumed to operate in the developing CNS. In the developing
organism, appropriate levels of thyroid hormone are required for normal development of the
brain. Although data regarding the potential effects of ETU on this aspect of brain development
are not available, we do know from studies of hypothyroid rats and human neonates that low T4
levels and elevated TSH levels lead to a variety of growth problems and deficits in brain
development. Studies of neonates born with congenital hypothyroidism, primarily due to iodine
deficiency, have shown that the nature and severity of the abnormalities is dependent on the
degree of thyroid hormone deficiency, as well as the duration of the deficiency; the abnormalities
1
-------
can be prevented if the level of thyroid hormone is corrected early enough.
Mice also can develop centrilobular hepatocellular cytomegaly and rumors when exposed to
ETU. This response has not been noted in rats. The mode of action is largely unknown, but the
species difference is thought to be due to the mouse specific metabolism of ETU via the flavin-
dependent mono-oxygenase (FMO) system. The FMO-binding of ETU metabolites to mouse
liver proteins may contribute to the chronic hepatotoxicity in mice.
When administered to pregnant rats, ETU can cause a variety of malformations of the CNS and
skeleton. The malformations can result following repeated doses, or following a single exposure
of a slightly higher dose. Similar malformations have only been noted in hamsters following
exposure to much higher doses. The malformations have not been observed in mice, rabbits,
guinea pigs, or cats even at doses 10-40 times that required to produce an effect in rats. Little is
known about the mode of action of the ETU-induced malformations. The observed species
difference is probably due to the wide species variability in the metabolism of ETU, as well as
some differences in the intrinsic sensitivity of the embryonic cells to ETU. It has been shown
that the malformations are due to ETU and not a metabolite of ETU. In contrast to the effects on
later CNS development that are thyroid dependent, the malformations are not the result of ETU-
induced hypothyroidism. Rather, ETU appears to cause necrosis of the undifferentiated
migrating neuroblast; the necrosis progresses with time and can lead to the formation of
hydrocephaly and other CNS malformations. While the necrosis is a requisite step in the
formation of the malformations, the developing rat can tolerate a certain degree of necrosis
without subsequent abnormal development. In vitro studies have also shown that ETU inhibits
the differentiation of limb bud cells, although at much higher concentrations than is required to
inhibit the differentiation of midbrain cells; whether this inhibition is related to the formation of
skeletal malformations is uncertain.
-------
Questions for Discussion
1) Given what is known about mode of action, are there commonalities among endpoints that
would be useful for quantitative analyses? For which endpoints should a common quantitative
analysis be conducted? For which endpoints should a separate analysis be conducted?
2) What additional information would be useful for quantitative analysis?
3) In the absence of this information, are any of the data sets presented in Tables 1-8 useful for
quantitative analysis? •
4) Are dose and duration of exposure important considerations? If so, for which endpoints and
how should they be handled?
5) In the absence of case-specific PBPK models, how should dose be adjusted for extrapolation
to humans? Does choice of a specific endpoint influence this decision? Why or why not?
6) What endpoint(s) would be useful for dose-response modeling in the observable range? Does
mode of action information influence this choice?
7) What quantitative method is recommended for low level exposures? Does this vary for •
different toxicities? Does mode of action information influence the choice of models?
8) If an RfD were to be calculated, does mode of action information influence choice of
uncertainty factors orinfluence uncertainties about data gaps?
-------
SUMMARY OF KEY STUDIES
A. THYROID
The thyroid is the major target organ of ETU. Several chronic studies have examined the
potential role of ETU in thyroid carcinogenicity, and these are described below. In addition,
several subchronic studies have been conducted to delineate the relationship between thyroid
hormone levels, thyroid hyperplasia, and thyroid tumors. These are described below in the
section on mechanistic studies.
Cancer Studies
Several studies have examined the potential role of ETU in thyroid carcinogenicity. Dietary
administration of 350 ppm ETU for 18 months resulted in a significant increase in thyroid
follicular cell carcinomas in male (15/26 versus 0/20 in controls) and female (6/26 versus 0/20 in
controls) CD rats; increases were not noted at 175 ppm. Another study also demonstrated
increased incidences of thyroid follicular cell adenomas in CD rats after administration of dietary
levels of 125 ppm for 1 or 2 years and follicular cell carcinomas at dietary levels of 250 and 500
ppm for 1 or 2 years; similar increases were not noted at levels of 5 or 25.
The NTP conducted a series "of studies in F344 rats and B6C3F1 mice to determine the potential
carcinogenicity of ETU.following standard 2-year adult exposures, combined perinatal and adult
exposures, arid perinatal exposures only. The data for male and female rats are summarized in
Tables 1 and 2, respectively. For the adult only exposures, rats were administered dietary
concentrations of 0, 83, or 250 ppm ETU and mice were administered concentrations of 0, 330,
or 1000 ppm for 2 years. Survival was adequate in all groups. Following 9 months of treatment
in rats, the incidence of follicular cell hyperplasia was significantly increased in both sexes of
both treated groups of rats, but there were no increases in adenomas or carcinomas. After 2 years
of treatment in rats, the incidence of thyroid follicular cell adenomas was significantly increased
in males at 83 and 250 ppm and in females at 250 ppm. The incidence of thyroid follicular cell
carcinomas was significantly increased in males and females at 250 ppm. T-4 levels were
decreased and TSH levels were increased at 83 and 250 ppm; the magnitude of change increased
from the 9 month to the 2 year time points.
In mice, exposure to 330 and 1000 ppm ETU resulted in significant increases in diffuse
cytoplasmic vacuolization of the follicular epithelium in both sexes following 9 months and 2
years of treatment. Follicular cell hyperplasia was not noted at 9 months; after 2 years of
exposure, hyperplasia was significantly increased in females at 330 ppm and in both sexes at
1000. Follicular cell adenomas were significantly increased in both sexes at 1000 ppm, but
carcinomas were significantly increased only in females at 1000 ppm!
For the "perinatal only "exposure portion of the study, female rats were exposed to 90 ppm and
female mice to 330 ppm for one week prior to mating through lactation and the pups were
-------
exposed to the same level from weaning until 8 weeks of age at which time exposure was
discontinued. In rats, there was a marginal increase in the incidence of follicular cell hyperplasia
in males and females; however, perinatal only exposure was not associated with thyroid tumors.
In mice, perinatal exposure was not associated with any lesions in the thyroid.
For the combined perinatal and adult exposures, FO female F344 rats were administered dietary
concentrations of 9, 30, or 90 ppm, and female B6C3F1 mice were administered concentrations
of 30,110, or 330 ppm ETU for one week prior to mating through weaning of the Fl litters. The
Fl pups were given the same diet as the dams from weaning until 8 weeks of age and then
administered adult (Fl) doses; the Fl adult concentrations were 25, 83, and 250 ppm for rats,
and 100, 330, and 1000 ppm for mice. The effect of perinatal exposure was determined by
comparing groups with varying FO concentrations and constant adult Fl exposure of 83 or 250
ppm in rats and 330 and 1000 ppm in mice.
In rats, there was no increase in thyroid tumors at the exposure level of 9:25, 30:83, or 90:83
ppm when compared to the 0:83 group. However, the incidence of follicular cell hyperplasia in
males, but not females, receiving 90:83 ppm was significantly higher than that in the 0:83 ppm
group. In contrast, comparison of the 0:250 and 90:250 ppm groups showed a significant
increase in follicular cell adenomas and carcinomas in males, and of carcinomas in females
exposed perinatally at 90 ppm.
In mice, combined perinatal exposure of 110 or 330 ppm with adult exposures of 330 or 1000
ppm was associated with increased incidences of thyroid lesions similar to those observed
following adult-only exposure. The only exception was a significant increase .in the incidence of
thyroid adenomas in females at 330:330 ppm compared to the 0:330 ppm group, and a marginal
increase in the incidence of follicular cell hyperplasia in males at 330:330 ppm as compared to
the 0:330 ppm groups.
Thyroid Mode of Action Studies
The increased sensitivity of the rat thyroid (compared to the mouse) to the effects of ETU is
thought to be due to metabolic differences in the two species. The effects of ETU on the thyroid
are thought to be mediated by the following mechanism. ETU inhibits the enzyme thyroid
pefoxidase which is required for the iodination of T-4. The pre-formed T-3 and T-4 continue to
be secreted, but as the supply is exhausted in the thyroid, the blood concentration decreases.
This results in increased secretion of TSH from the pituitary gland which produces a
hyperplastic, highly vascularized thyroid gland. However, this compensatory mechanism is
insufficient as long as ETU is present, and the cycle continues. Eventually the loss of thyroid
homeostasis leads to tumor development. The timing and nature of the thyroid lesions, as well as
the reversibility of the lesions, are dose and duration dependent. The following data are
generally consistent with this proposed mode of action.
-------
Toxicokinetics
ETU is rapidly absorbed from the gastrointestinal tract, is metabolized with little accumulation hi
the tissues except for the thyroid, and is eliminated rapidly. Elimination includes a substantial
amount of unmetabolized ETU. In rats, 18% of an oral dose was eliminated in the urine at 6
hours and 43% by 24 hours, and essentially all was excreted by 48 hours. In guinea pigs, 18%
was eliminated at 6 hours and 27% by 24 hours. The feces of rats and guinea pigs contained
negligible amounts of ETU. Similar results have been reported for mice, while, in monkeys,
55% was eliminated in the urine within 48 hours and 1.5% in the feces. ETU and its metabolites
have been found to have a half-life of about 28 hours in monkeys, 9-10 in rats, and 5 hours in
mice.
There are substantial species differences in the metabolism of ETU. Mice respond to ETU with
increased hepatic P450 and aniline hydroxylase activities, while these enzymes are markedly
reduced in the rat following ETU exposure. The mouse, but not the rat, metabolizes ETU via the
flavin-dependent mono-oxygenase system. The major metabolites identified in the urine of rats
24 hours after oral dosing were 63% unchanged ETU, 18% ethylene urea (EU), 5% imidazolone,
2% imidazoline, and 12% other metabolites. Mice metabolize ETU to EU and other unknown
metabolites.
ETU accumulates in the thyroid and the concentration of ETU and/or its metabolites in the
thyroid is dose-dependent; however, in rats the level of ETU in the thyroid does not increase
appreciably when the daily dose is increased above 50 mg/kg. Dietary administration for 1 week
with subsequent withdrawal of ETU from the diet led to an 80-94% reduction in the radioactivity
in the thyroid after 17 days.
Mutagenicity
With the exception of isolated positive responses reported with Salmonella typhimurium strain
TA1535, results of bacterial mutation studies with Escherichia coli and S. typhimurium have
been negative. Results, of studies with yeast showed some potential for induction of mitotic
aneuploidy, gene conversion, and DNA damage. No induction of sex-linked recessive lethal
mutations was observed in germ cells of male Drosophila melanogaster treated by feeding or
injection. In mammalian cells in vitro, ETU was negative for induction of chromosomal
aberrations, sister chromatid exchanges, and unscheduled DNA synthesis. Positive results were
reported in a mouse lymphoma assay for induction of trifiuorothymidine resistance in L5178Y
cells. In vivo tests for induction of micronuclei or sister chromatid exchanges in bone marrow
cells of mice were negative, as were tests for induction of dominant lethal mutations or sperm
abnormalities.
-------
Inhibition of Thyroid Peroxidase
In vitro studies have shown that ETU inhibits thyroid peroxidase. Inhibition of thyroid
peroxidase-catalyzed reactions by ETU occurs only in the presence of iodide ion with
concomitant oxidative metabolism to imidazoline and bisulfite ion. Inhibition ceases upon
consumption of ETU with no loss of enzymatic activity and negligible covalent binding of ETU
to thyroid peroxidase.
This mechanism is quite different from other inhibitors of thyroid peroxidase such as
methimazole and propylthiouracil which cause suicide inactivation via covalent binding to the
prosthetic heme group. Studies with PTU have shown that monkeys are much less sensitive than
rats; in an in vitro study, the concentration of PTU required to produce the same level of thyroid
peroxidase inhibition was 100 times greater for monkey enzyme than for rat enzyme. Since ETU
inhibits thyroid peroxidase by a different mechanism, it is not clear whether this species
difference would also apply to ETU.
Altered Hormone Levels and Thyroid Hyperplasia
Several studies have examined the relationship between dose and duration of treatment, and
altered thyroid homeostasis. Freudenthal et al (1977) administered Sprague-Dawley rats
(20/sex/group/time point) dietary concentrations of 0, 1, 5, 25, 125, or 625 ppm ETU for 30/60,
or 90 days. At 30, 60, and 90 days, serum T-3 and T-4, concentrations and thyroid 125I uptake
were measured; TSH concentrations were measured at 30 days. The data for the 30 and 90 day
time points are presented in Tables 3 and 4, respectively. Exposure to 625 ppm resulted in
mortality, reduced body weight, excessive salivation, rough and bristly hair coat, and scaly skin
texture. Serum T-3 and T-4 levels and 125I uptake were significantly reduced at 625 ppm at all
time points, and serum T-4 levels were significantly reduced at 125 ppm at all time points; TSH
levels were significantly increased at 30 days at 125 and 625 ppm. Absolute and relative thyroid
weights were significantly increased at 125 and 625 ppm at all time points, and histologic
examination showed mild (125 ppm) to moderate (625 ppm) follicular cell hyperplasia. In
addition, adenomas were noted after exposure to 625 ppm for 90 days.
Graham and Hansen (1972) administered male Osborne-Mendel rats dietary concentrations of 0.
50, 100, 500, or 750 ppm ETU for 30, 60, 90, or 120 days. At each sacrifice, 131I uptake was
measured 4 and 24 hrs post-injection and thyroid weights were recorded. Thyroids from animals
exposed for 90 days were examined histologically. 131I uptake at 4 hours post-injection was
significantly decreased at all time points for animals in the 500 and 750 ppm groups. There was
a dose-related decrease in 131I uptake at 24 hours post-injection at all exposure levels for all time
points; statistical significance was achieved at 500 and 750 ppm at all time points, and at 100
ppm after 30, 60, and 120 days. There was a dose-related increase in absolute and relative
thyroid weights at all dose levels at all time points; statistical significance was achieved at 100,
500, and 750 ppm at 30 days, at 500 and 750 ppm at 90 days, and at all levels after 60 and 120
days. Histological examination after 90 days of exposure showed no effects in the 50 ppm
-------
group and very slight to slight hyperplasia at 100 ppm. At levels of 500 and 750 ppm, there was
moderate to marked hyperplasia; adenomas were also noted at both levels.
Graham et al (1973, 1975) administered CD rats (68/sex/group) dietary concentrations of 0, 5,
25,125, 250, or 500 ppm for 2 years. Ten rats/sex/group were sacrificed at 2, 6, and 12 months.
At each time point, rats were injected with 131I and thyroid uptake was measured 24 hrs later. No
consistent pattern was noted. After 2 and 6 months of treatment, relative thyroid weights were
significantly increased in males at 250 and 500 ppm and in females at 125, 250, and 500 ppm.
After 12 months of treatment, relative thyroid weights were significantly increased in males and
females at 125, 250, and 500 ppm. At 6 months, thyroids were examined histologically from the
500 ppm group. The thyroids of all males were hyperplasic, and several adenomas and 1/5
carcinomas were present. Females also had hyperplasic thyroids and 2 had carcinomas. At 12
months, animals from all groups were examined histologically. The vascularity of the thyroid
was increased at all exposure levels in both sexes. Adenomas were present in males at 125 ppm
and carcinomas were evident at 250 ppm; in females carcinomas were evident at 500 ppm.
NTP conducted 13-week studies in F344 rats and B6C3F1 mice. Rats were administered dietary
concentrations of 0, 60,125, 250, 500, or 750 ppm ETU and mice were administered
concentrations of 0, 125,250, 500, 1000, or 2000 ppm. In the rat study, the final mean body
weights of male rats that received 500 or 750 ppm were 10 and 32% less than control,
respectively; the final mean body weights of females were reduced by 10% at concentrations of •
60-500 ppm and 28% at 750 ppm. Diffuse follicular cell hyperplasia was observed in all treated
animals, and focal follicular cell hyperplasia was observed at dose levels of 250 ppm and above
in males and at 750 ppm in females. In the mouse study, there were no treatment-related effects
on mean body weights. Diffuse follicular cell hyperplasia was significantly increased in both
sexes at doses of 500 ppm and greater.
In the range-finding study for the NTP's perinatal cancer studies, female rats were administered
dietary concentrations of 0, 8, 25, 83, or 250 ppm ETU for one week prior to mating through
weaning; the weanlings were administered these same concentrations from weaning to 8 weeks.
Mice were administered concentrations of 0, 33, 100, 330, or 1000 ppm ETU. In the rat study,
dose-related, minimal to moderate, diffuse follicular cell hyperplasia was observed in the male
weanlings at 25-250 ppm and in females at 83 and 250 ppm. In addition, follicular cell
adenomas were observed in 4/10 males at 250 ppm. In the mouse study, diffuse follicular cell
hyperplasia was noted in both sexes at 1000 ppm.
Reversibility
Several studies have been conducted to examine the reversibility of ETU-induced thyroid effects.
Rose et al (1980) reported that the effects of feeding rats 125-625 mg/kg/day for 2-12 weeks
resulted in dose-related suppression of T3 and T4 (with corresponding TSH elevation) and
thyroid hyperplasia. The hormone levels and thyroid hyperplasia were reversible within 22
weeks of placing on control diets. Arnold et al (1983) administered Sprague-Dawley rats diets
8
-------
containing 0, 75,100, or 150 ppm ETU for 7 weeks; animals were then examined after a
recovery period of 2, 3, or 4 weeks. After 7 weeks of exposure, linear dose X time contrasts
showed a significant dose-related linear increase in absolute and relative thyroid weight for both
sexes. Mean T4 blood levels were significantly reduced at 150 ppm (only group measured). The
linear increase in thyroid weight decreased in males after exposure ended, although a positive
dose-related effect was still apparent after 4 weeks. This same tendency toward reversibility of
the thyroid weight was noted in females at 2 and 3 weeks after exposure ended, but not at 4
weeks.
Graham et al (1975) administered CD dietary levels of 0, 5, 25,125, or 625 ppm ETU for 2
years. After 66 weeks of treatment, 3 animals from each group with palpable enlarged thyroid
glands were given the control diet for the remaining 38 weeks of the study. No indication of
morphological reversibility was observed in these animals, although the number examined was
small.
In the perinatal exposure portion of the NTP cancer bioassays, the FO dams were exposed to 90
ppm ETU for a week prior to mating until weaning; the Fl pups were then exposed to 90 ppm
until they were 8 weeks old at which time treatment was discontinued. When the animals were 2
years old, there was an increased incidence of follicular cell hyperplasia in males and females;
T3, T4, and TSH levels were comparable to control levels. Thus, it would appear that under
these exposure .conditions, some level of hyperplasia'is not reversible even when the hormone
levels are within the control range. Perhaps more importantly, it also appears that some level of
life time hyperplasia. at least in the unaltered'hormone levels, is not associated with tumor
formation.
B. PITUITARY
Hyperplasia and rumors of the pituitary gland pars.distalis have been observed in chronic mouse
studies. These lesions have not been noted following subchronic exposures. The cancer studies
are summarized belou.
Cancer Studies
In the NTP adult only exposure study, there was a significant increase in the incidence of focal
hyperplasia of the pituitary pars distalis in males at 1000 ppm and a significant increase in
adenomas in males and females at 1000 ppm. Lesions were not increased at the 9-month
sacrifice, or following exposure for 2 years to 330 ppm. There were no lesions associated with
perinatal only exposure to 330 ppm. Combined perinatal exposure of 110 or 330 ppm with adult
exposure to 330 or 1000 ppm ETU was associated with an incidence of pituitary lesions similar
to that observed with the adult only exposures. Therefore, perinatal exposure did not appear to
increase the incidence of pituitary lesions.
-------
Pituitary Mode of Action Studies
The pituitary lesions are thought to be caused by the anti-thyroid action of ETU. Cells in the
pars distalis secrete TSH, and the continued stimulation of these cells by low levels of circulating
T3 and T4 probably leads to hyperplasia and tumor formation. The reason that mice, but not rats,
are affected is unknown.
C. LIVER
Centrilobular hepatocyte cytomegaly has been observed in rats and mice exposed to high doses
of ETU, and liver tumors have been observed in mice. The studies are summarized below.
Subchronic Studies
'In the NTP 13-week study, F344 rats (10/sex/group) were administered dietary concentrations of
0, 60, 125, 250, 500, or 750 ppm ETU and B6C3F1 mice (10/sex/group) were administered
dietary concentrations of 0, 125, 250,500,1000, or 2000 ppm ETU. In the rat study,
centrilobular hepatocyte cytomegaly was observed in 7/10 males and 10/10 females at 750 ppm.
In the mouse study, it was observed in 4/10 females and 10/10 males at 500 ppm, and in all
animals exposed to 1000 or 2000 ppm. No tumors were observed in either species.
Cancer Studies
Innes et al (1969) examined the carcinogenic potential of ETU in 2 strains of hybrid mice,
C57B1/6 X c3h hybrid and C57B1/6 X AKR hybrid. The animals were given 215 mg/kg ETU
by gavage daily from day 7 to day 28 after birth, and then given dietary concentration of 646
ppm ETU for 18 months. Increased incidences of hepatomas were observed in males and
females.
Liver lesions observed in the NTP study in B6C3F1 mice are summarized in Table 5. In the
adult only exposures, centrilobular cytomegaly was significantly increased in males and females
at 1000 ppm following 9 months exposure; it was observed in males at 330 and 1000 ppm and in
females at 330 ppm after 2 years exposure. The incidence of hepatocellular carcinomas was
significantly increased in males and females at 330 and 1000 ppm following 2 years of exposure.
There were no liver lesions associated with perinatal only exposure to 330 ppm.
\
There were no liver lesions associated with a combined perinatal and adult exposure level of
33:100 ppm ETU. Perinatal exposure to 110 or 330 ppm combined with adult exposure to 330 or
1000 ppm was associated with an increased incidence of centrilobular cytomegaly and
hepatocellular carcinomas similar to that observed with the adult only exposures. Therefore,
perinatal exposures did not appear to increase the incidence of liver lesions.
10
-------
Liver Mode of Action Studies
Little is known about the mode of action of ETU-induced liver tumors in mice. It has been
suggested that the species difference noted between rats and mice is due to the mouse specific
metabolism of ETU via the flavin-dependent mono-oxygenase (FMO) system. The FMO-
binding of ETU metabolites to mouse liver proteins may contribute to the chronic hepatotoxicity
in mice.
D. MISCELLANEOUS TUMORS
In the NTP study that examined the combined perinatal and adult exposures, there was an
increased incidence of Zymbal gland neoplasms and mononuclear cell leukemia that were not
observed following the adult only or perinatal only exposures. A significant increase in Zymbal
gland neoplasms were observed in 5/50 males at 90:250 ppm versus 1/50 in the control group (p
< 0.05). Mononuclear cell leukemia was observed in 35/50 males at 90:83 and 29/50 males at
90:250 ppm versus 22/50 controls (p < 0.05); it was observed in 25/50 females at 90:250 ppm
versus 18/50 controls (p < 0.05). No information is available concerning the mode of action.
E. DEVELOPMENTAL EFFECTS UNRELATED TO THYROID
Prenatal Studies
Numerous oral prenatal developmental toxicity studies of ETU have been conducted. ETU has
been shown to be a potent developmental toxicant in rats causing CNS malformations at low
doses and limb abnormalities at higher doses. In the rat, ETU causes necrosis of the CNS which
progresses into hydrocephaly and a variety of other defects. Most of the hydrocephalic pups die
during postnatal life, and those that survive exhibit motor impairment and a hopping gait. The
defects have been observed following single as well a*s repeated oral doses, and they occur in the
absence of any toxic effects on the dam. There are tremendous species differences in the
magnitude and type of developmental effects induced by ETU. A similar spectrum of effects has
only been observed in hamsters at much higher doses than needed in the rat. Prenatal exposure
to ETU causes little or no developmental toxicity in cats, guinea pigs, mice, or rabbits, and
effects have only been noted following exposure to high doses of ETU. In the rat, developmental
toxicity has been noted following a single oral dose of 30 mg/kg and a repeated dose of 10-20
mg/kg/day on gestation days 6-15. In contrast, the lowest reported developmentally toxic dose
of ETU in mice is 1600 mg/kg for a single dose and 200 mg/kg/day for repeated dosing on days
6-15 of gestation. In hamsters, developmental effects have been noted following a single oral
dose of 1200 mg/kg and repeated doses of greater than 100 mg/kg/day on gestation days 4-9.
The prenatal studies are summarized in Table 6.
Mode of Action Studies
Little is known regarding the mode of action of the ETU-induced malformations. These effects
11
-------
do not seem to be mediated by altered maternal thyroid function. ETU (and not one of its
metabolites) is the compound responsible for the malformations. The species differences appear
to be related mainly to differences in maternal metabolism, but there are also minor differences
in the susceptibility of the embryonic cells among rats and mice, the only species examined. The
migrating undifferentiated neuroblasts appears to be a target cell of ETU; ETU causes necrosis of
these cells, but the mechanism leading to the necrosis is unknown. A certain amount of necrosis
has been observed at doses that do not lead to malformations. These conclusions are supported
by the following studies.
/ ^
Toxicokinetics
i
When 14C-ETU was administered to pregnant rats, radioactivity was uniformly dispersed
between plasma and erythrocytes of the mother, and the radiolabel binding of erythrocytes was
found to be reversible. The radioactivity in the embryo was dispersed uniformly with no
evidence of binding to DNA, RNA, or protein of the embryo. In pregnant rats orally dosed on
day 15 of pregnancy with 240 mg/kg 14C-ETU resulted in maximum radioactivity in maternal
blood 0.5 to 2 hr postdosing with 93% of the activity associated with the cellular membrane of
erythrocytes. .Radioactivity in the whole fetus was similar to the radioactivity in maternal tissues
at 6 and 12 hr postdosing, but negligible 24 hr after dosing. An oral dose on day 12 of pregnancy
was readily absorbed with a peak radioactivity level at 2 hr postdosing; none appeared to be
incorporated into the crude protein fraction of maternal serum. By 24 hr, the activity in the urine
accounted for about 80% of the administered dose. The radioactivity in the fetus peaked at 2 hr
postdosing and showed a 18-fold reduction at 24 hr; none was incorporated into the fetus protein
fraction.
In pregnant mice orally dosed with 240 mg/kg 14C-ETU, the concentrations in the maternal
tissues, fetus, and placenta were similar at 3 hr postdosing. At 6 and 12 hrs, the radiolabel in the
maternal tissues declined, but at a rate faster than the decline of radiolabel in the maternal rat.
After dosing pregnant mice with 240 mg/kg ETU containing 100 uCI/kg of 14C-ETU and
pregnant rats with 240 me Ten ETU containing 50 uCI/kg 35S-ETU, values for tl/2 of ETU
elimination, postdosing time of peak level occurrence, and peak level of radioactivity in the
blood were 5.5 and 9.4 hr. 1.3 and 1.4 hr, and 136.7 and 214.6 |ig/g for mice and rats,
respectively. x
As stated above in relation to the thyroid,, there are substantial species differences in the
metabolism of ETU. Mice respond to ETU with increased hepatic P450 and aniline hydroxylase
activities, while these enzymes are markedly reduced in the rat following ETU exposure. The
mouse, but not the rat, metabolizes ETU via the flavin-dependent mono-oxygenase system. The
major metabolites identified in the urine of rats 24 hours after oral dosing were 63% unchanged
ETU, 18% ethylene urea (EU), 5% imidazolone, 2% imidazoline, and 12% other metabolites.
Mice metabolize ETU to EU and other unknown metabolites. The major metabolites identified
in the urine of cats 24 hours after oral dosing were 28% unchanged ETU, 4% EU, and 64% S-
methyl ETU.
12
-------
Toxicokinetic differences may in part account for the species differences in the teratogenic
potential of ETU. The absence of a clear teratogenic activity in the cat may be due to conversion
of most of the ETU to S-methyl ETU and in the mouse may be due to a shorter half-life with
faster metabolic degradation of ETU. However, there also appear to be toxicodynamic
differences, at least between rats and mice, which may also contribute to the species difference in
response to ETU. The studies supporting this are described below.
Potential Role of Maternal Thyroid
The inhibitory effect of ETU on the thyroid raised the possibility that functional alteration of the
maternal thyroid may be the cause of the ETU-induced malformations. To investigate this
possibility, Lu and Staples (1978) administered euthyroid and hypothyroid
(thyroparathyroidectomized) rats 40 mg/kg/day ETU on days 6-15 of gestation. The results
indicated that the malformations produced after administration of ETU were not due to alteration
of maternal thyroid function. However, maternal hypothyroidism increased the background level
of malformations, and altered the spectrum of malformations in response to ETU both
qualitatively and quantitatively.
Identification of Proximate Teratogen
Several studies have been conducted to ascertain whether ETU or a metabolite of ETU is
responsible for the malformations in the rat. In one study, the effects of pretreatment with
metabolic modifiers on ETU-induced malformations was examined. None of the modifiers was
associated with developmental tdxicity when administered alone. Pretreatment of pregnant rats
on gestation day 13 with SKF-525A; an inhibitor of P-450, increased the incidence of
malformations compared to those given 60 mg/kg ETU alone, and were similar to that observed
following treatment with 120 mg/kg ETU. Pretreatment with 40, 60, or 80 mg/kg of sodium
pentobarbital injected once or twice daily on days 9-12, or 20 mg/kg/day of methyl cholanthrene
on days 11-13, failed to alter the effects of 60 mg/kg ETU given orally on gestation day 13.
These results suggest that ETU, and not a metabolite, are the cause of the malformations in rats.
In vitro whole embryo culture studies are also consistent with the suggestion that ETU is the
proximate toxicant. The role of maternal metabolism in modifying the teratogenicity of ETU
was assessed by adding hepatic S-9 fractions from Aroclor 1254-induced rats and mice to whole
embryo culture. Rat S-9 had no effect on ETU teratogenicity, but mouse S-9 virtually eliminated
the formation of abnormalities typical of ETU. ETU-typical defects were observed in embryos
exposed to ETU and mouse S-9 which had been treated with carbon monoxide to inactivate its
monooxygenase system, indicating that the mouse S-9 was metabolizing ETU. Other whole
embryo culture studies have shown similar results.
Target Cells
Since ETU causes several CNS malformations, several in vivo and in vitro studies have been
13
-------
conducted to determine whether neuronal cells are a specific target of ETU. In one study,
pregnant rats were given an oral dose of 15 or 30 mg/kg on gestation day 12. Some dams were
sacrificed 12, 24, 48 and 72 later; the remaining were allowed to litter and the pups were
followed until 80 days postnatally. Fetuses in the 30 mg/kg group had karyorrhexis in the
germinal layer of the basal lamina of CNS extending from the thoracic spinal cord to the
telencephalon 12 hours after treatment. The initial degenerative changes were observed in the
undifferentiated migrating neuroblast. By 48 hours, the spinal cord showed obliteration and
duplication of the central canal and disorganization of the germinal and mantle layers. In the
brain, the ventricular lining was focally denuded and the nerve cell proliferation was
disorganized. By 80 days postnatally, 50% of the pups had hydrocephaly and died. Fetuses in
the 15 mg/kg group had cellular degeneration in the CNS that was restricted to single cells or
small groups of cells widely dispersed in the germinal layer. These changes did not result in
hydrocephaly or mortality indicating that a certain degree of necrosis was compatible with
further development.
In another study, ETU, orally administered as a single 30 or 45 mg/kg dose on gestation day 18,
was found to induce necrosis of neuroblasts in the fetal CNS after 18 and 24 hrs of dosing and a
high incidence of hydrocephalus in postnatal pups at both doses. ETU was then administered at
a single dose of 80 or 120 mg/kg on day 18, and rat fetal brains were trypsinized and dissociated
into a cell suspension. The total number of viable cells was significantly reduced at both dose
levels. When grown as monocell layers, there was a marked decrease in neuronal cells and
increase in non-neuronal cells as compared to the controls.
In another study, Wistar rat embryos were cultured from days 11-13 of gestation in the presence
of 30, 150, or 300 ug/mL or midbrain (MB) and limbbud (LB) cells were cultured at
concentrations ranging from 30 to 600 fig/ml. There were dose-related increases in the incidence
of head, limb, face, and tail abnormalities in the whole embryo culture experiments. In the cell
culture experiments, there was a dose-related increase in the inhibition of the differentiation of
MB and LB cells; MB cells were far more sensitive to ETU than LB cells.
