You are here:
Differentiating pathway-based toxicity from non-specific effects in high throughput data: A foundation for prioritizing targets for AOP development.
Citation:
Fay, K., J. Swintek, Dan Villeneuve, S. Edwards, M. Nelms, B. Blackwell, AND G. Ankley. Differentiating pathway-based toxicity from non-specific effects in high throughput data: A foundation for prioritizing targets for AOP development. SETAC North America, Minneapolis, MN, November 12 - 16, 2017.
Impact/Purpose:
One of the EPA’s priorities in toxicity testing is to advance the field away from whole animal exposure studies to high-throughput, primarily in vitro, chemical screening. To this purpose, the US EPA’s ToxCast program was established (Dix et al. 2007, Judson et al. 2010, Kavlock et al. 2012). This program has screened over 9000 chemicals for hundreds of biological effects, generating a tremendous amount of data. For the data to be useful for hazard assessment, the in vitro effects must be related to outcomes at the organism or population levels. Adverse Outcome Pathways (AOPs) describe a sequential chain of causally-linked events at different levels of biological organization, ultimately resulting in adverse health or an ecotoxicological effect. As such, AOPs offer a mechanistic means to connect biological effects measured in vitro to outcomes relevant for chemical risk assessment. The AOP discovery and development task (CSS 17.01) recognizes the need to develop AOPs for ToxCast assay targets. The present work utilizes the statistical evaluation of ToxCast data conducted by Judson et al. (2016), to prioritize those assays providing pathway-specific effects data for AOP development. Judson et al. statistically defined concentration thresholds for ToxCast chemicals above which they elicit general cytotoxic or effects. This meta-analysis identifies the ToxCast assays frequently responding to chemicals at concentrations well below their ‘cytotoxic burst’ concentrations, ostensibly providing more pathway-specific effects information. This analysis offers a means to prioritize ToxCast biological targets for AOP development, which builds on FY16 products and supports the FY18 product: application of AOPs to complex mixture assessments.
Description:
The Environmental Protection Agency has implemented a high throughput screening program, ToxCast, to quickly evaluate large numbers of chemicals for their effects on hundreds of different biological targets. To understand how these measurements relate to adverse effects in an organism, scientists develop adverse outcome pathways (AOPs). An AOP organizes existing data from the fields of toxicology, physiology and biology to describe how the chemical perturbation of a protein or similar molecular target can trigger effects at the cellular, organ and organismal levels. Because AOP creation is time-consuming, there is a need to prioritize the ToxCast assays providing the most useful chemical-biologic interaction information for AOP development. This study evaluates the quality of the data being generated by the ToxCast assays and distinguishes those assays providing specific information on changes to important biological pathways from others measuring general cytotoxicity. The ToxCast chemical screening approach enables the rapid assessment of large numbers of chemicals for biological effects, primarily at the molecular level. Because adverse outcome pathways (AOPs) connect biomolecular effects with potential adverse outcomes at the level of the individual or population, they offer a means to support chemical hazard assessment using ToxCast data. Arguably, AOP development is most critical in relation to specifically-acting toxicities for which effective structure-based prediction models are currently lacking. Previous work identified a ‘cytotoxic burst’ phenomenon wherein large numbers of the ToxCast assays begin to respond at or near concentrations that elicit cytotoxicity. The concentration range over which the burst occurs is statistically definable. We conducted a meta-analysis to identify which assays were frequently responding at concentrations well below the cytotoxic burst and are, ostensibly, providing the most pathway-specific effect information. Assays were ranked by the fraction of chemical hits below the burst concentration range compared to the number of chemicals tested, as well as by diagnostic-odds ratio, resulting in a ranking of potentially important, target-specific ToxCast assays. The assay prioritization based on the cytotoxic burst definition indicated numerous assays with targets previously identified for AOP development (e.g., thyroid peroxidase, peroxisome proliferator-activated receptor γ, estrogen receptor α, aromatase) along with several novel targets (e.g., progesterone receptor, monoamine oxidases A and B, norepinephrine transporter). Additional analyses identified chemicals that elicited assay activity only at concentrations within the cytotoxic burst region suggesting a non-specific mode of action. These chemicals were anticipated to be classified as likely to elicit toxicity largely through narcosis, which was largely supported when the chemicals were cross-referenced with other Mode of Action classification/prediction tools (e.g., Assessment Tools for the Evaluation of Risk; ASTER). These analyses help to differentiate which pathway-specific assay targets to prioritize for AOP development. They also provide insight into the utility of tools which integrate chemical occurrence data with the high-throughput screening data, such as the EPA’s Exposure Activity Ratio Calculator. The content of this presentation neither constitute nor necessarily reflect US EPA policy.