Investigation into LiverChipTM as a 3D hepatocyte model for clearance prediction

The quantitative prediction of drug clearance from human in vitro data remains challenging. Recently, novel liver cell culture models have been developed that more accurately mimic the in vivo 3D environment and potentially maintain hepatocyte specific functions longer. LiverChipTM is a dynamic 3D cell culture platform in which hepatocytes can be cultured up to 14 days. The aim of the present study was to evaluate this novel 3D perfused hepatocyte system in terms of its quantitative capability to predict hepatic drug clearance. Human hepatocytes were cultured under dynamic culture conditions in LiverChipTM scaffolds for up to 10 days. Secretion of albumin and urea was measured using a commercially available ELISA and colorimetric assay, respectively. Functional activity of CYP and UGT metabolising enzymes was determined using multiple probe substrates, including both low and high clearance compounds at a single substrate concentration (5 µM). Intrinsic metabolic clearance (CLint) values were calculated from parent medium depletion rates and/or metabolite formation rates for tolbutamide (CYP2C9), diclofenac (CYP2C9), midazolam (CYP3A4), alprazolam (CYP3A4), phenacetin (CYP1A2) and gemfibrozil (UGT2B7). Functional activity of OATP1B transporters was assessed from initial rosuvastatin uptake rate data (up to 3 min) in hepatocytes cultured in LiverChipTM for 10 days. Albumin secretion in hepatocytes cultured in LiverChipTM was stable for up to 10 days of culture. Urea synthesis decreased with time, but urea medium concentrations were higher in dynamic cultures compared to static 2D cultures. CLint values ranged from 0.6 ± 0.15 to 27.7 ± 4.2 µL/min/M cells for alprazolam and gemfibrozil, respectively; clear resolution between low and high clearance drugs was achieved for CYP2C9 and CYP3A4 substrates. Tolbutamide Clint values in hepatocytes cultured for 3 and 6 days were comparable, demonstrating the preservation of metabolic functionality of at least CYP2C9 over the culture time. In vitro – in vivo extrapolation of metabolic clearance data resulted in good concordance of hepatic CLint for tolbutamide (7.4 mL/min/kg), alprazolam (1.6 mL/min/kg) and gemfibrozil (71 mL/min/kg). Significant under-prediction was observed for high clearance compounds (diclofenac, 42.5 mL/min/kg; midazolam, 14.4 mL/min/kg; phenacetin, 8.3 mL/min/kg). Rosuvastatin total uptake clearance values in hepatocytes cultured in LiverChipTM for 10 days amounted to 8.1 µL/min/mg protein and decreased by 70 % in presence of the OATP1B inhibitor rifamycin SV. In summary, liver-specific functions are well maintained in hepatocytes cultured up to 10 days under dynamic conditions in LiverChipTM. Moreover, functional activity of major CYP enzymes, UGT enzymes and OATP1B transporters was demonstrated and illustrates the potential of LiverChipTM to quantitatively predict drug clearance, in particular for low clearance compounds.