Human pluripotent stem cell-derived 2D neural tissue samples for predictive neurotoxicity screening

There is a growing need for fast and accurate methods for testing developmental neurotoxicity across industrial, pharmaceutical, and environmental chemical exposures. Current approaches, such as in vivo animal studies, and assays of animal and human primary cell cultures, suffer from challenges related to time, cost, and applicability to human physiology. Prior research demonstrated success employing machine learning to predict developmental neurotoxicity using gene expression data collected from complex human 3D tissue models exposed to various compounds, but the complexity of 3D tissue models require extensive expertise and effort to employ. While a 3D tissue model is more physiologically accurate, by focusing only on the goal of constructing an assay of developmental neurotoxicity, we propose that a simpler 2D tissue model may prove sufficient. We thus compared the accuracy of predictive models trained on data from a 2D tissue model with those trained on prior dataset from a more complex 3D tissue model, and found the accuracy of the 2D model to be substantially better than the 3D model. Furthermore, we found that the 2D tissue model is more robust and consistent under stringent gene set selection, whereas the 3D tissue model suffers substantial degradation of accuracy. While both approaches have advantages and disadvantages, we propose that our described 2D tissue model has the potential to serve as a valuable tool for decision makers when prioritizing neurotoxicity screening.