You are here:
Phenotypic screening of the ToxCast chemical library to classify toxic and therapeutic mechanisms
Kleinstreuer, N., J. Yang, E. Berg, T. Knudsen, A. Richard, M. Martin, D. Reif, R. Judson, M. Polokoff, D. Dix, R. Kavlock, AND K. Houck. Phenotypic screening of the ToxCast chemical library to classify toxic and therapeutic mechanisms. Nature Biotechnology. Elsevier Science, New York, NY, 32(6):583-591, (2014).
Addressing the safety aspects of drugs and environmental chemicals has historically been undertaken through animal testing. However, the quantity of chemicals in need of assessment and the challenges of species extrapolation require the development of alternative approaches. Our approach, the US Environmental Protection Agency's ToxCast program, utilizes a large suite of in vitro and model organism assays to interrogate important chemical libraries and computationally analyze bioactivity profiles. Here we evaluated one component of the ToxCast program, the use of primary human cell systems, by screening for chemicals that disrupt physiologically important pathways. Chemical-response signatures for 87 endpoints covering molecular functions relevant to toxic and therapeutic pathways were generated in eight cell systems for 641 environmental chemicals and 135 reference pharmaceuticals and failed drugs. Computational clustering of the profiling data provided insights into the polypharmacology and potential off-target effects for many chemicals that have limited or no toxicity information. The endpoints measured can be closely linked to in vivo outcomes, such as the upregulation of tissue factor in endothelial cell systems by compounds linked to the risk of thrombosis in vivo. Our results demonstrate that assaying complex biological pathways in primary human cells can identify potential chemical targets, toxicological liabilities and mechanisms useful for elucidating adverse outcome pathways.
Here we have tested 776 chemicals using panels of human primary cells with active signaling pathways relevant to respiratory, skin, immune and vascular exposure sites. This data set included reference chemicals with known modes of action that were correctly predicted by the assays, pharmaceuticals whose intended targets were identified and novel targets hypothesized, and environmental chemicals that were mapped to putative toxicity pathways. It is remarkable that, given the relatively limited diversity of the cell systems used and the nonbiased selection of test chemicals with respect to those cell systems utilized, so many different important mechanisms were uncovered. There is the potential to cover a much broader swath of toxicology by expanding the panel of assays beyond the current limited approach, and to construct models that derive significant associations between the in vitro based predictions and therapeutic or toxicological outcomes in vivo. Continued exploration of these data as a potential guide for human exposure limits will need to include the additional considerations of pharmacokinetics and metabolism as well as the inherent variability of exposure levels between individuals. This study has demonstrated the ability to use alternative methods combined with existing knowledge to classify compounds for potential mechanisms and to broaden our understanding of relevant doses and polypharmacology. This type of complex in vitro assay panel based on human primary cells may serve as a first-tier alternative to animal testing using a human-relevant system for chemical screening and prioritization with applications to toxicity testing and drug discovery.
Record Details:Record Type: DOCUMENT (JOURNAL/PEER REVIEWED JOURNAL)
Organization:U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
NATIONAL CENTER FOR COMPUTATIONAL TOXICOLOGY