Toxicodynamic Differences Between Rats and Mice
Although most of the species differences can be attributed to differences in metabolism, there is
some evidence that rat embryos are intrinsically more sensitive to ETU than mouse embryos. An
in vitro study examined the effects of ETU on the differentiation of midbrain cells from rat and
mouse embryos. Differentiation of mouse MB cells was inhibited at concentrations of ETU 3-10
fold higher than needed to inhibit rat midbrain cells. A whole embryo culture study showed that
mouse embryos are susceptible to ETU-induced teratogenesis, manifesting the same types of
abnormalities observed in rats, at concentrations 2-3 times higher than those needed to produce a
comparable spectrum of effects in rats. Thus, there is some difference in the sensitivity to ETU,
but it is of insufficient magnitude to fully account for the 10-40 fold difference reported in vivo
studies.
14-
-------
F. DEVELOPMENTAL EFFECTS RELATED TO THYROID
Animal Studies
Range-finding developmental toxicity studies were conducted as part of the series of chronic
studies conducted by the NTP (summarized in Table 6). In the rat study, F344 female rats were
administered dietary concentrations of 0, 8, 25, 83, or 250 ppm ETU for one week prior to
mating until weaning; the weanlings were then administered the same dietary concentrations
until scheduled sacrifice at 8 weeks of age. Some of the dams were sacrificed on gestation day
18 for evaluation of potential prenatal effects. There were no effects on the number of
implantations, litter size, number of live or dead fetuses, mean fetal weights, external
abnormalities, or placental weights in the treated groups. There was reduced survival from day
0-4 postnatally at 250 ppm. Dose-related thyroid follicular cell hyperplasia was observed in
males at 25-250 ppm and in females at 83 and 250 ppm. Thyroid follicular cell adenomas were
observed in 4/10 males at 250 ppm, and cytoplasmic vacuolization of the pituitary pars distalis
was noted in 7/10 males at 250 ppm.
A similar study was conducted in B6C3F1 mice except concentrations of 0, 33,100, 330, and
1000 ppm ETU were administered. No developmental effects were noted in litters examined at
day 17 of gestation. Postnatal survival from day 7-28 was reduced at 1000 ppm. Thyroid
follicular cell hyperplasia was noted in 7/10 males and females at 1000 ppm; and centrilobular
hepatocellular cytomegaly was noted in 8/10 males and 7/10 females at 1000 ppm.
Mode of Action Studies
The mode of action of the thyroid and pituitary lesions is probably the same as that described
above for the adult lesions. However, it is not known whether the lesions are due to exposure
prenatally, postnatally. or a combination of prenatal and postnatal exposure. Studies of normal rat
development have shown that radioiodine is concentrated in the thyroid follicle by 17 days of '
gestation, and by 20 days the gland is actively synthesizing thyroid hormones. Consequently,
serum T4 levels increase appreciably between day 20 and parturition. At birth, serum TSH
levels are low. but rise rapidly to reach a peak at about 8 days postpartum. Likewise, serum T4
levels are low at birth but rise to a peak at about 15 days postpartum before falling to adult levels.
Thus, it is likely that the thyroid effects observed in the NTP studies are due primarily to
postnatal exposure to ETU. However, it should be noted that in humans the maturation of the
thyroid occurs much earlier in gestation. Fetal thyroid hormone synthesis probably begins
around 10-12 weeks of gestation and then increases throughout the remainder of gestation.
Serum T3 has been detected as early as 15 weeks and then continues to rise. TSH levels remain
low until 20 weeks and then plateaus for the remainder of gestation. Thus, given these species
differences, it is likely that the thyroid effects of ETU may be more important in humans than
rats during gestation.
15
-------
Other Potential CNS Effects Due to Altered Thyroid Status
Although no studies have been conducted in humans or animals to assess the potential thyroid
mediated effects of ETU on the developing nervous system, it is quite likely that the CNS may
be affected given what is known about the role of thyroid hormone in development, as well as
known developmental consequences of hypothyroidism in humans. The critical period during
which appropriate thyroid hormone levels are absolutely essential for normal brain development
is from 18 days of gestation until 21 days ppstpartum in the rat, and late gestation until 1-2 yrs in
humans. In the cerebrum, this period is associated with proliferation of axons and dendrites,
synapse formation, gliogenesis, and myelination. In the cerebellum, this period encompasses all
of the above events as well as the majority of cell proliferation. Hypothyroidism during this
period can cause serious damage to the development and organization of the brain. There is not
general agreement whether thyroid hormone plays an important role in CNS development prior
to this. The precise role of maternal thyroid hormone in the developing nervous system is not
well established, and it is therefore possible that altered maternal thyroid homeostasis may cause
some effects prior to the established "critical period".
Untreated congenital hypothyroidism is known to have severe effects on neurological
development. The severity of the effects are correlated with the magnitude of the deficiency, the
apparent time of onset of the deficiency, and the age at which appropriate replacement therapy is
begun. In addition to the mental retardation and growth retardation that can occur, other
common neurological disorders include deafmutism, poor coordination and balance, abnormal
fine motor movements, speech problems, spasticity, tremor, and hyperactive deep tendon
reflexes. In humans, development is largely corrected if replacement therapy begins at birth;
however, there are some studies suggesting that children with severe congenital hypothyroidism
may have some behavioral and learning disabilities even when replacement therapy begins at
birth. The American Thyroid Association Guidelines for Newborn Screening recommends the
following screening and treatment for newboms. Newborns with T4 levels less than the 10th
percentile should have a TSH assay. Newborns with low T4 levels (2 SD below the mean for the
normal range, usually below 10 ng/dL) and normal TSH yames seldom have thyroid
insufficiency. However, there are some cases with low T4 levels and delayed TSH increase so
these newborns should be followed. Infants with low T4 and elevated TSH (greater than 40
mU/L2) are considered hypothyroid until proved otherwise.
16
-------
TABLE 1
Thyroid Hormone Levels and Thyroid Lesions in Male F344 Rats
. Following 9 Months and 2 Years Exposure to ETU
Endpoint
9 months
T3 (ng/dL)
T4 (ug/dL)
TSH (ng/mL) ,
Follicular
Cell
Hyperplasia
Follicular
Cell
Adenoma
FO:F1 Concentration (ppm)
0:0 90:0 9:25 0:83 30:83 90:83 0:250 90:250
101±9
5+0.3
211+24
0/10
0/10
104 + 5
5.1+0.3
221 ±27
4/9
0/9
NA
NA
NA
1/10
0/10
88 ±6
3.3 ±0.2*
261 ±37
10/10*
0/10
70 ±4*
3.4 ±0.2*
308 + 31
8/10*
0/10
65 ±7*
3.3 ±0.2*
325 ± 70
10/10*
0/10
97±4
3.2 ±0.1*
340 ±65
10/10*
0/10
95+4
2.7 ±0.1*
331 ±33
10/10*
3/10
2 years
T3 (ng/dL)
T4 (ug/dL)
TSH (ng/mL)
Follicular
Cell
Hyperplasia
Follicular
Cell
Adenoma
Follicular
Cell
Carcinoma
75 ±4 .
3.1 ±0.3
241+23
4/49
(1.3)
0/49
1/49
81+5
3 ±0.2
240 ± 30
12/49*
(1.3)
1/49
3/49
NA
NA
NA
13/46
1/46
2/46
93 ±7
2.4 ±0.2
308±55
30/46*
(2.1)
9/46*
3/46
84 ±9
1.9 + 0.2*
744+148*
35/47*
(2.1)
10/47*
4/47
87 ±7
2.1 ±0.3*
984 ± 234*
47/50*
(2.1)
8/50*
6/50*
75 ±10
1.8±0.1*
2874 ± 729*
41/50*
(3.9)
23/50*
26/50*
51±6*
1.8 ±0.3*
1543 ±
826*
39/50*
(3.5)
34/50*
44/50*
* Statistically different from the 0:0 ppm group
17
-------
TABLE 2
Thyroid Hormone Levels and Thyroid Lesions in Female F344 Rats Following 9 Months and 2 Years Exposure to ETU
Endpoint
9 months
T3 (ng/dL)'
T4 (ug/dL)
TSH (ng/mL)
Follicular
Cell
Hyperplasia .
Follicular
Cell
Adenoma
FO:F1 Concentration (ppm)
0:0 90:0 9:25 0:83 30:83 90:83 0:250 90:250
150 ±6
4.1 ±0.2
162 ±8
0/10
0/10
167 ±8
4.1 ±0.2
' 178±9
0/9
0/9
NA
NA
NA
0/10
0/10
1 1 1 ± 6*
2 ± 0.2*
260 ± 27*
5/10*
0/10
107 + 7*
1.9 + 0.2*
288 + 26*
10/10*
0/10
120 ±5*
2.5 ±0.1* •
396 ±54*
10/10*
0/10
150 ±5
2.5 + 0.2*
241 ±22*
10/10*
0/10
117 + 8*
2.2 ± 0.2*
421 ±55*
10/10*
1/10
2 years
T3 (ng/dL)
T4 (ug/dL)
TSH (ng/mL)
Follicular
Cell
Hyperplasia
Follicular
Cell
Adenoma
Follicular
Cell
Carcinoma
109 + 7
2.9 + 0.2
338 ±48
0/50
1/50
2/50
145 + 15
2.9 ±0.2
236' ±34
8/48*
(1.3)
0/48
0/48
NA
NA
NA
15/49
0/49
1/49
.137 + 9
2.7 + 0.1
516±78
33/44*
(1.8)
6/44
1/44
113±9
2.5 ±0.2
511±27*
30/46*
(2.0)
5/46
1/46
124 + 4
2.6 ±0.2
629 ± 142*
41/47*
(2.1)
7/47
2/47
72 ±13* '
2.5 ±0.2
769 ±104*
45/49*
(2.7)
28/49*
8/49*
121 ±9
1.7±0.1*
1371 +
345*
47/50*
(2.8)
29/50*
17/50*
Statistically different from the 0:0 ppm group
18
-------
TABLE 3
Thyroid Hormone Levels and Thyroid Hyperplasia in Sprague-Dawley Rats
Following 30 Days of Exposure to ETU
Endpoint
Males
T3 (ng percent)
Males
T4 (ug percent)
Males
TSH (uUI/ml)
Females
T3 (ng percent)
Females
T4 (ug percent)
Females
TSH (uUI/ml)
ETU (ppm)
0 1 5 25 125 625
76 ±12
5 ±1.7
6.7 ±2.5
83 ±16
3.8 + 1.4
6 ±4.1
82 ±13
5.1±1
6.4 ±0.8
91±11
3.5 ±1.0
4.5 ±0.9
79 ±8
4.7 ±0.4
6.7 ±1.4
88 ±13
2.9 ±0.9
4.9 ±1.4
-67±16
5.6 ±1.1
7.3 ±1.5
86 ±15
3.8 ±0.8
5.1 ±1.3
71 ±12
2.6 ±0.4*
23.3 ±5.9*
104 ±16
2.1 ±0.5*
18.3 ±4*
57 ±4*
0.9 ±0.6*
14.3 + 0.9*
58 ±10*
1.1 ±1.0*
14.6 ±1.9*
Degree of Hyperplasia
(Sex not specified)
0
>1
1
>2
2
>3
3
4/23
9/23
10/23
0/23
0/23
0/23
0/23
0/20
7/20
13/20
0/20
0/20
0/20
0/20
1/20
6/20
13/20
0/20
0/20
0/20
0/20
0/20
9/20
11/20
0/20
0/20
0/20
0/20
0/20
0/20
18/20
1/20
1/20
0/20
0/20
0/20
0/20
2/20
4/20
12/20
2/20
0/20
19
-------
TABLE 4
Thyroid Hormone Levels and Thyroid Hyperplasia in Sprague-Dawley Rats
Following 90 Days of Exposure to ETU
Endpoint
Males
T3 (ng percent)
Males
T4 (ug percent)
Females
T3 (ng percent)
Females
T4 (ug percent)
ETU (ppm)
015 25 125 625
72 ±22
4.5 ±0.8
107 ±25
3.3 ±0.8
69+10
4.0 ±1
117±18
2.5 ±0.7
76 ±13
5±1
105 + 17
3 ±0.7
79 ±13
3.8 ±1
109+12
2.9 + 0.7
86 ±15
2.3 ±0.4*
•106 + 16
1.6 ±0.3*
28 ±14*
1.1 ±0.6*
35 ±4*
1.1 ±0.6*
Degree of Hyperplasia
(Sex not specified) . , .
0
>1
1
>2 .
2
>3
3
. 12/24
9/24
3/24
0/24
0/24
0/24
0/24
6/20
7620
8/20
0/20
0/20
-0/20
0/20
6/20
8/20
6/20
0/20
0/20
0/20
0/20
3/20
8/20
9/20
0/20
0/20
0/20
0/20
1/20
0/20
14/20
3/20
1/20
0/20
1/20
0/20
0/20
1/20
1/20
12/20
3/20
1/20
20
-------
TABLE 5
Liver Lesions in B6C3F1 Mice Following 9 Months and 2 Years Exposure to ETU
Endpoint
9 months
Males
Centrilobular
Cytomegaly
Males
Hepatocellular
Adenoma
Females
Centrilobular
Cytomegaly
Females
Hepatocellular
Adenoma
FO:F1 Concentration (ppm)
0:0 330:0 33:100 0:330 110:330 330:330 0:1000 330:1000
0/10
0/10
0/10
0/10
0/10
1/10
NA
NA
NA
NA
NA
NA
0/10
0/10
0/10
NA
8/10*
0/10
NA
NA
5/10*
0/10
NA
NA
10/10*
2/10
10/10*
2/10
10/10*
1/10
9/10*
1/10
2 years
Males
Centrilobular
Cytomegaly
Males
Hepatocellular
Adenoma
Males
Hepatocellular
Carcinoma
Females
Centrilobular
Cytomegaly
Females
Hepatocellular
Adenoma
Females
Hepatocellular
Carcinoma
0/49
11/49
13/49
0/50
2/50
2/50
1/49
6/49
8/49
0/50
1/49
5/49
6/33
6/33
4/33
2/29
2/28
2/28
36/50*
16/50
19/50
11/50
34/50
31/50*
33/50*
15/47
15/47
8/50
35/50
23/50*
29/49*
20/49
19/49
9/50
17/50
48/50*
25/50*
9/50
45/50*
0/50
14/50
47/50*
40/50*
15/49
45/49*
8/50
17/50
48/50*
Statistically different from the 0:0 ppm group
21
-------
TABLE 6
Summary of Developmental Toxicity Studies of ETU
Species
Sprague-
Dawley Rat
Wistar Rat
Wistar Rat
Wistar Rat
Dosing
Period
6-20
Gavage
6-15
Gavage
7-20
Gavage
21-42
days
before
pregnancy
until day
15 of
pregnancy
Gavage
Dose
0,5,10,20,
30, 40, 80
mg/kg/day
0,5,10,20,
40,80
0,5,10,20,40
'-
•
0,5,10,20,40
Maternal
Effects
80: 25%
mortality,
1 BW
80: mortality
None
None
Developmental Effects
< 10: J BW
< 20: hydrocephalus
< 30: Reduced
ossification
< 40: encephalocele
< 80: cleft palate,
kyphosis, micromelia,
hemimelia,
oligodactyly,
syndactyly, edema,
micrognathia
< 10: kinky tail
< 20: exencephaly,
abnormal flexion
hindlimb
< 40: micrognathia,
oligodactyly, short tail
< 80: coloboma eyelids,
hemimelia
< 10: exencephaly
< 20: hydrocephaly,
kinky tail, abnormal
flexion hindlimb
< 40: micrognathia,
short tail
< 20: exencephaly,
micrognathia, abnormal
flexion hindlimb, kinky
tail
<40: short tail
Reference
Chernoff
et al (1979)
Khera, 1973
Khera, 1973
Khera, 1973
-------
Species
Wistar Rat
Wistar Rat
Sprague-
Dawley Rat
F344 Rat
F344 Rat
CD-I
Mouse
Dosing
Period
12
Gavage
14
Gavage
18
Gavage
1 week
prior to
breeding
until
gestation
day 18
Diet
1 week
prior to
breeding
until 8
weeks
postnatal
Diet
6-15
Gavage
Dose
0,15,30
0, 15,30,45
0, 30, 45
0, 8, 25, 83,
250 ppm
0, 8,25,83,
250 ppm
0,100,200
mg/kg/day
Maternal
Effects
None
None
None
None
None
<100:I
relative liver
wt
Developmental Effects
15: no pup mortality by
80 days postnatal
30: 50% pup mortality
and hydrocephaly by 80
days
15: no pup morality by
80 days postnatal
30: 90% pup mortality
and hydrocephaly by 80
days; survivors had
motor impairment and
hopping gait.
45: 100% pup morality
and hydrocephaly with
4 weeks postnatal
30: hydrocephaly in
57/73 pups
45: hydrocephaly in
73/73 pups
None
-
25 ppm: 4/10 males
thyroid follicular cell
hyperplasia
< 83 ppm: all males and
females thyroid
follicular cell
250ppm: males
1 survival day 0-4; 4/10
males follicular cell
adenoma; 7/10 males
pituitary pars distalis
cell vacuolization
200: 1 supernumerary
ribs
Reference
Khera and
Tryphonas,
1985
Khera and
Tryphonas,
1977
Khera, 1987
NTP, 1992
NTP, 1992
Chernoff
et al (1979)
-------
Species
B6C3F1
Mouse
B6C3F1
Mouse
New
Zealand
White
Rabbit
Hartley
Guinea Pig
Golden
Hamster
Cat
Dosing
Period
1 week
prior to
breeding
until
gestation
davl?
Diet
1 week
prior to
breeding
until 8
weeks
postnatal
Diet
-
7-20
Gavage
6-24
4-9
Gavage
16-35
Gavage
Dose
0, 33, 100,
330, lOOOppm
0,33,100,
330, 1000 ppm
0, 10, 20, 40,
80 mg/kg/day
0,50, 100
mg/kg/day
0,25,50, 100
mg/kg/day
0,5,10,30,
60, 120
Maternal
Effects
None
None
None
None
None
< 10: JEW,
tremors,
hindlimb
paralysis,
mortality
Developmental Effects
None
1000 ppm: 1 survival
day 7-28; 7/10 males
and females thyroid
follicular cell
hyperplasia; 8/10 males
and 7/10 females
centrilobular
cytomegaly of liver
80: 1 resorptions, D
relative brain wt
None
None
None in surviving
animals
Reference
NTP, 1992
NTP, 1992
Khera, 1973
Chemoff et
al (1979)
Chemoff
et al (1979)
Khera and
Iverson, 19
24
-------
TABLE 7
Prenatal Effects of ETU in Sprague-Dawley Rats following Exposure on Gestation Days 6-15
Observation
No. Pregnant
(term)
Fetal weight (g)
Hydrocephalus
Encephalocele
Cleft Palate
Kyphosis
Limb defects
Digit defects
Edema
Micrognathia
Dose (mg/kg/day)
0
27
4.2 ±0.1
0
0
0
0
0
0
0
0
5
9
4.4 ±0.2
0
0
0
0
0
0
0
0
10
19
3.9±0.1**
0
0
a
0
0
>,
0
0
0
20
31
3.9±0.1**
12/5*
0
0
0
0
0
0
0
30
11
3.8 + 0.1**
38/10***
0
0
0
0
0
0
0.
40
11
A
3.4 ±0.1**
33/10***
13/3*
0
7/2
0
4/1
0
0
80
8
2.6 + 0.1*
13/7***
59/7***
10/4**
30/7***
8/2*
21/7***
20/5***
5/4**
25
-------
Human Health Risk Assessment:
Case Study on Ethylene Oxide
Colloquium #2
June 3-4,1998
Risk Assessment Forum
U.S. Environmental Protection Agency
-------
ETHYLENE OXIDE (EtO) CASE STUDY
EXECUTIVE SUMMARY
Ethylene oxide (EtO) is widely used as an intermediate in chemical synthesis and as a
sterilant. The primary route of human exposure to EtO is by inhalation. EtO is a metabolite of
ethylene which is produced endogenously, and also can be produced from exogenous exposure to
ethylene.
EtO is readily absorbed from the respiratory tract of both humans and laboratory animals into
the blood and is uniformly distributed throughout the body. Clearance is rapid with half-lives
reported to be about 60 min in the human, 12 min in the rat, and 3 min in the mouse. EtO is
metabolized primarily by two pathways. One is conjugation with glutathione; the other is
hydrolysis to ethylene glycol and subsequent metabolism to carbon dioxide. There are species
differences in the metabolism of EtO. Physiological based pharmacokinetic (PBPK) models
indicated that glutathione conjugation accounts for about 10% of EtO in humans, with most of
the remainder undergoing hydrolysis. The percentage of glutathione conjugation is 50% in the
rat, and 75% in the mouse. PBPK models predicted that at equivalent exposure concentrations of
EtO (< 200 ppm), the effective dose, measured as area under the curve (blood EtO vs. time) is the
same for the human, rat, and mouse.
Available data in humans and laboratory animals indicates that EtO induces a variety of acute
and chronic effects including eye and upper respiratory tract irritation, hematoxicity,
neurotoxicity, developmental and reproductive toxicity, genetic toxicity, and cancer. The
mechanisms by which EtO induces a wide range of toxicity are not completely understood. EtO
is an effective alkylating agent of both DNA and proteins. EtO exposure results in genetic
damage and leads to somatic and germ cell mutations, specific DNA adducts, increased
micronuclei formation in mice and humans, and increased sister chromatid exchanges in
peripheral lymphocytes of rats, rabbits, monkeys, and humans. In general, the degree of damage
is correlated with the level and duration of exposure. Therefore, some of the toxic effects are
probably mediated by alkylation of DNA. It is also possible that EtO affects the peripheral and
central nervous system by interference with metabolism of neuronal perikaryon or axonal
transport, thus inhibiting delivery of essential metabolites to nerve terminals.
In humans, toxic effects occurring after acute and short-term exposure to high concentrations
(> 200 ppm) include eye and upper respiratory tract irritation, nausea, vomiting, diarrhea,
headache, dizziness, malaise, and fatigue, muscle weakness, and signs and symptoms of
peripheral neuropathy. Chronic exposure to low 8-hour TWA concentrations (1 ppm) is
associated with the same effects as acute exposure, probably due to daily high level excursions.
Two epidemiologic studies provided suggestive evidence that occupational exposure to EtO is
associated with adverse reproductive outcomes- spontaneous abortions, pre-term births, and post-
-------
term births. Epidemiologic studies conducted to assess the effect of exposure to EtO on mortality
due to malignant neoplasms in chemical factories or sterilizer facilities have produced mixed
results regarding excess cancer risks. The most frequently reported association from EtO
exposure has been lymphatic and hematopoietic cancer, although this appeared to be only
marginally to nonsignificantly elevated. Human studies also showed EtO exposure is associated
with genetic damage including increases in the frequency of sister chromatid exchange or SCE,
and chromosomal aberration in the blood lymphocytes, and in the induction of micronuclei in
bone marrow cells of exposed workers. In general, the degree of damage is correlated with level
and duration of exposure. Increases in genetic damage were not observed when cells were
analyzed 90 days following exposure suggesting that repair had occurred. No genetic damage
was seen after chronic exposure to concentrations less than 0.1 ppm.
Animal data indicate that mice are the most sensitive species to the acute toxic effects of EtO
(LC50 of 600 ppm in mice compared to 4000 ppm in rats). Lethal concentrations of EtO cause
irritation .to the eye and upper and lower respiratory tract, neurological effects manifested by
absence of tail and toe pinch reflex and startle reflex, ataxia, semiconsciousness, and
convulsions. Death was due to respiratory failure, which was likely due to CNS toxicity. Similar
effects were observed in animals surviving acute exposure to EtO. The effects were usually
reversible within a few days after inhalation exposure depending on the concentration of EtO.
Clinical signs observed after repeated exposures are similar to those observed after single
exposure. In addition to irritation and neurological effects, growth retardation, mild anemia, and
pathologic lesions in the adrenal gland, thymus, kidney and spleen also occurred in exposed rats.
and mice to EtO vapors.
Several studies showed that EtO vapors cause developmental and reproductive effects at
concentrations > 50 ppm in rats and > 200 ppm in mice. EtO was carcinogenic in both rats an
mice at a number of sites at concentrations of 50 and 100 ppm. EtO is genotoxic in mammalian
germ cells as evidenced by induction of dominant lethal mutations and heritable translocatipns in
male mice (> 200 ppm). The concentration-response curve for both endpoints are not linear but
are markedly concave upward. EtO is also genotoxic in somatic cells as indicated by induction of
unscheduled DNA synthesis (UDS), gene mutation, SCEs, chromosomal aberrations in human
cells, and gene mutation, micronuclei, chromosomal aberrations, and cell transformation in
rodent cells in vitro.
The mechanisms by which EtO induces a wide range of toxicity are not completely known.
Some of the toxic effects including genetic damage and cancer are probably mediated by
alkylation of DNA, which can alter the structure and functional activities of genes and
chromosomes. It is likely that protein alkylation of protein is involved in testicular toxicity and
developmental toxicity. EtO affects the peripheral and central nervous system possibly by
interference with metabolism of neuronal perikaryon or axonal transport, thus inhibiting delivery
of essential metabolites to nerve terminals.
-------
These characteristics make EtO an interesting chemical to consider in developing models for
quantitative risk estimation. EtO's short half-life obviates the need for complex pharmacokinetic
models and considerations of long-term internal exposure even from acute external exposures,
simply due to prolonged circulation of the uncleared chemical from the target tissue site.
Moreover, the fact that the parent compound (and not a metabolite) is primarily responsible for
the effects observed further reduces the need for consideration of the effects of metabolism on
EtO's toxicity. On the other hand, EtO's apparent mechanism of action, i.e., the formation of
adducts with DNA and proteins, suggests a potential for prolonged retention of the agent at the
cellular target site, once target exposure and interaction has occurred.
EtO has a variety of acute and chronic effects, including contact site irritation, hematoxicity,
reproductive and developmental toxicity, neurotoxicity, and cancer. Several epidemiological
studies have been conducted on the mortality of workers potentially exposed to ETO. In general,
the most frequently reported association from ETO exposure has been lymphatic and
hematopoietic cancer, although this appeared to be only marginally to nonsignificantly elevated.
Given the number of studies and the less-than obvious association of exposure with effect in the
individual studies, would analysis of the database benefit from Monte Carlo or similar
techniques?
Based upon studies in animals, EtO appears to cause genetic damage throughout the life
cycle (e.g. gametes, in utero, adult) which directly effects other health endpoints. The dominant
lethal effects result in a reduction in litter size, due to the death of the developing embryo around
the time of implantation. Carriers of heritable translocations generally show reduced fertility in
the first generation. Moreover, in the event of a balanced reciprocal translocation, there is an
increase in developmental anomalies, including fetal death and malformations, in offspring sired
by the translocation carriers.
With regard to carcinogenesis, EtO was found to cause tumors of the forestomach, mainly
squamous-cell carcinomas, in female rats treated via intragastric intubation. By inhalation, EtO
produced dose-related tumors of the brain, mononuclear-ceil leukemia in rats of both sexes and
peritoneal mesotheliomas in the region of the testis and subcutaneous fibromas in the males. EtO
also induced multiple tumors in mice by inhalation including lung tumors, tumors of the
Harderian gland in both sexes, and uterine adenocarcinomas, mammary carcinomas, and
malignant lymphomas in female mice.
Questions for Discussion
1. Given what is known about mode of action, are there commonalities among toxicities that
would be useful for quantitative analyses? For which toxicity should a separate quantitative
analysis be conducted?
With regard to ethylene oxide, is there any reason to propose different mechanisms of action
for the various endpoints?
-------
2. What additional information would be useful for quantitative analysis of the various toxicities?
For example, is consideration of the entire spectrum of mutational changes, such as the
induction of gene mutations, structural chromosome mutations, and numerical chromosome
alterations (i.e., aneuploidy) important?
3. Are dose and duration of exposure important considerations? If so, for which responses and
how should they be handled?
For ethylene oxide, can dose-equivalency be assumed for all species at various biological
levels of exposure (i.e., administered dose, delivered dose, biologically-effective dose)? How
such cellular functions as repair mechanisms become a part of the overall process of quantitative
analysis?
4. What dose metric would be appropriate for dose response modeling in the observable range?
Does this differ for different toxicities?
5. What quantitative method is recommended for low-level exposures? Does this vary for
different responses? Does mode of action information influence choice of models?
6. If a RfC were to be calculated, does mode of action information influence choice of
uncertainty factors or influence uncertainties about data gaps?
SUMMARY OF KEY STUDIES
HUMAN
Exposure to high concentrations of ethylene oxide gas is irritating to the eyes, while exposure
to aqueous solutions can produce injury to the eye and skin. Reports of respiratory effects (e.g.,
bronchitis) in workers with different exposures are mixed. Central nervous systems effects are
frequently reported, including headache and nausea. Other studies have reported peripheral
neuropathy, impaired hand-eye coordination, and memory loss. Studies of occupationally
exposed women have reported mixed results for increased incidences of spontaneous abortion.
EtO exposure has been associated with lymphatic, hematopoietic, pancreatic, and stomach
cancer.
Acute Exposure Effects
As summarized in IARC (1994), several human studies indicate that acute exposure can have
effects on several different physiological parameters. In a study of five sterilizer operators
exposed to over 700 ppm (1280 mg/m3) for periods up to O.Shrs, Deleixhe et al. (1986) found
moderate to severe "intoxication", headache/vertigo, myasthenia, and indigestion'. These
symptoms were gone by 21 days after the exposure. From day 9-11 after exposure, hemolysis
was noted that lasted until day 16. Bryant et al. (1989) reported that exposures to as little as 10.7
-------
ppm (19.6 mg/m3) may be associated with eye and skin irritation. Deschamps-et al. (1992)
indicated that 20% of workers exposed to greater than 700 ppm (1280 mg/m3) for 4h/day over 4
days developed persistent nonimmunological asthma.
Chronic Exposure Effects / Neurotoxicity
When exposed for periods of from six months to 14 years to concentrations below 10 ppm,
there were on significant effects on immunological, hematological or biochemical parameters
(summarized IARC, 1994). However, exposures over 0.5-20 years to "a few daily short-term
peaks ... 250-700 ppm" (8-hr TWA concentrations of <1 to 4.7 ppm [< 1.83-8.6 mg/m3]) lead to
symptoms of peripheral neuropathy and personality dysfunction and cognitive impairment.
Spontaneous Abortion -
Hemminki et al. (1982, 1983) reported on a case-control study that examined the association
of exposure to specific sterilizing agents (EtO, glutaraldehyde, formaldehyde) with spontaneous
abortion. EtO exposure was associated with an increased rate of spontaneous abortion. The
other two agents showed no such association.
Carcinogenicity
Occupational exposure to EtO has been associated with elevated risk of lymphatic and
hematopoietic cancer. In animal studies, oral exposure to EtO has been associated with tumors
of the forestomach in female rats. A variety of tumors have been observed following inhalation
exposure in rats and mice. The mode of action of EtO-induced carcinogenesis is not completely .
understood. However, it is known that EtO is efficiently absorbed from the respiratory tract into
the blood and is widely distributed throughout the body to all tissues. It is a reactive electrophile
that forms adducts with proteins in humans as measured by hemoglobin adducts in exposed
workers. EtO also forms adducts with proteins, as well as DNA in experimental animals.
Several epidemiological studies have been conducted on the mortality of workers potentially
exposed to EtO. The populations studies fall into two groups, sterilant and chemical workers. In
general, sterilant workers are less likely to have occupational exposure to other chemicals. The
studies are summarized below, as well as in Table 5 (as reviewed by IARC, 1994 and Shore et
al.,1993).
Sterilant Workers
Overall, mortality from lymphatic and hematopoietic cancer was only marginally elevated in
the largest U.S. study on sterilant workers. Three other studies of workers (two in Sweden and
one in the U.K.) involved in sterilization showed nonsignificant excesses of lymphatic and
hematopoietic cancer.
