The Toxicologist - Society of Toxicology

726 UPDATED ALUMINUM PHARMACOKINETICS IN
INFANTS FOLLOWING EXPOSURES THROUGH DIET
AND VACCINATION.
R. J. Mitkus, M. O. Walderhaug and M. Hess. U.S. FDA/CBER, Rockville, MD.
Aluminum is a ubiquitous element that is released into the environment via volcanic
activity and the natural breakdown of rocks on the earth’s surface. Exposure of
the general population to aluminum occurs primarily through the consumption of
food, antacids, and buffered analgesics. Exposure to aluminum in the general population
can also occur through vaccination, since vaccines often contain aluminum
salts (frequently aluminum hydroxide or aluminum phosphate) as adjuvants. Since
some lay public have hypothesized that aluminum in vaccines may pose a risk to infants,
we wanted to develop an up-to-date analysis of the safety of aluminum adjuvants.
We therefore conducted a literature search and found that in an effort to
evaluate the relative contribution of childhood vaccines and diet to aluminum levels
in infants, Keith et al. (2002) previously analyzed the pharmacokinetics of aluminum
for infant dietary and vaccine exposures and compared the resulting body
burdens to the minimal risk levels (MRLs) established by the Agency for Toxic
Substances and Disease Registry (ATSDR). We updated the analysis of Keith et al.
(2002) with a current pediatric vaccination schedule recommended by the US
Advisory Committee on Immunization Practices, a more recent aluminum retention
function from human volunteers, and an adjustment for the kinetics of aluminum
efflux at the site of injection. Using these updated pharmacokinetic parameters
we found that the body burden of aluminum from vaccines and diet
throughout an infant’s first year of life is one to three orders of magnitude lower
than the “safe level” calculated using the MRL for dietary aluminum. We conclude
that episodic exposures from vaccines that contain aluminum adjuvant have been
and continue to be extremely low risk to the pediatric population and that this contributes
to the high benefit of continued aluminum adjuvant use in vaccines.
727 ASSESSING THE ROBUSTNESS OF CHEMICAL
PRIORITIZATIONS BASED ON TOXCAST CHEMICAL
PROFILING.
A. Wilson, S. Gangwal, M. Martin, R. Judson, D. Dix and D. Reif. National
Center for Computational Toxicology, Office of Research and Development, U.S. EPA,
Research Triangle Park, NC .
A central goal of the U.S. EPA’s ToxCast program is to provide empirical, scientific
evidence to aid in prioritizing the toxicity testing of thousands of chemicals.
The agency has developed a prioritization approach, the Toxicological Prioritization
Index (ToxPi), that calculates a comprehensive toxicity potential and a relative
priority rank by incorporating information from ToxCast in vitro bioactivity data
(high-throughput screening results from over 500 diverse assays), inferred toxicity
pathways, in vitro to in vivo dosimetry estimates, chemical structural descriptors,
and exposure considerations. Here, we explore the robustness of the prioritization
assessing potential endocrine activity of 309 chemicals in the face of several sources
of variation: 1) changes in the chemical makeup of the experiment, 2) missing data,
and 3) spurious (false-positive) assay results. Bootstrap resampling was used to assess
the effects of alternative chemical sets. Although missing data was not an issue
in Phase I, it may be a concern in subsequent phases and in certain data domains
(e.g. exposure data). To address this concern, we simulated both missing-at-random
and missing-by-domain datasets for comparison with complete data. A similar approach
was taken to assess the potential impact of false-positive assay results.
Generally, the higher-scoring chemicals tended to be less sensitive to alternative
chemical sets but were more sensitive to missing values and false positives than
lower-scoring chemicals. However, initial results for all experiments showed 95%
confidence intervals with mean width representing less than one decile, indicating
that the multivariate endocrine rankings are relatively stable in the face of anticipated
levels of common sources of data variation. This robustness, which is essential
to a reliable prioritization scheme, arises out of the comprehensive nature of the
scores, in that no single datum wields ultimate influence. This abstract does not necessarily
reflect U.S. EPA policy.
728 TOXPLORER: A COMPREHENSIVE
KNOWLEDGEBASE OF TOXICITY PATHWAYS USING
ONTOLOGY-DRIVEN INFORMATION EXTRACTION.
I. Shah 1 , A. Singh 2 , C. Haugh 1 , J. Jack 1 , R. Judson 1 , T. Knudsen 1 , M. Martin 1
and J. Wambaugh 1 . 1 NCCT, U.S. EPA, Research Triangle Park, NC and 2 Lockheed
Martin, Research Triangle Park, NC .
Realizing the potential of pathway-based toxicity testing requires a fresh look at
how we describe phenomena leading to adverse effects in vivo, how we assess them
in vitro and how we extrapolate them in silico across chemicals, doses and species.
