vvEPA
United States
Environmental Protection
Agency
Research and
Development
Washington, DC 20460
EPA/600/9-90/038
June 1990
Research to Improve
Health Risk Assessments
(RIHRA) Program
-------
-------
EPA/600/9-90/038
June 1990
Research to Improve Health Risk
Assessments (RIHRA) Program
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
Printed on Recycled Paper
-------
-------
FOREWORD
To realistically assess the human health risks associated with exposures to environmental
pollutants, we must understand a number of important physical, chemical, and biological processes
and mechanisms. This suggests that real progress in enhancing the accuracy and precision of health
risk assessments depends on the products of a research program focused on the underlying
mechanisms of exposure, dose, and effects. Although much of the U.S. Environmental Protection
Agency's (EPA's) base research program has been targeted historically on these issues, substantial
benefits can still be gained from instituting a new, more focused research program aimed specifically
at reducing the critical uncertainties associated with health risk assessments.
The goal of this program, known formally as the Research to Improve Health Risk Assessments
(RIHRA) Program, is to generate research results that will significantly improve the EPA's ability to
assess human health risks. The RIHRA program is structured so that the research products are
germane to a cross-section of the Agency's regulatory programs and risk assessment needs. We are
confident that the mission-oriented research described in this document will provide the EPA with
credible scientific findings that are relevant for risk assessment decisions.
Ken Sexton, Sc.D.
Director, Office of Health Research
Chairman, RIHRA Committee
-------
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY
INTRODUCTION i
TOPIC 1 -ANALYSES OF UNCERTAINTY IN RISK ASSESSMENT 1-1
1.1 Issues Covered 1-1
1.2 Development of a Framework 1-1
1.3 Research Strategy , 1-1
TOPIC 2 INTEGRATED EXPOSURE ASSESSMENT 2-1
2.1 Issues Covered 2-1
2.2 Gathering Exposure Data 2-3
2.3 Project Selection Strategy 2-6
TOPIC 3 PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS 3-1
3.1 Issues Covered 3-1
3.2 Project Selection Strategy 3.4
TOPIC 4 BIOLOGICALLY BASED DOSE-RESPONSE MODELS 4-1
4.1 Issues Covered 4-1
4.2 Project Selection Strategy 4.3
4.3 Decision-Tree Strategy for Prioritizing Research Needs 4-4
4.4 Summary of the Proposed Research Program 4-6
-------
APPENDIX A
TABLE OF CONTENTS (cont.)
TOPIC 1: ANALYSES OF UNCERTAINTY IN RISK
ASSESSMENT PROJECT DESCRIPTIONS
A-l
APPENDIX B
TOPIC 2: INTEGRATED EXPOSURE ASSESSMENT
PROJECT DESCRIPTIONS
B-l
APPENDIX C
TOPIC 3: PHYSIOLOGICALLY BASED PHARMACOKINETIC
MODELS PROJECT DESCRIPTIONS C-l
APPENDIX D
TOPIC 4: BIOLOGICALLY BASED DOSE-RESPONSE MODELS
PROGRAM OVERVIEWS AND PROJECT DESCRIPTIONS D-l
-------
EXECUTIVE SUMMARY
The U.S. Environmental Protection Agency (EPA) must assess the health risks of environmental
exposures in order to make regulatory decisions that safeguard public health. The Agency is limited
in its ability to assess health risks quantitatively because of a lack of understanding about the
underlying biological, chemical, and physical processes that determine exposures and effects. Without
sufficient knowledge of these processes, uncertainties are introduced into the risk assessment process
that allow wide interpretation of the limited available data. Thus, the current database may support
diametrically opposed assessments of risk, each with dramatically different ramifications for related
regulatory decisions.
Targeted, long-term research is needed to reduce these uncertainties in health risk assessment.
To meet this need, EPA's Office of Research and Development (ORD) established a systematic and
integrated Research to Improve Health Risk Assessments (RIHRA) program. This research program
is designed to provide critical data on the relationship between exposure, dose to target tissue
(delivered dose), and associated health effects. The program emphasizes laboratory and field research
to improve understanding of basic biological mechanisms, especially as they relate to our ability to
extrapolate from one set of circumstances (e.g., animals exposed to short-term, high concentrations)
to another (e.g., humans exposed to long-term, low concentrations). In implementing an integrated
and systematic research effort, the RIHRA program builds on existing capabilities within ORD and
ongoing intramural efforts. Data gathered in projects under the RIHRA program will enhance our
ability to quantify the human health risks associated with environmental exposures.
-------
-------
RIHRA
Introduction - i
INTRODUCTION
BACKGROUND
The U.S. Environmental Protection Agency
(EPA) is responsible for protecting public health from
the adverse effects of exposures to environmental
agents. In developing regulations to protect public
health, however, EPA currently relies on quantitative
assessments of the health risks associated with
different pollutants. The current lack of understand-
ing about the underlying biological, chemical, and
physical processes that determine exposures and
effects hobbles EPA's ability to make these assess-
ments. Often, quantitative risk assessment is entirely
precluded because of the paucity of appropriate data.
In other cases, risk assessment is possible, but numer-
ous assumptions representing "fall back" or "default"
positions must be applied due to critical data gaps.
Application of these assumptions fosters enough
uncertainty regarding the interpretation of the avail-
able data that diametrically opposed risk assessments
can be made on the basis of the same information.
In 1988, Congress recommended that EPA's
Office of Research and Development (ORD) establish
a systematic and integrated Research to Improve
Health Risk Assessments (RIHRA) program. Al-
though much of ORD's ongoing research focuses on
this issue, the Agency saw benefit in a more formal,
structured approach. Consequently, $3 million in
fiscal year (FY) 88 and $10 million in FY89 ($7
million for health; $3 million for ecology) were
earmarked for development of a targeted, coherent
research program to reduce the uncertainties in the
risk assessment process.
COMPLETE RISK MODEL
To understand the human health risk associated
with environmental exposures, the principal relation-
ships between the various sources of a pollutant and
the pollutant's effects on a target population must be
defined. Establishing these connections for a particu-
lar target and pollutant is a critical task in any effort
to reduce total risk from that pollutant. We can
construct the causal chain joining a pollutant source
to effect—and understand the biological, chemical,
and physical processes that underlie human exposure
and response to the pollutant—through an evaluation
of each of the four components of, or links in, the
chain:
• Source of the pollutant
• Movement of the pollutant from the source to the
target resulting in exposure of the target to the
pollutant (NOTE: RIHRA is concerned primarily
with the human population as target; a compan-
ion program on ecological risk assessment is in
preparation)
» Dose received by the target caused by this
exposure
• Adverse effects resulting from the dose
The informational output for one component in
the chain is the input for the next. A mathematical
formulation representing this entire chain would
constitute a complete risk model for the pollutant
RISK ASSESSMENT FRAMEWORK
Assessing the risk associated with exposure to a
given pollutant requires a sufficient understanding of
each of the four links in the complete risk model. If
the necessary information for a single link is missing,
the chain between source and effect cannot be fully
characterized. It is possible to focus on a single link
in the chain in developing mitigation strategies—for
example, strategies aimed at the sources or movement
patterns of a pollutant or the behavior of a target may
effectively mitigate risk. Estimates of the residual
health risk, however, must still be derived after
implementation of such strategies.
The steps in risk assessment mirror the links in
the chain joining the human population to pollutant
source(s). Given the inadequacy of current databases,
these steps also represent key issues in environmental
health research:
• Exposure Assessment What environmental
exposures occur or are anticipated to occur for
relevant human populations?
• Hazard Identification/Dose-Response Assess-
ment Does the agent cause an adverse health
effect? What is the site(s) of toxic action within
the body and what is the pollutant dose delivered
to that site(s)?
-------
RIHRA
Introduction - ii
Risk Characterization. What are the relation-
ships among exposure, target dose, and adverse
health consequences?
UNCERTAINTIES IN RISK ASSESSMENT
The causal pathway between pollutant source and
heath effect cannot be fully characterized for every
possible environmental scenario. It is not possible to
fully research every set of exposure conditions in
relation to every possible pollutant, or to examine the
particular susceptibility of every human subgroup to
effects from a pollutant In addition, much data on
response to pollutants must be gathered from labora-
tory animals under an entirely different set of expo-
sure conditions than humans experience. The most
critical problem in risk assessment—how to make
accurate extrapolations—develops from these limita-
tions. An unacceptable level of uncertainty is current-
ly associated with efforts to extrapolate observed
effects from one set of circumstances (e.g., cancer
incidence in rats subjected to high, chronic exposures
in controlled experiments) to an entirely different set
of circumstances (e.g., individual excess cancer risks
in humans experiencing intermittent, low-level expo-
sures).
Researchers encounter these uncertainties while
extrapolating from species to species, from one
individual or subgroup to another individual or
subgroup within a particular species, or from one set
of exposure conditions to another. The uncertainties
spring not only from the insufficiency of data, but
also from a lack of fundamental understanding about
the relevant underlying physical (e.g., atmospheric
dispersion characteristics, human activity patterns),
chemical (e.g., chemical reactions and transforma-
tions), and biological (e.g., metabolism, disease
processes) mechanisms that affect the validity of the
extrapolation assumptions.
For example, most risk assessments in the past
have assumed by default that an equivalency of
response exists between animals and humans. As one
proceeds from the molecular level or biochemical
event toward injury at the tissue or organ level, that
assumption may become less tenable. A sufficiently
high dose can produce damage at any level from the
target molecule to the intact organ, resulting in
various disease states. On the other hand, host
defense and/or compensatory systems could intervene
to prevent or repair the damage. Disease outcome is
also heavily dependent on the size of the delivered
dose and dose-time relationships. The lack of infor-
mation on these related events results in significant
uncertainty in current risk assessments and, as a
result, the use of a variety of default assumptions.
Under RIHRA, careful consideration of the
assumptions made at various stages of the risk assess-
ment process has highlighted areas in which research
can improve risk assessments by providing scientifi-
cally defensible data to replace these assumptions.
Even if the assumptions cannot be refuted, the data
may support more defensible defaults. In either case,
such research will increase our confidence in making
health risk assessments.
DEVELOPMENT OF THE
RIHRA PROGRAM
Selection of Topics
The RIHRA program is designed to focus limited
resources into a structured and integrated program for
reducing uncertainties in the risk assessment process,
thus enhancing our ability to quantify the human
health risks associated with environmental exposures.
ORD has taken three significant steps to focus the
program:
1. Collective judgment and understanding of the
risk assessment process was used to narrow the
scope of the program. Accordingly, the RIHRA
program will focus on elucidating the relation-
ship between exposure, internal dose to the target
tissue, and associated health effects. The RIHRA
program does not encompass efforts to measure
ambient concentrations of pollutants for exposure
assessments or to characterize sources; instead, it
studies concentrations of pollutants at the point
of human contact as well as the health effects
associated with that contact. Nor does it encom-
pass efforts in hazard identification—which is an
area of research covered by the ORD base
program (see Program Implementation).
2. Special emphasis will be given to noncancer
health endpoints, such as neurological, pulmo-
nary, and reproductive effects. This emphasis is
appropriate because EPA is increasingly called
on to estimate these kinds of risks for environ-
mental exposures to a variety of pollutants. In
general, the risk assessment issues are less well-
defined and articulated for noncancer than for
cancer effects.
3. The program will involve extensive short-term
and medium- to long-term research planning to
achieve the necessary stability. Stability is
-------
RIHRA
Introduction - iii
required because research to reduce the uncer-
tainties of risk assessment is iterative in nature;
that is, the research progresses in multiple stages,
with significant outputs at each stage.
In light of this focusing effort, ORD chose four
major topics for inclusion in the program (the rela-
tionship between these topics and factors in the risk
assessment process can be seen in Figure 1-1):
Topic 1: Analysis of Uncertainty in Risk
Assessment. The degree of uncertainty associa-
ted with the qualitative and quantitative aspects
of risk assessment is often poorly understood.
Research related to this topic addresses clarifying
and quantifying the uncertainty associated with
each assumption.
Topic 2: Integrated Exposure Assessment.
Exposure assessment (i.e., contact between chemi-
cals and humans) is based on either ambient or
Major Factors in
Health Risk Assessment
Major Components
Of RIHRA
Environmeniaf
Levels
Applied Dose
Delivered Dose
(target)
Health Effects
Uncertainty Analyses
Exposure I
Assessment 1
Physiologically Based
Pharmacokinetic
Models
Biologically Based
Dose-Response
Models
RISK ASSESSMENT
Figure 1-1: Relationship between the major factors in health risk assessment and the major components
of the RIHRA Program.
-------
RIHRA
Introduction - iv
biological measurement Ambient measurements
can be further subdivided into direct (e.g., indi-
vidual monitoring) and indirect (combining human
activity data with pollutant measurements in
important microenvironments) approaches. Im-
provements are needed in the quality and consis-
tency of data used to assess exposure with either
approach.
Topic 3: Physiologically Based Pharmaco-
kinetic (PB-PK) Models. The concentration of
a pollutant to which a human is exposed is usual-
ly not the same as the concentration of pollutant
delivered to the target organ (the dose). A
number of mechanisms, many poorly understood,
affect the transport of the pollutant through the
portal of entry (e.g., the lung for an inhaled
pollutant) to the target organ(s). In risk assess-
ments, however, the exposure concentration of a
pollutant is often used as a surrogate for the dose
concentration because the data on the latter are
not available. Better dose data will help reduce
the uncertainties associated with extrapolating
effects from one route of exposure to another,
from chronic to acute exposure, from high to low
exposure, and from one species to another.
While biological factors must also be considered
when assessing risk, first scientists must account
for the effects of the duration, magnitude, and
frequency of the exposure on the dose to the
target organ (i.e., pharmacokinetics). PB-PK
models can be used to establish a quantitative
relationship between exposure and dose delivered
to a target site in animals and humans under a
variety of conditions.
Topic 4: Biologically Based Dose-Response
(BB-DR) Models. A key element in the risk
assessment process is to estimate the incidence of
a specific health effect in human subpopulations.
Because human exposure data are usually not
available, scientists must extrapolate the risk of
health effects from data gathered with laboratory
animals. Research under this topic will be
heavily oriented to mechanistic studies aimed at
facilitating more biologically valid extrapolations
from animal data, and thus reducing uncertainties
in this step of the risk assessment process. The
development of accurate BB-DR models will
allow scientists to better define a quantitative
relationship between the dose of a pollutant and
associated health effects in humans.
Selection of Projects
To determine whether projects were appropriate
for inclusion in the RIHRA program, ORD applied
the following four criteria in the selection process:
1. Does the project focus on major significant
uncertainties and knowledge gaps in Agency risk
assessments? Important areas for research
include assumptions and extrapolations that are
used frequently and in which we have little
confidence.
2. Does the project have a reasonable probability
of success? The research must be technically
feasible in terms of existing expertise, resources,
and knowledge.
3. Would results of the research directly support the
needs of the Agency's risk assessors? Scientists
in EPA and elsewhere should be able to apply
the results of the research in risk assessments. In
aggregate, the projects were chosen to produce a
mixture of short- and long-term products for use
by the Agency in risk assessments.
4. Will the project data be amenable to wide appli-
cation? Results should be useful in a broad
context rather than limited to specific situations.
Ideally, the short-term products of this research
can be immediately applied to Agency needs and
can also be used as the logical building blocks
for long-term improvement of risk assessment
methodologies.
Furthermore, the projects chosen had to be
consistent with the EPA mission and represent the
kind of work that EPA and ORD are expected to
perform; that is, the projects must provide results that
are appropriate to EPA's legislative mandates and
regulatory authorities, and are useful to Program
Offices. These same criteria will be applied to the
choice of RIHRA projects in the future.
Program Implementation
The EPA Multimedia Research Committee—and,
more specifically, the RIHRA Subcommittee—has
formal responsibility for the RIHRA program. The
subcommittee, which is composed of the Office
Directors (and other key designees) from the partici-
pating ORD Offices (see Figure 1-2), designs, imple-
ments, and manages the program. In addition, the
subcommittee has two other tasks. First, it coordi-
nates the intramural (e.g., Program Offices, Science
Advisory Board, Risk Assessment Council) and
-------
RIHRA
Introduction - v
extramural (e.g., outside peer-review groups) reviews
of the RIHRA program, and ensures that the reviews
are used to strengthen and improve the research
program. Second, it has established working groups
for each of the major RIHRA research topics, com-
posed of one permanent member from each participat-
ing Office and a varying number of other ORD
scientists. The working groups integrate input from
their respective Offices to develop a focused research
plan (down to the project level) for each topic. These
research plans are reviewed and approved by the
subcommittee.
Implementation of the RIHRA research plans
under each topic highlights the following charac-
teristics of the program:
• RIHRA complements the ORD base research
program. The ORD base research program is
funded primarily through ORD Research Com-
mittees for each medium (e.g., air, water). These
committees, which are composed of ORD and
Program Office personnel, primarily support
targeted research on risk assessment issues or
chemical-specific topics to meet immediate needs
of Program Offices. The RIHRA program also
supports short-term research, but it concentrates
on longer-term efforts that may be augmented by
base program capabilities.
RIHRA comprises many cross-media projects.
The RIHRA program categorizes projects in a
hierarchy: projects fall under issues and issues
fall under topics. RIHRA research activities,
however, often contribute important data to
efforts under more than one topic (e.g., PB-PK
modeling as well as exposure assessment) and to
examinations of more than one medium. In
contrast, most projects developed under the ORD
base program are designed to answer needs of a
specific Program Office—and thus focus on a
single medium.
ORGANIZATION OF DOCUMENT
This document summarizes the research that is
both under way and proposed for the RIHRA pro-
gram. This summary is organized by the fundamental
topics chosen for the program: analysis of uncertainty
in risk assessment (Section 1), integrated exposure
Program Office:
Risk Assessment^
Council
Science Advisory Board
Multimedia
Research Committee
RIHRA Subcommittee
"(Participating ORD Office Directors)'
Extramural
Peer Review
Working Groups
(One Permanent Member/Office/Topic)
Topic 1
Uncertainty
Topic 2
Exposure
Topic3
Physiologically Based
Pharmacokinetic Models
Topic 4
Biologically Based
Dose-Response Models
Figure 1-2: Proposed implementation scheme for the Research to Improve Health Risk Assessments
(RIHRA) program.
-------
RIHRA
Introduction - vi
assessment (Section 2), physiologically based pharma-
cokinetic models (Section 3), and biologically based
dose-response models (Section 4). Each section
provides further detail and a research strategy for
work under that topic. Project descriptions are
included in the Appendix for each topic. Note that
program overviews for work in different fields (e.g.,
neurotoxicology, developmental toxicology) under
Topic 4 are provided in Appendix D.
-------
RIHRA
Topic 1 -1
TOPIC 1
ANALYSES OF UNCERTAINTY IN RISK ASSESSMENT
1.1 ISSUES COVERED
The importance of understanding uncertainty in
the estimates of risk produced by Agency risk asses-
sors cannot be overstated. Without such an under-
standing, risk assessors—and for that matter, risk
managers—fall prey to the "tyranny of numbers" that
can affect their ability to put risk characterization,
both qualitative and quantitative, in proper perspec-
tive. Such lack of perspective can lead to inappropri-
ate comparisons of risk across chemicals and can
result in faulty decision-making with regard to chemi-
cal-specific issues.
The issue of uncertainty analysis cuts across all
of the RIHRA topics; but for Topic 1, the specific
goal is the development of a research and data
collection strategy to deal with generic aspects of the
issue. A framework is needed to delineate the sour-
ces and magnitudes of uncertainty associated with
Agency risk assessments, and to explore various
approaches for describing or reducing these uncertain-
ties. Within this framework, researchers will examine
whether a particular uncertainty is of sufficient
magnitude to significantly impact regulatory decisions.
If it is, the type of uncertainty (e.g., extrapolation,
source assessment) will be described and, if possible,
quantified. Efforts will also include investigation of
what the end users of risk assessments (i.e., decision
makers) need with respect to assessment of these
uncertainties, both quantitatively and qualitatively. As
a result of this work, researchers will be able to
identify the sources of uncertainty in risk assessment
and to explore various approaches to reducing these
uncertainties.
Projects under this topic might include explora-
tion of possible approaches to analyzing uncertainty
that build on such formal disciplines as decision
analysis theory or Monte Carlo simulation. A number
of approaches have been recognized as potentially
useful, but further work is needed to select the most
appropriate ones.
task of identifying projects for potential funding in
FY89 and FY90. The workshop, which took the form
of a "Seminar on the Use of Analytical Methods for
Uncertainty and Sensitivity Analyses in Pharmaco-
kinetic Studies and Risk Assessments," was held at
the Dulles Marriot Hotel in Virginia on December 7
and 8,1988. Participants were informed of the Agen-
cy's intent to develop a framework for evaluating
uncertainties in risk assessments, and to identify
innovative research aimed at describing and/or reduc-
ing uncertainties. Presentations by ten invited partici-
pants were followed by a discussion session of
interested Agency participants.
A number of approaches to an overall framework
for classifying uncertainties were presented and
discussed. The consensus opinion preferred a scheme
presented by Max Henrion of Carnegie Mellon
University, which has now been adopted as a means
of stratifying uncertainties and categorizing research
proposals in this issue area. The scheme is based on
sources of uncertainty, as listed here:
1. Random error/statistical variation
2. Systematic error/subjective expert judgment
3. Disagreement between experts
4. Linguistic imprecision
5. Interindividual variability
6. Model approximations/model simplifications
7. Model structure/fundamental assumptions
In addition to discussing framework development,
Agency participants provided feed-back on specific
research proposed by the invited experts. This
process has resulted in the research strategy described
in the Section 1.3.
The elements of this framework allow for a
comprehensive evaluation of uncertainties in risk
assessments. In addition, each of these elements can
be evaluated with regard to available or proposed
research to address uncertainties that fall within it.
This approach leads naturally into the development of
a research strategy.
1.2 DEVELOPMENT OF A FRAMEWORK
In view of the breadth and depth of Topic 1, a
scoping workshop of active investigators in the area
of uncertainty analysis was set up to develop a
strategic framework; the group was also given the
1.3 RESEARCH STRATEGY
To develop a research strategy for uncertainty
analysis within the limited resources available under
Topic 1, ORD took into account ongoing work in the
-------
RIHRA
Topic 1-2
base program (see Introduction) and assessed which
projects would be best accomplished under other
RIHRA topics. Workshop participants preferred a
concentration on the development of structured
presentations of issues in uncertainty analysis and on
the elements of the framework related to use and
evaluation of expert judgment (#2 and #3), linguistic
imprecision (#4), the influence of interindividual
variability on uncertainty (#5), and implications of
model approximations/simplifications (#6). ORD
determined that work on random error/statistical
variation (#1) and on fundamental model structure
(#7) were more properly the purview of Topics 2, 3,
and 4 of the RIHRA program. These latter elements
of uncertainty are integral parts of data collection and
model validation, which are major components of the
research efforts under the other topics. Several
projects presented by experts at the workshop were
compatible with the expressed emphasis of this
strategy; others were judged not to be as pertinent.
From the ten proposals made at the workshop, those
pursued for FY89 are presented in Table 1-1. A
detailed description of Topic 1 projects is included in
Appendix A.
" " /"" T- TABLE 1-1 ' ' ""'
PROJECTS FUNDED UNDER TOPIC 1
ANALYSES OF UNCERTAINTY IN RISK ASSESSMENT
TOPIC/ PROJECT
ISSUE OFFICER
1,2
Farland
"Guidelines"
Farland <* - Addressing Uncertainties
for Exposure
Farland ^_ Interindividual Variabl^y
-"\ and Uncertainty
STATUS
Coftp* + National Academy
•of Sciences
Coop - Lawrence Livetutote
National Lab
Coop ~ Lawrence Livetniote
National Lab
Farland Evaluation of Communication Coop - Harvard
35
90
90
35
FY89
FY89
FY 89
*Coop = Cooperative Agreement
-------
RIHRA
Topic 2 -
TOPIC 2
INTEGRATED EXPOSURE ASSESSMENT
2.1 ISSUES COVERED
Information about the number of people exposed
to pollutant(s) and thek degree of exposure is an
essential component of any risk assessment Estimat-
ing human exposure, however, has proved difficult
and may be the single largest source of uncertainty in
environmental risk assessments.
Over the past several years, researchers at EPA
and elsewhere have assessed exposure and dose to the
target organ(s) in three ways, one direct and two
indirect. Direct techniques require real-time measure-
ments of contact intensity by personal monitors, such
as radiation badges, active devices that pump and trap
volatile chemicals, or chemical analysis of the
amounts and contamination levels of food and water
ingested. Indirect measurement can either be predic-
tive, using models for pollutant behavior and/or
human (ecological) behavior; or dosimetric (under
limited conditions), using body burden and knowledge
of pharmacokinetics to back-calculate the exposure
from observed levels in the body. All of these
methods are limited by associated uncertainties.
Dosimetry is discussed under Topic 3.
Predictive exposure assessment techniques are
appealing because, by using them, regulators can
evaluate the impacts of regulatory options on risk.
Exposure models can not only estimate concentrations
of pollutants in the various media, but can also link
particular environmental concentrations to exposure
levels in target populations with different activity
patterns. Such models are needed because humans
are mobile in the environment, and thus assuming a
constant level of exposure over time for an individual
or population is at best an approximation, and at
worst a major misrepresentation of real-time exposure.
In the past, insufficient data on human activities and
activity patterns and a lack of methods to incorporate
this information have been weak links in risk assess-
ment—and thus sources of great uncertainty. The
first two issuer discussed under this topic (2.1 and
2.2) therefore deal directly with reducing that uncer-
tainty through development of models that can accu-
rately predict exposure, taking human activity patterns
into account.
Measurements of source contributions to pollut-
ant levels have also been too limited. Pollutants have
customarily been measured only in carrier media (e.g.,
outdoor air, streams, soil, or food). This approach
relies on the assumption that limiting pollutants in
important carrier media will bring about the desired
protectiori of public health and welfare—for example,
that efforts should concentrate on reducing outdoor
sources of pollution to the point that observed outdoor
air quality meets national standards (Clean Air Act).
RIHRA exposure models will support an alterna-
tive approach, termed risk-based environmental
management, in which all sources that contribute to
exposure and risk are evaluated. In other words,
carrier media of long-standing concern (e.g., outdoor
air, drinking water) and sources of pollutants operat-
ing through these media are included in the strategy,
but so are building materials (e.g., adhesives, carpets,
floors), consumer products (e.g., furniture, sorbents,
pesticides), and other sources of pollution (e.g., soils
and dust, vehicular traffic) if these sources contribute
to the total risk from a pollutant Implementing risk-
based environmental management involves:
• Defining whose risks are going to be reduced;
the target, or receptor, of the risks could be a
human population, an animal species, or an eco-
logical system
• Determining the causes of the risk
• Reducing the risks to acceptable levels by con-
trolling the causes of the risk in a manner suited
to thek relative importance; for example, if
source A of a pollutant contributes 95% of the
risk and source B only 5%, then most control
efforts should concentrate on source A
This integrated exposure assessment approach
leads to improved efficiencies in reducing risk be-
cause it focuses resources on the sources of the
greatest amount of risk. Also, allocating risk reduc-
tion efforts by the relative contribution of each source
to total risk and assessing the cost to society of each
component of the risk reduction effort can increase
cost-effectiveness.
Another major source of uncertainty covered
under this topic is the means by which exposure data
are gathered in the field. For risk assessors to reach
conclusions with minimal uncertainty, data-gatiiering
methods must be appropriate to the use of the data,
whether in predictive assessments or dkect measure-
ments of exposure. For example, how do a series of
short-term duration exposure peaks differ in a risk
-------
RIHRA
Topic 2 - 2
context from lifetime long-term exposures at fairly
constant but lower levels? And how can this informa-
tion be best measured in the field and incorporated
into the exposure assessment? Efforts under Issue 2.3
address this concern, and are related to work proposed
under Issue 4.3 (which addresses how information
from different exposure conditions would be used in
a dose-response relationship). Taken together, results
from projects under Issues 2.3 and 4.3 will reduce
uncertainty in risk assessments by making the assess-
ments more closely describe real events.
A final major uncertainty source associated with
exposure estimation is the inconsistency in assump-
tions, methodology, and parameter values used by
various assessors for similar exposure situations. This
variation often leads to substantially different esti-
mates of exposure or risk for the same situation (for
example, from EPA and a company sponsored by a
potentially responsible party on a Superfund site).
Work under Issues 2.1 and 2.3 is improving EPA's
approach to exposure assessment by gathering data
that can be used consistently across many assessments
and by standardizing the approach to estimations of
population exposures for many commonly evaluated
situations (e.g., incinerators, waste sites, indoor air).
Issue 2.1: Human Exposure Models
ORD will develop and validate human exposure
models designed to generate realistic predictions of
exposure to chemicals using data on human activity
patterns and exposure sources. This type of model
links the concentrations of pollutants to which people
are exposed in various microenvironments (i.e.,
microenvironmental exposures) with the time spent in
those microenvironments (human activity patterns),
and integrates this, information with the pollutant
doses associated with the length and intensity of the
exposure. Human exposure models should address all
the microenvironments in which people visit or reside
(e.g., homes, stores, places of religious worship,
schools, buses, automobiles, workplaces), and must
take into account multiple possible routes of exposure
(air, food, drinking water). Indoor and in-transit
microenvironments are particularly important. Re-
search in this area will include developing these
models using existing data (2.1.1), and also validating
these models (2.1.2).
2.1.1: Development of Human Exposure Models
Research in this area will focus on constructing
human exposure models for important pollutants that
EPA regulates (e.g., respirable particles, volatile
organic compounds, semivolatile organics, formalde-
hyde) using the best microenvironmental data and the
best activity pattern data available. If microenviron-
mental data are unavailable, researchers will conduct
special field investigations to construct the needed
submodels. The human activity pattern and micro-
environmental concentration values will combine
human activity, time budget (see Issue 2.2), and
microenvironmental concentration data using a gener-
alized exposure equation. Researchers will incorpo-
rate suitable pharmacokinetic models into the expo-
sure models to estimate body burden and dose for the
regulated pollutants.
2.7.2: Model Validation
The human exposure models will be validated by
comparison of the resulting predictions to field data
collected in total human exposure field studies for a
variety of situations. Uncertainty ranges around the
predictions will be characterized. Then, teams will
develop guidance concerning the appropriateness and
the accuracy/uncertainty of the models for various
exposure situations.
Issue 2.2: Human Activity Patterns
Research efforts will characterize human activity
patterns in order to improve exposure analyses in risk
assessments. Human activity patterns, sometimes
called "time budgets," are records of what people do,
where, when, and for how long. Recent field studies
of human exposure, such as the Total Exposure
Assessment Methodology (TEAM) studies, have
shown that an individual's activities are critical in
determining exposure to environmental pollutants.
Once sufficient human activity data have been
gathered, the Agency will be able to realistically
predict individual exposures through the use of
probabilities. Significantly, the Agency will no longer
have to rely solely on a worst-case scenario in pre-
dicting a population exposure. Instead, with activity
pattern information, risk assessors can develop a
frequency distribution for exposure, and thereby
support predictions concerning "average" exposure,
exposure in the 90th percentile, and so on.
Issue 2.3: Database on Indirect
Exposure Parameters
Researchers will expand the database on the
parameters used to make indirect exposure estimates
and will clarify how to use these estimates. More
information is needed on the ranges and distributions
of parameters used in .indirect exposure assessments,
-------
RIHRA
Topic 2 - 3
such as ingestion rates, exposure durations, contact
rates, and short-term versus long-term exposures. The
Regions have repeatedly requested guidance on such
topics to improve consistency in risk assessment.
Guidance is also needed on ways to apply these
factors in creating different scenarios, such as a
typical and a worst-case exposure level.
Two concerns that will receive special attention
under this issue relate to 1) predicting the incidental
soil ingestion of children based on site-specific
parameters such as ground cover and weather, and 2)
determining the contribution of short-term peaks to
total exposure. Soil ingestion by children, an activity
that has stirred significant controversy regarding its
importance, is an important exposure route in many
risk assessments. For many pollutants, short-term
peaks are extraordinarily important contributors to
exposure. Besides their contribution to exposure,
short-term peaks affect the dose-response portion of
risk assessment (see Issue 4.3). Teams will pursue
methods to measure short-term peaks and then use
this information to establish dose-response relation-
ships.
2.2 GATHERING EXPOSURE DATA
The ORD base exposure program (see Introduc-
tion) has emphasized these research areas:
• Developing and evaluating methods for conduct-
ing total human exposure studies for several
pollutants or pollutant classes (e.g., volatile
organic compounds [VOCs])
• Developing instrumentation
These two efforts together form the scientific basis
from which to develop the exposure model framework
(Figure 2-1). Embodied within this framework is the
monitoring systems approach, which includes moni-
toring and statistical design, quality assurance objec-
tives, information requirements, Office of Manage-
ment and Budget (OMB) clearance, and information
assessment
To fill the critical information gaps that remain
in exposure assessment, further efforts in both direct
and indirect measurements must be made. Only with
HUMAN
EXPOSURE
PREDICTIONS
WITH KNOWN
PRECISION
MODEL
VALIDATION
MICROENVIRONMENTAL
SUBMODELS
M ICROENVIRONMENTAL
DATA
CO
TEAM
voc
TEAM
CHARACTERIZE
UNCERTAINTY
TOTAL EXPOSURE
MODELS
ACTIVITY
PATTERNS
NOPES
TEAM
PART4CLE
TEAM
OTHERS
Figure 2-1: Exposure assessment framework.
-------
RIHRA
Topic 2-4
this two-pronged approach will it be possible to
estimate total population exposures to environmental
pollutants with a known accuracy and precision.
These estimates are essential for the Agency to
accurately prioritize the chemicals requiring assess-
ment and to effectively apply a risk-based manage-
ment strategy in the face of multiple pollutant sources
and multiple pathways of exposure.
To generate the exposure and dose estimates
necessary for evaluating the routes of exposure by
which pollutants reach human beings, appropriate
research programs as well as methodologies to assess
individual pollutants or classes of pollutants must be
developed. The first of these methodologies is
virtually complete (for carbon monoxide), and two
others are partially developed (for VOCs and pesti-
cides). Many more, however, must be developed—
for example, for particles, organics on particles, polar
organic compounds, and other regulated pollutants.
At the basis of an integrated exposure assessment
methodology is the idea of a three-dimensional bubble
around the target human being. If a pollutant in one
of the four possible carrier media makes contact with
this "bubble," the target is said to have experienced an
"exposure" to that pollutant at that instant of time
(Figure 2-2). This instantaneous exposure is ex-
pressed quantitatively as a concentration (mass/vol-
ume) in a particular geophysical medium (mass) at a
particular instant of time (time units); and the average
exposure is the average of the concentration at the
surface of the bubble over some appropriate averaging
time. Some pollutants can reach humans through
only one carrier medium (e.g., carbon monoxide
through air). Others, such as lead and chloroform,
can reach humans through two or more media. When
multiple pathways of exposure are involved, a per-
son's exposure is the sum of concentrations of the
pollutant in each carrier medium at a particular instant
of time.
AIR
WATER
GUT
FOOD
SOIL
Figure 2-2: Conceptual model of total human exposure.
-------
RIHRA
Topic 2 - 5
Once implemented, an integrated exposure
methodology for a given pollutant would be used to
provide information on the exposures of the target
population through all environmental media regardless
of the route of exposure (i.e., inhalation, ingestion, or
skin absorption). These data will elucidate the
number of people exposed, their levels of exposures,
and the pollutant sources or other contributors respon-
sible for the exposures. Studies completed over the
last few years underline the value of the integrated
exposure approach: the findings have already had a
great impact on the Agency's policies and priorities in
relation to most critical pollutant concerns.
Making the human target the crux of the total
exposure model is an innovative approach. Using this
methodology, ORD scientists first consider all routes
of exposure by which a pollutant may reach a human
target, then focus their effort on those particular
routes relevant for a particular pollutant of concern.
With this structure, they can gather accurate informa-
tion on the concentrations present and the movement
of pollutants through the exposure routes of utmost
importance.
Direct and Indirect Measurements Used in Inte-
grated Exposure Assessments
The complementary role of direct and indirect
measurements in estimating nationwide human expo-
sures to pollutants is at the basis of any integrated
exposure assessment methodology. Direct measure-
ments of the exposure of a target population to a
pollutant are taken as follows:
• Selection of a representative and sufficiently
large random sample of the population on the
basis of a carefully planned statistical design
• Measurement of the pollutant concentrations
reaching the population sampled through all
relevant environmental media
• Development of estimates, with known precision,
about the exposures of the larger population from
which the sample was drawn
• Measurement of body burden (breath or blood
concentrations) to verify the exposure estimates
and infer a dose
• Statistical analysis of information from the
respondents' diaries (e.g., activities and locations
visited)
• Identification, if possible, of likely sources,
microenvironments, and human activities that
contribute to exposures
This approach to exposure measurements (i.e.,
TEAM) is invaluable for determining exposures and
sources of exposure for the specific population
sampled. EPA has completed a number of successful
TEAM field studies addressing carbon monoxide,
VOCs, airborne paniculate matter, and pesticides.
Other such studies are underway.
Indirect measurements, on the other hand, are
used to extrapolate data to much larger populations
than can be measured directly. Exposure modeling is
used to measure and understand the basic relation-
ships between causative variables and resulting
exposures, usually in particular microenvironments.
These models incorporate data collected in the field,
and the exposure predictions derived from them are
intended to supplement results from direct studies and
to extend and extrapolate these findings to other
locales and other situations. Exposure models are not
the traditional dispersion models for predicting
outdoor concentrations; they are instead designed to
predict the exposure of a mobile human being to a
particular pollutant. Thus, required model inputs
include:
• People's typical activities and time budgets
» Likely pollutant concentrations in places where
the people spend time (e.g., home, work, school,
in transit)
• Amounts of food and water consumed
• Likely concentrations of the pollutant in question
in food and water
An example of a recently developed exposure
model is the Simulation of Human Air and Pollution
Exposure (SHAPE) model, which was designed to
predict exposures of urban populations to carbon
monoxide. This model, which is similar to the
National Ambient Air Quality Standards Exposure
model, is based on the carbon monoxide concentra-
tions measured in a Washington, DC - Denver,
Colorado, study that assessed the relative contribu-
tions to exposure from commuting, cooking, cigarette
smoke, as well as other factors. A model such as
SHAPE can be validated by showing that it accurately
predicts exposure distributions as directly measured,
such as by a TEAM field study. After this validation,
the model can be used in a new city—unaccompanied
by a field study—to predict that population's exposure
to the pollutant in question (the new city's data on
-------
RIHRA
Topic 2-6
human activities, travel habits, and outdoor concentra-
tions are used as input). In the future, models such as
SHAPE are planned for application to pollutants other
than carbon monoxide (e.g., VOCs, household pesti-
cides) and for other routes of exposure. The goal of
this effort is to estimate total exposure frequency
distributions for the entire country, or for major
regions.
2.3 PROJECT SELECTION STRATEGY
The general selection strategy for RIHRA pro-
jects, described in the introduction to this document,
provided the framework from which more specific
project criteria in each research area could be devel-
oped. For exposure-related activities, the selected
projects had to meet these additional criteria:
" A long-term direction and goals consistent with
RIHRA objectives
» A cohesive, logical, defensible design
• Likely production of results that would lead to
iterative improvements in risk assessment
• Links to the base exposure program (see Intro-
duction)
» Contribution, along with other projects, to a
reasonable balance across the RIHRA issues,
routes of exposure, and sources of exposure
Table 2-1 provides information on the funded pro-
jects.
