Employing Probabilistic Hazard Assessment Approaches to Compare Sensitivity of ToxCast Estrogen Agonist Assays

Environmental and human health implications of endocrine disrupting chemicals (EDCs), particularly xenoestrogens, have received extensive study. Multiple in vitro and in vivo techniques have been developed to assess estrogenicity. However, the relative sensitivity of these approaches and agreement of their conclusions remain inconclusive. As little to no toxicity data exist for a vast number of compounds, the U.S. Environmental Protection Agency (USEPA) created the ToxCast program to prioritize chemicals requiring additional toxicity information with high throughput screening assays. In addition, the Endocrine Disruptor Screening Program plans to use ToxCast data for identifying potential EDCs, and multiple assays pertain to endocrine disruption, particularly estrogenicity. Probabilistic Hazard Assessment (PHA) approaches, such as chemical toxicity distributions, can be used to measure the relative sensitivities of these assays by predicting the likelihood of chemical classes eliciting specific toxicities at or above environmentally relevant concentrations. The PHA in this study examined the comparative sensitivity of 19 assays for estrogen agonists representing a diverse group compounds from USEPA’s ToxCast dataset and noted considerations for assay selection. Overall, these assays had various levels of consistency, and reporter gene assays were among more sensitive designs. ToxCast estrogen agonist assays also had a low probability of detecting compounds below an E2 threshold for the MCF-7 human breast ademocarcinoma proliferation assay, which suggests the relative sensitivity of ToxCast estrogen agonist assays is lower than other in vitro models