We developed the ToxPlorer framework to extract experimental evidence from
the literature about the effects of chemicals in living systems and to coherently synthesize
prior knowledge into semantic networks. This was accomplished in four
main steps. First, we developed an ontology to formally describe functional relationships
in toxicology, which include molecules and their interactions, but also cell
types, cellular processes and behaviors, phenotypes and histological effects. Second,
we systematically analyzed the text of 655,271 PubMed abstracts about the mammalian
liver using 363,472 diverse entities in our ontology. Third, we used this information
to focus on a subset of 23,244 abstracts about nuclear receptor-mediated
hepatocarcinogenesis. Out of 241,944 sentences a subset of 3,712 sentences from
2,199 abstracts produced more than 100,000 putative relationships of relevance.
Fourth, we used ontology-driven information extraction tools to find less than
5,000 semantically-valid assertions about chemical-induced hepatotoxicity. By
manually curating this information we found evidence relating 501 chemicals, 671
genes, 121 cell-events and 38 histological lesions (to date). Our findings recapitulate
many of the events involved in nuclear receptor-mediated direct and indirect
hyperplasia, formation of preneoplastic foci and development of neoplastic lesions
in rodents. ToxPlorer is publicly available along with tools for analyzing and interactively
reconstructing putative toxicity pathways.
This abstract does not necessarily reflect US EPA policy.
729 BELIEF THEORY IN MECHANISTIC
PHARMACOVIGILANCE: USING THE BIOMEDICAL
LITERATURE TO CONFIRM ADVERSE DRUG EVENT
REPORTS FROM ELECTRONIC HEALTH RECORDS.
E. Ahlberg Helgee and S. Boyer. Computational Toxicology, AstraZeneca and EU-
ADR Consortium, Mölndal, Sweden.
Mining of electronic health records (EHRs) for indications of adverse drug events
carries with it a high false positive rate. In order to assess possible biological mechanisms
underlying these positive ‘signals’ the EU-ADR project has embarked on an
extensive program of biomedical literature mining to enhance the interpretation of
adverse event signals from the 35 million European EHRs used in the project. We
have assessed the utility of belief theory to objectively combine the signals arising
from EHRs and from the biomedical literature into a single ‘confirmed signal’.
Diverse sources of evidence such as this can be combined into a single predictor in
a number of ways, such as simple consensus voting methods, weighted consensus
estimates and full probability Bayesian estimates. However, few methods have the
ability to handle both full probability ‘objects’, like Bayesian probability models,
and models in which the underlying probability structure is partly or completely
unknown. The combination of EHR-based signals and the biomedical literature,
however, contains a mixture of full probability estimates and partial/unknown
probabilities. Combining these types of evidence therefore requires a mathematical
structure that manages both types of probability expressions. Dempster-Shafer theory
deals with measures of “belief” as opposed to probability. It also introduces an
explicit formulation for the unknown or uncertain state. We have applied
Dempster-Shafer Belief Theory to a number of known signals from EHR data including
upper GI bleeding and rhabdomyolysis. The results suggest that this approach
can be used to combine signals from EHRs with biomedical text mining in
order to build a robust, biological mechanism-based pharmacovigilance system that
could enhance the ability to simultaneously monitor both EHRs and the literature
to provide earlier warning of adverse events following drug treatment.
730 PREDICTING ADAPTIVE RESPONSE TO FADROZOLE
EXPOSURE: COMPUTATIONAL MODEL OF THE
FATHEAD MINNOWS HYPOTHALAMIC-PITUITARY-
GONADAL AXIS.
M. Breen 1, 2 , D. Villeneuve 1 , G. Ankley 1 , K. Watanabe 3 , M. Breen 1 , A. Lloyd 2
and R. Conolly 1 . 1 U.S. EPA, Research Triangle Park, NC, 2 North Carolina State
University, Raleigh, NC and 3 Oregon Health and Science University, Beaverton, OR.
Exposure to endocrine disrupting chemicals can affect reproduction and development
in both humans and wildlife. We are developing a mechanistic mathematical
model of the hypothalamic-pituitary-gonadal (HPG) axis in female fathead minnows
to predict dose-response and time-course (DRTC) behaviors for endocrine effects
of fadrozole. The model includes two feedback regulatory loops within the
HPG axis that mediate adaptive responses to endocrine stress. One regulatory loop
controls the secretion of luteinizing hormone (LH) and follicle-stimulating hormone
(FSH) secretion from brain, and the other regulates LH and FSH receptor recycling
in ovary. Fathead minnows were exposed to fadrozole at 3 or 30 ug/L for 8
days followed by a 20-day recovery phase, with samples collected for plasma 17βestradiol
(E2) and vitellogenin concentrations during exposure and post-exposure.
Adaptive changes in plasma E2 levels occurred during exposure and overshoot occurred
post-exposure. Comparing the model-predicted DRTC with experimental
SOT 2011 ANNUAL MEETING 157