Overview of Sourceto Effects Modeling for Susceptible Children Populations

1
Overview of Source-to Effects Modeling for
Susceptible (Children) Populations Halûk
Özkaynak1, Hugh Barton2, Bob Sonawane3 U.S.
Environmental Protection Agency, Office of
Research and Development, 1National Exposure
Research Laboratory 2National Health and
Environmental Effects Research Laboratory
3National Center for Environmental Assessment
Methods/Approach
Results/Conclusions
Figure 1. Elements of Source-to-Effects Modeling
Science Question
The EPAs Office of Research and Development
(ORD) has been developing and applying
source-to-effect models to estimate childrens
exposures to environmental pollutants and
formulating appropriate physiologically based
pharmacokinetic (PBPK) models for evaluating
risks to children at various early life-stages.
Figure 1 shows the main elements of ORDs
source-to effects modeling framework that is
being applied to the study of childrens risks
resulting from exposures to environmental
chemicals. Modeling Exposures and Dose In the
area of developing models to estimate exposures
of young children to pollutants, ORD has been
developing a state-of-the-art aggregate exposure
and dose model, the Stochastic Human Exposure and
Dose Simulation (SHEDS) model (see Figure 2 for a
model flow diagram for SHEDS-Air Toxics) to
examine childrens exposure to a variety of
inhalation and multi-media multi-pathway
pollutants, such as air toxics, a number of the
organophosphate pesticides, and arsenic from
contact with CCA treated wood (e.g., Zartarian et
al. 2003). The SHEDS model is designed to
interface with more sophisticated
source-to-concentration (e.g., the indoor
fugacity model for pesticides) and
exposure-to-dose (e.g., ERDEM or the Exposure
Related Dose Estimation Model) models (see also
Agg/Cum Risk topic). In the development,
evaluation and dissemination of the exposure and
dose models, ORD has been collaborating with
various academic partners (e.g. Rutgers and
Harvard University) and with several private and
public organizations, for example, the
International life Sciences Institute (ILSI), the
Cumulative and Aggregate Risk Evaluation System
(CARES) model group, World Health Organizations
International Program on Chemical Safety
(WHO/IPCS). Formulating PBPK Models for
Children In formulating appropriate PBPK models
for children, ORD has been evaluating both human
and animal data sets that are relevant to
chemicals and exposure scenarios of concern for
children. Two major approaches are being used for
evaluating potential risks from early life
exposures, each of which has strengths and
limitations (Barton 2004). The first approach is
to extrapolate from adults to children based upon
quantifiable differences in exposure,
pharmacokinetics, and pharmacodynamics leading to
potential pharmacological activity or toxicity
(Ginsberg et al. 2004b). The second approach is
to extrapolate across species from animal
toxicity studies that involve exposures at early
life stages. Such studies include developmental
toxicity (in utero exposure), one- or
two-generation reproductive and developmental
toxicity, and neurotoxicity studies. Developing
PK Parameters for Children ORD researchers have
been conducting research using the first approach
for evaluating and incorporating child/adult
toxicokinetic differences in assessing risk to
several environmental toxicants. Specific
research questions addressed under this research
are 1) Are there differences in toxicokinetics
(TK) of xenobiotics between children and adults
due to physiological changes and the immaturity
of enzyme systems and clearance mechanisms? 2)
How these differences can be incorporated into
PBTK models that simulate fate of environmental
toxicants in both children and adults? While
there are very little PBTK data for environmental
agents in children, there are much more abundant
data for therapeutic drugs used in pediatric
practice. Using published literature, a
childrens PK database has been compiled by
ORD-sponsored research that compares PK
parameters between children and adults for 45
pharmaceutical agents. Extrapolation from
Animal Toxicity Studies Involving Early Lifestage
Exposures The second approach evaluates risks at
early ages by conducting cross-species
extrapolation from animal studies that involve
exposures at early life-stages (Figure 3).
However, a major challenge for modeling these
life stages is how to obtain datasets for model
parameterization (i.e. choices of physiological
and chemical-specific parameters), calibration
(i.e. assignment of values to the parameters),
and testing (i.e. using an independent dataset to
evaluate the success of the model beyond the
original data). Pregnancy and very limited
lactational models have been developed
previously. Therefore, postnatal development has
been the major focus of recent ORD research.
