Improvement in static and dynamic projections of drug-drug interactions caused by cytochrome P4503A time-dependent inhibitors through in vitro allosteric modulation by progesterone.

Abstract

Current drug discovery screens to assess the drug-drug interaction (DDI) risk caused by time-dependent inhibition (TDI) of cytochrome P450 (CYP) 3A4 are known to overpredict or produce false positives that do not translate in vivo. Recent work identified that inclusion of the allosteric modulator progesterone (PGS), at a concentration of 45 μM to human liver microsomal incubations, generated in vitro TDI values that replicated clinical DDI predictions for 2 well established mechanism-based inhibitors. Further application of this approach across a diverse set of compounds was undertaken in this study, with 56 molecules reported in literature as time-dependent inhibitors in vitro tested in the human liver microsomal TDI kinetic assay in the absence and presence of 45 μM PGS. No TDI signal was observed for 15 molecules under control conditions despite literature reports. For the remaining compounds observed to have a TDI signal under control conditions, presence of PGS modified the inactivation efficiency for 36 compounds and eliminated the TDI signal for 5 compounds that were false positives. In vitro kinetic values were incorporated into mechanistic static and dynamic physiologically based pharmacokinetic models to project DDIs. TDI parameters established in the presence of PGS decreased the magnitude of overprediction while maintaining a high sensitivity (96% and 100%) for the detection of TDI with improved specificity (69% and 89%) when using mechanistic static and dynamic models, respectively. Inclusion of PGS into in vitro TDI assays provides a simple, rapid, and cost-effective solution for identifying true CYP3A4 TDIs and improving TDI-related DDI predictions. SIGNIFICANCE STATEMENT: The impact of the previously determined optimal concentration of the allosteric modulator progesterone (45 μM) was evaluated across a set of 56 compounds reported to be time-dependent inhibitors in vitro. In vitro generated values were incorporated into mechanistic static and physiologically based pharmacokinetic models to predict extent of drug-drug interactions and compared to clinical reports. Inclusion of progesterone into the assay identified in vitro false positives and improved risk predictions.