Under Issue 2.1, researchers will formulate
exposure distributions for input into biological human
models (see RIHRA Topics 3 and 4). First they will
construct submodels for classes of pollutants the
Agency regulates by integrating:
• Human activity data
• Existing microenvironmental data
» Source-specific information to account for source/
activity combinations with expected high expo-
sures; this information includes exposures to tradi-
tional or regulated sources such as incinerators or
waste dumps as well as exposures through use of
consumer products
Model validation efforts, proposed for FY90-91,
will involve the following tasks:
• Comparison of data with exposure distributions
derived from total human exposure studies
• Special targeted field studies to identify (finger-
print) sources
• Comparison of data with reconstructive bio-
marker techniques
• Estimation of the transportability of each sub-
model for estimating exposures with different
exposure or source scenarios
The models will include equations that predict the
total exposure resulting from multiple routes. One of
the routes that will be examined is the passage of
toxic incinerator emissions through the air.
Under Issue 2.2, researchers will compare the
activity profiles of populations examined in TEAM
studies or other surveys to determine the variation
among sampling years and seasons and between
geographical regions. This work will be used to
assess whether human activity pattern information
from one place and time can be used to estimate
exposures in a different time and place.
In addition, activities under this issue should
show the extent to which existing national activity
data can be used for exposure assessment. For
example, these activity data should be valid indicators
of total time spent indoors but cannot be used to
estimate the time spent in activities associated with
high exposures (e.g., time spent using specific con-
sumer products). The end product of these projects
will be a representative national exposure profile data-
base. That database will then be used as input for the
human exposure model; model users could, for
example, pick from menus the data from a representa-
tive activity/microenvironment/subpopulation that
should be included in a particular estimate of total
exposure.
Projects under Issue 2.3 address specific informa-
tion gaps that, according to preliminary information,
are sources of variation in exposure estimates. Data
on ingestion and short-term peak exposure will be
used to refine the submodel components of the
exposure model. Researchers will also develop
specific detailed exposure distributions for input into
pharmacokinetic models, and these data will also be
used to interpret the relationship between short- and
long-term exposure.
Work under this issue will examine the frequency
and impact of pica behavior in children six years old
and younger. Other efforts will study the contribution
-------
RIHRA
Topic 2 - 7
of household dust and soil ingestion to the exposure
of children to metals and pesticides; the data gathered
will stand as direct measurements of source contribu-
tions that had previously only been modeled. Re-
search in these areas is intended to mitigate an
important limitation of past exposure assessments,
particularly related to soil cleanup activities, in which
large uncertainties regarding household dust and soil
ingestion were clearly but unavoidably ignored.
Researchers will also evaluate the contribution to
exposure associated with garages in private homes,
which have been identified as pollutant sources for
organics in indoor air and as possible sources of
polycyclic aromatic hydrocarbons. This latter class of
chemicals enters homes through the air (doors and
cold air intakes) and may also be tracked in on
clothing and shoes. The extent to which this source
adds to a person's total exposure will be assessed
through source characterization of garages and adjoin-
ing rooms in the residential environment. Again, data
from this work will be used to refine the submodel
components dealing with air and dermal/ingestion
exposure.
A detailed description of Topic 2 projects is
included in Appendix B.
TABLE 2-1
-
PROJECTS FUNDED UNDER TOPIC 2 ^
.EXPOSURE ASSESSMENT
TOPIC/ PROJECT
ISSUE OFFICER
2,1.1 Irwto/AREAL
Behar/EMSL-LV
2.1.2 - Schaum/OHEA
2.2.1 Behar/EM$L-LV
2.2.2 Nelson/AREAL
2.3.1 White/OHEA
2.3,2 White/OHEA
2.3.3 Bond/AREAL
2.3.4 Evans
2.3,5 Highsmith/AREAL
SHORTTTTLE
Model Refinements
Benzene Prototype
Procedures for Evaluating Multi-
media. Exposures to Incinerator
Stack Emissions
Comparative Analyses of Existing
Activity Pattern Databases
Develop National Database of
Activity Patterns for Exposure
Assessment
Pica Ingestion Rates in Children
Selection of Food Consumption
Rates
Measurements of Exposures to
House Dust and Soil Ingestion
Short-term Spatial Variation
of Aerosols and Organics
Exposure from Attached Garages
FY89
STATUS ' ($K.) DURATION
Coop* -Harvard Univ. 100 FY 89^1
UNLV
Contract: Technical 125 FY.89-91
Resources, Inc..
Contract: U. MD ,,, 150 FY 89-91
Coop-U. MD 150 FY$9-91
/•••.• **
Coop Fred Hutchinson 125 FY 89-91
Research Institute
Pending 0 FY 90-92 ,
Contract: Research 150 FY 89
Triangle Institute
Contract: Research 125 FY 89
Triangle Institute
Coop - Rutgers 75 FY 89
*Coop = Cooperative Agreement
-------
RIHRA
Topic 2 - 8
Future Directions
Proposed RIHRA exposure activities will be
judged by their contribution to a generic model that
can predict human exposure profiles through all media
using human activity patterns and measured (or
modeled) concentrations in significant microenviron-
ments. The iterative process of validating the model,
characterizing its performance, and reducing the
associated uncertainties will continue until the uncer-
tainties of the model results are not relatively signifi-
cant to the overall uncertainties of the health indica-
tors. RIHRA offers researchers the opportunity to
draw on results of the base exposure programs,
formalize these results into an exposure model, and
then validate the final product with directly measured
personal exposure distributions.
-------
R1HRA
Topic 3-1
TOPIC 3
PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS
3.1 ISSUES COVERED
Physiologically based pharmacokinetic (PB-PK)
models provide a bridge between exposure data
(Topic 2) and biological outcome (Topic 4). In
assessing dose and the effects of a given pollutant
dose at different levels (e.g., molecules versus tissues)
within the body, researchers commonly extrapolate
from experimental to real-life human dose levels.
The most critical uncertainties associated with this
process result from assumptions concerning:
• The comparability of exposure by different routes
of administration (route to route)
• The comparability of dose delivery for different
regimes of exposure, such as repeated versus
single dosing, or episodic, peak, and chronic
exposures totaling the same cumulative dose
(chronic to acute)
• The proportionality between external exposure
level and the resulting delivered dose for higher
exposure studies, compared to lower levels
typical of environmental exposure (high to low
dose)
• Scaling or translation of dose to determine
exposures yielding equivalent doses in different
species, especially when extrapolating toxic
effects in experimental animals to those expected
in humans (species-to-species)
PB-PK models are abstract: they represent a
system of postulates, data, and inferences presented
mathematically to describe the kinetics of the uptake
and distribution of chemicals both in laboratory
animals and in man. The key physiological variables
used in these models are based on underlying biologi-
cal processes and on the physics governing transport
kinetics; thus, using these models can lessen the
uncertainty associated with extrapolating from dose to
dose or route to route. In addition, species differ-
ences are incorporated in the model parameters, so
model users can more accurately extrapolate from
data gathered from laboratory animals to expected
response in humans.
To improve the accuracy of risk assessments,
better dose data are needed for input into these
models. Such data are necessary to reduce the
uncertainties associated with extrapolating from dose
to dose, route to route, or species to species. In the
past, measurements of applied dose (or exposure
concentration) have been used instead of data on
delivered dose (the dose delivered to the tissue or
organ of toxic action). The duration, magnitude, and
frequency of an exposure, however, greatly affect the
delivered dose; and thus the two measurements are
not interchangeable.
Examining dose levels, however, does not answer
all questions about extrapolation in risk assessment:
the equivalency of effects across species is deter-
mined not only by relative dose delivery but also by
any species differences in reactivity or susceptibility
to a given delivered dose. Similarly, the extrapolation
of effects from high to low doses or from acute to
chronic exposures depends not only on the delivered
dose differences in these circumstances, but also on
the relative toxicological effects of different degrees
and durations of tissue exposure to the proximate
toxicant.
Attempts to determine patterns of delivered dose
at particular tissue sites focuses attention on the need
to understand the mechanisms of action involved. For
example, the target site at which the delivered dose
should be described may not be clear. Depending on
the mechanism of action, the toxic response may be
a function of quantity of metabolite formed, the
number of adducts or other covalent reaction products
formed with crucial cellular macromolecules, the
extent of reversible binding to specific receptors, the
average and peak toxicant concentration, or duration
spent above a crucial concentration. There may be
countervailing repair processes of limited or perhaps
saturable efficacy. These other factors can and
probably do vary among species, strains, sexes,
previous histories of exposure, and physiological
condition of the subjects.
Thus, although gathering data on dose levels can
reduce a great deal of the uncertainty associated with
extrapolative procedures, other biological factors must
be addressed. Before this step can be taken, however,
researchers must account for and eliminate the con-
founding and obscuring effects caused by dose
delivery and pharmacokinetics. Progress in pharma-
cokinetics is central to the development of biological-
ly based risk assessments.
-------
R1HRA
Topic 3 - 2
Issue 3.1: Experimental Absorption and Biological
Parameter Data
To improve risk assessments, ORD must gather
more experimental and physiological data relative to
the effective dose responsible for biological effects.
Often both the amount and the biologically active
form of the toxicants are unknown. Without data on
the absorption, metabolism, transport, and elimination
of the parent compound and its metabolites, risk
assessors must use uncertainty factors that lessen the
accuracy of their assessments. Gathering these
experimental data can be coupled with studies on
toxic mechanisms of action.
5.7.7 Experimental Absorption Studies
Researchers must balance efforts to develop
theoretical models (see below) to predict the uptake
and distribution of chemicals with experimental
studies in which the uptake, distribution, and time
course for elimination of the parent chemical and
various metabolites are determined in the major body
organs for various species. Moreover, experimental
dosimetry studies will support the guidance and
design of biological experimentation in a number of
organ systems (e.g., reproductive, nervous, pulmo-
nary). Teams should use an integrated approach to
encompass oral, dermal, and inhalation exposures for
adult, neonatal, and fetal animals, with a linkage to
human studies where possible.
3.7.2 Physiological and Anatomical
Parameters Across Species
Physiologically based pharmacokinetic modeling
provides this key advantage: the model structure is
common across species; only the scale is changed.
The model developed for one species (along with the
attendant insights about delivered dose) can be
applied to another species if the various physiological
and anatomical parameters governing the kinetics of
the compound in the first species are replaced appro-
priately. Physiological parameters describe capacities
and volumes (e.g., organ weights, blood volume,
partition coefficients, lung capacity) and rates (e.g.,
blood flows, metabolic rates, ventilatory rates, elimi-
nation rates). Anatomical parameters describe the
structure of an organ, such as the size and number of
airways in the lung, and the number of glomeruli and
the structure of the nephrons in the kidney.
Once most of these values are determined, they
can serve as input data in the construction of models
for many compounds. Caution must be exercised,
however, when making generalizations if any evi-
dence suggests that the chemical being modeled can
itself influence the various physiological and anatomi-
cal parameters that help to define the model. Meta-
bolic parameters can be very chemical-specific, but
even these can benefit by characterization of major
biochemical pathways for metabolism of xenobiotic
compounds.
3.1.3 Influence of Varying Exposure Parameters
(Route, Duration, Rate) on Delivered Dose
Issue 3.1.1 describes experimental dosimetry
studies that can be associated with health effects
studies. High- to low-dose extrapolation may be
highly uncertain due to a lack of pharmacokinetic and
pharmacodynamic data. Also, one of the extrapola-
tions that must be made in quantitative risk assess-
ment is from the experimental dose regimen used in
an animal lexicological study (repeated dosing or
chronic exposure, usually for extended periods) to the
expected human exposure patterns (which may be
single exposure or chronic, episodic, or continuous).
The doses are usually compared on a total cumulative
dose basis, e.g., the total mg/kg or ppm/h of expo-
sure. However, both dose rate and dose level can
affect the pharmacokinetics of a compound and
hence the amount that is delivered to the target site.
For example, high dose levels may activate pathways
that at lower dose levels do not contribute to meta-
bolic conversions which are linked to the toxicity of
the compound.
As work progresses under Issues 3.1.1 and 3.1.2,
researchers will begin to examine the influence of
route, duration, and rate of exposure on delivered
dose in order to understand the uncertainties inherent
in extrapolating lexicological data obtained using one
exposure scenario and species to another. With data
from such projects, teams will develop guidance for
improving risk assessment methodologies.
Issue 3.2: Route-to-Route Exposure
Researchers need to identify the assumptions and
conditions that affect the scientific defensibility of
route-to-route extrapolation. For many chemicals,
risk assessments are not available for key routes of
exposure because the pharmacokinetic and toxicologi-
cal data needed to make these assessments are also
not developed. What may be available instead is
information on the lexicological effects of a chemical
associated with a different route of exposure. By
route-to-route extrapolation, scientists can use the
available data to predict exposure by the route of
interest PB-PK models are the best means of gelling
ihe maximum use of such data.
-------
RIHRA
Topic 3 - 3
Prior to expanding current efforts on route-to-
route extrapolation, the Agency will conduct a sym-
posium/workshop to 1) discuss the critical assump-
tions and limitations related to route-to-route extrapo-
lation; 2) provide specific guidance for risk assess-
ments; and 3) recommend research areas that would
facilitate route-to-route extrapolation. Such an effort
represents an extension of past activities of organiza-
tions like NAS to facilitate the use of pharmacokinet-
ics in risk assessments. Since various EPA Program
Offices extrapolate oral toxicity data to inhalation
reference doses, the meetings will focus on this issue.
Also, dermal versus oral absorption extrapolations are
a priority. ORD will seek Agencywide participation
in identifying questions for the meetings. Recommen-
dations arising from this exercise will be incorporated
into ongoing ORD research programs.
Issue 3.3: Theoretical Models
ORD scientists will explore theoretical models to
serve as a unifying structure on which intra- and
interspecies dosimetric comparisons can be made.
Development of theoretical PB-PK models for
chemicals will facilitate better estimates of dose-
equivalence across species for various rates and
durations of exposure. The development and model-
ing of pharmacokinetic information from various
species will enable researchers to determine the
pharmacodynamic differences among species. Im-
proved mathematical formulations will be derived for
the disposition of compounds following oral, dermal,
or inhalation exposure that incorporate both age- and
species-specific input parameters, such as partition
coefficients and blood flow, and the properties of the
molecules being considered. With validated PB-PK
models, researchers can scale exposure, dose, and
effects observed in one circumstance (e.g., in labora-
tory animals in a controlled environment) to com-
pletely different circumstances (e.g., human beings in
an uncontrolled ambient environment).
3.3.1: Theoretical PB-PK Models
With theoretical PB-PK models, researchers can
make intra- and interspecies dosimetric comparisons.
Although specific chemicals are often restricted
environmentally to only one or two exposure routes,
models are needed for all routes of exposure (oral,
dermal, and inhalation). A major impetus for the
application of pharmacokinetics to risk assessment is
the suspicion that experimental rodents may exhibit
quite different degrees of delivery of an applied dose
to the site of action than humans. This difference
would be due to the physiological and metabolic
processes in small mammals that occur at much
greater rates than in humans. While pharmacokinetic
differences are not the factor affecting a chemical's
potency in various species, extrapolation procedures
should account for differential dose delivery.
3.3.2: Structure-Activity Relationships in
Mechanistic Models
The actions of xenobiotic agents in a biological
system are a direct consequence of their molecular
properties and are produced by a variety of specific
molecular interactions and non- specific processes.
These agents may interact with receptors, enzymes,
and macromolecules involved in transport by fully or
partially mimicking the relevant properties of endoge-
nous substances or by evoking the detoxification
potential of the biological system. The scientific
basis for predicting risk to human health from specific
chemicals will be enhanced by pharmacokinetic
models that incorporate not only the physiological
parameters of the system being modeled, but also the
molecular structure and reactivities of the substances
involved. The latter information will provide impor-
tant insight into the transformation, distribution, and
deposition of the chemicals under consideration
because these properties depend on the biological
activity of the chemicals. Additionally, because the
potential risk from a chemical must often first be
assessed without all the relevant data, a modeling
approach that incorporates molecular properties has a
significant advantage.
3.3.3 Models Linking Exposure to Dose to
Biological Outcome
Even after estimates of delivered dose have been
made, a great deal of uncertainty still surrounds the
incorporation of these estimates into the quantitative
risk assessment methodology. The delivered dose
information will aid in the extrapolations, but factors
outside of pharmacokinetics must also be considered.
Species, for example, may have different degrees of
response to a given delivered dose—not only because
of possible idiosyncratic differences in defenses, but
also because the physiological processes affected by
the toxic agent are themselves subject to scale differ-
ences in different-sized mammals. When focusing
on the "biologically effective dose," scientists must
also incorporate knowledge of the biological reaction
to that dose to make proper use of the pharmaco-
kinetic data. Projects under this issue will examine
the interactive role of pharmacokinetic and phar-
macodynamic factors. The activities under this issue
are linked to those under Issue 4.2.3 (Interaction of
Exposure Parameters on Outcome).
-------
RIHRA
Topic 3-4
3.2 PROJECT SELECTION STRATEGY
The rationale and general strategy that drives
research in the area of PB-PK models has been
outlined in the Introduction. For PB-PK model-
related activities, the selected projects had to meet
these additional criteria:
• Long-term direction and goals consistent with
RIHRA objectives
• A cohesive, logical, defensible design
» Likely production of results that would lead to
iterative improvements in risk assessment
» Cohesiveness of the experimental and theoretical
modeling components of the program
• Appropriateness for the desired program mix of
projects (i.e., balance of highly relevant and
focused short-term projects with immediate
delivery of products and of projects that address
more generic and long-term RIHRA goals)
The projects can be evaluated on the basis of these
criteria as well as the strategy outlined in the Intro-
duction. Table 3-1 provides a listing of projects for
FY89 by RIHRA issue.
Under RIHRA Issue 3.1, Experimental Absorp-
tion Studies, research efforts provide a balanced
approach encompassing oral, dermal, and inhalation
routes of exposure. The ORD base program (see
Introduction) is surprisingly deficient in experimental
pharmacokinetic studies. To meet this recognized
need, a Dosimetry Branch was proposed as part of the
reorganization of the Health Effects Research Labora-
tory (HERL). Some of the work under this issue
explores research areas that were already being
investigated by the base program, and these efforts
were redirected under RIHRA support The ongoing
research of the Office of Health and Environmental
Assessment (OHEA) into the dermal permeation of
compounds bound to soils will also be complemented
by RIHRA support. Finally, some research under this
issue focuses on obtaining specific PB-PK data for
compounds with known multispecies carcinogenic re-
sponses.
, ", , TABLES.!
.. r , PROJECTS FUNDED UNDER TOPIC 3
totof'VxST^rtlt Oi'iff*1 A t" T V1 R AQirn PITA Ditjf A /*v?iv1rKTti*rf'iiVT *n*r/"irvEiĄ' 6 "
^TJpl l3pi.^Lt*_/uJ.V'O1>1/ i|rjtlJ\KJW./\^ljrfV|.|NlJ/jmr, IVIL/LJJ&4jo
TOPIC/
ISSUE
34,1
34.2
3,2
334
3.3.2
333
*Coop =
PROJECT
OFHCER
•>
Hoang .
Hall * t
Gerrity
Gerrity
Gerrity
White
Farlandt
Gerrity
Gerrity
Gerrity
White
Rablnowitz
Kavlock
Beliles
s ••
"• ** v
SHORT TITLE , ^
• « • '
. permeation Coefficients
Dermal Absorption
• Particles. Deposition/Clearance
Dosimetry of Aqueous Aerosols
PK of Inhafedflngested VOCs
PK and Known Multispecies
Cancer Potency
• Data Base of Biological Parameters
Rwte-to-Route
Inhalation RFD Methodology
Development PB/PK Model*
Model Validation arid Sensitivity
Computational Models
•s
Embryonic Dosimetry s %
, Biological Time for Teratogeaesis
Cooperative Agreement
f- ••' '
•>
'STATUS
^ •••-. ;/' -
Coop* - UCSF
Coop - UCSF Contract NSJ
Coop - Univ. NL Cardlma
'On-Site EMSl
Coop - Univ, Georgia
To be determined'
-y
To be determined
On-Site >fSI
On-Site NSI '
Qn-Site^NSI
Coop -Univ. Delaware
On-Site EHRT
Coop - Molecular Resv Ihst
Coop^NCSU
Coop - Childrens Hosp-Ctn
f ft f f f, <• t f r '*..
'
FY89
<$K)
$0
330
130
100
117
0
15
7$
0
240
50
150
200
70
.,
DURATION
FY 89-91
FY 89-91
FY 89-91
FY 89-91
FY 90-92
"
FY90
FY 89-90
FY 90-92
FY 89-91
FY 89-91
FY 89-91, ,
FY89-9J
FY 89-90 "
....
-------
RIHRA
Topic 3 - 5
Prior to the commitment of large amounts of
resources to Issue 3.2, a scientific consensus must be
built regarding the priority research concerns. Con-
troversy surrounds a number of questions related to
the issue of route-to-route extrapolation:
• How and when is it appropriate to use oral
toxicity data to establish an inhalation reference
dose?
• What are the assumptions that must be met for
dermal to oral extrapolation?
• What minimum data are needed to defend the
assumptions that support certain extrapolations?
Thus, RIHRA efforts under this issue are devoted to
convening workshops and expert panels to prioritize
concerns and to provide research recommendations.
Research projects under Issue 3.3 have been designed
to complement the experimental studies outlined
under Issue 3.1 (i.e., model formulation and validation
projects that are designed to build from the experi-
mental data gathered under Issue 3.1). Figure 3-1
depicts the three levels of research associated with
developing PB-PK models. Such models are used to:
» Estimate delivered dose
• Perform reconstructive exposure assessments
• Where appropriate, apply the results of biomark-
er studies to quantitative exposure and risk
assessments
• Assist in making route-to-route, dose-to-dose,
and scenario-to-scenario extrapolations
Level I contains specific experimental issues (see
projects under Issue 3.1) that must be resolved as a
basis for pharmacokinetic model formulation and
testing. To provide better input for PB-PK models,
needs such as the following must be addressed:
• Better and more detailed quantitative descriptions
required of the absorption processes at the major
portals of entry (e.g., developing good definitions
of dermal absorption and how to determine key
parameters in dermal absorption)
• Better route-specific descriptions required to
increase confidence in extrapolation
• Examination of existing methods for determining
model parameters (e.g., partition coefficients,
permeation constants, and metabolic rate con-
stants) and guidance for choosing the best meth-
od on a case-specific basis
• Determination of the most precise procedures for
extrapolating parameter values from one species
to another
Level II comprises efforts to formulate a model
and apply it to simple exposure scenarios. During
this model synthesis phase (see projects under Issue
3.3), the following steps would be taken:
• Establishment of criteria for model testing and
validation
• Extrapolation of models for interspecies compari-
sons (building on Level I efforts)
• Determination of different species parameters by
allometric scaling or use of accepted values
• Determination of parameters that do not lend
themselves to allometric scaling (e.g., metabolic
rate constants) on a case-by-case basis
• Development of criteria and guidance for the
interspecies transfer and application of phar-
macokinetic models
• Beginning development of higher-resolution
models (e.g., for estimating proximate dose near
the effective site of action) and of pharmacody-
namic models (e.g., for describing not only the
concentration of putative toxin but also the
alteration of physiologic processes such as
neurotransmitter blockage)
At Level III, efforts will be focused on applying
the outputs of the first two phases in exposure and
risk assessments. In particular, the methods will be
tried in more complex exposure scenarios, such as
involving mixtures; these tests should highlight
additional issues for resolution.
The Agency does not have to prioritize projects
at these three levels of model development for com-
pletion in a specific order. For example, pharmacoki-
netic models have already been applied to risk and
exposure assessment processes in several situations;
and, as a result, several of the issues outlined here for
Level I and II research were identified. When these
studies are completed, the methods used in these
previous assessments will be modified as appropriate.
Model research and application is thus an iterative
process, not a sequential passage from level to level
with a fixed end product.
-------
RIHRA
Topic 3 - 6
Significant effort in the base research program
has already been devoted to PB-PK modeling for
inhaled compounds. Other routes of exposure,
however, are under-represented in the program, and
only a few HERL scientists have the relevant experi-
ence to develop a coherent program in the missing
areas. Research, therefore, will provide resources to
fill this void through the on-site contractor to HERL.
Advantages of this approach include:
• Cost-effectiveness of using the staff depth and
capabilities presently available through the
contractor (i.e., whole teams of biophysicists,
computer programmers, pharmacologists, and
lexicologists are needed to develop PB-PK
models)
" Importance of the Dosimetry Branch proposed as
part of HERL as a focal point for the laboratory
on PB-PK research (i.e., because information on
dose to target will then be available for observed
biological effects, future health effects studies by
HERL will become even more useful to risk
assessors)
Investigators will also examine structure-activity
relationships in mechanistic models. Efforts in this
area can help synthesize the assessment of risk
through the examination of issues related to the
toxicity of chemical classes. The resources allocated
to this research area reflect an increased level of
effort over base activities, but work under RIHRA
will largely deal with classes of compounds that are
not presently addressed under the base program.
Efforts in PB-PK modeling and structure-activity
relationships can also be synergistic: advances in one
area can help to focus research issues and objectives
in the other. For example, once a structure-activity
relationship is established between a tested and
untested pollutant, predictions can be made regarding
the latter's activity (e.g., partition coefficients) that
are based on chemical structure; and once established,
these estimates can then be incorporated into a PB-PK
model.
To underpin the PB-PK models being developed,
researchers will synthesize and develop mathematical
and physiological bases for model validation, formula-
•• wv. 5. ' •• « %
•• ^, '• ' J
Uncertainty '" .Reconstructive exposure
analyses ^kV^mentandbfemartefs
of exposure
-- ,,,v " tevei); Specific
PB-PK pgrameter experimental
,s ; 'determination absotbtion studies,
-. ^ tnterspeciesexfrcpolation of PB-PK
- ' data and parameters
issues for Model Development
s^ S5 ••
> *V' ", ^ j, s
PB-PK modeE ; •. . Criteria for mode! ' ' M^ application: to Higher resolution
forroidafion v;^ testing and Devaluate different ^models to estimate
techniques : vaJidatfort , scenarios of exposure " proximate doses
• • - ,,\ •''•'< 'r '' - , """ ' near toe effective
,,' - -' "' '" ' ' site of action
- ; steyel )f: Mode) Synthesis and Application Techniques
lr - ' ' ' ,, v
Model applications to twTp
evaluate the Impact o!
jtiixtures and complex
exposure scenarios on s
exposure and risk assessments
l.evef III;
•.. .
V
^yitfrnate Utilization
Mode! applications for estimating
proximate delivered dose and
for intefspecies extrapolation
of dose'
fffft \
Figure 3-1: Levels of PB-PK modeling research.
-------
RIHRA
Topic 3-7
tion, and sensitivity analyses. They will also attempt
to link exposure to dose to biological outcome—a
goal that is key to the RIHRA program. The linkage
of exposure-dose-outcome data is necessary as inputs
into the risk assessment process; yet, success in this
area is difficult to achieve. Interdisciplinary research
teams must be brought together and then kept highly
focused on a single objective. In the area of develop-
mental biology, some linkage work has been under-
taken in the base research program as well as in some
RIHRA projects. Working models are being devel-
oped to focus interdisciplinary studies on the expo-
sure-dose-outcome linkage.
Because PB-PK models are the bridge between
Topics 2 and 4, efforts will be made to link closely
the research activities in all three areas. For example,
the Topic 2 exposure assessment projects will provide
the input driver functions for the PB-PK models that
are critical for the models' accuracy and usefulness.
The Topic 2 studies, in other words, will apportion
the exposure concentration of a given chemical or
class of compounds among the various routes (e.g.,
air, water, soil); and the PB-PK models will then take
those input values and apportion the dose of the
compound throughout the body. Without an accurate
understanding of exposure, there can be no accurate
estimation of delivered dose, and thus no accurate
understanding of biological outcome. In addition,
Topic 2 projects that assess sources of error in
exposure models will provide outputs for use as
inputs in PB-PK models developed under Topic 3.
Detailed descriptions of Topic 3 projects are
provided in Appendix C. Linkages between projects
are also highlighted there—for example, Project
3.1.1(1), on dermal absorption of para-substituted
phenols, is designed to provide animal and human
PB-PK data and is linked to Projects 3.3.1(2), 3.3.2-
(1), and 3.3.3(1).
-------
-------
RIHRA
Topic 4-1
TOPIC 4
BIOLOGICALLY BASED DOSE-RESPONSE MODELS
4.1 ISSUES COVERED
Biologically based dose-response modeling (BB-
DR) is the ultimate goal of quantitative risk assess-
ment. By explicitly modeling the underlying mecha-
nisms of toxicity, BB-DR models facilitate these
improvements in the risk assessment process:
• Extension of the observed dose-response
relationship to much lower doses. Low-proba-
bility responses cannot be observed in bioassays
of reasonable size. Mechanistic models, howev-
er, provide a means of relating the quantal dose-
response (which cannot be detected when events
are rare) to the underlying continuous mecha-
nisms (dose-effect) even down to very low levels
(measurement is limited only by the resolution of
the analytical techniques).
• Specific predictions about relative sensitivity
across species. The level of human risk is
seldom observable directly; thus, the sufficiency
of nonmechanistic cross-species extrapolation
methods cannot be established empirically.
Mechanistic models, however, allow application
of dose-response data on one species to another
because they are designed to account for quanti-
tative differences in targets and in the various
rates of attack, repair, and propagation of dam-
age.
• Provision of a framework for generating and
testing hypotheses about mechanism. Models
embody current knowledge or hypotheses about
underlying mechanisms of toxicity. As such,
they provide direction for their own validation
and improvement—in other words, the design of
the model highlights data gaps that must be filled
for further model refinement For a given set of
observations, hypothetical mechanisms can be
evaluated by varying the model parameters and
then testing the model's ability to predict experi-
mental results. Thus, modeling should proceed
in concert with the gathering of data and elucida-
tion of toxic mechanisms.
In recent years, progress has been made in
developing mechanistic models as a means of investi-
gating and improving the extrapolations used in risk
assessment. Such models describe the key physiolog-
ical and biological elements and processes, as well as
their interactions. Well-constructed models trace the
consequences of variation in any key element as it
differs among species or dose levels. Mechanistic
models provide a means to incorporate new data and
theories about special mechanisms of toxic action into
quantitative extrapolation. For example, the impact of
very different levels of chemical stimulation of
peroxisome proliferation in different species could be
accounted for by modeling the consequences to the
chain of biological processes, i.e., carcinogenesis.
In summary, the advantage of the mechanistic
approach is that it allows the overall process of
extrapolation to be broken up into its biological
elements. Experimental data can be developed as
independent elements, and the consequences to the
overall extrapolation determined. As knowledge of
the underlying processes improves, the biological
realism of the model improves, and an experimentally
testable basis is developed for extrapolations that must
currently be done on the basis of assumptions. The
models themselves are generic and are applicable to
a variety of chemicals, but they can be adjusted to
incorporate chemical-specific effects information
(such as metabolic differences across species or
particular mechanisms of toxicity).
Under the RIHRA program, this biologic/mech-
anistic orientation will guide the development of BB-
DR models. Research will be directed at determining
the conditions under which data obtained in test
species and test systems can be used to predict
toxicity in humans. Such efforts will reduce the
uncertainties associated with existing methodologies
as well as lead to the development of new BB-DR
models for human health risk assessment Ancillary
components of these activities will be the develop-
ment of protocols to 1) validate these models, and to
2) facilitate their application in risk assessment.
To develop BB-DR strategies, scientists must
both use existing data and generate new data; and the
extent to which each option is exercised will, in part,
depend on the state of mechanistic knowledge for a
given health effect. The issues to be explored in this
area—intra- and interspecies extrapolation, extrapo-
lation of health risk across different exposure scenari-
os, and delineation of multiple, contributory mecha-
nisms of toxicity—are associated with major uncer-
tainties. Research will reduce these uncertainties by
confirming or replacing assumptions with scientifical-
ly defensible data. Emphasis will be placed especial-
ly on pulmonary, reproductive/development, and
-------
RIHRA
Topic 4-2
neurological health effects, with some effort devoted
to cancer effects. Because the primary focus of this
research is to better understand the role of various
biological processes on chemically induced injury, the
models should be flexible enough to incorporate new
information as it is obtained.
Issue 4.1: Inter/Intraspecies Extrapolation
Risk assessors find considerable uncertainty
associated with the factors that are responsible for
differences in response within and across species.
Researchers must elucidate the critical physiologic
and mechanistic factors that contribute to the health
effects of concern in the risk assessment process.
Such research will improve the basis on which risk
assessors adjust for intra- and interspecies variability
in dose-response extrapolations.
4.1.1: Homologous Models
To determine the extent to which effects ob-
served in one species can be extrapolated to another,
researchers must ascertain whether effects in animals
are analogous (i.e., superficially similar) or homolo-
gous (i.e., resulting from a common mechanism of
action) to those in humans. Research emphasis will
be placed on evaluating species similarities and
differences in both mechanism and expression of a
given outcome. Thus, work in this area will not only
attempt to confirm the existence of homologous
mechanisms for inducing specific toxicities (i.e.,
disease), but also to define the degree of homology
in the expression of such disease (i.e., comparable
outcome).
4.1.2: Interspecies Sensitivities
Using pharmacokinetic models (Topic 3), scien-
tists will be able to determine the effective dose at a
given target site. However, even given equivalent
target doses, interspecies differences in sensitivity
may still affect the severity of the resulting health
effects. Research will focus on the degree to which
dose-effect functions for a given health effect (e.g.,
reproductive failure) differ across species, as well as
the degree to which the relative sensitivity of different
health effects (e.g., reproductive versus neural) vary
across species at a given dose. Efforts under this
issue will also determine the appropriate dose metric
for expressing and comparing a given dose across
species (e.g., mass/unit volume, mass/unit area).
4.1.3: Intraspecies Sensitivity
Work in this area will characterize the factors
that may contribute to different sensitivities in re-
sponse to chemical exposure among individuals of the
same species. Variables to be evaluated include age
of the individual (developing, adult, or aging organ-
ism), previous or current health status, and genetic
makeup. Such data will allow for a better estimate of
the differential probability and extent of a given
health risk for particular subpopulations. An impor-
tant aspect of this research will be determining how
these factors interact with pharmacokinetics to
produce intraspecies differences in sensitivity.
Issue 4.2: Exposure Scenarios
Major uncertainties hinder understanding of how
variations in dose-rate, concentration, and duration
of exposure to environmental pollutants affect toxico-
logical outcomes in humans. Research efforts in this
risk assessment issue will determine the effects of
varying route, dose, dose-rate, duration, and cumula-
tive dose on health outcomes. Attention will also be
directed toward defining the continuum of effects and
their lexicological significance as a function of
exposure. These efforts will be designed to develop
BB-DR models in which both data and assumptions
realistically reflect human exposure scenarios. Some
activities in this area are expected to be cross-cutting
with those under Topic 3 (i.e., characterizing the
behavior of the delivered dose under different expo-
sure conditions). Clearly, exposure scenarios are
important determinants of outcome for a number of
health endpoints. Research in this area is therefore
critically important to risk assessment and risk man-
agement
4.2.1: Mechanisms Across Dose
To extrapolate accurately, risk assessors must
know whether the mechanism of toxicity varies as a
function of dose. Testing protocols that evaluate
many lexicological endpoints use some approxima-
tion of the "maximum tolerated dose" as their high
dose, with lower doses being mathematically reduced
multiples of thai level. In evaluating dose-response
relationships, risk assessors often assume that the
mechanism of toxicily does nol vary as a function of
exposure scenario, and that novel or secondary
mechanisms do not influence outcome at these very
high exposure rates. Yet, the experimental subjects
may be exposed to conditions that might well exceed
their capacity to biotransform and excrete the active
moiety, might saturate or cause disruption of natural
protective/repair mechanisms, and/or might trigger
-------
RIHRA
Topic 4-3
nonspecific stress responses. Because of the similari-
ty of protocol designs, results from studies on the
validity of effects caused by very high dose exposures
cut across multiple areas of toxicological significance.
4.2.2: Sensitivity ofEndpoints as a Function of Dose
In a dose-response study, endpoints range' in
severity from biochemical alterations, to physiological
changes, to pathological conditions, to mortality. Risk
assessors must understand the progression of these
biological effects in terms of adaptive responses,
compensatory responses, and overtly adverse or
pathophysiological responses. Research projects
under this issue will define the full extent of respon-
ses throughout the experimental dose range, the
interrelationships among these effects, and their
biological significance.
4.2.3: Influence of Exposure Parameters on Outcome
At one extreme, chemicals exert their effects
when a critical body burden is exceeded, regardless
of the level and duration of exposure. At the other
extreme, the toxicity of a chemical depends on the
dose rate or duration of exposure. In the latter case,
the toxic effects of a short-term exposure to high
concentrations may be very different from those
produced by long-term, low-level exposure. In some
disciplines of toxicology, this issue translates into a
choice between the peak concentration and the cumu-
lative exposure as the trigger for inducing toxicity.
Researchers must therefore attempt to improve our
understanding of the interplay among rate, intensity,
and duration of exposure as they affect toxicological
outcomes so that the appropriate trigger can be
identified.
Issue 4.3: Mechanistic Variation
A variety of biological events may contribute to
the occurrence of a given health effect. Researchers
must develop dose-response models that factor in the
potential for different biological mechanisms to elicit,
initiate, or contribute to the health effects of concern.
To date, the primary efforts in this area have focused
on delineating the role of non-genetic events in the
development of dose-response models for cancer.
This effort will continue, and comparable consider-
ation will be given to the variety of mechanistic
pathways that contribute to mutagenic events or other
target organ toxicities.
4.2 PROJECT SELECTION STRATEGY
The rationale and general strategy that drives
research in the area of BB-DR models has been
outlined in the Introduction. To explore all of the
issues listed there, however, would well exceed
available resources for Topic 4 research. Thus, the
Office of Health and Environmental Assessment
(OHEA) and the Office of Health Research (OHR)
have identified and prioritized the research activities
that can be addressed realistically with available
resources and can provide data that impact on the risk
assessment process within a reasonable timeframe.
To facilitate the prioritization process, a decision-
tree strategy (see Section 4.3) was developed that
could be systematically applied by key investigators
in each specialty area (e.g., neurotoxicology or
cancer). Groups were asked to consider the critical
assumptions for each target system (process) and the
importance of resolving those uncertainties based on:
• Impact
Direct: Research will result in direct modification
of current risk assessment practices and/or posi-
tions. Work is usually short-term in timetable
and application.
Indirect: Research will provide necessary data for
other research efforts that address uncertainties in
risk assessment. Work is longer-term in timeta-
ble and impact.
• Feasibility (i.e., given limited resources, can the
work be done?)
• Timetable for accomplishing the work
The ability to achieve the goals outlined in the
Introduction differs across specialty areas as a func-
tion of state-of-knowledge. Investigators in all health
effects areas, however, were encouraged to:
• Develop a well-articulated basis for the research
focus within the area
• Clarify how the proposed work integrates with
existing base programs in the respective offices
and recommendations on interactions between
OHEA and OHR on specific projects.
Since 1986, OHEA has had a prototypic research
program on Reducing Uncertainties in Risk Assess-
ment (RURA). Those efforts have been folded into
the RIHRA program. Projects funded under RURA
are so designated in Appendix D.
