Prediction of the potential for drug-induced liver injury (DILI) in the early stages of drug development is important. We developed an organoid-based and functional endpoint method for accurate prediction of DILI. To this end, hepatic organoids (HOs) derived from human pluripotent stem cells (hPSCs) were cocultured with hepatic stellate cells (HSCs) and THP-1 macrophages in Matrigel domes to mimic the cellular and physiological environment of the human liver. To validate our hepatotoxicity prediction model, we selected 12 hepatotoxic reference compounds. As indicators, we used factors related to mechanisms of hepatotoxicity and markers thereof, including factors related to oxidative stress and proinflammatory cytokines. We plotted radar graphs and calculated the relative areas of polygons to analyze the effects of drugs with different degrees of hepatotoxicity. The drugs in the severe DILI group significantly increased the levels of factors related to oxidative stress (ROS, GSSH, and catalase) compared to those in the no and mild DILI groups. The drugs in the severe group significantly increased the levels of inflammation-related factors (IL-1, IL-6, and IL-10). The drugs in the mild and severe groups highly significantly increased the activities of ALT and AST and the level of ALB compared to those in the no DILI group. In summary, the drugs in the severe DILI group had significantly greater effects on the factors analyzed than those in the no DILI group. Therefore, our hepatotoxicity evaluation method is suitable for predicting DILI in the early stages of drug development.
Keywords: Drug-induced liver injury; Hepatic organoid; Hepatotoxicity.
© 2025. The Author(s).