Complementing in vitro screening assays with in silico molecular chemistry tools to examine potential in vivo metabolite-mediated effects
Citation:
Tan, C. Complementing in vitro screening assays with in silico molecular chemistry tools to examine potential in vivo metabolite-mediated effects. 2017 Society of Toxicology Annual Meeting, Baltimore, MD, March 12 - 16, 2017.
Impact/Purpose:
Invited presentation at the Society of Toxicology Annual Meeting in March 2017. This presentation summarizes an in silico approachthat enriches and complements in vitro methods, allowing for more accurate and efficient prioritization of chemicals based on hazard, through identification of inactive parents that could potentially generate active metabolites in vivo.
Description:
High-throughput in vitro assays offer a rapid, cost-efficient means to screen thousands of chemicals across hundreds of pathway-based toxicity endpoints. However, one main concern involved with the use of in vitro assays is the erroneous omission of chemicals that are inactive under assay conditions but that can generate active metabolites under in vivo conditions. To address this potential issue, a case study will be presented to demonstrate the use of in silico tools to identify inactive parents with the ability to generate active metabolites. This case study used the results from an orthogonal assay designed to improve confidence in the identification of active chemicals tested across eighteen estrogen receptor (ER)-related in vitro assays by accounting for technological limitations inherent within each individual assay. From the 1,812 chemicals tested within the orthogonal assay, 1,398 were considered inactive. These inactive chemicals were analyzed using Chemaxon Metabolizer software to predict the first and second generation metabolites. From the nearly 1,400 inactive chemicals, over 2,200 first-generation (i.e., primary) metabolites and over 5,500 second-generation (i.e., secondary) metabolites were predicted. Nearly 70% of primary metabolites were immediately detoxified or converted to other metabolites, while over 70% of secondary metabolites remained stable. Among these predicted metabolites, those that are most likely to be produced and remain stable were then inputted into a consensus model, using the 1,812 chemicals from the orthogonal assay as a training set, to predict ER binding and potency. The consensus model predicted a significantly lower number of active secondary metabolites, compared to active primary metabolites. This in silico approach enriches and complements in vitro methods, allowing for more accurate and efficient prioritization of chemicals based on hazard, through identification of inactive parents that could potentially generate active metabolites in vivo. Disclaimer: Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.
