Fiona Plait, Elise Dunzo, Noha Rayad, Laura Iavarone
Parexel International
Introduction: As part of the clinical development plan, a set of in vitro studies were undertaken to determine the drug-drug interaction (DDI) potential of a new compound. Studies using human microsomes and recombinant CYP enzymes, suggested that CYP3A4 was responsible for approximately 50% of the overall metabolism of the drug.
Objectives: To develop a PBPK model using GastroPlus® to assess the DDI liability of a new compound with 50% metabolism through CYP3A4 as victim of a strong CYP3A4 inhibitor and a strong CYP3A4 inducer.
Methods: A PBPK model was developed in GastroPlus®; In silico and in vitro data for the compound were utilized to build the PBPK base model, whereas clinical data from a Phase 1 study were used to verify and validate the model. Parameter sensitivity analyses were utilized during model development to understand how variation of key model input parameters can influence the modeling outcome. Overall prediction performance of the model was assess using an a priori fold-error (FE) value of two-fold for the key PK parameters. Once model development was complete, the base model was joined with predeveloped and validated model of 1) itraconazole, as a strong CYP3A4 inhibitor and 2) rifampin, as a strong CYP3A4 inducer. Ratios of Cmax and AUClast (DDI:alone) were evaluated for DDI potential using the predefined FDA criteria of <2, 2≤ratio<5 and ≥5, designated as mild, moderate or strong degrees of potential interaction, respectively.
Results: The model was developed using data from an X mg and 0.5X mg dose; predictability was assessed using the plasma PK parameters of Cmax, Tmax, AUClast, AUCinf, T1/2, C24, AUC(0-24), CL/F and Vz/F. For the X mg dose, the FE ranged from 0.64 to 1.62; only the FE for Vz/F fell below 0.5. For the 0.5X mg dose, the FE ranged from 0.61 to 1.73; FE for Tmax was 0.48. The model was then used to determine the predictability of 0.08X mg, 0.25X mg, 1.5X mg and 2.25X mg. The 0.08X mg dose generally under-predicted the clinical data; of the 5 PK parameters that passed criteria, the FE ranged from 0.51 to 1.81. All predicted PK parameters for the 0.25X mg dose passed the FE criteria and ranged from 0.56 to 1.51. Similarly, for the 1.5X mg dose, all predicted PK parameters for the 1.5X mg dose were well-predicted and ranged from 0.66 to 1.26. The 2.25X mg dose was also well predicted with FE ranging from 0.74 to 1.11.
The developed model at the X mg dose was utilized for the DDI simulations, as this is the primary intended target clinical dose. Plasma exposure for the new compound were simulated as single dose administration in the presence of the strong CYP3A4 inhibitor itraconazole administered at 400 mg once daily (QD) for 9 days (starting 7 days before coadministration to cover the whole disposition time of the new compound). The Cmax and AUClast geometric mean ratios (GMR) with and without itraconazole administration [90% confidence interval (90% CI)] were 1.08 (1.07, 1.24) and 1.33 (1.28, 1.41), respectively, indicative of a negligible DDI potential in the presence of a strong CYP3A4 inhibitor. In the presence of a strong CYP3A4 inducer, rifampicin (administered at 600 mg QD for 12 days; new compound administered on Day 10), GMR (90%CI) with and without rifampicin administration for Cmax and AUClast were 0.31 (0.27, 0.34) and 0.19 (0.17, 0.21), respectively. The predicted changes in the plasma exposure of the new compound during coadministration with a strong CYP3A4 inducer are considered to be moderate to severe.
Conclusions: PBPK modeling is a powerful tool that was used (1) to build a base model for the new compound and (2) to perform DDI predictions on this base model with strong CYP3A4 inhibitor and inducer, itraconazole and rifampicin, respectively. The degree of interaction was considered negligible for itraconazole and moderate to severe for rifampicin based on 50%-fraction metabolized by CYP3A4 (i.e. strong inhibitors appear to have a small effect on the compound with moderate fraction metabolized (fm), and strong inducers appear to have moderate to severe effect despite this moderate fm). In such cases, the recommendation would be to perform a clinical interaction study with a strong CYP3A4 inducer to evaluate potential reduction in exposure and consequent possible lower efficacy.
References:
[1] Clinical Drug Interaction Studies — Cytochrome P450 Enzyme- and Transporter-Mediated Drug Interactions Guidance for Industry | FDA
[2] Guideline on the investigation of drug interactions (europa.eu)
Reference: PAGE 30 (2022) Abstr 10061 [www.page-meeting.org/?abstract=10061]
Poster: Drug/Disease Modelling - Absorption & PBPK