In vitro approaches for modelling gut-liver crosstalk in drug-induced liver injury

Drug-induced liver injury (DILI) remains a major challenge in drug development and clinical practice due to its limited predictability in current preclinical models. Increasing evidence indicates that the gut-liver axis, through bidirectional interactions between the gut and the liver, plays an important role in modulating drug metabolism and hepatotoxicity. Nevertheless, the exact role of our gut, its microbiota and microbiota-derived metabolites in the manifestation of DILI remains poorly understood, highlighting the need for more human-relevant in vitro models taking along gut-liver interactions. To address this need, this review paper aimed to map the current landscape of in vitro gut-liver models and evaluate their potential applicability for future DILI research. A literature screening using the PubMed database identified 95 articles published on this subject between 2000 and 2026. The in vitro models were categorized into 5 groups, including monocultures, Transwell^® systems, hydrogel-based systems, organ-on-a-chip platforms and microfluidic bioreactors. Each category features distinct advantages and disadvantages with respect to physiological relevance, complexity, integration of gut microbiota interactions and industrial applicability. Despite these differences, several key gaps remain across all in vitro gut-liver models, including the limited incorporation of a gut microbiota component, insufficient modelling of interindividual variability and the lack of physiologically relevant flow conditions. Future optimization efforts of these models, combined with integration with in silico tools, may improve the prediction of human DILI, ultimately resulting in reduced late-stage drug attrition, less reliance on animal studies and the development of safer pharmaceutical drugs.