A human liver microphysiology platform for investigating physiology drug safety and disease models

摘要

This paper describes the development and characterization of a microphysiology platform for drug safety and efficacy in liver models of disease that includes a human, 3D, microfluidic, four-cell, sequentially layered, self-assembly liver model (SQL-SAL); fluorescent protein biosensors for mechanistic readouts; as well as a microphysiology system database (MPS-Db) to manage, analyze, and model data. The goal of our approach is to create the simplest design in terms of cells, matrix materials, and microfluidic device parameters that will support a physiologically relevant liver model that is robust and reproducible for at least 28 days for stand-alone liver studies and microfluidic integration with other organs-on-chips. The current SQL-SAL uses primary human hepatocytes along with human endothelial (EA.hy926), immune (U937) and stellate (LX-2) cells in physiological ratios and is viable for at least 28 days under continuous flow. Approximately, 20% of primary hepatocytes and/or stellate cells contain fluorescent protein biosensors (called sentinel cells) to measure apoptosis, reactive oxygen species (ROS) and/or cell location by high content analysis (HCA). In addition, drugs, drug metabolites, albumin, urea and lactate dehydrogenase (LDH) are monitored in the efflux media. Exposure to 180 μM troglitazone or 210 μM nimesulide produced acute toxicity within 2–4 days, whereas 28 μM troglitazone produced a gradual and much delayed toxic response over 21 days, concordant with known mechanisms of toxicity, while 600 µM caffeine had no effect. Immune-mediated toxicity was demonstrated with trovafloxacin with lipopolysaccharide (LPS), but not levofloxacin with LPS. The SQL-SAL exhibited early fibrotic activation in response to 30 nM methotrexate, indicated by increased stellate cell migration, expression of alpha-smooth muscle actin and collagen, type 1, alpha 2. Data collected from the in vitro model can be integrated into a database with access to related chemical, bioactivity, preclinical and clinical information uploaded from external databases for constructing predictive models.