Improved Prediction Of Clinical Drug‐Drug Interactions Using A Novel Numerical Method For Evaluation Of Time‐Dependent Inhibition Of Cytochrome P450

Time‐dependent inhibition (TDI) of cytochrome P450 (CYP), can lead to drug‐drug interactions (DDI) and has resulted in withdrawal of drugs in the past. The replot method is the preferred method for in‐vitro TDI analysis. This method often leads to over prediction of DDI. Hence accurate and timely prediction of in‐vivo DDI mediated by TDI is critical. We hypothesize that the poor prediction of TDI mediated DDI result from the complexities in TDI kinetics and improper in‐vitro analysis. We propose a new numerical method (1) which can incorporate complex CYP mechanisms. This method does not assume steady state and irreversibility and can better estimate K I and k inact , thus improving DDI predictions. In‐vitro TDI assays were conducted with human liver microsomes with the standard two‐step method as described previously(1). Verapamil (VER), diltiazem (DTZ), erythromycin (ERY) and their respective primary metabolites and troleandomycin (TAO) were tested as inactivators of CYP3A4 with midazolam as a substrate. The concentrations of 1‐hydroxy midazolam were measured by LC‐MS/MS. Several kinetic schemes were developed for different inactivators. Mathematica® 10.1. was used to estimate K I and k inact using both the replot and numerical methods. The ratio of the area under the curve (AUC) of the substrate in the presence (AUC i ) and absence (AUC) of each inactivator was predicted using a static model and parameters obtained from either replot or numerical methods. A better fit for TAO was obtained with the two‐binding site model. The K I and k inact for TAO with the numerical method were 0.73 μM and 0.003 min −1 and 2.26 μM and 0.008 min −1 for the first and second site respectively. K I and k inact obtained for ERY, N‐demethyl erythromycin (NDE), DTZ, N‐desmethyl diltiazem (MA), norverapamil (NV), and VER using numerical method were 10.4 μM and 0.004 min −1 , 4.17 μM and 0.01 min −1 , 44 μM and 0.003 min −1 , 1.18 μM and 0.003 min −1 , 6.8 μM and 0.007 min −1 and 11.4 μM and 0.007 min −1 respectively. K I and k inact obtained through replot method were 7.31 μM and 0.021 min −1 , 3.84 μM and 0.036 min −1 , 1.61 μM and 0.017 min −1 , 1 μM and 0.018 min −1 , 2.53 μM and 0.025 min −1 , 4.91 μM and 0.035 min −1 , 1.67 μM and 0.022 min −1 for ERY, NDE, DTZ, MA, NV, VER and TAO respectively. The numerical method was successfully applied to estimate K I and k inact and predict clinical DDIs. A total of 37 clinical DDIs were predicted with the four inactivators and CYP3A substrates. The average fold difference between observed and predicted was 4.86 for the replot method and 1.26 for numerical method. AUC ratio predictions with Midazolam as the substrate were 2.48 (numerical method) and 7.08 (replot method) vs 2.2 (observed) for ERY. Similarly, for DTZ AUC predictions were 1.01 (numerical method) and 3.64 (replot method) vs 1.6 (observed), for TAO it was 8.31 (numerical method) and 10.44 (replot method) vs 4.6 (observed), for VER it was 8.7 (numerical method) and 21.22 (replot method) vs 2 (observed). In conclusion, TDI kinetic parameters obtained from the numerical method resulted in improved prediction of DDI potential of the tested inactivators thereby demonstrating increased ability over the replot method to characterize TDI kinetics. Support or Funding Information This work was supported by National institute of health (NIH) grant GM1R01GM114369 and 1R01GM104178 This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .