PK and PK / PD Modeling of Bcl2 Inhibitor S65487 in Patients With AML and Investigation of Nonlinearity With Microdosing

S65487, a drug targeting the Bcl2/Bim protein complex implicated in cancer cell survival, was studied using pharmacokinetic/pharmacodynamic modeling. At a 1200 mg dose, simulations predicted a 97.6% maximum reduction in the formation of the Bcl2/Bim complex. A microdose study in healthy volunteers showed substantially different pharmacokinetic parameters compared to therapeutic doses in cancer patients, indicating slightly nonlinear drug behavior. The most plausible explanation was nonlinearity in drug disposition, modeled via a Michaelis-Menten approximation of target-mediated drug disposition. This optimized model reduced the discrepancy in describing microdose exposure from 5-fold to 1-fold, without affecting therapeutic dose descriptions, enabling better extrapolation from microdose to therapeutic dose results.

Study at a glance

Abstract

This work aims to assess the S65487-induced disruption of the B-cell lymphoma-2 (Bcl2)/Bcl2-like protein 11 (Bim) complex at therapeutic doses via a pharmacokinetic/pharmacodynamic model and to reconcile pharmacokinetic data from microdoses and therapeutic doses in a single population pharmacokinetic model. The first model consists of a linear 3-compartment pharmacokinetic model linked to an indirect response model with concentration-dependent inhibition of Bcl2/Bim complex formation. Simulations predicted a 97.6% maximum reduction in complex formation with 1200 mg of S65487, following a sustained administration schedule and using characteristics from phase I/II patients. Before clinical trials were conducted in cancer patients, a microdose study was conducted in healthy volunteers. Although S65487 exhibited slightly nonlinear pharmacokinetics at therapeutic doses, pharmacokinetic parameters estimated after microdose administration differed substantially. Thus, several hypotheses were explored to reconcile the microdose and therapeutic data within a unified PK model. Nonlinearity in drug disposition emerged as the most biologically plausible explanation. A Michaelis-Menten approximation of the target-mediated drug disposition model was applied to elimination and distribution processes to estimate parameters through saturable pathways. Performance of this second model, the microdose optimized pharmacokinetic model, was evaluated by comparing population predicted and observed area under the curve. This optimized model demonstrated significant improvement, reducing the difference in the description of exposure for the microdose study from 5- to 1-fold, without compromising the description at therapeutic doses. This modified model has a wider range of applicability across doses and could support the extrapolation of microdose study results to therapeutic doses in future research.