Navigating ADME profiling challenges in microphysiological systems: Evaluation of a liver-chip model for clearance...

Accurate prediction of the hepatic clearance in humans is essential during pre-clinical drug development. As conventional in vitro models do not replicate the complexity of the human liver physiology, hepatocyte-specific functions are rapidly lost resulting in limited assay sensitivity. In this study, a commercially available organ-on-chip model featuring a two-compartment microfluidic architecture was evaluated to estimate human hepatic clearance for a panel of 15 commercial drug compounds with diverse metabolic pathways and clearance rates. Upon further optimization of the model for the co-culture of primary human hepatocytes and liver sinusoidal endothelial cells, key system parameters - including compound permeability, non-specific binding, and evaporation - were systematically characterized. Despite optimized conditions, the human hepatic clearance was generally underpredicted by the liver-chip. The best performance was observed for compounds with low to moderate clearance while the underprediction was more pronounced for compounds exhibiting high metabolic turnover. By introducing a model-specific systematic scaling factor, more than 79% of clearance predictions fell within a three-fold range of observed human values. The study identified challenges in liver-chip systems for ADME applications that stem from specific design features and proposes optimization strategies for using liver-chips in metabolic stability assessments.