You are here:
Predictive models of prenatal developmental toxicity from ToxCast high-throughput screening data
SIPES, N. S., M. T. MARTIN, D. M. REIF, N. C. KLEINSTREUER, R. S. JUDSON, A. V. SINGH, K. J. CHANDLER, D. J. DIX, R. J. KAVLOCK, AND T. B. KNUDSEN. Predictive models of prenatal developmental toxicity from ToxCast high-throughput screening data. TOXICOLOGICAL SCIENCES. Oxford University Press, Cary, NC, 124(1):109-127, (2011).
We hope to use predictions from the present model, along with knowledge drawn from the literature, to further HTS and prioritizing the testing of environmental chemicals for developmental toxicity. Additionally we hope to identify specific embryonic organ systems or biological themes (e.g., limb bud, or embryonic vasculature development) for additional predictive models that capture developmental complexities. Focusing on particular embryonic organs or themes will narrow the relevant chemical and biological space, allowing for more detailed analyses. Additionally, in silico agent-based models can be used to integrate and visualize the process of tissue disruption, from genes to cellular processes to tissue level perturbations. Pathways and processes that are linked to particular developmental defects can be studied in these various models, and model outputs can be used for hypothesis generation. These models can also be used to identify missing ToxCast HTS assays, representing important developmental toxicity pathways and processes. These additional data would likely improve the performance of future predictive models of developmental toxicity.
EPA's ToxCast™ project is profiling the in vitro bioactivity of chemicals to assess pathway-level and cell-based signatures that correlate with observed in vivo toxicity. We hypothesized that developmental toxicity in guideline animal studies captured in the ToxRefDB database would correlate with cell-based and cell-free in vitro high-throughput screening (HTS) data to reveal meaningful mechanistic relationships and provide models identifying chemicals with the potential to cause developmental toxicity. To test this hypothesis, we built statistical associations based on HTS and in vivo developmental toxicity data from ToxRefDB. Univariate associations were used to filter HTS assays based on statistical correlation with distinct in vivo endpoints. This revealed 423 total associations with distinctly different patterns for rat (301 associations) and rabbit (122 associations) across multiple HTS assay platforms. From these associations, linear discriminant analysis with cross validation was used to build the models. Species specific models of predicted developmental toxicity revealed strong balanced accuracy (BA > 70%), and unique correlations between assay targets such as transforming growth factor beta (TGFβ), retinoic acid receptor (RAR), and G-protein-coupled receptor (GPCR) signaling in the rat and inflammatory signals, such as interleukins (IL1a and IL8) and chemokines (CCL2) in the rabbit. Species specific toxicity endpoints were associated with one another through common Gene Ontology (GO) biological processes, such as cleft palate to urogenital defects through placenta and embryonic development. This work indicates the utility of HTS assays for developing pathway level models predictive of developmental toxicity. [The views expressed in this article are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency.]