-------
RIHRA
Topic 4-4
4.3 DECISION-TREE STRATEGY FOR
PRIORITIZING RESEARCH NEEDS
FOR A GIVEN TARGET SYSTEM
A decision-tree strategy (Figure 4-1) was used as
one guide in evaluating proposals under Topic 4. The
research questions posed represented successive levels
of BB-DR model development. Within each disci-
pline, therefore, these levels can be used as a se-
quencing strategy in choosing the most appropriate
projects for immediate funding. For example, a
project that is designed to develop a BB-DR model
for a target system (Level in) in a test species might
not be the best choice for immediate funding if
sufficient homology data (Level I) have not been
gathered for that test species.
Level I: Assessment of Homology (Qualitative
Comparability) of Current Test Species
Work in this area defines the relative merits of the
current test species in terms of:
• Comparability of basic physiological processes to
humans
• Homology of endpoints to humans
This qualitative concern must be addressed before the
test species can be used in more quantitative research.
This research area should not be confused with work
in species-species extrapolation, which is directed at
calibrating species differences between an acceptable
animal model and humans.
Component. 1: Qualitative Homology of Current Test Species
Defines relative merils and adequacy of current test species. If qualitatively comparable, proceed to levels of analyses.
Component^! Accurate Derivation of Dose-Effect jFtelatlonsnfps for Existing Data*
Endpotnt Interrelationship , Progression of
and Co-influence Toxldty
Predicted Validity Low Dose
Measures to Funo.Toxieity
Defining Critical
, effect
Components: Risk Estimation*
""'• Statistically Derived
" Approaches
"* Derive High+Low Dose Model
MachanfSrri TftrdsftoTd/baCKaroUrvd
' •• ^ 'y ^ fr «'. "
' /'*'? '*
2nd fjer »Modat Modification
Irtter/lntnjspecies different Exposure ?
Difference Conditions
Novel Approaches
Ouanfify
Uncertainly
Factors;
Tor a gfren target system (process), research underComponents 2 amf 3 may proceed concurrently or sequentially.
^ A fwdbacftfoopbefweea components fe envisioned.
Figure 4-1: Components of strategy to be applied to a given target system (process): developing
research rationale and focus.
-------
RIHRA
Topic 4-5
Level II: Accurate Derivation (Expression)
of Dose-Effect Relationships for
Existing Data
Work in this area focuses on developing an empirical
base of data on underlying mechanisms and their
lexicological consequences for a given target Uncer-
tainties in this area concern identifying, analyzing,
and expressing existing data in a manner suited to
deriving the most accurate and relevant dose-effect
function for the toxic event Examples of possible
activities within this area include:
• Endpoint interrelationships and co-influences on
the expression of the dose-response curve
• Identification of mechanisms/processes that affect
the progression of toxicity across dose ,
• Predictive validity of low-dose sensitive mea-
sures to higher-dose functional outcomes
• Integration of the previous three types of activi-
ties to define critical effect
The primary goal of work in this area is the
establishment of empirical relations between compo-
nent processes. Data must be gathered, therefore, not
only on dose-effect relationships but on effect-effect
relationships (i.e., the influence of underlying pro-
cesses on one another) and on effect-response rela-
tionships (which empirically tie different levels of the
underlying continuous physiological effects to the
probability of quantal response). Results from pro-
jects in this area would feed back on the qualitative
models (by validating or refuting causal links) and
would also be the basis for generalization, parameter
estimation, and testing of the quantitative models.
Research in this area may have the greatest
impact (see the criteria listed at the beginning of this
section) of any work under Topic 4. Possible direct
impacts include better guidance to Program Offices
concerning the appropriateness of a given endpoint on
which to base a risk assessment, as well as better
control of the RfD process by clarification of the term
"critical effect." Possible indirect impacts include
provision of data for guiding the development of BB-
DR models.
Level III: Quantitative Approaches to
Risk Estimation
Efforts within this area would be concentrated on
reducing the uncertainties associated with the applica-
tion of either statistical approaches (e.g., RfD) or BB-
DR models for risk estimation by use of the empirical
relationships identified above. These activities are
interdependent.
Statistical Approaches
These approaches are likely to continue to
dominate the Agency's risk assessment process for
noncancer health effects. Therefore, research that
reduces the uncertainties associated with these meth-
ods would have a direct impact on the process of risk
assessment Projects of this sort usually have rela-
tively short timetables. Work following this approach
falls into these categories:
• RfD process: studies to calibrate inter- and intra-
species differences to provide statistical certainty
concerning the appropriateness and magnitude of
the uncertainty factors
• Development/application of novel mathematical
approaches: studies do not rely only on the No
Observed Adverse Effect Level (NOAEL);
instead, they use total dose-response data in
linear extrapolation and attempt to define the
dose associated with specific risk levels (e.g.,
benchmark approach)
BB-DR Models
Work in this area requires longer-term support
than statistical efforts. Although modeling projects
can continue over long period of time, their interme-
diate stages provide much potential benefit to the risk
assessment process. Following is a possible tiered
approach to model creation:
1. Derive a "working" high-to-low-dose extrapola-
tion model(s) for a given target process; required
efforts include:
• Identification and integration of existing
basic biologic information
• Support of work on mechanisms to identify
critical biologic processes
• Verification/integration of threshold for an
outcome into the BB-DR model
• Other efforts not yet defined
2. Use the model to define (and drive) subsequent
research designed to address data gaps and
recalibrate the model
-------
R1HRA
Topic 4-6
3. Adjust the model for inter- and intra-species
differences and different exposure conditions
It should be noted that some of the data required
under tier 3 (e.g., interspecies differences) may be
available from studies conducted under RIHRA Topic
3 (Pharmacokinetics).
4.4 SUMMARY OF THE PROPOSED
RESEARCH PROGRAM
Under Topic 4, research programs have been
funded in the following areas:
• Neurotoxicology
• Developmental/reproductive toxicology
» Pulmonary toxicology
» Genetic toxicology/cancer
• Generic, cross-cutting issues dealing with epide-
miology, severity of effects, duration of expo-
sure, and improving the RfD process
Groups responsible for each area were asked to
prepare a statement of focus and rationale as the basis
for prioritizing their projects. The rationales for
project choice in the areas of genetic toxicology/can-
cer and reproductive/developmental toxicology were
developed jointly by OHR and OHEA. The Topic 4
co-chairs, Drs. Robert Kavlock and Harold Zenick,
then used the program packages (including overview,
prioritization chart, and project descriptions of exist-
ing and proposed work) to recommend the most
important projects for Topic 4 resources.
During this final prioritization process,
following considerations were recognized:
the
Several of the proposed projects would provide
important data for a number of Topic 4 issues;
thus, assignment of the projects to a given issue
was occasionally arbitrary (see Table 4-1).
The rationale statements for programs on the
various target systems had to include projects
already incorporated into the RIHRA base pro-
gram (or the OHEA RURA program) as well as
new proposals. The emphasis and priorities of
the programs were therefore often shaped by
these prior commitments. Presumably,, as the
RIHRA program evolves, these constraints will
lessen.
• The nature of the projects proposed under Topic
4 suggested a number of possible interactions
between OHEA and Health Effects Research
Laboratory (HERL) personnel. These interac-
tions are being pursued.
Following is a brief synopsis of the programs for
each research area. The detailed rationale and project
descriptions for each of the areas noted here may be
found in Appendix D. See Table 4-1 for a listing of
the funded projects under Topic 4.
Neurotoxicology
The majority of RIHRA projects in this area
concentrate on inter- and intra-species comparability,
especially in terms of sensory and cognitive process-
es. Related work compares various quantitative
approaches to the RfD in estimating neurotoxic risks.
Additional projects address the influence of exposure
conditions on outcomes.
Several of the proposed projects—especially the
work examining quantitative models and evaluating
historical pharmaceutical databases—can provide
excellent vehicles for interaction between OHEA
scientists at ECAO-Cin and HERL. ECAO-Cin has
substantial experience in the areas of modeling and
the use of existing databases; NTD-HERL has the
expertise to evaluate the scientific credibility of
various approaches to neurotoxicity data.
Developmental/Reproductive Toxicology
HERL and OHEA groups have joined in devel-
oping and evaluating reproductive and developmental
risk assessment guidelines for focusing research
activities in this area. Under female reproductive risk
assessment, research efforts are addressing issues
concerning the homologous mechanisms associated
with processes of ovulation through implantation.
Under male reproductive risk assessment, efforts are
concentrating on defining endpoint relationships and
determining the functional significance of low-dose
measures of toxicity.
The developmental segment of the program
addresses endpoint issues, especially as related to
maternal and developmental toxicity. Because a more
substantial database is available on developmental
concerns, efforts are also being directed toward the
development of BB-DR models that incorporate
threshold concepts.
-------
RIHRA
Topic 4-7
TOPIC/
ISSUE
TABLE 44
PROJECTS FUNDED HINDER TOPIC 4
BIOLOGICALLY BASED DOSE-RESPONSE MODELS
PROJECT FY89
OFFICER SHOUT TITLE STATUS ($IC)
DURATION
NeurotoxicoloEY
4.1.1
4.L3
Hudnellb Homologous Sensory
Function
Stanford Homologous Learning
Stantonb Homologous Learning
PadiOa*1 Peripheral Neuropathy
MacPhaii1* Sucepfibility to
- Neuroloxicants
MacPhall1* Quantitative
Assessment
Jensenb Cholmergic Aging
4,2.3 Crofton? Exposure Scenarios
Developmental/Reproductive Toxicology
4.1.1
4,1.2
4.1.3
4.2.1
* Coop
tlAG;
Cummingsb Early Pregnancy Loss
Gray6 Probability of Infertility
Cooper1' Control of Ovulation
Kavlockb Renal Mechanisms
Rogers* Embryo Culture
Laskeyb Leydig Function
Rogersb Litter Effects
Chexnofi* Maternal Genome
Miller* Maternal Toxicity
Chemotf1' Low Dose Thresholds
Rogers'1 Zinc Metabolism
= Cooperative Agreement
= Inter-Agency Agreement
Northrop Services, Inc.
Northrop Services, Inc.
Coop* - UNC
Virginia Pblytechnical Institute
Northrop Services, Inc.
JAGt-Natl. last.
Mental Health
Coop -UNC
Coop - N.E. Louisiana Slate Univ.
Northrop Services, Inc,
Northrop Services, Inc.
Northrop Services, Inc.
Coop - Duke
Northrop Services, Inc,
Coop - Johns Hopkins
Exist NS1/NRC
Nat], Research Council
Northrop Services, Inc.
Natl. Research Council
Coop - UC Davis
,
122
85
110
225
81
120
160
106
106
56
131
125
152
100
106
66
71
S0_
75
«*»w
FY 89-90 (91)
FY 89-90
FY 89-90 <91)
FY 89-90 (91)
FY 89-90
FY 89-90
FY 89-90 {91}
FY 89-90 (91)
FY 89-90 $1)
FY 89-90 (91)
PY 89-90(91)
FY 89-90 (913
FY 89-90 (91)
FY 89-90(91)"
FY 89-90 (913
'FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
-------
RIHRA
Topic 4-8
TABLE 4-1 (Cont'd,)
TOPIC/ PROJECT
ISSUE OFFICER
Kimmel/
Kavlock^
Dellarco0
4.2.2 Kimmeld '
Cleggc
Pulmonary1
4.1.2 Hatch
4.1.3 Hatch
4,2 Keren
Genfetic Toxicology
4.1,2 Nesnowb
Moore*
Nesnow''
Farland*
Massb
Daniel
4.13 Baylissd
4,2.2 DeUarco"
43 Massb
Nesnowb
Chenc
Reese0
Chen*
Bayard3
Schoenyd
SHORT TltLB " "
BB-Dfc Models
Zygote Toxicity
Multiple Endpoints
Sperm-Fertility Relationship
Ozone Dosimeay
Lung Injury
Inflammatory Response
Rodent/Human Sensitivity
Mutation Induction
AZQ Adducts
. Fish Cancer D-R Models
Oneogene Activation
Hepalo/Nephro Carcinogens
Cohort Survivorship
DNA Alkylalions
MoolgaV./Knud.
Gene. Expression Alterations
Chromosomal Translocation
Initiation/Promotion
Class of BB-DR Models
D-R Promoters/Initiators
Carcinogen Combination
F
STATUS ,„
IAG * Natl. Center Toxicol. Res,
IAG - Oakridge
wNatI. Lab
IAG - HJth, Tox. Prog.
Coop - Georgetown
Coop - UC Davis
Exist NSI
Exist Environmental Monitoring
Services, tnc.
Exist Environmental Health
Research Testing, Inc.
Coop - U. Vt.
Coop -Bay lor
EPA-Duluth
Coop - Med, Coll, Ohio
Coop - Northern Kentucky Univ.
IAG - MIOSH
Coop - Coriell Inst.
Coop - Univ, Wisconsin-Madison
Coop - U. Pgh.
Proposed Coop
IAG - FDA
Varied
Varied
On-Site Cntrt
Y8Ł>
80
80
40
100
75
265
171
136
75
290
60
0
0
40
&
170
120
80
30
105
105
40
DURATION
FY89-90 (91)
FY 89-90 '
FY 89-90 (913
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (913
FY 89-90 (91)
FY 89-90 (91)
Starts
FY 90-91
Starts FY 90
FY 89-90
FY89
FY 89-90 (91)
FY 89-90 (91)
FY 89-90
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
-------
RIHRA
Topic 4-9
TABLE 4-1 (Cont'd.)
TOPIC/
ISSUE
PROJECT
OFFICER
SHORT TITLE
STATUS
FY89
C$K)
DURATION
Cross Issue/Cross Target
4.2,2
Swartoutd
DeRosa4
Everson
Hertzberg/Davis*
Farland"1
Quantifying RfD Process
Less-Than-Lifetime Exposure
Human Data
Severity of Effects
Workshops Risk Assessment
"Year in parenthesis denotes major review of progress
"OHRlead
COHEA lead under RIHRA
"OHEA RURA component
'Co-funded by OHEA and OHR
fOHR lead on all pulmonary projects
Varied
Varied
SRA
Varied
^7/
Varied.
prior to further
110
128
80
140
150
FY- 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
FY 89-90 (91)
'FY 89-90 (91)
commitments of funds in that year.
Pulmonary Toxicology
Efforts in this area focus on improving the
dosimetric and concentration-time relationships in
animal and human exposure models. The proposed
studies add critical components to existing studies by
enabling determination of target tissue dose in nonhu-
man primates exposed to ozone. Combining these
data with those already obtained in rat studies will
facilitate extrapolation of a body of existing toxico-
logical data to humans. Other studies are evaluating
a recently proposed model of concentration time-by-
time relationships for air pollutants through analysis
of bronchoalveolar lavage proteins as well as pul-
monary host defense systems. These sensitive indica-
tors of damage and repair capability provide a basis
from which to develop defensible regulations for a
variety of exposure scenarios.
Genetic Toxicology/Cancer
The primary issues in this area are:
• Mechanisms of action
• Interspccies sensitivity
• Biological modeling
Most of the new work focuses on somatic cells.
Whereas the FY88 HERL-RIHRA projects concen-
trate heavily on interspecies extrapolation, the new
work is directed toward the definition of basic mecha-
nisms (genetic and nongenetic) of cancer. Special
emphasis is placed on the identification of the events
of initiation and promotion.
This program represents ongoing discussions
between HERL and OHEA scientists and offers great
potential for the types of interaction envisioned under
the RIHRA program, namely 1) a feedback loop
between modeling efforts to identify data gaps; and 2)
laboratory work aimed at elucidating basic physiologi-
cal processes.
At the present time, no projects on endpoints
used in heritable mutagenesis risk assessment are
included in the program. The critical needs in this
area must be better delineated and coordinated with
ongoing, substantial programs such as the National
Toxicology Program (NTP). A heavy commitment of
resources is required to address these issues; a more
profitable approach may be to interact with and
contribute to the research activities of the NTP
program. Research identified in heritable mutagenic
-------
RIHRA
Topic 4-10
risk will be incorporated into the reproductive and
developmental research strategy.
Generic, Cross-Cutting Issues
Activities in this research area address a number
of issues with relevance across target systems (pro-
cesses), including:
» Comparing clinical, epidemiological, and experi-
mental databases to improve species extrapola-
tion
» Convening workshops to address specific issues
in the development and implementation of the
RIHRA program
• Developing improved methods for quantifying
risk (including the RfD)
» Factoring severity of effect and varying exposure
conditions into the risk characterization process
These issues are pivotal to any health risk assess-
ment. The research effort will benefit from greater
involvement of HERL/OHEA scientists. Special con-
sideration will be placed on target systems for which
the expertise is available within HERL/OHEA to
closely scrutinize model assumptions, approaches, and
results.
The OHEA/RURA program will fund the Work-
shop activities listed on Table 4-1. This project has
been listed under Topic 4; however, these funds are
presumably available for workshops addressing any of
the four Topic areas.
-------
RIHRA
Appendix A -1
APPENDIX A
TOPIC 1: ANALYSES OF UNCERTAINTY IN RISK ASSESSMENT
PROJECT DESCRIPTIONS
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Analyses of Uncertainty in Risk Assessment
Research Strategies
New RIHRA project
Review of "Guidelines for Quantifying Uncertainty in Risk Assessment"
General guidelines for quantifying uncertainties in risk assessment are regarded as a high
priority issue among Agency risk assessors. Resources for the Future (RFF) has
developed such guidelines using input from Federal and academic scientists. In addition,
they are applying these guidelines to case studies of current interest to the Agency. This
project will use the cooperative agreement with the newly formed Committee on Risk
Assessment Methods (CRAM) of the National Academy of Sciences to perform a detailed
peer review of these guidelines. A select committee of knowledgeable scientists will be
convened by CRAM with input from a Federal Liaison Group. A written report of the
committee's findings will be provided to the Agency.
A report of the comments of NAS peer reviewers will be presented to the Agency at the
completion of the review. This report will take the form of a project summary but may
be developed into a journal article depending on the nature of the review and comments
developed.
The Agency will benefit from the efforts of RFF to develop guidelines for uncertainty
analysis. In addition to providing input in the development of these guidelines, this
project will allow the guidelines to receive necessary scientific review and consensus,
which will be needed should the Agency choose to incorporate these guidelines by
reference into Agency documents. This project also illustrates the concept of leveraged
funding that will be important in carrying out research in uncertainty analysis under Topic
1 of RIHRA.
1 yr FY89: $35K
W. Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Title:
Analyses of Uncertainty in Risk Assessment
Research Strategies
New RIHRA project
Addressing Uncertainties for an Integrated Exposure Model
-------
RIHRA
Appendix A - 2
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
The goal of this project is to develop a strategy for evaluating the sources of and
characterizing the uncertainty in a TEAM-type of exposure scenario. The project will use
tetrachloroethylene in the air and ground water in a hypothetical community as an
illustration of the approaches to be applied. This project will be closely coupled with the
project described below (Issue: 1.1.2) in its dealing with the question of uncertainties in
projecting individual versus population exposure levels.
Transport Modeling. Using multimedia transport models such as GEOTOX, and single
media models for air and ground water dispersion in combination with measured data,
sources of the chemical will be quantified and relative magnitudes of uncertainties
associated with both the multimedia and single-media estimates of contaminant
concentrations attributable to a regional source for steady-state and dynamic conditions
will be characterized. This approach will allow the apportionment of estimated
uncertainty among contributions of model assumptions, transport properties, and chemical
properties.
Personal Exposure. For the compound under study, methods for characterizing pathway
exposure factors (PEFs) will be identified and uncertainties surrounding each PEF will be
characterized. Special attention will be paid to characterizing those PEFs that are both
highly uncertain and apparently important contributors to total exposure. TEAM data
developed by the Agency will be compared with an integrated exposure model approach
developed by LLNL. Uncertainty contributions from physical and chemical properties,
variations across microenvironments (such as within households), and variations among
individuals are expected to be the focus of this effort.
Results of such a study, combined with Agency research efforts in exposure assessment
under the ORD base and within RIHRA Topic 2, will provide a case study of application
of the principles of uncertainty assessment currently in use. These data will be presented
in a report to the Agency for use in evaluating the application of principles of uncertainty
assessment. Depending on the results of the study, and comments during review, the
report may be published in the peer-reviewed literature.
Results of this effort will integrate nicely with ongoing EPA work. The case study chosen
will have direct applicability to ORD's support for the Air Program in understanding and
characterizing exposures for volatile organic compounds, a high-priority issue. This
project will leverage work accomplished by ORD and will permit comparison between
collected data and model predictions.
1 yr FY89: S90K
W. Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Analyses of Uncertainty in Risk Assessment
Research Strategies
New RIHRA project
-------
RIHRA
Appendix A - 3
Title:
Description:
Identified Results:
Project Length
and Cost:
Project Officer:
Intel-individual Variability and Uncertainty in Risk Assessment
This project will focus on the element of interindividual variability in the overall
uncertainty framework and will characterize the implication of this parameter on
uncertainties in the case study under Project 1.1.2. This study will focus particularly on
the issue of added uncertainty in estimating individual risks as opposed to population
risks, given a range in interindividual variability.
Major Sources of Uncertainty. This project will address the major sources of uncertainty
within the post-exposure portions of the tetrachloroethylene assessment Residual
uncertainly and interindividual variability will be assessed quantitatively and characterized
using a model developed at LLNL. This final characterization will specifically address
the predicted uncertainty and interindividual variability distributions of risks to exposed
individuals and the distribution of uncertainty in corresponding population risk (i.e., in the
total number of predicted cancer cases).
Communication of Uncertainties. Results of the previously described work will be framed
into a readily understandable format, highlighting the distinction between interindividual
variability and other risk assessment uncertainties, on the one hand, and between
individual risk and population risk on the other. Novel approaches, both analytical and
graphical, will be employed for this effort.
Results of this project will be useful to the Agency in addressing two issues within the
framework of uncertainty in risk assessment: interindividual variability and linguistic
imprecision. The interaction of the two investigators from LLNL will provide a
comprehensive evaluation of uncertainty in the case study. As described above, this case
is important for the Air Program.
1 yr FY89: $90K
W. Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Title:
Description:
Analyses of Uncertainty in Risk Assessment
Research Strategies
New RIHRA project
Evaluation of Communication of Uncertainty in Risk Assessment
This project will provide sound survey information on how carcinogen classification
systems are interpreted by both the lay public and the scientific community. Both
presently used descriptors of cancer hazard and those under consideration for future use
will be analyzed. In addition, a selection of approximately a dozen case chemicals will
be used to explore the practical application of these descriptors to weight-of-the-evidence
situations. Groups of scientists will be selected and briefed using approaches validated
in earlier work at Harvard involving the influence of expert scientific judgment in cancer
risk assessment This project will be funded co-operatively with Dow Chemical company.
-------
RIHRA
Appendix A - 4
Identified Results:
Usefulness of Results:
Project Length
and Cost:
This project will result in a report that will provide insights concerning linguistic
imprecision in currently used terms for describing cancer hazards. Depending on the
results of the research and review comments, the report may be submitted for peer-
reviewed publication.
These results will be useful to the Agency as it re-considers approaches to classifying and
describing cancer hazards. Many consider this aspect of the risk assessment process to
be the most illustrative case of linguistic imprecision. Survey data will assist the Agency
in deciding how it might improve communication of cancer hazards under varying
weights-of-the-evidence. The project represents a joint effort between the Agency and
industry to address this issue.
1 yr FY89: $35K (plus $20K provided by Dow)
Project Officer:
W. Farland (OHEA) (202) 382-7315
-------
RIHRA
Appendix B -1
APPENDIX B
TOPIC 2: INTEGRATED EXPOSURE ASSESSMENT
PROJECT DESCRIPTIONS
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Integrated Exposure Assessment
2.1 Human Exposure Models
New
Modeling Human Exposure
The goal of this project is to improve the accuracy of predictions of exposure for
chemicals of interest to the Agency. Human exposure models employ computer
algorithms to combine the concentrations of pollutants people experience in various
microenvironments (microenvironmental exposures) with the times spent in those
microenvironments (human activity patterns). Although human exposure modeling is in
its infancy, several prototype models (e.g., SHAPE and NEM) have been developed based
on the early concepts developed by Fugas and Duan. To be effective, such models must
include all the microenvironments that people normally visit (for example, homes, stores,
schools, offices, work places, subways, buses, and automobiles) and must include multiple
routes of exposure (air, food, skin, and drinking water). Although several models have
been developed, none has been fully validated with real exposure data collected from
human exposure field studies. Such validation is necessary to determine the accuracy of
the models and to calibrate and redesign them to make predictions with known confidence
intervals. Recent findings from total exposure studies suggest how these models can be
developed further and improved.
Research on this topic will proceed by first constructing submodels for important
pollutants that EPA regulates (selecting, for example, from respirable particulates, volatile
organic compounds, semivolatile organics, formaldehyde) using the best
microenvironmental exposure field data available. Existing data bases and research
literature will be reviewed to construct these microenvironmental submodels, and ongoing
microenvironmental field study data that can be used will be carefully evaluated. Serial
autocorrelation problems in microenvironmental submodels will be addressed
(RTI/Harvard and NOAA).
Simultaneously, the best human activity patterns data currently available (for example,
John Robinson's 1986 national activity pattern survey, the 1988 California activity pattern
survey, and the activity surveys conducted in Cincinnati, Denver, and Washington, DC)
will be evaluated and coded for input into the model (UNLV).
The human activity pattern data and resulting microenvironmental submodels will be
combined in a human exposure model using a generalized exposure equation. Once
developed, this model will be validated by testing it with field data collected in total
exposure field studies for a variety of situations. Uncertainty ranges around the
predictions will be characterized (UNLV).
Short term: Evaluation of existing models and needed improvements.
Long term: A generalized human exposure model and submodels that have been
validated and the uncertainty well characterized. The written product will consist of a
-------
RIHRA
Appendix B - 2
Usefulness of Results:
Project Length
and Cost:
Project Officer:
on
research report and several peer-reviewed articles describing the model, its validation, and
applicability to various exposure situations.
These results can be used by all Program Offices in the agency that are involved in
estimating human exposures.
FY90 - reported results on modeling the impact of outdoor ambient air concentrations
measurements of total human exposure.
FY91 - reported results on the status of existing human exposure models.
FY92 - reported results on the prototype human exposure model combining multiple
pathway exposures.
An essential component of the risk assessment process is the estimation of human
exposure to chemicals. This model will improve Agency risk assessments by providing
a tool that will permit more accurate exposure assessments than previously possible. This
model also will allow the effect on risks to health of different Agency regulatory
strategies to be evaluated more accurately.
3yr FY89:$100K
Irwin (OMMSQA; AREAL-RTP) (919) 541-4567
Beher (OMMSQA; EMSL-LV) (702) 798-2216
Topic:
Issue:
Status:
Title:
Description:
Integrated Exposure Assessment
2.1 Human Exposure Models
Ongoing
Procedure for Evaluating Multimedia Exposures to Incinerator Slack Emissions
The goal of this project is to improve the process of modeling exposures by developing
and applying a standard PC computerized approach. Each year both commercial and
social activities in the United States produce more than 150 million tons of discarded
waste. Both hazardous and municipal solid wastes have been traditionally disposed of in
land areas dedicated to that purpose. Meanwhile, as the amount of wastes needing
disposal continues to grow with the increase in the U.S. population, restrictions concerning
land disposal are becoming more evident. As ocean disposal and land disposal of wastes
are being phased out, waste incineration becomes a more viable solution. Incineration of
wastes, both municipal and hazardous, is in most cases not 100% efficient in destroying
contaminants. Consequently, deposited stack emissions are drawing increasing attention
because deposited contaminants can cause potential health hazards and can create
multimedia pathways for human exposure. Deposited contaminants can contaminate soil
and the food chain. Runoff from contaminated soil can cause contamination of sediment
and water. Example facilities that can be of concern include hazardous waste incinerators,
sludge incinerators, lead smelter emissions, and municipal waste incinerators.
-------
RIHRA
Appendix B - 3
Identified Results:
A significant effort in the area of multiple pathway exposure analysis due to municipal
waste combustor emissions has been developed by EPA's Office of Air Quality Planning
and Standards and the Environmental Criteria and Assessment Office in Cincinnati, OH.
A draft document entitled "Methodology for the Assessment of Health Risks Associated
with Multiple Pathway Exposure to Municipal Waste Combustor Emissions" was released
in October 1986. The methodology consists of a series of environmental fate and
transport models that utilize the known physical and chemical properties of specific
pollutants to predict the:
atmospheric dispersion from stack emissions
potential for surface deposition and accumulation
movement of the settled pollutants through and into various environmental media
potential bioaccumulation of pollutants into trophic systems
potential for adverse effects on the vitality of natural ecosystems
potential for adverse effects on human health
With regard to evaluation of human health effects, the methodology estimates health risks
that result from inhalation of predicted ambient air concentrations of pollutants, ingestion
of pollutants deposited on the ground and bioaccumulated through the food chain,
ingestion of potable water or aquatic organisms contaminated by surface runoff and
leaching and percolation of settled pollutants into water supplies, and ingestion of soil
contaminated by deposited incinerator emissions.
The final document will address these items: 1) a review of the present methodology,
closely coordinating with ECAO-Cincinnati, to ensure that the latest analytical tools are
being used in the development of this project; and 2) a user-friendly PC program that will
incorporate much of the methodology of the ECAO document which has not yet been
programmed in a easy-to-use fashion.
Short term: A review of the present methodology.
Long term: A computer model that quantifies exposure assessment procedures applied
to several point source categories.
FY90 - reported results of the review of the current methodology.
FY91 - PC program for use in evaluating risks from incinerator stack emissions.
FY92 - PC program for use in evaluating risks from other emission sources.
Usefulness of Results: The final product should achieve the following results:
• create a standard procedure to evaluate human exposure to deposited toxic
contaminants originating from stack emissions
Project Length
and Cost:
Project Officer:
create a simplified PC model that will generate exposure and risk values under
numerous exposure scenarios for many different levels of stack emissions.
3 yr FY89: $125K
J. Schaum (OHEA-EAG) (202) 382-8909
-------
R1HRA
Appendix B - 4
Topic:
Issue:
Status:
Title:
Description:
Integrated Exposure Assessment
2.2 Human Activity Patterns
Ongoing
Comparative Analysis of Existing Activity Pattern Data Bases and Relationship to Specific
Microenvironments
The goal of this project is to utilize existing activity pattern data for predicting human
exposure. Human activity patterns are important for making exposure assessments, but
EPA has not adequately utilized past activity pattern data bases. Additional work is
required to review, interpret, code, assess, and understand activity pattern data bases
collected in previous field investigations in the U.S. and Europe.
This project will obtain and examine the data bases collected in activity pattern surveys
conducted in the previous five years to characterize differences in urban, state, and
regional activity patterns through comparisons with the activity patterns in cities where
TEAM studies have been carried out. The data bases selected for review include the
Robinson 1986 nationwide activity pattern survey, various urban and statewide activity
pattern studies, and European time budget studies.
Improved exposure predictions can be made if the following questions can be answered:
1. In what activity is an individual involved?
2. Where and when is the activity taking place in time and space?
3. What is the duration of the activity?
This research will associate each activity with a frequency distribution of
microenvironments in which the activity occurs. A detailed stratified structure of
location/activity data will be constructed. This work will enable the production of a
computer subroutine that can be used to model total human exposure to environmental
pollutants. The subroutine will generate human activity patterns and associated
microenvironments necessary for exposure estimation. Also, insufficiencies in human
activity pattern data will be identified during the proposed research. Such identification
will lead to suitable recommendations for future human activity pattern research.
Short term: Identify if additional activity pattern data need to be collected and, if so,
why.
Long term: Develop a valid data base for modeling exposure for the Agency's risk
assessment efforts.
FY90 - report the results of the comparison of the 1985 - 86 National Activity Patterns
Survey data and the 1987 - 88 California Activity Pattern Survey data.
FY91 - report the results of testing the hypothesis related to the influence air quality has
on activity patterns.
FY92 - report results from studies of children's activity patterns related to dermal and
inhalation exposure.
Usefulness of Results: Although past field studies have provided exposures for the population in a few U.S.
cities, those data are insufficient to predict annual exposures of the U.S. population. To
Identified Results:
-------
RIHRA
Appendix B - 5
Project Length
and Cost:
Project Officer:
extend the findings from these few cities to other parts of the country, activity patterns
in these cities must be compared with activity patterns throughout the U.S. These
analyses will allow us to generalize estimated exposures of the U.S. population for use
in Agency risk assessments.
3 yr FY89: $150K
J. Beher (OMMSQA; EMSL-LV) (702) 798-2216
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Integrated Exposure Assessment
2.2 Human Activity Patterns
New
Develop National Database of Activity Patterns for Exposure Assessment
This project will design and collect a national activity pattern data base specific to
exposure assessment. Although several large activity pattern data bases exist, none of
these contain all the activity variables and information required for human exposure
modeling. Use of representative probability sampling techniques will permit a new EPA
activity pattern survey on a national scale that will provide the data needed to meet
exposure modeling needs at a reasonable cost. Prior to conducting the large-scale survey,
the researchers will conduct a pre-test survey to evaluate questionnaires and data
collection techniques.
A small-scale pilot activity pattern field survey will be undertaken. Questionnaires and
diaries will be evaluated using indepth interviews and focus groups. Respondent-
following techniques will be tested to compare the quality of the data generated through
questionnaires, diaries, interviews, and telephone recall methods. In addition, an
automated diary datalogger will be evaluated as a possible technique for providing more
accurate activity pattern information than traditional pencil-and-paper or recall methods.
An automated sensor capable of monitoring the times spent in indoor and outdoor settings
will be included in this pilot pre-test study.
This project will also formulate sampling protocols for a multi-stage probability sample
of the U.S. population. The sample will be stratified to provide suitable urban and
regional information for exposure modeling.
Short term: A standardized exposure-related survey instrument.
Long term: A national exposure-related data base for use in exposure modeling.
FY90 - results from pilot testing new exposure-related survey instrument.
FY91 - a-survey of national activity patterns in the United States.
FY92 - results showing differences and similarities of activity patterns by geographic
region, socioeconomic status, seasons, and other factors.
-------
RIHRA
Appendix B - 6
Usefulness of Results:
Project Length
and Cost:
Project Officer:
One of the weakest components of human exposure models is inadequate human activity
pattern data, particularly on a nationwide scale. The activity pattern information provided
by the large-scale survey is "generic" in that it can support development of improved
human exposure for several different pollutants and classes of pollutants. This research
provides the basic tools required for conducting a large-scale activity pattern survey.
3 yr FY89: S150K
W. Nelson (OMMSQA; AREAL-ATP) (919) 541-3184
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Integrated Exposure Assessment
2.3 Indirect Exposure
Ongoing
Pica Ingestion Rates in Children
This project will examine the soil ingestion rates of children exhibiting pica behavior. In
work conducted to date, EPA has studied soil ingestion in a population of one hundred
children in the state of Washington (final report completed). Additionally, a follow-up
field study with a subgroup of these same children was conducted to determine how
measured ingestion rates vary over a longer time period and to make a first evaluation of
soil ingestion in adult subjects (parents of the participating children). At the same time,
a study carried out by the University of Massachusetts, earlier work from the Centers for
Disease Control, and research conducted in the Netherlands contribute to our
understanding of ingestion rates in "normal" children. However, a question regarded as
critical in many EPA assessments, the rate of soil ingestion by children having "pica"
behavior, is not addressed directly in this research. This project will support an
examination of soil ingestion in children exhibiting such behavior.
In planning to study pica children, the researchers found that the first major obstacle was
the identification of a population of children who exhibit pica behavior (regular ingestion
of non-food materials). Therefore, a first task is a study to determine the frequency of
pica behavior (as observed by parents) in a population of children six years and under.
This will be done through a survey of parents of young children in an area, or areas, to
be selected. Parents will be asked to report observed mouthing, soil ingestion, or other
pica behavior using a graded frequency scale. These data will allow an assessment of
both the prevalence of pica behavior and its nature. Since no expert consensus exists on
exactly what range of behaviors should be termed pica, this study will in effect develop
a working definition of the problem. Among the subjects identified as exhibiting
significant pica behavior, a subset will be selected, randomly, for measurement of soil
ingestion.
Short term: In FY89 a cooperative agreement will be initiated with the Fred Hutchinson
Cancer Research Center to conduct the field component of this research, establish field
procedures, and develop a working definition of pica.
-------
RIHRA
Appendix B - 7
Usefulness of Results:
Project Length
and Cost:
Long term: Report on prevalence of pica behavior in young children and soil ingested
by these children.
FY90 - identification of a sub-population of children with pica for measurement of soil
ingestion.
FY91 - report on prevalence of pica behavior in children in a selected metropolitan area.
FY92 - report on the effect of the prevalence of pica on the sample weights used in risk
calculations.
Estimates of soil ingestion are necessary for the majority of agency assessments
addressing contaminated sites and surfaces. Existing research has made much progress
toward identifying a typical or average level of soil ingestion by young children.
However, estimates of ingestion by children exhibiting pica are needed to improve the
basis for individual "reasonable worst case assessments." Additionally determining the
size of the actual pica population has much importance in the weighing of calculated risks.
This project will efficiently address these questions. EAG is aware of no other research
that will provide these data.
3 yr FY89: $125K
Project Officer:
P. White (OHEA-EAG) (202) 382-2589
Topic:
Issue:
Status:
Title:
Description:
Integrated Exposure Assessment
2.3 Indirect Exposure
New
Selection of Food Consumption Rates
This project is designed to improve the accuracy of the food consumption rates used in
current Agency risk assessments. Food consumption data play a major role in many
Agency exposure and risk assessments. Assumptions about the quantity of contaminated
fish consumed can strongly influence risk estimates from water pollutants. Contaminated
surface sites or leaking landfills can in turn contaminate local water bodies and thus risk
assessments become dependent on fish consumption rates. Animals grazing or vegetables
raised in contaminated soils may be consumed locally or distributed commercially, with
exposures dependent on levels of consumption of these foods. The process of setting
tolerances for pesticide residues involves selecting consumption rates for the particular
crops.
Over the years different EPA offices have developed varied food consumption estimates
to deal with these and related problems. However, the agency has not conducted a
systematic examination of the methodologies appropriate for deriving these estimates. A
number of large-scale studies have examined food consumption patterns across the U.S.
population; however, application of these study results to EPA needs can be problematic.
-------
RIHRA
Appendix B-8
Identified Results:
Usefulness of Results:
A particular difficulty is that food consumption data are most commonly obtained for
short periods of time with, for example, individuals reporting their intakes of food over
a three-day period. EPA's needs, on the other hand, are for long-term or "usual"
consumption data: Individual risk estimates are dependent on assumed lifetime average
food consumption rates.
In this project we will build on existing research being carried out by other federal
agencies to develop a recommended strategy for estimating rates of consumption of foods
and apply this strategy to form estimates of consumption rates for a small number of
sample foods. Food survey experts at the U.S. Department of Agriculture are currently
sponsoring basic statistical research to determine methodologies for estimating usual
consumption patterns. Similarly, the National Academy of Sciences and the Federation
of American Societies for Experimental Biology have sponsored recent fundamental
research on means of obtaining usual food consumption rates. On initial examination, the
results of this research are encouraging. In this project we will select the most promising
of the statistical procedures developed, make appropriate modifications, and address the
long-term consumption rates for foods of particular EPA interest. Much existing research
is oriented toward evaluating human nutrition, and alternate approaches may be required
for evaluating foods consumed with low frequency.
