Abstract

Predicting transporter-mediated in vivo hepatic drug clearance (CL) from in vitro data (IVIVE) is important in drug development to estimate first in human dose and the impact of drug interactions and pharmacogenetics on hepatic drug CL. For IVIVE, one can use human hepatocytes and the traditional MGPGL (i.e. mg of protein content per gram of liver tissue) approach. However, this approach has been found to consistently under-predict the observed in vivo hepatic drug CL. Therefore, we hypothesized that using transporter-expressing cells and the relative expression factor (REF), determined using targeted quantitative proteomics, will accurately predict in vivo hepatic CL of drugs. We have successfully tested this hypothesis in rats with rosuvastatin, which is transported by hepatic OATPs. Here we tested this hypothesis for another drug and another transporter, namely OCT1- mediated hepatic distributional CL of metformin. First, we estimated the in vivo metformin hepatic sinusoidal uptake CL (CLh,s,in ) of metformin by reanalysis of previously published human positron emission tomography imaging data. Next, using the REF approach, we predicted the in vivo metformin CLh,s,in using OCT1-transporter expressing HEK293 cells or plated human hepatocytes. Finally, we compared this REF-based prediction with that using the MGPGL approach. The REF approach accurately predicted the in vivo metformin hepatic CLh,s,in while the MGPGL approach considerably under-predicted the in vivo metformin CLh,s,in . Based on these and previously published data the REF approach appears to be superior to the MGPGL approach for a diverse set of drugs transported by different transporters.

Significance Statement This study is the first to use OCT1-expressing cells and plated hepatocytes to compare proteomics-informed REF approach with the traditional MGPGL approach to predict hepatic uptake CL of metformin in humans. The plasma membrane abundance corrected proteomics-informed REF approach accurately predicted the PET-imaged metformin hepatic uptake CL, whereas the MGPGL approach consistently underpredicted this CL.