You are here:
Modeling Reproductive Toxicity for Chemical Prioritization into an Integrated Testing Strategy
Citation:
MARTIN, M. T., T. B. KNUDSEN, K. A. HOUCK, R. JUDSON, R. J. KAVLOCK, AND D. J. DIX. Modeling Reproductive Toxicity for Chemical Prioritization into an Integrated Testing Strategy. Presented at Annual Meeting of the Society of Toxicology, Washington, DC, March 06 - 10, 2011.
Impact/Purpose:
The ability of this model to broadly identify reproductive toxicants, in conjunction with other predictive models spanning endocrine and developmental toxicity, will inform on chemical testing prioritization and will lead to more efficient tiered and targeted testing strategies.
Description:
The EPA ToxCast research program uses a high-throughput screening (HTS) approach for predicting the toxicity of large numbers of chemicals. Phase-I tested 309 well-characterized chemicals in over 500 assays of different molecular targets, cellular responses and cell-states. Of the 309 chemicals in Phase I of ToxCast, 247 have been tested in a multigeneration reproductive toxicity study and are stored in the relational database, ToxRefDB. A total of 94 chemicals were identified as reproductive toxicants, based on having achieved a reproductive lowest observed adverse effect level (rLOAEL) and having observed, at minimum, a second generation reproductive effect. Each assay was assessed for its univariate association with the identified reproductive toxicants. Significantly associated assays were used for the subsequent predictive modeling. Using linear discriminant analysis (LDA) and ten-fold cross-validation, a model was produced capable of identifying reproductive toxicants with 78% accuracy (74% and 72% average training and test set balanced accuracy, respectively). Cell-free binding and cell-based transcription factor activation assays targeting steroidal and non-steroidal nuclear receptors, including estrogen, androgen, and peroxisome proliferator-activated receptors, alpha and gamma, were the major components of the model. Cytochrome P450 inhibition assays, including aromatase, and cell-based markers of growth factor stimulation were other positive indicators of reproductive toxicity potential. Assays targeting the pregnane X receptor (PXR) were the only negative indicators in the model and may provide evidence of rapid chemical metabolism and thus a lower likelihood of reproductive toxicity. The ability of this model to broadly identify reproductive toxicants, in conjunction with other predictive models spanning endocrine and developmental toxicity, will inform on chemical testing prioritization and will lead to more efficient tiered and targeted testing strategies. (This abstract does not necessarily reflect U.S. EPA policy.)