A particular issue to be examined is fish consumption rates. Work is underway to
compile recent data on fish consumption rates, which will form the starting point for
improving fish consumption estimates. EAG is cooperating with the USDA in developing
questions for a planned survey of national food consumption patterns. This work may
yield an added bonus for this project by providing a means to define "regular consumers"
of particular food products, allowing survey data to be separately broken out for these
groups. In this case, some of the analytical effort in this project will focus on evaluating
these results.
Short term: Report on appropriate methodology for estimating accurately long-term food
consumption rates.
Long term: Improved estimates of food consumption rates used in Agency risk
assessments.
FY90 - report on statistical methodology for addressing EPA needs for long term food
consumption data.
FY91 - recommended methodology of assessing fish and seafood consumption rates.
FY92 - report on modifications to short surveys for improving the surveys' ability to
identify regular consumers of food products.
As indicated above, food consumption rates are central to many important risk
assessments conducted by EPA. In some cases large uncertainties in intake rates are
important factors in the uncertainty in the bottomline risk estimate. For example, fish
consumption rates ranging from a few grams per day to a few hundred grams have been
used in assessments for contaminated water bodies. This research will lay the groundwork
for developing methodologies by which reliable and appropriate consumption estimates
can be made. In addition, explicit analysis of data on fish consumption and consumption
of other selected food will aid immediate Agency risk assessment needs.
-------
RIHRA
Appendix B - 9
Project Length
and Cost:
Project Officer:
2 yi FY90: $100K
P. White (OHEA-EAG) (202) 382-2589
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Integrated Exposure Assessment
2.3 Indirect Exposure
New
Measurements of Exposures to House Dust and Soil Ingestion
The long-term goal of this project is to accurately measure the contribution of house dust
and soil ingestion to children's exposures to pesticides and metals. The amount of
household dust and soil ingested by toddlers may be a major contributor to their exposure
to certain metals and pesticides. Little is known concerning either the amount that they
ingest or the levels of metals and pesticides in household dust One reason is the absence
of reproducible sampling methods for household dust. Recent work by EMSL-RTP (now
AREAL) has resulted in the development of a sampler that collects reproducible amounts
of dust while preserving the integrity of the sample for later chemical analysis. These
samples will be used in a study—under controlled conditions—of several homes selected
to represent a range of possible environmental conditions. Children will be observed
during their household activities and areas where they play will be sampled before and
after play. Amounts of dust and soil on hands will also be measured and analyzed for
elemental and chemical content.
Short term: Report on construction and monitoring protocol for deployment of a house
dust sampler.
Long term: Report of measurements obtained in a pilot test of the methodology. (Further
long term research will not be funded.)
FY90 - report on the measurement of pesticides in house dust
The reports will be incorporated into scenarios of total human exposure, particularly the
exposure of children, through air, water, food, and dust ingestion for those pollutants
(metals, pesticides) that have several different pathways of exposure.
1 yr FY89: $150K
A. Bond (OMMSQA; AREAL-ATP) (919) 541-4329
-------
RIHRA
Appendix B -10
Topic:
Issue:
Status:
Title:
Description:
Identified Results: *
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Integrated Exposure Assessment
2.3 Indirect Exposure
New
Short-Term Peak Exposure
This project is designed to develop an estimate of short-term peak volatile organic
compound (VOQ exposure. Several studies have indicated that short-term peak exposures
to many pollutants may be extremely important contributors to total exposure. For
example, a single hour spent stripping paint has been calculated to result in exposure to
as much methylene chloride as a lifetime exposure to ambient levels. Clearly, such
extreme exposures need to be accounted for in any realistic risk assessments.
Therefore, this project is designed to measure peak short-term exposures to a variety of
pollutants in the most important microenvironmental situations. Emphasis will be placed
on the most common microenvironments (measured in terms of millions or even hundreds
of millions of persons visiting those microerivironments per year), and on the highest
expected peak exposures, in order to obtain maximum efficiency.
The first pollutants to be studied will be a set of VOCs, including chloroform, other
chlorinated compounds, and aromatic and aliphatic hydrocarbons. Crucial microenviron-
ments for these substances are the shower (50% of a person's inhalation exposure to
chloroform occurs during the 10 minutes in the shower.McKone, 1987); and use of
household products (e.g., spray cleaners, air fresheners). Each of these activities will be
studied in a careful quantitative way to allow much more precise estimation of exposures.
Short term: Based on available data, estimate the short-term peak to mean ratio for the
VOCs measured in TEAM studies. Long- term research in this area will not be funded.
FY90 - report on results on the use of new short-term methods in measuring peak
exposures and peak levels observed in selected microenvironments.
These studies will provide data for use in risk assessments. These data can be input into
models based on activity patterns, time budgets, and microenvironmental concentrations
to estimate the proportion of total exposure contributed by the given microenvironment
under different exposure scenarios.
1 yr FY89: $125K
G. Evans (OMMSQA; AREAL-RTP) (919) 541-3124
Topic:
Issue:
Status:
Integrated Exposure Assessment
2.3 Indirect Exposure
New
-------
RIHRA
Appendix B -11
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Measuring Contribution of Exposure from Attached Garages
This project is designed to develop accurate estimates of the source contribution
originating from attached garages emitting into the indoor residential microenvironment.
The importance of the attached garage as a route of exposure to a number of toxic and
carcinogenic compounds has only recently been understood. Many people store gasoline,
kerosene, cleaning products, paints, pesticides, lawn mowers, snow blowers, and other
potential emitters of noxious vapors in their attached garages, in addition to the
automobile itself. A recent study of benzene exposure has identified the attached garage
as a possible source of the elevated concentrations observed in homes.
An attached garage will be stocked with the products under investigation and
measurements will be taken both in the garage and in the living quarters to determine the
impact of each source. Measurements will be made with canisters for VOCs and PUF
cartridges for pesticides. Qualitative and quantitative identifications of a large number of
compounds will be made to apportion sources using established "signatures" for each
major source category. (Initial work to establish such signatures will be done for those
source categories for which information is inadequate.) Air exchange between garage and
home, as well as between the garage and the outdoors and the home and the outdoors,
will be carried out using the multiple-tracer PFT technique developed at Brookhaven.
Actual emission rates of some or all of the major source categories will be measured in
headspace or chamber studies, to compare with calculated emission rates using the air
exchange information. Studies of effect of evaporative emissions will also be undertaken
to estimate the contribution of the car to overall exposure levels.
Short term: (FY90) a report of the relationship between emission rates, garage-home air
exchange rates, and human exposure will be established for a variety of garage scenarios.
This project will provide realistic estimates of the impact of an attached garage on indoor
air quality levels. In turn, these estimates will improve total human exposure estimates
for use in Agency risk assessments.
1 yr FY89: $75K
R. Highsmith (OMMSQA; AREAL-RTP) (919) 541-7828
-------
-------
RIHRA
Appendix C -1
APPENDIX c
TOPIC 3: PHYSIOLOGICALLY BASED PHARMACOKINETIC MODELS
PROJECT DESCRIPTIONS
Note: No description is provided here for Project 3.1.1(6) because it is a support, not a research, task.
Topic: Physiologically Based Pharmacokinetic Models
Issue:
Status:
Title:
Description:
3.1 Experimental Absorption and Biological Parameter Data
Ongoing
Experimental Studies for Models of Dermal Absorption of Chemicals
The goal of this project is to improve the accuracy of predictions of dermal absorption
and disposition for chemicals of interest to the Agency. Parallel experimental studies will
be conducted in animals and in humans to examine and evaluate models for predicting
human skin penetration.
Any model—whether animal or mathematical—intended for use in a regulatory setting
must have a sound biological foundation. This study therefore 1) examines the
relationship between factors such as chemical structure and skin absorption and 2)
compares the output from existing models for quantifying the dermal penetration with data
gathered from human volunteers. The models examined include:
• model skin lipid membranes (rotating diffusion cell)
• in vitro methods using mouse and rat skin
• rat in vivo model
• human skin grafted athymic nude mouse model
Initially, a series of para-substituted phenolic compound will be used that were selected
as part of a Cooperative Agreement with the University of San California at San
Francisco. These will be used to examine the relationship between chemical structure and
skin absorption, including the substituents for hydrophobic (Hansch), electrostatic
(Hammett), steric (Taft E), dispersive (molar refractivity), and dermal penetration.
One of the critical parameters involved in the assessment of risk is the potential for
exposure of the target population to the agent(s) being evaluated. The dermal route of
exposure is the major pathway for many environmental compounds, especially pesticides.
At present, there are no generally accepted laboratory models for the determination of
potential human dermal absorption. Proposed methods include quantitative structure
activity relationship (QSAR) models (Project 3.3.2(1)), in vivo animal models (using a
variety of animals species), and in vitro systems that rely on both viable and nonviable
animal and human skin. The usefulness of these models for extrapolating human exposure
levels has been questioned because of:
• the role of large interspecies differences in skin structure
• the lack of relevant metabolic capabilities in the in vitro methods
This project has been designed to collect data using the most promising methods for
assessing skin absorption so that the accuracy with which the various approaches predict
human response can be evaluated. In addition, the usefulness of the QSAR approach and
-------
RIHRA
Appendix C - 2
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
other factors in predicting dermal penetration and improving understanding of the
mechanisms of skin transport will be examined.
Short term: Evaluation of methods to predict skin penetration in humans
Long term: Assessment of QS AR approach for predicting dermal penetration of phenols
and other classes of chemicals; identification of factors (e.g., dose, chemical form)
associated with predicting dermal ingress.
These results can be used by all Program Offices in the Agency that are involved in risk
assessment to increase the precision of target tissue dosage estimation.
FY89 - reported results on the model lipid membrane system with phenols, the in vitro
mouse skin studies, and the solid state studies
FY90 - reported results on the dermal penetration of the substituted phenols in rats
FY91 - reported results on the high-molecular-weight studies and on the dermal absorption
of phenols in humans
FY92 - reported results on the dermal ingress of phenols in the human skin grafted
athymic nude mouse model
Hazard evaluation is often made on the basis of structure-activity relationships. This
research effort is exploring biological and mathematical approaches for predicting
percutaneous penetration on the basis of chemical structure and properties as well as
generating a database that can be used in constructing PB-PK models for the phenols.
This task links with 3.3.1(2) by providing pharmacokinetics that will assist in interpreting
the embryonic dosimetry results.
3 yr FY89: $330K
L. Hall (OHR/HERL) (919) 541-2774
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
Ongoing
Experimental Deposition in Man of Inhaled Insoluble Particles for Development and
Validation of PB-PK Models
This project will provide a comprehensive database on the deposition and fate of inhaled
particles in humans. The output of this database will feed into pharmacokinetic models
of inhaled particles (Project 3.3.1(2)), which are intended to provide quantitative
predictions of the dose of toxic particles to the lungs and other organ systems in humans.
-------
RIHRA
Appendix C - 3
Identified Results:
The three primary components of the study are:
• Experimental determination of lung mucociliary clearance of inhaled insoluble particles
in human volunteers
• Experimental determination of alveolar clearance of inhaled insoluble particles in
human volunteers
• In vitro studies of the dissolution rates of toxic particles
Efforts under the first component will examine the patterns of mucociliary clearance of
inhaled particles as a function of particle diameter in a population of healthy human
volunteers. Mucociliary clearance of inhaled particles occurs within the first day
following an acute paniculate inhalation. The data gathered will help explain particle
retention as a function of time during the first hours following exposure. The retention
of particles 24 hours after inhalation will also be measured; particles retained for that
length of time are deposited in alveolar regions of the lungs.
The particles the subjects will inhale will be either ferric oxide or teflon particles labelled
with the radioisotope Tc-99m. Following inhalation, a gamma camera that provides a
two-dimensional representation of the activity distribution will be used to measure the
radioactivity retained in the lungs. The initial particle deposition distribution and
clearance patterns can thus be measured simultaneously.
Activity in the head and stomach will also be measured as a function of time to quantitate
translocation of particles from the lungs and extrathoracic airways to the stomach. The
technology required for this work has already is already available at the Clinical Research
Branch. Transport rates of mucus within individual airways will be assessed by placing
isolated quantities of particles within individual airways either through introduction of a
fiberoptic bronchoscope or by inhalation of aerosol boluses. The latter is the more
desirable method because it is noninvasive; however, the results need to be confirmed.
Work on the second component of this task will examine the slower patterns of alveolar
clearance of inhaled particles as a function of particle diameter in a population of healthy
human volunteers. Longer-lived radioisotopes such as In-111 will be used because
particle retention will be measured over a period of several days rather than several hours.
More sensitive radiation detectors will therefore be required because smaller amounts of
initial activity will be used.
Activities associated with the third component of this study will determine the rates of
dissolution and release of toxic compounds that adhere to toxic particle surfaces. These
studies will be conducted in vitro utilizing fluids that are either obtained directly from
human volunteers by bronchoalveolar lavage, or manufactured and tested for
biocompatibility with bronchoalveolar lavage fluid. In vivo studies may also be used to
measure the disappearance rates of soluble particles and to determine the remaining
fractions of compounds in particle and dissolved states. This last component of the study
will provide the final link in the chain of events between inhalation of particles and
release of constituent compounds within the lungs.
4th Q/89 - Interim results from aerosol deposition and mucociliary clearance study and
from the particle dissolution study. These data will provide important information to
program offices on establishing chronic doses to tracheobronchial epithelium from inhaled
particles of various solubilities. They will also comprise important input into the
-------
RIHRA
Appendix C - 4
pharmacokinetic modeling project (3.3.1(2)), which will enable accurate predictive
dosimetry.
FY90 - Final data from aerosol deposition and mucociliary clearance study and from the
particle dissolution study. Interim data on alveolar deposition and clearance of particles.
The data, even though available only for a limited range of particle sizes, will provide
program offices with vital information on the residence times of inhaled particles in the
gas exchange regions of the lungs—the site of high chronic particle lung burdens. This
information will thus lead to better estimates of the dose of chronically inhaled particles.
Also, the data will be used as input into the pharmacokinetic modeling project (3.3.1(2))
to further enhance the accuracy of predictive dosimetry.
FY91 - Final data on the deposition and alveolar clearance of inhaled particles over a
complete range of particle sizes. Both directly, and through the pharmacokinetic
modeling, these data will enable program offices to predict the dose of inhaled particles
to various regions of the respiratory tract as a function of particle size, breathing pattern,
route of inhalation, and particle solubility.
Usefulness of Results: For a known ambient particle concentration of a toxic aerosol, our ability to predict the
potential toxic effects in humans depends on our knowledge of the initial distribution and
subsequent fate of these particles within the lungs. The initial distribution of inhaled
particles is highly dependent on their size and the pattern of breathing. The fate of
deposited particles depends on the region of the lungs in which they are deposited and on
the physicochemical properties of the particles. Particles deposited in the conducting
airways of the lungs are transported cephalad toward the larynx by mucociliary transport,
at which point they are swallowed. Particles deposited in the gas exchange region of the
lungs are cleared from this region by a variety of processes. Particles can be engulfed by
macrophages and translocated either to the ciliated airways of the lungs or into interstitial
spaces. Pinocytosis by alveolar epithelial cells is also possible, as is direct translocation
of free particles either to the ciliated airways or into the interstitium and blood.
Overlaying all of these processes is the mechanism of particle dissolution, whereby the
particles decompose into their chemical constituents. The residence times of particles and
their chemical constituents in different regions of the lungs—and the rate of their delivery
to other organ systems—depend on the detailed kinetics of the particle transport and
clearance mechanisms. Although some experimental information has been gathered on
particle clearance mechanisms, it is not sufficient for providing accurate quantitative
dosimetric information. Data resulting from this project will include particle loads within
respiratory tract compartments as a function of particle and breathing parameters, as well
as clearance rate constants for these compartments. Particle dissolution is also included
as a clearance compartment. These data will form the basis of modeling the chronic
retained dose of soluble and insoluble particles. The models will provide predictive dose
information on compounds in paniculate form and thereby improve heath risk assessment.
Project Length
and Cost:
Project Officer:
3 yr FY89: $130K
T. Gerrity (OHR/HERL) (919) 541-2567
-------
RIHRA
Appendix C - 5
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
Ongoing
Dosimetry of Hygroscopic Aerosols in Humans
This project will provide data on the respiratory tract deposition of hygroscopic aerosols
in spontaneously breathing humans. This information will then be used to validate models
of hygroscopic aerosol deposition—which can, in turn, be applied to animal species and
thus serve as an extrapolation link for acid aerosol dose between animals and humans.
Respiratory tract deposition of insoluble aerosols in humans is measured by comparing
the particle number concentration between inspiration and expiration. The instrumentation
required to accomplish these measurements requires that the particle size between
inspiration and expiration not change. For insoluble aerosols, this restriction does not
pose a problem. Hygroscopic aerosols, however, can alter in size as they traverse the
humid environment of the respiratory tract. Standard methodologies for measuring
particle deposition in the respiratory tract cannot therefore be applied. A new technique
of aerosol characterization that simultaneously—and in real time—measures particle size
and concentration is currently under development at the University of Arkansas under
funding from EPA. This methodology should be applied to the measurement of
hygroscopic aerosol particle deposition in the human respiratory tract. Emphasis will be
on acid aerosols because these compounds are ubiquitous in urban air and are of
immediate regulatory interest.
This research effort will be multistaged. The first set of experiments will measure the
deposition of aerosols, such as sulfuric acid, in humans inhaling through a mouthpiece.
The deposition studies will examine factors affecting deposition: tidal volume, ventilatory
flow rate, temperature and RH of the aerosol, and acid species. These initial studies will
help to test and refine models of acid aerosol deposition currently under development at
HERL.
The second set of experiments will measure the deposition of hygroscopic aerosols under
conditions that better simulate natural spontaneous breathing. This phase requires the
development of an aerosol delivery system that will allow subjects to breathe orally,
nasally, or oronasally, and that will simultaneously measure the inspired and expired
concentrations of aerosol. These data, which will be of more direct regulatory
significance because the breathing conditions will be more natural, will provide a picture
of the population variability of deposition.
In a third set of experiments, the subjects will be exposed to hygroscopic aerosols in a
chamber so as to induce acute responses such as retardation of mucociliary transport and,
possibly, bronchoconstriction (in asthmatics only). Hygroscopic aerosol deposition in
these subjects will be measured simultaneously so that a relationship between delivered
dose and response can be determined.
4th Q/89 - Data on the deposition of sulfuric acid aerosol in humans as a function of
flow, tidal volume, temperature, and RH during mouthpiece breathing. These data will
allow direct computations of sulfuric acid doses to respiratory tract epithelium under a
variety of ambient conditions that could affect particle characteristics.
-------
RIHRA
Appendix C - 6
Usefulness of Results:
Project Length
and Cost:
Project Officer:
FY90 - Data on the deposition, by mouthpiece, of acid aerosols as a function of acid
species. Data on the respiratory tract deposition of acid aerosols by spontaneously
breathing humans inhaling the aerosol orally, nasally, or oronasally. This information will
facilitate better quantitation of delivered acid aerosol doses based on more physiologically
normal conditions for breathing.
FY91 - Data on the relationship between deposited acid aerosol dose and acute health
effects. These data will help the Office of Air Quality Planning and Standards to assess
the variability of delivered dose and of response in the population at large; as a result,
better predictions will be made of the population impact of acid aerosol health effects.
Experiments in humans have shown that exposure to low concentrations of sulfuric acid
aerosols can produce acute retardation of mucociliary transport in the periphery of the
lungs. Low concentrations of sulfuric aerosol may also induce bronchoconstriction in
persons with asthma. Animal studies have shown similar results with respect to
retardation of mucociliary transport; in addition, prolonged exposure has been shown to
permanently depress mucociliary transport. Because such transport of inhaled particles
is an important lung defense mechanism, this depression could have significant health
implications. To extrapolate the chronic animal study data to humans, estimates must be
made of the dose of acid aerosol delivered to lung tissue in both humans and animals.
Little is known of the dosimetry of hygroscopic aerosols in unencumbered, spontaneously
breathing humans. One reason for the paucity of data is the aqueous nature of the
aerosols: during respiration, they change in size and thus the manner in which these
particles are distributed within the different compartments in the lungs also changes.
Recent technological advances in aerosol measurement now allow simultaneous
measurements of particle size and number, so that aqueous aerosol deposition can now
be studied. Data from this project will be of extreme interest to the Office of Air Quality
Planning and Standards for estimating the risk associated with acid aerosol exposure in
humans. These data will allow the extrapolation of data from chronic acid aerosol
exposure in animals to humans. Moreover, advancements in understanding the deposition
of hygroscopic particles will be of interest to other Program Offices as well.
3 yr FY89: $100K
T. Gerrity (OHR/HERL) (919) 541-2567
Topic:
Issue:
Status:
Project Code:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
Ongoing
3.1.1(5)
Permeation Coefficients and Comparison of Dermal Absorption Models
PB-PK models that account for dermal absorption by two different approaches. First,
sufficient data and modeling may be available for the compound after exposure by another
-------
RIHRA
Appendix C - 7
Identified Results:
Usefulness of Results:
route. In this case, a forcing function is written as the mathematical description
characterizing the blood concentration after dermal absorption. This approach would
require laboratory experimentation to determine the blood levels after dermal exposure.
The necessary parameters, however, are easily determined. Models derived in this manner
would be useful to risk assessors attempting to estimate delivered or target dose in
animals for the dermal exposure situation when the bioassay experiments were performed
by another route. The potential difficulty in extrapolating from one species to another
and, in some cases, from one dose to another is the primary disadvantage of this
technique.
A second and more global approach is to write PB-PK models that describe the
physiology and thermodynamics of the dermal absorption process. Even for the simplest
of models, the key required parameter is the permeation coefficient across the skin.
Although numerous reports have described methods for determining this parameter, but
several questions remain. It is not clear, for example, which in vitro model is best for
determining the in vivo permeation coefficient. Nor is it apparent whether animal or
human skin is best suited for determining these coefficients. Further, if animal skin is
chosen, which skin is a good surrogate for the human condition? Or, if human skin is
used, from what region of the body should it be taken, or should cultured skin cells be
used instead? Once a proper method is chosen, these coefficients can be accurately
determined and then used in PB-PK models. A framework is necessary that outlines the
appropriate method for determining the necessary parameters for given conditions and
available data.
Several compounds with varying physical properties will be chosen. PB-PK models
describing transfer across the skin and deposition in the body will be written for these
compounds. The literature will be searched for pharmacokinetic data that can be used to
validate and refine the pharmacokinetic models. Following model formulation and
examination of existing data, the experiments necessary for filling data gaps will be
identified and the appropriate protocols, written. On-going research involves reviewing
existing literature and consulting with experts regarding the various in vitro methods of
determining permeation. From this research, a framework will be prepared that indicates
the best methods for determining permeation coefficients for specific compounds,
available data, and exposure conditions.
Several PB-PK models for compounds that can potentially be absorbed through the skin.
A framework for using such models to predict the extent and rate of dermal absorption
and the disposition to target organs after dermal absorption. A synopsis of the various
methods used for determining transdermal permeation and a hierarchy for choosing the
best method.
Describing and predicting the degree of dermal absorption of xenobiotics using current
methodologies is difficult and results in uncertain estimates. It is not clear which
laboratory procedures for either whole animal or in vitro testing are applicable under
which conditions. Further, even with laboratory determinations, it is not clear how to
extrapolate to human exposure conditions. In recent years, PB-PK models have been used
with success to predict disposition of xenobiotics in the body and at potential target sites
within the body. The key parameter in models predicting disposition after dermal
absorption is the permeation coefficient This study will provide methodologies for the
best and most efficient manner of determining permeation coefficients and formulating
and implementing PB-PK models.
-------
RIHRA
Appendix C - 8
Project Length
and Cost:
Project Officer:
3 yr FY89: $50K
K. Hoang (OHEA-EAG) (202) 382-2059
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
New
PB-PK Data and Models for Compounds with Known Multispecies Carcinogenic
Responses
Recently, through an Interagency Agreement, OHEA and the Department of Defense
reviewed exposure and carcinogenesis data for about twenty compounds known to be
carcinogenic in more than one species, including humans. The goal is to determine if
animal-based risk estimates, as they are now performed, are within reasonable limits of
risk estimates based on human data. These data were also used to determine the best dose
metric for extrapolating risk to humans from estimates derived in animals. Often the data
were taken from a variety of exposure scenarios for both animals and humans. Thus,
establishing comparability of exposures is difficult. Establishing an equivalent dose metric
among species requires knowledge of the actual target tissue dose after the various
exposure regimens. With those data, exposures at very different concentrations and with
different regimens could be equated, based on the resultant internal doses. PB-PK models
offer an excellent opportunity to accomplish this task. For most compounds, however,
adequate models have not yet been developed.
Representative compounds will be chosen for formulation of PB-PK models. After a
review of the pharmacokinetic data and formulating models for the compounds, any data
gaps for the compounds (e.g., partition coefficients or metabolic rate constants) will be
identified. Those compounds with the most complete pharmacokinetic data will then be
chosen for complete study. The models formulated and validated for these compounds
will be used to calculate the actual target dose in the various species for which
carcinogenic responses occur. With these data, the best dose metric can be chosen for
each compound. At that point, efforts using both inhouse and extramural resources can
begin to resolve the problem of equivalent dose metric across species. Outside resources
will concentrate on reviewing and obtaining the necessary data, while inhouse staff will
formulate and implement the PB-PK models and apply the results to the question of dose
equivalency. Necessary laboratory experiments will be performed either inhouse or by
outside resources.
A review and collation of the available pharmacokinetic data for several human and
animal carcinogens. PB-PK models for some compounds. These models can be used to
determine the impact of different exposure regimens on risk. As a result, the importance
of pharmacokinetic nonlinearities—which may not be identical at different exposure levels
and regimens—can be accounted for in future site-specific risk assessments. Ultimately,
this project would begin to resolve the question of equivalent dose metrics between
species and would also provide a design for future study of this problem, even with
endpoints other than carcinogenesis.
-------
RIHRA
Appendix C - 9
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Perhaps the greatest controversy concerning risk assessment techniques involves the
uncertainty with which risk assessors extrapolate between species. No method for
equilibrating dose between the test species and a species at potential risk (usually the
human) has ever been universally accepted. The problem is particularly difficult because,
even in those few cases in which response data exist, the actual dose to the target tissue
is not known. Another complicating factor is the lack of equivalence between the
exposure regimens for each of the tested species and the species at risk. This project will
forge a reliable methodology for determining target tissue doses under the various
exposure conditions, thus allowing comparison of response vis-a-vis dose across a myriad
of species.
3 yr FY90: $75K
P. White (OHEA-EAG) (202) 382-2589
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
Ongoing
Pharmacokinetics of Inhaled and Ingested Volatile Organic Compounds
The objective of this project is to develop and validate PB-PK models intended to
accurately predict the concentration of halocarbons in blood and tissues over time
following oral or inhalation exposure. The models can be used in dose-to-dose, route-to-
route, and species-to-species extrapolations; thus, they can be used to estimate the internal
organ/tissue exposure from a known or hypothesized external exposure scenario for
humans. The health effects of this internal exposure can then be estimated using the
human organ/tissue dose-response relationship.
PB-PK models have been developed and validated on the basis of results from:
Intravenous injection
Inhalation exposure (at two vapor concentrations)
Oral bolus
Gastric infusion (at doses equivalent to inhalation exposure)
Multiple oral doses
With these methods and results as a basis, studies will be performed to delineate the time
course of tissue levels in rats during pulmonary and oral exposure to several different
halocarbons at equivalent doses. Empirical measurements will be made of important
physiological parameters, e.g., in vivo tissue-blood partition coefficients, clearance values,
metabolic constants, and tissue volumes (especially in adipose tissue). PB-PK models for
inhaled or ingested halocarbons will be developed, validated, and modified on the basis
of the database obtained. Validation of the models will be done with low-level exposures,
including simultaneous oral and pulmonary exposure. Once the models have been
validated, the data will be extrapolated to human exposures, and the results compared with
existing data on the chosen compounds.
-------
RIHRA
Appendix C -10
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
A large pharmacokinetic database, consisting of inhalation-and-ingestion absorption and
distribution data for several volatile organic compounds. Validated pharmacokinetic
models, based on both oral and pulmonary exposure, together with their extrapolations to
humans and comparison with human data. A set of measured physiological parameters
important in PB-PK modeling. Finally, a framework for developing a complete
pharmacokinetic approach that incorporates both ingestion and inhalation exposures.
The validated PB-PK models produced will be used as input into biologically based dose-
response (BB-DR) models and thus for health risk assessments. The models can be used
in high-to-low-dose, route-to-route, and species-to-species extrapolations, and thus to
estimate the internal organ/tissue exposure form a known or hypothesized external
exposure scenario for humans. The health effects of this internal exposure are then
estimated using the human organ/tissue dose-response relationship, which may be based
on animal experimental data, animal PB-PK models, and human-animal relative sensitivity
from (often) tissue and cell culture experiments. Validated PB-PK models enable the
extrapolation of experimental results from animals in laboratory controlled environments
to humans in ambient uncontrolled environments, thereby reducing the uncertainty in
health risk assessments.
3 yr FY89: $117K
T. Gerrity (OHR/HERL-RTP) (919) 541-2567
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Physiologically Based Pharmacokinetic Models
3.1 Experimental Absorption and Biological Parameter Data
New
Database Synthesis of Existing Biological Parameters
This project uses available expertise to develop consensus opinions on the values of
certain physiologic parameters commonly used in PB-PK models. As a result of an
Interagency Agreement, staff at Oak Ridge National Laboratories have summarized
literature values for physiological parameters such as organ volumes, blood flows, and
body weights for several mammalian species. The parameters chosen generally fit into
the simpler PB-PK models. This project funds the EPA portion of an Interagency
Agreement with the Food and Drug Administration (FDA) and the Department of Defense
(DOD) to empanel experts selected by the American Physiological Association to review
the Oak Ridge work, expand the chosen parameters, and arrive at a consensus of ranges
for the values of the parameters. The panel, which will be composed of national experts
in physiology, will issue a report.
Report detailing range of values for physiologic parameters such as blood flow and organ
volume in several mammalian species that are commonly used in lexicological testing.
Having standard, peer-reviewed values for physiologic parameters used in PB-PK
modeling would greatly improve confidence in the models. The goal of pharmacokinetic
modeling is the reduction of the uncertainties inherent in risk assessment. The models,
-------
RIHRA
Appendix C -11
Project Length
and Cost:
Project Officer:
however, have their own intrinsic uncertainties. Accurately establishing as many of the
model parameters as possible is vital to the models' usefulness. Some parameters are
case-specific, while others (such as the physiologic parameters, have ranges that can be
narrowed significantly. To date, examination of the literature has revealed no consensus
is available on appropriate ranges for these parameters. This project will establish such
a consensus, thus reducing the model variation that results from the arbitrary selection of
parameter values.
1 yr FY90: $15K
W. Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.2 Route-to-Route Extrapolation
New
Guidance on Limitations and Use of Route-to-Route and Other Extrapolations in EPA
Risk Assessments
PB-PK models have become useful tools for estimating target-tissue doses and thus
represent a critical advance in risk assessment methodology. Based on the premise that
a particular target-tissue dose will have the same biological effect as an equivalent target-
tissue dose achieved by another exposure regimen, these models may be used to predict
delivered doses under conditions of exposure different from those initially tested,
including different:
• Routes of administration
« Exposure concentrations
• Time patterns of exposure
Risk assessment extrapolations such as these require equations and biologic parameters
sufficiently detailed to describe the differences in absorption by various routes and to
account for the structural and physiologic differences between species.
Panels of experts will be commissioned to:
• Discuss the critical parameters, assumptions, and limitations related to route-to-route
extrapolation
• Provide specific guidance on the use of PB-PK models and associated algorithms in
risk assessments
• Recommend research areas in which research efforts would facilitate future
extrapolations
The research recommendations and guidance arising from this exercise will be
incorporated into ongoing ORD research programs and risk assessment methodologies.
-------
RIHRA
Appendix C -12
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Guidance in the form of a report on the use of PB-PK models and algorithms,
incorporating key parameters identified by consensus. Recommendations on research
areas that will facilitate the various extrapolations required in risk assessment (e.g., route-
to-route, high-to-low exposure concentrations, and intermittent dosing regimens to chronic
scenarios) for use in focusing ORD research programs on filling critical data gaps.
Databases used in risk assessments often lack sufficient lexicological and pharmacokinetic
information for the route, dose rate, and regimen of exposure relevant to the human
exposure scenario. Consequently, current methodologies use various assumptions and
default values to extrapolate to the desired situation using other available data. These
approaches neither systematically incorporate known anatomic and physiologic interspecies
differences in the routes nor utilize pharmacokinetic or pharmacodynamic properties of
the agent and species in question. In addition, the approach for such extrapolations often
differs across Program Offices. Improved approaches are needed that model the
absorption/distribution/retention processes in sufficient detail to estimate the target-tissue
dose for various exposure scenarios. As a result of such work, the Agency will not have
to request extensive, time-consuming, and often expensive disposition studies for each
agent and scenario of interest.
This project will provide expert scientific input to achieve consensus on the criteria,
assumptions, and key parameters and limitations of application that are requisite for
performing route-to-route or other extrapolations in risk assessments. Guidance in the
form of a state-of-the-art report that explicitly defines limitations on application will
ensure consistent use across Program Offices. Recommendations on research that would
facilitate such extrapolations and strengthen the existing assumptions or eliminate default
values will help to focus ORD research programs and ensure that Agency risk assessment
methodologies evolve with scientific credibility.
2yr FY89:$75K
T. Gerrity (OHR/HERL-RTP) (919) 541-2567
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
New
Regional Retained Dose Ratios for Adjustment of Respiratory Tract Burdens Across
Species and Use in Inhalation Reference Dose Methodology
Empirical data and existing theoretical models will be systematically analyzed, with an
emphasis on aerosol exposures, to determine the anatomic and physiologic parameters
necessary for the quantitation of an inhaled dose. Development of a model that
incorporates factors to account for the regional (e.g., extrathoracic, tracheobronchial, and
pulmonary) deposition in the respiratory tract as well as the fate of the inhaled particles
over time will be based on these parameters. Factors that will be incorporated, as
feasible, to account for the fate of inhaled relatively insoluble particles include:
-------
RIHRA
Appendix C -13
Identified Results:
Usefulness of Results:
• Mucociliary transport and clearance rates
• Alveolar clearance rate (phagocytosis/translocation by macrophages/dissolution/tree
particle translocation)
• Particle solubility
• Chemical activity
Certain data gaps for individual species will be filled empirically by contract—for
example, consistent methodological procedures will be used to determine surface areas and
clearance rates of the respiratory regions to ensure compatible and precise estimates for
model input across species.
Output of the proposed model will be estimates of retained doses in the various regions
of the tract for a given species used experimentally. Calculation of the ratios of these
regional retained doses between a given experimental species and humans provides a
factor for scaling an aerosol concentration to which the animal was exposed for
interspecies differences in retention of the inhaled dose; thus, it will then be possible to
more accurately estimate the equivalent exposure concentration for humans.
Specifications on how to apply these ratios as scaling factors (such as guidance on which
ratio is appropriate for an observed toxicologic effect either locally hi the lungs or
systemically) and limitations on their application will be explicitly stated. The
characterization of anatomic and physiologic parameters and the filling of critical data
gaps across species, involved in the development of the aerosol model, will provide the
basis for mass transport estimates needed to expand and refine existing gaseous uptake
models. A gas and vapor model that accounts simultaneously for characteristics such as
solubility, reactivity, and metabolic transformation will then be developed.
Short-term - An analytical model for derivation of regional deposited dose ratios by
which animal experimental exposure concentrations can be scaled to human equivalent
conditions for accurate inhaled aerosol dose estimation used in calculation of reference
doses. The model will be used to generate regional retained dose ratios in a tabular
format as part of a support document to the Inhalation Reference Dose Methodology. The
technical support document will also provide guidelines on the application and explanation
of the limitations for the procedural use of these ratios in the RfD methodology.
Long-term - Sensitivity and interaction analyses of the model parameters versus time to
provide insight for supporting guidance on the limitations of acute versus chronic
extrapolations. Analysis of the effect of concentration on the parameters to assess high-to-
low dose extrapolations. Similar model development for various gases according to
characteristics (and their combinations) such as: solubility, reactivity, and metabolic
transformation. (Because the majority of agents of interest to Program Offices are gases,
development of this model is urgent.) Support documents for inhalation RfD
methodologies giving guidelines on dose adjustment for gas and vapor agents.
The Agency's intention to use reference doses as the principal values in assessing risks
for noncarcinogenic health effects critically depends on accurate extrapolation of inhaled
doses from animals to humans. The following, in other words, are needed:
• Precise estimates of the retained dose received by an animal species in a given
exposure concentration
• Accurate extrapolation of that animal dose to an exposure concentration that would
result in an equivalent retained dose in humans
The modeling effort under this project will provide ratios of precise regional retained dose
estimates in animals to those of humans for use in scaling experimental aerosol exposure
-------
RIHRA
Appendix C -14
Project Length
and Cost:
Project Officer:
concentrations to human equivalent conditions. These ratios thus provide the required
dose estimates and extrapolation necessary for the operational derivation of the reference
dose for accurate risk assessment of inhaled aerosols. The modeling effort will initially
focus on aerosol deposition and fate because vast amounts of data and many theoretical
models currently exist due to past efforts to quantify radionuclide exposures. However,
identification and characterization of the anatomic and physiologic parameters necessary
to estimate deposited doses of aerosols will support similar model development for mass
transport of gases and vapors to more precisely estimate the retained doses of these
agents. Accurate estimation of an inhaled dose and knowledge of governing parameters
is also necessary before any dose partitioning for analysis of mixtures can be
accomplished, a procedure that will be needed for analysis of risk on a source category
basis. Analysis of the interaction of these parameters over time will provide insight on
the limitations for acute versus chronic applications, while analysis of how certain factors
are affected by concentration (e.g., clearance rates) will provide guidance on high-to-low-
dose extrapolations; both are crucial outstanding areas of uncertainty in risk assessment.
3 yr FY90-92: $225K total
T. Gerrity (OHR/HERL-RTP) (919) 541-2567
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
Ongoing
Development of PB-PK Models Following Inhalation, Ingestion, or Dermal Exposure to
Toxic Compounds
This project takes data output from experiments on uptake and clearance of particles that
are inhaled, ingested, and dermally applied. A thorough review of current knowledge
regarding pharmacokinetics of particles administered by any route will be conducted.
Because mechanisms of action vary and often result in first-order linear, second-order
linear, or nonlinear kinetics, no universally applicable PK-PB model can be constructed.
Models for general classes of chemicals, however, can be developed that provide a
framework for work on individual chemicals in the class. In addition, the number of
compartments in the mathematical model is appropriately determined by the level of
understanding of the biological processes involved, the intended regulatory applications
and their societal impacts, and the extent to which the model can be validated. The scope
of work of this project is intended to provide a personnel skill mix to extend the capability
of theoretical PB-PK modeling to OHR research efforts involving oral, dermal, or
inhalation exposure. Current capabilities are largely restricted to dosimetry of inhaled
compounds.
PB-PK models will be developed for various experimental efforts, such as the para-
substituted phenols from dermal studies (Project 3.1.1(1)) and the human inhalation of
• insoluble particles (3.1.1(3)): In addition, for key toxicologic studies that are part of the
OHR base program, some project efforts will encompass the development of PB-PK
-------
RIHRA
Appendix C -15
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
models; this will allow future OHR research programs to provide risk assessors with the
appropriate dosimetry models for evaluating the importance of observed biologic effects.
