Physiologically Based Pharmacokinetic Modeling of the Drug-Drug Interaction Between CYP3A4 Substrate Glasdegib and Moderate CYP3A4 Inducers in Lieu of a Clinical Study.

Glasdegib (DAURISMO) is a hedgehog pathway inhibitor approved for the treatment of acute myeloid leukemia (AML). Cytochrome P450 3A4 (CYP3A4) has been identified as a major metabolism and clearance pathway for glasdegib. The role of CYP3A4 in the clearance of glasdegib has been confirmed with clinical drug-drug interaction (DDI) studies following the coadministration of glasdegib with the strong CYP3A4 inhibitor ketoconazole and the strong inducer rifampin. To evaluate potential drug interactions with CYP3A4 modulators, the coadministration of glasdegib with a moderate CYP3A4 inducer, efavirenz, was evaluated using physiologically based pharmacokinetic (PBPK) modeling using the Simcyp simulator. The glasdegib compound file was developed using measured physicochemical properties, data from human intravenous and oral pharmacokinetics, absorption, distribution, metabolism, and excretion studies, and in vitro reaction phenotyping results. The modeling assumptions, model parameters, and assignments of fractional CYP3A4 metabolism were verified using results from clinical pharmacokinetics (PK) and DDI studies with ketoconazole and rifampin. The verified glasdegib and efavirenz compound files, the latter of which was available in the Simcyp simulator, were used to estimate the potential impact of efavirenz on the PK of glasdegib. PBPK modeling predicted a glasdegib area under the concentration-time curve ratio of 0.45 and maximum plasma concentration ratio of 0.75 following coadministration with efavirenz. The PBPK results, in lieu of a formal clinical study, informed the drug label, with the recommendation to double the clinical dose of glasdegib when administered in conjunction with a moderate CYP3A4 inducer, followed by a resumption of the original dose 7 days post-discontinuation.

Prediction of Human Brain Penetration of P-glycoprotein and Breast Cancer Resistance Protein Substrates Using In Vitro Transporter Studies and Animal Models.

Four P-glycoprotein (P-gp) and breast cancer resistance protein (BCRP) substrates with human cerebrospinal fluid (CSF) concentrations and preclinical neuropharmacokinetics were used to assess in vitro-in vivo extrapolation of brain penetration in preclinical species and the ability to predict human brain penetration. Unbound brain (Cb,u), unbound plasma (Cp,u), and CSF compound concentrations (CCSF) were measured in rats and nonhuman primates (NHPs), and the unbound partition coefficients (Cb,u/Cp,u and CCSF/Cp,u) were used to assess brain penetration. The results indicated that for P-gp and BCRP dual substrates, brain penetration was severally impaired in all species. In comparison, for P-gp substrates that are weak or non-BCRP substrates, improved brain penetration was observed in NHPs and humans than in rats. Overall, NHP appears to be more predictive of human brain penetration for P-gp substrates with weak or no interaction with BCRP than rat. Although CCSF does not quantitatively correspond to Cb,u for efflux transporter substrates, it is mostly within 3-fold higher of Cb,u in rat and NHP, suggesting that CCSF can be used as a surrogate for Cb,u. Taken together, a holistic approach including both in vitro transporter and in vivo neuropharmacokinetics data enables a better estimation of human brain penetration of P-gp/BCRP substrates.