S65487 PK/PD model predicts 97.6% Bcl2/Bim complex reduction and reconciles microdose/therapeutic data

Targeting anti-apoptotic proteins like Bcl2 is a crucial strategy in treating hematological malignancies such as Acute Myeloid Leukemia (AML). However, developing drugs with complex pharmacokinetics (PK), especially those exhibiting nonlinearity, presents significant challenges. Reconciling PK data obtained from early-stage microdose studies in healthy volunteers with therapeutic dose data from cancer patients is vital for accurate dose extrapolation and efficient drug development. This gap often leads to discrepancies, hindering the prediction of drug behavior and efficacy at clinically relevant concentrations.

Researchers developed a pharmacokinetic/pharmacodynamic (PK/PD) model to assess S65487's disruption of the Bcl2/Bim complex. The initial model was a linear 3-compartment pharmacokinetic model linked to an indirect response model, featuring concentration-dependent inhibition of Bcl2/Bim complex formation. This model utilized characteristics from phase I/II AML patients. Separately, a microdose study was conducted in healthy volunteers. To reconcile the substantial differences observed between microdose and therapeutic dose PK, a second, 'microdose optimized pharmacokinetic model' was developed. This refined model incorporated a Michaelis-Menten approximation of target-mediated drug disposition, applied to both elimination and distribution processes, to account for saturable pathways.

The initial PK/PD model predicted a substantial 97.6% maximum reduction in Bcl2/Bim complex formation with 1200 mg of S65487, following a sustained administration schedule in phase I/II AML patients. However, pharmacokinetic parameters estimated from microdose administration in healthy volunteers differed substantially from those at therapeutic doses, with S65487 exhibiting slight nonlinearity. The subsequent 'microdose optimized pharmacokinetic model', which accounted for nonlinearity in drug disposition via a Michaelis-Menten approximation, significantly improved data reconciliation. This optimized model reduced the discrepancy in describing exposure for the microdose study from a 5-fold difference to a 1-fold difference, without compromising the accuracy of description at therapeutic doses. This demonstrates the model's enhanced ability to unify PK data across a wide range of doses.

The optimized model reduced the difference in microdose exposure description from 5-fold to 1-fold, while maintaining accuracy at therapeutic doses.