-------
• Steenland et al. (1991) conducted a retrospective mortality study on 18,254 workers (55%
female and 45% male) employed for at least 3 months at 14 U.S. facilities producing
sterilized medical supplies and spices. The average follow up was 16 years. The average 8-
hour TWA concentration of ETO was 4.3 ppm for sterilizer operators and 2.0 ppm for other
exposed workers. Mortality from lymphatic and hematopoietic cancer was only marginally
elevated, but a significant trend was found, especially for lymphatic leukemia and non-
Hodgkin's lymphoma. For exposure of 1 ppm over a working lifetime (45 years), a rate ratio
of 1.2 was estimated for lymphatic and hematopoietic cancer (Stayner et al. 1993).
• • Hogstedt et al. (1979b) reported three cases of leukemia- two cases of myeloid leukemia and
one case of Waldenstrom's macroglobulinemia, that had occurred between 1972 and 1977
(0.2 cases expected) among 240 workers (77 women and 163 men) at a Swedish factory
where hospital equipment was sterilized. ETO air measurements ranged from 2-70 ppm; the
estimated TWA was 20 ppm. A follow up study (Hogstedt 1988) provided an update for
1978-1982 and reported one additional case of leukemia.
• Hagmar et al. (1991) studied 2170 workers (861 men and 1309 women) employed for at least
one year in two Swedish plants where disposable medical equipment sterilized with ETO was
produced. Air concentrations of ETO measured in the early 1960s and 1970s dropped from
about 40 ppm in one facility and from 75 ppm in another to < 0.2 ppm by 1985. These
subjects were followed up to 1986 for mortality, and from 1972 and 1985 for cancer
registration. One case of polycythemia vera and two cases of lymphomas were found (2.0
hematopoietic cancers expected).
• Gardner et al. (1989) followed up 1405 British workers (394 men and 1011 women) in 8
hospitals that started to use ETO sterilizers between 1962 and 1972 through 1987. Air
concentrations of ETO measured in the 1970s were < 5 ppm with peak exposures of about
400 ppm during loading and unloading of sterilizers in the hospitals. There were 32 deaths
from cancers (30 expected). These included 2 deaths from stomach cancer (1.7 expected),
two from non-Hodgkin's lymphoma (0.6 expected), and none from leukemia (0.8 expected).
Studies on chemical workers
There are eight studies of chemical workers exposed to ETO. In a study of chemical workers
exposed to ETO at two plants in the U.S., the mortality rate from lymphatic and hematopoietic
cancer was elevated, but the excess was confined to a small subgroup with only occasional low-
level exposure to ETO. Five studies found excesses of lymphatic and hematopoietic cancer,
which were significant only in two studies. No excesses in cancer mortality rate were found in
two studies. A few studies of chemical workers exposed to ETO show an increased risk for
stomach cancer, which was significant only in one study.
Because of the possibility of confounding occupational exposures, less weight can be given
to the findings from studies of chemical workers. Nevertheless, they are compatible with the
-------
small but consistent excesses of lymphatic and hematopoeitic cancer found in studies of
sterilization personnel.
• A series of studies was carried out on a cohort of 2174 male workers employed between 1940
and 1978 at two chemical plants in West Virginia, USA, which produced and used ETO
(Greenberg et al., 1991; Benson and Teta, 1993; Teta et al. 1993). Production of ETO by the
chlorohydrin method began in 1925 and was phased out in 1957, while production by direct
oxidation began in 1937 and continued until 1971. After 1971, the plants continued to use
ETO brought in from elsewhere. The 8-hr TWA concentration of ETO measured in 1976 was
less than 1 ppm but ranged up to 66 ppm. The cohort was followed to the end of 1988 and .
vital status was ascertained for more than 98% of subjects.
Among the 278 workers in a department which produced ethylene chlorohydrin and
propylene chlorohydrin, but in which there was no ETO production and only occasional ETO
use, there were 8 deaths from leukemia (2.7 expected) and 8 from pancreatic cancer (1.6
expected). In the follow up study (Teta et al. 1993), men who worked in the ethylene
chlorohydrin unit were removed from the study. For the remaining cohort, there were no
excesses of leukemia or pancreatic cancers. The SMR for brain and nervous system cancer,
and stomach cancer were non-significantly increased.
• Hogstedt et al. (1979a) studied 175 male workers at a Swedish chemical plant where ETO .
had been produced by the chlorohydrin process. ETO exposure levels were estimated to
range from 5-27 ppm between 1950-1963 and between 0.5-5 ppm after 1963. The cohort was
followed from 1961 to 1977. The authors reported two leukemias among exposed workers
versus 0.2 expected. A significant excess of stomach cancer was also observed (5 observed
vs. 0.6 expected). In a follow up study (Hogstedt 1988) in which the cohort was extended to
1985, there were three leukemias vs. 0.3 expected, and 9 stomach cancer vs. 1.3 expected.
• Hogstedt et al. (1986) also studied a cohort of 12,8 workers at a Swedish chemical plant
which used the direct oxidation process. Estimated ETO exposure levels dropped from 1-8
ppm during 1963-1976 to 0.4 -2 ppm during 1977-82. One case of chronic myeloid leukemia
(0.2 expected) was observed.
• Gardner et al. (1989) studied 1471 workers (all but one were male) in four British chemical
companies that produced or used ETO. Two companies began producing ETO by the
chlorohydrin process for 10-15 years starting in the 1950s before shifting to the direct
oxidation process. The third company only used ETO by the direct oxidation process during
1960-1981. The fourth company used ETO in the manufacture of derivatives since 1959.
Environmental and personal monitoring since 1977 had shown a TWA of < 5 ppm for almost
all jobs. The vital status of workers was > 98% complete. A small excess of leukemia
mortality was observed (3 observed vs. 1.3 expected) among the chemical workers: An
analysis of leukemia by duration of exposure showed no trend.
-------
Morgan et al. (1981) followed 767 male workers who had potential exposure to ETO from
1955-77 at a production plant in Texas, U.S. ETO was manufactured by the chlorohydrin
process between 1948-64; direct oxidation process was used since 1958. The workers at this
plant had been employed for at least 5 years; 55% of the workers were employed for > 20
years. Measurements of ETO in 1977 in the production area showed air concentrations of
ETO of < 10 ppm. Vital status was ascertained for 95%. Nonsignificant excesses were seen
of cancers of the pancreas (3/0.8) and the brain (2/0.7), and Hodgkin's disease (2/0.35); no
death from leukemia was found (0/0.7).
A follow up of this cohort by Divine (as reported in Shore et al., 1993) was extended to 1985
and achieved a 99.7% follow up rate. There were nonsignificant excesses of brain cancer
(3/1.1) and hematopoietic cancers (3/3.0); all three of the hematopoietic cancers were
Hodgkin's disease.
Bisanti et al. (1993) studied a cohort comprising of 1971 chemical workers licensed to handle
ETO during 1938-84 in the two regions of Northern Italy. Mortality follow up was 99.2%
complete. The mean length of follow up was 9.8 years. No information was available on
worker exposure levels at individual facilities. There were significant excesses of mortality
from lymphosarcoma and reticulosarcoma (4/0.6, p < 0.05) and nonsignificant excesses of •
leukemia (2/1.0) and and stomach (5/4.1).
Thiess et al. (1981) examined the mortality of 602 male workers who had been employed for
at least six months by a German company in the production of ETO and propylene oxide.
ETO was produced by the clorohydrin process in 1928 until 1965 when the direct oxidation
process was introduced. Industrial hygiene measurements during 1978-80 showed that the
average ETO concentration was < 4 ppm. The average follow up of this cohort was 14 years;
the vital status of workers was complete for > 97%. Reported cancer deaths included one case
of myeloid leukemia and one case of lymphatic sarcoma. An analysis by length of exposure
did not reveal any associations.
Kiesselbach at al. (1990) studied 2658 male workers from six German chemical companies
who were exposed to ETO for at least one year between 1928 and 1981 (most had been
exposed after 1950). This study included eligible workers and deaths from the study by
Thiess et al. (1981) and updated the Thiess cohort for an additional 1.5 years. Exposures to
workers ranged from one to 42 years, with a median of 9.6 years. Exposure levels which were
categorized according to types of jobs were characterized only about two thirds of the
workers. Vital status was ascertained for 97.6% of the cohort. The median length was 15.5
years. There were no significant excesses of leukemias nor hematopoietic cancers.
-------
LABORATORY ANIMAL
Laboratory animal studies have demonstrated that EtO is a reproductive toxicant, a
developmental toxicant, a neurotoxicant and a carcinogen, demonstrating a range of effects
similar to that observed in humans (IARC, 1994; Kimmel, C. et al, 1984). Most prominent are
the studies on the mutagenic effects of EtO and the resultant effects on reproduction,
development and carcinogenesis. EtO has been shown to be an effective alkylating agent of both
DNA and proteins, and it has been proposed that this alkylation is a primary mechanism by
which EtO exerts its toxicity (reviewed in Dellarco et al., 1990).
The dominant lethal effects result in a reduction.in litter size, due to the death of the
developing embryo around the time of implantation. Carriers of heritable translocations are
generally viable, and show reduced fertility in the first generation. Moreover, in the event of a
balanced reciprocal translocation, there was an increase in developmental anomalies, including
fetal death and malformations, in offspring sired by the translocation carriers.
With regard to carcinogenesis, EtO was found to cause tumors of the forestomach, mainly
squamous-cell carcinomas, in female rats treated via intragastric intubation. By inhalation, EtO
produced dose-related rumors of the brain, mononuclear-cell leukemia in rats of both sexes and
peritoneal mesotheliomas in the region of the testis and subcutaneous fibromas in the males. EtO
also induced multiple tumors in mice by inhalation including lung tumors, tumors of the
Harderian gland in both sexes, and uterine adenocarcinomas, mammary carcinomas, and
malignant lymphomas in female mice. EtO induced local sarcomas in female mice following
subcutaneous injection. No increases in skin tumors were found in female mice, however, in a
limited dermal study.
Reproductive/Developmental Effects:
Effects on sperm - dominant lethals / heritable translocation:
EtO induces dominant lethal mutations and heritable translocations in mice and rats (Embree
et al., 1977; Generoso et al., 1980). Extending these findings to study the dose-response effects,
Generoso et al. (1990) exposed male (C3Hxl01)F, mice to 165, 204, 250, or 300 ppm EtO by
inhalation for 6 hrs per day, 5 days per week, for 6 weeks; and then daily beginning at week 7 for
2.5 weeks. During the last 10 days and for 1 day after the last exposure, exposed males were
mated to T stock and (SECxC57BL)F, females. Dominant lethal mutations and heritable
translocations were measured. The dose-response curves for both endpoints were not linear, but
were markedly concave upward.
The induction of dominant-lethal mutations is shown in Table 1 and Figure 1 (taken from
Generoso et al., 1990; p!28). Statistically significant increases in dominant lethals were
detected at all but the lowest exposure concentration. Even at the lowest exposure level, there
appears to be a difference from the control values, although the authors note that the range in the
10
-------
exposed groups is within that of the controls. The ICPAEMC1 criteria for a positive dominant
lethal response were used, placing an emphasis on a significant reduction in the average number
of living embryos per female with a corresponding change in either the average number of dead
implants per female or the proportion of females with one or more dead implants (ICPAEMC,
1983). These criteria were first met at the 204 ppm exposure level, where a reduction in the
average number of living embryos and an increase in the number of females with at least one
dead implantation were observed . These effects were magnified in a dose-dependent manner at
250 and 300 ppm, and there was a significant reduction in total implantations at 300 ppm. The
authors suggest that this latter effect was associated with mutations in spermatids and
spermatozoa that were preimplantation lethals. Both stocks of females showed similar responses
except at 300 ppm where the T-stock appeared to respond to mating with exposed males with a
greater percentage of dominant lethals.
The frequency of heritable translocations was significantly increased over controls in all
exposure groups (Table 2, taken from Rhomberg et al., 1990, pp 106). The incidence of
heritable translocations induced by EtO shows a clear dependence on the exposure concentration
(Figure 2, taken from Generoso et al, 1990, p 129). The frequency of translocation carriers in the
exposed groups was defined as all semisteriles plus all steriles after correction for steriles in the
respective control groups. The authors note that this method is justified since previous studies
demonstrated that a high proportion of semisterile males are translocation carriers, arid a random
sampling in the current study showed that all semisterile males were translocation carriers.
I
As in the dominant lethal study, there appeared to be a higher frequency of translocation
carriers when males from the 300 ppm exposure group were mated to T-stock females as
opposed to (SEC X C57BL)F, females. The authors indicate that this is consistent with their
previous results, and suggests that the T-stock oocyte generally has a lower capacity for repairing
lesions induced by chemical mutagens either in its own genome or that from the fertilizing
sperm.
Effects on gestation and developmental endpoints:
EtO is a developmental toxicant (Kimmel, C, et al., 1984). EtO exposure during periods of
development have been associated with increased prenatal death, structural malformations,
reduced litter size and fetal weight, and reduced implantation and increased gestation length.
Three studies carried out in the 1980's were particularly important in initially defining the
developmental toxicity of inhaled EtO. A diagramatic representation of the study designs and
relative exposure periods is shown in Figure 3.
1 International Commission for Protection Against Environmental Mutagens and
Carcinogens.
11
-------
Snellings et al. (1982a) carried out a standard developmental toxicity study in which pregnant
female rats were exposed by inhalation to 0, 10, 33,100 ppm for 6 hrs/day on gestation days 6-
15 (the period of organogenesis). There were no specific endpoints of maternal toxicity reported.
However, Hackett et al. (1982) reported that there were no effects on food consumption or
maternal body weight under a similar exposure regimen of ISOppm EtO on gestation days 7-16.
None of the exposure concentrations had any effect on fetal viability or incidence of
malformations. There was a significant reduction in the fetal weight of both males and females
at the 100 ppm concentration. Thus, a no-observed-adverse-effect level can be established at 33
ppm for developmental toxicity as observed in this study.
Snellings et al. (1982b) also carried out a single generation reproductive study, in which both
male and females rats were exposed to 0,10, 33,100 ppm EtQ for 6 hrs/day, starting 12 weeks
prior to fertilization and continuing through 21 days following parturition.2 No effects were
observed in the parents at any of the exposure concentrations. However, other studies (Hackett et
al..jl982; Snellings et al., 1984) have reported body weight reductions in animals exposed to as
little as 33 ppm over a similar period of time. Thus, the lack of an effect on parental body weight
over the pre- and postmating exposure period must be viewed with some caution. With regard to
the effects on gestation and developmental endpoints, there was a reduction in litter size and an
increase in gestation length at 100 ppm. There was also a significant decrease in postnatal day
21 body weight, although only in males at 33 ppm; not at the 100 ppm level. Thus, it was not
indicative of a dose-response effect, and an apparent no-observed-adverse-effect level would be
33 ppm.
The studies of Generoso and his colleagues (Generoso et al., 1987) have shown that the
period of gestation just after fertilization is particularly sensitive to EtO. Female mice were
exposed by inhalation to 1200ppm for l.Shrs (1800 ppm-hrs) at 1, 6, 9 and 25 hrs after mating.
At 1 hr and 6 hrs after mating, there was increased pre- and post-implantation loss, hydropia, and
morphological abnormalities when the embryos were examined on gestation day 17, just prior to
parturition. The malformations included defects of the eye, palate, heart, abdominal wall,
extremities, and tail. The incidence of these effects was reduced when exposure began 9 hrs after
mating, returning to control levels by 25 hrs after mating (2-cell stage). Effects on fetal body
weight were not reported. This study by itself was inadequate to provide information on the
dose-response nature of EtO toxicity. However, it supported the findings of developmental
toxicity in the Snellings studies, and as importantly, demonstrated that brief exposures during the
one-cell stage of embryonic development could result in structural defects, as well as death.
2 Prior to mating, both males and females were exposed 5 days/week. Following mating,
females continued exposure, but changed to a schedule of 7days/week. Exposure was stopped
from gestation day 20 until five days after parturition. At postnatal day 5, exposure of the dams,
separate from their litters, was continued for 6 hrs/day, 7 days/week until postnatal day 21:
12
-------
In a follow-up to this study, an EPA-sponsored study (Kimmel, G., et al., unpublished) to
further examine the dose-response nature of EtO developmental toxicity was carried out, using a
study design similar to that of Generoso et al. (1987). Female (C3H/R1 X C57BL)F, mice were
exposed by inhalation to EtO at 600, 900, and 1200 ppm for 1.5 hours (i.e., 900, 1350, and 1800
ppm-hrs, total exposure), 6 hours after the end of the 30-min mating period. As noted above,
Generoso et al. (1987) demonstrated that of the four exposure times examined over the first day
of gestation, the 6-hour post-mating exposure time resulted in the greatest effect as measured by
post-implantation loss. Uterine analysis was carried out on gestation day 17. The fetuses were
then fixed in buffered formalin for skeletal analysis.
The dose-response data are summarized Table 3 for effects on intrauterine viability, growth,
and development. There were no significant effects of exposure on implantation at any of the
concentrations tested. There was^n exposure-related effect on viability. A statistically with
increase in the percent of post-implantation loss and a concomitant decrease in the number of
live fetuses per female were observed at the 900 ppm and 1200 ppm exposure levels. There was
a significant trend for the higher exposure concentrations to be associated with an increase in
fetal weight. This was likely associated with the decrease in litter size and consequent larger
surviving offspring due to reduced competition for maternal support. "Affected" implants were
also calculated to determine whether a different dose-response pattern would result from using
this more inclusive measure of developmental toxicity. There was a statistically significant dose-
response trend beginning at 900 ppm, identical to the effect levels for the individual endpoints of
post-implantation loss and growth.
Table 4 summarizes the dose-response effects of the ethylene oxide exposure on external
morphology observed in the viable fetuses. Dose-related increases in hydropia and in structural
alterations of the limb or tail were observed. The changes were statistically significant for both
of these end points at 1200 ppm and for hydropia at 900 ppm. Limb/tail defects approached
statistical significance (p = .06) at 900 ppm and those of the eye approached statistical
significance at 1200 ppm (p = .06). There was no dose-related effect on abdominal abnormalities
and the incidence of exencephaly did not increase over control for any exposure level. The
predominance of hydropia. limb/tail and eye defects is similar to the findings of Rutledge et al.
('89) at the 6-hr postmating treatment period. The incidence of abdominal effects, however, was
not as great in the current study. Evaluation of skeletal development was carried out as an
extension of this study and the preliminary findings have been reported by Polifka et al. ('91).
The complete analysis of the effects of ethylene oxide on skeletal development is detailed in the
accompanying paper (Polifika et al., '93). In general, ethylene oxide exposure resulted in
alterations in ossification, with a notable increase in the incidence of stemebral effects, including
cleft sternum.
13
-------
Carcinogenicity:
EtO was tested for carcinogenicity in one experiment by oral administration in rats, in two
experiments by inhalation in rats, and two experiments by inhalation in mice. It was also tested
in single studies in mice by skin application and by subcutaneous injection.
Oral Studies
• Groups of 50 female Sprague-Dawley rats (100 days old) were administered ETO (99.7%
purity in commercial vegetable oil) by gastric intubation at 7.5 or 30 mg/kg body weight,
twice weekly for 107 weeks. Control animals were either untreated or treated with vegetable
oil alone (50 per group). The survival rate of the high dose group was lower than that of the
control groups. There were dose-related increases in the incidence of forestomach tumors
(0/0, 0/0,21/50,46/50 for the two control groups, low, and high dose animals, respectively).
The forestomach tumors identified in the low and high dose animals, respectively, as follows:
squamous cell carcinomas' (8/50 and 29/50), fibrosarcomas (0/50 and 2/50), carcinomas in
situ (4/50 and 4/50), papillomas, hyperplasia or hyperkeratosis (8/50 and 9/50). There was no
increased tumor incidence at other sites in treated animals over that in controls (Dunkelberg,
1982).
Inhalation Studies
• In a study as reported by Snellings e\al. (1984) and Garman et al. (1985), groups of 120 male
and female Fischer 344 rats (8 weeks of age) were exposed by inhalation to ETO (purity >
99.9%) vapor at 10, 33, 100 ppm for 6 hours per day, five days per week for two years (Table
6). Two control groups, each of 120 animals per sex were exposed in inhalation chambers to
room air. During month 15 of exposure, mortality increased in both treated and control
groups due to a viral sialodacryoadenitis.
The incidences of brain tumors, classified as "gliomas, malignant reticulosis and granular-
cell tumors" were significantly increased in treated animals of each sex at 18 and 24-25
months of exposure (males: 1/181 control, 0/92 low dose, 3/86 mid dose, 6/87 high dose;
females: 0/187 control, 1/94 low dose, 2/90 mid dose, 2/78 high dose). Statistically
significant increases in mononuclear-cell leukemia (MCL) were also found in treated animals
of both sexes at 24 months (males: 13/97 control, 9/51 low dose, 12/39 mid dose, 9/30 high
dose; females: 11/116 control, 11/54 low dose, 14/48 mid dose, 15/26 high dose).
In male rats, there were also significant increases in peritoneal mesotheliomas which were
originated in the testicular serosa (2/97 control, 2/51 low dose, 4/39 mid dose, 4/30 high
dose). High dose males also had increases incidence in subcutaneous fibromas (3/97 control,
9/51 low dose, 1/39 mid dose, 11/30 high dose).
14
-------
Similar tumor findings were observed in the study by Lynch et al. (1984) in which groups of
80 male weanling Fischer rats were exposed by inhalation to ETO (purity of 99.7%) vapor at
0 (filtered air), 50, or 100 ppm for 7 hours per day, five days per week for two years (Table
7). The mortality rate was increased in the two treated groups over that in controls. Gliomas
of the brain were significantly increased in the high dose animals (0/76 control, 2/77 low
dose, and 5/79 high dose). Peritoneal mesotheliomas in the region of the testis developed in
3/78 control, 9/79 low dose, and 21/79 high dose; the increase was significant for the high
dose group. Mononuclear-cell leukemia (MCL) was observed in 24/77 control, 38/79 low
dose, and 30/76 high dose; the incidence of MCL was significant in the low dose group but
the increase could not be ascertained in the high dose groups owing to excessive mortality.
In a screening bioassay conducted by Adkins et al. (1986), groups of 30 female A/J mice (8-
10 weeks old) were exposed by inhalation to ETO (99.7% pure) at 0, 70 or 700 ppm for 6
hours per day for 5 days per week for 6 months. A positive control group consisted of 20
animals received a single intraperitoneal injection of urethane (Ig/kg body weight). Survival
rates at the end of the 6-month period were: 30/30 (untreated control), 28/30 (low dose),
29/30 (high dose), and 19/20 (positive control). The numbers of animals with pulmonary
adenomas among survivors were: 8/30 (untreated control), 16/28 (low dose), 25/29 (high
dose), 19/19 (positive control). Similar results were observed in a second experiment, in
which the low dose group was omitted. The number of animals with pulmonary tumors
among survivors were: 8/29 (untreated control), 12/28 (200 ppm ETO), 19/19 (urethane).
In a study conducted by the National Toxicology Program (NTP, 1987), groups of 50 male
and 50 female B6C3F1 mice (8 weeks of age) were exposed by inhalation to 0, 50, or 100
ppm ETO (>99% pure) for 6 hours per day, five days per week for up to 102 weeks (Table 8).
Survival rates at the end of the study were: 28/50 (control males), 31/50 (low dose males),
34/50 (high dose-males); 25/50 (control females), 24/50 (low dose females), 31/50 (high dose
females). Mean body weights of treated males and females were similar to those of controls.
There were statistically significant increases in the incidences of alveolar/bronchiolar
carcinomas and combined lung tumor incidences (carcinomas and adenomas) in treated
animals of each sex. The respective tumor incidences were: 6/50 and 11/50 control males,
10/50 and 19/50 low dose males, 16/50 and 26/50 high dose males; and 0/49 and 2/49 control
females, 1/48 and 5/48 low dose females, 7/49 and 22/49 high dose females.
Statistically significant increases in the incidences of papillary cystadenoma of the Harderian
gland were also observed in treated animals of each sex (males: 1/43 control, 9/44 low dose,
8/42 high dose; females: 1/46 control, 6/46 low dose, 8/47 high dose). In addition, one
papillary cystadenocarcinoma of the Harderian gland was found in a high dose male mouse
and in one low dose female mouse. In female mice, there were also statistically significant
increases in the incidences of malignant lymphomas, uterine adenocarcinomas, and
15
-------
mammary gland carcinomas (lymphomas: 9/49 control, 6/48 low dose, 22/49 high dose; uterine
cancer: 0/49 control, 1/47 low dose, 5/49 high dose; mammary tumors: 1/49 control, 8/48 low
dose, 6/49 high dose).
SUMMARY OF
RELEVANT MECHANISTIC / MODE OF ACTION INFORMATION
The reaction of EtO with nucleophilic molecules raises concern over its potential toxicity
(reviewed by Dellarco et al., 1990). EtO reacts directly with a variety of cellular
macromolecules, including DNA, and has been shown to alkylate protein and DNA at exposure
levels encountered occupationally. EtO is an effective mutagen in a variety of organisms ranging
from bacteria to mammalian cells. There is also a positive correlation between EtO exposure and
human somatic cell cytogenetic damage. EtO is not only effective at producing somatic cell
mutation, but also at inducing genetic damage in germ cells.
Alkylation products (adducts) of the reaction of ethylene oxide with blood proteins, including
hemoglobin, can be readily followed in humans and animals providing an internal measure of
exposure from both endogenous production and exogenous sources (reviewed in I ARC, 1994).
Formation of DNA adducts in rat tissues is linear over the range 1-30 ppm for 6 hours. The
reactive parent is removed by reaction with cellular nucleophiles, metabolism, or exhalation.
The metabolic pathways reduce the chemical's reactivity by hydrolysis or by conjugation with
glutathione. Inhalation exposures lead to dose dependent depletion of glutathione at sufficiently
high concentrations (e.g. 20% depletion at 100 ppm for 4 hr and 60 - 70% depletion at 600 ppm
for 4 hr). Urinary metabolites are derived from the oxidative and glutathione conjugation
processes. Rats conjugate ethylene oxide to a greater extent than mice; rabbits appear to be
incapable of this reaction. Elimination may be slower in cases of high exposure where
glutathione is depleted. The fact that the studies with gastric intubation and dermal application
did not result in tumors away from the site of exposure indicate that EtO cannot reach sufficient
internal concentrations by these routes to initiate carcinogenesis, either because it reacts with the
DNA or proteins in the tissue of those sites or because there is local metabolism that minimizes
EtO's toxicity.
EtO exposure results in genetic damage and leads to somatic and germ cell mutations,
specific DNA adducts, increased micronuclei formation in mice and humans, and increased sister
chromatid exchanges in peripheral lymphocytes of rats, rabbits, monkeys, and humans. In
general, the degree of damage is correlated with the level and duration of exposure. Transmitted
germ-cell chromosomal effects appear to be limited to postmeiotic cells; EtO does not appear to
produce effects at the level of the gene in stem cells (noted in Generoso et al., 1990). Greater
numbers of hemoglobin and DNA adducts occur per unit of exposure in rats and mice at high
concentrations (>33 ppm) than at lower concentrations (noted in IARC, 1994, pi38).
16
-------
Information on the impact of repair mechanisms is limited, although this component of the
mode of action will ultimately have to be considered in modeling toxicity. In the dominant lethal
study reported by Generoso et al. (1990), there appeared to be a higher frequency of translocation
carriers when males from the 300 ppm exposure group were mated to T-stock females as
opposed to (SEC X C57BL)F, females. The authors indicate that this is consistent with their
previous results, and suggests that the T-stock oocyte generally has a lower capacity for repairing
lesions induced by chemical mutagens either in its own genome or that from the fertilizing
sperm.
REFERENCES
Bryant, H; Visser, N; and Yoshida, K. (1989). J. Soc. Occup. Med. 39: 101-106.
Dellarco, V; Generoso, W; Sega, G; Fowle, J; Jacobson-Kram, D. (1990). Environ. Molec.
Mutatgenesis, 16: 85-103.
Deleixhe, A; Balsat, A; and Laurent, C. (1986). Arch. B. Med. Soc. Hyg. Med. Tr. Med. Leg.
44:478-488 (French). ' •
Deschamps, D; Rosenberg, N; soler, P; Maillard, G; Foumier, E; Salson, D; and Gervais, P.
(1992). Br. J. Ind. Med! 49: 523-525.
Embree, J; Lyon, J; Mine, C (1977). Toxicol Appl Pharmacol 40: 261-267.
Generoso, W; Cain, K; Krishna, M; Sheu, C; Gryder, R (1980). Mutat Res 73: 133-142.
Generoso, W; Rutledge, J; Cain, K; Hughes, L; and Braden, P (1987). Mutation Research,
176:269-274,
Generoso, W; Cain, K; Cornett, C; Cacheiro, N; and Hughes, L (1990). Environ. Molec.
Mutatgenesis. 16: 126-131.
Hackettetal. (1982)
Hemminki, K; Mutanen, P; Saloniemi, I; Niemi, M; and Vainio, H. (1982). Br. Med. J. 285:
1461-1463.
Hemminki, K; Mutanen, P; and Niemi, M; (1983)..Br. Med. J. 286: 1976-1977.
17
-------
ICPAEMC (1983). International Commission for Protection Against Environmental Mutagens
and Carcinogens Committee 1 Final Report, Mutat Res 114:117-177.
IARC, International Agency for Research on Cancer (1985). Allyl Compounds, Aldehydes,
Epoxides and Peroxides. Volume 36.
IARC, International-Agency for Research on Cancer (1994). Some Industrial Chemicals.
Volume 60.
Kimmel, C; Laborde, J; and Hardin, B (1984). In: Toxicology and the Newborn. Kacew and
Reasor, eds.; Elsevier Science, Amsterdam.
Rhomberg, L; Dellarco, V; Siegel-Scott, C; Dearfield, K; and Jacobson-Kram, D (1990).
Environ. Molec. Mutatgenesis, 16: 104-125.
Snellings, W; Maronpot, R; Zelenak, J; and Laffoon, C (1982a). Toxicol Appl Pharmacol 64:
476-481.
Snellings, W; Zelenak, J; and Weil, C. (1982b). Toxicol Appl Pharmacol 63: 382-388.
18
-------
TABLE 1. EtO Induction of Dominant-Lethal Mutations
I-K)
concentration
(ppin)"
165
Control
204
Control
250
Control
300
Control
' Slock of
females'1
T-stoek
(SECx 101 )F,
T-stock
(SI-Cx 101)!',
T-stock
(SliCx 10I)F,
T-stock
(SF.Cx 101 )F,
T-stock
(SI-Cx 101)1',
T-stock
(SF.Cx 101)1-,
T-stock
(SI-Cx I01)F,
T-stock
(SI-Cx 101)F,
Mated
females
40
31
45
31
44 .
38
39
40
40
29
39
29
36
31
39
34
Pregnant
females
35
21
38
25
39
31
38
36
35
28
37
24
27
21
34
27
Implants
,pcr
pregnant
female
8.3
8.2
8.1
8.6
8.3
8.4
8.4
9.1
7.9
8.0
8.4
8.4
5.2**
7.2
8.3
8.0
Living
embryos
per
pregnant
female0
6.3
7.6
6.7
8.3
6.1
7.5* '
7.1
8.6
5.4**
6.0**
7.0
7.9
2.7**
4.2**
6.7
7.7
Dead
implants
(%)
24
7
18
3
27
12
17
6
32
25
16
6
48
42
19
4
No. of .
females with
one or more
dead implants'1
27
11* ,
27
6
36**
21**
25
14
• 32**
24**
24
9
25**
18**
9
6
Dominant
Icthals
(%)'
6
8
_
14
13
_
23
24
_
60
45
" Male mice were exposed by inhalation 6 hours a day on weekdays for 6 weeks and then daily beginning the seventh week for 2.5 more weeks.
b Males were mated to T-stock females during 6-10 days prior to ending the daily exposure and to (SEC x 101)F, females during the remainder of exposure period and for one day
afterwards.
c Comparisons between treatment and control groups are by one-sided Mann-Whitney nonparametric analysis.
d Comparisons between treatment and control groups arc by one-sided chi-squarc test in a 2 x 2 contingency table.
°.Percent dominant Icthals = 11-living embryos per pregnant female (cxpcrimental)/Iiving embryos per pregnant female (control)] x 100.
*/'<0.05. • . . .
