Identifying Metabolically Active Chemicals Using a Consensus Quantitative Structure Activity Model for Estrogen Receptor Binding
Citation:
Leonard, J., C. Stevens, K. Mansouri, D. Chang, AND C. Tan. Identifying Metabolically Active Chemicals Using a Consensus Quantitative Structure Activity Model for Estrogen Receptor Binding. SOT Annual Meeting, Baltimore, MD, March 12 - 16, 2017.
Impact/Purpose:
The study summarized in this abstract can be used to increase confidence in in vitro screening results, through identification of potential parents that are inactive under in vitro conditions but that might become metabolically active in a whole organism.
Description:
Endocrine disrupting chemicals (EDCs) are abundant throughout the environment and can alter neurodevelopment, behavior, and reproductive success of humans and other species by perturbing signaling pathways related to the estrogen receptor (ER). A recent study compared results across 18 ER-related assays in the ToxCast™ in vitro screening program to predict the likelihood of a chemical exhibiting in vivo estrogenic activity, with the purpose of eliminating chemicals that may produce a false signal by interfering with the technological attributes of an individual assay. However, flaws in in vitro assay design can also prevent induction of signal activity by EDCs. Another reason for not observing activity for some EDCs in in vitro assays is that metabolic activation is required to perturb ER-related pathways. In the current study, 1,024 chemicals were identified as lacking ER activity after establishing a consensus across each of the 18 ER-related in vitro assays, and nearly 2,000 primary and 3,700 secondary unique metabolites were predicted for these chemicals. The ER binding activity for each metabolite was then predicted using an existing ER activity quantitative structure activity relationship (QSAR) consensus model. Binding activity was predicted for 2-3% of the metabolites within each generation. Of the inactive parent compounds generating at least one metabolite predicted to have ER-binding activity, nearly 30% were found to have metabolites from both generations with such activity. About 10-25% of these inactive compounds with active metabolites (e.g., azobenzene and p-cresol) are considered to be estrogenic in vivo, thus identifying their active metabolites illuminates their safety concerns. The approaches presented here can be used to increase confidence in in vitro screening results, through identification of potential parents that are inactive under in vitro conditions but that might become metabolically active in a whole organism. Disclaimer: This abstract has been cleared by the EPA but solely expresses the view of the authors.
