Overview of Use of Mechanistic Data in Risk Assessment

1
Overview of Use of Mechanistic Data in Risk
Assessment Hugh A. Barton1, Jeff Gift2, Resha
Putzrath2, R. Woodrow Setzer1 and Chris
Saint3 U.S. Environmental Protection Agency,
Office of Research and Development, 1National
Health and Environmental Effects
Laboratory, 2National Center for Environmental
Assessment, 3National Center for Environmental
Research
Results/Conclusions
Methods/Approach

• A set of guiding principles for the use of
  mechanistic information describing the toxicity
  pathway from exposure through dosimetry and
  toxicodynamic processes to response has been
  articulated that is applicable to many toxic
  effects.
• Principles for evaluating the human relevance of
  animal mode of action information have been
  articulated, as well as issues for application to
  other children.
• Case studies have been described for several
  modes of action and chemicals acting through
  them. These provide examples of how to evaluate
  mode of action issues, implement mode of action
  analyses in risk assessments, and address risk
  assessments for specific chemicals.

Science Question
The EPAs Office of Research and Development
(ORD) has been developing and applying general
guidance on the use of mechanistic data in risk
assessment (Figure 1), as well as developing
scientific data on specific modes of action for
use in evaluation of risk associated with
specific chemicals. The need is not for the
development of a single methodology for all
situations, but rather it is the need for
consistent application of the pertinent
information on toxicity, dosimetry, mode of
action and exposure in all risk assessments
regardless of adverse health outcome and chemical
pollutant. The research is designed to develop a
consistent set of principles and guidelines for
drawing inferences from scientific information by
better understanding the key events in the
toxicity pathways from source to response for a
range of adverse health outcomes and describing
how they will be addressed in risk assessment
applications (Figure 2). Risk Assessment
Guidance EPA has a substantial record of
developing guidance applicable to risk assessment
(Figure 1). Two major ongoing efforts in this
area are highlighted in this session guidelines
for cancer risk assessment and guidance and
software for application of dose response
modeling (i.e., Benchmark Dose). The draft
Revised Guidelines for Carcinogen Risk Assessment
contain the most extensive discussion of
mechanistic data and its role in risk assessment
to date. Although these guidelines directly
address cancer risk assessment due to the
historical priority given this endpoint, the
framework they contain for evaluating mode of
action data and its relevance to humans,
including children, is generally applicable
regardless of endpoint. Statistical analysis of
dose-response relationships is a key tool
applicable to any toxicity endpoint. The
Benchmark Dose software and guidance represent a
significant effort to implement a range of
different mathematical models applicable to
endpoints with different characteristics (e.g.,
reproductive or developmental studies in which
litter effects need to be evaluated, adult
subchronic and chronic animal bioassays).
Dose-response analysis can be conducted using
exposure doses/concentrations or using internal
dose metrics obtained by pharmacokinetic modeling
and applicable to the mode of action for the
endpoint being analyzed. Risk Assessment Tools As
risk assessment moves towards more comprehensive
descriptions of the processes and events leading
from sources through to exposures, internal
doses, and health outcomes (Figure 3), additional
tools are desirable. EPA has collaborated in
research with academic institutions including
Clark University and the University of Washington
to develop new approaches and tools. Human
variability affects virtually every step in the
source to effects continuum. While it is one of
the major elements addressed under the topic of
susceptible populations, better descriptions of
human variability are essential for more
completely characterizing risks to human
populations in general (Hattis et al., 1999a,b).
Modeling the toxicodynamic processes leading
from the biologically effective dose to the
adverse outcome or disease is a particularly
challenging area for the development of new tools
(Figure 3). While this is an area that is
beginning to explode with the advent of new
omics technologies to generate broad
descriptions of perturbations of biological
systems, the biological understanding of how
systems level behaviors create different
dose-response behaviors for different modes of
action is still in its early stages. Clonal
growth modeling has historically been applied to
cancer analysis, but more recently there has been
interest in its potential applications for other
endpoints such as neurodevelopmental toxicities
arising from changes in cell growth,
differention, and death (Gohlke et al., 2002,
2004). Mode of Action Case Studies Evaluations of
mode of action and approaches for its use in risk
assessment have been ongoing in a variety of
contexts. Collaborative efforts among Agency
scientists and colleagues from other government
departments, academia, and industry from the
United States and abroad have been ongoing
through organizations such as the International
Life Sciences Institute (Cohen et al., 2003 Meek
et al., 2003 Klaunig et al., 2003 Slikker et
al., 2004a,b) and under the auspices of the
Society of Toxicology (Bogdanffy et al., 2001)
all with funding, in part, from the U.S. EPA.
These efforts typically involve case studies for
chemicals causing toxicities through a variety of
modes of action and efforts to develop general
principals from those case studies. This case
study approach is also a major component of the
Agencys approach. The perchlorate risk
assessment described in this session serves as an
example of the application of the expertise of
Agency scientists to a state-of-the-art analysis
for a chemical on which ORD is not actively
engaged in research. The several other topical
areas also described in this session represent
examples where ORD is actively involved in
laboratory or epidemiological research to address
modes of action (e.g., Ah receptor agonism,
leuteinizing hormone disruption, oxidative
stress, modulation of metabolism by cytochromes
P450) including combinations of modes of action
(e.g., the variety of activities associated with
arsenic).

