Mechanistic Pharmacokinetic Modeling for the Prediction of Transporter-Mediated Disposition in Humans from Sandwich Culture Human Hepatocyte Data

TLDR

Transporter-mediated clearance is increasingly sought to reduce CYP450-mediated clearance, yet PBPK prediction for such compounds remains less established than for passively mediated drugs. The study proposes a prediction strategy for novel OATP substrates. Using sandwich culture human hepatocyte data, the authors estimated uptake and biliary efflux parameters, then fitted in vivo data to derive empirical scaling factors for active uptake and efflux, which were incorporated into a PBPK model. The model overpredicted exposure for six of seven OATP substrates, but after applying empirical scaling factors (58 for uptake, 0.061 for efflux) it accurately captured multiphasic plasma profiles, demonstrating the utility of in vitro data for human PK prediction.

Abstract

With efforts to reduce cytochrome P450-mediated clearance (CL) during the early stages of drug discovery, transporter-mediated CL mechanisms are becoming more prevalent. However, the prediction of plasma concentration-time profiles for such compounds using physiologically based pharmacokinetic (PBPK) modeling is far less established in comparison with that for compounds with passively mediated pharmacokinetics (PK). In this study, we have assessed the predictability of human PK for seven organic anion-transporting polypeptide (OATP) substrates (pravastatin, cerivastatin, bosentan, fluvastatin, rosuvastatin, valsartan, and repaglinide) for which clinical intravenous data were available. In vitro data generated from the sandwich culture human hepatocyte system were simultaneously fit to estimate parameters describing both uptake and biliary efflux. Use of scaled active uptake, passive distribution, and biliary efflux parameters as inputs into a PBPK model resulted in the overprediction of exposure for all seven drugs investigated, with the exception of pravastatin. Therefore, fitting of in vivo data for each individual drug in the dataset was performed to establish empirical scaling factors to accurately capture their plasma concentration-time profiles. Overall, active uptake and biliary efflux were under- and overpredicted, leading to average empirical scaling factors of 58 and 0.061, respectively; passive diffusion required no scaling factor. This study illustrates the mechanistic and model-driven application of in vitro uptake and efflux data for human PK prediction for OATP substrates. A particular advantage is the ability to capture the multiphasic plasma concentration-time profiles for such compounds using only preclinical data. A prediction strategy for novel OATP substrates is discussed.