Assessing in vitro long-term liver metabolism models to characterize intrinsic clearance determination in NGRAs

In Next Generation Risk Assessments, measurements of hepatic intrinsic clearance (CLint , in vitro ) and metabolite(s) formation can markedly impact in vitro to in vivo extrapolations and subsequent testing of parent and/or metabolites in in vitro bioassays. It is especially important for slowly metabolized chemicals to use the most suitable model and experimental design to ensure that the IVIVE is accurate. We evaluated four long-term hepatic models (sandwich cultures of primary human hepatocytes (PHH), HepatoPac, HµREL and TruVivo), as well as PHH suspensions, using salicylic acid, 1-methyl-2-pyrrolidinone (NMP) and valproic acid. Assay designs differed in terms of positive and negative controls, cell densities and incubation durations. All models generated a measurable CLint , in vitro value for salicylic acid i.e. depletion >20%. Differences may be due to donor differences and the cell number. HepatoPac generally resulted in higher values compared with the other models. NMP was not metabolized by HµREL and TruVivo models and a reliable value was only possible using HepatoPac. NMP was metabolized by sandwich cultures; however, the half-life was much longer than the incubation duration. To derive a more robust value, follow up testing could include longer incubation times. Valproic acid was only metabolized in one of the three models in which it was detected, namely sandwich cultures. In conclusion, PHH sandwich cultures, HepatoPac, HµREL and TruVivo models are superior to PHH suspensions with respect to maintaining XME activities over days and for measuring CLint, in vitro values for slowly metabolized chemicals. To gain regulatory acceptance and better interpret data from different labs, we recommend increasing the number of chemicals tested and harmonization of protocols.