**/'
-------
5O-
3O-
2O-
1O-
15O 2OO
EtO CONCN. (ppm)
25O
3OO
Fig. 1. concentration-response curves for induction of dominant lethal
mutations in male germ cells. A. matings with T-stock females; O. matings
with (SEC x IOI)F, females.
20
-------
TABLE 2. Mouse Heritable Translocation Test Results for EtO*
Exposure
(ppm)
0
165
204
250
300
Combined
0
165
204
250
300
T-stock female
Translocation
- incidence
1/1,451
14/610
28/399
41/354
33/100
1/2,068
.32/1,143'
52/1,021
88/812
109/427
(%)
0.07
2.29
7.02
11.58
33.00
0.05
2.79
5.09 .
10.84
25.53
(SEC x C57BL)
F, female
Translocation
incidence (%)
0/617 0
18/533 3.38
24/622 3.86
47/458 10.26
76/327 23.24
>
*(C3H x 101) F, male mice were placed in a semidynamic inhalation exposure chamber made of
glass and exposed to 165, 204, 250, or 300 ppm EtO for 6 hours per day, 5 days per week for 6
weeks and then daily beginning at week 7 for 2.5 weeks. During the last 10 days of exposure
and 1 day after the last exposure, treated males were mated to T stock and (SEC x C57BL)F,
females. The fertility test method of identifying translocation carriers was employed (Generoso
etal., 1990).
21
-------
4O-|
30-
o
en
-------
Figure 3. Diagram of EtO Study Designs
SNELLINGS ET AL. (1982): DEVELOPMENTAL TOXICITY STUDY - RATS
0,10, 33,100 PPM, 6 HRS/DAY
6 15 20
FEMALES
MALES
SNELLINGS ET AL. (1982): REPRODUCTIVE STUDY - RATS
0, 10, 33,100 PPM, 6 HRS/DAY, 5 DAYS/ WEEK
FEMALES
MALES
GENEROSO ET AL. (1987): BRIEF EXPOSURE STUDY - MICE
1200 PPM, 1.5 MRS ON MATING DAY
FEMALES
MALES
17
I
M
23
-------
Table 3. Summary of Dose-Response Effects of EtO (Kimmel, G. et al., unpublished)
CONTROL
600 ppm
900 ppm
1200ppm
n
140
58
57
44
Impl/dam
11.3
11.5
12.4
10.7
Live/dam
10.6
10.7
8.9*
42*
Postimp
loss1
6.0
7.4
28.1*
61.0*
Fetal
Wt2
1.10
1.12
1.30*
1.25*
Affected1
6.8
8.6
32.6*
66.1*
'percent per litter
2grams
24
-------
Table 4. DOSE-RESPONSE EFFECT OF ETHYLENE OXIDE
WITH RESPECT TO ALTERATION OF SPECIFIC DEVELOPMENTAL ENDPOINTS
Control
6002
900
1200
Litters
n
140
58
57
44
Hydropia
x (Range) .
0.0(0)
0.5(0-14)
3.5 (0-50)*
8.7 (0-67)*
Limb or Tail
x (Range)
0.2 (0-8)
0.1 (0-8)
' 3.4 (0-33)
10.8(0-67)*
Eye1
x (Range)
0.3(0-11)
0.2 (0-9)
2.9 (0-50)
5.4 (0-66)
Abdominal
x (Range)
0.1 (0-8)
0.3(0-10)
0.3 (0-9)
2.0 (0-33)
Exenceph.
x (Range)
0.4(0-13)
0.5(0-10)
0.4 (0-20)
0.0 (0)
All reported as mean percent per litter & range of percent per litter
'Includes open or missing eye
2parts per million
*dose level is included in statistically significant trend
25
-------
Table 5. Summary of cpidcmiological findings on ethylene oxide
Reference
(country)
Hogstedt et at
(I986);llogstcdt
(1988) (Sweden)
Hogstedt et al.
(I979a, 1986);
Hogstedt (1988)
(Sweden)
Hogstedt et al.
(1986) (Sweden)
Morgan et al.
(1981)(USA)
Type of plant; study
period; number of
subjects; minimal period
employed; follow-up
Production of sterilized
supplies; 1978-82; 203
subjects; 1 year; 100%
Ethylene oxide production
plant (one facility); 1961-
85; 241 subjects, of which
89 "full-time operators"; 1
year; 100%
Lthylene oxide production
(one plant); 1964-81; 355
subjects; 1 year; 100%
Production of ethylene
oxide; 1955-77; 767 men;
5 years; around 95%
No. of .
deaths
5
34
8
46
No. of
cancers
4
2
14
5
2
1"
11
2
0
RR
(2.5)
(15)
(2.3)
(8.3)
(10)
.
0.72
5.7
0
95% CI
(0.68-6.4)
(1.9-56)
(1.3-4.8)
(2.9-21)
(1.2-36)
.
0.36-1.3
0.64-21
0-5.2
Site
All neoplasms
L&H
All neoplasms
Stomach
Leukaemia
,
Leukaemia
All neoplasms
Hodgkin's
disease
Leukaemia
Comments
Estimated average past exposure in
storage room was 20 ppm; one leukaemia
was part of a cluster which had originally
prompted the study.
Estimated average exposure before 1963,
5-25 ppm; mortality rates shown only for
"full-time operators" (high-exposure
group); no overall excess tumour mortality
among workers with intermittent exposure
or those unexposed; excess mortality from
stomach cancer (4 deaths, SMR, 6.67) and
from leukaemia (1 death; 0.2 expected)
among workers with intermittent exposure
The one case of leukaemia (0.16 expected)
was in a maintenance worker with
multiple exposures; average exposure in
1963-1976, 1-8 ppm; after 1977, 0.4-2
ppm
High percentage of deaths of unknown
cause (9%); limited information on
manufacturing processes and exposure
concentrations; exposures probably below
10 ppm with occasional peaks to 6000
ppm; nonsignificant excess risks from
cancer of the pancreas and cancers of the
central nervous system
26
-------
Reference
(country)
Divine (unpub-
lished); reported
by Shore et al.
(1993) (USA)
Kiesselbach et
al. (1990)
(Germany)
Gardner et al.
(1989) (UK)
Benson & Teta
(1993) (USA)
Type of plant; study
period; number of
subjects; minimal period
employed; follow-up
Updating of Morgan et cil.
(1981); 1955-85; 99.7%
Chemical plants (8
facilities); 1928-82; 2658
men; 1 year; 97.6%
Production or use of
ethylene oxide (4
facilities); 1956-87; 1471
subjects; no minimal
employment; around 98%
Hospital sterilization units
(8 hospitals); 1964-87;
1405 subjects; no minimal
exposure; around 98%
Work in a chlorohydrin
unit and potential exposure
to ethylene oxide (2
facilities); 1940-88; 278
men; no minimal
employment; 98%
No. of
deaths
Not
applica-
ble
268
157
69
147
No. of
cancers
3
0
68
14
5
2
53
3
3
2
32
2
0
2
40
1
8
4
RR
(1.0)
(0)
0.97
1.4
1.0
0.85
1.1
0.7
2.3
(1.9)
I.I
1.2
0
(3.5)
1.3
(0.7)
2.9
(3.5)
95% CI
(0.21-2.9)
(0.0-3.4)
0.76-1.2
0.75-2.3
0.32-2.3
0.10-3.1
(0.85-1.5)
(0.15-2.1)
(0.47-6.6)
(0.23-7.0)
0.73-1.5
0.15-4.3
0-4.9
(0.42-13)
0.93-1.8
0.02-3.9
1.3-5.8
0.96-8.9
Site
Hodgkin's
disease
Leukaemia
All neoplasms
Stomach
L&H
Leukaemias
All neoplasms
Stomach
Leukaemia
Non-Hodgkin's
lymphoma
All neoplasms
Stomach
Leukaemia
Non-Hodgkin's
lymphoma
All neoplasms
Stomach
L&H
Leukaemia
Comments
No information on exposure
concentrations or on nature of production
processes; most of study population of
Thiess et al. included.
Average exposure after 1997 was to less
than 5 ppm (<1 ppm in many jobs), with
occasional peak exposures of several
hundred ppm; highest mortality from
leukaemia among subjects with definite
exposure to ethylene oxide; risk increased
with latency of exposure; non-significant
excess risks for cancers of the oesophagus,
lung and bladder.
Updating of study by Greenberg et al.
(1990), including only workers ever
employed in the chlorohydrin department;
excess of pancreatic cancer (8 deaths,
SMR, 4.9; 95% CI, 1.6-11).
27
-------
Reference
(country)
Teta el at.
(1993)(USA)
Steenland ct al.
(I991);Stayncr
c/o/. (1993)
(USA)
Type of plant; study
period; number of
subjects; minimal period
employed; follow-up
Production or use of
cthylene oxide (2
facilities); 1940-1988;
1896 men; no minimal
employment; 99%
Production of sterilized
medical supplies and
spices (14 facilities);
1943-87; 18 254 subjects;
3 months; 95. 5%
No. of
deaths
431
1117
No. of
cancers
110
8
7
5
343
11
36
13
(16)
RR
0.86
1.6
0.59
I.I
0.90
0.95
L06
0.97
(1.3)
95% CI
0.71-1.0
0.69-3.2
0.24-1.2
0.35-2.5
0.81-1.0
0.45-1.7
0.75-1.5
0.52-1.7
(0.76-2.2)
Site
All neoplasms
Stomach
L&II
Leukaemia
All neoplasms
Stomach
L&H
Leukaemia
(Non-
Hodgkin's
lymphoma;
ICD9 200, 202)
Comments
Average exposure in producing
departments <1 ppm, but occasionally up
to 66 ppm 8-h TWA. Updating of study by
Greenberg et al. (1990), excluding
workers ever employed in the
chlorohydrin department; in an internal
comparison with workers in the same
complex, a two- to three-fold increase in
leukaemia risk was observed for workers
exposed for more than 10 years to
ethylene oxide.
Recent average exposure of sterilizer
operators was 4.3 ppm, that of other
workers was 2.0 ppm; no significant trend
in mortality from L&H with duration of
exposure; mortality from L&H increased
with latency (SMR at ^2*0 years since first
exposure, 1.8 [95% Cl, 0.94-3.0]); test for
linear trend, p = 0.03; increased risk for
L&H with cumulative exposure (for
results by cumulative exposure, see Table
9); mortality from kidney cancer was also
elevated (SMR, 1.8, 13 deaths) and
increased with latency.
28
-------
Reference
(country)
Mngnar et al.
(1991) (Sweden)
Bisanti ct al.
(1993) (Italy)
Type of plant; study
period; number of
subjects; minimal period
employed; follow-up
Production of disposable
medical equipment (2
facilities); 1964-86; 2 170
subjects; 1 year; 98.2%
Workers licensed (o handle
cthyleiie oxide; 1940-84;
1971 men; 1 year with
licence; 99.2%
No. of
deaths
15
76
No. of
cancers
21"
3
0
43
6
2
4
5
RR
0.78
1.5
0
1.3
2.5
1.9
6.8
1.2
95% CI
0.49-1.2
0.32-4.5
0-7.4
0.98-1.8
0.91-5.5
0.23-7.0
1.9-17
0.40-2.9
"
Site
All neoplasms
L&H
Stomach
All neoplasms
L&H
Leukaemias
Lympho- and
reticulosarcoma
Stomach
Comments
Average estimated exposure of sterilizers,
around 40 ppm in 1970-72, less than 1
ppm in 1985; packers, around 35-50 ppm
in 1970-72, less than 0.2 after 1985; no
trend in risk with increasing cumulative
exposure but only 0.2 expected cases of
L&H in "high" exposure group (> 1 ppm-
year).
Increased mortality from all types of
cancer; no increase in risk for L&H with
latency or duration of exposure; risk for
L&H highest among workers licenced
only for ethylene oxide (5 deaths; SMR,
7.0; 95% CI, 2.3-16); no information on
exposure levels.
RR, risk estimate: standardized mortality ratio, SMR, unless otherwise specified; CI, confidence interval; L&H, neoplasms of the lymphatic and haematopoietic
tissues
"Cancer cases,.standardized incidence ratio
29
-------
Table 6. Inhalation Carcinogenicity Studies of EtO in F344 Rats (Snellings et al., 1984;
Carman et al. 1985)
Concentrations (ppm)
0
10
33
100
Brain Tumors
(gliomas, malignant
reitulosis, granular-
cell tumors)
Male Female
1/181 0/187
0/92 1/94
3/86 2/90
6/87* 2/78*
Mononuclear-Cell
Leukemia
Male Female
13/97 11/116
9/51 11/54
12/39 14/48
9/30* 15/26*
Peritoneal
Mesothelioma
Male
2/97
2/51
4/39
4/30*
30
-------
Table 7. Inhalation Carcinogenicity Studies of EtO in Male F344 Rats (Lynch et al. 1984)
Concentrations
(ppm)
0
50
100
Brain Gliomas
0/76 •
2/77
5/79*
Mononuclear-cell
Leukemia
24/77
38/79
30/76*
Peritoneal
Mesothelioma
3/78
9/79
21/79
Table 8. Inhalation Carcinogencity Sudies in B6C3F1 Mice (NTP, 1987)
Concentrations
(ppm)
0
50
100
Lung tumors
(alveolar/bronchio-
lar-adenomas &
carcinomas)
Male
11/50
19/50
26/50*
Female
2/49
5/48*
22/49*
Harderian Glands
Tumors (papillary
cystadenoma)
Male
1/43
9/44*
8/42*
Female
1/46
6/46*
8/47*
Malignant
Lymphoma
Female
9/49
6/48
22/49*
Mammary
Gland
Carcin-
omas
Female
1/49
8/48*
6/49*
-------
FRAMEWORK FOR HUMAN HEALTH RISK ASSESSMENT
Colloquium #2
Case Study: Trichloroethylene (TCE)
Executive Summary
TCE is one of the halpgenated ethylenes, used in many diverse manufacturing industries
as solvents, carriers, or extractants; in dry cleaning textiles; in metal cleaning and degreasing; in
textile manufacture; as insulating fluids/coolants; and as chemical intermediates. Human
exposure to TCE generally occurs via inhalation, ingestion, and dermal contact.
TCE is rapidly absorbed from the gastrointestinal tract and through the lungs. Absorption
of the vapor through the skin is negligible. Once absorbed, TCE is rapidly distributed throughout
the body, preferentially to adipose tissues. TCE is metabolized primarily in the liver but also in
the kidney. The major pathway is oxidative metabolism leading to the formation of chloroacatic
acids. A minor pathway in rodents and humans involves the formation of mercapturic acids via
the GST pathway. Metabolism plays an important role in the toxicity of TCE because many of its
metabolites are themselves toxic. Many differences among species in their responses to TCE
exposure may be attributed to differences in the rates at which they metabolize the parent
compound.
Based on effects reported in humans and laboratory animals, the primary targets for TCE
toxicity appear to be the nervous system, liver, and kidneys. Inhalation of TCE can produce toxic
effects in mouse lungs, but the specific targeting of the lungs in exposed humans does not seem
to be a major effect. Available studies show no consistent effect of TCE on the human
reproductive system. There is little evidence of toxic effects in developing rats and mice. Data
regarding the genotoxicity of TCE suggest that it is a very weak, indirect mutagen.
TCE causes neurological effects in humans after acute exposure to high levels in the
workplace and in controlled studies in human volunteers. Neurological effects from TCE
exposure included dizziness, drowsiness, impaired motor coordination, visual perception, and
cognition. Neurological effects (e.g. increasing rearing activity, transient ataxia) were observed
in rats after acute inhalation or oral exposures to high doses of TCE. Neurological symptoms
including ataxia, lethargy, convulsions, and hind-limb paralysis were also reported in rats
following chronic oral exposure to TCE. The mechanisms of TCE-induced neurological effects
are not known but are likely to be mediated by its action on disruption of cellular phospholipid
membrane of neurons.
There is some evidence for TCE-induced hepatic effects (liver failure and necrosis) in
humans following accidental or intentional exposure to relatively high levels. Impairment of liver
functions and enlarged livers have been reported in occupationally exposed workers. Exposure
1
-------
concentrations in these studies were not reported. Liver enlargement is the primary hepatic effect
seen in TCE-exposed rats and mice after oral and inhalation exposure. Histological alterations
associated with liver enlargement included cellular hypertrophy and necrosis which occurred at
higher doses.
There is also some human evidence suggesting that TCE exposure is associated with
elevated risks of cancer. Several occupational, community cohort and case-control studies have
been conducted to evaluate toxicity from TCE exposure. Conclusions drawn from many of these
studies are somewhat limited due to the presence of confounding factors, such as smoking and
altered health state. In addition, many of the subjects included in these studies were exposed to
other compounds in addition to TCE. Results from the three most informative studies
consistently indicate an excess relative risk for cancer of the liver and biliary tract, and non-
Hodgkin's lymphoma. Studies of structural chromosomal aberrations, aneuploidy, and sister
chromatid exchange in peripheral lymphocytes of workers exposed to TCE were inconclusive.
TCE has been shown to induce liver tumors in mice following chronic oral or inhalation
exposure. Liver tumors were not induced in rats under similar exposure conditions. The liver
toxicity and tumors induced in mice exposed to TCE appear to be related to the induction of
peroxisome proliferation by its metabolite, trichloroacetic acid (TCA). TCA has been shown to
induce hepatic peroxisome proliferation in rodents and induce liver tumors in mice. Differences
among species in response to TCE exposure appear to reflect differences in their metabolic
pathways and production of TCA. Mice metabolize TCE more efficiently than rats or humans.
Mechanisms by which peroxisome proliferation may induce cancer are unclear, although it has
been postulated that the generation of increased levels of reactive oxygen species in peroxisomes
may cause indirect DNA damage. The general background of chronic cellular injury, necrosis,
and regenerative cell growth common to peroxisome proliferation may result in sustained DNA
synthesis, hyperplasia, and eventually cancer.
Direct exposure to other TCE metabolites including dichloroacetic acid (DCA) and
chloral hydrate also induce liver tumors in mice, providing support to the theoretical mechanism
of toxic metabolites in TCE-induced liver tumors in mice. These metabolites are not potent
inducers of hepatic peroxisome proliferation, suggesting other mechanisms might be involved.
No evidence of renal toxicity has been observed in people exposed acutely to high vapor
levels of TCE. Mild changes in renal function have been reported in some workers
occupationally exposed to TCE. Chronic inhalation and oral exposure of rats to TCE has resulted
in increased kidney1 weights, minimal to mild cytomegaly, and karyomegaly of the renal tubular
epithelial cells. Renal tumors were also induced in male rats. The mechanisms by which TCE
causes kidney tumors in rats are not known. The kidney effects of TCE in rats do not appear to
be related to an increase in alpha-2u-globulin, and are likely mediated by metabolic activation in
the kidney of a glutathione-conjugated metabolite, N-acetyl-dichlorovinyl-cysteine (DCVC) and
subsequent (3-lyase cleavage metabolism of DCVC. DCVC has been shown to be highly
nephrotoxic and mutagenic in the Ames test. Mice are much less sensitive to the renal effects of
-------
TCE. No renal tumors were found in TCE-exposed mice. This may be due to the fact that
glutathione conjugation, and subsequent DCVC formation is more efficient in rats than in mice.
Inhalation of TCE also induces lung tumors in mice but not rats. Differences between rat
and mouse lung tumor induction may be attributed to differences in lung morphology. Clara
cells are more abundant in mice and distributed in the bronchi and bronchioles, while those of the
rat are located lower in the lung, where their exposure is reduced. TCE-induced lung tumor is
thought to be mediated through the formation and accumulation of chloral in the Clara cells.
These cells lack the capacity to metabolize chloral to trichloroethanol and the subsequent
accumulation of chloral leads to marked vacuolization of the cells. The relevance to humans has
not clearly been established. It has been pointed out that Clara cell morphology of the rat lung is
more similar to humans than mice.
In an effort to provide a quantitative analysis of risk based on the mechanistic data
discussed above, several PBPK models for TCE have been developed in rodents and humans
following oral and inhalation exposure. The models include descriptions of the three principle
target tissues for cancer in animals: liver, kidney and lung. Because the toxicities resulting from
TCE exposure are due to metabolites, oxidative and conjugative metabolism are included in the
relevant compartments (i.e., liver, lung and kidney). The PBPK models are validated against
data from mice, rats and humans following oral and inhalation exposure, including blood, tissue
and breath concentrations of TCE and the metabolites.
-------
Case Specific Questions for TCE
1) What seems to be the series of events leading to each observed toxic response? Are there any
reversible step in the process? Can an irreversible step be identified in each process?
2) Given that TCE induced toxicities are mediated through metabolites, are there common
biological responses across toxicities that would be useful for quantitative analyses?
3) Which of the above selected responses is most relevant to human regarding specificity
(response concordance) and sensitivity (dose range of response)?
4) What additional information would be useful for quantitative analysis?
5) Are dose and duration of exposure important considerations? If so, for which toxicity and how
should they be handled?
6) What response(s) would be useful for dose-response modeling in the observable range for
each toxicity? How does mode of action information influence this choice? Given the availability
of the PBPK models, what would be the appropriate dosimeters for the toxicity observed in the
liver, the lung, and the kidney? Which quantitative models should be used for the observed data?
7) Given what is known about the mode of action for each toxicity, what quantitative approach
would-be recommended for characterizing risk associated with low level exposures (i.e. beyond
the observable range) for each toxicity?
8) If a RfD or MoE were to be developed, which factors should be considered to account for
uncertainties in risk assessment?
-------
CASE STUDY: TRICHLOROETHYLENE
I. Sources and Exposure
TCE is one of the halogenated ethylenes, used in many diverse manufacturing industries
as solvents, carriers, or extractants; in dry cleaning textiles; in metal cleaning and degreasing; in
textile manufacture; as insulating fluids/coolants; and as chemical intermediates. The general
U.S. population is exposed to TCE via inhalation, ingestion, and dermal absorption.
II. Pharmacokinetics
Rodent and human studies indicate TCE is rapidly absorbed from the GI tract and
through the lungs, whereas absorption of the vapor through the skin is negligible. Once absorbed,
TCE is rapidly distributed throughout the body, preferentially to the fat. TCE is metabolized
primarily in the liver but also in the kidney.. The major pathway is oxidative metabolism leading
to the formation of chloroacatic acids. A minor pathway in rodents and humans involves the
formation of mercapturic acids via the GST pathway. The complete metabolic pathway is
outlined in Figure 1.
TCE is metabolized through the P450 pathways to chloral and chloral hydrate (CH). CH
is metabolized rapidly in both humans and experimental animals to TCOH and TCA. Urinary
metabolites from the P450 pathways include oxalic acid, TCOH, TCOG, TCA, and DCA. Mice
have consistently higher rates of biotransformation than rats. TCA has a longer plasma half-life
in humans than in rodents, presumably because there is more binding to plasma proteins in
humans. Much of an administered dose of TCA is excreted unchanged in the urine or rats and
mice. Reductive dechlorination and glutathione conjugation are involved in the formation of the
urinary metabolites, oxalate and thiodiacetic acid. DCA is metabolized in humans and
experimental animals, and oxalate, thiodiacetic acid and unchanged DCA are excreted in urine.
DCA clearance is decreased in humans after repeated administration. Species differences in the
clearance of DCA are observed in rodents: clearance in rats is much slower^than in mice.
Through the GST pathway, DCVC, DCVG, and N-AcDCVC are formed, with the latter excreted
as urinary metabolite in rats. The (3-lyase metabolism of DCVC generates reactive thiol and
subsequent species which might initiate several possible toxic modes of action in the kidney.
Qualitatively, the pathways of biotransformation in humans and animals are identical,
with most metabolites identified in experimental animals also found in humans. However,
quantitative difference exists and likely contribute to the observed species difference in the
toxicities associated with TCE exposure. For example, mice metabolize TCE more efficiently
than rats or humans. The maximum rate of the in vitro metabolism of TCE in humans is 1/3 that
in the rat and 1/4 that in the mouse. Urinary excretion of TCA after repeated dosing of TCE over
five days is constant in rats, but increases steadily in humans; while TCOH excretion increases in
rats and remains constant in humans.
-------
ICE
C y. j^//fV»/l A[ctt4.\
[trichloroethylene oxide]
oxalic
acid
_r.(hydjoxyac_etyj)
aminoethanol
chloral
chloral
hydrate
/ (CH)
Dichloroacetyl
chloride
trichloroethanol trichloroacetic
(EQH) acid OCA)
v/
trichloroethanol
glucuronide
(TCOG)
\
Dichloroacetic
acid (OCA)
PA-
dichloro vinyl
glutathione
(DCVG)
\/
&1,2-dichlorovinyl
cysteine (DCVC)
> monocholoacetic
acid (MCA)
try u*J o
S-1.2-dichlorovinyl-
N-acetylcysteine
(DCVAO
[RS" ] : reactive species
: urinary excretion
-------
III. SUMMARY OF KEY STUDIES
A. Genetic Toxicology
Studies on the mutagenicity of TCE have been performed in bacteria, fungi, yeast, and in
cultured mammalian cells. TCE was not mutagenic in bacterial mutagenicity assays. In cultured
mammalian cells TCE did not induce sister chromatid exchanges in Chinese hamster ovary cells
and did not induce DNA repair in primary cultures of rat hepatocytes. Thus, there is no
convincing evidence that TCE is mutagenic or genotoxic. However, several studies demonstrated
the mutagenicity of DCVC and DCVG. These metabolites are mutagenic in the Ames test and
also induce UDS and DNA double strand breaks (see section below on kidney toxicity).
Cytogenetic damage in lymphocytes was observed in a study of 28 male degreasers
exposed to 206 ppm (1106 mg/m3) TCE, with nine reported to have >13% hypodiploid cells in
cultured peripheral lymphocytes (Konietzko et al., 1978), as compared to normal controls
exposed to 116 ppm (623 mg/m3). The rate of hypodiploid cells was 10.9% in exposed group as
compared to 6.5% in control group. The exposed workers also had a fivefold higher mean rate of
chromosomal breaks per 100 mitoses than the controls. Similar differences on chromosomal
aberrations between exposed and control groups were observed in another study, although sperm
counts and frequencies of abnormal sperm heads and of sperm with two fluorescent Y bodies
were similar in both groups.
No increase in sister chromatid exchanges was seen in peripheral lymphocyte of workers
exposed to TCE for ten years, with TCE concentration of 19.1 - 1066.4 mg/1 total trichloro
compounds found in spot urine sample collection. Similar results are found in several other
studies, although smoking was found to elevate the sister chromatid exchanges in exposed vs.
control groups. A summary of these studies is given in Appendix A
B. Liver
Several occupational and community cohort, and case control studies were conducted to
evaluate the carcinogenic effects of TCE. Although conclusions from many of these studies were
limited due to the presence of confounding factors, such as smoking, altered health state, and
exposure to other solvents, results from the three most informative studies consistently indicate
an excess risk of liver and biliary tract cancer. These studies combined found a total of 16 cases
of primary liver cancer, which excludes cancer of the biliary tract, where 9.5 were expected.
Occupational cohort studies found SIR (2 studies) and SMR (2 studies) greater than or equal to
1.0 for most cancers. Some elevated SIR and SMR were found for liver and biliary tract
(SIR=1.4-1.9, SMR=0.94-1.9). For the case control studies, elevated primary liver cancer was
found from patients in the Finnish.Cancer Registry during 1974-1981, with the range of OR =
1.8 - 3.4, especially high in women. Thus, an association between TCE exposure and liver cancer
has been found in humans. Table B-l summarizes the epidemiological studies showing liver-
effects in human.
-------
Animal studies indicate a definite risk of liver cancer with exposure to TCE. Chronic
exposure to TCE induces liver cancer in mice but not rats following oral and inhalation exposure.
An increase in the formation of hepatocellular adenomas and carcinomas in mice chronically
exposed to 1,000 and 2000 mg/kg/day or 300 and 600 ppm TCE was found. Incidences were
much higher following com oil gavage dosing than inhalation.
Animal studies suggest that the formation of liver tumors is mediated through TCA and
possibly DC A, and species differences are probably due to differences in the formation of TCA.
Mice are more prone to the formation of liver tumors and metabolize TCE more rapidly than rats
or humans, which are less sensitive to the formation of hepatic tumors and are less efficient at
metabolizing TCE. TCA and DC A induce hepatocellular carcinomas when administered in
drinking water to male mice. DCA-induced tumors in male mice following exposures ranging
from 2 to 20 mmol/L in drinking water (365-576 days) resulted in the formation of tumors that
were predominately eosinophilic whereas TCA exposures of the same concentrations and
duration results in tumors that were predominately basophilic.
A dose-related increase in the incidence of malignant tumors and precancerous lesions
was found when mice were exposed to between 1 and 5 ug/L TCA with as little as 12 months of
treatment. These same conditions, however, did not induce tumors in rats. A 52-week drinking
study (1 or 2 ug/L DC A and TCA) resulted in the formation of hepatoproliferative lesions in
male mice, including hepatic nodules, adenomas and hepatocellular carcinomas. DCA treatment
resulted in enlarged livers which were characterized by marked cytomegaly and massive
accumulation of glycogen, whereas TCA resulted in small increases in cell size, a more modest
increase in the accumulation of glycogen and marked accumulation of lipofusin. Areas of focal
necrosis throughout the liver was seen with DCA treatment but not TCA treatment. Both TCA
and DCA induce peroxisome proliferation.
Although TCE causes liver tumors in mice but not rats, toxicities observed with acute and
subchronic TCE exposure in mice and rats have included increased liver weight to body weight
ratio, hypertrophy, small increases in serum levels of liver enzymes and limited necrosis. These
effects were dose dependent both for severity and incidence over dose ranges of approximately
50 to 2,000 mg/kg/day (oral gavage) and 25 to 600 ppm (inhalation). Hepatomegaly was usually
observed in mice treated with hepatocarcinogenic doses of TCE and has been seen following
treatment with doses as low as 100 mg/kg/day for 6 weeks, and in mice exposed to 1, 2.5 or 5
g/L TCE in drinking water for 6 months. Additional early liver effects from TCE exposure
included hypertrophy primarily due to proliferation of the subcelluar organelles, including
peroxisomes. Induction of CYP450s involved in lipid and xeniobiotic metabolism also occurs.
Chronic studies in multiple rat strains report no significant liver pathology.
Similar toxicities were found in animals exposed acutely and subchronically to TCA and
DCA. TCA is known to cause a range of effects in the liver via the peroxisome proliferator-
activated receptor and TCA induces peroxisome proliferation in male mice at the same dose
range that it induces hepatic rumors. Increases in liver weight were found and were linear with
8
-------
dose. Hepatomegaly was observed at TCA doses as low as 0.3 g/L or 100 mg/kg/day in mice.
Short term exposure to 0, 0.5 and 6 g/L DCA in drinking water for 30 days resulted in inhibition
of mitosis, alterations in cellular metabolism and a shift in ploidy class.
In summary, the formation of liver tumors in mice following TCE exposure is likely
mediated by TCA and/or DCA. The more sensitive species, mice, metabolize TCE to TCA and
DCA more efficiently than rats or humans. Moreover, TCA and DCA induce tumors when
administered in the drinking water. The blood concentrations achieved with these studies were
the same as those concentrations achieved following TCE exposure that result in tumor
formation. That is, TCA concentrations resulting from TCE exposure were equal to the
• carcinogenic concentrations when TCE was administered alone. Both TCA and DCA cause
peroxisome proliferation. TCA also causes an increase in lipid peroxidation and DCA exposure
results in marked cytomegaly and a large accumulation of glycogen. This cytomegaly is
associated with the development of focal areas of necrosis which in turn leads to high levels of
cellular proliferation.
Table B-2 to B-4 summarizes animal and molecular effects of TCE and its metabolites which
show several types of liver effects from both short and long term oral/inhalation exposure.
C. Kidney
Kidney toxicity has been reported sporadically in humans and data on the renal effects of
TCE in humans is very limited. Studies at one factory where workers were frequently exposed to
high concentrations have found tubular degeneration and increases in kidney carcinomas.
Concentrations were not measured, so estimates of possible concentrations have been based upon
reports of neurological effects such as dizziness. Several aspects of kidney disease in exposed
factory workers have been studied. Among those workers with kidney cancer, all had varying
degrees of tubular damage. Comparable kidney cancer patients without high exposures to TCE
showed tubular damage in about a half of the cases. Alterations in a kidney-specific tumor
suppressor gene were observed in 100% of the TCE exposed workers while these alterations
were observed in 33 to 55% of those with kidney cancer but not exposed'to the chemical. Table
C-l summarizes the human studies showing kidney effects.
