Student Chung-Ang University, Kyonggi-do, Republic of Korea
Abstract: The development of reliable preclinical models for assessing drug-induced toxicity is essential for improving the clinical and economic efficiency of drug development while ensuring patient safety. Human induced pluripotent stem cells (hiPSC)-derived neural organoids are considered effective resources for personalized medicine and neurotoxicity assessment due to their physiological similarity to human neural cells. In this study, we established a system that can evaluate drug toxicity in real time using patient hiPSC-derived neural organoids. To achieve this, we investigated 1) generating two cell lines capable of monitoring apoptosis in real time, 2) developing a high-efficiency system for producing hiPSC-derived neural organoids, 3) conducting neurotoxicity assessments of 12 FDA-failed drugs within a microfluidic-concave chip, and 4) incorporating AI-based machine learning algorithms to predict drug neurotoxicity using the generated database. As a result, the neurotoxicity predictions obtained from hiPSC-derived neural organoids exhibited over 80% concordance with clinical outcomes. Furthermore, neurotoxicity predictions of iPSC-derived neural organoids demonstrated over 80% similarity to clinical outcomes, proving to be a rapid, effective and cost-efficient system that can be applied to new drug development and safe patient-specific therapies.
Funding Source: This research was supported by the National Research Foundation (NRF) (No. RS-2023-00220207) and the BK21 FOUR funded by the MOE of Korea.
