Clearance prediction with three novel plated human hepatocyte models compared to conventional suspension assays:...

Incubation of drugs with suspension hepatocytes (SH) to determine intrinsic clearance is common in drug discovery. However, the limited duration of SH assays hampers clearance assessment of metabolically stable compounds. In turn, this has driven the development of alternative in vitro approaches to generate intrinsic clearance estimates. Culturing primary hepatocytes with supportive cells as co/tricultures has been shown to maintain morphology, viability, and drug-metabolizing enzyme function for weeks, permitting extended incubations. Another assay from our laboratory is the preloaded hepatocyte assay (preload assay), which involves preloading plated monoculture hepatocytes with compounds and measuring the loss from cells in drug-free media. This approach increases analytical sensitivity compared to assays that measure bulk compound loss in the cells plus medium. We conducted a systematic evaluation of the ability of coculture, triculture, and preload assay models to predict human in vivo clearance for 50 predominantly low-clearance compounds with a range of physicochemical properties, including equal numbers of compounds following or violating Lipinski's rule of 5, across 3 hepatocyte donors. The results were compared with SH. Co/tricultures exhibited lower inter-donor differences compared to the preload and SH assays, likely due to the blunting of environmental cues after 5 days in culture prior to compound introduction. All 3 plated models significantly reduced the number of compounds with insufficient turnover to calculate CL_int,u compared to SH (SH: 40%; preload: 18%; cocultures: 8%; tricultures: 4%), exhibited strong interexperimental reproducibility and robust predictions of blood clearance (preload: 26/41; cocultures: 31/46; tricultures: 30/48 within 3-fold of observed). SIGNIFICANCE STATEMENT: Preloading plated hepatocytes with compounds and measuring the loss in drug-free media, or culturing hepatocytes with supportive cells as co/tricultures, facilitate quantitation of metabolically stable compounds in substrate depletion assays compared to suspension hepatocytes (SH). All 4 models exhibit robust estimates of CL_int,u and CL_b, but plated models allowed assessment of several compounds found to be too stable to evaluate in SH.