Abstract

Relative expression factors (REFs) are used to scale in vitro transporter kinetic data via in vitro–in vivo extrapolation linked to physiologically based pharmacokinetic (IVIVE-PBPK) models to clinical observations. Primarily two techniques to quantify transporter protein expression are available, immunoblotting and liquid chromatography–tandem mass spectrometry. Literature-collated REFs ranged from 0.4 to 5.1 and 1.1 to 90 for intestinal P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP), respectively. The impact of using human jejunum–Caco-2 REFs for P-gp (REF_iP-gp) and BCRP (REF_iBCRP), generated from the same samples and using different proteomic methodologies from independent laboratories, on PBPK outcomes was assessed. A 5-fold decrease in REF_iP-gp for a single oral dose of digoxin resulted in a 1.19- and 1.31-fold higher plasma area under the curve and C_max, respectively. All generated REF_iP-gp values led to simulated digoxin C_max values within observed ranges; however, combining kinetic data generated from a different laboratory with the 5-fold lower REF_iP-gp could not recover a digoxin-rifampicin drug-drug interaction, emphasizing the necessity to obtain transporter-specific kinetic estimates and REFs from the same in vitro system. For a theoretical BCRP compound, with absorption taking place primarily in the jejunum, a decrease in the REF_iBCRP from 2.22 (University of Manchester) to 1.11 (Bertin Pharma) promoted proximal intestinal absorption while delaying t_max 1.44-fold. Laboratory-specific differences in REF may lead to different IVIVE-PBPK outcomes. To understand the mechanisms underlying projected pharmacokinetic liabilities, it is important to assess the potential impact of bias on the generation of REFs on an interindividual basis within a target population.