(M-107) Physiologically-Based Pharmacokinetic Modeling of Maribavir Incorporating Metabolism by Cytochrome P450, Glucuronidation, and Hepatic Uptake, for Prediction of Victim Drug-Drug Interaction Potentials

Howard Burt, Ph.D. – Senior Director, PBPK Consultancy, Certara UK Limited; Ingrid Michon, Ph.D. – Principal PBPK Consultant, Certara UK Limited; Ivy Song, Ph.D. – Executive Director, Quantitative Clinical Pharmacology, Takeda Development Center Americas Inc

Objectives: To (1) improve a previously developed PBPK model [1] for maribavir (TAK-620) by including all known metabolic and disposition pathways, including CYP3A4-, CYP1A2-, and UDP-glucuronosyltransferase (UGT)-mediated metabolism, renal clearance, and uptake via organic cation transporter 1 (OCT1); (2) use the validated model to assess drug-drug interaction (DDI) liability of maribavir as a victim of CYP3A4-, CYP1A2- and UGT-mediated metabolism in combination with CYP3A4 inducers.

Methods: (1) Model development: A combination of in vitro ADME and clinical PK data of a single 400 mg oral dose of maribavir in healthy subjects (without or with concurrent ketoconazole) were used to develop the PBPK model in Simcyp v22. (2) Model validation: Assignment of parameters in the model was validated with clinical PK data at 400 to 1600 mg of single and multiple doses. The contribution (fractions metabolized [fm]) of pathways was independently validated with clinical DDI data with rifampicin. (3) Model application: The model was used prospectively to predict DDI outcomes with rifampicin (strong CYP3A4, CYP1A2, UGT1A1 inducer), phenobarbital (strong CYP3A4 inducer), phenytoin (strong CYP3A4 inducer and CYP1A2 inducer), carbamazepine, rifabutin and efavirenz (moderate CYP3A4 inducers).

Results: (1) Model development: Maribavir is metabolized by CYP3A4 (fm 53%), CYP1A2 (fm 27%) and UGT1A1 (fm 20%). Renal excretion accounts for < 1% of total clearance. Hepatic uptake through both OCT1 and passive permeability occurs prior to hepatic metabolism. The previously undefined fm of 61% [1] was eliminated. (2) Model validation: Validation was achieved as simulated AUC, Cmax, and Ctrough and exposure ratios were within 0.80x to 1.25x of observations, for all dose levels of single and multiple dose maribavir, without or with rifampicin coadministration. (3) Model application: Coadministration of maribavir with 600 mg QD rifampicin leads to a reduction in maribavir AUC, Cmax, and Ctrough to 0.46x, 0.70x and 0.16x of levels vs maribavir alone. Coadministration of maribavir with commonly prescribed regimens of phenobarbital, phenytoin and efavirenz leads to reduction in maribavir AUC, Cmax, and Ctrough to 0.33x - 0.50x of levels vs maribavir alone. Coadministration of maribavir with commonly prescribed regimens of carbamazepine or rifabutin leads to reduction in maribavir AUC, Cmax, and Ctrough to 0.50x - 0.90x of levels vs maribavir alone.

Conclusions: An improved PBPK model of maribavir was constructed to incorporate all known pathways of maribavir. The improved PBPK model is considered more scientifically rigorous, as it eliminated the previously unspecified fm of 61%. The model was validated with clinical data and then used to predict maribavir in victim DDI with concurrent CYP3A4 inducers. The results could inform dosing decisions with maribavir in such scenarios with concomitant CYP3A4, CYP1A2 and UGT inducers.

Citations: [1] Chen G, Sun K, Michon I, Barter Z, et al. Physiologically-Based Pharmacokinetic Modeling for Maribavir to Inform Dosing in Drug–Drug Interaction Scenarios with CYP3A4 Inducers and Inhibitors. J Clin Pharmacol 2024 Jan;64(1):80-93