Conflicting results have been found in acute and subchronic studies in animals. Mice
gavaged with 1,100 mg/kg/day. and rats gavaged with 1,000 mg/kg/day, for 3 weeks showed no
evidence of nephrotoxiciry.' In another study, exposure of males rats and mice to 1,000
mg/kg/day for 10 days resulted in elevated cyanide-insensitive palmitoyl CoA oxidase activity in
the kidneys, which is indicative of peroxisome proliferation but not cytotoxicity. Increased
kidney weight but no gross pathological effects were seen in rats given 660 mg/kg/day in the
drinking water for 6 months.
Chronic exposure to TCE induces renal toxicity. Daily administration of 550 to 1,100
mg/kg/day in rats and 1,200 to 2,300 mg/kg/day in mice resulted in treatment related chronic
nephropathy, characterized by degenerative changes in the tubular epithelium. Chronic daily
-------
gavage doses of 500 and 1,000 mg/kg in rats and mice resulted in toxic nephrosis, characterized
by cytomegaly, and cytomegaly of the renal tubular cells coupled with toxic nephropathy. A 52-
week corn oil gavage (250 mg/kg/day TCE) resulted in an increase in renal tubular nucleocytosis
in male rats.
Although.mice display some renal toxicity following acute and chronic TCE exposure,
rats are more sensitive to the formation of kidney tumors. In an NTP study, F344 rats
administered 0, 500 or 1000 mg/kg/day TCE 5 days a week for up to 103 weeks (50 males and
50 females) had an increase in the formation of renal tubular cell adenocarcinomas in the high
dose male group. In a similar study (0, 500, or 1000 mg/kg/day, 5 days/week, 103 weeks), an
increase in the incidence of renal tubular cell adenomas and carcinomas occurred in several
different rats strains. However, interpretation of this study was limited due to reduced survival.
Toxic nephropathy observed in these animals was characterized by cytomegaly, karyomegaly,
and toxic nephrosis of the tubular epithelial cells in the inner renal cortex. The severity of
cytomegaly was proportion to the dose and duration of dosing in animals that died early.
Inhalation studies reveal similar findings. An increase in the formation of renal tubular cell
adenocarcinomas was observed in the high dose group of rats exposed to 0, 100, 300 or 600 ppm
TCE(7hr/dy, 5 dy/wk, 104wks). .
- f '
•s
Although the incidence of renal neoplasms in TCE exposed male rats was not always
statistically significant (p>0.5) relative to concurrent controls, the production of the lesions is
considered to be evidence of a carcinogenic effect in rats, assuming special importance because
they are a rare type of tumor. This conclusion is generally accepted. (Renal adenocarcinomas
have never been observed in the Sprague-Dawley colony in Italy or in any control rats examined
by the NTP, and renal tumors occur only rarely in F344 and O-M rats, according to NTP
historical control data). No increase in renal tumors has been reported in female rats, although a
rare renal tubular cell tumor has been observed in studies in which male rats showed in increase
in these tumors. Other carcinogenicity studies in rats did not produce renal tumors. No renal
tumors have been observed in carcinogenicity studies in hamsters. Table C-2 summarizes some
of these studies.
The formation of kidney tumors in rats is likely mediated by the metabolites formed from
the GSH conjugation pathway, including DCVG, DCVC and DCVC sulfoxide. DCVC and the
reactive species generated from its metabolism are nephrotoxic and nephrocarcinogenic. The
relative importance of this pathway in the formation of renal tumors in dependent on whether
concentrations of these metabolites are high enough following TCE exposure to result in toxicity.
TCE concentrations in the kidney are comparable or higher than that presented to the liver, thus
through interorgan metabolism and biotransformation and concentration in the kidney, the renal
dose of reactive intermediate metabolites is likely to be appreciable as compared to other organs
of the body. In addition, the rat kidneys posses the capacity to metabolize TCE via GSH
conjugation and PBPK models have been used to illustrate that the kidneys are exposed to
significant concentrations of TCE. Thus, this pathway is relevant to the formation of kidney
tumors. . -
10
-------
The kidney tumors formed following TCE exposure are very rare and a single mode of
action has not been identified, but studies in animals indicate that mutagenicity and cytotoxicity
from DCVC are involved. Peroxisome proliferation and accumulation of Alpha-2u-globulin
induced nephropathy have been ruled out as mechanisms but DCVC has been shown to be highly
nephrotoxic and mutagenic in the Ames test. Rats exposed to TCE either by com oil or inhalation
show characteristic signs of proximal tubular damage, such as elevated levels of urinary N-
acetyl-b-glucosaminidase, GOT and glucose excretion, and blood urea nitrogen. Similarly,
DCVC or DCVG administered intraperitoneally to male rats resulted in an increase in blood urea
nitrogen and urinary glucose excretion. Rats exposed chronically to TCE (550 to 1,110
mg/kg/day) showed chronic neuropathy, characterized by degenerative changes in the tubular
epithelium and cytomegaly.
Most of the studies examining the early biological effects have focused on DCVC or
DCVG administered as the parent compound although some have looked at TCE. Cytotoxicity
from TCE and DCVC exposure occurs secondary to oxidative stress, which is characterized by
GSH depletion, lipid peroxidation and oxidation or alkylation of protein sulfhydryl groups.
DCVC causes a disturbance in CA2" ion homeostasis, alterations in mitochondria! function,
including inhibition of mitochondrial macromolecular synthesis and DNA damage. DCVC also
induces the repair-proliferative response which may lead to kidney damage and the development
of neoplasias.
The relevance of the formation of kidney tumors in rats to humans has not been
established. A recent study reported blood levels of DCVG, a precursor of DCVC, in humans
exposed to occupational!)' relevant concentrations of TCE (4-hour exposure to 50 and 100 ppm
TCE). Sex-dependent differences in were also found; peak blood levels in men were 2-fold
higher than in women and were reached sooner than in females. Since male rats are more
susceptible to the nephrotoxic and nephrocarcinogenic effects of TCE and also have a higher rate
of GSH conjugation in the liver and kidney, these finding suggest that men may be at a greater
risk of developing nephrotoxicity from TCE exposure.
In summary, studies illustrate that at high doses, DCVC produces oxidative stress,
protein and DNA alkylation, and mitochondrial dysfunction. Cytotoxicity occurs secondary to
inhibition of active transport mechanisms and marked ATP depletion, and acute tubular necrosis
occurs. At lower doses, mild changes in mitochondrial function and oxidative stress, as well as
selective alkylation of protein and DNA occurs. These processes lead to changes in hemeostatic
processes in the cell and alter gene expression and cell growth.
D. Lung
Epidemiological studies found an increase in susceptibility to pulmonary functions in
humans exposed to TCE vapors. This effect has also been found in mice.' Table D-l gives a
summary of these studies.
11
-------
Lung tumors are found only in mice, not in rats, and only observed in inhalation studies
only. These observations come from two studies: inhalation study in female ICR mice (males
are not studied) at 150 and 450 ppm (mid and high doses), and inhalation studies in male Swiss
mice (no female) at 300 and 600 ppm (mid and high doses), and in female B6C3F1 mice (not
male) at 600 ppm (high dose only).
The formation of lung tumors is mediated through the formation and accumulation of
chloral in the Clara cells. These cells lack the capacity to metabolize chloral to trichloroethanol
and the subsequent accumulation of chloral leads to marked vacuolization of the cells.
Acute exposure (30-minutes) to 500 ppm in mice resulted in the vacuole formation and
endoplasmic reticulum dilation hi the Clara cells of the bronchiole tree. Similar results occurred
following a 6-hour exposure to 100 ppm. Additional early effects observed include alterations,
such as foci of perivascular inflammation on small pulmonary veins. Rats, however, showed no
histopathological changes in these studies, and in a 6-week or 90-day exposure to 700 ppm TCE.
Repetitive exposure CHL has been shown to be highly genotoxic in a number of studies.
i
The relevance to humans has not been clearly established. Lung tumors are not found in
rats exposed to TCE, which is likely due to differences in the morphology of lung: Clara cells are
more abundant in the mouse than rat and are located in the bronchi and bronchioles. In the rat,
these cells are located in lower lung and are subject to less exposure. The rat lung morphology is
more similar to the mouse than to the human indicating that humans may not be susceptible to
the formation of lung tumors. Table D-2 summarizes some of the animal studies showing lung
effects.
E. Reproductive and Developmental Effects
Studies on the various reproductive and developmental effects of TCE have yielded
conflicting results. One study found an increase in miscarriage among nurses exposed to TCE
and other chemicals in the workplace, although no specific association with TCE was found.
Another study found no increase in malformations in the children of 2000 fathers and mothers
exposed to TCE via inhalation. An association, but no direct cause and effect relationship, was
found between elevated levels of chlorinated hydrocarbons, including TCE, in drinking water
and congenital heart disease in children of exposed parents.
In one study, semen specimen from workers (15) using TCE for degreasing for more than
20 hrs per week were compared to those from unexposed physicians (14). There was no
difference between the two groups in terms of sperm count or morphology, but the exposed
group had a small, statistically non-significant increase in the prevalence of mature spermatozoa
containing two fluorescent Y bodies, which may indicate Y-chromosomal nondisjunction (EBE).
Reproductive and developmental toxicity studies in animals are limited, but suggest
associations with cardiac anomalies and eye malformations. The developmental effects of TCE
are largely associated with the oral route of exposure. Eye malformations, including a reduction
12 •'
-------
in or absence of the ocular bulge, have been found in rats exposed in utero (days 6-15) at high
doses (1125 and 1500 mg/kg/day). Cardiac malformations have also been found in drinking
water studies and upon direct placement of TCE in the uterus. An increase in abnormal sperm
morphology in mice exposed to TCE by inhalation was found. Effects on implantations, litter
size and fetal resorption or other similar measures of reproductive success were found at high
oral doses (1000 mg/kg/day). Although the mode of action is not known with certainty, TCA and
DC A are implicated in the developmental toxicities.
F. Neurotoxicity
Neurological effects are associated with exposures to a wide range of concentrations of
TCE in air. Anesthesia required approximately 2,000 ppm. Controlled-studies with volunteers
exposed for short times (hours) found neurological effects including sleepiness, reductions in
motors skills, and altered rates of breathing and heart beat. One study (200 ppm for 7 hours for 5
days) reported mild fatigue and sleepiness. A slight trend toward slower pulse rate was found at
27 and 81 ppm for 4 hours and no effect on heart beat or breathing rates was found at exposure to
200 ppm TCE for 2.5 hours. Exposure to 110 ppm for 8 hours resulted on decreased
performance on skills tests. Controlled studies with exposure to the metabolites, CH and TCOH,
report similar effects. The neurotoxic effects of DC A observed repeatedly in experimental
animals have rarely been documented in clinical trials. Drowsiness is a fairly frequent side-effect
of DC A and has been observed in healthy volunteers, adults with type II diabetes and patients
with lactic acidosis.
G. Immiinotoxicology
Impaired immune function has been observed at high oral and inhalation exposures in
animals, including reduced spleen cell number or fractional spleen weight, decreased leukocyte
count. The mode of action for the immunological effects is unknown.
IV. Quantitative analysis
A. Physiologically Based Pharmcokinetic Models
Several PBPK models have been developed for TCE in rodents and humans following
oral and inhalation exposure. The models include descriptions of the three principle target tissues
for cancer in animals, liver kidney and lung, in addition to fat, rapidly and slowly perfused
compartments. Oxidative and conjugative metabolism, as described above, are included in the
relevant compartments (i.e., liver, lung and kidney). The PBPK models are validated against
data from mice, rats and humans following oral and inhalation exposure, including blood, tissue
and breath concentrations of TCE and the metabolites. Several possible dosimeters could be
selected for each endpoint, including concentration, AUC (area under the curve) of either TCE or
its metabolites in blood, urine or relevant tissues.
13
-------
B. Biologically Based Dose Response Model for Liver Tumor
Recent study by Bull and colleagues indicate that TCE may induce liver tumors by
selection and promotion. They demonstrated that TCE and its metabolites (DCA and TCA)
induce single strand breaks (SSB) in DNA in mouse liver in vivo. As noted by Nelson and Bull
(1988), the induction of SSB in DNA has been associated with both initiation and promotion
events in chemically induced carcinogenesis. In Bull et al (1990), male and female mice were
administered DCA and TCA in drinking water at concentration of 1 or 2 g/1 for up to 52 weeks.
Suspension of DCA at 37 weeks resulted in the same number of hepatocellular proliferative
lesions(HPL) at 52 weeks that would have been predicted on the basis of total dose administered.
However, none of these lesions progressed to carcinomas, indicating that continuing DCA
treatment is necessary for conversion to carcinomas. A possible explanation of this observation is
that DCA induces progression of malignant tumor cells.
14
-------
Appendix A: Genetic and related effects of TCE and metabolites
Table A-l. Summary of genetic and related effects of TCE and metabolites ,
A: aneuploidy; C: chromosomal aberrations; D: DNA daniage; DL: dominant lethal mutation; G: gene mutation; I: inhibition of intercellular communication; M:
micronuclei; R: mitotic recombination and gene conversion; S: sister chromatid exchange; T: cell transformation
+/-: considered to be positive/negative for the specific end-point and level of biological complexity
?: considered to be equivocal or inconclusive
(I: only I valid study is available)
TCI2 without
imitagcnic stabilizers
TCI2 with mut stab,
or of uncertain purity
CM
DCA
TCA
in vitro
Animal cells
D G S M C A T I
-1 +1 + -1 +1 +1
+ -1 -1 +1
i- +1 + -1
1
-1 +1
Human cells
D G S M C A T I
-I
?!-•
-1 - ?! + +
-1
-1
in vivo
Animal cells
D G S M C DL A
+ -1 -1 + -1 -1
-1 + -
? + . +
?
? ? +1
Human cells
D G S M C A T I
15
-------
Appendix B: Liver studies
Table B-1: Epidemiologic studies of liver and biliary passages
Exposure Type-Level
Urinary TCA
Urinary TCA
Job matrix-cum.
exposure
Qualitative inference
of TCE exposure, No
IH data used to infer
exposure
Comments
Incidence, liver only
Years since first
urinary measurement
Incidence, liver only
Too few cases to
examine E-R trends
Stronger RR seen in
most recent follow-up
period (RR=2.3)
Logistic regression
Only deaths among
pensioned employees
included in analyses.
Relative Risk
2.27(0.74-5.29)
0-9 yr, 0 deaths
10-19 yr, 1.74(0.21-
6.29)
20+ yr, 6.07 (1.25-
17.7)
Urinary TCA
<100^mol/L, 1.64
,(0.20-5.92)
100+Aimol/L, 2.74
(0.33-9.88)
1.4(0.4-3.6)
1.3(0.5-3.4)
cf , ppm-yr:
<5, 1.1 (0.3-4.1)
5-25, 0.9 (0.2-4.3)
>25, 0.7(0.2-3.2)
$, ppm-yr:
<5, 1.6(0.2-18.2)
5-25, 0 deaths
>25, 2.3 (0.3-16.7)
Ever versus never
exposure to TCE,
OR=0.54 (0.1 1-2.63)
Reference
Anttilaetal. (1995)
Axelsonetal. (1994)
Blair et al. (In press)
Greenland etal. (1994)
16
-------
Job matrix
Cumulative exposure
Cox proportionate
hazards modeling.
97(35.9-213.1)
Ever vs. Never,
RR= 1.5 (0.56-3.9)
Peak exposure,
RR=0.98 (0.29-3.35)
Cum. exposure,
<2000,RR=2.12
(0.59-7.66)
>2000, RR=1.19(0.34-
4.16)
Morgan et al.
(submitted)
17
-------
Table B-2. Liver effects from short and long term exposure to TCE and metabolites in experimental animals
Species (sex)
Dose
Duration
Liver effects
Reference
TCE
Mice
136C3FI
(Swksold)
20 (M)
50 (M)
20 (F)
50 (F)
B6C3FI
(8 wks old)
50 (M)
50 (F)
B6C3FI
(M) (2 studies)
(F)
Species (sex)
mg/kg COG
5 days/vvk
0
1169
2339
0
869
1739
mg/kg COG
0
1000
0
1000
ppm inhalation
0
100
300
600
0
100
300
600
Dose
90 wks
78 wks
78 wks
90 wks
78 wks
78 wks
103 wks
103 wks
103 wks
103 wks
78 wks
78 wks
78 wks
78 wks
78 wks
78 wks
78 wks
78 wks
Duration
survival
8/20
36/50
22/48
20/20
42/50
39/47
Mepatocellular
N&A
7/48
14/50
4/48
16/49
Tumor incidence
3.3%
1.1%
3.3%
6.2%
2.2%
4.4%
4.4%
10%
Hepatocellular
Carcinoma
1/20
26/50
31/48
0/20
4/50
11/47
Mepatocellular
Carcinoma
8/48
31/50
2/48
13/49
Tumor incidence
18.9%
21.1%
30%
23.3%
Liver effects
NCI, 1976
NTP, 1990
Maltoni, 1986
Maltoni, 1988
Reference
18
-------
I36C3FI
NMRI (M & F)
NMRI (M & F)
1CR (M & F)
Swiss-Cox
Swiss-Cox
mg/kg
0,100,400
ppm inhalation
0,100,500
ppm inhalation
0,37,75,150
300
ppm inhalation
0,50,150,450
ppm inhalation
0
100
300
600
inhalation
(ppm)
0,100,400, 1600
6 wks
1 8 months
(observed
30 months
30 days,
24hrs/day
104 wks
(obs 107
wks)
7 lir/day
5d/wk
\
6 wks, 5
days/wk
increase relative liver weight, enlarged liver
cells at 100 and 400 mg/kg
no significant tumor incidence
increase liver weight in M and F at >75 ppm
(NOAEL = 1 50), liver weight and enzyme
change return to normal after 4 months
following treatment
no significant tumor incidence
hepatoma
M
4/90
2/90
8/90
13/90
enlarged hepatocytes
at 400 ppm, 1600 ppm
no effects in F
centrolobular
necrosis at 1 600
ppm
Henschler et
al., 1980
Kjellstrand et
al., 1981,
1983a
Fuduka et al,
1983
Maltoni, 1986
Buben and
O' Flaherty,
1985
Rat
19
-------
Species (sex)
Osborne-Mendel
(M&F)
AC! (M&F)
August/M&F
Marshall/M&F
Osborne-
Mendel/M&F
F344/M&F
WISTAR/M&F-
SD/F
SD
Dose
oral, COG
B-549&1097
mg/kg
oral, COG 500,
1000 mg/kg
oral, COG 500,
1000 mg/kg
oral, COG
500, 1000
mg/kg
oral, COG
500, 1000
mg/kg
oral, COG
500, 1000
mg/kg
inhalation
0, 100, 500 ppm
inhalation 50,
150, 450 ppm
100,300,600
ppm
inhalation
Duration
78wks
(HOwks)
103 wks
(103\vks)
103 wks
(103 wks)
103 wks
(103 wks)
103 wks
(103 wks)
103 wks
(103 wks)
18 months
(36)
months
104 wks
(107 wks)
104 wks
(until
death)
Liver effects
no effect at all doses
no effect at all doses
no effect at all doses
>
no effect at all doses
no effect at all doses
no effect at all doses
no effect at all doses
no effect at all doses
no effect at all doses
Reference
NCI, 1976
Henscliler, et.
a!., 1980
NTP, 1988
NTP, 1988
NTP, 1988
NTP, 1990
Henscliler, et.
a!., 1984
Fukuda et al., .
1983
Maltoni et al.
1986
TCA
Mice
20
-------
Species (sex)
B6C3FI (M)
B6C3F1 (M)
I36C3FI (M)
B6C3F1 (F)
Dose
0
5g/L
0
1 g/L
2 g/L
2g/L
0
0.05 g/L
0.5 g/l,
4.5 g/L
5 g/L
0
0.35
1.2
3.5
0
0.35
1.2
3.5
Duration
61 wks
52 wks
3 7 wks .
60-95 wks
60 wks
.95 wks
52 wks
81 wks
-
Hepatocellular
Nodules & Adenomas
2/22
8/22
1/35
5/11
1 5/24
2/11
not reported
nr
nr
nr
nr
1/40
6/40
3/19
2/20
2/90
14/53
12/27
18/18
Hepatocellular
carcinoma
0/22
7/22
0/35
2/11
4/24
3/11
6.7-10%
22%
38%
87%
55%
0/40
0/40
0/19
5/20
2/90
0/53
5/27
5/18
Reference
Herren-
Freund et al.,
198
Bull et al.,
1990
Daniel et al,
1993
Pereira, 1996
DCA
Species (sex)
Dose
Duration
Hepatocellular
Nodules & Adenomas
Hepatocellular
carcinoma
Reference
21
-------
B6C3FI (M)
B6C3FI (M)
B6C3FI (M)
B6C3F1(F)
B6C3F1 (F)
Species (sex)
0
5 g/L
1 g/L
2 g/L
2 g/L
0
0.5 g/L
3.5 g/L
5 g/L
o
0.05 g/L
0.5 g/L
Ox 104\vks
0.5 g/L
0 x 52 wks
0.28
0.93
2.8
0x81 wks
0.28
0.93
2,8
Ox 104 wks
0.5 g/L
3.5 g/L
Dose
61 wks
52 wks
37 wks
60 wks
75wks
Duration
25/26
2/11
23/24
7/11
0/10
-
12/12
27/30
2/28
4/29
3/27
1/20
12/24
1/40
0/40
3/20
7/20
2/90
3/50
7/28
16/19
nr
nr
nr
Hepatocellular
Nodules & Adenomas
21/26
-
5/24
0/11
0/10
-
8/12
25/30
-
-
-
2/20
15/24
0/40
0/40
0/20
1/20
2/90
0/50
1/28
5/19
1/39
1/25
23/25
Hepatocellular
carcinoma
Herren-
Freund et al.,
198
Bulletal.,
1990
DeAngelo et
al, 1991
Daniel et'al.,
1992
Pereira, 1996
Schroeder et
al., 1997
Reference
Rats
22
-------
F344(M)
F344 (M)
0
0.05 g/L
0.5 g/L
2.4 g/L
0
0.05 g/L
0.5 g/L
2.4 g/L 0
0
0.05
0.5
1.6
Chloral Hydrate
Mice
C,7BLXC,I1F,
(single dose to
neonatal mice)
I36C3F1
0 mg/kg
5
10
0
1 g/L
60 wks
60 wks
60 wks
60 wks
1 04 wks
104 wks
104 wks
104 wks
104 wks
92 wks
92 wks
92 wks
104 wks
1 04 wks
0/7
O/
0/7
26/27
1/23
0/26
-
not done
1/33
0/26
5/29
4/28
0/7
0/7
0/7
1/27
0/23
0/26
3/29
not done
1/33
0/26
3/29
6/28
Richmond et
al., 1995
DeAngelo et
al., 1996
0/19
2/9
3/8
1/20
8/24
2/19
1/9
3/8
2/20
11/24
Rijhsinghani .
etal., 1986
Daniel et al.,
1992
23
-------
Table B-3. Mutation frequency and spectra with codon-61 of Ha-ras of B6C3F1 mice treated with TCE
and its metabolites
Chemical
#H-ras61/
tumors
Mutation
Frequency
gin
CAA
lys
AAA
arg
CGA
leu
CTA
Male mice
Spontaneous
hepatocarcinoma"
TCE
DCA
1 g/Lx 104wksb
3.5 g/Lx 104wksb
5 g/L x 76 wksc
0.5 g/L x 90 wks
2 g/L x 52 wks
combined '•
TCA
4.5 g/Lx 104wksc
2g/L,x52wksd
1.9 g/Lx 90 wks
combined
179/333
39/76
6/13
16/33
40/64
10/28
7/26
79/164
5/11
16/30
2/11
23/52
0.54
0.51
0.46
0.48
0.63
0.36
0.27
0.48
0.45
0.53
0.18
0.44
150(0.45)
34 (0.45)
7 (0.54)
17(0.52)
24 (0.37)
18(0.64)
19(0.73)
85(0.52)
6 (0.55)
14 (0.47)
9 (0.82)
29 (0.56)
106(0.32)
12(0.16)
7 (0.08)
3 (0.09)
11 (0.17)
4(0.14)
3 (0.12)
22(0.13)
4(0.36)
7 (0.23)
0(0)
11 (0.21)
50(0.15)
10(0.13)
3 (0.23)
8 (0.24)
14(0.22)
2 (0.07)
2 (0.08)
29(0.18)
1 (0.09)
6 (0.02)
1 (0.09) .
8(0.15)
21 (0.06)
17(0.22)
2(0.15)
5(0.15)
15 (0.23)
4(0.14)
2 (0.08)
28(0.17)
0(0)
3(0.1)
1 (0.09)
4 (0.08)
Female mice - .
Spontaneous
hepatocarcinoma3
DCA<
3.5 g/Lx 104 wks
33/49
1/22
0.67
0.05
16(0.33)
21/22
(0.95)
17(0.35).
0(0)
12 (0.27)
0(0)
4(0.08)
1 (0.05)
" Maronpot et al. (1995) Toxicology 10!, 125-156
b Ferreira-Gonzalez et al. (1995) Carcinogenesis 16, 495-500
c Anna et al., (1994) Carcinogenesis 15, 2255-2261
d Orner et al., (1998) In press
24
-------
Table B-4: Studies reporting noncancer tumor effects
Species
Dose route & matrix Doses (mg/kg/d)
Endpoints
References
Oral studies
Rats
Rats
Mice
Mice
Mice
Mice
Mice
aqueous emulsion (5% emulphor)
0,100,250,400
micro encapsulated in diet
0,600, 1300,2200,4800
corn oil gavage
0, 100, 200, 400, 800, 1600, 2400, 3200
corn oil gavage
0,500,1000,1500
corn oil gavage
0,50, 100,200,500, 1000,2000
groundnut oil gavage
0, 500, 1000, 2000
corn oil gavage or aqueous emulsion
0,600, 1200, 2400 (M)
0,450,900, 1800(F)
com oil gavage
0,250,500, 1200,2400
drinking water
0, 18,2*17, 393,660.(M)
0, 18, 193, 437, 793 (F)
LW/BW ratio
LW/BW ratio
LW/BW ratio
serum enzyme levels
LW/BW ratio
DNA/cell histopathology
LW/BW ratio
palmitoyl CoA oxidation
LW/BW ratio,
a-aminolevulinic
dehydratase
histopathology
LW/BW ratio
LW/BW ratio
LW/BW ratio
Borzelleca et al.,
1990
Melnick et al.,
1987
Buben.and
O'Flaherty, 1985
Elcombe et al.,
1985
Elcombe, 1985
Goeletal., 1992
Merrick et al.,
1989
Scott etal., 1982
Tucker et al.,
1982
Inhalation study
Rats
Rats,
mice
730 ppm (8 h/d, 5 d/wk, 6 wks)
37-300 ppm continuously for 30 days
no gross pathological liver
effects
LW/BW increase, mice
more sensitive than rats or
gerbils, largely reversible
in 30 days following
exposure
Prendergast et
al., 1967
Kjellstrand et al.,
1981, 1983a
25
-------
Appendix C: Kidney
Table C-l: Epidemiological study of Kidney effects
Author
Anttilaetal. (1995)
Axelsonetal. (1994)
Blair et al. (In press)
Greenland etal. (1994)
Henschler et al. (1995)
Exposure Type-Level
Urinary TCA
Urinary TCA
Job matrix-cum.
exposure
-
Qualitative inference
of TCE exposure, No
IH data used to infer
exposure
Qualitative inference
Relative Risk
0.87(0.32- 1.89)
Years since first
exposure:
0-9,0.53(0.01-2.95)
10-19, 1.39(0.45-3.24)
20+, 0 cases
1.16(0.4-2.5)
1.6(0.5-5.1)
25, 1.2(0.3-4.8)
?, ppm-yr:
<5, 0 deaths
5-25,9.8(0.6-157)
>25, 3.5 (0.2-56.4)
OR=0.99 (0.30-3.32)
SIR=11.15(4.49-
23.00), comparison
with Danish Cancer
Registry
SIR=9.66 (3. 14-22.55),
comparison with
former G.D.R. Cancer
Registry
Comments
Incidence
PER:SIR=1.82(0.22-
6.56; 2 cases)
Incidence, no E-R or
duration of exposure
analyses (for kidney)
Stronger RR seen in
most recent follow-up
period (RR=2.6)
Cumulative exposure,
Poisson regression
analysis
Logistic regression
analysis for ever
exposure to TCE
Incomplete
identification of
cohort, incomplete
ascertainment of
deaths, lack of
exposure data for total
cohort.
26
-------
Morgan et al.
(submitted)
Job matrix
131.9(57.0-259.9)
Ever vs. Never,
RR=1.1 (0.51-2.58)
Peak exposure,
RR=1.89 (0.85-4.23)
Cum. exposure,
<2000,'RR=0.31
(0.04-2.36)
>2000,RR= 1.59 (0.69-
3.71)
Cumulative exposure
Cox proportionate
hazards modeling.
MAN Document
(1997)
Questionnaire obtained
job history
OR=13.42(3.50-51.39)
Logistic regression
analysis controlled for
effects of age, sex,
smoking, body mass
index, blood pressure
and diuretic use.
Cases identified
between 1988 and
1992. All controls
identified in 1992.
Possible bias
introduced since TCE
usage decreased over
this time period.
27
-------
Table C-2: Kidney effects in rats exposed to TCE
Species (sex)
Dose
Duration
Kidney effects
Reference
Oral studies
F344(M&F)
F344(.M&F)
F344 (M & F)
Osborne-Menclel
Marshall, ACI,
August, Osborne-
Mendel (M & F)
Oshorne-Mendel
(M&F)
mg/kg/day COG
C(Xi, mg/kg/day
0
500
1000
DW, 393 and 793
mg/kg/day
mg/kg/day COG
549, 1097 .
nig/kg/day COG
0,500,1000
mg/kg/day COG ~
0
500
1000
13\vks, 5d/wk
101 wks, 5d/wk
6 months
78 wks, 5 d/wk
103\vks
no effect at 500, cytomegaly renal tubular seen in
5/10 F at 1000 mg/kg/day, and in M at 2000
mg/kg/day
nephrosis and
cytomegaly
M F
.19/49 49/49
49/49 49/49
renal tubular cellular
adenocarcinoma (M only)
0/48 ' .
0/49
3/49 (not significant)
increase kidney weight in M at 393 nig/kg and in
F at 793 mg/kg
chronic nephropathy, characterized by
degenerative changes in tubular epithelium
renal cytomegaly > 80% in all treated M&F,
toxic nephropathy in 17-80% of treated groups
no difference in kidney toxicity between M&F
renal tubular
cellular
hyperplasia
0/50
5/50
3/50
adenoma
0/50
6/50 (not significant)
1/50
NTP, 1990
NTP, 1990
Tucker, 1982
NCI, 1976
NTP, 1988
NTP, 1988 (same
study as above)
28
-------
Inhalation
SD
ppm inhalation
0
100
300 .
600
104 wks, 5d/wk,
7hr/day
cytokaryomegaly
M
no effect
16.9%
77.7%
Renal tubular
adenocarcinoma
3.1%inM
Maltoni
29
-------
Appendix D: Lung effects
Table D-l: Epidemiological studies of lung effects
Author
Anttilaetal. (1995)
Axelsonetal. (1994)
Blair et al. (In press)
Greenland etal. (1994)
Exposure Type-Level
Urinary TCA
Urinary TCA
Job matrix-cum.
exposure
Qualitative inference
ofTCEexposure.no
IH data used to infer
exposure
Relative Risk
0.92(0.59-1.35)
Years since 1st
exposure:
0-9 yr, 1.19(0.59-2.13
1 0-19 yr, 0.67 (0.30-
1.26)
20+ yr, 1.11 (0.36-
2.58)
Urinary TCA
<100^mol/L, 1.02
(0.58-1.66)
lOO+^mol/L, 0.83
(0.33-1.71)
0.69(0.31-1.30) •
0.9(0.6-1.3 )
25, 1.1 (0.7-1.8)
$ , ppm-yr:
<5, 0.6 (0.1-2.4)
5-25,0.6(0.1-4.7)
>25, 0.4 (0.1-1.8)
OR=1.01 (0.69-1.47)
Comments
Incidence
-
Incidence, E-R analysis
not presented for
lung/bronchus
Cumulative exposure,
Poisson regression
analysis
Logistic regression
analysis for ever
exposed toTCE
30
-------
Morgan et al.