PB-PK models for various exposure routes (e.g., dermal, oral, inhalation). The PB-PK
modeling for deposition and clearance of insoluble particles will receive initial emphasis
to link with the efforts on inhalation reference doses (Project 3.3.1(2)). Also, the
modeling effort needed to link exposure-dose-developmental biologic effect (links to
3.1.1(1) and 3.3.3(1) will be enacted in FY89.
The assessment of the potential health risk from toxic chemicals depends on animal
toxicology studies establishing an organ dose-response relationship. The extrapolation
from the results of these studies to the potential for health impacts in humans requires the
ability to relate the human exposure conditions directly to organ doses. One of the most
promising means by which such extrapolations can be made is through mathematical
modeling of the pharmacokinetics of compounds that are delivered by the inhalation, oral,
and dermal routes.
To have the predictive power necessary for providing answers to regulatory questions,
these models must be firmly based on quantitative data. All of the projects under Topic
3 will provide databases to support mathematical modeling. Resulting data will permit
predictions of ambient concentrations of particles and vapors to which humans are
exposed and that would result in specific doses of the compounds to systemic organs.
3 yr FY89: $240K
T. Gerrity (OHR/HERL-RTP) (919) 541-2567
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
Ongoing
Development of Mathematical and Physiological Bases for Model Validation, Formulation,
and Sensitivity Analyses
Pharmacokinetic model formulation requires an understanding of physiology,
thermodynamics, biochemistry, and the specific database in question. No clear guidance
is available on how to properly formulate a model for specific conditions. For example,
modelers will often structure a particular model in a certain manner because the available
data cannot be understood in any other way. This project will accomplish several tasks:
• Form a summary of pharmacokinetic models used to describe deposition and discuss
the criteria used for selecting or formulating each model (summaries gathered by
discussion with experts, literature review, inhouse implementation of existing models,
and peer review of reports)
• Formulate guidance on what information must be gathered and what questions must
be answered to formulate a PB-PK model
-------
RIHRA
Appendix C -16
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
• Provide guidance for model validation
» Develop a decision tree approach to give guidance on model formulation in specific
cases (this type of decision analysis lends itself to expert system automation)
During the first year, work will begin on just a portion of the PB-PK model formulation.
In vivo metabolic rate constants are often case-specific and difficult to extrapolate
between species. In vitro generated metabolic rate constants offer promise of ease of
experimentation and the ability to use human tissues for determination. Recently, BAG
initiated research to determine the best way to measure such constants in vitro as well as
mathematical models for incorporation into global PB-PK models. Results suggest that
several models pould be used for incorporating in vitro derived values. Predictive ability
is a function of the extraction ratio of the compound. Depending on the approximate
value of the extraction ratio, differences between the models apparently may or may not
be significant. An expert system will be drafted that will draw on a developed knowledge
base of rules and extraction ratio values stored in the database to reach conclusions
regarding the best mathematical in vitro to in vivo extrapolation model.
Another option is to build an expert system that helps the user decide which type of
pharmacokinetic model is needed for a reconstructive dose/exposure assessment, given
certain site-specific data.
From this work and the PB-PK model summaries, other expert systems would be
developed to help choose other components of model formulation. Ultimately, an
extensive expert system would be constructed that would encompass the several phases
of model building.
For the first two years of the project only:
Short-term - Summaries of types of PB-PK models and the conditions for which they are
appropriate. Prototype expert system to select the best mathematical model for
incorporating in vitro rate constants into in vivo models. Prototype expert system to
decide the type of pharmacokinetic model that is most feasible for a reconstructive
dose/exposure assessment.
Long-term - Guidance on model formulation criteria and expert system to aid in model
formulation and simplification.
This project will help establish a consensus on the best way to formulate and validate PB-
PK models. In addition, the expert system approach will encourage all users to take full
advantage of existing data. Such an approach will ensure that, because of the inherent
logic in expert systems, the validity of assumptions used in model formulation can be
accurately evaluated. Expert systems can simultaneously accommodate qualitative and
quantitative information in the decision-making process.
3 yr FY89: $50K
P. White (OHEA-EAG) (202) 382-2589
-------
RIHRA
Appendix C -17
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
Ongoing
Computational Molecular Models for the Deposition, Disposition, Transformation, and
Activity of Specific Chemical Classes
The action of a xenobiotic chemical in a biological system results from its interaction with
the molecules in the biological system. Its deposition, transformation, disposition, and
activity at its molecular site of action depend on its specific interaction with endogenous
molecules and nonspecific chemical processes. All of these microscopic actions determine
the macroscopic activity of the xenobiotic. The interactions and nonspecific processes that
determine these actions depend on the same physicochemical properties that determine
chemical reactions.
In this project, computational models will be developed based on a description of the
underlying molecular processes that determine the pharmacokinetics and
pharmacodynamics of specific chemical classes. A class is defined as chemicals that
appear to act by the same mechanism, not just by structural similarity. Members of
classes for which relevant data exist, as well as for which we have knowledge of the
underlying biomolecular processes, will be used to initiate the modeling. Emphasis will
be placed on any available information about the molecular processes that determines
differential activity within the class. These models will then be used to predict the
pharmacokinetics and activity for untested chemicals within the class. The transformation
of polyaromatic hydrocarbons to their active intermediates and the stability of these
intermediates will be the relevant molecular process modeled to initiate this effort.
Assessment of the potential for health effects from individual chemicals within a class is
important for Superfund and other Program Offices.
Quantum mechanics, along with methods to compute physical properties, will be used to
model the relevant molecular characteristics. Many of the relevant methods already exist
or are being developed in other tasks.
Another relevant class of chemicals for study by these methods is the short-chain acids
and chemicals that are transformed to short-chain acids (e.g., glycol ethers, phthalates,
alcohols). For these chemicals, the important processes to be modeled are the
transformation deposition and activity. A model for the substituted phenols will be
developed to compare with the results from Project 3.1.1(1). As the project progresses,
other chemical classes may be added.
1. Methods to predict the transformation, stability, and covalent bonding to biopolymers
of the active intermediates of PAHs and similar chemicals. Intermediate journal
articles on these methods and applications to specific chemicals in the class.
2. Models for the transport and activity of short-chain acids and molecules that are
transformed into short-chain acids. These models could be used to perform
assessments of specific chemicals within a class and as a tool for understanding the
health effects of class members.
3. Models for the transport or activity of chemical classes as yet unspecified, particularly
classes of chemicals for which the health effects are receptor-mediated (e.g., dioxins,
some pesticides).
-------
RIHRA
Appendix C -18
Usefulness of Results:
Project Length
and Cost:
Project Officer:
In many cases, risk assessments must be performed without all the requisite data. In some
instances, even data on the physicochemical properties are unknown. Causal
computational models that predict from molecular structure the distribution,
transformation, and molecular-level activity of the members of specific chemical classes
will provide rational input to these assessments. These models will also aid in the design
of a bioassay strategy for the chemical class that rationally treats assessment needs and
should operate in a feedback mode with the development of bioassay data (for instance,
the data on substituted phenols).
The knowledge gained about the underlying mechanism of action will aid in extrapolation
from bioassay data to situations removed in dose regime, species, etc., from the
experimental circumstance. PAHs, the initial class of chemicals for study, are important
for Superfund, Hazardous Waste, and other Program Offices. The class of chemicals that
includes short-chain acids and the chemicals transformed to short-chain acids is of
importance to Toxics, Superfund, and perhaps other Offices. At least two, possibly
related endpoints in this class are of interest for assessment. If other classes of chemicals
are chosen for modeling, both the chemical class and endpoint will be chosen because of
their importance to risk assessment needs.
3 yr FY89: $150K
J. Rabinowitz (OHR/HERL-RTP) (919) 541-5714
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
Ongoing
Embryonic Dosimetry and Pharmacokinetic Studies in Developmental Toxicity
Within the Perinatal Toxicology Branch, a major research project is underway to evaluate
structure-activity relationships for the induction of developmental toxicity in several
chemical classes (para-substituted phenols and branch-chained carboxylic acids) using both
in vivo (Chemoff/Kavlock Assay) and in vitro (whole embryo culture) bioassays.
Biological activities of 27 phenols in vivo and a subset of 12 in whole embryo culture
have been completed and submitted as an Administrator's Item in August 1988. That
work is attempting to link specific molecular parameters (e.g., lipophilicity, electronic
effects of substituents, molar refractivity) with the potential for causing adverse effects
in the conceptus.
Building on this large database, this project will establish basic pharmacokinetic data on
the processing of the congeners in the two systems, obtaining information on metabolism,
maternal plasma kinetics, whole body clearance, and deposition of the congeners at the
target site within the embryo. The structure-activity studies, with their emphasis on cross-
assay comparisons of effects, suffer from the inability to determine exactly what
concentrations are being delivered to the embryo. Potency measurements are now based
on administered dose, and improvements in quantifying the structure-activity relationships
should be expected if the potencies can be expressed on delivered dose. This project will
-------
RIHRA
Appendix C -19
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
provide the resources to establish the link between administered dose, delivered dose, and
biological effect for a well-defined set of chemical congeners.
Physiologically based models of a series of chemical congeners in the maternal/fetal unit.
These models should provide direct evidence of how molecular properties affect placental
passage and embryonic development within the chemical class under study. Using the in
vitro system, this effort will determine the role of metabolism in the induction of adverse
effects, and these data will then be compared to in vivo observations. This work will
produce the first side-by-side analysis of the chemical dosimetry and reactivity in in vivo
and in vitro assay systems. The first chemical class for study is also the subject of
structure-penetration studies for the dermal route of exposure; therefore, the data derived
from the oral studies in this proposal will provide information on route-route differences
in pharmacokinetics.
To improve the methods by which teratogenic risk is currently extrapolated, attention must
be placed on determining those factors that contribute significantly to the interspecies
differences in the teratogenic responses that are observed following chemical exposure.
These factors must be either pharmacodynamically and/or pharmacokinetically based; and
determining their relative importance, even for one or two chemicals, is an enormous
undertaking. This project will provide a unique pharmacokinetic database that relates the
influence of physical-chemical properties in deposition in the maternal/embryonic unit at
the time of critical susceptibility and how that aspect affects the expression of
developmental toxicology. Combining the in vivo and in vitro data will provide important
direct information on the interpretability of the embryo culture system. The project can
be expanded in the future to provide comparable data for the same congeners in other
whole animal and in vitro developmental toxicity assays.
3 yr FY89: $200K
R. Kavlock (OHR/HERL-RTP) (919) 541-2326
Topic:
Issue:
Status:
Title:
Description:
Physiologically Based Pharmacokinetic Models
3.3 Theoretical Models
Ongoing
Pharmacokinetic and Biological-Time Based Models for Teratogenesis
Pharmacokinetic models for fetal concentration, based on published and unpublished data,
have been developed for simulation of maternal-blood concentrations and fetal
concentration of hydroxyurea in rats and primates. When the dose is expressed in terms
of biological time (units of percent of the period of organogenesis) instead of real time,
the estimated potency (response) seems to be consistent with the experimental response
data. Several organic acids (e.g., trichloroacetic acid) have been identified as the
proximate teratogen. Both pharmacokinetic and teratogenic data from both primates and
rodents regarding maternal blood concentration and fetal tissue and fluid concentrations
(sometimes in terms of both parent compounds and reactive metabolites) are available.
-------
RIHRA
Appendix C - 20
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
The apportionment between the fetal fluids and the fetal tissue is probably the result of
the plc,, of the acid. Changes in the cell turnover are influenced by the cellular pH.
Unpublished information will be gathered. Data ,on fetal body weight and amounts of
fetal fluids during the period of organogenesis as well as actual combined fetal response
(instead of divided response) will be gathered. Models will be developed based on
maternal blood concentration, fetal fluid concentration, and pka-influenced fetal
concentration. Simulations will be compared to experimental results concerning fetal
concentration and response when the dose is expressed in terms of biological time. The
use of cell-tumover-based models will be explored as a basis for adjusting the sensitivity
of the fetus with respect to teratogenic responses and, perhaps, in utero exposure to
carcinogens.
Collation of important pharmacokinetic and physiologic data such as fetal body weight
and sizes of fetal fluids during the period of organogenesis and actual combined fetal
response instead of divided response: fetal death, specific incidence of anomalies. PB-PK
models describing disposition in the fetus and pharmacodynamic models using cell-
turnover rates as an endpoint for sensitivity determination.
No pharmacokinetic and pharmacodynamic models relating maternal exposure to fetal
outcome exist. This project will provide a basis for developing a methodology for
rationally assessing the possible damage to the embryo/fetus. This proposal deals with
higher-resolution endpoints than simple placental crossing to assess fetal/embryonic
sensitivity to teratogens and intrauterine carcinogens.
2 yr FY89: $70K
R. Beliles (OHRA-HHAG) (202) 382-5898
-------
RIHRA
Appendix D -1
APPENDIX D
TOPIC 4: BIOLOGICALLY BASED DOSE-RESPONSE MODELS
PROGRAM OVERVIEWS AND PROJECT DESCRIPTIONS
D.I NEUROTOXICOLOGY
Increasing emphasis and efforts have been
placed in the past few years on identifying and
characterizing the hazards of neurotoxic chemicals.
These developments have come about largely as a
consequence of a growing recognition of the number
of chemicals that have been either shown or suspected
to produce behavioral and/or neurological impairments
in humans. Among the evidence gathered are data
from laboratory studies indicating that many chemi-
cals cause a variety of neuronal impairments, ranging
from subtle functional changes to frank structural
alterations. These trends have been well documented
in the scientific literature and have led a number of
expert scientific associations and committees to
recommend an increased commitment to identifying,
characterizing, and regulating neurotoxic chemicals.
Recently, EPA has also focused an increasing
amount of attention on neurotoxicity. In 1985, for
example, the Office of Toxic Substances (OTS)
published a broad set of test guidelines for evaluating
neurotoxic compounds. Many of these guidelines are
now being applied, either through test rules or consent
agreements, to the evaluation of compounds under
section 4 of the Toxic Substances Control Act (TSCA).
Some of these methods are also being applied to the
evaluation of compounds under section 5 of TSCA,
for example through the Data Gaps project of OTS.
Additional guidelines have either been promulgated
(e.g., developmental neurotoxicity) or are under
development (e.g., eleclrophysiology). Intra-Agency
committees that include Neurotoxicology Division
(NTD) scientists are being formed to review existing
guidelines and make recommendations concerning the
need for additional guidelines. The heightened
Agency awareness of neurotoxicity as a health effect
of regulatory concern has highlighted the need for
neurotoxicity risk assessment guidelines. An intra-
Agency workshop that includes NTD scientists has
also recently been formed to develop guidelines for
making risk assessments on the basis of neurotoxicity
data.
D.I.I Neurotoxicology Division Research
The Neurotoxicology Division has developed
and maintained a program of research to provide
scientific support to the activities of the Program
Offices. Broadly speaking, NTD has ongoing pro-
grams to: 1) develop and validate methods for identi-
fying and" characterizing neurotoxic hazards; 2)
provide data on the neurotoxicity of key chemicals or
chemical classes of concern to the Program Offices;
and 3) produce an empirical basis for interpreting the
significance of laboratory results and extrapolating
them to humans. In the course of these activities,
NTD has recognized that the risk assessment process
can be improved by a research program targeted to
address certain recurrent generic issues of concern.
To develop such a program, NTD has both undertak-
en new research directions and refocused a portion of
its ongoing research. A description of recently
initiated and proposed efforts follows.
D.l.l.i Current RIHRA Projects
1. Considerable effort is being focused on the
issue of inter/intra-species extrapolation (Issue 04).
Particular emphasis is being placed on empirically
determining the degree of homology that exists
between human sensory and cognitive processes and
their animal counterparts (Issue 0401). These studies
involve direct laboratory comparisons of human and
animal data using neurophysiological and behavioral
test methods. A successful outcome of these studies
will greatly enhance our confidence that results
obtained from laboratory animals are predictive of
effects seen in humans under similar exposure condi-
tions.
2. A portion of NTD's research also focuses
on a direct comparison of the susceptibility of rats
and hens to organophosphate-induced delayed neuro-
toxicity in order to determine whether rats can be
substituted for the traditional "hen test." Successful
completion of this project will provide an empirical
basis for specifying rats as an alternative and biologi-
cally more significant species for identifying neuro-
pathic organophosphates.
-------
RIHRA
Appendix D - 2
3. Research is also underway to determine
intra-species differences in sensitivity to neurotoxic
compounds (Issue 0403). This research is specifically
designed to compare basic neurobehavioral processes
in F344 and Long-Evans rats using behavioral and
biochemical test methods. This research will be
supplemented by electrophysiological tests of nervous
system function. Because industry and contract
laboratories frequently use F344 rats in toxicity
testing but NTD research has focused almost exclu-
sively on Long-Evans rats, data gathered in this
project will be significant for species comparisons and
cross-species extrapolation. For example, the F344s
are albino rats known to have neurological anomalies,
and Program Offices need to know whether the
neurotoxicity data collected with these rats are indica-
tive of effects seen in other strains of rat such as
Long-Evans.
4. Research is also underway in NTD to
determine the extent to which the manifestations and
mechanisms of neurotoxicity depend on the particular
exposure conditions (Issue 0403). Specifically, this
project is designed to determine the relationship
between level arid duration of exposure in producing
neurotoxicity using behavioral, biochemical, and
morphological test methods. This research will
provide an empirical basis for predicting neurotoxicity
under one set of exposure conditions on the basis of
data available for another set of exposure conditions.
This research will also indicate whether short-term
high-level exposure studies may be substituted for the
more costly long-term low-level studies currently
required by OTS and the Office of Pesticide Programs
(OPP).
D.l.1.2 Proposed Projects
The proposed projects were selected to both
augment and complement current RIHRA efforts.
1. The project on quantitative risk estimation
directly supports multiple RIHRA issues. This project
is specifically designed to evaluate and compare new
methods for quantitatively estimating risk with the
traditional approach of determining No Observed
Effect Levels (NOELs) and then applying uncertainty
factors. Considerable interest has been displayed in
the scientific community and within the Agency in
this type of approach. For example, in Cincinnati
Dourson and colleagues are currently evaluating
alternative methods for quantitative risk estimation,
but these do not include applications to neurotoxicol-
ogy. The results of this project will therefore not
only complement ongoing efforts within the Agency,
but will also provide a basis for further applications
and evaluations using additional methods for assessing
neurobehavioral integrity.
2. The project on homologous models of
sensorimotor function is specifically designed to
compare the results of reflex testing in humans and in
animals. The particular sensorimotor reflex chosen
for this study has already been shown to occur in a
wide range of species and to be sensitive to a variety
of toxicants in laboratory animals. Owing to its
noninvasive nature, the technique offers great promise
for rapid and accurate neurotoxic hazard identification
and characterization in human populations. This
project will complement current efforts to evaluate
homologies in sensory processes using electrophysio-
logical techniques, and in cognitive processes using
behavioral techniques.
3. The project on the role of age as a determi-
nant of the sensitivity and severity of effects of
cholinesterase-inhibiting pesticides was designed to
specifically address a research need recently identified
by the Technical Panel on Cholinesterase Inhibition to
the Risk Assessment Council (12 September 1988
memo from P. Preuss to V. Newill). This project will
provide needed information on the importance of
including an additional uncertainty factor, based on
age of exposure, in assessing the neurotoxic risks
posed by this class of pesticide.
4. Finally, the project on historical data is
designed to assemble existing pharmaceutical data on
the neurobehavioral effects of centrally acting chemi-
cals in multiple species, including humans wherever
possible. This project will provide direct evidence of
the adequacy of traditional uncertainty factors to
compensate for variations in susceptibility both within
and between species.
In summary, the proposed projects represent a
blend of efforts to empirically determine the relation-
ship of neurotoxic effects seen in different species,
including humans, and to both improve upon existing
methods and identify new and promising methods for
risk assessment. The results of these projects will
therefore support and complement current efforts
within the Program Offices to assess the risks associ-
ated with neurotoxic chemicals.
-------
R1HRA
Appendix D - 3
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Homologous Models for Evaluating Sensory Function
This project is designed to establish the relationship between measures of human sensory
function and measures taken from laboratory animals. This problem may be addressed
by parallel studies in laboratory animals and humans that directly compare measured
endpoints. Sensory-evoked potentials are electrophysiological recordings taken from the
sensory areas of the brain following a sensory stimulus that are thought to be good
measures of sensory function. They can be recorded from any species, including humans
and rats. The problem to be investigated is how closely the rat and human potentials are
related, and to what extent changes in one species predict those in the other. This
approach involves parallel studies in both species examining: 1) the response to
manipulation of stimulus parameters, 2) the response to certain drugs, and 3) the
relationship to other measures of sensory function. If sensory-evoked potentials recorded
from man and rats show similar changes in response to stimulus and drag treatments, then
there will be better grounds for concluding that the responses measured from the two
species reflect similar functional mechanisms, and that changes reported in the sensory-
evoked potentials of animal species are predictive of changes in human sensory function.
1. Internal progress reports
2. Final report describing the nature of the project, the results, and recommendations for
using these types of sensory assessments in determining the risks for humans exposed
to neurotoxic chemicals. These reports will be supplemented by peer-reviewed
publications in the open literature.
A lack of homologous models for neurotoxic endpoints has been identified as one of the
major sources of uncertainty in the risk assessment process (RIHRA). The Program
Offices of EPA, including OPTS, must decide the likely human consequences of exposure
to compounds based on animal data, even though the extent to which the measured
endpoints in the animal species coincide with adverse consequences in humans is often
unknown. This problem may be addressed by parallel studies in laboratory animals and
humans that directly compare measured endpoints. If measures of sensory function
recorded from man and rats shown similar changes in response to a variety of treatments,
then EPA will have better grounds for concluding that the responses measured from the
two species reflect similar functional mechanisms. Successful completion of this project
should improve the ability to perform cross-species extrapolation of sensory neurotoxic
effects. This information will help Program Offices, including OPTS, determine the
extent to which changes reported in animal species are predictive of changes in sensory
function of humans.
3 yr FY89: $122K
K. Hudnell (OHR/HERL-RTP) (919) 541-7538
-------
RIHRA
Appendix D - 4
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Homologous Models for Assessing Learning and Memory
This project is designed to improve animal-to-human extrapolation in the study of
chemical-induced learning and memory disorders, in order to reduce uncertainty in
evaluating potential risk posed by chemicals for producing these disorders in adult and
developing organisms. To accomplish this objective, we will employ two basic learning
and memory paradigms that have recently been used in comparative studies of cognition
involving animals and humans. Research with human infants asks whether a range of
memory phenomena that can be demonstrated with these tests emerge during ontogeny
in a manner which would be predicted from maturation of their neural substrates and/or
maturational profiles of behavior that occur in non-human animals. Additional work with
animals will attempt to alter memory development by disrupting neural maturation with
lesions, CNS teratogens, neurotoxicants, and environmental chemicals with known or
suspected neurotoxicity. In human infants, we are trying to identify special "high risk"
subject populations, neurological cases, or victims of accidental chemical exposure for
study. In human adults, we will examine task performance in healthy volunteers
administered various neurally active drugs, in neurological populations, and in populations
subject to occupational chemical exposure. Parallel studies will be conducted in animals
that will model the chemical treatment or neurological insult.
The products will include internal progress reports and a final report describing the nature
of the project, the results and recommendations for using these types of cognitive
assessments in determining the risks for humans exposed to neurotoxic chemicals. These
reports will be supplemented by peer-reviewed publications in the open literature.
Cognitive effects are often reported in humans following exposure to neurotoxic
environmental pollutants. To predict human cognitive impairments, risk assessors must
usually evaluate data obtained from laboratory animals. The tests applied to laboratory
animals often appear similar to tests which can be applied to humans, and many
superficially parallel methods for evaluating these kinds of effects are currently in use.
However, it is not clear to what extent changes recorded in animals can be extrapolated
to humans, and risk assessors will therefore have difficulty using animal data. These
studies are designed to determine the extent to which certain forms of learning and
memory share similar properties in animals and humans, and the extent to which they may
share a common neural substrate. Understanding these relationships will reduce
uncertainty in inferring human neurotoxic risk based on animal data.
Syr FY89:$85K
M.E. Stanton (OHR/HERL-RTP) (919) 541-7783
-------
RIHRA
Appendix D - 5
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Learning and Memory in Human Neonates
This project is designed to improve animal-to-human extrapolation in the study of
chemical-induced learning and memory disorders, in order to reduce uncertainty in
evaluating potential risk posed by chemicals for producing these disorders in adult and
developing organisms. Specifically, this project represents the human infant component
of OHR Task #1375. In order to accomplish this objective, we will employ two basic
learning and memory paradigms that have been used in comparative studies of memory
involving animals and humans. Research with human infants asks whether a range of
memory phenomena that can be demonstrated with these tests emerge during ontogeny
in a manner that would be predicted from maturation of their neural substrates and/or
maturational profiles of behavior that occur in non-human animals. In addition to
understanding the basic behavioral processes, this project will also improve our
understanding of the degree to which cognitive deficits in humans can be modeled using
laboratory animals. We will try to identify special "high risk" human infant populations,
neurological cases, or victims of accidental chemical exposure for study.
The products will include progress reports and a final report describing the nature of the
project, the results and recommendations for using these types of cognitive assessments
in determining the risks for human infants exposed to neurotoxic chemicals. These reports
will be supplemented by peer-reviewed publications in the open literature.
Cognitive effects are often reported in humans following exposure to neurotoxic
environmental pollutants. To predict human cognitive impairments, risk assessors must
usually evaluate data obtained from laboratory animals. The tests applied to laboratory
animals often appear similar to tests which can be applied to humans, and many
superficially parallel methods for evaluating these kinds of effects are currently in use.
However, it is not clear to what extent changes recorded in animals can be extrapolated
to humans, and risk assessors will therefore have difficulty using animal data. This
project is part of a larger series of studies designed to determine the extent to which
certain forms of learning and memory share similar properties in animals and humans, and
the extent to which they may share a common neural substrate. Understanding these
relationships will reduce uncertainty in inferring human neurotoxic risk based on animal
data.
2yr FY89:$110K
M.E. Stanton (OHR/HERL-RTP) (919) 541-7783
Topic:
Issue:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
-------
RIHRA
Appendix D - 6
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Ongoing
Cross-Species Validation of a Risk Assessment Method for the Characterization and
Prediction of Peripheral Neuropathy
This project is designed to determine the relationship between a neurotoxic syndrome seen
in two different species of test animal. Specifically, this project will compare the
neurochemical and neuropathological effects of organophosphate esters in hens and rats
to determine the extent to which inhibition of neurotoxic esterase in both species is
predictive of later pathology. This research is necessary to validate the use of rats as an
alternative test species to screen for neuropathic organophosphate esters. Rats and hens
will be exposed acutely to various organophosphorus-containing compounds as well as
additional selected toxicants. Levels of inhibition of neurotoxic esterase (NTE) will be
determined in the brain and spinal cord of each species shortly after dosing (24-48 hr).
Additional animals will be sacrificed approximately two weeks later for neuropathological
assessment of the peripheral nervous system and spinal cord. The extent of pathology will
be correlated with the degree of NTE inhibition and compared across species.
Compounds selected for this validation effort include a matrix of those that either do or
do not produce inhibition of NTE and those that either do or do not produce later
neuropathology.
The products will include internal progress reports and a final report describing the nature
of the project, the results and recommendations for substituting rats for hens as the
required species for identifying neuropathic organophosphorus-containing compounds.
These reports will be supplemented by peer-reviewed publications in the open literature.
Many industrial and pesticidal formulations contain organophosphates, and many
organophosphorus compounds produce delayed neuropathy. EPA routinely requires the
hen test for evaluating the potential of a compound to produce this type of neuropathy.
However, the hen is not used for the other forms of testing required by the Agency.
Historically, the rat has been considered refractory to this type of neuropathy. However,
recent evidence from NTD suggests that rats are indeed susceptible. This research will
be directed to determine the extent to which organophosphate-induced delayed neuropathy
is similar across species. More specifically, the research will determine the extent to
which rat data can be used as a more efficient and biologically based substitute to predict
human risk.
3 yr FY89: $225K
S. Padilla (OHR/HERL-RTP) (919) 541-3956
Topic:
Issue:
Status:
Title:
Biologically Based Dose-Response Models
4.1.3 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Intraspecies Differences in Susceptibility to Chemical-Induced Neurotoxicity
-------
RIHRA
Appendix D - 7
Description:
Identified Results: .
Usefulness of Results:
Project Length
and Cost:
Project Officer:
This project will help determine whether it is necessary to specify a particular type of rat
in assessing the neurotoxicity of chemicals. To accomplish this objective, neurobehavioral
function will be compared in F344 and in Long-Evans rats. F344 is the strain that is
widely used in routine toxicity testing, whereas Long-Evans rats are one of the most
widely used stocks in neurotoxicology research. The types of neurobehavioral function
assessed will include temperature regulation and metabolism, learning, and performance.
Concurrent measurements of neurotypic and gliotypic proteins will provide biochemical
markers of nervous system integrity. Parametric studies will assess whether the
neurobehavioral functions, as well as distribution and levels of neurotypic and gliotypic
proteins, are similar in both strains of rat. Chemical exposures using prototype
neurotoxicants will next be arranged and the responses of the two species will be
compared. Analytical chemistry will be employed to determine whether brain levels of
neurotoxicants producing effects are comparable in the two species.
This project will result in internal progress reports and a final report describing the test
conditions, results, and recommendations for requiring a particular species of rat for
neurotoxicity testing purposes. These reports will be supplemented by peer-reviewed
journal articles in the open literature.
Although albino strains such as the F344 are widely used in routine toxicity assessments,
many albinos have neurological abnormalities suggesting they may be inappropriate for
neurotoxicity testing. By determining whether qualitative and/or quantitative differences
exist in neurobehavioral function between F344s and Long-Evans rats, and whether the
two species are affected similarly by neurotoxicant exposures, this project will give risk
assessors an empirical basis for stipulating the particular type of rat to be used in
neurotoxicity testing.
3 yr FY89: $8 IK
R. MacPhail (OHR/HERL-RTP) (919) 541-7833
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
Multiple Issues
New
Quantitative Risk Assessment in Neurotoxicology
This project is designed to determine the relative value of traditional and novel approaches
to quantitative risk assessment based on neurobehavioral data. Laboratory animals will
be exposed to neurotoxic environmental pollutants, and dose-response functions will be
established using behavioral measures of neurotoxicity. The experimental design will vary
the number of animals used per exposure group and the amount of variability in
behavioral performance obtained within and between animals. The resulting dose-
response curves will be analyzed mathematically using curve-fitting techniques in order
to calculate ADIs. The resulting ADIs will be compared to those determined using the
traditional approach of adjusting NOELs using uncertainty factors.
-------
RIHRA
Appendix D - 8
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
This project will result in internal progress reports and a final report describing the test
conditions, models, results, conclusions, and recommendations for using these two
approaches for quantitative risk assessment using neurobehavioral endpoints. The
conclusions and recommendations will be supported by peer-reviewed journal articles in
the open literature.
The traditional approach to risk assessment involves determining NOELs, then adjusting
exposure levels downward with uncertainty factors in order to arrive at an ADI. For
example, the NOEL is routinely divided by 10 in order to compensate for individual
differences in sensitivity. An alternative approach to risk assessment has recently been
introduced (e.g., Crump, 1984; Dews, 1986) in which the ADI is calculated by fitting
mathematical models to dose-response data, and then estimating the dose corresponding
to some specified small amount of additional risk. This project will directly compare the
utility of these two different approaches for making risk assessments in neurotoxicology.
Program Offices (e.g., OPTS) are beginning to require submission of data on the
neurotoxic potential of chemicals. The products of this project will therefore greatly
improve the Agency's ability to make risk assessments in order to efficiently and
effectively regulate levels of exposure to neurotoxicants.
3 yr FY89: $120K
R. MacPhail (OHR/HERL-RTP) (919) 541-7833
Topic:
Issue:
Status:
Title:
Project Status:
Description:
Biologically Based Dose-Response Models
4.1.3 Inter/Intraspecies Extrapolation: Interspecies Sensitivity
Ongoing
Age as an Uncertainty Factor in Assessing Sensitivity and Severity of Effects Produced
by Cholinesterase-Inhibiting Compounds
New
The objectives of this project are twofold. First, the project will determine the extent to
which the severity of effects produced by cholinesterase-inhibiting compounds is
dependent upon the age of exposure. Second, the project will determine whether there
are any principles (e.g., metabolism) governing the extent to which different age subjects
are variously sensitive to the cholinesterase-inhibiting properties of cholinesterase
inhibitors. Data will be collected in rats. To address the severity question, different
groups of rats will be exposed to a variety of cholinesterase-inhibiting compounds during
different developmental stages, from prenatal to late adulthood. The relationship between
level of cholinesterase inhibition and neurobehavioral and neurochemical effects will be
.compared across age groups. To address the sensitivity question, the relationship between
exposure and cholinesterase inhibition will be explored during different developmental
stages. Neurobehavioral endpoints to be measured will be determined between NTD staff
and the awardee.
-------
RIHRA
Appendix D - 9
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Two products are envisioned, both reports. One report would provide an assessment of
the extent to which age affects sensitivity to cholinesterase-inhibiting compounds. The
second report would provide an assessment of the extent to which age affects the severity
of effect produced by equivalent levels of cholinesterase inhibition. These reports will be
supplemented by publication of the results in peer-reviewed journals.
Several Program Offices (e.g., ODW, OPTS, OSWER) perform risk assessments on
substances whose principal effect is inhibition of the cholinesterase enzymes. For several
years, the Risk Assessment Forum has had an active Technical Panel whose mission was
to develop a uniform Agency policy for applying uncertainty factors when performing risk
assessments on cholinesterase-inhibiting compounds. At a recent colloquium convened
by the Risk Assessment Forum to discuss the Technical Panel's report, age-dependent
severity and age-dependent sensitivity were highlighted as issues that have not been
addressed adequately. Insufficient data are currently available to determine whether any
principles can be applied to cholinesterase-inhibiting compounds to address these areas
of uncertainty.
3 yr FY89: $160K
K. Jensen (OHR/HERL-RTP) (919) 541-1560
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Biologically Based Dose-Response Models
4.2.3 Exposure Scenarios: Interaction of Exposure Parameters on Outcome
Ongoing
Exposure Scenarios and Neurotoxic Outcome .
This project is designed to determine how variations in exposure may affect the
neurotoxicity of chemicals., To accomplish this objective, we will determine how level
and duration of exposure to neurotoxic compounds jointly affect nervous system integrity.
Laboratory animals (rats) will be exposed to neuropathic chemicals and nervous system
integrity will be evaluated using morphological, biochemical, and functional
(neurobehavioral) tests. Systematic variation of both the level and duration of
neurotoxicant exposure will be used to determine whether neurotoxic outcome varies with
exposure scenario. Analytical chemistry will also be included to determine whether
exposure-related changes in neurotoxicity are due to differences in target-site sensitivity
or a simple accumulation of the toxicant at the target site.
This project will result in internal progress reports and a final report describing the nature
of the project, results, and recommendations regarding the extent to which neurotoxic
effects under one set of exposure conditions can be extrapolated to other exposure
conditions. These reports will be supplemented by peer-reviewed journal articles
published in the open literature.
Risk assessment often involves predicting neurotoxic effects following long-term, low-
level exposure on the basis of data from short-term, high-level exposures. Uncertainties
in predicting neurotoxic hazards could be greatly reduced by empirically determining the
-------
RIHRA
Appendix D -10
Project Length
and Cost:
Project Officer:
relationship between level and duration of exposure, and whether the mechanism(s)
underlying neurotoxicity change as a function of exposure scenario. An additional
potential benefit of this project is the possibility of being able to require short-term, high-
level exposure testing in lieu of more costly longer term, low-level exposure testing.
3 yr FY89: $106K
K. Crofton (OHR/HERL-RTP) (919) 541-2672
D.2 DEVELOPMENTAL/REPRODUCTIVE
TOXICOLOGY
This section provides the background and
rationale for the areas of focus for the reproductive,
developmental, and germ cell mutagenesis research
under RIHRA. It represents a synthesis of HERL and
OHEA directions and reflects activities in the existing
FY89 OHR RIHRA and OHEA/RURA base programs
as well as newly funded projects.
Throughout the existence of the EPA, Program
administrators have been faced with making decisions
on the potential of environmental pollutants to ad-
versely affect the reproductive processes of humans.
The Agency developed animal testing guidelines for
reproductive and developmental toxicity in the late
1970s and, six years ago, published -them in final
form (Pesticides Assessment Guideline, Subdivision
F. Hazard Evaluation: Human and Domestic Animals,
EPA-540/9-82-025). Minor modifications to the
protocols were made later by the Office of Toxic
Substances. In addition, the Office of Toxic Sub-
stances has recently indicated a commitment to
require testing and regulation on the basis of heritable
genetic risk (Fed. Reg. 53L247:51, 847-851, 856,
1988).
The risk assessment process for developmental
toxicants was standardized in 1984 with publication in
the Federal Register of the "Proposed Guidelines for
Health Assessment of Suspect Developmental Toxi-
cants." These guidelines were issued in final form in
1986 and revised and published for public comment
in Spring, 1989. Final publication is planned for
1990. Proposed guidelines for Male Reproductive
Risk Assessment and for Female Reproductive Risk
Assessment were published in the Federal Register in
June, 1988. Publication of those guidelines in final
form is also anticipated in 1990. Guidelines for germ
cell mutagenicity risk assessment were published in
final form in Fall, 1986. The current status of repro-
ductive, developmental, and mutagenicity risk assess-
ment is described in these sets of guidelines.
These documents also delineate fundamental
research areas for which the directing of resources
would most likely improve our ability to use animal
testing data in risk assessment In particular, the
developmental and reproductive risk assessment
guidelines describe several areas of uncertainty
inherent in the risk assessment/risk characterization
process that could benefit from a well-designed
research program. An integrated program would
evaluate those processes necessary to maintaining the
reproductive capability of each parent, including the
production of normal offspring.
In germ cell mutagenesis research, investiga-
tions on the male risk have dominated: the proposed
program should encompass research on both sexes.
Furthermore, the RIHRA program will be coordinated
with the National Toxicology Program, which is
recognized as a major research program in this field.
The preparation of the developmental, reproduc-
tive, and mutagenicity guidelines has benefited from
a number of recent workshops, symposia, and position
papers (Agency-sponsored and otherwise) that have
identified critical research needs. Part of this effort
has included commissioning a document that provides
a detailed critique of research needs in 16 different
areas of developmental toxicology. Not only does
this document form a logical basis for expanding the
existing EPA research efforts in developmental
toxicology, but many of these research areas have
high priority in reproductive risk assessment as well.
In discussions between the developmental,
reproductive, and genetic lexicologists of OHR and
OHEA, several mutually agreeable research needs
were identified in accordance with the issues de-
scribed in Topic 4. The key areas concentrated on
homologous mechanisms to strengthen inter-species
-------
RIHRA
Appendix D -11
extrapolation (Issue 4.1.1) and the development of
biologically based dose response models for high-to-
low dose extrapolation. Elements of this latter
activity relate to Issues 4.2.1 and 4.2.2 and include
ascertaining the interrelationship and co-influence of
different events expressed over the dose-response
range (e.g., the relationship between maternal and
developmental toxicity); establishing the predictive
validity of low-dose, sensitive measures to higher
dose, functional outcomes; and identifying basic
mechanisms and determining the commonality of such
mechanisms across dose. These activities are viewed
as critical to deriving the most valid dose-response
function upon which to base BB-DR models.
