Real-Time Bioenergetic Analysis in a Microphysiological Human Hepatic Function Model for Cell-Stressor Identification

Drug-induced liver toxicity is difficult to detect using animal models, leading to clinical failures and post-marketing drug withdrawals. Also, current in vitromodels lack key human liver functionalities (animal primary hepatocytes or cancerous/immortalized human lines) or show enhanced prediction but low throughput, high run costs and lack or conformity (primary human hepatocytes).

Here we demonstrate a dual cell type liver organoid, spontaneously organizing in a collagen gel matrix from E6/E7^LOWhuman hepatocytes and endothelial cells. We follow its formation and spatial cell associations, revealing a liver-like zonation profile, as evident by the differential distribution of the enzymes G6pase (periportal marker) and CYP2E1 (perivenous marker). Ample evidence is provided of its physiologically relevant transcriptional profile and liver functional metrics (albumin, urea production etc). We further utilized these organoids in an improved liver-on-chip microphysiological platform that can assess in parallel the gross mitochondrial/metabolic effects of multiple substances and physiological stressors. The system automatically identifies the organoids in a 384well format and measures in-situdrug/stressor-induced oxygen consumption fluctuations by two-frequency phase modulation of organoid-embedded phosphorescent microprobes. The readings revealeffects on mitochondrial function – where the main bulk of oxygen utilization happens – and cell metabolism, even at sub-lethal doses. We finally present a perfusion bioenergetic platform, based on the same organoids and oxygen sensing, but with medium flow simulating blood flow sheer in liver sinusoids and also metabolite sensing (glucose, lactate, glutamine) for complete liver-metabolic profiling. Both systems provide high predictive capacity and represent a significant advancement in pharmaceutical toxicological assessment and liver modelling.