Prediction of toxicity risk and symptom-specific adverse effects using MEA measurements of human iPSC-derived neurons

Abstract

In vitro human iPSC-derived neuronal systems are an alternative platform for neurotoxicity testing to animal models and primary cultures. Microelectrode array (MEA), measurement system of the electrophysiological activity, are suitable to evaluate the neurotoxicity of compounds. We have previously reported the electrophysiological responses to known neurotoxicity compounds using MEA in human iPSC-derived neurons. However, the identification of analytical parameters to detecting toxicity of unknown compounds remains an important issue. In this study, we evaluated the toxicity risk of more than 100 compounds. We identify the analytical parameters enabling the separation of responses between neurotoxicity and negative control, and the separation the mechanism of action by using principal component analysis. By comparing the optimized analytical parameters of each testing compound with the standard deviation (SD) of negative control, we can predict general neurotoxicity in relatively quantitatively scale, for example, low risk for lower than SD range, medium risk for 2xSD, and high risk for over 2xSD. The neurotoxicity risk of compounds associated with adverse event reports was detected with a probability of 91.0%, demonstrating that this evaluation method enables highly sensitive and specific prediction of toxicological risks. Additionally, it was shown to specifically detect not only seizure-related risks but also adverse effects such as psychiatric symptoms and headaches. This predictability can help select appropriate concentration levels to avoid toxicity/ adverse effects.