• The critical science questions identified by ORD
  in the area of using mechanistic data in risk
  assessment were
• What mode/mechanisms of action (MOA) are
  important for understanding the impact of
  environmental stressors on human health?
• What are the attributes (e.g., shape of the
  dose- response, species specificity) of the MOA
  that impact risk assessment?
• How do we measure, model, and/or predict the key
  attributes of the MOA that could impact risk
  assessment?
• How do we incorporate mechanistic tools into
  risk assessment?
• Assessments of health risk from exposures to
  environmental agents have been performed quite
  differently depending upon whether the adverse
  health outcome is cancer or a noncancer (e.g.,
  neurological, reproductive, developmental) one.
  To a great extent this practice has been based
  upon a limited understanding of the modes of
  action of toxic substances involved in the
  production of the different classes of adverse
  health outcomes. The considerable increase in
  knowledge of the key events involved in the
  induction of different disease types is changing
  this knowledge base and the implications of this
  are increasingly being implemented in overall
  risk assessment guidance, tools (i.e., methods,
  models, and measurements), and analyses for
  specific chemicals.

Impact and Outcomes
The roles of tissue dosimetry and mode of action
in risk assessment, particularly dose-response
assessment, are increasingly prominent in EPA
guidance documents and risk assessments for
chemicals, reflecting the level of importance of
research activities in ORD. Both the draft
cancer guidelines and supplemental guidance for
early-life exposures are organized around the
concept of mode of action. Implementation of
benchmark dose analysis is an important tool for
analyzing dose-response relationships, whether
based upon exposures or internal doses predicted
by pharmacokinetic modeling. Risk analyses for
specific chemicals including atrazine,
chloroform, perchlorate, and butoxyethanol have
been modified substantially through the use of
mode of action data
Future Directions

• Development of case studies addressing additional
  modes of action is ongoing. Examples describe in
  this session include modulation of xenobiotic
  metabolizing enzymes and oxidative stress. These
  case studies require collection of laboratory
  data to characterize the mode of action following
  which implications for risk assessment can be
  evaluated.
• Develop guidance for evaluation and
  implementation of pharmacokinetic modeling in
  risk assessment to articulate the lessons learned
  from case examples that have used this approach.
• Develop and evaluate approaches to probabilistic
  analyses of risks for a range of noncancer
  endpoints utilizing information describing human
  variability.

References
Bogdanffy MS, Daston G, Faustman EM, Kimmel CA,
Kimmel GL, Seed J, Vu V. Harmonization of cancer
and noncancer risk assessment proceedings of a
consensus-building workshop. Toxicol Sci. 2001
May61(1)18-31 Cohen SM, Meek ME, Klaunig JE,
Patton DE, Fenner-Crisp PA. The human relevance
of information on carcinogenic modes of action
overview. Crit Rev Toxicol. 200333(6)581-9
Gohlke JM, Griffith WC, Faustman EM. The role of
cell death during neocortical neurogenesis and
synaptogenesis implications from a computational
model for the rat and mouse. Brain Res Dev Brain
Res. 2004 Jul 19151(1-2)43-54. Gohlke JM,
Griffith WC, Bartell SM, Lewandowski TA, Faustman
EM. A computational model for neocortical
neuronogenesis predicts ethanol-induced
neocortical neuron number deficits. Dev Neurosci.
200224(6)467-77. Hattis D, Banati P, Goble R,
Burmaster DE. Human interindividual variability
in parameters related to health risks. Risk Anal.
1999 Aug19(4)711-26. Hattis D, Banati P, Goble
R. Distributions of individual susceptibility
among humans for toxic effects. How much
protection does the traditional tenfold factor
provide for what fraction of which kinds of
chemicals and effects? Ann N Y Acad Sci.
1999895286-316. Klaunig JE, Babich MA,
Baetcke KP, Cook JC, Corton JC, David RM, DeLuca
JG, Lai DY, McKee RH, Peters JM, Roberts RA,
Fenner-Crisp PA. PPARalpha agonist-induced rodent
tumors modes of action and human relevance. Crit
Rev Toxicol. 200333(6)655-780 Meek ME, Bucher
JR, Cohen SM, Dellarco V, Hill RN,
Lehman-McKeeman LD, Longfellow DG, Pastoor T,
Seed J, Patton DE. A framework for human
relevance analysis of information on carcinogenic
modes of action. Crit Rev Toxicol.
200333(6)591-653 Slikker W Jr, Andersen ME,
Bogdanffy MS, Bus JS, Cohen SD, Conolly RB, David
RM, Doerrer NG, Dorman DC, Gaylor DW, Hattis D,
Rogers JM, Setzer RW, Swenberg JA, Wallace K.
Dose-dependent transitions in mechanisms of
toxicity case studies. Toxicol Appl Pharmacol.
2004 Dec 15201(3)226-94. Slikker W Jr,
Andersen ME, Bogdanffy MS, Bus JS, Cohen SD,
Conolly RB, David RM, Doerrer NG, Dorman DC,
Gaylor DW, Hattis D, Rogers JM, Woodrow Setzer R,
Swenberg JA, Wallace K. Dose-dependent
transitions in mechanisms of toxicity. Toxicol
Appl Pharmacol. 2004 Dec 15201(3)203-25
Research Goals

• There are three major research goals for using
  mechanistic data in risk assessment.
• The first is to develop generalized guidance for
  Agency risk assessment activities on the use of
  qualitative and quantitative approaches that
  facilitate the use of mode of action data in
  risk assessments.
• The second goal is to develop new tools for
  incorporating mechanistic information into risk
  assessment.
• The third goal is to develop sufficient
  information for modes of action of importance to
  the Agency that specific approaches can be
  developed to use the information in risk
  assessment.
• These latter case studies for specific modes of
  action and specific chemicals ultimately feedback
  to the generalized guidance by helping inform
  whether that guidance is adequate to address the
  specific modes of action or needs modification.