You are here:
Computational approaches to integrate DNT NAMs for fit-for-purpose identification of DNT hazard (SOT March 2022)
Citation:
Carstens, K., T. Shafer, AND K. Paul-Friedman. Computational approaches to integrate DNT NAMs for fit-for-purpose identification of DNT hazard (SOT March 2022). Society of Toxicology 61st Annual Meeting and ToxExpo 2022, San Diego, California, March 27 - 31, 2022. https://doi.org/10.23645/epacomptox.19585813
Impact/Purpose:
Presentation to the Society of Toxicology Annual Meeting March 2022. The purpose of this presentation is to share and receive feedback on work to integrate multiple assay endpoints for developmental neurotoxicity new approach methdologies into a relevant indicator for developmental neurotoxicity screening. This presentation is part of a symposium titled 'New Approach Methods for Functional Developmental Neurotoxicity' which will be presented at the Society of Toxicology annual meeting in 2022. The abstract and the proposal will be reviewed by the meeting planning committee.
Description:
Current developmental neurotoxicity (DNT) hazard assessment relies on in vivo testing that is resource intensive and lacks information on key cellular processes. To address these limitations, DNT new approach methodologies (NAMs) are being evaluated for their utility to inform DNT hazard, including: functional microelectrode array network formation assay (NFA) to evaluate neuronal network formation; and high-content imaging to evaluate proliferation, apoptosis, neurite outgrowth, and synaptogenesis. This work applies computational approaches to address three related hypotheses: (1) a broad screening battery will provide a sensitive marker of potential DNT bioactivity; (2) evaluating selective bioactivity may provide a more specific indicator of the functional processes underlying DNT; (3) some subset of the DNT NAM endpoints may optimally classify DNT in vivo reference chemicals. The dataset contained 57 assay endpoints, with a union set of 92 screened chemicals, including 53 in vivo DNT reference positives and 13 putative negatives. K-means clustering of selectivity revealed five chemical clusters with distinct DNT-relevant activity and identified DNT reference compounds with 66% sensitivity and 92% specificity, capturing 18 false negatives. We used a supervised machine learning approach to identify the most informative endpoints in classifying a DNT reference chemical. The five most important features included three cytotoxicity endpoints, network formation, and neurite outgrowth endpoints.This work highlights current obstacles in DNT NAM data interpretation, including a limited set of reference negatives and the question of whether cytotoxicity should be considered off-target or is relevant to deriving key DNT-related points-of-departure. Together, these data emphasize the importance of an integrated analysis that combines computational approaches, a broad chemical set, and a diverse suite of assays to demonstrate the fit-for-purpose utility of DNT NAMs for identification, hazard characterization, and prioritization of chemicals. This abstract does not reflect EPA policy.
URLs/Downloads:
DOI: Computational approaches to integrate DNT NAMs for fit-for-purpose identification of DNT hazard (SOT March 2022)
CARSTENS_DNT_SOT22_28MAR2022_SUBMIT_STICS.PDF (PDF, NA pp, 2827.084 KB, about PDF)