Strategy for CYP3A Induction Risk Assessment from Preclinical Signal to Human: a Case Example of a Late-Stage Discovery Compound.

Purpose: The exposure of G2917 decreased by four-fold at oral doses of 100 mg/kg twice daily for seven days in cynomolgus monkeys. Additional investigative work was conducted to understand: (1) the causes for the significant reduction in G2917 exposure in monkeys; (2) the extrapolation of in vitro induction data to in vivo findings in monkeys, and (3) the relevance of this pre-clinical finding to humans at the projected human efficacious dose.

Methods: Pharmacokinetic and induction potency (in vitro and in vivo) of G2917 in monkeys, and the in vitro human induction potency were studied. The hepatic CYP3A biomarkers 4β-hydroxycholesterol (4β-HC) and 6β-hydroxycortisol/cortisol ratio (6β-OHC/C) were monitored in in vivo studies. The static mechanistic model was used to quantitatively understand the in vitro-in vivo extrapolation (IVIVE) on the magnitude of induction retrospectively. Physiologically based pharmacokinetic (PBPK) modeling was used to predict the human pharmacokinetics and induction-based drug-drug interactions (DDI).

Results: All in vitro and in vivo data indicate that the significant reduction in exposure of G2917 in monkeys is caused by auto-induction of CYP3A. The mechanistic understanding of IVIVE of G2917 induction in monkey provides higher confidence in the induction risk prediction in human using the PBPK modeling. PBPK model analysis predicted minimum auto-induction and DDI liability in humans at the predicted efficacious dose.

Conclusions: The learning of this example provided a strategy to address the human CYP3A induction risk prospectively when there is an auto-induction finding in preclinical toxicology study.