In Vitro Screening of Environmental Chemicals for Targeted Testing Prioritization: The ToxCast Project

TLDR: A statistically significant inverse association between the number of pathways perturbed by a chemical at low in vitro concentrations and the lowest in vivo dose at which a chemical causes toxicity is found.

Abstract: There are thousands of environmental chemicals, including many industrial chemicals and pesticidal active and inert ingredients, with the potential for significant human exposures but for which toxicity information is either limited or nonexistent (Judson et al. 2009). This data gap is due largely to the high cost and length of time required to conduct animal testing in rodents and other species. A complete set of regulatory tests for a single chemical (including those for carcinogenicity and for chronic, reproductive, and development toxicity) uses thousands of animals and costs millions of dollars. In addition, traditional animal tests often yield limited information on mechanism of action, and hence on the cellular pathways that could lead to toxicity in humans. Such mechanistic information is key to moving beyond default approaches for extrapolating from high-dose animal toxicity tests to estimation of human risk at realistic exposure levels. There is a pressing need to screen the large backlog of chemicals for their potential toxicity and, ultimately, their contribution to human diseases. The National Research Council (2007) advocated the use of mechanistically informative in vitro assays based on human cells or human cell constituents that measure effects on “toxicity pathways” leading to human disease. The U.S. Environmental Protection Agency (EPA), through its ToxCast program (Dix et al. 2007) and the Tox21 collaboration with the National Toxicology Program and the National Institutes of Health Chemical Genomics Center, is pursuing similar objectives and applying many of the ideas represented in the National Research Council report (Collins et al. 2008; Kavlock et al. 2009). ToxCast is a large-scale experiment using a battery of in vitro, high-throughput screening (HTS) assays, applied to a relatively large and diverse chemical space, to develop methods to predict potential toxicity of environmental chemicals at a fraction of the cost of full-scale animal testing. Three major goals of ToxCast are to a) identify in vitro assays that can reliably indicate alterations in biological processes of relevance to in vivo toxicity; b) develop signatures or prediction models based on multiple assays, along with computed or available chemical properties, that can achieve higher predictive power than single assays or chemical structure alone; and c) use these combined in silico and in vitro assay-based signatures to screen large numbers of previously untested environmental chemicals. The ToxCast data set provides a rich resource for identifying chemically induced changes in biological pathways that are associated with in vivo end points and that could potentially lead to human disease. Chemicals whose properties and assay profiles match these predictive signatures can be prioritized for more in-depth testing, which may include nontraditional, mechanism-focused in vivo tests. In this article, we provide an overview of the entire ToxCast phase I assay results data set and present initial analyses and findings.