Physiological parameters for growing rats have
been compiled in an electronic database this has
been contributed to an ORD sponsored and
ILSI-organized collaborative effort among
government Agencies, academics, and industry to
evaluate the state of knowledge for physiological
parameters for modeling growing mice, rats, and
humans. Chemical-specific parameters can be
estimated using in vitro or in vivo studies. In
collaboration with ORD and the Air Force Research
Laboratory, the age-dependency of partition
coefficients for volatile organic compounds with
a range of physicochemical properties are being
determined in rat and human blood and tissues.
ORD has recently prepared a draft document
entitled Framework for Childrens Health Risk
Assessment (see Childrens Risk Assessment
topic). This internal document is now being
revised following an internal review in October
2004. Soon to be sent for external peer-review,
this document describes ORDs approach for
assessing childrens risk from exposures to
environmental pollutants. ORDs source-to-effects
modeling research, which supports this effort,
has identified key age-dependent factors that
influence exposures and toxicokinetics and
toxicodynamics of select classes or types of
chemicals that children may come in contact with
during different periods of early development.
The SHEDS exposure and dose model was applied to
the organophosphate (chlorpyrifos) and wood
preservative (arsenic) case studies of regulatory
interest to EPA (Zartarian et al 2003). These
applications demonstrated the important
contributions of hand-to mouth and
hand-to-object-to-mouth exposure pathways for
infant and young childrens exposures to
multimedia pollutants. Likewise, the ERDEM model
has also been applied to chlorpyrifos and to
malathion and selected volatile organic compounds
(e.g. trichloroethylene) in air and water (see
Agg/Cum Risk topic). ORD also developed a case
study for malathion PBPK model for children. The
childrens PK data base complied by ORD has
enabled comparison of child and adult PK function
across a number of cytochrome P450 (CYP)
pathways, as well as certain Phase II conjugation
reactions and renal elimination (Hattis et al.
2003). The database provides an opportunity for
calibrating and validating PBTK models for
various chemicals, using examples as theophylline
and caffeine (Ginsberg et al. 2004b). These drugs
are particularly useful case studies because the
clearance mechanism in neonates and infants is
considerably different (slower) than adults.
The physiological parameters database for
growing rats has already been compiled in an
electronic form. The chemical-specific tissue
partitioning data, which has been assembled by
ORD and the Air Force Laboratory, demonstrate
only minor differences across age groups. By
contrast, modeling of serum protein binding for
estrogenic compounds demonstrates substantial
differences across ages and species due to high
affinity binding to alpha-fetoprotein (rats) and
serum hormone binding globulin (humans), while
lower affinity albumin binding is age, but not
species, dependent. Thus, determinants of tissue
distribution must be carefully considered in the
analysis of pharmacokinetic and pharmacodynamic
linkages and potency of a compound by lifestage.
Conazole fungicides are now the subject of an
integrated toxicokinetic and toxicodynamic
research effort in ORD (see Harmonization topic).
Assessment of chemical-specific risk for children
requires several types of information on
exposures and effects occurring during early life
stages. In particular, chemical risk assessments
now rely upon an analytical framework, which
includes characterization of exposure,
pharmacokinetics (or toxicokinetics) and
pharmacodynamics (or toxicodynamics) and response
(cf. Barton 2004, Ginsberg et al. 2004a, Barton
et al. 1998). In order to evaluate the basis of
differential risk to children from environmental
stressors, ORD researchers are implementing an
interdisciplinary research program that attempts
to address all the four components of this
framework. The critical science questions that
were identified by ORD under the
source-to-effects modeling for susceptible
populations, including children, were 1) How
can we improve the linkage between emissions,
exposure and dose estimation for predicting
target tissue response or biologically effective
dose response for children? 2) What are the best
approaches for developing physiologically-based
pharmacokinetics models and evaluating potential
health risks to children, either using human
studies data, or by extrapolating across species
using animal toxicity data that involve exposures
at early life stages?
