S4: Pathway-based models for rapid screening of endocrine disruptors

Pathway-based models for rapid screening of endocrine disruptors

S3: Pathway-based models for rapid screening of endocrine disruptors, OpenTox USA 2017

Nicole Kleinstreuer




Deputy Director


Nuclear hormone receptors are critical components of the endocrine system and are implicated in the pathogenesis of developmental and reproductive impairment, obesity, diabetes, and certain cancers, with mounting evidence for their role in the developmental origins of health and disease. People are potentially exposed to thousands of environmental chemicals, and some of these are known or suspected to interact with the endocrine system, in particular the estrogen and androgen pathways via their corresponding nuclear hormone receptors. Therefore, screening for endocrine disrupting chemicals (EDCs) is both a scientific and a regulatory priority. Given the complexity of the in vivo tests currently used to identify and fully characterize EDCs, only a small fraction of the chemical exposure universe has been studied in any depth. This has led to the development of in vitro and in silico approaches, including the federal Tox21 and ToxCast high throughput screening (HTS) programs, allowing rapid testing of thousands of chemicals and identification of potential EDCs. Under these programs, diverse in vitro HTS assays measure perturbations of the estrogen receptor (ER) and androgen receptor (AR) pathways at multiple points (e.g. receptor binding, co-regulator recruitment, gene transcription and protein production) in multiple cell types using several different analytical technologies (e.g. fluorescence, luminescence, electrical impedance). A certain number of chemicals could be expected to act as true agonists or antagonists in these pathways, but there are also chemicals that are known to interfere with the various assay technologies through false signals such as auto-fluorescence or cytostatic mechanisms. Here, we leveraged orthogonal assays that help distinguish non-specific activity from interaction with the intended target, and built computational network models that integrate the data and predict whether chemicals are true agonists, antagonists, or are acting via interference pathways. These computational toxicology approaches were then validated using a performance based approach against large sets of reference chemicals spanning a wide range of potencies. The validated ER pathway model was accepted by the U.S. Environmental Protection Agency (EPA) for ER bioactivity screening as an alternative to multiple lower-tier approaches, including the rodent uterotrophic bioassay. Recent work has shown excellent performance of the AR pathway model against a group of well-characterized reference chemicals, and these results are also being considered as a replacement for existing test methods. Approximately 1850 chemicals to date have been screened and run through the ER and AR pathway models, and a subset of environmental chemicals including pesticides, plasticizers, flame retardants, and personal care product ingredients have been identified for their endocrine disrupting potential. Further, experimental reverse toxicokinetic measurements are being used to parameterize models for in vitro to in vivo extrapolation (IVIVE) to assess assay performance against reference chemicals with demonstrated adversity in vivo. These efforts are providing reliable pathway-based predictions for thousands of chemicals and transforming endocrine disruption screening.