Similarly for CYP2D6, the PBPK model provided excellent predictions for mexiletine in both EM and PM subjects, with a slight underprediction of drug clearance with multiple dosing in EM subjects. whole range of DDI studies investigated. The models are provided on GitHub (GitHub Inc., San DL-Adrenaline Francisco, CA, USA), expanding the library of publicly available qualified whole-body PBPK models for DDI predictions, and they are thereby available to support potential recommendations for dose adaptations, support labeling, inform the design of clinical DDI trials, and potentially waive DL-Adrenaline those. (value; gray area encompasses the standard deviation interval of the predicted profiles; (c) predicted and observed mean (standard deviation) mexiletine concentrations with (red) and without (blue) fluvoxamine. Additional comparisons of model-predicted vs. observed DDI outcomes can be found in Supplementary Material S3. CYP, cytochrome P450. Good predictions were also obtained in DDI simulations with fluvoxamine as a CYP2C19 inhibition perpetrator and omeprazole or S-mephenytoin as substrate (Figure 3aCc). Strong inhibition of CYP2C19 by the twice-daily co-administration of fluvoxamine led to an observed 5.3-fold increase and predicted 6.0-fold increase in the AUC of omeprazole in extensive metabolizers (EM), and an observed 21% increase in AUC and predicted 0% increase in PM. The difference between the observed increase of 21% and predicted increase of 0% is within bioavailability limits and thus may be explained by interindividual variability. Predictions of the inhibitory potential of fluvoxamine on CYP2C19 were excellent for both EM and PM subjects (Figure 3a,b): the ratios for mean predicted/observed AUC and Cmax were all around 1 (range: 0.83C1.13 for AUC and 0.89C1.33 for Cmax; Table 3). Open in a separate window Figure 3 Example comparisons of model-predicted and observed CYP2C19 DDI outcomes: (a) observed and predicted omeprazole concentrations with and without fluvoxamine coadministration in homozygous extensive metabolizers; (b) observed and predicted omeprazole concentrations with and without fluvoxamine coadministration in homozygous poor metabolizers; and (c) simulated and observed concentration-time course for moclobemide with and without omeprazole coadministration. Additional comparisons of model-predicted vs. observed DDI outcomes can be found in Supplementary Material S3. Table 3 Overview of results of DDI predictions of template models. = 0.8697 nM, b = 100 min?1, c = 200 min?1. 3.2. Omeprazole Similar to other approaches from the literature [11], a PBPK model for omeprazole as racemate and for the single enantiomers esomeprazole and R-omeprazole was developed with two parallel metabolic pathways via CYP2C19 and CYP3A4 as linear processes. Furthermore, a CYP2C19 autoinhibitory process via time-dependent inhibition was included. The omeprazole model was intended to be used as a substrate or perpetrator. Data from 34 clinical studies or study subgroups after intravenous DL-Adrenaline and oral administration using different formulations were used for model development and qualification, with nine studies or study subgroups assigned to the training set (Supplementary Material S1.2). CYP2C19 expression DL-Adrenaline in gut was lowered to account for the higher-than-expected oral bioavailability of the R-omeprazole with expression levels from PK-Sim databases [12]. Model predictions described well the observed concentrationCtime profiles after single and multiple doses esomeprazole/R-omeprazole in both CYP2C19 EMs and PMs. In addition, DDI simulations with omeprazole as perpetrator and moclobemide as a victim substrate demonstrated a good prediction of DL-Adrenaline moclobemide levels (Figure 3c). Predicted/observed ratios were in excellent agreement (1.07 for AUC, 0.89 for Cmax; Table 3). DDI simulations with omeprazole as victim substrate and moclobemide as perpetrator also demonstrated a good prediction of moclobemide action on CYP2C19 and CYP3A4 (Table 3). Although Cmax levels for EMs were slightly underpredicted (predicted/observed ratio: 0.77), predictions for PMs were excellent (predicted/observed ratio: 1.02), whereas AUC levels were slightly underpredicted in both patient populations (predicted/observed ratio: 0.79 for EMs and 0.86 for PMs; Table 3). 3.3. S-mephenytoin The classical path of PBPK modeling was not possible, as no PK data were available in humans after intravenous dosing, and there were only limited data following oral administration. Therefore, a PBPK model for S-mephenytoin as COPB2 a victim substrate was developed with an unspecific liver clearance scaled from an in vivo measured apparent clearance [13,14] as the main clearance process and without further fitting of any other model parameter. For qualification, data from five clinical studies after oral administration using different studies were used (Supplementary Material S1.3). DDIs with fluvoxamine as perpetrator were simulated, and simulations were used to investigate several values found in.