Modeling Drug Induced Liver Injury with Patient-Derived Liver Organoids and Microfluidic Chips

Drug-induced liver injury (DILI) is an infrequent but important cause of both acute and chronic liver disease. The symptoms of DILI can be vague and non-specific, and may not appear until weeks or months after drug administration. While most cases of DILI are mild and resolve by stopping the offending medication, in some cases it can lead to liver failure requiring transplantation or death. Additionally, there are no drugs or therapies that can reverse the liver damage caused by DILI. An estimated 22% of clinical trial failures and 32% of market withdrawals of novel molecular entities are due to unanticipated hepatotoxicity. The combination of iPSC derived human liver organoids (HLOs) from DILI patients with organ-on-chip technology has the potential to revolutionize DILI risk prediction and may ultimately lead to the development of safer and more effective medications. We have established a biobank from 15 well-characterized U-M patients enrolled in DILIN along with methods to engineer liver tissue that can recapitulate their specific disease phenotype to enable mechanistic inquiry and establish improved animal-free methods for predicting DILI. We propose to fully develop a next-generation DILI risk prediction platform to improve predictive accuracy and improve our understanding of the etiopathogenesis and biomarkers of DILI due to specific drugs or HDS products. We hypothesize that HLOs from DILI patients will provide a novel and scalable platform for risk prediction and mechanistic inquiry. We will test this hypothesis with three Specific Aims: (1) expand DILI patient biobank through more patient enrollment and increase the number of culprit drugs to characterize more mechanisms of DILI, (2) increase the model complexity by incorporating additional liver cell types including same-patient immune cells, cholangiocytes, and an endothelial barrier that will be more physiologically active to enable high accuracy in retrospective prediction of the culprit drug, (3) perform state-of-the-art single-cell mechanistic studies involving selected drugs including amoxicillin-clavulanate and link genotype and transcriptional profiles to cellular phenotypes following drug exposure. The top identified targets will be validated by CRISPR genome editing, functional assessment, and differential hepatotoxicity for DILI drugs.