Recommended approaches in the application of toxicogenomics to derive points of departure for chemical risk assessment
Citation:
Farmahin, R., A. Williams, B. Kuo, N. Chepelev, Russell S. Thomas, T. Barton-Maclaren, I. Curran, A. Nong, M. Wade, AND C. Yauk. Recommended approaches in the application of toxicogenomics to derive points of departure for chemical risk assessment. Archives of Toxicology. Springer, New York, NY, 91(5):2045-2065, (2017).
Impact/Purpose:
• Agency Research Drivers -The US EPA has to prioritize chemicals for screening under the Frank R. Lautenberg Chemical Safety for the 21st Century Act. One approach for prioritization of large numbers of chemicals is using high-throughput screening (HTS) assays. HTS assays can rapidly assess chemical activity across a large number of endpoints, providing insights into chemical toxicity. • Science Challenge – Commercially available HTS assays focus largely on pharmacologically-relevant endpoints, limiting the ability of HTS approaches to inform environmental chemical toxicology. Expanding the biological coverage of HTS assays is necessary to use HT approaches for prioritization environmental chemical screening. It has recently been proposed that gene expression data can be used for dose-response analysis, but there is no consensus to select relevant genes for each chemical. Because a very large number of genes can be affected by chemical exposure, the process of selecting informative changes in gene expression is a major challenge. • Research Approach – We used published gene expression data to evaluated approaches for selecting informative changes in gene expression. These approaches were compared them to three previously proposed approaches for identifying relevant gene expression changes. • Results – We found that our gene expression change approaches to dose response analysis were remarkably aligned with in vivo results. A subset of our approaches met standard statistical criteria and would qualify as effective estimates of in vivo dose response. • Anticipated Impact/Expected use – Our results suggest that transcriptional response can be used as an efficient alternative approach for point of departure selection for chemical risk assessment.
Description:
ABSTRACT:Only a fraction of chemicals in commerce have been fully assessed for their potential hazards to human health due to difficulties involved in conventional regulatory tests. It has recently been proposed that quantitative transcriptomic data can be used to determine benchmark dose (BMD) and estimate a point of departure (POD). Several studies have shown that transcriptional PODs correlate with PODs derived from analysis of pathological changes, but there is no consensus on how the genes that are used to derive a transcriptional POD should be selected. Because of very large number of unrelated genes in gene expression data, the process of selecting subsets of informative genes is a major challenge. We used published microarray data from studies on rats exposed orally to multiple doses of six chemicals for 5, 14, 28, and 90 days. We evaluated eight different approaches to select genes for POD derivation and compared them to three previously proposed approaches. The relationship between transcriptional BMDs derived using these 11 approaches were compared with PODs derived from apical data that might be used in a human health risk assessment. We found that transcriptional benchmark dose values for all 11 approaches were remarkably aligned with different apical PODs, while a subset of between 3 and 8 of the approaches met standard statistical criteria across the 5-, 14-, 28-, and 90-day time points and thus qualify as effective estimates of apical PODs. Our results suggest that transcriptional response can be used as an efficient alternative approach for point of departure selection for chemical risk assessment.
