A Prediction Method for P-glycoprotein-Mediated Drug-Drug Interactions at the Human Blood-Brain Barrier From Blood Concentration-Time Profiles, Validated With PET Data.

The purpose of this study was to establish physiologically based pharmacokinetic models to predict in humans the brain concentration-time profiles and P-glycoprotein (Pgp)-mediated brain drug-drug interactions between the model Pgp substrate (R)-[11C]verapamil (VPM), the model dual Pgp/breast cancer resistance protein (BCRP) substrate [11C]tariquidar (TQD), and the Pgp inhibitor tariquidar. The model predictions were validated with results from positron emission tomography studies in humans. Using these physiologically based pharmacokinetic models, the differences between predicted and observed areas under the concentration-time curves (AUC) of VPM and TQD in the brain were within a 1.2-fold and 2.5-fold range, respectively. Also, brain AUC increases of VPM and TQD after Pgp inhibitor administration were predicted with 2.5-fold accuracy when in vitro inhibition constant or half-maximum inhibitory concentration values of tariquidar were used. The predicted rank order of the magnitude of AUC increases reflected the results of the clinical positron emission tomography studies. Our results suggest that the established models can predict brain exposure from the respective blood concentration-time profiles and rank the magnitude of the Pgp-mediated brain drug-drug interaction potential for both Pgp and Pgp/BCRP substrates in humans.