The priority and level of inquiry into each of
these issues was a function of toxicologic, physiolog-
ic, and mechanistic knowledge available in the areas
of male and female reproduction and developmental
toxicity as well as the relationship to base program
activities. Thus, extensive activity in the delineation
of mechanisms and dose-response modeling is pro-
posed for developmental toxicity—an area that would
benefit from a more substantial database. On the
other hand, the reproductive toxicology efforts con-
centrate primarily on obtaining fundamental informa-
tion on species comparability and endpoint interrela-
tionships. BB-DR modeling is seen as a subsequent
undertaking in reproductive toxicology.
The assessment of female reproductive risk is
the least developed of these areas and poses the
greatest uncertainties. Thus, emphasis in this area is
directed toward defining the mechanistic homology
between conventional test species and humans for the
critical events of ovulation, fertilization, embryo-
uterine interactions prior to implantation, and the
implantation process (Issue 4.1.1). Work in this area
will improve our ability to predict basic reproductive
failure in the human female, including very early
pregnancy loss. Equally important, this work will
examine the dose-response sensitivity and relationship
of the various endpoints for these events (Issues 4.2.1
and 4.2.2). These latter topics have high priority in
the male reproductive and developmental areas as
well.
The research proposed in the male reproductive
area also addresses issues of species comparability but
has advanced to a more quantitative level, i.e., the
actual derivation of interspecies extrapolation factors.
Proposed work in rodents on agents known to alter
human sperm production (e.g., via endocrine mecha-
nisms) will provide dose-response data that can be
compared to human data obtained in extensive work
sponsored by OHEA under its RURA program. Such
information will be useful in calibrating the inter-
species uncertainty factor for this target and provide
important basic information for BB-DR modeling for
male reproductive toxicants.
Additional inter-species research is directed at
delineating the predictive relationship between normal
sperm number and fertility. In test species, sperm
measures are more sensitive than fertility indices and
are endpoints that can be more readily evaluated than
reproductive success in humans. However, the
relationship between sperm measures and male
fertility is unknown. The proposed projects address
this issue in rat, rabbit, and man and will provide
critical data for interpreting the biologically functional
significance associated with sperm alterations. Such
information will promote the use of these measures in
the regulatory process by reducing the uncertainty as
to health significance associated with such data.
Again, the clarification of endpoint interrelationships
will contribute to subsequent development of BB-DR
models.
The developmental toxicology research concen-
trates primarily in two areas: 1) endpoint interrela-
tionships (e.g., the relationship between maternal and
developmental toxicity) and 2) the generation of BB-
DR models including research to test basic mechanis-
tic hypotheses. Laboratory efforts related to the
former activity will complement and extend a current
OHEA/RURA project that is evaluating data from the
NTP continuous breeding and developmental toxicolo-
gy programs. This effort is examining the relation-
ship between parental toxicity and reproductive/devel-
opmental consequences. The modeling efforts and
associated laboratory activities will address several
issues including the incorporation of the threshold
concept. These efforts in BB-DR modeling are
particularly timely. The sole application of the RfD
approach for noncancer endpoints has been criticized
within and outside the Agency. However, recent
efforts to develop other statistical approaches have
only been partially successful since they also fail to
incorporate basic biological processes. The work
proposed in the area will provide a more biologically
defensible basis for estimating human developmental
risk.
As described in the "Guidelines for Mutagenici-
ty Risk Assessment," concern is widespread that
exposure of humans to environmental mutagens may
induce genetic lesions which can be expected to result
-------
RIHRA
Appendix D -12
in higher frequencies of spontaneous abortion, lower
birth weight, congenital anomalies, and genetic
disease among offspring. For example, studies over
the last two decades have well documented the
substantial contribution of chromosomal anomalies
(both structural and numerical) to adverse human
reproductive and developmental effects. Thus,
research focused on the delineation of mutagenic and
other genetic mechanisms relevant to adverse repro-
ductive outcomes is viewed as an important compo-
nent in reproductive and developmental toxicity
research.
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanisms Across Dose
New OHEA/RURA project to be co-funded with HERL
Development of Biologically Based Dose-Response Models for Developmental Toxicity
Past efforts have involved the application of statistical models to data from standard
developmental toxicity studies with some attempts to take biological considerations (e.g.,
litter effects) into account. However, little has been done to evaluate the data available
on the underlying biological processes and the toxic effects of particular chemicals and
to integrate this into an approach to biologically based dose-response modeling. This
project will make use of model developmental toxicants for which some mechanistic/site
of actions data are available; incorporate pertinent information on such processes as cell
replication, migration, repair, cell death, etc.; and integrate this information with available
pharmacokinetic data to derive mathematical models that best incorporate these factors.
This project will involve the solicitation of proposals with new or unique approaches to
dose-response modeling in this area. This effort will require the cooperation of basic
scientists and risk assessment experts, including experts in biologically based dose-
response modelling.
1. Identification of appropriate model compounds for use in this effort
2. Identification of the basic developmental biology principles that are pertinent to
this approach
3. Development of experimental models using the data from 1 and 2 above
4. Identification of data gaps that require additional laboratory efforts
The Agency's guidelines for developmental toxicity risk assessment have called for the
development of new approaches to the extrapolation of dose-response data, and many
scientists and regulators have criticized the use of the RfD for noncancer endpoints.
Although several efforts to develop statistical approaches to this problem have been
underway in OHEA and OHR, they only partially deal with the biological processes
underlying developmental toxicity. This project could pave the way toward identifying
the data available for use in biologically based models and delineating research gaps to
be investigated further. Development of appropriate biologically based dose-response
models would provide significantly greater confidence in the estimation of human risk.
3 yr FY89: $160K (50% to be co-funded by HERL)
-------
RIHRA
Appendix D -13
Project Officers:
C. Kimmel (RDTB/OHEA) (202) 382-7331
R. Kavlock (DTD/HERL) (919) 541-2321
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Biologically Based Dose-Response Models
4.2.2 Exposure Scenarios: Sensitivity of Endpoints as a Function of Dose
Ongoing OHEA/RURA project
Quantitative Dose-Response Models for Multiple Endpoints in Reproductive and
Developmental Toxicology
The estimation of risk in reproductive and developmental toxicology utilizes primarily a
qualitative approach, with the application of uncertainty factors to the NOAEL/LOAEL
being the only quantitation currently done. This approach has been criticized by scientists
both inside and outside the Agency, who have suggested that a more quantitative approach
to estimating human risk is needed.
Several efforts have been undertaken along this line over the past two years, some of
which will be continued during the next two years. In particular, a model proposed by
Rai and Van Ryzin was applied to data from NTP developmental toxicity studies. This
model allows the evaluation of the effect of litter on the final outcome. Refinements in
the model were evaluated, and work on the incorporation of the intracluster correlation
into the Rai and Van Ryzin model was completed.
More recently, efforts have focused on the relationship of dose-response curves for
multiple endpoints in reproductive toxicology. The initial efforts have been to identify
the specific parameters to be evaluated and to identify databases available with appropriate
data for use in this effort. An IAG with the NTP has been established in order to make
use of their databases. The application of the Rai and Van Ryzin model to individual
development endpoints has been explored using the NTP database. In addition, through
a cooperative agreement with SIMS (Dr. Louise Ryan, Harvard), some analytical
approaches for comparing endpoints of developmental toxicity have been explored.
This project will be a continuation of the past efforts in this area, and will focus
particularly on the evaluation of endpoint relationships and the development of models for
comparing dose-response curves and for low-dose extrapolation.
1. Identification of the endpoints available for comparison in reproductive and
developmental toxicity databases
2. Development of methods for comparison of dose-response curves for multiple
endpoints of reproductive and developmental toxicity
3. Development of approaches for combining data from continuous and binomial
endpoints of reproductive and developmental toxicity
Quantitative models are needed that appropriately relate dose to response, that relate
responses to each other, and that are capable of estimating the risk at any dose level,
particularly at the potentially low levels of human exposure. This project will evaluate
-------
RIHRA
Appendix D -14
Project Length
and Cost:
Project Officer:
the relationship of multiple endpoints of reproductive and developmental toxicity to each
other and evaluate the possible influence of some endpoints (e.g., maternal toxicity) on
others (e.g., developmental effects). This information will provide an improved
understanding of the data from these types of studies which will be invaluable in the risk
assessment (both qualitative and quantitative) for reproductive and developmental
toxicants.
3 yr FY89: $40K
G. Kimmel (RDTB/OHEA) (202) 382-5978
Topic: .
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanisms Across Dose
New RIHRA project
Mutagen-Induced Developmental Toxicity Following Exposure of Mouse Zygotes
Recently, certain mutagens have been shown to induce high incidences of fetal
malformation and death when the zygote stage is exposed. This finding raises questions
regarding the danger to women who may be exposed to chemicals in the time period
immediately following conception. Most lexicologists and clinicians consider the zygote
as susceptible to chemically induced toxic effects but relatively impervious to
developmental effects. Extensive work has already ruled out maternal toxicity, gene
mutations, and chromosomal aberrations as underlying causes of the fetal malformations
and death. Nevertheless, a genetic mechanism is still strongly suggested by stage
specificity of the effect, as well as by the results of a reciprocal egg-transfer study. It is
proposed that this mechanism involves a gene dosage or genetic transposition mechanism.
The implications of this observation, therefore, are extremely important in the assessment
of risk for developmental toxicity by showing that the window of susceptibility is not
limited to organogenesis and paving the way for mechanistically tying a dose-response
relationship with a dose-effect. This project will generate additional information defining
the dose-response curve of these mutagen-derived zygote effects and the genetic
mechanism(s) involved. Such information is essential for developing risk extrapolation
procedures for these mutagen-induced fetal effects.
1. Data on several different concentrations of a mutagen(s) to allow for the extrapolation
of dose-response trends
2. Mechanistic information describing the underlying causes of those trends, which will
be useful in developing a model that is biologically reflective
In developmental toxicity risk assessment, a NOAEL/RfD approach to assessing risk is
traditionally used. Although a threshold mechanism is usually assumed, the underlying
causes of developmental toxicity are largely unclear. Thus, to advance developmental risk
assessment in the area of biologically based dose-response models, research is needed to
-------
RIHRA
Appendix D -15
Project Length
and Cost:
Project Officer:
mechanistically tie dose-response relationships with dose effects. The above proposal is
an attempt to accomplish this for mutagen-induced developmental toxicity.
This work is viewed as having a great impact on the approaches that are used to assess
developmental toxicity caused by mutagens by providing a framework for generating and
testing hypotheses about mechanism. In addition, it will result in an improvement of
current risk assessment procedures by extension of the observability of the response to
much lower doses.
2 yr FY89: $80K
V. Dellarco (GTAB/OHEA) (202) 382-7332
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.2/4.2.2 Endpoint Sensitivities and Interspecies Sensitivities
New RIHRA project
Quantitative Extrapolation for Human Male Spermatotoxicity: Relationship Between
Production of Normal Sperm and Fertility.
The overall goal is to develop the capability to predict potential effects on human fertility
from semen evaluations in humans, from effects on fertility in a test species, or from
effects on sperm parameters in a test species.
To predict potential effects on human male fertility, it is necessary to know the
quantitative relationships between number of normal sperm produced and fertility in both
humans and test species. Additionally, information on the relationships between different
measures of sperm production in test species is needed. This project, along with a parallel
project with rabbits, will produce those data. Using those data, quantitative extrapolation
methods will be developed to replace uncertainty factors in assessment of risk for male
spermatotoxicants. Once those methods are established, data from humans or test animals
exposed to specific agents or mixtures can be utilized, along with adequate exposure
information, to assess potential effects of exposures on fertility of men. This type of
approach has been identified as a need in the preparation of the U.S. EPA Guidelines for
Assessing Male Reproductive Risk. The approach should be useful in predicting the
extent of effects caused by exposure of human males to spermatotoxic agent(s) in or near
waste disposal sites, in the workplace or in the environment. Thus, the results of this
projects should be of substantial value to the Agency.
In this project, relationships between number of normal sperm produced and time required
to achieve conception will be quantified. Human couples desiring to have a child will be
recruited. Multiple semen samples will be obtained from each male prior to intercourse
at regular intervals around the predicted day of ovulation. Abstinence intervals between
ejaculations and timing relative to predicted day of ovulation will be controlled. Multiple
endpoints will be examined individually and in combination to assess sperm quality.
Urinary estrone and pregnanediol-glucuronides, plus LH and hCG will be measured.
Pregnancy will be determined by hCG, missed menses, and by ultrasonic examination.
-------
RIHRA
Appendix D -16
Identified Results:
Usefulness or Results:
In the absence of pregnancy, multiple cycles will be studied. Using the data from this
study and the parallel project with rabbits, the appropriate conversion factors and
equations will be developed to describe the relationships that can then be used to predict
effects on fertility.
Specific Aims:
1. Utilize critical analyses of sperm number and quality in men derived from couples
attempting conception which will allow determination of the quantitative relationships
between number of "normal" sperm and fertility (time required for conception) in
humans.
2. Use the data from 1) to predict the effects of reductions in number of "normal" sperm
on human fertility (inter-endpoint extrapolation).
3. Combined with the data from the parallel experiments with rabbits, develop
extrapolation factors to predict effects on human fertility from effects on "normal"
sperm production in rabbits (interspecies extrapolation).
1. Characterization of the relationships between number of normal spermatozoa produced
by human males and measures of fertility
2. Calculation of quantitative extrapolation factors for predicting effects on human
fertility from measurement of number of normal sperm produced
3. Combined with results from a parallel project with rabbits, calculation of interspecies
extrapolation factors for predicting effects on human fertility from effects on number
of normal sperm produced in a test species (rabbit)
Currently, when risk assessment is conducted involving toxic effect on the male
reproductive system, uncertainly factors are applied to a No Observed Adverse Effect
Level (NOAEL) to conservatively estimate an exposure level that should be safe for
humans. Often, only data from test species are available, and uncertainty factors must be
used for interspecies adjustment as well as for protection of more sensitive individuals.
Experimentally derived extrapolation factors are preferable to uncertainty factors because
they can more accurately reflect the differences found.
This project is designed to contribute to achieving that goal. To do this, the relationships
between production of normal sperm and fertility in both humans and test species must
first be characterized more rigorously than has been done previously. This project will
determine the quantitative relationships between endpoints that reflect sperm number
and/or sperm quality and fertility in humans. Efforts elsewhere (independent of this
project), will provide similar information with a test species (rabbit). With the
interspecies and inter-endpoint relationships available, risk assessments can convert
measurements of sperm production to a prediction of fertility status, either when fertility
data are not available or when the fertility results are judged to be too insensitive based
on other available information. This would have particular application in these situations:
• With suspected human exposures, semen samples may be obtained from the men, but
it is unlikely that sufficient information on fertility will be available. If the
quantitative relationship between number of normal sperm produced and fertility were
known for humans (supplied by data derived from this project), the risk to fertility for
men could then be estimated.
-------
RIHRA
Appendix D -17
Project Length
and Cost:
Project Officer:
• Knowledge of the relationships between number of normal sperm and fertility for both
humans and test species can provide the capability to extrapolate between the test
species and humans.
• With quantitative extrapolation factors available to convert from results on other
endpoints reflecting effects on sperm production, it would be possible to apply inter-
endpoint and interspecies extrapolations to predict effects on humans from data on
sperm production in a test species. The extrapolation capabilities produced by this
project and the parallel project with rabbits can be incorporated into risk assessments
that utilize spermatotoxic effects as the critical effect or used to evaluate effects
observed in biomonitoring. The approach should be useful for individual chemicals
or for mixtures.
In addition to the capabilities described above, this project has important ramifications for
in clinical medicine. Those benefits will be obtained from the project as currently
designed and will not require any additional expense to EPA. Those ancillary benefits
include:
1. Data on the distributions of semen characteristics for a normal population of human
males plus fertility information for those same males
2. Data, based on prospective testing, for objectively determining human male fertility
parameters
3. Improved ability to diagnose male factor infertility based on laboratory evaluation of
semen
4. Improved ability to evaluate the efficacy of certain male contraceptives by evaluation
of semen
5. Improved understanding of the timing of events in early pregnancy from use of
biochemical and ultrasonic monitoring methods, and ascertain any association with
objective laboratory-determined parameters of human sperm
6. Association of the presence of antisperm antibodies with the numbers and normality
of sperm found in periovulatory cervical mucus specimens; an indication of the
relationship between those properties and fertility may also be obtained
3 yr FY89: $100K
E. Clegg (RDTB/OHEA) (202) 475-8914
Topic:
Issue:
Status:
Title:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Homologous Models for Very Early Pregnancy (VEP) Loss in Humans
-------
RIHRA
Appendix D -18
Description:
Identified Results:
Usefulness of Results:
In mammals, successful pregnancy occurs only when the early developmental events
including fertilization and cleavage result in the formation of a normal embryo that is
transported to the uterus at the appropriate time and implants normally in a receptive
uterus. These early events depend upon the production and time-appropriate release of
healthy gametes (sperm and oocytes) and the maintenance of a physiologically sound
tubal/uterine environment. Perturbation of any of these events may result in pregnancy
termination, even before the pregnancy is detectable.
Recent progress has been made in our ability to detect early pregnancy in humans, most
notably by assaying for hCG, the human pregnancy marker, using extremely sensitive
immunological tests (Wilcox et al., 1988). Pregnancy surveillance studies using this assay
and others, have shown that as many as two-thirds of potential pregnancies fail between
the time of ovulation and implantation. The high incidence of VEP loss attests to the
vulnerability of early developmental events and the potential for environmental chemicals
to act during this period. Indeed, this subject received a great deal of emphasis at a recent
NIOSH/NIEHS sponsored symposium on "Assessing Reproductive Hazards in the
Workplace" (Cincinnati, OH, June, 1988). Unfortunately, the causes of VEP loss in
humans are largely unknown. This is because it is difficult, if not impossible, to access
fertilization and early development in women.
In the area of reproductive risk assessment, little or no attention has been given to the
characterization of VEP loss in animal models as a means of detecting toxicant-induced
VEP loss. This is surprising since both in vivo and in vitro methods for doing so have
been available for years. This project will focus on the validation of rodent animal
models to evaluate very early developmental events. This approach is justifiable on the
grounds that the basic biological/molecular mechanisms underlying early embryonic
development and implantation are similar in humans and rodents. Included in the
validation process will be the definition of critical periods wherein toxicants may perturb
these early developmental events, and determination of mechanisms whereby xenobiotics
may cause VEP loss. Obviously, an animal model is needed to produce dose-response
data on compounds known or suspected of causing very early pregnancy loss. Initially,
compounds with a wide range of potential actions will be examined. These will include
compounds that directly perturb cellular processes (e.g. cell division) as well as those that
may alter the hormonal milieu during the peri-implantation period.
This approach will also permit discrimination between VEP due to embryo vs. maternal
effects. Certainly, the database generated by this approach is critical for the prediction
and hopefully characterization of human risks associated with reports of infertility or
repeated VEP loss in toxicant-exposed women.
1. Recommended methods and protocols for assessing toxicant-induced very early
pregnancy loss in rodents.
2. Report on critical periods of exposure associated with specific VEP effects.
3. Report on mechanisms underlying VEP loss, common to humans and rodents, at the
level of fertilization, preimplantation embryonic development and transport, ovarian
function and hormonal support of implantation, and uterine receptivity.
Effective female reproductive risk assessment has been encumbered on two fronts: 1)
lack of information on the incidence and causes of VEP loss in humans, and 2) lack of
data on homologous mechanisms of VEP loss in animals and humans. The products of
this project will address both needs. By improving our ability to predict VEP loss in
humans, based on the dose-response data obtained from animals, uncertainties associated
-------
RIHRA
Appendix D -19
Project Length
and Cost:
Project Officer:
with assigning specific "margins of safety" should be reduced. Furthermore, this research
will identify an array of endpoints with improved sensitivity and mechanistic value that
may be added to the relatively insensitive fertility measures cunuitly used in hazard
identification. Such endpoints may be considered for inclusion into the EPA Female
Reproductive Risk Assessment Guidelines.
3 yr FY89: $106K
A. Cummings (OHR/HERL-RTP) (919) 629-5194
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Reproductive Toxicants and the Hormonal Control of Ovulation
In the mammalian female, appropriate endocrine signals are essential for the production
of gametes by the ovary. As spontaneous ovulators, both the human and rat exhibit
regular ovarian cycles. In both species, the key endocrine event controlling ovulation is
the appearance of the midcycle surge of luteinizing hormone (LH) from the pituitary into
the general circulation. This clearly defined surge stimulates the meiotic and cytoplasmic
maturation of the oocyte and triggers those follicular events that lead to die release of
oocytes.
Various pharmacological manipulations (e.g. administration of the alpha-adrenergic
blocking agent, phenozybenzamine) or exposure to substances of abuse (e.g. ethanol,
tetrahydrocannabinol or morphine) can disrupt the surge and, as a consequence, block
ovulation in several species including humans. Moreover, there is evidence (reviewed by
Mattison, 1985) of menstrual alterations after occupational exposure to a variety of
toxicants that may be linked to disruptions in pituitary regulation of ovarian function.
Therefore, the proposed research will focus on the influence of environmental compounds
on the hypothalamic-pituitary control of pvulation, an area in which there is a substantial
data gap.
In the area of reproductive risk assessment, virtually no information is currently available
on the effect of toxicants on the neuroendocrine control of ovulation. Since the key event
in this process is the mid-cycle surge of luteinizing hormone (LH) from the pituitary
gland, it is surprising that there is such a paucity of information because the surge is
readily amenable to study. This is also disturbing, since an understanding of these effects
is critical for the characterization of human risks associated with reports of menstrual
irregularities or infertility in toxicant-exposed women. The proposed studies will evaluate
the effect of selected environmental compounds on the ovulatory surge of LH in the rat.
Since there is a large degree of homology between the rodent and human in the
mechanisms controlling this event, and in the critical nature of the LH itself, the results
of these studies will provide a valuable and important database for immediate use in risk
assessment, as well as insight concerning potential effects of specified compounds on
reproductive competence in the human female. The data obtained will provide definitive
-------
R1HRA
Appendix D - 20
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
information for use in reducing the current uncertainties associated with female
reproductive risk assessment
The proposed studies will employ an acute, in vivo administration of test compounds in
different dose levels at specific times during the ovarian cycle and the identification of
changes in LH secretion. Possible associated changes in other reproductive hormones will
also be monitored. Initial selection of test agents will focus on those known to have
adverse reproductive effects. This approach will provide a means to determine the way
in which a compound affects the process of ovulation and the presence of any critical
periods that may exist over the ovarian cycle during which an effect of exposure is more
pronounced.
These studies will permit the identification of the immediate target tissue within the
reproductive system (i.e., brain, pituitary or ovary). Initial experiments will investigate
the effect of in vivo treatment on the LH surge. Once a compound is identified as having
an adverse effect, subsequent in vitro experiments will examine potential impairments in
hormonal release from hypothalamic and pituitary explants in order to determine the
appropriate mechanism(s) of action. Furthermore, this research could lead to further
studies assessing the impact of delays in ovulation on gamete viability, altered
implantations, and possible abnormalities in fetal development that may result from
changes in the critical timing of oocyte release.
1. Recommended methods for the assessments of toxicant-induced disruptions in the
ovulatory surge of luteinizing hormone
2. Report on the important temporal parameters associated with a compound's effect on
the ovulatory surge of LH (and oocyte release) and the establishment of a critical
database for such an effect
3. Report on the mechanisms underlying any observed alterations in the LH surge, using
data collected from studies of hormonal release in vitro from pituitary and
hypothalamic tissue
Because the outlined approach will establish a much needed database for assessing the
effect of environmental toxicants on gamete release, it is an important component of the
study of female reproductive toxicology. Moreover, since these studies address the
process of ovulation, as opposed to implantation and pregnancy maintenance, the data will
permit an evaluation of a compound's effect on those events that occur prior to, and are
necessary for, conception. Such information will fill a critical data gap that currently
exists and reduce the attendant uncertainties associated with female reproductive risk
assessment
Mattison, D.R. Clinical manifestations of ovarian toxicity. In: R.L. Dixon (ed.),
Reproductive Toxicology. Raven Pr., N.Y. (1985), pp. 109-130.
3 yr FY89: $131K
R. Cooper (OHR/HERL-RTP) (919) 541-4084
-------
RIHRA
Appendix P - 21
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
Neuronal and Hormonal Factors in Normal and Abnormal Kidney Development
Deficits in the functional development of organ systems following in utero or early
postnatal exposure to xenobiotic agents is one of the four manifestations of developmental
toxicity. However, as noted in the Developmental Toxicity Risk Assessment Guidelines,
relatively little attention has been paid to these adverse health effects. While some
progress has been made in evaluating adverse health effects on the development of the
central nervous system, a need exists to develop methodologies and evaluate the potential
magnitude of effects for other systems. This project is an effort to expand our
understanding of the pathogenesis of altered development of the kidneys. The expertise
in physiological assessment of renal development available at HERL is being united with
the expertise in biochemical and pharmacological development of the kidneys available
at the Department of Pharmacology at Duke University Medical Center to provide an in-
depth analysis of the issue.
This project will evaluate the hypothesis that disturbances in the development of either
the endocrine or autonomic nervous system can cause altered morphological, biochemical,
and physiological maturation of peripheral organ systems. That is, the very systems
geared to maintaining homeostasis in the adult organism may also provide critical trophic
influences on the development of organ systems in the perinatal animal. If true, the
findings would demonstrate that the period of differentiation and activation of the
autonomic these developmental events lie primarily in the third trimester. However, in
the species used to evaluate potential developmental toxicity, they reside in the early
postnatal period of development Therefore, our present testing guidelines (which are
focused on prenatal exposures and evaluations) would need to be and endocrine systems
represent key critical periods in ontogeny. In the human, modified to accommodate them.
In addition, since the effects on the organ systems would occur after the period of
organogenesis, no gross morphological indications of the perturbation would be apparent.
Thus, biochemical and physiological approaches are required to evaluate the hypothesis.
Because the events under study occur in the early postnatal phase of rodent development,
easy experimental manipulation of the developmental processes involved is possible. To
evaluate the hypothesis, three different sets of neuronal or hormonal perturbations will be
employed: 1) global sympathectomy (via systemic administration of 6-hydroxydopamine)
or central catecholaminergic lesioning via intracisternal administration) immediately after
birth; 2) hyper- and hypo-thyroidism and exogenous glucocorticoid therapies (treatments
known to influence the onset of sympathetic innervation); and 3) the effects of teratogens
with potential action on the kidney mediated through the hormonal or neuronal alterations.
For each of these models, four test batteries will be conducted that will establish neuronal
status (degree of impulse flow, reflex stimulations, etc.), tissue responsiveness (renal
cyclic ADH stimulation), and finally differentiation of renal tissue (nucleic acids, protein
synthesis, ODC, polyamines and histology). The application of the test batteries to the
experimental models will provide a comprehensive picture of how neuronal and hormonal
input contribute to the development of renal structure and function and show conclusively
whether teratogenic insult to the kidney may arise from primary insult to neuronal or
hormonal systems.
-------
RIHRA
Appendix D - 22
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
1. Compilation of reports detailing the effect of direct manipulation of the developing
autonomic nervous system on the biochemical and physiological development of the
kidney
2. Compilation of reports detailing the effect of manipulation of thyroid and adrenal
function on the biochemical and physiological development of the kidney
3. Compilation of reports detailing the involvement of the neuronal and endocrine
systems in chemically induced alterations in the functional development of the kidneys
As stated in the Developmental Toxicity Risk Assessment Guidelines, the need exists to
understand the potential of chemicals to affect the functional development of organ
systems following exposure during the perinatal period. The risk assessor will benefit
from this project in several ways: 1) by the project's focus on assessment of postnatal
organ function and how that function can be permanently perturbed by exposure to
xenobiotic agents at what may amount to be a new critical period; 2) the exposure being
evaluated (the early neonatal period) is a time period in which little lexicological data is
available; and 3) the mechanistic approach to determine how organ dysfunctions could
arise in the absence of gross morphological disturbances. These items will provide
important new information on how functional decrements in organ performance might be
induced in developmental toxicity assays and whether our present testing strategies are
adequate to detect these adverse effects.
3 yr FY89: $125K
R. Kavlock (OHR/HERL-RTP) (919) 541-2326
Topic:
Issue:
Status:
Title:
Project Status:
Description:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
New
Embryo Culture: A Unique in vitro Tool to Investigate Critical Issues in Interspecies
Extrapolation in Teratology
New
Risk assessors in the area of developmental toxicity must often confront the wide disparity
in reactivity of laboratory species to chemical exposure. In the absence of data to the
contrary, the customary solution to this problem is to use the most sensitive species for
establishing the NOEL. Uncertainties in the process would be reduced if the
pharmacokinetic and/or pharmacodynamic factors responsible for these species differences
were understood. Such knowledge would allow a more logical selection as to which
species provides the most meaningful indicator of potential human risk.
The in vitro culture of embryos in co-culture with primary hepatocytes offers one
approach to determining the relative contribution of metabolism and target tissue
sensitivity in teratogenesis. Mouse, rat and hamster embryos can now be routinely
-------
R1HRA
Appendix D - 23
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
explanted from the uterus at the 4-12 somite stage and grown in culture. During the 48-
hour period in which embryonic development is comparable to the in vivo situation, the
embryos are isolated from the influence of maternal metabolism, kinetics, and placenta
transport. Addition of the chemical to the culture medium allows direct determination of
the effects of the agent on the embryo. Co-culturing primary hepatocytes with the
embryo, in turn, facilitates understanding of the role of metabolism in teratogenesis
(embryos of this developmental stage lack most Phase I metabolizing pathways). The
hepatocytes need not be of the same strain or species as the embryo, and in practice,
matrices of hepatocytes/embryo combinations can be constructed which can be used to
contrast metabolic versus tissue sensitivity factors in the etiology of birth defects.
Hepatocytes offer a major improvement over the widely used S-9 preparations as a
metabolizing system, both in terms of duration and variety of metabolites produced. This
co-culture system was developed and validated at the HERL laboratory. While not yet
in use, human metabolic systems, derived from either placentas or liver biopsies, can be
incorporated into the system. Results from this approach would enable direct comparison
of interspecies metabolism in teratogenesis. Initial experiments will focus on chemicals
which have been demonstrated to induce varying results in rodent/rabbit developmental
toxicity studies. Likely candidates include endrin (a hamster but not mouse or rat
teratogen); meclizine (a mouse but not rat teratogen), ethylenelhiourea (a rat but not
hamster, guinea pig or mouse teratogen), phenytoin (a strain-specific mouse teratogen),
as well as several "unidentified" species-specific teratogens supplied by researchers in the
pharmaceutical industry interested in evaluating the system.
1. Reports describing quantitative dose-response information for heterologous
hepatocyte/embryo combinations on a chemical by chemical basis. These reports will
shed light on the relative contribution of metabolism versus embryonic sensitivity as
explanations for observed strain or species differences in teratogenesis.
2. Summary evaluation of the use of the hepatocyte/embryo culture system in assisting
selection of the most appropriate species for use in extrapolation to humans.
The presence of marked species differences in the embryonic response to some
environmental agents reduces the certainty of their respective risk assessments. For
example, resolution of the potential human teratogenicity of the pesticide carbaryl has not
occurred despite many years of debate within the Agency. For that chemical, the dog
stands out from nearly ten other laboratory species as being a sensitive species. Whether
the majority of negative species, or the reaction of the dog should be used to predict
human risk has been subject of a yet unresolved debate. Similar, but less dramatic
examples of species differences in teratogenic susceptibility are not rare. This project
offers one possible approach to examining the underlying mechanisms behind these at
times dramatic differences. The demonstration of either a unique species-specific
metabolic pathway or a species-specific "embryonic receptor" responsible for such
observations would be a major contribution in determining the most appropriate factors
to consider in assessing human risk. While this project has certain limitations due to the
limited number of species that can be cultured and the limited duration of the culture
period itself, it does offer hope for determining the mechanism responsible for species
differences in teratogenic susceptibility for at least some chemicals. As such, it is a
potentially powerful tool for improving the scientific basis of risk assessments in the field
of developmental toxicity.
3 yr FY89: $152K
J. Rogers (OHR/HERL-RTP) (919) 541-5177
-------
RIHRA
Appendix D - 24
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
New
Utilization of Testicular Perfusion to Characterize the Effects of Leydig Cell Toxicants
on Spermatogenesis in Several Species of Laboratory Animals
The testis of the male is divided into two morphologically distinct compartments that are
functionally dependent upon each other's hormonal secretions. Spermatogenesis (sperm
production) occurs in the tubules of the testes, the first compartment, and this process is
dependent upon testosterone that is secreted by the Leydig cells in the interstitial space
between the tubules. Alterations of Leydig cell functioning may result in azoospermia and
infertility. For example, Spermatogenesis declines in aging men as the Leydig cell
population falls below the threshold level necessary to maintain sperm production. The
hypogonadal male, where the pituitary does not secrete the hormones (FSH and LH)
essential for normal testicular function, is another example of the importance of Leydig
cell function. These men have low sperm counts, in part, because in the absence of LH
the Leydig cells do not secrete sufficient testosterone to maintain Spermatogenesis.
The recognition of the critical relationship between intra-testicular testosterone levels and
Spermatogenesis has prompted the REAG/EPA to fund a study examining the decline in
sperm counts in men following the administration of exogenous testosterone. High serum
testosterone levels provide negative feedback in pituitary LH secretion with the subsequent
decline in LH, Leydig cell function and Spermatogenesis. The RTB/HERL plans to
conduct a similar study using rats to that comparative dose-response relationships between
administered testosterone and sperm production will the available. The present study will
also examine the effects of Leydig cell toxicants (testosterone and EDS) on
steroidogenesis in various mammalian species using the perfused whole testes from
several species (rats, rabbits, and possible primates) following in vivo and in vitro
administration of the compounds. The rabbit has several biochemical and morphological
homologies with primates that the rat does not afford. Primate and human testes will be
utilized if available. This task will also compare the effects of EDS administration in
young, middle-aged and old male rats to determine if a safety factor of 10 for intraspecies
variation in susceptibility is sufficient
For these studies, whole testes will be perfused and the functional capacity of the Leydig
cells will be evaluated using a variety of biochemical and morphological measures after
both in vivo and in vitro EDS administration. In the in vivo studies Spermatogenesis will
be monitored by light microscopy, and daily sperm production, epididymal sperm reserves,
and sperm morphology will be measured. In concert, the testes will also be examined and
changes in Leydig cell numbers and stage-specific degeneration of the seminiferous
epithelium will be quantified. In vitro the specific site of toxicant action in the Leydig
cell will be determined using HPLC analysis of steroid substrate accumulation in the
biosynthetic pathway.
This task will define the quantitative relationship between Leydig cell function and
Spermatogenesis in different mammalian species. The results of this research, in
conjunction with other Agency studies (REAG study on the effects of testosterone on
human Spermatogenesis and a similar HERL study using rodents) will enable risk
assessors to extrapolate effects on Leydig cell function, seen in toxicology studies using
rats, to potential effects in man.
-------
RIHRA
Appendix D - 25
Usefulness of Results:
Project Length
and Cost:
Project Officer:
This information will be useful to the Program Offices that receive reproductive data
because it will enable them to determine if the effects on hormone synthesis, seen in
rodents studies, will produce adverse effects on human sperm production, resulting in
infertility in man. The database will be useful to REAG because it provides multiple
reproductive endpoints across doses in two mammalian species.
3 yr FY89: $100K
John Laskey (OHR/HERL-RTP) (919) 541-2782
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.3 Inter/Intraspecies Extrapolation: Intraspecies Sensitivities
Ongoing
Mechanisms of Litter Effects:
Teratology
Assessing the Reliability of Dose-Response Models in
A major problem in interspecies extrapolation of experimental teratology data to the
human species is that rodents have many offspring per pregnancy, and the response among
the fetuses in a litter can be quite variable. Since it is the pregnant animal, and not the
fetuses, that is treated with the test compound, the individual fetuses cannot be treated as
independent events statistically. Furthermore, it is often observed that the variation among
the litter responses at a given dose is greater than one would expect from the variation of
the responses within the litters. These issues become important in developing
mathematical dose-response models. Is intraliuer variability always less than interlitter
variability? What are the biological bases for litter effects? Are these effects similar
across compounds? It is possible to incorporate such so-called "litter effects" in statistical
models of teratology data, but one needs to have an idea of how such effects occur, and
how large one might expect them to be in our test species.
We will first study litter effects in the mouse using the known teratogen dinocap. The
major malformation produced is cleft palate, and we will analyze the within-litter
distribution of cleft palates following doses of dinocap. We will also record te position,
weight, and sex of each fetus in the uterus. These data should allow us to measure
correlations between fetuses in a litter, including within vs. between litter comparisons and
analysis of "neighborhood effects" (whether neighboring fetuses in a litter respond more
similarly to a toxicant than do non-neighboring fetuses). These results will provide
information on how to incorporate overall litter effects into dose response models.
A possible cause of litter and neighborhood effects is variation among dams in the dose
delivered to their uterus and to different pups within a uterus. In the second part of the
study we will quantify pharmacokinetic variation by measuring the delivered dose of
selected compounds to litters and to individual fetuses within litters. These measurements
should allow us to quantify the contribution of pharmacokinetic variation to litter effect,
both through variation in delivered dose and as a component of position and neighborhood
effects.
-------
RIHRA
Appendix D - 26
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
A major criticism of all current models is the lack of incorporation of biological
mechanisms responsible for intra- and interlitter variability. Any practical, biologically
based dose-response model will have to address two problems: accounting for among-
litter variability and predicting mean responses for any given dose. The results of these
studies should contribute to the development of such models, with the aim of reducing the
uncertainty involved in developmental toxicity risk assessment
There are two components of this project One is the actual animal and laboratory work
involved in gathering the needed biological data. This effort will be supported by one
NSI person. The second component is statistical design and mathematical modeling of
the data. This work will be completed by a National Research Council research associate.
1. Report on the prevalence of the "neighborhood effect", its magnitude, and its
contribution to the litter effect
2. Report on the extent to which pharmacokinetic variation can contribute to the observed
neighborhood and litter effects
3. Assessment of the value of, and methods for, incorporation of information from (1)
and (2) above into practical, biologically based, dose response models for
developmental toxicity data
As indicated above, modeling of rodent developmental toxicity data is complicated by the
fact that these species have multiple offspring per litter. Using a simple mean value to
represent the response of a litter results in the loss of information concerning intralitter
variation. This research will lay the groundwork for 1) assessing the impact of litter and
neighborhood effects on variability seen in developmental toxicity data, and 2) developing
methodologies for incorporating these variables into biologically based dose-response
models.
3 yr FY89: $106K
J. Rogers (OHR/HERL-RTP) (919) 541-5177
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.3 Inter/Intraspecies Extrapolation: Intraspecies Sensitivities
New
Intraspecies Variability in Litter Responses
Individual Variation in the Maternal Genome
to Developmental Toxicants: Role of
It has long been hypothesized that a significant component of intra-species variation in
the maternal/litter response to teratogenic agents is due to individual genetic variation.
This research project will evaluate the degree of individual variability resulting from
exposure to a variety of teratogens administered at different times to outbred strains in
two commonly used laboratory species (rat and mouse).
-------
RIHRA
Appendix D - 27
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Standard teratology bioassays require that the maternal animal be killed shortly before
term so that the fetuses can be removed for morphological examination. This procedure
does not allow for any evaluation of the relationship of individual maternal variability
(e.g., in pharmacological parameters) to developmental outcome. The proposed series of
studies will directly address this critical issue by utilizing a strategy involving the repeated
testing of a series of unrelated compounds known to induce developmental toxicity that
may be ascertained by postnatal examination. Individual animals will be exposed to the
same compound and dose during three successive pregnancies. Outbred female mice and
rats will be bred to identified inbred males and teratogens administered at established
effective doses. Data will be analyzed to determine the degree of individual constancy in
teratogenic response across breedings; the relationships(s) of litter load and of maternal
weight to fetal outcome; and the degree of similarity in the above factors for the different
compounds and species tested.