Figure 2. SHEDS-Air Toxics Model Flow
Impact and Outcomes
ORD had significant success in generating basic
information and models that would support study
of childrens exposures to a number of chemicals
of regulatory concern. For example, the
SHEDS-Wood model developed by ORD was used by the
Office of Pesticides Program in their regulatory
evaluation of exposures and risks to children
from contacting CCA treated wood on playsets and
decks. The research in the area of toxicokinetic
modeling for children is ongoing in ORD with
related efforts at other institutions (e.g., CIIT
Centers for Health Sciences, ILSI, Clark
University, Connecticut Department of Health).
ORD scientists have presented and organized
workshops at professional society meetings (e.g.,
2005 Society of Toxicology Annual Meeting) to
promote generation of needed information and
methods by the toxicology community to support
early life pharmacokinetic modeling. More
reliable source-to-effects models will allow the
Agency to replace or revise the default
uncertainty factors typically used in its current
risk assessments for children and to make
informative risk management decisions.
Research Goals
Future Directions

• In order to understand potential health risks to
  children from exposures to environmental
  chemicals of concern, it is necessary to
  investigate quantitative relationships between
  exposure, absorbed dose and the biologically
  effective dose. The main goals of the ORDs
  current source-to-effects modeling research
  program for children under the susceptible
  subpopulations area, are to
• identify and quantify routes and pathways of
  exposures of concern for children,
• develop new tools that can describe the
  biological basis for differential sensitivity in
  children, and
• develop and apply scientifically robust human
  exposure and risk analysis framework that
  incorporates models, databases and analytical
  tools in estimating exposure, dose and health
  risks to children.
• This information is being provided to risk
  assessors and policy-makers both within the
  Agency and elsewhere, so that appropriate source
  and exposure mitigation strategies can be
  considered, in order to reduce risks to children
  exposed to environmental stressors.

ORD is in the process of extending the exposure
and dose-response modeling tools to study
cumulative risks resulting from exposures to
mixtures or multiple chemicals, such as the
pyrethroid or nmethyl carbamate pesticides.
Using animal or adult human data for
extrapolation to different life stages have been
informative and helpful for constructing
chemical-specific PBTK models. However,
applicability of these models to other chemicals
of interest remains to be determined through
future research. Additional research is being
performed to determine whether children may be
more susceptible than adults due to toxicokinetic
or due to toxicodynamic differences, or both.
On-going and future research planned by the ORD
under the Computational Toxicology Program and in
the exposure and PBPK modeling areas should
improve cross-species extrapolation of effects
and the prediction of health risks to children,
resulting from exposures to chemicals of concern.
Figure 3. Life Stage and Species Extrapolations
References
Barton, H.A. Computational Pharmacokinetics
during Developmental Windows of Susceptibility. J
Toxicol Environ A (in press). Barton, H. A.,
Andersen, M. E., and Clewell, H. J., III 1998.
Harmonization developing consistent guidelines
for applying mode of action and dosimetry
information to cancer and noncancer risk
assessment. Human EcolRisk Assess
475-115. Ginsberg, G., Slikker, W., Bruckner J.
and Sonawane, B. 2004a. Incorporating childrens
toxicokinetics into a risk framework.
Environmental Health Perspect. 112272-283. Ginsbe
rg, G., Hattis, D., Russ, A., and Sonawane, B.
2004b. Physiologically based pharmacokinetic
(PBPK) modeling of caffeine and theophylline in
neonates and adults implications for assessing
children's risks from environmental agents. J
Toxicol Environ Health A 67297-329. Hattis, D.,
Ginsberg, G., Sonawane, B., Smolenski, S., Russ,
A., Kozlak, M., and Goble, R. 2003. Differences
in pharmacokinetics between children and adults.
II. Childrens variability in drug elimination
half-lives and in some parameters needed for
physiologically-based pharmacokinetic modeling.
Risk Analysis 23117-142. Zartarian,V.G., Xue J.,
Özkaynak H., Dang W., Glen G., Smith L.,
Stallings C. 2003. Probabilistic Exposure
Assessment for Children Who Contact CCA-Treated
Playsets and Decks Using the Stochastic Human
Exposure and Dose Simulation Model for the Wood
Preservative Exposure Scenario (SHEDS-Wood).
Draft Preliminary Report. Reviewed at the EPA
Office of Pesticide Programs FIFRA (Federal
Insecticide, Fungicide, Rodenticide Act) Science
Advisory Panel (SAP) meeting, December 3-5, 2003.
EPA/600/X-04/089