(submitted)
Job matrix
109.8(89.1-134.0)
Cumulative exposure
Statistically significant
.elevated risk observed
for bronchitis,
emphysema, and
asthma.
Hardell et al. (1994)
Questionnaire obtained
information on
occupational and non-
occupational exposures
OR=3.4 (1.3-42)
Univariate analysis
Logistic regression
analysis showed
elevated OR for "high
grade" exposure to
organic solvents
(OR=3.5; 1.7-7.1)
31
-------
Table D-2: Lung cancer effects from long term exposure to TCE in experimental animals
Species (sex)
Dose
Duration
Lung tumors
Reference
TCE
Mice
ICR
(7 wks old)
49-50 F
Swiss
(11 wks old)
90 M/F
B6C3F1
(12 wks old)
90M/F
ppm
0
50
150
450
ppm
0
100
300
600
ppm
0
100
300
600
5 days/wk,
7 hr/d
104 wks
104 wks
104 wks
104 wks
5 days/wk,
7 lir/day
78 wks
78 wks
78 wks
78 wks
5 days/wk,
7 hr/day
78 wks
78 wks
78 wks
78 wks
lung adenocarcinoma
1/49
3/50
8/50
7/46
lung tumors
effects in M only
10/90
11/90
23/90
27/90
no effects in M in all
dose groups
lung adenoma &
adenocarcinoma
6/49
-
13/50 (not significant)
1 1/46 (not significant)
no effect in all dose
groups in F
lung tumors in F only
4/90
6/90
7/90
15/90
Fukuda, 1983
Maltoni, 1986, 1988
Maltoni, 1986, 1988
32
-------
FRAMEWORK FOR HUMAN HEALTH RISK ASSESSMENT
Colloquium #2
Case Study: Vinyl Acetate
Executive Summary
Vinyl acetate is a synthetic organic ester with a wide range of uses. The 1990 Clean Air
Act Amendments include vinyl acetate as one of the 189 hazardous air pollutants listed
under Title III because it has been shown to be a respiratory tract toxicant in experimental
species.
Vinyl acetate induces nasal tumors in rats, but not mice, following long-term inhalation
exposure. Tumors have also been shown in the portal of entry (buccal cavity, esophagus,
and forestomach) after oral administration at high concentrations (10,000 ppm), but the
focus of this case study is inhalation exposure. All but one of the nasal tumors were
observed at the terminal sacrifice indicating a late-life dependency of tumor formation.
Non-neoplastic lesions of the nasal cavity were observed in both rats and mice. These data
show that the upper respiratory tract (URT), and in particular the olfactory mucosa, is the
primary target of vinyl acetate toxicity in both rats and mice. This portal-of-entry toxicity
is not surprising given that vinyl acetate is metabolized to acetic acid and acetaldehyde by
nasal carboxylesterase. Rats are chosen as the most sensitive experimental species because,
while olfactory degeneration was shown in both species at comparable exposure
concentrations, tumors were observed only in rats.
Olfactory degeneration is proposed as a sentinel lesion in rodents based on dosimetry
considerations and on the pathogenesis continuum believed to be involved in tumor
development. Degeneration of the olfactory epithelium is rapidly followed by an induced
proliferative response, which is likely the driving force or rate limiting step behind the
observed tumor formation. Dose-response modeling of olfactory degeneration is thereby
proposed as a relevant basis for human health risk assessment of both noncancer toxicity
and carcinogenesis. Olfactory degeneration is mechanistically linked to the in situ
formation of acetic acid in olfactory sustentacular cells and the consequent loss of control
over intracellular pH. Acetaldehyde, a known clastogen (chromosome breaking agent) and
inducer of sister chromatid exchanges, is also formed during the hydrolysis of vinyl acetate.
Acetaldehyde's involvement in the development of nasal tumors is unclear, and is more
-------
likely to occur only at high exposures where saturation of its detoxification would occur.
Acetaldehyde is rapidly oxidized almost exclusively to acetic acid by NAD+-dependent
aldehyde dehydrogenase. Airflow dynamics and the distribution of enzymes which
metabolize vinyl acetate are shown to be key determinants of uptake and distribution in the
tissues of the URT of the rat.
Mechanistic studies have been conducted to test the hypothesis that both the genotoxicity
and degenerative cytotoxic effects were related to carboxylesterase-mediated hydrolysis of
vinyl acetate to acetaldehyde and acetic acid. These studies support the hypothesis that the
cytotoxic and carcinogenic effects of vinyl acetate are related to the carboxylesterase-
mediated formation of acetic acid, a strong cytotoxicant, and acetaldehyde, a clastogen.
In vitro studies on both vinyl acetate and acetaldehyde suggest that neither induces point
mutations. Research on the stability of acetaldehyde-induced DNA-protein crosslinks, show
the crosslink is unstable at physiological temperature and pH (t,/2 = 6.5 hr) raising the
possibility that the carcinogenic effect of vinyl acetate is less dependent on acetaldehyde-
induced DPXL and more dependent on acetic acid-induced cytotoxicity. Furthermore,
research on the effects of pH on clastogenic activity in vitro show that low pH alone can
induce clastogenic responses similar to those induced by vinyl acetate. Therefore,
conditions under which intracellular pH is maintained in a physiological range such that
cytotoxicity-induced cell proliferation is prevented will likely minimize any potential
contribution of metabolite, acetaldehyde, to the formation of DNA protein crosslinks and
potential consequent clastogenesis.
Cell proliferation studies showed a rebound response of olfactory epithelium to extended
vinyl acetate exposure (1 day vs. 5 days, vs. 20 days). These results suggest that
restorative cell proliferation within the basal cell compartment, to. replace lost sustentacular
cells, becomes the driving event for neoplastic growth. Thus, in olfactory epithelium, the
carcinogenic response to vinyl acetate exposure appears to be driven largely by a cytotoxic
proliferative mechanism.
Determinants of uptake and nasal tissue dose were evaluated extensively including
quantitative estimation of kinetic constants governing carboxylesterase and aldehyde
dehydrogenase activities and histochemical localization of their cellular distributions.
Separate methods of analysis of carboxylesterase activity suggested that the enzyme is
-------
localized in nasal tissue in functionally distinct compartments. Experiments utilizing a
unique in vitro whole tissue gas uptake system demonstrated that vinyl acetate is almost
completely metabolized in the most superficial compartment of olfactory epithelium
(sustentacular cells). Therefore, sustentacular cells are likely the primary target of vinyl
acetate-induced olfactory toxicity. Also of importance is the observation from the
histochemical analyses that in olfactory epithelium, the basal cells, which are stem cells for
proliferative regeneration and presumably are progenitor cells for neoplastic growth, are
devoid of carboxylesterase activity.
A physiologically-based model of the URT has been developed that describes vinyl acetate
vapor deposition and metabolism, and acetic acid-induced changes in intracellular pH in the
rat. Dosimeters generated from the modeling showed a pattern that is consistent with the
overall mechanistic hypothesis. The intracellular pH of olfactory epithelium was predicted,
through simulations, to drop at external exposure concentrations above 50 ppm. This is
consistent with observations from the 2-year inhalation bioassay that olfactory degeneration
occurs at concentrations above 50 ppm. Therefore, dosimeters related to intracellular pH
(total amount of acetic acid formed, and final proton concentration in olfactory tissue)
appear to be tenable measures of tissue dose on mechanistic grounds.
-------
Case Specific Questions for Vinyl Acetate
I. Does the existing database support the URT lesions as the sentinel toxicity for
inhalation exposures to vinyl acetate?
II. Can the cytotoxic changes caused by vinyl acetate exposure be considered as
sequentially linked to the observed tumor outcome? What are the key
considerations to characterize the conditions of hazard (e.g., high dose versus low
dose)? How do the genotoxic data factor in this characterization?
III. What mechanistic data are most relevant to characterizing tumor outcome? Which
would be useful for dose-response modeling in the observable range? What are the
implications of the mode of action information for extrapolation of risk to low
dose?
IV. Given the availability of the PB-PK model, which dose metrics should be
considered for the dose-response analysis? Does, this choice of dose metric address
consideration of the role of exposure duration?
V. What are the uncertainties in using these data to characterize human risk?
VI. Should an RfC be developed separately? If an RfC or MOE were to be developed,
which factors should be considered to account for uncertainties in the
extrapolations applied0
VII. What mechanistic data would be useful for development of risk estimates of
exposures via the oral route?
-------
I Introduction
Vinyl acetate monomer is a synthetic organic ester with a wide range of uses including
application in polyvinyl acetate emulsion of latex paints and as a copolymer with ethylene in
adhesives, paper, and paper board coatings. Vinyl acetate is also used in the manufacture of
polyvinyl alcohol. Because vinyl acetate has been shown to be a respiratory tract toxicant in
experimental species and is emitted from production facilities, concern exists over its potential
adverse human health effects. The 1990 Clean Air Act Amendments include vinyl acetate as
one of the 189 hazardous air pollutants listed under Title m (42 U.S.CA §7412(b)).
Vinyl acetate is a highly flammable, colorless liquid with an acrid, ether-like sweetish odor. It is
soluble in most organic solvents and moderately soluble in water. The physical properties are
summarized in Table 1. A typical commercial sample of technical vinyl acetate has a purity of >
99.8% (w/w) and may contain trace quantities of water, acetic acid, and acetaldehyde.
Hydroquinone is typically added at 1.5 - 20 ppm to inhibit polymerization (ECETQC, 1991).
II Summary of Key Studies
A. Effects in Humans
In an occupational study by Deese and Joyner (1969), no adverse effects associated with long-
term occupational exposure were detected in a review of medical records and multiphasic
examinations of workers in three vinyl acetate production units of a chemical plant. The mean
concentration of vinyl acetate in the air of the units was 8.6 ppm based on a range of 0 to
49.3 ppm. No significant eye or throat irritation was noted below 10 ppm. There is inadequate
evidence in humans for the carcinogenicity of vinyl acetate based on epidemiology studies
(IARC, 1995).
-------
Table 1.
Physical and Chemical Properties of Vinyl Acetate
Name:
IUPAC name:
Synonyms:
Chemical Abstracts Index name:
CAS Registry No.
Formula:
Molecular weight:
Density:
Vapor pressure:
K^
Conversion factors (20°C, 760 mm Hg):
Vinyl acetate
Vinyl acetate
Acetic acid, vinyl ester
Acetic acid, ethenyl ester
1 - Acetoxy ethene
1-Acetoxyethylene
Ethanoic acid, ethenyl ester
Ethenyl acetate
Ethenyl ethanoate
Vinyl acetate monomer
Vinyl A monomer
Vinyl ethanoate
Acetic acid, ethenyl ester
108-05-4
86.09
0.932 at 20°C
108mmHgat25°C
0.73
1 ppm = 3.57mg/m3
1 mg/m3 = 0.28 ppm
B. Laboratory Bioassay Data
Non-Cancer Respiratory Tract Effects
Exposure of mice or rats for up to four weeks to concentrations greater than 150 ppm, or
500 ppm, respectively, produced clinical signs consistent with irritation of the respiratory tract
(Owen, 1979a, b). Concentrations as high as 2,000 ppm caused a decreased rate of weight
gain, eye and nose irritation, and increased numbers of lung macrophages in rats (Gage, 1970).
Rats and mice were exposed for up to three months to'vinyl acetate vapors ranging from
50 ppm to 1,000 ppm. At 1,000 ppm decreased body weight gain and increased lung weights
were accompanied by histopatholbgical changes in the lower respiratory tracts of rats and in the
entire respiratory tract of mice (Owen, 1980a,b). Inflammatory and metaplastic changes were
apparent in the respiratory epithelium of mice. The no-observable adverse effect level
(NOAEL) in these -studies was 200 ppm in rats and 5.0 ppm in mice.
-------
The effects of chronic exposure (0, 50, 200, and 600 ppm) to vinyl acetate in rats and mice
(60/sex/group) were evaluated (Bogdanffy el al, 1994). The study also included three
satellite groups of 10 of each species and sex for interim evaluations and recovery studies
(data not shown in tables). There were no exposure-related tumors observed in any satellite
group animal. Exposure-related effects in both species were confined to the respiratory
tract. There was no evidence of systemic toxicity or systemic oncogenicity.
Non-neoplastic lesions in rats. A summary of significant non-neoplastic lesions of
the respiratory tract is presented in Table 2 and Figure 1. In the respiratory epithelium of
the nasal cavity, no non-neoplastic treatment-related changes were observed. The most
prominent and consistent compound-related nasal lesion consisted of thinning of the
olfactory epithelium accompanied by basal cell hyperplasia. In severe cases, low cuboidal
cells lined the thickened submucosa in the olfactory region. In less severe cases,
proliferating basal cells were covered by epithelium that resembled respiratory epithelium.
In most animals of the 600 ppm groups, these changes were associated with submucosal
edema and with loss of nerve bundles and Bowman's glands or with hyperplasia of glandular
structures. Infiltrations of inflammatory cells in the epithelium and submucosa as well as
leukocytic exudate were seen regularly in 600 ppm-exposed animals.
In 200 ppm groups, the location of the lesions described above were often restricted to, or most
pronounced in, the anterior part of the dorsal meatus. This region is normally covered by
olfactory epithelium. In the 600 ppm concentration group the lesions extended to the posterior
part of the olfactory epithelium. Focal squamous metaplasia of olfactory epithelium (without
keratinization) was observed in many 600 ppm-exposed rats and was located mainly on the
top of the dorsal lamellae of the ethmoturbinates.
Regeneration of the olfactory epithelium was evident in many rats of the 200 ppm groups and in
a few rats of the 600 ppm group. The regenerated epithelium was seen as a layer of stratified
undifferentiated epithelium containing small foamy structures resembling nerve bundles and
groups of epithelial cells containing yellow-brown pigment resembling acinar cells of the
Bowman's glands. The regenerating epithelium was most prominent in the anterior part of the
dorsal meatus.
-------
Table 2
Summary of Statistically Significant Non-neoplastic Changes in Lungs and Nose Rats: Main Studv
Concentration (ppm):
Longs:
Bronchial exfoliation
very slight
slight
moderate
Intraluminal fibrous projections
very slight
slight
moderate
severe
Pigment macrophage
very slight
slight
moderate
Peribronchiolar/perivascular lymphoid
aggregates
very slight
slight
moderate
Nose:
Olfactory epithelial atrophy
very slight
slight
moderate
severe
Olfactory epithelial squamous metaplasia
very slight
slight
moderate
severe
Olfactory epithelial regeneration
very slight
slight
moderate
Olfactory epithelial inflammatory cell
infiltrate
very slight
slight
moderate
Epithelial nest-like infolds
very slight
slight
moderate
severe
Control
(58)
0
0
0
0
0
0
0
1
1
0
5
15
1
(59)
0
0
. 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
(59)
0
0
0
0
0
0
0
0
. 3
0
1
18 *
4
(60)
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Incidence of Lesions
Males
200 600
(60)
0
0
0
0
0
0
0
0
3
1
0*
21
1
(59)
4
47»«»
2
0
0
' 0
0
0
3
30*"
2
0
0
0
0
1
15«*«
1
(60)
8«»
26«««
2
16*"
14»«.
1
0
0
33»«.
2
0*
14
2
(59)
0
7*
33"*
10"
2
12"
9"
1
0
1
0
1
7*
1
0
0
5
5
Other than Tumors*
Control 50
(60)
0
0
0
0
0
0
0
t
0
6
0
0
11
2
(60)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
(60)
0
0
0
0
0
0
0
0
4
0
1
14
1
(60)
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
• o
0
Females
200
(60)
0
0
0
0
0
0
0
0
1
0
2
14
2
(60)
4
23"*
0
0
5
0
0
0
3
16*"
3
0
0
0
1
0
5
0
600
(59)
0 .
4
0
3
28*"
8**
1
1
10
4
0
23*
5
(59)
0
18"*
30***
3
•4
26***
7"
0
2
7**
,. 0
0
5*
1
0
0
5*
2
-------
Table 2 (continued)
Summary of Statistically Significant Non-neoplastic Changes in Lungs and Nose Rats: Main Study
Concentration (ppm):
Olfactory epithelial leukocvtic exudate
very slight
slight
moderate
severe
Basal cell hyperplasia
very slight
slight
moderate
severe
Turbinate leukocvtic exudate
very slight
slight
moderate
severe
Submucosal inflammatory cell infiltrate
slight
moderate
severe
very severe
Control
0
0
0
0
2
, 0
o
0
0
4
3
0
V
2
1
0
0
50
0
0
0
0
5
0
0
0
2
8
6
0
0
3
0
0
Incidence of Lesions
Males
200 600
0
0
0
0
3
40«.
11"*
0
0
5
3
0
1
1
0
1
0
11**»
2
1
1
2i«»«
22«»
2
0
5
8
1
2
6
1
0
Other than Tumors*
Control 50
0
0
0
0
0
0
0
0
1
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
3
1
0
0
0
0
0
Females
200
1
0
1
0
7*
24"*
3
0
3
3
1
0
0
0
0
0
600
0
5*
3
0
0
35"*
16"*
0
,
1
7
7"
0
0
5*
0
0
Figures in parenthesis represent the number of animals from which this tissue was examined microscopically. Significance of differences in a pairwise
(Fisher's) test between each treatment and control incidence is indicated by *p<0.05, **p<0.01, "*p<0.001.
In the anterior part of the nose, signs of rhinitis, such as leukocytic exudate, hyperplasia of the
epithelium covering the nasal and maxillary turbinates, and epithelial and submucosal infiltrates
of inflammatory cells were observed to about the same incidence and severity in test and control
animals.
The lack of non-neoplastic effects reported for the respiratory epithelium of rats is remarkable
considering the appearance of two inverted papillomas in this region and two squamous cell
carcinomas in the anterior (non-olfactory) region in male rats of the 600 ppm exposure group.
It is unlikely that the inverted papillomas arose spontaneously. In a retrospective study of
material in the NTP archives, Brown (1990) reported a spontaneous papilloma incidence (not
otherwise specified) of 0/1596 male rats and 0/1643 female rats. The appearance of non-
neoplastic effects in the respiratory epithelium of rats may have been masked by the pathology
-------
induced by tumor formation. Nasal respiratory epithelium of mice was reported to be affected
by vinyl acetate (see below). It is also possible that non-neoplastic effects were evident at an
earlier time during the study and were subsequently repaired. Evidence for this is as follows.
Higher exposure concentrations have been shown to affect respiratory epithelium. Exposure of
rats for up to 4 weeks show a very low incidence of respiratory epithelial damage and repair
(see Cell Proliferation Effects, below). In-vitro cytotoxicity studies discussed below show that
vinyl acetate has the potential to produce cytotoxicity in rat respiratory epithelium, but at
relatively high concentrations (Kuykendall et al., 1993). Therefore, it appears that the
respiratory epithelium is susceptible to the cytotoxic effects of vinyl acetate, but is substantially
more resistant than the olfactory epithelium. These data also suggest that the respiratory
epithelium is capable of rapid repair and adaptation.
10
-------
Incidence of Olfactory Epithelial Atrophy in Rats
100
0-
100
200
300
400
500
600
Vinyl Acetate Concentration (ppm)
Figure 1. Dose-response for olfactory degeneration (atrophy). The incidences are for males and females combined,
regardless of severity qualification. AH lesions observed at 50 ppm were considered either "very slight" or "slight".
No compound-related non-neoplastic changes were seen in larynx or trachea. Treatment-
related changes in the lower respiratory tract were restricted to male and female rats of the
highest exposure concentration and generally involved the bronchi and bronchioli. Bronchial
exfoliation of the lining epithelium was observed in many 600 ppm group males and in a few
females, without showing apparent associated acute bronchitis. In addition, intraluminal fibrosis
was observed. This lesion was characterized by fibrous plaques and buds covered by normal
bronchial epithelium that projected into the lumen of the airways. An increased incidence of
macrophages laden with brown pigment granules located in the main bronchi, in bronchioli, in
alveolar spaces, and in the interstitium was observed in the 600 ppm groups.
11
-------
In general, treatment-related nasal and lower respiratory tract lesions similar to those seen in the
main study were present in the interim and recovery groups and occurred to about the same
incidence and severity as in the main study. In the lower respiratory tract, however, bronchial
exfoliation was not observed in any of the interim or recovery group rats.
Vinyl acetate is metabolized to acetic acid and acetaldehyde. Acetaldehyde is also a nasal
toxicant (Woutersen et al, 1986). Morphologically, the non-neoplastic lesions induced by vinyl
acetate bear only slight resemblance'to those induced by acetaldehyde. The main similarities
include a high incidence of olfactory epithelial atrophy and basal cell hyperplasia of the olfactory
epithelium (Woutersen et al, 1984; 1986). Olfactory epithelial atrophy induced by
acetaldehyde appears as early as 28 days after exposure of rats to 400 ppm acetaldehyde
(Appelman et al, 1982). Squamous metaplasia was found with high incidence among rats
exposed to >750 ppm acetaldehyde (lower levels were not tested) or >200 ppm vinyl acetate.
In the case of vinyl acetate on the other hand, squamous metaplasia was not associated with
keratinization. A particularly interesting difference observed between the non-neoplastic effects
of vinyl acetate and acetaldehyde, under the conditions of their respective bioassays, is the lack
of pronounced effect on nasal respiratory epithelium in rats exposed to vinyl acetate. Squamous
metaplasia, with or without keratinization, and simple epitheliomatous hyperplasia was observed
in respiratory tissue of rats exposed to 1500 ppm acetaldehyde.
The non-neoplastic. lesions induced by vinyl acetate bear greater resemblance to lesions induced
in rodents by inhaled organic acids and esters. Examples include propylene glycol monomethyl
ether acetate, ethyl acrylate, methyl acrylate, n-butyl acrylate, formic acid, and acrylic acid
(Miller etal., 1981; Miller eta!., 1984; Miller et al., 1985b;Reininghause/a/., 1991;National
Toxicology Program, 1992). For all of these compounds the critical lesion of the nasal passages
is degeneration of the olfactory epithelium, primarily of the epithelium lining the dorsal meatus.
Respiratory epithelium is generally less sensitive.
The strongest data set linking the mechanism of vinyl acetate-induced non-neoplastic nasal
lesions to that of other inhaled esters is the work on dibasic esters. Keenan et al., (1990)
showed that 13 week exposures of rats to dibasic esters mixtures produces degeneration of only
the olfactory epithelium while Lee et al (1992) showed that high concentrations of dibasic
esters (5900 mg/m3 aerosols plus unspecified amounts of vapor) damage both respiratory and
olfactory epithelium. The pathogenic responses were similar to that of vinyl acetate in that the
lesions progressed from reduced olfactory epithelial thickness and "degeneration to a reparative
state of hyperplasia and/or metaplasia with prominent basal cell mitotic activity. In the case of
12
-------
dibasic esters, the carboxylesterase-rich sustentacular cell was shown at the ultrastructural level
to be the primary target of cytotoxicity (Trela et al., 1992).
'Non-neoplastic lesions in mice. A summary of the statistically significant non-neoplastic
lesions in mice is presented in Table 3. In general, the morphology of the non-neoplastic lesions
observed in the nasal cavity of mice was similar to that of rats, however, several specific
differences were noted. In mice, some atrophic areas of the olfactory epithelium were
accompanied by foci of respiratory epithelium (respiratory metaplasia). This type of respiratory
epithelial metaplasia occurred locally both at the dorsal meatus in the mid-region and at the
dorsal parts of the nasal cavity in the ethmoturbinate region. The ciliated cells often appeared to
be continuous with the ciliated lining epithelium of the ducts of the underlying Bowman's glands.
In rats, areas of regeneration of the olfactory epithelium were often accompanied by a
keratinizing squamous epithelium and epithelial nest-like infolds.
Another notable difference between rats and mice was the appearance of non-neoplastic lesions
in the respiratory epithelium of mice. Focal non-keratinizing squamous metaplasia of
respiratory epithelium of the maxilloturbinates and lateral wall of the nasal cavity at the
naso/maxilloturbinate region, and occasionally of olfactory epithelium at the dorsal meatus, was
observed in the 600 ppm mice. Also, eosinophilic hypertrophic sustentacular cells along with
local loss of sensory cells was observed in all gVoups, including controls, but occurred more
frequently in mice of the 200 ppm and 600 ppm groups.
13
-------
TableS
Summary of Statistically Significant Non-neoplastic Changes in Lungs and Nose of Mice: Main Study
Concentration (ppm):
Lungs:
Accumulation of alveolar macrophages
very slight
slight
moderate
severe
Intra-alveolar eosinophilic material
very slight
slight
moderate
severe
Accumulation of brown pigmented
macrophages
very slight
slight
moderate
Intraluminal fibroeptthelial projections
very slight
slight
moderate
Bronchial gland dilatation
Bronchial/bronchiolar eprthelial flattening
and'or exfoliation
very slight
slight
moderate
severe
Bronchial/bronchiolar eprthelial
disorganization
very slight
slight
moderate
Nose:
Inflammatory exudate
Mucosal inflammatory infiltrate
Submucosal gland hyperplasia
slight
moderate
Olfactory eprthelial atrophy- (mainly dorsal
mearus)
very slight
slight
moderate
severe
Incidence of Lesions Other than Tumors*
Control
(51)
5
10
0
1
0
3
0
0
2
0
0
0
0
0
14
0
1
0
0
0
0
0
(52)
0
1
3
0
0
0
0
0
50
(51)
1
2*
4
1
0
1
0
0
2
0
0
1
0
0
16
0
0
0
0
0
0
0
(48)
0
0
3
0
0
0
0
0
Males
200
(56)
4
4
8"
4
3
1
0
0
1
5
1
2
0
0
26
0
0 -
0
0
0
, 0
0
(53)
2
0
28***
8"
2
5
28*"
4
600
(S3)
3
7
4
0
1
19*"
10"
2
11*
12*"
1
3
17*"
3
17
4
25*"
7*
0
0
11"
4
(50)
15*"
12" •
25*"
15*"
0
0
2
3
Control
(56)
• 5
3
2
1
0
0
0
0
3
1
0
1
0
0
8
0
0
0
0
0
0
0
(56)
0
1
2
0
0
2
0
0
50
(55)
2
8
1
3
0
0
0
0
5
1
0
0
2
0
17
0
0
0
0
0
1
0
(57)
0
2
5 -
0
0
4
0
0
Females
200
(55)
6
4
1
1
2
0
1
0
1
4
0
0
1
0
20*
i
0
0
0
0
'
0
0
0
(55)
1
0
42"*
7"
0
8
26***
1 4
600
(51)
1
10
12"
1
1
7**
15***
i
2
, •:.}**•
2
6
19"*
• 7"
15
4*
28***
4*
1
5'
18*"
0
(51)
5"
5
35"*
13"*
0
0
0
1
14
-------
'Table 3 (continued)
Summary of Statistically Significant Non-neoplastic Changes in Lungs and Nose of Mice: Main Study
Concentration (ppm):
Olfactory epithelial atrophy (widespread)
'
slight
moderate
severe
Squamous metaplasia at the
naso/maxillouirbinate region
very slight
slight
moderate
severe
Replacement of olfactory by respiratory
epithelium
slight
moderate
severe
Trachea/bronchi:
Epithelial hyperplasia
Incidence of Lesions Other than Tumors* ;
Control
0
1
0
0
1
0
0
0
0
0
(49)
0
50
0
0
0
0
1
1
0
0
0
0
(46)
0
Males
200
1
8»
4
0
2
0
0
5
1
0
(51)
2
600
0
5
39...
0
13"
H»*
0
11*"
0
0
(48)
19***
Control
0
0
0
0
4
0
0
0
0
1
(55)
1
50
0
0
0
0
2
0
0
0
1
0
(56)
1
Females
200
0
12«*
2
0
0
0
0
15"*
5*
0
(52)
0
600
0
5*
45"*
1
13*
6"
1
10*"
10*"
0
(48)
11*"
Figure in parenthesis represent the number of animals from which this tissue was examined microscopically. Significance of differences in a pairwise
(Fisher's) test between each treatment and control incidence is represented by *p<0.05, **p<0.01, ***p<0.001.
In mice (53 and 83 week interim sacrifices), the pattern of lesions was similar to that of the main
study although the incidence and severity of the lesions was not as great at the terminal
sacrifice.
Tracheal epithelial hyperplasia was significantly increased in incidence in 600 ppm mice of the
main study. A few mice of the 600 ppm group showed trachea! epithelial flattening and/or
exfoliation, metaplasia, or intraluminal fibroepithelial projections, similar to the treatment-related
changes observed in the intrapulmonary conducting airways.
As with the rats, the treatment-related changes of the conducting airways of mice occurred only
in the 600 ppm group. There was flattening and/or exfoliation of the bronchial and bronchiolar
lining epithelium, without obvious evidence of an associated inflammatory response. Moreover,
intraluminal fibroepithelial projections, seen as finger-like projections,
15
-------
plaques, and buds, protruding into the lumen of the bronchi and bronchioli were observed. The
projections were lined by flattened epithelium and they incorporated a stromal component.
Epithelial disorganization of the bronchial and bronchiolar epithelium was defined as the
presence of foci or areas of dedifferentiated lining epithelium, seen as a pleomorphic picture of
swollen basopilic epithelial cells showing pronounced nucleoli, together with relatively flattened
or cuboidal epithelial foci, suggesting regeneration, and occasionally multilayered or hyperplastic
foci. These changes were grouped since they generally occurred together in the same area and
appeared to be stages of a process of continuous degeneration and regeneration. Focal
metaplasia of the bronchi/bronchioli was occasionally observed. A single male of the 600 ppm
group showed a small area in the alveolar tissue with comifying squamous metaplasia.
In the alveoli there were compound-related accumulations of alveolar (foamy) macrophages in
600 ppm females. Accumulation of brown pigmented macrophages was observed in the
600 ppm group in both sexes and perhaps also in males of the 200 ppm group. Moreover, there
was intra-alveolar accumulation of eosinophilic material. Occasionally this material was taken
up by macrophages, resulting in an eosinophilic appearance of the macrophages. Although no
statistically significant increases in non-neoplastic lesions were noted in the larynx of mice, one
female in the 600 ppm group showed a focus of squamous epithelial hyperplasia with dysplastic
changes. Several other females showed epithelial hyperplasia.
In general, treatment-related changes occurred to about the same degree and severity in mice of
the interim sacrifice groups as occurred in those of the main study. Among mice of the
recovery groups the results were also similar to the main study except that the severity of the
lesions appeared to be slightly lower. Squamous metaplasia in the naso/maxilloturbinate region
and compound-related inflammatory exudate in the nasal cavity were not observed in the
recovery groups of mice.
Other Non-neoplastic endpoints
Irvine (1980) exposed rats (24 mated females/concentration) to 0, 52, 198, or 1004 ppm
to VA on gestation days 6 through 15 for 6 hr/day. Toxicity in the dams exposed to the highest
concentration only was noted as a decrease in body weight gain of 10-12% from day 10 of
gestation to the end. Fetal growth retardation (decreased mean litter weight, mean fetal weight,
and crown/rump length; increase in retardation of sternebral and occipital ossification) also
occurred only at the highest concentration.
16
-------
An oral (drinking water at 0, 200, 1000, and 5000 ppm) 2-generation study in rats (Shaw,
' 1987) indicated a marginal effect (not statistically significant) on reproductive performance in
males dosed at 5000 ppm. No effects in offspring parameters were noted at any concentration.
Neoplastic Effects
Neoplastic lesions in Rats. A summary of the significant neoplastic lesions of the
respiratory tract is presented in Table 4 and Figure 2 (Bogdanffy et al, 1994). A total of twelve
tumors of the nasal cavity were observed. Four of them were classified as benign inverted,
endophytic papillomas and were found in the 600 ppm males only; one was classified as a
benign exophytic papilloma and was found in one male in the 200 ppm concentration group.
The papillomas were characterized by pseudoacinar structures with cuboidal to columnar
epithelium and, in some cases, multilayered epithelium with atypical cells and flattening in some
areas.