1. These studies will yield data sets that will allow analysis of the degree of intra-species
variability that may be due to individual factors.
2. The above analysis should allow the empirical placement of confidence limits on the
degree of inter-litter variation that can be ascribed to intra-species genetic variation.
3. The data generated in the studies should therefore allow regulatory personnel to
investigate and/or improve the rationale for the "100-fold" safety factor often used in
setting exposure limits for environmental agents. This safety factor assumes a 10-fold
intra-species variability due to genetic differences, an assumption not firmly based on
data generated in the test species employed in the laboratory.
4. Peer-reviewed publications on these studies, and the presentation of the data at
scientific meetings.
The extrapolation of developmental toxicity data generated in standard laboratory
bioassays often involves the setting of safety factors. These, in turn, are based in large
part, on presumed inter- and intra- species variability. Generally, factors of 10 are used
for each of these sources of experimental variability. While the accumulated data from
studies of developmental toxicity have allowed scientists to gain some insight into inter-
species variability, the question of intra-species variability as it may impact responses to
developmental toxicants has remained largely unanswered. The proposed studies will
directly address this issue for its enlargement. Such data can then be utilized to justify
and/or alter the safety factors employed to account for the component of variability due
to individual genetic differences.
1 yr FY89: $66K (for NRC Fellow)
N. Chernoff (OHR/HERL-RTP) (919) 541-2651
Topic:
Issue:
Status:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanisms Across Dose
New
-------
RIHRA
Appendix D - 28
Title:
Description:
Identified Results:
Usefulness of Results:
Maternal Toxicity and Nonspecific Stress in Developmental Assessment
EPA offices currently determine if a compound is a developmental toxicant by evaluating
data that has been collected using standardized testing. These guidelines suggest multiple
dose levels with the highest dose level being one that induces some form of overt
maternal toxicity. Maternal toxicity is usually defined as weight loss or lethality.
Uncertainty exists in the risk assessment of suspected developmental toxicants when
fetotoxic effects occur only at levels producing an alteration in maternal health status.
When agents are fetotoxic only at doses that are also maternally toxic, it is particularly
difficult to extrapolate from high to low dose. Thus greater uncertainty exists in the
ranking of these agents for selective developmental toxicity and for further toxicity testing
when the Lowest Observed Effect Level is the same for the adult and the developing
organism.
This uncertainty is partly due to the lack of understanding of the relationship between
maternal and developmental toxicity. It is currently difficult to separate direct fetotoxic
effects of an agent from those effects that are produced by an alteration in the maternal
compartment Current information is inadequate to assume that developmental effects at
a maternally toxic dose only result from maternal toxicity. Such agents may in fact be
directly active in both populations but be effective by different mechanisms. The
uncertainty in risk assessment exists because there is no adequate database concerning the
influence of maternal effects on the dose-response curve in the fetus. Because maternal
toxicity is not usually an area of study in and of itself but is rather a component of the
standard teratology protocol, there is little available data on the specific developmental
effects associated with defined alterations in the maternal compartment Thus, the lack
of a database on the effect of maternal toxicity and/or nonspecific stress factors on the
subsequent embryonic and postnatal development processes impedes the risk assessment
decision-making process.
This project focuses on the maternal compartment and will provide data on the specific
developmental effects associated with defined alterations in the maternal compartment.
The health status of the dam will be altered in several ways, including the use of agents
known to induce specific types of toxicity an/or conditions producing general or
nonspecific stress. Concurrent controls will include non-pregnant animals and dams that
are allowed to go to term for offspring evaluation. Biochemical, physiological, and
functional indices will be used to evaluate the organ systems involved in maternal toxicity
and nonspecific maternal stress. Collaborative portions of this project will include
teratological and histological evaluations.
1. Report that outlines the adverse developmental endpoints associated with specific
maternal alterations
2. Report containing database to demonstrate the frequency of each category of adverse
developmental effects resulting from maternally toxic and nontoxic doses of well-
characterized doses
The goal in dose-response assessment is to define the relationship of the dose of an agent
and the occurrence of developmental toxic effects. This relationship is then used to
extrapolate from high doses administered in experimental animals to the low exposure
levels expected for human contact with the agent in the environment. As noted in the
Developmental Risk Assessment Guidelines, it is difficult to correlate developmental
effects with maternal toxicity, especially in the cases where developmental and maternal
toxicity coexist. Information provided by this research will reduce the uncertainty
associated with ranking agents for specific developmental toxicity and aid in the decision
-------
RIHRA
Appendix D - 29
Project Length
and Cost:
Project Officer:
of whether specific chemicals should undergo further testing. Future research
approaches/strategies will be developed from the data produced by this project.
3 yr FY89: $71K
D. Miller (OHR/HERL-RTP) (919) 541-4186
Topic:
Issue:
Status:
Title:
Description:
Results:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanisms Across Dose
New
Maternal Stress Response to Chemical Exposure: Impact on Zinc Status During Pregnancy
General maternal stress, including that induced by toxic levels of chemicals, has been
demonstrated to adversely affect fetal development. Further, there is evidence to suggest
that stress affects zinc homeostasis. Zinc deficiency has been found to be teratogenic in
all species where it has been investigated, including humans. These studies will
investigate the relationships of zinc status, maternal stress response, and adverse
developmental outcome in experimental animals. If maternal zinc homeostasis is impacted
by the stress response, the zinc status of animals prior to and during pregnancy could be
a critical factor in determining Utter sensitivity. Components of compound-induced stress
(i.e., corticosterone levels, induction of metallothionein synthesis, and effects on thymus,
adrenal, spleen, and whole animal weights) will be measured following exposure of
pregnant rats and mice to diverse chemical toxicants. The levels of stress will be
correlated with the degree of developmental toxicity including effects on postnatal survival
of offspring. Interactions of zinc status with chemical toxicity and/or stress will be
assessed by measuring maternal and fetal tissue zinc concentrations in exposed and control
animals fed dietary zinc at levels varying from deficient to supplemented. If significant
interactions between compound-induced stress and zinc homeostasis are demonstrated,
further work will define the most sensitive tissue and/or intracellular sites by following
the absorption and distribution of 65Zn. All of the trace element and metallothionein
analyses, and a portion of the gross developmental toxicity studies, will be done at the
Institution receiving the Cooperative Agreement
1. Maternal/fetal tissue levels of zinc and developmental outcome following chemically-
induced stress under conditions of varying dietary zinc levels
2. Data on effects of chemically induced stress on maternal organ weights and plasma
corticosterone, and metallothionein concentrations in maternal and fetal tissues
3. Effects of chemically induced maternal stress on the absorption and distribution of
Zn65
4. Data derived from these studies will be presented at scientific meetings and published
in appropriate toxicology journals
-------
RIHRA
Appendix D - 30
Usefulness of Results:
Project Length
and Cost:
Project Officer:
The proposed research will yield data of value in two areas of regulatory toxicology. The
first concerns the relationship(s) of zinc homeostasis, stress, and adverse developmental
outcome in the laboratory setting. Standard teratology bioassays require exposures to test
agents at levels producing maternal toxicity and attendant stress during pregnancy.
Increased knowledge of the role(s) of zinc in the maternal and fetal responses to agent-
induced stress will improve our ability to interpret data obtained at these high dose levels.
The second area concerns the extrapolation of experimental data to humans. Zinc is an
integral part of over 100 critical enzymes, and in the United States it is thought that a
substantial percentage of women of childbearing age may not receive adequate levels of
zinc in the diet.
Given these facts and the putative relationship between zinc status, stress, and
developmental outcome, the proposed studies should generate data which will help define
the role of stress and/or zinc deficiency in intraspecific variation in sensitivity to
developmental toxicants. These data will therefore allow the Agency to consider dietary
zinc levels in human subpopulations during the process of extrapolation and the
determination of safety factors.
3 yr FY89: $75K
J. Rogers (OHR/HERL-RTP) (919) 541-5177
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanisms Across Dose
New
The Potential Existence of Low-Dose Thresholds for Developmental Toxicity:
Quantification of Cell Death, Cell Cycle Perturbations, and Repair Capacity in the
Mammalian Embryo
To create biologically based dose response models for developmental toxicity risk
assessments, knowledge is needed concerning the existence of doses below which
biologically significant effects are not seen. The critical question of the possible existence
of such threshold doses in teratology has never been answered and is essential for the
development of biologically relevant mathematical models for low dose risk assessment
by regulatory agencies. The proposed research will utilize novel approaches to determine
if such thresholds exist for developmental toxicity. Historically, research efforts directed
at this question have involved the use of extremely large numbers of animals to extend
the dose-response curve to lower levels. Such "mega mouse" studies have not been
successful since they have not addressed the central issue of the repair capacity of
embryos following xenobiotic-induced damage.
The proposed research uses a combination of a sensitive assay for embryonic cell death
using vital dye staining of whole embryos, flow cytometric cell cycle analysis, and a
detailed analysis of the most sensitive target structure(s) of the fetus including gross
examination, skeletal and cartilaginous analysis, and histopathology. These studies will
initially examine the role of cell death and cell cycle perturbations in the developmental
-------
RIHRA
Appendix D - 31
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
toxicity of several classes of developmental toxicants. By using cell death and cell cycle
effects as endpoints at multiple dose levels, we will be able to quantitate the repair
capability of the developing conceptus. This will yield information on the presence of
thresholds for different classes of teratogens. Studies done in collaboration with
laboratories studying embryonic cell death and flow cy tometric analysis of cell replication
cycles have indicated that such an approach is viable. While HERL has standard
teratology bioassay and flow cytometric capabilities, it lacks expertise in sensitive assays
for the identification of embryonic cell death. The bioassay for cell death, and the
detailed analysis of fetal morphology will be accomplished at the Institution receiving the
Cooperative Agreement.
1. A report on the effects of an alkylating agent (e.g. cyclophosphamide) on patterns of
cell death, cell cycle alterations, and repair capability in the murine embryo at dose
levels ranging from embryolethal to "sub-teratogenic."
2. Similar data on other classes of teratogens as that obtained with the alkylating agents.
This body of data may enable evaluation of the general applicability of the threshold
concept for developmental toxicity.
3. The data derived from these studies will be presented at scientific meetings and will
be published in peer-reviewed toxicology journals.
4. The data will be presented to appropriate regulatory personnel in the Program Offices
of the EPA in Washington, D.C. through seminars documenting the progress of this
research effort.
The quantitative risk assessment of developmental toxicants suffers from a lack of
biologically valid mathematical models for extrapolation of laboratory data derived at high
dose levels to the low dose levels relevant in environmental exposures. One of the major
problems in the establishment and validation of such models is the question of the
existence of threshold dose biological markers of toxic effects occurring at doses below
those producing gross terata offers the most viable approach to the elucidation of low dose
effects and/or the presence of threshold phenomena. The goal of the research is to
establish whether thresholds for developmental toxicity are a general phenomenon. The
conclusions of these studies will provide a biological basis for the inclusion or exclusion
of the threshold concept in the construction of dose response models.
3 yr FY89: $50K
N. Chernoff (OHR/HERL-RTP) (919) 541-2651
Topic:
Issue:
Status:
Title:
Biologically Based Dose-Response Models
4.1.1 Inter/Intraspecies Extrapolation: Homologous Models
Ongoing
The Relationship Between the Probability of Infertility in the Male Rodent and Ejaculated
Sperm Counts
-------
RIHRA
Appendix P - 32
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Reductions in sperm counts have been detected in workers exposed to environmental
pollutants like DBCP following complaints of long-term infertility. However, the utility
of routine on-site surveillance of ejaculated sperm counts (EJC) to determine if
occupational exposure has produced adverse effects on sperm measures is limited, except
in the most extreme cases like DBCP, because of the extreme variability in EJCs within
and between individuals. It is therefore essential to test chemicals for their potential to
induce adverse effects on sperm measures and fertility in laboratory species and to
extrapolate such effects to man.
To reduce the uncertainty in extrapolating effects seen in rodents to man, data must be
acquired on a common endpoint. In man the most common reproductive data available
through the years has been on sperm counts from fertility clinics. In fact, sufficient data
were available on sperm motility and there is a paucity of human data on other
reproductive endpoints (testicular biopsy for example). However, in the rodent EJCs have
been difficult to collect and analyze and, because of a lack of a common reproductive
endpoint, uncertainty exists on how to extrapolate effects seen in rodents to the human.
Recently, however, a few studies have described a methodology that surmounts the
problem of collecting an ejaculate from the male rodent (Peitz and Olds-Clark, 1986;
Zenick et al., 1984).
With these newly developed techniques, the relationship between the EJC of rodents will
be compared with infertility. This project will define the relationship in rodents between
EJC and infertility, compare the curves between the human and rodents, and determine
the relative value of other endpoints (testis weights, serum hormones, histology, testicular
sperm production, caudal sperm reserves, sperm morphology and motility). The EJC will
be manipulated through direct alterations of CNS, hypothalamic, pituitary, or gonadal
function. This information is currently unavailable and will enable us to provide
quantitative estimates of uncertainty associated with rodent to human extrapolation in male
reproductive toxicology.
Short-Term - Development of the mathematical relationship between ejaculated sperm
count and infertility in the rat using a model of congenital germ cell agenesis and
comparison of this function with that described for humans.
Long-Term - Development of a critical database that compares the relationship between
ejaculated sperm counts and infertility using chemical insults that target different male
reproductive tissues. Determination of the mathematical relationship of other reproductive
endpoints to the ejaculated sperm count and infertility.
The results of this task will be useful to both REAG/OHEA and the Program Offices.
REAG will utilize this database to model the relationships between measures of target
organ impairment and reduced reproductive success. Using the database developed in the
current task, the Program Offices will be able to predict the adverse effect on fertility
from changes seen in other reproductive endpoints. This is important because the
Program Offices often receive data indicative of reproductive alterations in the absence
of a fertility assessment.
3 yr FY89: $56K
E. Gray (OHR/HERL-RTP) (919) 541-7750
-------
RIHRA
Appendix D - 33
D.3 PULMONARY
By virtue of its functional intimacy with air and
its varied constituents, the lung is a target and/or
conduit for potentially toxic inhalants. Although our
concern resides in human health, the bulk of existing
toxicology that are acute or short-term. Hence, there
are at least two major uncertainties in our understand-
ing of inhalant toxicity: the dosimetric/extrapolative
linkage between experimental animals and man, and
the inter-relationship between exposure and duration
in the elicitation of response. The overall objective of
this initial program is to establish a framework that
has some general applicability to these issues. The
projects being initiated will not produce final or
universal answers, but will be systematic studies with
these issues as primary hypotheses rather than retro-
spective attempts to coalesce the data in a poorly
thought through context.
Ozone is one of a few toxicants for which there
is multi-species and multi-endpoint toxicology and for
which it is also possible to do experimental human
studies. With the aid of a novel nonradioactive
tagging procedure, it is possible to acquire data on
tissue and cell target dose which can be correlated
with biologic effect Obviously, more invasive detail
can be obtained in animals and at present considerable
data exist for the rat Limited studies are to be
conducted in humans that can address lavagable cell
dose data which can be related to inflammation and
immune function. Nonhuman primates offer the
opportunity to study dose/effects using identical
techniques and permits invasive assay of endpoints as
in the rat. Moreover, the larger size of the animal,
not to mention the similarity to man in lung structure,
increase both the sensitivity and precision of the
isotope analysis. Collectively, the diversity of expo-
sure across these animal species, the dosimetry which
includes man though in a more limited fashion, and
detailed cell biology assay will allow construction of
a credible model of lung injury which has a cross-
species dose and response framework for the purpose
of extrapolation and risk assessment.
The inter-relationship of concentration and time
is particularly important to the risk assessment pro-
cess. This relationship is especially critical to the
RfD process, which attempts to extrapolate over
lifetime periods. Recently, the assessment of re-
sponse over shorter periods as might be incidentally
encountered has drawn more attention. The some-
what fragmented existing data suggest that, in the
induction of lung injury, concentration dominates time
in very acute periods. However, as concentration falls
and time of exposure is extended, a more complex
interaction function is observed, at least in prelimi-
nary work with ozone. Does this relationship hold
with other air toxicants? endpoints? and over what
periods? Can we develop a mathematical function
that can be incorporated into the risk assessment
process which can include concentration/time relation-
ships?
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
New
The Use of Inflammatory and Immune Response Indicators to Study Exposure-Response
Relationships and Animal-to-Human Extrapolations
This project will provide information regarding the effects of pollutants on hose defense
systems in humans. It is intended to increase our understanding of exposure-response
relationships and animal-to-human extrapolation as they relate to host defense systems
The response of similar target tissues in humans and rodents to chemical insults will be
determined. Tissues from rodents and humans will be exposed in both in vivo and in
vifro settings. Cellular, biochemical, and molecular changes will be compared
Specifically, this effort will examine the effects of pollutants on human and animal
-------
RIHRA
fix D - 34
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
immune cells with particular focus on pulmonary alveolar macrophages. These cells are
the body's first line of defense against inhaled pollutants and microorganisms.
Macrophages also produce enzymes that can lead to inflammation in the lungs, as well
as tissue destruction. Thus, any adverse effect of pollutants on macrophages could lead
to an increased susceptibility to infection as well as respiratory problems. To achieve the
goals of exposure-response relationships and animal-to-human extrapolation as they relate
to hose defense systems, exposures to pollutants of interest will be performed in animals
and humans under controlled chamber conditions as well as in culture using human and
animal inflammatory and immune cell types.
Comparative data on the effects of pollutants on human and animal host defense systems
for risk assessors to judge differences in species sensitivity.
Because host defense effects are often among the most sensitive effects examined in risk
assessments, the results of this project will help reduce uncertainties in an important
endpoint (species sensitivity). The complementary research conducted in animals and
humans in vivo and in vitro using sensitive biological endpoints relevant to hose defense
mechanisms, will result in a better assessment of potential health effects associated with
exposure to pollutants. ,
3 yr FY89: $171K
H.S. Koren (OHR/HERL-RTP) (919) 966-6254
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.2/4.1.3 Inter/Intraspecies Extrapolation: Inter/Intraspecies Sensitivities
Ongoing
Establishment of Intraspecies and Interspecies Linkages Between Delivered Dose and
Toxic Effect- A Human/Monkey/Rat Comparison Using a Prototype Inhalant, Ozone
This project will determine the relative responsiveness (sensitivity) of humans versus
monkeys and rodents by correlating markers of both dose and toxic response in the three
species. There are few toxicants for which a large multi-species and multi-endpoint
toxicology database exists, and which can at the same time be feasibly studied in human
clinical or epidemiological studies or in all species of laboratory animal. Ozone, which
is such a toxicant, is being studied as a prototype inhaled pollutant in both human clinical
studies and rat studies now in progress at EPA, Research Triangle Park. Techniques
recently developed by EPA are being used which determine the dose of ozone (using
oxygen-18 labeling techniques) delivered to pulmonary cells obtained by bronchoalyeolar
lavage. This measurement of dose is being correlated with quantitative immunological,
biochemical, morphological, and molecular markers of toxicity. The ratio of delivered
dose to toxic effect observed is interpreted as the "sensitivity" to ozone of a given cell
or tissue.
The proposed cooperative agreement will provide data from studies with nonhuman
primates that will be complementary to the rodent and human data. The addition of the
-------
RIHRA
Appendix D - 35
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Possible Grantees:
Project Officer:
nonhuman primate data will add important morphological data on the airways and lung
parenchyma that cannot be obtained with humans or rodents because of the invasiveness
of scale problems. The availability of pulmonary lavage techniques makes it possible to
sample human and animal cells which have been exposed in vivo to ozone. Using new
labeling technologies, it will be possible to evaluate both the relative does and the
sensitivity of different species to ozone, and by analogy, to other inhaled pollutants. The
endpoints chosen will provide information on a broad spectrum of adverse health effects,
including impairment of host defense to microorganisms, as well as reversible and chronic
pulmonary health effects.
July 1990: Report entitled "Tissue dose effect studies of ozone in humans, rhesus
monkeys and Fisher rats: implication^ for human risk assessments of inhaled reactive
gases."
This project will establish for the first time a link between human clinical studies data,
nonhuman primate data, and rat toxicology data. It will do this at the tissue level, i.e.,
using measurements of the dose actually delivered to tissue and a multiplicity of toxic
response measurements. The animal data are essential to provide an understanding of the
full array of health effects of toxicants; however, the linkage of dose and sensitivity to the
human must be made in order to accurately assess risk in humans. Questions about the
applicability of rat data, or nonhuman primate data to the human, will be answered by this
project.
2 yr FY89: $75K
Dr. Charles Plopper, Dr. Dallas Hyde, Dr. Jerold, Last School of Veterinary Medicine,
University of California Davis, CA 95616; Dr. John R. Harkema, Lovelace Toxicology
Research Institute, P.O. Box 5890, Albuquerque, NM 87185
G. Hatch (OHR/HERL-RTP) (919) 541-2658
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.2.1 Exposure Scenarios: Mechanism^ Across Dose
Ongoing .
A Lung Injury Dose/Time-Response Model for Assessing Toxic Inhalants
A prototypic mathematical model initiated by OAQPS in an attempt to integrate data from
human phosgene exposures with animal data on B AL protein for risk assessment purposes
is replete with scientific and ethical uncertainties. Recently, another model describing the
CxT relationships of ozone on BAL protein has also been proposed but lacks through
concentration-time dependency information. A theoretical model needs to be integrated
and strengthened by mechanistic data to become scientifically acceptable for prediction
of human risk from combined animal and human data. The generation of validated
models describing high dose/low dose relationships in animals would be amenable to
critical study in the area of pulmonary toxicity. Pulmonary edema, caused by leakage' of
plasma from the blood into the air spaces, is sensitively and quantitatively detectable after
-------
RIHRA
Appendix D - 36
Identified Results:
Usefulness of Results:
high as well as low levels of acute pulmonary toxicant challenge. Hence, accumulation
of protein in response to damage and its disappearance, as is found in repair, could
provide a credible basis for addressing "generic" lung injury and its repair. Protein in
BAL of human subjects after exposure to oxidants has been reported and the kinetics of
leakage appear to be quite similar to those described in animals. The model developed
could provide a bioassay and extrapolative tool for risk assessment of noncarcinogenic
substances.
A lung injury dose-response model will be generated that is applicable to pulmonary
toxicity which permits intraspecies extrapolation between high and low dose and between
single and repeated exposures. The model would have particular utility in addressing the
poorly understood issue of acute to chronic relationships of toxicant lung damage and
would also provide a means to generically examine a common index of lung damage
likely with all inhalants.
From a matrix exposure design derived from existing preliminary data, a mathematical
model will be developed which describes empirically the kinetics and CxT relationships
of lung bronchoalveolar lavage fluid (BAL) protein accumulation, an indicator of alveolar-
capillary permeability. The application of the model will then be explored by
quantification of the deviations between expected and measured values under various CxT
exposure scenarios. Rat data will form the basis of the model; the general applicability
of the model will also be tested in other species. Other ongoing studies which employ
ClC4C-phosgene and CICSO-ozone will examine whether using the "retained dose" rather
than the "exposed dose" of toxicant will enable "normalization" of interspecies variations.
While the empirical model is being validated, mechanistic experiments will be performed
for the purpose of extending the model from a strictly empirical one to one which
explains permeability changes in theoretical terms. Questions that will be addressed
include: 1) "What is the potential influence of the protein turnover in the alveoli on the
toxicity measurements?" 2) "What molecular probes provide optimal detection of alveolar
permeability changes?" and 3) "Does repeated exposure reduce the sensitivity to
permeability changes and does this alter is utility for predicting chronic disease?"
1. Report on the mathematical relation between exposure concentration, duration, and
time post exposure which will be validated for two inhaled toxicants. The model's
accuracy in predicting measurements will be quantified in rats and spot-tested in three
other species.
2. A report describing the feasibility of normalizing dose-response using "retained dose."
The limitations of the existing model for risk assessments for phosgene as well as its
refinement criteria will be included.
3. A description of analogies between criteria pollutants (ozone and nitrogen dioxide) and
phosgene, and quantitative comparison of other toxicity endpoints (quantitative
morphology, physiology, etc.) with the measurement of BAL protein. Such compar-
isons will test the broad applicability of the model.
4. A report describing the utility of the model for studies of complex atmospheres and
varied exposure scenarios. Repeated exposure scenarios will be studied using the
model to predict cumulative damage, and experiments will be planned to investigate
the linkage between acute and chronic exposures.
The generation of a validated model describing high dose/low dose relationships in
animals would be amenable to critical study in the area of pulmonary toxicity. Such a
model would assist OAQPS in problems of extrapolating between animals and humans,
-------
R1HRA
Appendix D - 37
Project Length
and Cost:
Project Officer:
between acute exposures of different lengths and conditions, and between other endpoints
of toxic pulmonary response. The model would be applicable to Criteria or Air Toxic
pollutants, both gaseous and paniculate, which cause acute pulmonary toxicity. This
model will benefit any Program Office in need of data and/or models of lung toxicity for
regulatory processes. The RfD process would be greatly advanced with this model and
the minimal data ultimately needed to utilize it for a vast array of potentially toxic
substances to humans.
2yr FY89:$256K*
*Costs include inhalation engineering support.
Costs include technical biological support to conduct project.
G. Hatch (OHR/HERL-RTP) (919) 541-2658
D.4 GENETIC TOXICOLOGY
The principles and concepts applied to cancer
risk assessment are discussed in the 1986 "Guidelines
for Carcinogen Risk Assessment" (which are currently
being updated). These "Guidelines" point out major
areas of uncertainty encountered during the assess-
ment of human risk from exposure to environmental
agents. Because cancer is an endpoint of major
concern in regulatory decisions, the uncertainties
associated with cancer risk characterization are
important problematic issues that need to be addressed
in a well-designed research program. The research
areas of focus are in accordance with the issues
described in Topic 4 and concentrate on the develop-
ment of biologically based dose-response models for
high-to-low dose extrapolation and response equiva-
lence across species. These research needs relate to
Issues 4.1: Inter/Intraspecies Extrapolation and 4.3:
Mechanistic Variation, and contain aspects of Issue
4.2: Exposure Scenarios with respect to mechanisms
across dose and sensitivity of endpoints as a function
of dose.
A major uncertainty encountered in the risk
assessment process concerns the validity of using the
linearized multistage model (a curve-fitting approach)
for high dose-to-low dose extrapolation. A focus of
research in this area should be on the development of
alternative mathematical models that incorporate
information on key biologic processes and endpoints
currently thought to be involved in carcinogenesis as
opposed to statistical curve fitting. As pointed out in
the "1986 Guidelines," progress in this area is depen-
dent upon a better understanding of the underlying
mechanisms that lead to neoplasia. Of major impor-
tance in this respect is a better understanding of those
components of the cancer process, initiation, promo-
tion, and progression, which are universally critical to
the development of neoplasia, irrespective of the
particular carcinogenic agent.
Although the mechanism of action of chemical
and physical carcinogens is not known with certainty,
almost all human carcinogens that have been exam-
ined to date appear to be mutagenic or reactive with
DNA. It is this characteristic that currently permits
EPA's use of inferences to predict carcinogenic
activity. There remains a subset of carcinogens that
do not appear to be DNA-reactive, however. This
class, the so-called "nongenotoxic" carcinogens,
presents an added uncertainty with respect to appro-
priate risk assessment models because they may not
be best described by linear, nonthreshold low-dose
multistage models. In addition, there exist agents that
are not by themselves carcinogenic but can signifi-
cantly enhance the yield of tumors or accelerate or
modulate the progression of premalignant to malig-
nant lesions (ie.., tumor promoters) or the biological
characteristics of malignant cells (progressors). Very
little information exists on the mechanism of action of
"nongenotoxic" carcinogens. In some cases, com-
pounds (e.g., amitrol activated by peroxidative sys-
tems) thought to be nongenotoxic have been shown to
possess DNA-damaging activity. In other cases,
nongenotoxic carcinogens will alter genetic material
through indirect or non-DNA reaction mechanisms
(e.g., physical interference by asbestos in chromosom-
-------
RIHRA
Appendix D - 38
al segregation, modification by 5-azacytidine of
pathways involved in hypomethylation of DNA).
While definitive identification of specific steps
in carcinogenesis will remain elusive, a deliberate
research focus is proposed on the 1) identification of
intermediate biological endpoints of carcinogenesis, 2)
the generation of data on these endpoints, and 3) the
application of this information for the development of
biologically based mathematical models for low-dose
extrapolation. Research efforts should concentrate on
non-DNA reaction pathways or possible indirect
pathways by which target gene expression (especially
oncogenes and tumor suppressor genes) is changed.
It should be emphasized, however, the generation of
data is not necessarily a prerequisite for mathematical
model development; it is recognized that model
development may, in fact, help define important
parameters.
Risk assessment for carcinogens is a complex
process that relies on experimental data from animals
and where possible on human cancer epidemiologic
studies. It is frequently necessary to assume equal
sensitivity for tumor induction in rodents and humans.
Hence, another Topic 4 issue is the uncertainty
surrounding species-to-species sensitivity extrapola-
tion. Extrapolation of data from animals to humans
with a reasonable level of confidence requires that the
mechanism of action of carcinogens and the carcino-
genesis process in humans and experimental species
be similar. Some of the essential characteristics of
the process of carcinogenesis can probably be deter-
mined by modeling and comparing the time course,
dose-response, proliferation of stem cells and their
predecessors, and the kinetics of their differentiation
in relevant animal models as compared to what is
known about the incidence and time course of human
cancers. The testing of the assumption that humans
are as sensitive as experimental animals to the chemi-
cal induction of tumors can also be investigated by
techniques measuring certain genotoxic endpoints
(e.g., methods to detect gene mutations in lympho-
cytes, cytogenetic techniques measuring sister chrom-
atid exchanges, chromosomal aberrations including
micronuclei) which can be evaluated in both experi-
mental animal models and humans.
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.2.2 Exposure Scenarios: Sensitivity of Endpoints as a Function of Dose
Ongoing OHEA/RURA project
The Quantitative Relationship Between DNA Alkylations and Stable Chromosomal
Damage
For exposed populations of animals or humans, the level of DNA and nuclear protein
adducts and chromosome aberrations induced in peripheral blood lymphocytes (PEL) and
the level of formation of stable translocations induced in germ cells should each have a
direct quantitative relationship to the level of EtO exposure. Quantitative data correlating
these three endpoints—levels of DNA or protein alkylations, initial levels of chromosome
aberrations, and levels of induced stable chromosomal rearrangements following exposure
to various levels of EtO in any single test system—is not available. This project proposes
to quantitative the dose and dose-rate relationship between EtO exposure, the formation
of nuclear DNA and protein alkylations, the formation of chromosome aberrations at first
metaphase, and the induction of stable translocation.s at the 5th population doubling level
in the human fibroblast cell line IMR-90. If the quantitative relationships observed in
IMR-90 cells are the same or similar to those in PEL and germ cells and if these
relationships are similar for mice and humans, the data generated in this study will aid in
assessing the risk of heritable DNA damage following the demonstration of a given level
of either DNA or protein adducts or chromosome aberrations in PEL in EtO-exposed
human populations.
-------
RIHRA
Appendix D - 39
Usefulness of Results:
Project Length
and Cost:
Project Officer:
The proposed research project with Coriell Institute is designed to contribute to EPA's
efforts in developing biologically based dose response models for health risk assessment
by aiding in our understanding of the underlying causes of toxicity. Such biologically
reflective risk models will provide more meaningful estimates of health risk associated
with exposure to a toxic chemical. Ethylene oxide (EtO) was chosen for these studies
because its physical/chemical properties and metabolism are relatively well understood.
Moreover, there is a large database on the mutagenicity of EtO, including heritable germ
cell chromosome translocations in mice and information concerning its ability to bind to
or alkylate DNA and protein in animals and humans. The work at Coriell will aid in our
efforts to develop a mathematical model for assessing risk associated with chromosome
translocations. Chromosomal translocations have been associated with infertility, cancer,
and other medical disorders.
FY89: $95K
V. Dellarco (OHEA/GTAB) (202) 382-7332
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.3 Mechanistic Variation
Ongoing OHEA/RURA project
A Class of Biologically Based Dose-Response Models
The project is designed to develop a BB-DR model that includes the Moolgavkar and
Venson (1979) model as a special case. The Moolgavkar and Venson model makes some
undesirable biological assumptions and may not be applicable to data from animal
bioassays that are used in cancer risk assessment. The mathematical model which
requires numerical solution for some equations has already been developed. Our current
effort is to develop a numerical algorithm to solve equations and to develop a computer
software for constructing dose-response models.
1. A document describing mathematical derivation of the model and procedure for using
the computer software to construct BB-DR models
2. Computer software
3. Results will be submitted for publication
The BB-DR model and computer software developed from this research will enhance
CAG's ability in cancer risk assessment.
2 yr FY89: $75K
C. Chen (OHEA/CAG) (202) 382-5719
-------
RIHRA
Appendix D - 40
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.1.3 Inter/Intraspecies Extrapolation: Intraspecies Sensitivities
Ongoing OHEA/RURA project
Cohort Survivorship Analysis
In retrospective/prospective cohort studies, large deficits of mortality will often be seen
after the completion of an analytical phase. Several researchers have tried to explain these
deficits by citing the healthy worker effect and preferential survivorship. Others have
developed differing opinions regarding the reasons for these deficits and how these effects
alter the direction of the differences, e.g., lengthy employment results in an "excess" of
mortality. This study will utilize well-documented occupational cohorts that have been
detected health effects. The beginning and ending period of follow-up of the study cohort
will be represented by times A and B, respectively. A time Z will be chosen to represent
a point in time during this period, such that A
-------
RIHRA
Appendix D - 41
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
member (alkaline phosphatase), and possibly others, of a class of unique cell surface
proteins is lost from prospective tumor cells during the preneoplastic phase of the cancer
process. The carcinogen-induced loss of this marker protein has been shown to occur
across species (rats, mice, dogs, and humans) and carcinogen classes (nitrosamines, furans,
and 2AAF). Because the new system duplicates these observations in vitro and focuses
on a specific molecular event—the loss of a marker protein—it provides the opportunity
to study the early actions of initiators and promoters at the molecular level under carefully
controlled conditions.
This study will provide insights into initiation and promotion mechanisms and thus help
to distinguish these two important components of the cancer process. It also provides an
in vitro surrogate for the cancer process in animal species and humans.
These products will directly aid in developing biologically based dose-response models
by providing data for better distinguishing between initiators and promoters. Equally
important, the system opens up the possibility of pursuing modeling using in vitro
methodologies.
2 yr FY89: $30K
D. Reese (OHEA/GTAB) (202) 382-7342
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Biologically Based Dose-Response Models
4.3 Mechanistic Variation
New RIHRA project .-...;'
Biologically Based Dose-Response Modeling for Chromosome Translocation—An
important element in cancer and genetic risk assessments
The objectives of this research are:
1. To construct a biologically based dose-response (BB-DR) model for heritable
chromosome translocation. The model will be applied to data of heritable translocation
induced by ethylene oxide (EtO) which is part of the EPA's ongoing effort on genetic
risk assessment.
2. The mathematical formulation on the chromosome translocation will also be utilized
to construct a cancer BB-DR model which assumes that the chromosome translocation
is an initiation event in the context of initiation/promotion/progression carcinogenesis.
This research supports the EPA's ongoing effort on arsenic risk assessment (please see
the attachment) and BB-DR modeling in general.
3. To develop numerical algorithms and computer software for model construction.
1. A documentation of mathematical derivation of model and procedure for using the
computer software
-------
RIHRA
Appendix D - 42
Usefulness of Results:
Project Length
and Cost:
Project Officer:
2. A computer software for model construction
3. Results that will be submitted for publication
Chromosome translocation is an important component in cancer and genetic risk
assessments. The proposed research is the first attempt to construct a BB-DR model for
chromosome translocation. The product will enable OHEA to construct a biologically
based dose-response model on heritable translocation and will enhance OHEA's existing
ability in cancer risk assessment.
2 yr $80K/yr
C. Chen (OHEA/CAG) (202) 382-5719
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.3 Mechanistic Variation
Ongoing OHEA/RURA project
i • i
Generalized Dose-Response Model for Tumor Promoters Initiators
This project focuses on developing and testing a generalized dose-response model for
carcinogens which act as promoters, especially liver tumor promoters, since more
nongenotoxic agents which EPA regulates as carcinogens produce liver tumors in animals.
Included in this development is a method which incorporates the results from initiation-
promotion assays into the generalized MoolgavKar-Venzon-Knudson model. This model
also allows risk extrapolation for carcinogens which possess both initiating and promoting
capabilities.
..'.'•:.' ! '
Products will be a new tested model for quantitative risk assessment which incorporates
the mechanism of initiation-promotion. There will also be a computer program to solve
for the necessary parameters and low dose risk estimates.
Two advances here are: 1) use of cancer models which allow for various nongenotoxic
mechanisms for carcinogenesis and 2) use of results from initiation-promotion shorter term
assays in quantitative risk estimation.
3 yr FY89: $105K
S. Bayard (OHEA/CAG) (202) 382-5722
-------
RIHRA
Appendix D - 43
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose Response Modelling
Cross-Issue
Ongoing
Improved Methods for Combination of Quantitative Risk Estimates for Carcinogens
The goal of this work is to provide methods for incorporating a wide range of information
into Agency carcinogen risk assessments. To date, a variety of methods have been rather
inconsistently applied in those instances wherein there exists more than one data set
suitable as the basis for a quantitative estimate. These have included choosing the highest
quantitative estimate, rarely combining data prior to calculation of an estimate, and
combining estimates across studies, across sexes or across species. This project has had
a biological and a mathematical focus. The first step has been to examine existing
carcinogen risk assessments for those wherein this was an option for combined risk
estimation, cataloging the combination methods and identifying rationales for combining
or choosing a single estimate. This has been followed by an effort to evaluate sources
of valuability in data sets and evaluating risk estimates both from a biological as well as
from a statistical standpoint. Concurrently, there is taking place an analysis of methods
for preparing combined risk estimates; for example, the theoretical appropriateness of
taking geometric means of upper limits values resulting from application of a linearized
mutltistage procedure. Another approach being investigated is the application of statistical
tests for validity of combining data sets prior to application of an extrapolation model.
This work is proceeding through use of both extramural resources and Agency scientists
and mathematicians. A workshop on results is planned for Spring of 1990.
Specific guidance on use of multiple data sets for quantitative risk assessments for
carcinogens in the form of reports and published manuscripts. It is expected that results
of this work will directly impact revisions to the Guidelines for Risk Assessment of
Carcinogens.
Carcinogen risk assessments based on all applicable data will warrant more confidence
by risk assessment specialists and be better able to stand the test of public opinion and
the courts. Guidance as to validity of combination methods is needed by ORD and
program offices to produce such estimates.
3yr FY89:$40K
R. Schoeny (QHEA/ECAO-Cin) (513) 456-7544
W: Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
Ongoing
-------
RIHRA
Appendix D - 44
Title:
Description:
Identified Results:
Usefulness of Results:
Determination of Relative Rodent-Human Interspecies Sensitivities to Chemical
Carcinogens/Mutagens
The effect of chemical exposures will be measured by several complementary methods
using a combination of morphological, genetic, cytogenic, enzymatic, and molecular
biological markers. Detailed dose-response relationships will be modeled according to
biological and statistical concepts to obtain the desired extrapolation (sensitivity)
constants.
