Population PK and PK/PD Analysis of Intravenous Investigational Agent MLN9708 in Solid Tumors and Lymphoma Patients
Neeraj Gupta, Karthik Venkatakrishnan
Clinical Pharmacology, Millennium Pharmaceuticals, Inc., Cambridge, MA, USA
Objectives: MLN9708 is an investigational potent, reversible and specific 20S proteasome inhibitor in phase-1 and phase-1/2 clinical trials. The objectives of this analysis were to characterize the population pharmacokinetics of intravenous MLN9708 and explore the effect of body size on PK variability. A PK/PD (20S proteasome activity) model was also developed to describe the concentration-effect relationship for target inhibition in humans.
Methods: MLN9708 dose was based on body surface area (BSA) for both twice-weekly and weekly dosing schedules in solid tumor and lymphoma patients (N=42). Plasma concentrations and whole blood 20S proteasome activity data available after single and multiple doses for 168 hours were used for the analysis. Population PK and PK/PD data were analyzed using NONMEM VII. Various covariates including BSA, age, sex, and race were tested. Standard goodness-of-fit diagnostics, posterior predictive checks and bootstrap analysis were done to evaluate the adequacy of models.
Results: MLN9708 population pharmacokinetics was described by a three-compartment model with first order elimination. Body size descriptors (Body weight and BSA) were found to be a significant covariate only on peripheral volume of distribution (p<0.01). Since both BSA and body weight are correlated, BSA was kept in the final model for ease of clinical translatability of the results since BSA based dosing was used in phase-1 studies. The estimates of pharmacokinetic parameters for a typical individual were 2.15 L/hr for clearance, 5.4 L for central volume of distribution (V1) and 419 L for peripheral volume of distribution (V2). Inter-individual variability (IIV) was approximately 45% for CL, 46% for V1, and 39% for peripheral volume of distribution (V2), with only 20% of IIV in V2 explained by BSA. Based on simulations using the final pop PK model, there were no differences in concentration-time profiles or exposures (AUC or Cmax) between the BSA-based dose (1.76 mg/m2) and the flat-fixed dose (3.3 mg). PK/PD relationship was explained by an inhibitory sigmoidal Emax model. IC50 obtained from the model was 21.1 ng/mL (IIV=30%).
Conclusions: Population PK analysis showed that BSA did not significantly impact AUC or Cmax, supporting a switch from BSA-based dosing to flat-fixed dosing in clinical development. An inhibitory sigmoid Emax model adequately describes the concentration-20S activity relationship.