Incidence of Nasal Tumors in Rats
•a-
10-
§ 8-1
2-
0-
11/118
1/119
.100 200 300 ' 400 500
Vinyl Acetate Concentration (ppm)
600
Figure 2. Dose-response for tumor incidence. All tumors included The one tumor observed at 200 ppm was a
benign exophytic papilloma.
17
-------
The tumors were seen in various regions of the nose (Table 5). One was seen in the respiratory
area on the lateral wall in zone C, one in the respiratory area of the lateral meatus in zone A, and
three in the ethmoturbinate olfactory region in zone E. The other tumors all were observed in
the 600 ppm group and were classified as malignant carcinoma in situ (one case) or squamous
cell carcinomas with varying degrees of keratinization (six cases). The carcinoma in situ was
found in a male rat in the olfactory area in zones C and D. One small squamous cell carcinoma
was found in another male rat in zone B in the ventral floor normally covered by cuboidal cells.
In one male rat a squamous cell carcinoma was noticed in the maxilloturbinate area in zone B.
One small squamous cell carcinoma occurred in the ethmoturbinate olfactory area (zone E) in a
female rat. Three large squamous cell carcinomas were observed in female 600 ppm rats.
These tumors obstructed one side of the nasal cavity and were characterized by invasive growth
in nasal bones, soft tissues, and the maxillary sinus. The origin of these tumors could not be
established. A subsequent review of these three tumors indicated that one of these might have
arisen from respiratory regions, but its size and invasiveness prevented a definitive assignment
(KT Morgan, personal communication).
In the larynx a squamous cell carcinoma was found in one female rat of the 600 ppm group. No
tracheal or treatment-related lung tumors were found in the terminal sacrifice animals nor were
any neoplasms observed in the 53 week, 83 week, or recovery groups.
Neoplastic lesions in mice. No treatment-related tumors were observed in the nose,
larynx, trachea, or other tissue besides lung of mice of the main study group nor in any airway
tissue or other tissue of mice in the satellite groups. A single moderately invasive squamous cell
carcinoma was found in a major bronchus of the lung of a male of the 600 ppm group while a
single adenocarcinoma occurred in the lung of a male of the control group. These were not
statistically significant.
18
-------
Table 4
Summary of Statistically Significant Neoplastic Changes in Lungs and Nose of Rats:
Main Study
Concentration (ppm):
Lungs:
Well differentiated adenoma [B]
Nose:
Inverted papilloma [B]
Squamous cell carcinoma [M]
Papilloma [B]
Carcinoma in situ [M]
Total benign tumors
Total malignant tumors
Total nasal tumors
Larynx:
Squamous cell carcinoma [M]
Incidence of Tumors (Numeric)1
Control
(58)
0
(59)
0
0
0
0
0
0
0
(59)
0
Males
50
(59)
0
(60)
0
0
0
0
0
0
0
(60)
0
200
(60)
0
(59)
0
0
1
0
1
0
1
(60)
0
600
(60)
0
(59)
4
2
0
1
4
3
7**
(60)
0
Control
(60)
1
(60)
0
0
0
0
0
0
0
X
(60)
0
Females
50
(60)
0 '
(60)
0
0
0
0
0
0
0
(60)
0
200
(60)
0
(60)
0
0
0
0
0
0
0
(60)
0
600
(59)
0
(59)
0
-4
0
0
0
.4
4
(59)
1
Figures in parenthesis tepieseiit the number of animals from which this '"""* was examined microscopically. [B] — benign, [M] = malignant Significance of
differences in a pairwise (Fisher's) test between each treamient and control incidence is represented by **(p<0.01).
19
-------
TableS
Regional Distribution of Vinyl Acetate-Induced Nasal Tumors in Rats'
Site of Origin Tumor Type
Olfactory region 2 inverted papillomas
1 exophytic papilloma
1 squamous cell carcinoma .
1 carcinoma in situ
Respiratory region 2 inverted papillomas
Cuboidal epithelium of zone B* 2 squamous cell carcinoma
Unknown 3 squamous cell carcinoma
' The histopathologicaJ evaluation of the nasal cavity included four cross-sections as follows: slightly
posterior to the upper incisor tooth at the level of the first palatine fold, at the level of the incisive papilla, and slightly
anterior to the first molar. These cross-sections are approximately equivalent to those described by Mery et aL (1994)
as follows: section 5,9, 11, and 19, respectively. The nasal lesions were recorded for five anatomical zones: zone A,
maxilJoturbinates: zone B. nasolachrimal duct medial/ventral to incisor tooth root; zone C, palatine fold region, with
anterior region of olfactory epithelium; zone D, maxillary sinus region, anterior to ostiunx and zone £, ethmoturbinate
region. ,
* Zone B is located within a cross section of the nose at the level of the nasolachrimal duct medial/ventral to
the incisor tooth root.
20
-------
Comparison of Tumorigenic Responses of Vinyl Acetate and Acetaldehyde
The tumorigenic response of rats to vinyl acetate and acetaldehyde are similar only in that both
produce squamous cell carcinomas. The squamous cell carcinomas arise mainly from
respiratory regions in rats exposed to high concentrations of acetaldehyde and from both
respiratory and olfactory regions in rats exposed to vinyl acetate (Table 6). These two
compounds differ in that acetaldehyde also produces adenocarcinoma, with a greater incidence
than squamous cell carcinoma, and arising primarily from olfactory regions. It is not entirely
surprising that the tumorigenic response produced by the two compounds are not identical.
First, acetaldehyde was tested at higher concentrations than vinyl acetate increasing the
likelihood that acetaldehyde detoxication pathways (in terms of removal of genotoxic species),
such as aldehyde dehydrogenase, are saturated. Second, acetaldehyde differs in its water
solubility and reactivity which is likely to affect differences in flow dynamics. Third,
acetaldehyde demonstrates genotoxic activity (e.g., clastogenesis) without metabolic activation
and, therefore, the active species is available to all nasal mucosal cells. Vinyl acetate on the
other hand requires metabolic activation to elicit clastogenic activity and was tested at lower
exposure concentrations where aldehyde dehydrogenase activity is less likely to be saturated.
Carboxylesterase liberates acetaldehyde and acetic acid from vinyl acetate. However,
carboxylesterase is localized only within specific cell types (Bogdanffy et al, 1987). Studies on
the mechanism of action presented below show that acetic acid is responsible for the cytotoxic
effects of vinyl acetate. Thus, replicating epithelial and seromucous gland cells of respiratory
mucosa and sustentacular cells (these may not be capable of replication), and Bowman's gland
cells of olfactory mucosa are apt to be primary sites of metabolic activation of vinyl acetate and
targets for toxicity. It is only under high concentration exposure scenarios, when acetaldehyde
detoxication capacity is overwhelmed, and subsequent cellular proliferation is induced that a role
for acetaldehyde in vinyl acetate carcinogenesis might be expected. Nevertheless, the cellular
targets of the two compounds would be expected to differ.
21
-------
Table 6
Vinyl Acetate vs. Acetaldehyde
Comparison of Carcinogenic Activitieb
Species tested
Species with positive response
No observed adverse effect level
(NO AEL) for tumors
Lowest positive exposure
concentration (LOEL)
- Response at LOEL1^
Types of rat nasal tumors'
Adenocarcinoma
Papilloma
Squamous cell carcinoma
Carcinoma in situ
Time of earliest tumor
Vinyl Acetate
Rat, mouse
Rat
200 ppm
600 ppm (rat)b
9% (11/1 19)
0 50 200 600
n= 119 120 119 118
-
1 4
- 6
1
24 Months (Week 1 03; benign
exophytic papilloma in a 200 ppm
male). All other tumors observed at
terminal sacrifice (weeks 106-107)
Acetaldehvde"
Rat, hamster
Rat, hamster
not determined
750 ppm (rat)
24% (24/98)
0 750 1500 3000/1000
99 100 106 102
- , 22 57 42
1 - -
1 1 . 15 32
3 6
<. 1 2 Months (malignant
adenocarcinoma in a 750 ppm
male). Squamous cell carcinomas
and carcinoma in situ were also
observed at < 12 months of
exposure.
' Dau from Woutersen et al, 1986
" One benign exophytic papilloma was observed in a 200 ppm male
' Males and females combined after two years of exposure
" Benign and malignant tumors combined
22
-------
III. Mode of Action and Determinants of Response
This section provides additional data that provide a basis on which to propose a mode of
action.
Cytotoxic Activity
The proposed mode of action was investigated in a series of in vitro experiments
(Kuykendall et al, 1993). The first hypothesis tested is that vinyl acetate-induced
cytotoxicity in nasal tissues is a carboxylesterase-dependent process. The second hypothesis
is that either acetic acid, acetaldehyde, or both are the primary cytotoxic metabolites. To
test these hypotheses, an in vitro assay for nasal tissue cytotoxicity was utilized. This assay
has been useful in previous studies aimed at elucidating the mechanism of toxic action of
dibasic esters (Trek and Bogdanffy, 1991a,b). The assay is based on measurement of
cytoxicant-induced release into the incubation medium of the intracellular enzyme acid
phosphatase. Previous research has shown a tight correlation between acid phosphatase and
early ultrastructural change in sustentacular cells (Trela et al., 1992).
To study the role of carboxylesterase in the cytotoxic effects of vinyl acetate, rats were
pretreated with a nonspecific esterase inhibitor, bis(p-nitrophenyl) phosphate (BNPP)
(Heymann and Krisch, 1967). BNPP pretreatment for three days prior to tissue collection had
no cytotoxic effect on maxilloturbinate (lined with respiratory epithelium) or endoturbinate-1
(lined with olfactory epithelium) tissues (Table 7). Vinyl acetate (50 mM) induced an
approximately 3- to 4-fold increase in acid phosphatase release from both turbinate types.
Pretreatment with BNPP attenuated the vinyl acetate-induced cytotoxic response. Following
BNPP pretreatment, 50 mM vinyl acetate induced only an approximate 2-fold increase in acid
phosphatase release. Vinyl acetate treatments caused a reduction in media pH as made obvious
by the visible, time- and concentration-dependent, change in the color of the media pH
indicator.
Because BNPP pretreatment attenuated the cytotoxic response in both turbinate types, it was of
interest to determine if BNPP pretreatment also inhibited metabolism of vinyl acetate.
Administration of BNPP inhibited the release of acetaldehyde into the media approximately
59% or 37% in maxilloturbinate and endoturbinate-1 tissues, respectively (Table 8). The
hydrolysis of vinyl acetate produces acetaldehyde and acetic acid. To assess the role of .
acetaldehyde in the cytotoxic effects of vinyl acetate, turbinates were incubated for 1 hr in
media alone or in media containing semicarbazide, an aldehyde scavenger. Semicarbazide alone
23
-------
was slightly, but not significantly, cytotoxic to both turbinate types (Table 9). Inclusion of
semicarbazide in the incubation media offered no protection from vinyl acetate-induced
cytotoxicity to either maxilloturbinate or endoturbinate-L tissues.
Table 7
Effect of BNPP Pretreatment on Vinyl Acetate-Induced Cytotoxicity:
Release of Acid Phosphatase into Media*
Pretreatment:
In vitro Treatment:
Maxilloturbinate
Endoturbinate-l
Saline
Control
7.4±0.4b
6.8 ±0.7
Saline
Vinyl Acetate
27.6±3.5C
21.3 ±1.9"
BNPP
Control
8.0 ±0.9
6.3 ±1.5
BNPP
Vinyl Acetate
17.2±1.3*e
15.8±1.74e
Rats were pretrealed with saline, a 5% (maxillotiirbinate), or a 10% (endoturbinate-1) suspension of BNPP in saline (refer to
Kuykendall el al, 1993). Nasal explants were incubated for 1 hr in media containing 50 mM vinyl anflatf and media was assayed for acid phosphatase
activity.
b Values are expressed as mean percentage of acid phosphatase release (n=4) ± S.E.M.
' Statistically different from saline, control (p< 0.05).
" Statistically different from BNTP. control (p< 0.05).
' Statistically different from saline vuiyl amali* treatment (p < 0.05).
Table 8
Effect of BNPP Pretreatment on Vinyl Acetate Metabolism:
Acetaldehvde Release into Media'
Pretreatment: Saline BNPP.
MaxilJoturbuiaie 25.7±4.2b 10.6±1.6C
Endoturbinate-l - • 34.9±3.0 22.1±0.8C
1 Rats were pi eu mini uuri saline, a 5% (maxilloturbinate), or a 10% (endoturbinate-1)
suspension of BNTP in saline < tvlcr to Ku\i;endall el al. 1993). Nasal explants were incubated for 20
min in WN1E containing 50 mM vuiyl aceta^ and media was assayed for acetaldehyde.
6 Values are expressed as mM acetaldehyde; mean i S.E.M (n=4).
' Statistically different from saline control (f 0.05).
24
-------
Table 9
Effect of Semicarbazide on Vinyl Acetate-Induced Cytotoxkrty:
Release of Acid Phosphatase into Media*
In vitro Treatment:
Maxilloturbinate
Endoturbinate-1
Control
5.8±l.h
6.1 ±1.7
Vinvl Acetate
17.9±2.8c
18.3±3.5C
Semicarbazide
9.6 ±1.2
9.9 ±2.6
Semicarbazide ±
Vinvl Acetate
19.0 ±2.2
29.3 ± 9.0d
"Turbinates were isolated from untreated rats (refer to Kuykendall et al. 1993). Nasal explains were incubated for 1 hr in media containing SO mM vinyl
acetate and media was assayed for acid phosphatase activity.
" Values are expressed as mean percentage of acid phosphatase release (n=4) ± S.E.M
' Statistically different from control (p< 0.05).
4 Statistically different from Semicarbazide control (p< 0.05).
To the contrary, a slight but not statistically significant increase in acid phosphatase release
was noted in endoturbinate-1 tissues incubated with vinyl acetate and semicarbazide relative
to those incubated with vinyl acetate alone. To test the cytotoxic potential of acetaldehyde
and acetic acid in nasal turbinates, maxilloturbinate and endoturbinate-1 tissues were
incubated for 1 hour in media with or without 50 mM acetaldehyde (Table 10) or acetic acid
(Table 11). Acetic acid, but not acetaldehyde, was cytotoxie at this concentration., These
studies demonstrate that vinyl acetate is cytotoxic to nasal turbinates, that the
carboxylesterase-mediated metabolism of vinyl acetate is necessary for cytotoxicity, and that
acetic acid, not acetaldehyde is the principal cytotoxic metabolite.
Genotoxic Activity
In vivo mutagenesis studies with vinyl acetate have been, in general, negative especially
when tested at nonlethal levels by the inhalation route (Table 12). Induction of erythrocyte
micronuclei has been demonstrated only when tested by the oral and i.p. routes at lethal
levels. In vitro mutagenesis assays with prokaryotes are also generally negative. However,
vinyl acetate has been reported to be mutagenic or clastogenic with or without an exogenous
source of enzymatic metabolism in a number of in vitro assays employing cultured human
lymphocytes, mouse L5178Y lymphoma cells, or Chinese hamster ovary cells
(Jantunen et al., 1986; He and Lambert, 1985; Kirby, 1983; Maki- Paakkanen and Norppa,
1987). These studies suggested a clastogenic effect possibly similar to that induced by
acetaldehyde. Alkaline elution studies with human leukocytes show some DNA-crosslinking
activity (He and Lambert, 1985). Oral administration of [vinyl-U-14C] vinyl acetate showed
25
-------
the association of radioactivity with hepatic nucleic acid and nuclear proteins but no specific
adducts could be identified (Simon et al, 1985b).
Role of Reduced IntraceUular pH
The clastogenic efifects observed in in vitro studies with mammalian cells could be, at least in
part, the result of reduced intracellular pH that results from the liberation of acetic acid from
vinyl acetate. -Morita (1995) has shown that low pH (pH 6.6) leads to chromosomal
aberrations and sister chromatid exchanges in Chinese hamster ovary cells and that these effects
are S- phase dependent. Thus, cells in a highly replicating population might be extra sensitive to
low pH-induced clastogenesis. Neutralization of the media abolished the clastogenic activity
(Morita et al., 1990). These observations supported the work of Sipi et al. (1992) who showed
that addition of organic acid metabolites of a variety of vinyl esters to the culture media reduced
the pH of the media 0.5-1.0 units and facilitated induction of sister chromatid exchanges in
whole blood human lymphocytes. However, Sipi et al: (1992) who studied vinyl acetate
specifically also noted that vinyl acetate-induced sister chromatid exchange results could not be
explained solely by the acetic acid-induced reduction in media pH. Since the studies by Morita,
Sipi, and colleagues relied on measures of extracellular (media) pH, it can be inferred that the
effect on clastogenic activity from reductions in,intracellular pH would likely be at least as
pronounced, if not more pronounced. It is readily envisaged that vinyl acetate could cross the
plasma membrane and be hydrolyzed intracellularly leading to the trapping and intracellular build
up of acetic acid. Acetic acid would be extensively ionized under physiological conditions.
26
-------
Table 10
Effect of Acetaldehyde on Nasal Explant Cytotoxicity:
Release of Acid Phosphatase into Media*
In vitro Treatment:
Maxilloturbinate
Endoturbinate-1
Control
8.6 ± 2.1"
9.1 ±1.7
Acetaldehyde
8.3 ± 0.7
8.7 ± 1.4
' Turbinates were isolated from untreated rats (refer to Kuykendall el aL, 1993). Nasal
explains were incubated for 1 hour in media containing 50 mM acetaldehyde and media was assayed
for acid phosphatase activity.
b Values are expressed as mean percentage of acid phosphatase release (n=4) + S.E.M.
Table 11
Cytotoxic Effects of Acetic Acid on Nasal Explant
Release of Acid Phosphatase into Media*
In vitro Treatment:
MaxiUoturbinate
Endoturbinale- 1
Control
6.1 ± 0.7*
7.4 ± 1.1
Acetic Acid
24.0 ± 1.3
25.9 ± 1.8
"Turbinauts were isolated trorr. untreated rats (refer to Kuykendall etaL, 1993). Nasal explains were
incubated for 1 hour IT. meOia containing SO mM acetaldehyde and media was assayed for acid
phosphatase activir.
' Values are enfrcsaeC ti mctr. percentage of acid phosphatase release (n=4) + S.E.M.
c Suiiai-ali. OiS.Trre from control (p< 0.05).
27
-------
Table 12
Summary of the Genotoxitity Data on Vinyl Acetate*
Test System
SOS Chromotest, E. Coli, PQ37
DNA-protein crosslink. E coliHB 101pUC13, filter binding
S. typhimurium TA100, reverse mutation
5. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1537, reverse mutation
S. typhimurium TA1537, reverse mutation
S. typhimurium TA1537, reverse mutation
' S. typhimurium TA1538, reverse mutation
S. typhimurium TAJ 530, reverse mutation
S. typhimurium TA1 530, reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA98. reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA98, reverse mutation
DNA-protein crosslink, nasal epithelial cells, in vitro
DNA crosslink, alkaline elution, purified human lymphocytes.
in vitro
Cell transformation SA7/Syrian hamster embryo cells
Gene mutations, mouse tymphoma L5 178Y cells
Sister chromatid exchange, Chinese hamster ovary cells, in vitro
Sister chromatid exchange, human lymphocytes, in vitro
Sister chromatid exchange, human isolated lymphocytes in vitro
Sister chromatid exchange, human lymphocytes in vitro
Micronucleus test, human lymphocytes, in vitro
Chromosomal aberrations, human lymphocytes, in vitro
Chromosomal aberrations, human lymphocytes, in vitro
Sister chromatid exchange, mice cells, in vivo
Micronucleus test, mouse bone marrow, in vivo
Micronucleus test, mouse bone marrow, in vivo
Micronucleus test, rat bone marrow, in vivo
Meiotic micronucleus test, mice, in vivo
DNA binding, rat hepatccytes. in vivo ("C-label)
DNA binding, rat hepatocyifis, in vivo ("C-label)
Sperm morphology, Fl mice, in vivo
Resole
Without With
exogenous exogenous
metabolic metabolic
' activation activation
.
-*•
-
-
-
-
-
-
-
-
- '
-
-
-
-
-
-
-
-
-
-
+ 0
+ 0
+ 0
+ 0
+ -f
+ 0
+ 0
+ 0
+ 0
-H 0
+ 0
•<-
-*•
-
-
-
-
-
-*-
Dose"
(LED/HID)
8.6mg/ml
1-100 mM
1,000 ug/pl
10,000 u/pl
500u/pl
30 umol/pl
>4xl03M/pl
Vapour
1000 ug/pl
10,000 ug/pl
30 nmol/pl
1000 ug/pl
10,000 ug/pl
30 umol/pl
1000 ug/pl
> 4 x Iff M/pl
Vapour
1000 ug/pl
10.000 fig/'pl
500ug
-------
acetaldehyde is not mutagenic in the Ames test (Rosenkranz, 1978; Sasaki and Endo, 1978), but
shows activity in a DNA-repair test inKcoli (Rosenkranz, 1978), is mutagenic in the mouse
lymphoma test (Wangenheim and Bolcsfoldi, 1988) and induces chromosomal aberrations
and sister chromatid exchanges in mammalian cells (Bird et al., 1982; He and Lambert,
1985; Obe and Ristow, 1977). Unlike vinyl acetate, acetaldehyde has been reported to
induce gene mutations at the HPRT locus in human lymphocytes (He and Lambert, 1990).
Acetaldehyde was weakly mutagenic in this system as concentrations ranging from 200 urn
to 2400 /zm were necessary to produce a positive response. Sequencing of the HPRT gene in
the mutant clones suggested that the majority of mutations were partial deletions or
rearrangements. These results are consistent with the conclusion that acetaldehyde is
clastogenic and that DNA-protein crosslinks (DPXL) are the putative genotoxic lesion. The
possibility that acetaldehyde is the active genotoxic metabolite of vinyl acetate is supported by
the studies of He and Lambert (1985), which showed a striking similarity in the time- and
concentration-dependent effects of the SCE-frequency with the two substances. The authors
assumed that ester hydrolysis of vinyl acetate occurs within the cell, because the addition of
purified carboxylesterases to the extracellular environment had no effect on the SCE-frequency
caused by vinyl acetate in vitro.
Role of Acetaldehyde
To further study the possibility that the genotoxic activity of vinyl acetate can be ascribed to the
release of acetaldehyde following carboxylesterase-mediated activation, Kuykendall et al.
(1993) evaluated the effect of carboxylesterase inhibition by BNPP on the induction of DPXL in
rat nasal epithelial cells. Preincubation of cells with 0.1, 0.5, and 1 mM BNPP inhibited vinyl
acetate-induced crosslink formation in a dose-dependent manner in epithelial cells from both
respiratory and olfactory tissues. There was a 76% and 78% reduction of 25 mM vinyl acetate-
induced crosslink formation by 1 mM BNPP in treated respiratory and olfactory cells,
respectively, confirming the dependence of crosslink formation on carboxylesterase-mediated
hydrolysis.
29
-------
Table 13
Summary of Factors to Consider in Evaluating the
Role of Acetaldehyde in Vinyl Acetate Carcinogenesis
+ Role of acetaldehyde in the mechanism of vinyl acetate clastogenic action is supported by a similar clastogenic
pattern of acetaldehyde.
+ Acetaldehyde and vinyl acetate have similar time- and concentration-dependent effects on SCE frequency.
-i- Work of Kuykendall et al. (1993) shows that DNA protein crosslinks (DPXL) in nasal cells are inhibited by
BNPP pretreatment suggesting that vinyl acetate-induced DPXLs are related to intracellular formation of
acetaldehyde.
In vitro experiments by Kuykendall et al. show that acetaldehyde-DPXL are unstable (fM = 6.5 hours)
Lam and Heck data show sigruficant increase in acetaldehyde-DPXL at 1000 ppm (no significant DPXL at 100
or 303 ppm)
Rigorous search for adducts b\ Hemminki and Suni (1984) show only reversible Schiflfs base reactions
Conclusion: Role of meuhobcalK -liberated acetaldehyde becomes significant only at high vinyl acetate exposure
concentrations where DPXLs contribute; to cytotoxicity, and possibly some mutagenicity that is not clearly supported
by data, but only under condiuons of induced cell proliferation.
Lam et al. (1986) studied the effects of inhalation exposure of rats to acetaldehyde on the
induction of DPXL In this study, there was no significant increase in the amount of DNA
crosslinked to protein a; L 303 ppm acetaldehyde in either respiratory or olfactory mucosa
following a single 6 hr exposure or 5 days of exposure. DPXL levels were significant in
respiratory mucosa after a single 6 hour exposure to 1000 ppm acetaldehyde. The DPXL levels
in respiratory mucosa v. ere similar after 5 days of exposure. DPXL levels were significant in
olfactory mucosa only after 5 days of repeated exposure to 1000 ppm. Thus, the respiratory
epithelium appears to be considerably more resistant to the cytotoxic and other effects of DPXL
than the olfactory mucosa (Lam et al (1986), Appelman et al, 1982; Woutersen et al, 1984).
The comparison further suggests that cytotoxicity, rather than DPXL levels, may be the primary
determinant in the mechanism of action of acetaldehyde with some secondary contribution to
genetic damage (in particular, clastogenicity).
The efficiency of formation of the acetaldehyde-DNA crosslink, and its stability was compared
30
-------
to that of formaldehyde. Using an in vitro system composed of plasmid DNA and calf thymus
histone as a model, formaldehyde-induced DPXL were formed with an efficiency 14 times
greater than that of acetaldehyde (Kuykendall and Bogdanfly, 1992a). Furthermore, the
acetaldehyde crosslinks were unstable under physiological conditions decaying with a half-life of
approximately 6.5 hours (Kuykendall and Bogdanfiy, 1992b). That is, approximately 5 x 10"4 of
the original amount of crosslinks formed would be present at the time of replication.
Interestingly, low pH facilitated the acetaldehyde-induced crosslinks which is probably a
consequence of a tighter association of the histone proteins with DNA (Kuykendall and
Bogdanfiy, 1992b). Free amino groups on histones would be more extensively ionized at low
pH facilitating their association with negatively charged phosphate groups on DNA. These
observations support those of Morita (1995) discussed above which show increased clastogenic
activity in Chinese hamster ovary cells at low pH.
In general, the results of functional assays for mutagenic activity of acetaldehyde are not
consistent with the induction of point mutation, which might be expected from some type of
DNA adduct of acetaldehyde, but suggest clastogenicity related to DNA-DNA, and DPXL
(Dellarco, 1988). A recent report showed the induction of 6-thioguanine resistance in normal
human fibroblasts following 5 hr cultures in the presence of 5 mM acetaldehyde
(Grafstrom et al, 1994). However, neither 1, 2.5, 8, nor 10 mM acetaldehyde produced this
response under the same conditions. There are only two reports in the literature investigating the
potential for acetaldehyde to form DNA adducts in vitro. The first work was conducted by
Hemminki and Suni (1984) in which nucleosides were incubated for 20 hours in pure
acetaldehyde. Acetaldehyde was found to bind and form an unstable, reversible reaction
product, principally with guanosine, which is believed to be a Schiffbase condensation product
on the N2 exocyclic amino group. This product was stabilized by addition of sodium borohydride
and the reduction product was identified as N2-ethylguanosine. These results have been
confirmed by Vaca et al. (1995) who demonstrated the low rate of reactivity and stability of the
acetaldehyde adducts and that only the reduced form of the adduct is somewhat stable under
reducing conditions. More recently, Fang and Vaca (1997) presented evidence of low levels
(approximately 2-3 adducts per 107 nucleotides) of N2-ethyl-3 '-deoxyguanosine monophosphate
in peripheral blood cells (granulocytes and lymphocytes) of alcohol-intoxicated humans.
Formation of Schiffbase intermediates with primary amines is a well characterized reaction of
aldehydes and is believed to be the first step in DPXL formation (Feldman, 1979; Ohba et al.,
1979). In vitro studies have shown that the first reaction product of acetaldehyde, in the
31
-------
sequence of steps leading to DPXL formation, is not with DNA but with amino acid residues,
principally e-amino groups of lysine, which is then followed by condensation with free amino
groups of DNA, principally guanine (Kuykendall and Bogdanffy, 1992b, 1994).
These data suggest that the role of acetaldehyde in vinyl acetate carcinogenesis becomes
significant only at high concentrations where DPXLs contribute to cytotoxicity and possibly a
clastogenic effect, and only under conditions of induced cell proliferation.
Role of Epoxide Intermediates
Theoretically, oxidation of vinyl acetate by mono-oxygenases could lead to acetoxy oxirane, the
epoxide of vinyl acetate, a substance which is mutagenic in the Ames test without metabolic
activation (Simon et a/., 1986). However, several lines of evidence lead to the conclusion that
an epoxide is not formed in appreciable quantities to have a role in vinyl acetate carcinogenesis.
First, unlike acetoxy oxirane, vinyl acetate is not genotoxic in the Ames test. Second, studies by
Norppa et al (1985), Laib and Bolt (1986), Simon et al. (1986) and Fedtke and Wiegand
(1990) support the idea that vinyl acetate is rapidly split by esterases and is therefore not readily
available for epoxidation. Moreover, the half-life of acetoxy oxirane in phosphate buffer (pH
7.8, at 37°C) is only 2.8 min and its mutagenicity is abolished completely by S9 mix
(Simon et a/., 1986). Third, vinyl acetate did not induce hepatic foci, of cellular alteration
(ATPase, GGTase) when administered intraperitoneally to neonatal rats either with or without
phenobarbital promotion (Laib, and Bolt, 1986). Other vinyl compounds which undergo
oxidation at the n bond to form a reactive epoxide metabolite, such as vinyl chloride and vinyl
carbamate, are positive in this test.
Cell Proliferation Effects
The effects of vinyl acetate .exposure on nasal epithelial cell proliferation were evaluated in rats
exposed for 1, 5, or 20 days to 0, 50, 200, 600, or 1000 pp'm (Bogdanffy et al., 1997a).
Exposure to vinyl acetate produced lesions in the olfactory epithelium of rats exposed to 600 or
1000 ppm (Table 14). The severity of olfactory epithelial lesions was concentration-related and
decreased along an anterior-posterior gradient. The severity of the lesions increased with
extended durations of exposure.
Following one exposure, lesions were characterized by degeneration, necrosis, and exfoliation
of olfactory epithelial cells. Areas of the olfactory mucosa most severely effected were the
dorsal one-third of the nasal septum and dorsolateral wall (i.e. areas just distal to the dorsal
arch), Masera's organ, and the medial most extent of the ethmoid turbinates.
32
-------
Olfactory mucosal lesions following 5 or 20 exposures were primarily those of post-necrotic
repair and adaptation. Lesions were characterized by regenerative hyperplasia of olfactory
epithelium along with attenuation and/or disorganization of the olfactory mucosa. Occasional
areas of squamous metaplasia were also present. Degeneration and atrophy of olfactory
nervebundles in the olfactory lamina propria were discernible following 20 exposures.
Cell proliferation responses were observed in both respiratory and olfactory epithelia at 1 day of
exposure (Figure 3). The responses were statistically significant in the 600 and 1000 ppm
groups at these times. Following 5 days of exposure, the proliferation responses subsided with
only slight increases noted at the 1000 ppm level. Following 20 days of exposure, the
respiratory epithelium appeared to have reached a point of adaptation to exposure while the
olfactory epithelium rebounded with a second wave of proliferation. This unusual time course
illustrates the distinctiveness of the responses of respiratory and olfactory epithelia and
highlights the concept that these tissues are separate organs. Transient cell proliferation
responses have also been observed in respiratory epithelium of rats exposed to 6 ppm
formaldehyde for up to six weeks (Monticello and Morgan, 1994).
33
-------
Table 14
Histopathological Observations of Rats Exposed to Vinyl Acetate
for up to Four Weeks"
Exposure Concentration
1 Exposure
Olfactory Epithelial
Degeneration
Respiratory Epithelial
Degeneration
5 Exposures
Olfactory Epithelial
Degeneration
Olfactory Epithelial
Regenerative Hyperplasia
Respiratory Epithelial
Regenerative Hyperplasia
20 Exposures
Olfactory Epithelial Degeneration ,
Olfactory Epithelial Regenerative
Hyperplasia
Respiratory Epithelial Regenerative
Hyperplasia
Squamous metaplasia
0
0
0
0
0
0
0
0
0
0
50
0
0
0
0
0
0
0
0
0
200
0
0
0
0
0
0
0
0
0
600
'•.5
0
3
5
0
5 .