New molecular methods will be utilized to understand the fundamental interspecies
differences between the response of rodent and human cells to genotoxic agents. A
systematic evaluation of genotoxic responses will allow us to determine how genotoxic
effects in rodents extrapolate to similar effects in humans. Research has already indicated
that human cells may be more capable than rodent cells of repairing at least some DNA
lesions, implying that human cells may be less sensitive to genotoxic agents. Molecular
techniques (i.e., Insertion of human DNA repair genes and specific target genes) will be
used to define, quantify, and understand the mechanisms of the specific interspecies
differences. Such techniques have already proven genotoxicity. Quantitating these
differences will reduce the uncertainty in the interpretation of in vitro and in vivo rodent
data and allow a more realistic and accurate assessment of human risk from particular
environmental exposures.
One of the major assumptions in the cancer risk assessment process is the assumption that
humans will respond to the same extent as do rodents to the carcinogenic effects of
chemicals (i.e., equivalent tumor incidence, tumor multiplicity, and tumor latency at
equivalent doses). The uncertainty in the risk assessment process created by this
assumption can be quite large ranging up to several orders of magnitude. A research
program aimed at reducing this uncertainty by defining the relative response
characteristics of similar tissues from both rodent and human sources for different classes
of chemicals will help to reduce the uncertainty associated with this assumption and its
detrimental effects on the risk assessment process.
Deliverable Title: Report and journal article defining potential fundamental mechanistic
differences in the comparative response of rodent and human cells to genotoxic agents.
Deliverable date: 9/90
Deliverable Title: Report and journal article defining rodent cell test models which have
been modified (i.e., by the addition of human DNA repair genes and/or metabolic
activation genes) to make suitable for mechanistic comparison with human cells.
Deliverable Date: 7/92
Deliverable Title: Report and journal article defining species sensitivity constants for
specific chemical exposure to rodent and human cells.
Deliverable Date: 8/92
This research is directly applicable to the RIHRA process. The agency currently uses the
assumption that humans will respond to the same extent as do rodents to the
carcinogenic/mutagenic effects of chemicals. These assumptions include equivalent tumor
incidence, tumor multiplicity, and tumor latency at equivalent doses. The uncertainty in
this assumption can make the ultimate risk assessment incorrect up to several orders of
magnitude. By comparing the genotoxic response characteristics of similar rodent and
human tissues we can determine for particular classes of chemicals and particular types
of genetic events (of the type associated with malignancy) the species sensitivity
-------
RIHRA
Appendix D - 45
Project Length
and Cost:
Project Officer:
constants. These interspecies sensitivity constants will provide information directly
applicable to the risk assessment process.
3 yr FY89: $136K
S. Nesnow (OHR/HERL-RTP) (919) 541-3847
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
Ongoing
Dose-Response Relationships for Human Gene Mutation
The objective of this Cooperative Agreement is to compare the ability of human X-linked
and autosomal genes to quantitative genotoxic damage from chemicals which are
clastogenic (break chromosomes). This information will then be evaluated quantitatively
with data from the same genes in rodents to assess relative rodent/human responses to
genotoxic agents.
The major histocompatibility complex (HLA) of chromosome 6 will be developed for use
as a marker for induced genotoxic damage. A selection system for HLA mutants will be
analyzed. When the HLA system is developed, the ability of the HLA markers to detect
clastogenic damage will be evaluated and compared to that of the hgprt locus. This
comparison will include the relative magnitude of the induced frequency as well as
karyotypic and molecular analysis of mutants. The response of human cells to genotoxic
agents will be compared to that obtained in the currently used rodent cell systems.
This basic information concerning the gene mutation response in rodent and human cells
will be combined with other aspects of the rodent-human interspecies response to
genotoxic agents project. The information gained in the integrated project will allow the
determination of sensitivity constants for the rodent and human cells. This comparison
will provide information concerning the relative sensitivity of rodent and human cells to
carcinogenic/mutagenic chemicals.
One of the major assumptions in the risk assessment process is that humans will respond
to the same extent as rodents to the carcinogenic effects of chemicals. A systematic
evaluation of genotoxic response (including specific gene loci) will allow us to determine
how genotoxic effects in rodents extrapolate to similar effects in humans. This
Cooperative Agreement will develop the genetic markers and the comparative marker
results necessary to evaluate genotoxic effects in humans.
Deliverable Title: Determination of dose-response relationships for human cell mutation
using the HLA and HGPRT markers.
Report and Journal Article
Deliverable Date: 9/91
-------
RIHRA
Appendix D - 46
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Deliverable Title: Comparison of rodent and human gene mutation response to specific
classes of chemicals
Report and Journal Article
Deliverable Date: 2/92
This research is directly applicable to the RIHRA process. The Agency currently uses the
assumption that humans will respond to the same extent as do rodents to the
carcinogenic/mutagenic effects of chemicals. These assumptions include equivalent tumor
incidence, tumor multiplicity, and tumor latency at equivalent doses. The uncertainty in
this assumption can make the ultimate risk assessment incorrect up to several orders of
magnitude. By comparing the genotoxic response characteristics of similar rodent and
human tissues we can determine for particular classes of chemicals and particular types
of genetic events (of the type associated with malignancy) the species sensitivity
constants. These interspecies sensitivity constants will provide information directly
applicable to the risk assessment process.
1 yr FY89: $75K
M. Moore (OHR/HERL-RTP) (919) 541-3933
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
Ongoing
Dosimetric Analysis of Chemicals in Human and Rodent Tissue Response Models
To determine the dosimetric relationships between exposure, dose to the DNA of target
tissue, and genotoxic effects. These relationships will be used to define the use of DNA
adducts as internal target dose measures. The use of DNA adducts as target dose may
more accurately define specific dose response relationships with respect to dose and will
be useful in future species-species and route-route extrapolations.
To determine the dosimetric relationships between exposure and dose to the target and
genetic effects, this research effort will expose rodents, isolate peripheral lymphocytes,
and determine both sister-chromatid-exchange frequencies, a measure of genetic damage,
and DNA adducts, a measure of exposure dose. Preliminary studies will commence with
animals exposed by gavage or I.P. in order to clarify the parameters or treatment time,
DNA adduct persistence and repair, and structural identification of the individual adducts.
The chemical selected for study is Diazaquone (AZQ), a cancer theraputic agent. A
study population has been identified who are currently being treated with AZQ. Blood
samples will be obtained from these patients and SCE and DNA adducts quantitated.
Similar studies will be performed using exposed rodents. DNA adducts will be identified
by use of organic synthetic techniques. DNA adduct/SCE response curves will be
generated from humans and rodents exposed in vitro and in vivo. The data will be fitted
to the parallelogram model and extrapolation constants derived.
-------
RIHRA
Appendix D - 47
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
The development of exposure, dose, and genetic effects relationships for genetically active
chemicals will allow a more precise extrapolation of risk from rodent data to the human
population. Although this project is centered on genetic effects, the development of
exposure-dose relationships using blood lymphocytes is also applicable to other
lexicological endpoints where systemic exposure is a requirement for the induction of
adverse health effects. The product of this research will be an improved method for
species-species extrapolation in risk assessment.
DNA-adduct-response potency parameters for AZQ exposed rodents: 7/91 (report and
journal article)
DNA adduct response potency parameters for exposed hurrians and derivation of
interspecies sensitivity constants: 8/92 (report and journal article)
The experimental derivation of interspecies (rodent-human) sensitivity constants for
specific classes of chemicals will be a major step in reducing the uncertainties in cancer
risk assessment. This uncertainty is due to the assumption that man is equal to the most
sensitive rodent species in terms of sensitivity to chemical carcinogens. The results of this
project on AZQ-DNA adduct dosimetry will feed into an existing program on interspecies
sensitivity using AZQ as an agent to which exposed tissues can be obtained from rodents
and humans.
3 yr FY89: $290K
S. Nesnow (OHR/HERL-RTP) (919) 541-3847
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
New
Dose-Response Relationships at Low Carcinogen Concentrations in a Small Fish Model
This study is designed to further evaluate the small fish model, using the Japanese
Medaka, for its utility as a carcinogenesis bioassay system. The primary objectives of this
effort are 1) to determine bioassay parameters to measure a carcinogenic incidence below
the 1% level, 2) to describe the dose-response curve for diethylnitrosamine-induced
hepatic neoplastic lesions in this system, and 3) to further validate Medaka as a model for
human carcinogenic potential.
The study will consist of preliminary experiments to, among other things, determine
appropriate age and exposure parameters for maximizing response, characterize
pharmacokinetics and dosimetry, evaluate progression of DEN lesions, standardize
histopathological examination, and determine the statistical validity of a proposed design
for the definitive bioassay.
The study will be carried out by investigators at the Gulf Coast Research Laboratory
under a Cooperative Agreement with that institution. Principal investigators for this
-------
RIHRA
Appendix D - 48
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
project are experienced in large-scale experiments with this model system. The project
will be jointly funded from both Federal and industrial sources. Currently, funding for
the project from the Federal side is available through Interagency Agreements with the
Army and Air Force in addition to RIHRA funds. A project oversight group made up of
representatives from funding partners has been established to interact with the
investigators at GCRL.
This project will be important in the overall effort to validate the small fish model
carcinogenesis bioassay. More specifically, the project is expected to contribute to our
understanding of DEN carcinogenesis as a surrogate for other genotoxic carcinogens and
is expected to provide data on dose response at incidence levels of 1% or less.
The results of these studies will have direct benefit for ongoing fish work at ERL-
DULUTH. In addition, a statistically significant dose-response curve for carcinogenicity
at low levels will be useful in evaluating current approaches to low-dose extrapolation for
all cancer risk assessments.
FY89: $60K
W. Farland (OHEA/CAG) (202) 382-7315
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Biologically Based Dose-Response Models
4.1.2 Inter/Intraspecies Extrapolation: Interspecies Sensitivities
New
Interspecies Extrapolation for Oncogene Activation and Adduct Formation
This effort will use specimens of human bronchus from surgical or autopsy materials and
rodent lungs, specifically Strain A/J mice, as the experimental tissues for extrapolation and
comparison experiments; Tissues will be exposed to five different environmental suspect
human carcinogens, either in vivo (for Strain A/J mice) or in explant culture (human
bronchus and mouse lung tissues) and will remain in culture until the development of
preneoplastic lesions is observed. The lesions will be quantitated and then excised, or
some will be removed for histopathologic analysis. Lesions that have been excised will
be subjected to nucleic acid extraction, and the DNA amplified at least 100,000 fold by
the polymerase chain reaction (PCR) to yield enough DNA for direct sequencing by the
Sanger dideoxy technique. Alterations in the sequence of the ras family of oncogenes
(H,K, and N-ras) will be examined. The identity and quantity of DNA adducts will be
determined and their relationship to DNA sequence changes in the ras oncogene family
will be investigated.
1. Identification of adducts produced by human carcinogens in Strain A/J mouse lungs
(journal article) 9/91
2. Comparison between the adduct profile produced by several human carconogens in
human bronchus and Strain A/J mice (journal article) 5/92
-------
RIHRA
Appendix D - 49
Usefulness of Results:
Project Length
and Cost:
Project Officer:
3. Ras oncogene mutations in human lung lesions and Strain A/J mice and the
relationship to adduct formation (journal article) 9/92
This project is designed to determine if changes at the DNA sequence level induced by
relevant human carcinogens are similar in humans and rodents. Therefore, this project
concerns species-to-species extrapolation of the effects of carcinogens in an important
human cancer target tissue: the lung. At present, we do not know how early in the cancer
process that ras oncogene alterations take place, or if the temporal activation pattern is
similar in rodents and humans. Insight into these questions will lead to determination of
whether the mechanism of action of carcinogens is similar in humans and rodents, and this
evaluation will undoubtedly bear on judgments as to what kinds of risk estimation models
are appropriate for these human carcinogens.
2 yr FY90: $135K
M. Mass (OHR/HERL-RTP) (919) 541-3514
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.3 Mechanistic Variation
New
Derivation of the Parameters of the Moolgavkar-Knudson Two-State Carcinogenesis
Model Using Carcinogens and Tumor Promoters
This project is designed to provide information primarily on the effect of tumor promoters
and nongenotoxic carcinogens on the development of stages in the carcinogenesis process.
Normal and premalignant cell populations and their kinetics of replication will be
quantitatively analyzed. • /
Derivation of the parameters of the Moolgavkar-Knudson model requires experimental
systems in which the number of normal target cells are known, the number of
intermediate cells are recognized, the transition rates are observable, the states of terminal
differentiation and proliferation can be recognized and numbers quantitated, and the
malignant lesion can be scored. Experimental models where recognition of these stages
of cells is possible include rat liver focus induction, rat mammary tumor models, mouse
skin models in vivo or in vitro, or rat tracheal cells in vivo or in vitro. Tissues chosen
from the appropriate experimental model will be exposed to carcinogens to generate the
initiated "intermediate cell," and the transition rate constant (mu-1) will be calculated from
appearance of markers of the initiated phenotype known for each tissue. The constants
relating terminal cell differentiation and proliferation (alpha and beta), which will give
information on the effect of tumor promoters, might be calculated by classical
autoradiography. The transition rate constant between intermediate cells and malignant
cells (mu-2) can be quantitated biologically. Data will be examined for appropriateness
of fit to the Moolgavkar-Knudson model, a modified model will be constructed or other
models will be examined, and the relevant constants extracted.
-------
RIHRA
Appendix D - 50
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Experimental and epidemiologic studies have provided strong evidence that carcinogenesis
is a process occurring in more than one stage. The Moolgavkar-Knudson model predicts
that the first stage, initiation, is a mutation involving transition of a normal cell to an
"intermediate cell" with a defect in control of cell proliferation. The numbers and
proportions of these cells within the normal population are dependent upon the
intermediate cell's program to terminally differentiate and die, or to remain proliferative;
both these activities (induction of terminal differentiation and proliferative activation) are
known characteristic actions of tumor promoters and antipromoters. This implies that the
number of intermediate (initiated) cells can be regulated by tumor promoters. A second
mutation is postulated to be responsible for the intermediate cell's conversion to a
malignant cell that is strictly committed to proliferation. Therefore, cancer can be
envisioned as a process that can have genetic components (creation of intermediate and
malignant cells by mutation), and epigenetic components such as regulation of the number
of offspring of intermediate cells by tumor promoters. The formal relationship has been
defined by the Moolgavkar-Knudson incidence equation, which uses constants to define
the probabilities of intermediate cell and malignant cell creation, and rates of proliferation
or differentiation.
Determination of the transition rate constant for generation of intermediate cells for
several carcinogens: 8/91 (report and journal article)
Determination of the effect of tumor promoters on constants alpha and beta: 8/92 (report
and journal article)
The Moolgavkar-Knudson model for assessing the biological effects of carcinogens and
tumor promoters: Application to several data sets and implications to risk assessment:
8/93 (report and journal article)
This research effort has direct application to the RIHRA process in that the Moolgavkar-
Knudson model provides a means by which to quantitate the effects of carcinogens and
tumor promoters, and their postulated differential effects on initiated cells. The Agency
has made regulatory decisions involving agents such as gasoline vapors, trichloroethylene,
and dioxin whose mechanism of action as carcinogens do not fall into the category of
genotoxic; therefore, it is necessary that the Agency develop proper guidelines based
upon the mechanism of action of these compounds that will ensure more appropriate risk
estimations. The project has direct benefit for the Agency's need to develop relevant risk
assessments. Genotoxic carcinogens, which are present in all environmental media over
which the Agency has purview, appear to be significant human health hazards, and our
understanding of their mechanism of action is poor.
3 yr FY89: $170K
M. Mass (OHR/HERL-RTP) (919) 541-3514
Topic:
Issue:
Status:
Biologically Based Dose-Response Models
4.3 Mechanistic Variation
New
-------
RIHRA
Appendix D - 51
Title:
Description:
Identified Results:
Comparative Studies on Changes in Gene Expression During Carcinogenesis in Human
and Rodent Cells after Exposure to Nongenotoxic Carcinogens
This project will examine through molecular techniques the changes in gene expression
occurring in both human and rodent cells exposed to nongenotoxic carcinogens. The
purpose of the effort is to determine if similar mechanisms are responsible for the cancer
process in the two species.
The mechanisms by which nongenotoxic carcinogens produce carcinogenic effects are not
fully known in either rodent or human tissues. Many diverse mechanisms have been
proposed. Although the initial mechanisms of action often differ with respect to the
initiation of cancer, some commonalities may exist in the stages of the carcinogenic
process after the initial chemical insult It is these stages that need to be measured by
state-of-the-art molecular biological techniques, especially at the level of gene
transcription and gene expression. Possible endpoints include the induction of protein
kinase c or other enzymes associated with nongenotoxic carcinogenesis, alterations in the
expression oncogenes or tumor marker genes, and alterations in the structure of
oncogenes.
Primary mesenchymal or epithelial cells of rodent and human origin will be cultured and
exposed to nongenotoxic carcinogens. The effects of exposure will be measured at the
molecular level in terms of alterations in gene expression of those genes thought to be
associated with the induction or progression of the cancer process. Histological,
morphological, and enzymatic changes will be examined to correlate with the changes in
gene expression markers, particularly with respect to preneoplastic or neoplastic changes.
If carcinogenesis assays are not available for a specific human cell type, such assays could
be developed by use of gene transfer techniques.
Uncertainty in carcinogenesis risk assessment arises, in part, from a lack of understanding
about the underlying chemical and biological mechanisms that are responsible for the
development of the cancer cell. Nongenotoxic carcinogens are more difficult to study
from a mechanistic point of view due to the initial diversity of potential mechanisms. By
investigating and measuring the events subsequent to the initiation of the cancer process,
such as the changes in gene transcription and gene expression that occur in the cell
•leading to tumorigenesis, this research effort may uncover a more unifying mechanism
that would simplify the assessment of these chemicals as well as make it easier to explain
differences between human and rodent responses.
Determination of the changes in gene expression in human cells after exposure to select
nongenotoxic carcinogens (peroxisome proliferators, hormones): 8/91 (report and journal
article)
Determination of the changes in gene expression in rodent cell after exposure to select
nongenotoxic carcinogens (peroxisome proliferators, hormones): 10/91 (report and journal
article)
Determination of the changes in gene expression in human cells after exposure to select
nongenotoxic carcinogens (tumor promoters, hyperplasiogens): 7/92 (report and journal)
Determination of the changes in gene expression in rodent cell after exposure to select
nongenotoxic carcinogens (tumor promoters, hyperplasiogens): 9/92 (report and journal
article)
-------
RIHRA
Appendix D - 52
Usefulness of Results:
Project Length
and Cost:
Project Officer:
In the short term, the judicious selection of chemicals to study based on chemicals
requiring or soon to require regulatory action, will provide extremely important
information needed for the risk assessment process of those chemicals. The long-term use
of these products could be the simplification of the risk assessment process of
nongenotoxic carcinogens through a unifying mechanistic hypothesis and through a more
facile extrapolation of data from cancer studies in rodents to the human situation.
3 yr FY89: $120K
S. Nesnow (OHR/HERL-RTP) (919) 541-3847
Topic:
Issue:
Title:
Description:
Biologically Based Dose-Response Models
4.1 Inter/Intraspecies Extrapolation
Molecular Markers as Indicators of Chemically Induced Altered Foci
The principal objectives of this research are to evaluate the following factors in mouse
liver and rat liver and kidney following treatment with non-genotoxic carcinogens:
1. Ability of tumor markers to consistently detect altered foci, nodules and carcinomas
from mouse and rat tissues
2. Growth rates and time-to-appearance of nodules and carcinomas
3. Frequency of phenotypic complexity as number of markers expressed/ normal tissue,
altered foci, nodules or carcinomas per unit time
4. Involvement of normal and activated p21 ras in the carcinogenic process.
The project will be initiated to complement the Biochemical and Molecular Toxicology
Branch's studies into the carcinogenicity of drinking water chemicals with information on
the molecular mechanisms underlying the nongenotoxic carcinogenic process. Research
will focus on the ability of tumor markers to consistently detect altered foci, nodules and
carcinomas and measurement of growth rates and time-to-appearance of nodules and
carcinomas. Six enzyme-altered foci will be studied as potential markers for
hepatocarcinomas in the B6C3F1 mouse and the F-344 rat:
1. p21 ras (both normal and activated)
2. pi 10 myc
3. Tumor-associated isozyme of aldehyde dehydrogenase (BALDH)
4. Glutathione-S-transferase (GST-P)
5. Gammaglutamyl transpeptidase (GGTase)
6. GIucose-6-Phosphatase (G-6-P)
Previous research on a large number of samples has yielded valuable insights for future
research directions. The studies have progressed to the point where reliable results have
been obtained with several immunohistochemical markers. Continued studies on livers
from animals exposed to chlorinated acetic acids or phthalate esters are expected to more
-------
RIHRA
Appendix D - 53
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
clearly define the markers most useful for quantitation and interpretation of the
carcinogenesis process.
Future research will involve more emphasis on different markers and use of
'autoradiography to estimate growth rates of foci and nodules. These estimates are
'considered worthwhile because previous research indicated some markers are much more
useful than others, and will contribute to development of mathematical models in
carcinogenesis.
1. Establish molecular marker profiles for preneoplastic foci, nodules and carcinomas in
mouse liver (12/90) and rat liver (8/91) and kidney (6/92): (reports and journal
articles)
2. Determine growth rates of altered foci, nodules and carcinomas in mouse liver (2/91)
and rat liver (12/91) and kidney (9/92): (reports and journal articles)
3. Biologically based models for assessing effects on dichloroacetic acid and phthalate
esters in mice (2/91) and rats (12/91): (reports and journal articles)
Data from the proposed studies should contribute to an understanding of non-genotoxic
carcinogenesis mechanisms, development of mathematical models of carcinogenesis and
assist the Agency in extrapolating test results to low environmental levels. These data can
be used to test mathematical computer models using the biologically based two-stage
carcinogenesis model (Moolgavkar model). This model is attractive for risk assessment
because it is based on, and constrained by, biological events at the cellular level.
Immunohistochemical techniques appear especially well suited to mathematical modelling
because they permit clear identification of putative preneoplastic lesions from normal
background tissue to facilitate quantitative size determination by computer-assisted image
analysis. Understanding the mechanisms of action and development of mathematical
models of carcinogenesis may permit more accurate prediction of risks associated with
environmental exposure to chemicals based on animal studies.
3 yr FY90: $90K
B. Daniel (OHR/HERL-Cin) (513) 576-7401
D.S CROSS-CUTTING ISSUES
In developing the research focuses under Topic
4, several activities were identified that would benefit
the risk assessment process across target systems
(processes). The genesis of much of this research lies
in ECAO-Cincinnati, where those scientists have had
the lead in implementing the RfD process. This effort
has included the identification of critical data gaps,
the resolution of which would improve quantitative
risk assessment in general and specifically the RfD
methodology. Recognition of these research needs is
reflected in support (OHEA/RURA) for the continua-
tion of efforts to improve the statistical basis for the
derivation of the RfD and the development of proto-
cols to better estimate health risks associated with
short-term exposures. Additional, proposed projects
to complement and expand those activities would
address the characterization and quantification of
severity of toxic responses and the further develop-
ment of models that incorporate varying exposure
durations into the risk estimation.
Two other "generic" projects have been pro-
posed. One is directed at convening workshops to
address specific issues that are identified in the
development and implementation of the RIHRA
program. The second project attempts to benefit from
-------
RIHRA
Appendix D - 54
the extensive regulatory databases accumulated within
the Agency (e.g., in OPP). Research efforts would
entail organization of those data for subsequent
evaluation on a variety of issues, including pharmaco-
kinetics, species comparability, SAR relationships,
and the relationship of various toxicities over varying
dose and exposure conditions. (To implement this
"database" project would require substantial resources
and involvement of a number of Program Offices. As
such, funding is not recommended under the RIHRA
program. However, it is important to present this
project in light of its substantial impact if a coordinat-
ed Agency effort could be funded.)
Most of the environmental pollutants considered
by EPA for regulation are noncarcinogens, i.e., the
major toxic effect is something other than cancer.
The risk assessment procedures for noncarcinogens,
however, lag behind those for carcinogens in terms of
quantitative development. In 1980, EPA published
the Guidelines for Ambient Water Quality Criteria,
which outlined risk assessment procedures for both
carcinogenic and general noncarcinogenic health
effects. Since 1980, only minor changes have been
implemented for noncarcinogens. For carcinogens,
EPA generates a dose-risk relationship that can be
applied to any exposure level. In contrast, for noncar-
cinogens, EPA generates a Reference Dose (RfD).
Exposures corresponding to intakes below the RfD are
presumed to be without appreciable risk, while those
above the RfD are presumed to be linked to some
unknown risk. One reason for this difference in
procedures is the difference in complexity of the
corresponding toxicity data. Cancer data usually
include only one type of effect—malignant tumors—-
which are usually considered fatal. In contrast,
noncancer toxicity commonly involves several types
of effects, in different organs, and of differing levels
of severity. Modeling the multivariate noncancer
response constrained by the presumed existence of a
theoretical threshold is much more difficult than
modeling cancer risk.
One weakness inherent in current procedures is
that duration of exposure is not explicitly factored
into the risk estimation. For example, the calculation
of the RfD only involves one selected data point from
the entire set of toxicity data. One recent EPA effort
to clarify and improve the noncancer risk assessment
is the preparation of Agency-consensus Risk Assess-
ment Guidelines for Noncarcinogenic Health Effects.
In these Guidelines and in recent OHEA publications,
models are discussed that allow characterization of
health risk over a range of dose levels and exposure
durations. Research on these models will substantial-
ly change the procedures for noncancer risk estima-
tion, improving both the accuracy and breadth of
application of the risk estimates.
Both qualitative and quantitative guidance is
needed to define adverse health effects and to charac-
terize their biological significance (i.e., severity).
Accordingly, the severity project involves both
biological and statistical issues. Some of the research
topics include the relationship between subtle effects
from short-term exposure and more severe effects that
follow long-term exposure; mechanistic relationships;
statistical and graphical methods for evaluating
multivariate relationships (e.g., dose and duration
predicting multiple effects).
The development of the noncancer guidelines
has been greatly facilitated by several workshops and
symposia sponsored during the past five years by
ECAO-Cincinnati. In the numerous papers and
recommendations from these meetings, the dominant
quantitative issue is the involvement of more, if not
all, of the toxicity data in the risk estimation proce-
dure. The research tasks described herein address
some of those recommendations, focusing on the
characterization of toxic severity, on the inclusion of
exposure duration as a model covariate, and on the
calculation of probabilistic risk of unspecified toxic
effects.
The OHEA/RURA base program on generic
issues in FY89 includes specific projects that repre-
sent extensions of work initiated in FY88 on probabil-
istic characterization of uncertainty factors, use of
dose-response models for estimating RfDs, sever-
ity/adversity characterization efforts, and use of less-
than-lifetime exposure data in risk assessment. The
base program for FY89 will provide explicit examples
of alternatives to the current RfD procedure. The
severity projects proposed herein will also demon-
strate alternatives and improvements to the current
procedure. When both sets of projects (RURA base
and proposed RIHRA) are completed in FY90 or
FY91, the results should be combined into a collec-
tion of proposed revisions to the Guidelines for
Noncarcinogenic Health Effects (in draft, planned
completion, FY90).
-------
RIHRA
Appendix D - 55
Topic:
Status:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Workshops on Identified Issues in RIHRA Research Program
New Project Funded Under FY89 OHEA/RURA
This activity is designed to convene workshops of experts to address specific issues
identified in the development and conduct of the RIHRA program. The topics will often
represent areas for which research strategies need to be developed that identify and
prioritize the most important questions. Research efforts can then be directed to answer
these questions. NAS and ELSI are currently designated as workshop coordinators.
Each workshop will produce a strategy document that will be utilized by the RIHRA
Subcommittee in research planning (and resource allocations).
The major benefactor of these workshops will be the RIHRA program and, in turn,
Agency risk assessors. For several areas of uncertainty, there is no existing consensus
formula as to the best approach. These workshops will serve to clarify these issues for
a given topic and produce a research strategy that represents the most current scientific
knowledge. Such guidance will reduce the probability of embarking on haphazard, less
credible research projects. The end result will be the efficient use of RIHRA resources
and the production of a quality product that will further the risk assessment process.
Continuous FY89: $150K; RIHRA Committee will set out-year amounts
W. Farland (OHEA) (202) 382-7315
Topic:
Issue:
Status:
Title:
Description:
Biologically Based Dose-Response Models
4.2.2 Exposure Scenarios: Sensitivity of Endpoints as a Function of Dose
New RIHRA project
Severity of Toxic Response Stratification and Modeling Project
Essential to the use of severity data in risk assessment is the identification of scientifically
credible criteria to define an effect as adverse, and to establish the relative severity of
each effect The range of potential lexicological effects may be construed as of uncertain
significance,adverse, compensatory or adaptive. This judgement will vary as a function
not only of the observed effects, but perhaps most importantly, the target organ ortarget
system involved. In addition, for purposes of hazard assessment severitystratification
within the adverse effects would be of significant utility. The toxicology of the various
organ systems (e.g. hematopoietic, immunologic, and nervous) need to be reviewed in the
context of establishing criteria that constitute or define an adverse effects of these organ
systems. A fundamental aspect of this effort is to establish on an organ-specific basis the
relationship between histopathological progression, functional impairment of the organ
system,and ultimately, the organism. This effort will build upon ongoing activity in the
area of systemic toxicity to define both qualitative and quantitative guidelines to stratify
the spectrum of critical effects in terms of their degree of adversity. The general
approach will be to analyze the existing literature for mechanistic theories or statistical
medels, and to consult experts in the appropriate fields to develop guidelines for the
-------
RIHRA
Appendix D - 56
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officers:
interpretation and modeling of effects with respect to severity. This research effort will
lead to a practical, qualitative ranking scheme that allows for systematic determination of
equivalent severity scores for the spectrum of effects likely to be reported in the literature,
as well as statistical models that can be used to describe a dose-severity relationship for
use in risk assessment Reversible or transient effects and "u-shaped" dose-response
curves, which are often reported but generally dismissed as not significant, will also
receive special attention. Some chemicals, such as some metals, are essential elements
and consistently display such a curve. The u-shaped curves are particularly difficult to
evaluate and model. The peculiar influence of threshold parameters on model behavior
is not will understood and needs to be investigated on both statistical and numerical,
grounds, since many such models must often be constrained at the extremes of the dose-
response relationship.
1. Identification of standard criteria to define adverse effects
2. Development of Agency-wide guidance for qualitatively stratifying endpoints and
integrate the relationship between histopathological progression, organ dysfunction and
organismal disability
3. Development of an overall severity ranking scheme and documentation for major organ
toxicity
4. Development and improvement of statistical and graphical methods for evaluating
multivariate relationships
5. Development of the following monographs:
a) correlation of in vitro and in vivo immunotoxicity
b) significance of reversible/transient effects
c) mechanistic explanations for u-shaped dose-response curves for essential and
nonessential chemicals
d) numerical stability and robustness of categorical regression models with and
without threshold parameters
e) concordance of lexicological endpoints among species
Applications of the results of this research are both qualitative and quantitative. A
potential quantitative application arises directly from the foregoing in that no allowance
is made for adjusting the RfDf or severity of effect, except when the RfD is based on a
LOAEL. In the latter case, limited guidance as to the magnitude of the adjustment is
given. The issue is significant because of the wide range of severity of first-observed
effects among lexicological studies. A biologically based severity ranking scheme would
provide a basis for rational adjustment of the RfD on this basis. In addition, for essential
chemicals, guidance can be given to ensure thai the RfD does nol drop below Ihe
essential intake level. The primary product of this research effort, the ranking scheme,
is also essential for the development and implementation of dose-risk models. The RfD
does not provide any indication of the health risk for higher doses. In contrast, many
regulatory efforts in EPA evaluate health risks at existing rather than target, exposure
levels and could iherefore benefil from the availability of such risk models.
3 yr FY89: $140K
Richard Hertzberg, Deborah McKean (OHEA/ECAO-Cin) (513) 569-7531
M. Davis (OHEA/ECAO-RTP) (919) 541-4162
-------
RIHRA
Appendix D - 57
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Biologically Based Dose-Response Models
4.2.2 Exposure Scenarios: Sensitivity of Endpoints as a Function of Dose
Ongoing OHEA/RURA project
Research to Support Use of Short-Term, Episodic and Other Less-Than-Lifetime Exposure
Data in Risk Assessment
The evaluation of risk from exposure to toxic substances forsome fraction of a human
lifetime is the subject of ongoing research by the EPA's ECAO. Reference Doses (RfDs)
have been used for subchronic and chronic exposures and Health Advisories have been
used to address risks from 1-Day, 10-Day andLonger Term exposures in drinking water.
Both methods require adequate lexicological data for the derivation of health-based
numbers. Reportable Quantities for Acute Mammalian Toxicity address acute exposures
and result in a number useful for priority setting. The RQ is of limited use to risk
assessors because it does not project "safe" and "unsafe" doses. In the absence of data
for the exposure duration of concern, it is often necessary to project theoretical thresholds
from limited data sets. A more useful method would be to generate estimated dose-
response curves from limited data. Toward these ends, a number of studies have
examined the relationship between measures of acute toxicity (LD50, ED50, etc.) and
experimentally determined NOAELs. As proposed, this "ratios approach" would be used
to estimatea NOAEL for a chemical from limited data. It is not anticipated that this
method be used in lieu of current RfD methodology. The probit-slope of the dose-
response curves for a variety of compounds has also been examined for a variety of
compounds in order to determine the feasibility of estimating the dose-response curve for
a variety of endpoints based upon limited data. Both approaches include the creation of
a computerized toxicity data base by accessin other existing data bases. At the opposite
end of the spectrum are chemicals with a very large toxicity data base. Two procedures
are being developed to facilitate visual interpretation of complex and diverse data sets.
Graphics for plotting severity of effects data for dissimilar experiments on a common,
human-equivalent scale f dose and duration have been used to identify combinations of
dose and duration which may result in adverse effects. Three dimensional (3-D) graphics
are also being developed for use when data sets are very large and complex. Statistical
techniques are also being explored for generating confidence intervals around a 3-D dose-
response surface. Research is also being conducted on a modification of existing
nonparametric generalized rank methods for examining the effect of exposures across
multiple organ systems in which the degree of severity cannot be quantified. Another
project examines the relationship between temporal patterns of exposure and the severity
of effects. The problem stems from the fact that the observed effects are related not only
to the total dose but also to the dosing scenario.
1. For each project, a computer program will be generated along with guidance for its use
and application to risk assessment. Monographs on the following topics:
2. A ratios approach to estimation of NOAELs based upon acute toxicity data
3. Estimation of the probit slope of the dose-response curve based upon acute toxicity
data
4. 3-D graphics for trend analysis in risk assessment
5. Graphic analysis of dissimilar data in risk assessment confidence intervals for
lexicological data arrayed in three dimensions
-------
RIHRA
Appendix D - 58
Usefulness of Results:
Project Length
and Cost:
Project Officers:
6. Novel statistical procedure for quantitative risk analysis at multiple lexicological
endpoints
7. The relationship between temporal dosing patterns and duration and the severity of
observed effects.
Specific data gaps and uses for products are described with each project description above.
All of these projects will contribute to improving health risk assessment to a substantial
degree. For each project, proposed guidance is expected to be developed as a result of
the peer review process. In general, the methods developed will permit estimates of
NOAELs and dose-response curves based upon limited data; facilitate trend analysis,
response prediction, and significance testing of large data sets and data from dissimilar
experiments; and provide guidance for both investigators and risk assessors on the
chemical-class-specific effects of dosing patterns and duration of exposure.
3 yr FY89: $128K
C. Cubbison, C. DeRosa (OHEA/ECAO-Cin) (513) 569-7531
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Biologically Based Dose-Response Models
Cross-Issue/Cross-Target
Ongoing OHEA/RURA project
Reference Dose: Investigating Underlying Probability
The objectives of this project are to determine probability density functions of uncertainty
factors (UF's) and the statistical variablility of the NOAEL, and to test various theoretical
dose response models that may improve the current Reference Dose (RfD) methodology.
The approach to the former goal is to collect and analyze data from toxicology studies
(at the dose group level) in order to determine the experimental variability underlying the
defined uncertainty factors. Dose response models will be investigated as a means 'to
expressing the uncertainty in the NOAEL, or replacing the NOAEL with some other
estimate. Models (or procedures) to be tested currently include Crump's benchmark
approach, Brown's exact NOAEL procedure, and Hertzberg's categorical regression
model.
Several manuscripts discussing specific areas of uncertainty and dose-response modeling
(4. papers completed FY89, 4 papers in preparation FY90); computerized data base of
experimental studies for estimation of various uncertainties.
The probabilistic limits on the RfD by analysis of its subcomponents and/or the
calculation of more mathematically precise RfDs by way of modeling are particularly
important goals for improving the credibility of quantitative risk assessment. The ability
to express the RfD in a statistical framework would greatly facilitate the interpretation and
use of this estimate in risk management decisions.
-------
RIHRA
Appendix D - 59
Project Length
and Cost:
Project Officer:
2yr FY89:$110K
J. Swartout (OHEA/ECAO-Cin) (513) 546-7811
Topic:
Issue:
Status:
Title:
Description:
Identified Results:
Usefulness of Results:
Project Length
and Cost:
Project Officer:
Biologically Based Dose-Response Models
4.2.2 Exposure Scenarios: Sensitivity of Endpoints as a Function of Dose
New RIHRA project
Refining Risk Assessment Models by Investigating Critical Issues in Interspecies
Extrapolation
This project is designed to refine risk assessment models by comparative analyses of
existing clinical, epidemiologic, and toxicology data concerning the effects of well-
characterized chemical exposures. Extensive animal and human data are typically
generated in the process of development and evaluation of new pharmaceutical agents and
other consumer products. These data, the product of many millions of dollars of research,
may include short-term and long-term mortality and carcinogenesis testing in experimental
animals, organ system toxicity, dose-limiting toxicity, as well as extensive
pharmacokinetic and other biologic data in both animals and man, and epidemiologic data
concerning individual differences and long-term effects in humans. Because the focus of
the use of the toxicologic data tends to be the development of clinical trials, and the
results of the clinical trials ultimately speak for themselves, rigorous comparisons of the
toxicologic and clinical findings have rarely been undertaken. For many classes of agents,
even relatively simple associations, such as whether the dose-limiting organ toxicity was
similar in animals and humans, have not been well described. Where the findings in
human and toxicologic systems coincide, they reinforce the utility of the toxicologic
models; where differences are observed, they indicate critical needs for further refining
the toxicologic models. While most effort will use existing data, focusing on
pharmacoepidemiologic data, existing data may be supplemented by attempts at filling
critical data needs under this project and by the development of protocols for filling larger
data gaps in other studies. We plan to pursue several aspects of these investigations
through collaborations with federal agencies such as the FDA and NCI or other sources
possessing the required data.
Interspecies comparisons of effects of well-characterized exposures in animal systems and
in humans.
An understanding of the concordance or discordance between well-defined exposure/effect
links in humans with exposure/effect links predicted by experimental systems widely used
for risk assessment should provide vital data for validating and refining approaches used
for risk assessment
3 yr FY89: $80K
R. Everson (OHR/HERL-RTP) (919) 541-1963
*U.S. GOVERNMENT PRINTING OFFICE 1991-548-187/25624
-------
-------
-------
-------
-------
------- |