DECODING THE GENETIC AND METABOLIC BASIS OF DRUG-INDUCED LIVER INJURY USING IN SILICO SIMULATIONS AND PATIENT-DERIVED HEPATOCYTE ORGANOIDS

Abstract

Drug-induced liver injury (DILI) is a leading cause of drug withdrawal and acute liver failure, driven by complex interactions between drug compounds and patient-specific genetic and metabolic factors. This study presents an integrated approach combining in silico molecular docking with in vitro patient-derived hepatocyte organoids to evaluate compound toxicity profiles. A panel of 180 compounds was computationally screened for binding affinity against hepatic enzymes, and high-risk candidates were tested on organoid cultures. Results demonstrated that compounds with strong binding affinity (ΔG < −10 kcal/mol) were associated with significantly reduced cell viability and increased markers of cytotoxicity, including LDH release and caspase-3/7 activation. Importantly, patient-derived organoids revealed genotype-dependent differences in toxicity response, underscoring the need for personalized toxicology. Multivariate analysis and clustering further distinguished three hepatotoxicity phenotypes, revealing that compound toxicity cannot be solely predicted by binding affinity. A series of 12 complex visualizations confirmed these trends and highlighted the interdependencies among toxicity indicators. This study establishes a novel hybrid model that leverages artificial intelligence, molecular simulation, and organoid-based experimentation to enhance the predictive accuracy of DILI assessment. The integration of these techniques provides a scalable, ethical, and clinically relevant alternative to traditional toxicity models and holds promise for streamlining the drug development pipeline.