5
0
0
1000
5
1
5
'
5
2
5
5
1
1
Five male rats were exposed nose-only for periods of 6 hrs per day, 5 days per week.
34 .
-------
Cell proliferation following 1 eapoare
Re^iratory
CHadsry
0 ZD «D BCD BOD KID
Viryl Aoetde QJIIJU trAjon(pprn)
Cell proliferation following 20 exposures
Vinyl Ao*ttt* Conc«rcraian (ppm)
Cen ptf iteration following 5 exposures
-dfacttry
Viryf AoeOe Oonoertrstion (pprn)
Figure 3. Cell proliferation responses in respiratory and
olfactory epithelium following 1,5, or 20 exposures to
vinyl acetate. Respiratory regions included the medial
and lateral aspects of the nasoturbinate at level n (Young,
1981). Olfactory regions included the dorsal meatus and
posterior portion of the nasal septum at level IV.
The cell proliferation responses could be interpreted as a two phase reaction to exposure. The
first is characterized as chemical insult followed by early regenerative repair (exposure days 1-
5). Early proliferative responses of respiratory and olfactory epithelium have been noted for
nasal irritants such as formaldehyde and acrylic acid (Swenberg et al, 1986). With regard to
effects on respirator)' epithelium, only one case of minimal degeneration in five rats exposed
was noted at 1000 ppm following 1 six hour exposure, and two cases of regenerative
hyperplasia were observed at 1000 ppm following 5 days of exposure. Complete recovery of
the respirator^' epithelium was observed following 20 days of exposure. The minimal response
and recovery in respiratory epithelium is consistent with the lack of non-neoplastic effects in
respiratory epithelium following 2 years of exposure. The lack of histopathology and cell
proliferation responses following 5 or 20 exposures suggests that the second phase of reaction
to exposure includes biochemical adaptation. In olfactory epithelium, the rebound response
35
-------
observed after 20 exposures coupled with expressions of active degeneration suggest adaptive
response apparent in respiratory epithelium is either not possible or not extensive enough in
olfactory epithelium.
The data from these experiments are instructive in several ways. First, the histopathology
demonstrates a strong anterior-to-posterior gradient. Such a gradient would be expected for
water soluble or reactive (i.e., metabolized) materials. Second, the gradient of response moves
anterior-to-posterior with increasing concentration. Such a gradient would be expected for
materials in which deposition is metabolically dependent. As vinyl acetate concentration
increases, fractional deposition decreases (i.e. greater distal penetration in the airway) due, in
part, to saturation of the metabolism-dependent component of deposition (Morris, 1995).
Finally, the data suggest that strong proliferative effects in the nasal cavity are observed early in
the sequelae of toxic responses to vinyl acetate exposure. These responses are likely to have
influence on the subsequent generations of epithelial cells that populate the nasal cavity during
the lifespan of the animal, and support a mode of action that is strongly dependent on induced
cellular proliferation.
IV. Determinants of Uptake and Tissue Dose
One of the key determinants of uptake for vinyl acetate is carboxylesterase metabolism located
in the olfactory epithelium of the URT.. This metabolism accounts for the proximal to distal
scrubbing of the chemical from the inspired airstream as well as the distribution pattern of
lesions observed in respiratory tract. A PB-PK model of vinyl acetate uptake and metabolism in
rats and humans has been developed using data derived from in vitro (rats and humans) and
in vivo (rats) kinetic analyses of deposition, uptake, and metabolism.
Carboxylesterase
The presumed mechanism of vinyl acetate nasal toxicity includes a significant role for the
carboxylesterase-mediated metabolism to acetic acid. The metabolism of vinyl acetate is
presented in Figure 4:
36
-------
Q ^ Carboxylesterase
O
„. , 'OH
Vtnyl Acetate Vinyl Alcohol
.^ Acetic Acid
.-. Aldehyde Dehydrogenase
Acetaldehyde
Figure 4 Established metabolic pathways for vinyl acetate.
Nasal mucosa of several species possesses active xenobiotic metabolizing systems that include
oxidative cytochrome P-450 systems, reductases, alcohol and aldehyde dehydrogenases, phase
n hydrolytic and conjugation enzymes, flavin monoxygenases, and carboxylesterases (Dahl and
Hadley, 1991). Several of these enzymes have been localized histochemically and are found in
discrete cell types within respiratory and olfactory mucosae (Bogdanfiy, 1990).'
Simon et al. (1985a) showed that vinyl acetate is hydrolyzed in preparations from rat and
human plasma, and rat liver and lung. The rate of hydrolysis of vinyl acetate has also been
measured in horhogenates of nasal respiratory and olfactory mucosa (Bogdanfiy and Taylor,
1993). Rat nasal carboxylesterase catalyzes the hydrolysis of vinyl acetate with great efficiency.
V,,^ values for rat and mouse respiratory tissue ranged from 22 to 46 mmoles/min/mg. This is
about the same as, or somewhat higher than, that obtained from rat liver microsomes
(23 mmoles/min/mg), and significantly greater than rat lung microsomes (6.2 mmoles/min/mg)
and rat and human plasma (0.56, and 0.69 mmoles/min/mg, respectively). Vam values for rat
and mouse olfactory tissue were considerably higher than other tissues and ranged from 95 to
254 mmoles/min/mg.
K,,, values obtained for rat and mouse nasal tissues ranged from 0.30 to 1.07 mM. K,,, values
obtained by Simon et al (1985a) were about the same or higher than nasal tissue: rat liver
microsomes, 0.73 mM; rat lung microsomes, 6.1 mM; rat plasma, 4.0 mM; and human plasma,
7.1 mM. Therefore, the increased capacity of rat nasal tissue to catalyze the metabolism of
vinyl acetate, coupled with the lower K,,, of nasal tissue relative to other tissues, yields a highly
efficient system for the in situ production of acetic acid and acetaldehyde. The efficiency of
hydrolysis of vinyl acetate in rat nasal tissue can be compared by considering the intrinsic
metabolic clearance, V/K. In rat and mouse respiratory tissue, V/K ranged from 52 to 79
37
-------
L/min/mg. In olfactory tissue, V/K ranged from 270 to 469 L/min/mg. Calculating this ratio
from the data of Simon et al. (1985a) yields the following values: rat liver microsomes, 32
L/min/mg; rat lung microsomes, 1.02 L/min/mg; rat plasma, 0.14 L/min/mg; and human
plasma, 0.10 L/min/mg. Since the liver data of Simon et al. (1985a) was collected using an
isolated microsomal fraction while the data reported here was derived from whole homogenates,
the differences between nose and liver carboxylesterase are likely to be even greater. Thus rat
olfactory tissue is the most efficient tissue type catalyzing the hydrolysis of vinyl acetate.
The results of the kinetic studies with tissue homogenates do not offer an explanation for the
species difference in sensitivity to the carcinogenic effects of vinyl acetate. Vinyl acetate was
hydrolyzed with approximately equal efficiency in respiratory or olfactory tissue of both rats and
mice. An explanation for the species difference in carcinogenic susceptibility must depend on
other factors such as differential rates of DNA damage and repair, cell proliferation kinetics, and
respiratory physiological factors influencing deposition of vinyl acetate vapor in upper airway
tissues.
Aldehyde Dehydrogenase
Nasal aldehyde dehydrogenase is present in specific cell types .of both respiratory and olfactory
mucosa although the activity is greater in respiratory mucosa (Bogdanffy et al, 1986;
Casanova-Schrnitz et al., 1984). Thus, acetaldehyde produced from vinyl acetate would also
be expected to be oxidized to acetic acid provided the cell type containing carboxylesterase also
contains aldehyde dehydrogenase. Casanova-Schmitz et al. (1984) studied the kinetics of
aldehyde dehydrogenase in rat mucosal homogenates. The authors observed the presence of
two isoforms in both the respiratory and olfactory mucosa, one of which may catalyze the
oxidation of both formaldehyde and acetaldehyde. The higher K^ isozyme, responsible for the
oxidation of acetaldehyde had a specific activity approximately five to eight times greater in
homogenates of respiratory (128 nmoles/min/mg protein) than olfactory tissue (28
nmoles/min/mg protein). K^ values obtained for the two tissues were similar (20 ± 3, and 22 ±
1 mM for respiratory and olfactory, respectively). Therefore, the respiratory mucosa appears to
be better equipped to detoxify acetaldehyde produced from the hydrolysis of vinyl acetate. This
is an important observation because it suggests that as vinyl acetate concentration is reduced
from high experimental levels to low ambient levels, and the deposition pattern moves
progressively towards the anterior of the nasal cavity (consistent with the concentration-related
gradient in nasal lesion formation noted above and the anticipated deposition pattern which will
be discussed below), an increasing fraction of vinyl acetate is deposited in respiratory epithelium
38
-------
where acetaldehyde metabolites are more readily detoxified.
Histochemical Distribution of Vinyl Acetate-Metabolizing Enzymes
Nasal carboxylesterase and aldehyde dehydgrogenase activities are critical enzymes in the
proposed mechanism of metabolic activation and detoxication of vinyl acetate. A comparison of
the cellular distributions of carboxylesterase and aldehyde dehydrogenase helps in
understanding uptake and possibly tissue sensitivity. Carboxylesterase activity is histochemically
detectable in all epithelial cells and seromucous glands of respiratory mucosa (Bogdanfiy et al,
1987). In olfactory mucosa, carboxylesterase is present in sustentacular cells, basal cells, and
Bowman's glands. Aldehyde dehydrogenase is present in all epithelial cells of respiratory
mucosa, but is present in only basal cells and Bowman's glands of the olfactory mucosa and only
at minimally detectable levels (Bogdanfiy et al., 1986). This comparison suggests acetaldehyde
metabolites produced in the surface epithelium will be converted to acetic acid to a greater
extent in respiratory epithelium than olfactory epithelium. Further, the mode of action studies
presented above showed that acetaldehyde was not cytotoxic, indicating little contribution of
acetic acid derived from acetaldehyde oxidation to the overall mechanism of vinyl acetate-
induced cytotoxicity.
Further support for this conclusion comes from a comparative analysis of the V/K ratios of
carboxylesterase vs. aldehyde dehydrogenase derived from the respiratory tissue homogenate
experiments. The former is in the range of 50 L/min/mg protein while the latter is
approximately 6x10"* L/min/mg protein. Thus, the contribution of carboxylesterase to the
total amount of acetic acid generated intracellularly is significantly greater than that of aldehyde
dehydrogenase oxidation of acetaldehyde.
Ultimately, the complete expression of toxicity and neoplasia will be critically dependent on the
balance between several competing mechanisms in the different mucosae. The first is
cytotoxicity induced by acetic acid from vinyl acetate hydrolysis. The second is metabolic
incorporation and general detoxication of acetic acid. The third is detoxication of acetaldehyde.
Acetaldehyde is detoxified through both aldehyde dehydrogenase oxidation and through binding
to cellular macromolecules and thiols, such as glutathione. Glutathione has been shown
histochemically to be present in all epithelial cells of respiratory and olfactory mucosae
(Keller efa/., 1990).
With regard to the histochemical distribution of the various enzymes involved in the
39
-------
mechanisms of activation and detoxication of vinyl acetate in human nasal tissue, Lewis et al.
(1994) have studied the distribution of carboxylesterase activity immunohistochemically in
respiratory mucosa. The cellular pattern of activity was' similar to that of rat respiratory
epithelium with diffuse reactivity noted in ciliated and secretory cells of the luminal respiratory
epithelium. Histochemical staining in human olfactory tissues for carboxylesterase activity, or
for aldehyde dehydrogenase in any human nasal tissue, has not been reported.
Physiologically-Based Modeling of Vinyl Acetate Uptake and Metabolism
Inhaled chemicals can be extracted in the nasal cavity where they are then metabolized and
absorbed into the systemic circulation. Inspired air follows distinct paths in the URT, resulting
in asymmetric ventilation to various regions, and the nasal mucosa consists of a variety of cell
types each having a different metabolic activity toward the inhaled chemical. Vinyl acetate
exposure induces non-neoplastic lesions in the rat nasal cavity with degeneration of the
olfactory epithelium as the critical response. In vivo experiments show that the severity of
olfactory epithelial lesions decreases along an anterior-posterior gradient. Dividing the olfactory
region into numerous compartments in the axial direction should allow a better approximation
of the proximal to distal scrubbing of the chemical from the inspired air stream and thus account
for the distribution of the lesions observed in the nasal cavity.
To accurately capture the flux to the sensitive regions of the nasal mucosa that are at a higher
risk of tissue damage due to vinyl acetate, requires a high degree of compartmentalization of the
URT. The model of Plowchalk et al. (1997) was extended by constructing a five-compartment
model of the rat nasal cavity and a four-compartment model of the human nasal cavity. The
airflow is split into the two demonstrated distinct pathways in the URT: lateral/ventral and
dorsal/medial (Kimbell et al., 1993). To better characterize the vinyl acetate flux to the apical
regions of the olfactory tissue, the five-compartment model divides the olfactory region into two
compartments; a small dorsal anterior compartment and a larger posterior compartment. The
respiratory mucosa on the ventral side is also divided into an anterior and a posterior
compartment, resulting in five tissue compartments, similar to the model structure proposed by
Frederick et al. (Fig. 5). Since the human nasal cavity has only a small area covered with
olfactory mucosa, one olfactory compartment on the dorsal side is used in the equivalent human
PB-PK model. In addition to representing the nasal mucosa using more compartments the
current model also incorporates air phase resistance to mass transfer from the lumen to the
airmucus interface. This is an improvement over the previous model by Plowchalk et al.
(1997) that assumed equilibrium between the air and the mucus phase.
40
-------
Nasal carboxylesterase and aldehyde dehydrogenases are critical enzymes in the mechanism of
metabolic activation and detoxification of vinyl acetate. These enzymes are located in specific
cell and tissue types within the nasal cavity. The extent to which vinyl acetate is extracted from
the air stream and metabolized is dependent upon the compliment of enzyme activity in the
various tissue sub-compartments. Each of the tissue compartments in the model are further
subdivided into a number of subcompartments to represent the mucus layer and the various cell
types of the tissue. The histochemical localization described earlier was used to distribute the
enzymatic activities of carboxylesterase and aldehyde dehyrodgenase within the compartments
of each tissue stack.
41
-------
Arterial Blood
Blood
Figure 5. Schematic representation of the PB-PK model for vinyl acetate extraction and metabolism in the nasal cavity.
The basic model structure is adapted from Morris (1993). Regional gas phase mass transfer coefficients and
compartment sizes are as defined in Frederick et al (1998). Tissue enzyme distribution and estimation of intracellular
pH are as in Plowchalk et al. (1997). The basic model structure is similar for rats and humans except that the human
model does not contain an olfactory 2 compartment Physiological and metabolic constants for each model are species-
specific. Cm = concentration entering the nose, C^ = concentration entering the nasopharynx; CE = carboxylesterase;
AldH = aldehyde dehydrogenase; AAld = acetaldehyde; AA = acetic acid
A high-affinity/low-capacity carboxylesterase pathway was included in the model. The kinetic
constants for this pathway were obtained by numerical optimization against deposition data
collected at a flow rate of 100 mL/min. Model estimates of fractional deposition (Figures 6 and
7), absolute deposition, and expired vinyl acetate and acetaldehyde concentrations (Figure 8)
were in good agreement with the experimental data as a result. The presence of both high- and
low-affinity isoforms of carboxylesterase in various species and tissues is common
(Morgan et al., 1994). K,,, values of approximately 25 and 400 ^im for the high- and low-affinity
carboxylesterase have been reported for rat liver and are consistent with those predicted from
the PB-PK model for nasal epithelium (55 nM). The optimized values for the respiratory and
42
-------
olfactory high affinity/low capacity term are: V^ 2.6 mg/hr; K,,,, 4.7 x 10"3 mg/mL (55
For expired acetaldehyde, systematic departures from the observations were evident at the
highest inspired vinyl acetate concentrations (Figure 8). However, model predictions were
consistently reasonable at the lower vinyl acetate concentrations. Thus, the model was
considered acceptable for low exposure dosimetry extrapolation.
c
o
'to
o
Q.
0)
TJ
"CD
c
g
"o
03
100
95
90
85
80
75
70
65
60
55
50
45
35
30
25
20
15
10
5
0
- -o- - Observed
—•—Tissue and mucus CE
—•—Tissue CE
—•— No metabolism
200
400
.600
800
1000 1200 1400 1600
VA Exposure Concentration (ppm)
Figure 6. Fractional deposition of vinyl acetate in the nasal cavity is nonlinear with exposure concentratioa
Simulation without nasal carboxylesterase (filled circles) indicate blood flow has little impact on vinyl acetate
depositioa With carboxylesterase present (squares) deposition is increased and addition of a high-affiniry/k>w
-------
f
¥
O Sonutatton
o 400 too 1700 mo o 400 too 1200 icoo
Vinyl Acetate Concentration (ppm)
0 400 »00 1200 1COO
Figure 7. Model simulations versus experimental observations of nasal extraction of vinyl acetate. Nasal extraction
studies were carried out at three flow rates, the studies conducted at a flow rate of 100 mL/min were used to optimize
the high affinity/low capacity metabolic pathway used in the whole nose model. Model simulations of the experiment
(open circles) were run for each exposure concentration and flow rate. Experimental observations (filled squares)are
the mean ± S.E. for an n = 4 or 5.
44
-------
S
—•—50mL/min
o Simulation
too
560-
500.
450
400.
360.
300.
BO
200
150
100
—•—100 mUtrm
O anMon
«00 800 UOO 1600 0 400 tOO 1200 1800
Vinyl Acetate Concentration (ppm)
400 800 1200 161
Figure 8. Model simulations versus experimental observations of acetaldehyde release into
expired air during the nasal extraction studies. Nasal extraction studies were carried out at three
flow rates. The studies conducted at a flow rate of 100 mL/min were used to optimize the high
affinity/low capacity metabolic pathway used in the whole nose model. Model simulations of
the experiment (open circles) were run for each exposure concentration and flow rate.
Experimental observations (filled squares) are the mean ± S.E. for an n = 4 or 5.
45
-------
Dosimeters Based on Mode of Action
Rat
Potential dosimeters for olfactory mucosa exposure to vinyl acetate, acetaldehyde, and acetic
acid include peak tissue concentrations, area under the concentration-time curve (AUC) and
total amount of metabolites formed. Simulations of 6-hour inhalation exposure to vinyl acetate
at concentrations of 50, 200, and 1000 ppm indicate olfactory tissue exposure of rats to vinyl
acetate is minimal compared to both acetaldehyde and acetic acid (Table 15). This is due to
rapid hydrolysis of vinyl acetate by carboxylesterase and is also the reason for the steep vinyl
acetate concentration gradient (6 orders of magnitude) predicted for olfactory tissue.
Table 15
Olfactory 1 tissue dosimeters predicted after simulations of a 6-hour inhalation
exposure of rats to 50-600 ppm vinyl acetate*
Exposure
Concentration
(ppm)
50
200
600
Steady-state Tissue
Concentration (ug/mL)
VA AAld AA
0.22 1.9
1.11 6.3
4.08 14.1
56.5
200.8
411.7
AUCQigxhr/mL)
VA AAld AA
1.33
6.67
24.5
12.0
38.0
84.6
339
1203
• 2467
Amount (mg) formed
per mL tissue
AAld AA
23 31
112 153
390 531
' Simulations were run at an inspirator)' flow rate of 197 mL/min.
VA = vinyl acetate
AALD = acetaldehyde
AA = acetic acid
Dosimeters from model predictions of vinyl acetate-induced intracellular acidification of
olfactory epithelium are presented in Table 16. The predicted pH; decreases and reaches a
steady-state which is a function of the predicted rate of H' formation by metabolism and
predicted rate of FT extrusion by the Na7H+ antiport. Thus, a 6-hr exposure to 50 ppm is not
expected to cause a significant increase in H" exposure or decreased pH; (A pHj <0.1). This
modeling prediction is in accordance with the observed NOAEL. A high incidence of olfactory
lesions were observed in rats after one 6 hour exposure to vinyl acetate at 600 ppm
(Bogdanfiy et al., 1997), which is consistent with the predicted pH; reduction (pHj = 6.91).
46
-------
1 death is a function of both the degree and duration of cellular acidification (Nedergaard et al.,
1991). Exposures of neuronal and glial cells, cells which can be considered analogous to
olfactory sensory and sustentacular cells to pHj 6.7 for less than 4 hours is not cytotoxic,
whereas 6-hr exposures to the same pH will cause cell death (10-20%). A single 6-hr exposure
of rats to 200 ppm was without effect, but chronic exposure of rats to 200 ppm produced a
significant incidence of olfactory degeneration. Therefore it appears from the intracellular pH
model that a pH; of 7.15, predicted by the model for a 200 ppm exposure is tolerated for only
short durations, but not over a lifetime of exposures.
Table 16
Dosimeters of intracellular acidification in olfactory
epithelium of the rat after a 6-hour exposure to vinyl acetate
Exposure
Cone. '
(ppm)
0
50
200
600
Final
Proton
Cone. (mM)
3.98 x 10*
4.79 x lO"8
7.08 x 10"*
1.23x lO'7
A Final
Proton
Cone. (mM)
0
0.8 xlO"8
+3.1 x 10-8
+8.3 x ID"8
Proton
AUC
(mmole x hr/L)
2.39 x 10'7
2.84 x 10'7
4.21 x 10'7
7.31 x 10'7
A Proton
AUC
(mmole x hr/L)
0
+0.45 x JO'7
+1.83x JO'7
+4.92 x 10'7
Final
PH
7.40
7.32
7.15
6.91
A Final
PH
0
-0.08
-0.25
-0.49
An interesting aspect of the dose-response curve for olfactory degeneration (Figure 1) is the
small rise in the response between 200 ppm and 600 ppm. Nasal air flow rate is an important
determinant of vinyl acetate delivery to nasal tissue and, hence, the degree of cellular
acidification. Factors influencing ventilation rate, such as respiratory depression induced by
exposure to high concentrations of irritating vapors may confound the interpretation of
exposure-response data. Respiratory rate depression was a determinant in the interspecies
differences in response to formaldehyde-induced nasal tumors (Barrow, et al., 1986). Recently,
the sensory irritation responses of mice to vinyl acetate exposure were measured (Dudek,
1996). The measured RDSO for vinyl acetate was 380 ppm. From these data, the respiratory
rate was predicted to be depressed approximately 68% at 600 ppm. Assuming a similar
response of rats to vinyl acetate, simulations of the 600 ppm exposure were conducted with a
68% reduction in minute volume. Replotting lesion incidence with the new dosimeter for the
600 ppm exposure reduced the sigmoidicity of the response curve. Therefore, depression of
47
-------
respiratory rate at high concentrations provides a possible explanation for the observed
nonlinearities in the response data in the high exposure concentration range (Figure 9).
— 60-
0-
—•— SmuMIM vntmm at AA twm*d i! IBJmL/tnrfi
- -O- • Smuktrt amount of AA torn*] it 37% of
nttmg ntitnrMorv nt«
100 200 300 400 500
Amount of Acetic Acid Formed (mg)
Figure 9. Simulations with respiratory rate reduced by 68% at only the 600 ppm exposure concentration (open
circles) reduced the total amount of acetic acid formed in olfactory tissue and resulted in a more linear correlation
between response and external exposures of 50 ppm (3% response) and 600 ppm (86% response). Solid circles
represent acetic acid estimates based on normal respirator)' rate at all exposure concentrations.
48
-------
Human
Estimates of dosimeters for human olfactory epithelium at 50 and 600 ppm are consistent with
the rat in that tissue exposure to vinyl acetate is minimal, while the greatest exposure is to acetic
acid (Table 17). Both predicted tissue concentration and AUC are greater in humans compared
to the rat when exposed to the same vinyl acetate concentrations suggesting that, relative to the
rat, human olfactory tissue may receive a greater dose of metabolites per unit vinyl acetate
exposure . However, predictions of pHp H* concentration and H* AUC all indicate similar
exposure of nasal tissue to reduced pH; in humans compared to rats when exposed to the same
external exposure concentration. Simulation of a 6-hour, 50 ppm exposure to vinyl acetate
resulted in a predicted final pH; of 7.32 in the rat compared to 7.35 in the human.
Table 17
Dosimeters of off acton1 epithelium exposure in humans after a simulated
6-hour inhalation exposure to vinyl acetate
Exposure
Concentration
(ppm)
50
200
600
Steady-state Tissue
Concentration (ug/mL) AUC (ug x hr/mL)
VA AAld AA VA AAld AA
0.2 10.6 82.7 1.35 63.6 495
1.7 38.9 345.6 10.1 233.3 2067.6
13.8 77.8 655.6 82.4 466.3 3921.6
Amount (mg) formed
per mL tissue
AAld AA
15.7 21.4
114.2 155.8
752.2 1025.5
• Simulations were run at an inspirator)- flow rate of 7.5 L/min
VA = vinyl acetate
AALD = acetaldehyde"
AA = acetic acid
49
-------
Table IS
Dosimeters of intracellular acidification in olfactory epithelium of humans after a
simulated
6-hour exposure to vinyl acetate
Exposure
Cone.
(ppm)
0
50
200
600
Final
Proton
Cone. (mM)
3.98 x ID"8
'4.72x ID"8
8.13 xlO-8
2.40 x 10'7
A Final
Proton
Cone. (mM)
0
0.74 x 10*
4.15x10*
2.00x10-7
Proton
AUC
(nunole x hr/L)
2.39 x 1C'7
2.79 xlO'7
4.72.x 10-7
1.40x10-6
A Proton
AUC
(mmole x hr/L)
0
0.26 x lO'7
1.51x10-7
8.68x10-7
Final
PH,
7.40
7.33
7.09
6.62 .
A Final
PH
0
-0.07
-0.31
-0.71
50
-------
APPENDIX D
Charge to the Participants
-------
Framework for Human Health
Risk Assessment Colloquia Series
Colloquium #2
CHARGE TO THE PARTICIPANTS
Background
There is a recognized need for the development of a framework for human health risk
assessment that puts a perspective on the approaches that are currently being practiced
throughout the Agency. In its 1994 report entitled Science and Judgement in Risk Assessment
(NRC, 1994), the NRC noted the importance of an approach that is less fragmented, more
consistent in application of similar concepts, and more holistic than endpoint-specific guidelines.
Both the NRC and the Agency's Science Advisory Board have raised a number of issues for both
cancer and noncancer risk assessment, that should be reconsidered in light of recent scientific
progress. In response to these needs, the Agency's Risk Assessment Forum is beginning the
long-term development of a human health risk assessment framework. As part of this effort, the
Risk Assessment Forum has invited you to participate in the second of two colloquia, which are
intended to bring together EPA risk assessors for a dialogue on various scientific and policy
issues pertaining to EPA's cancer and noncancer risk assessment approaches. The second
colloquium will focus on the role of mode of action information in developing descriptive
quantitative models, applicable to a variety of needs for carrying out a risk assessment.
Charge to the Participants
Prior to the second colloquium, each participant is receiving a single case study, a list of
general questions, and a list of the breakout groups. As in the first colloquium, the case studies
and accompanying questions will guide the discussions. The participants will spend the bulk of
the first day discussing their assigned case studies with specific focus on the case study questions
and the general questions for the plenary session. It is important that each participant review
their work group's case study and be prepared to add their scientific and regulatory expertise to
the work group discussion.
Each participant has been assigned to a specific breakout group. In making the group
assignments, -EPA sought to ensure a mix of expertise and Agency representation in each group.
Each breakout 'group will have a chair to facilitate the discussion and a rapporteur to capture the
consensus of the group. It is important that each of you participate in the breakout group to
which you have been assigned.
-------
APPENDIX E
s
General Questions for Plenary Session
-------
Framework for Human Health
Risk Assessment Colloquia Series
Colloquium #2
GENERAL QUESTIONS FOR PLENARY SESSION
Should a common quantitative analysis be conducted when there are commonalities
among toxicities?
In the absence of case-specific PBPK models, is there a common approach for dose
adjustment for interspecies extrapolation for all responses? Does this differ for different
routes of exposure?
In the presence of PBPK models, how does mode of action (MO A) information influence
the dose surrogate in characterizing toxicity? Can it be different for different responses?
In the absence of BBDR models, how does MOA information influence the default
approach(es) to characterize in quantitative terms the potential risk of toxicities at low
levels of exposure (i.e., beyond the range of observation)? Are there common default
approaches?
The 1996 Proposed Guidelines for Carcinogen Risk Assessment have recommended that
five factors be considered when determining the margin of exposure (MOE). These
include intraspecies variation, interspecies variation, nature of the response, steepness of
the dose-response curve, and biopersistence.
The current quantitative approach for noncancer effects generally involves development
of a single RfD/RfC for a "critical effect". Factors used include intraspecies variation,
interspecies variation, subchronic to chronic extrapolation, LOAEL to NOAEL
extrapolation, and completeness of the data base. An additional factor may be applied to
account for scientific uncertainties in the study selected for derivation of the RfD/RfC.
If the goal is to harmonize across toxicities, can a consistent set of factors be identified?
How does MOA information influence the choice of these factors?
-------
APPENDIX F
Breakout Group Assignments
-------
Framework for Human Health Risk Assessment Colloquia Series
Colloquium #2
Breakout Group Assignments
Wednesday and Thursday, June 3-4, 1998
Breakout Group 1
Versailles I Room
Ethylene Thiourea
Chair: Jennifer Seed
Rapporteur: Vicki Dcllarco
Barbara Abbott
Charles Abernathy
Kevin Crofton
Julie Du
Gary Foureman
Jennifer Jinot
Carole Kimmel
Jim Rowe
Gino Louis Scarano
Bill Sette
Breakout Group 2
Georgia Room
Ethylene Oxide
Chair: Gary Kimmel
Rapporteur: Kerry Dearfield
Eric Clegg
Jim Cogliano
Marion Copley
Penny Fenner-Crisp
Chris Lau
Bob Luebke
Cheryl Scott
Woodrow Setzer
Mark Stanton
Yin-Tak Woo
William Wood
Breakout Group 3
Connecticut Room
Trichloroethylene
Chair: Vanessa Vu
Rapporteur: Kim Hoang
Bob Benson
Carole Braverman
Chao Chen
Oscar Hernandez
Lee Hoffman
Aparna Koppikar
David Lai
Marc Mass
Bob McGaughy
Edward Ohanian
Gina Pastino
Breakout Group 4
Gallery Room
Vinyl Acetate
Chair: Annie Jarabek
Rapporteur: A mold Kuzmack
Donald Barnes
Karl Baetcke
Anne-Marie Burke
Terry Harvey
Richard Hill
Elizabeth Margosches
William Pepelko
Rita Schoeny
John Whalan
Paul White
Printed on Recycled Paper
-------
APPENDIX G
Agenda
-------
United States Environmental Protection Agency
Office of Research and Development
Risk Assessment Forum
Framework for Human Health
Risk Assessment Colloquia Series
Colloquium #2
Holiday Inn Bethesda
Bethesda, MD
June 3-4, 1998
Agenda
WEDNESDAY, JUNE 3, 1998
8:30AM Registration
9:OOAM Welcome Remarks William Wood
Risk Assessment Forum
U.S. Environmental Protection Agency (EPA)
Washington, DC
9:15AM Goals of the Human Health Risk Assessment Framework,
Introduction to Case Studies, and Colloquium Issues
and Charge to Breakout Groups Vanessa Vu
National, Center for Environmental Assessment
U.S. EPA
Washington, DC
9:45AM BREAK (Move to Breakout Rooms)
10:OOAM Breakout Groups Convene to Address Case-Specific Questions
12:OOPM LUNCH (on your own)
Printed on Recycled Paper
-------
WEDNESDAY, JUNE3, 1998 (continued)
1:OOPM Continue Breakout Group Discussions
2:30PM Status Report of Breakout Group Discussions
3:OOPM BREAK
3:15PM Continue Breakout Group Discussions
5:OOPM ADJOURN
THURSDAY, JUNE 4, 1998
8:3 0AM Review of Day Two Charge
8:35AM Breakout Group Reports and Discussions: Groups 1 and 2
10:OOAM .BREAK
1.0:15AM Breakout Group Reports and Discussions: Groups 3 and 4
12:OOPM LUNCH (on your own)
1:OOPM Plenary Session: Lessons learned and their applications to the development of
a Human Health Risk Assessment Framework
Moderators: Vanessa Vu and Gary Kimmel
3:30PM ADJOURN
------- |