Flat-Dosing Versus BSA-Based Dosing for MLN9708, An Investigational Proteasome Inhibitor: Population Pharmacokinetic (PK) Analysis of Pooled Data From 4 Phase-1 Studies

Investigational agent MLN9708 is a potent, reversible and specific 20S proteasome inhibitor. Both intravenous (IV) and oral administration are being studied on a twice-weekly (Days 1, 4, 8, and 11; 21-day cycles) and weekly (Days 1, 8, and 15; 28-day cycles) schedule. The twice-weekly doses studied are 0.125–2.34 mg/m² for IV and 0.24–2.23 mg/m² for oral dosing. Weekly doses tested are 0.125–1.4 mg/m² for IV and 0.24–3.95 mg/m² for oral dosing. MLN9708 completely hydrolyzes to the pharmacologically active MLN2238 in aqueous solutions. Here we report a population PK analysis of pts with solid tumors, lymphoma or multiple myeloma (MM) enrolled in 4 ongoing phase 1 studies of IV and oral dosing on a twice-weekly (Days 1, 4, 8, 11; 21-day cycle) and weekly (Days 1, 8, 15; 28-day cycle) schedule.

MLN9708 dose was based on body surface area (BSA) for both twice-weekly and weekly dosing schedules in solid tumor, lymphoma and MM patients (N=85). MLN2238 concentration-time data were analyzed using a non-linear mixed effects modeling approach (NONMEM^® VII), with Intel^® Visual Fortran Compiler, and the output was processed with SPLUS v8.1. Compartmental PK models were coded using the ADVAN 12 subroutine of NONMEM. A log-transform both sides approach was used. Estimation was by the First Order Conditional Estimation (FOCE) method with eta-epsilon interaction. Covariates tested included body weight, BSA, age, sex, dose and albumin on both clearance (CL) and volume, while creatinine clearance (CrCl), bilirubin, AST and ALT were only tested on CL parameters. Model validation and robustness were assessed using bootstrap simulations (N=1000) and visual predictive checks (N=100).

Mean age of 85 pts was 60.7 years, mean body weight was 80.0 kg, 87% of pts were Caucasian, and 58% were male. MLN9708 population PK was described by a three-compartment model with first order elimination. All model parameters were estimated with adequate precision when IV and oral data were fitted together. Of all the covariates tested, only body size descriptors (body weight and BSA) were found to be a significant covariate only on peripheral volume of distribution, V3 (p<0.01). Since BSA was more significant than body weight it was kept in the final model. Mean [%CV] CL (1.86 [7.0] L/h) and central volume of distribution (V2, 11.7 [13.1] L) were not affected by body weight or BSA. Inter-individual variability (IIV) was approximately 45% for CL, 82% for V2 and 30% for peripheral volume of distribution, with 17% of IIV in V3 explained by BSA. Absolute bioavailability (F) was estimated to be 62% based on this analysis. Based on simulations using the final population PK model, there was no difference in concentration-time profile or exposures (AUC or C_max) between the BSA-based and the flat doses.

Population PK analysis showed that the plasma concentration-time profile of MLN2238 can be well described by a three-compartment model with first order elimination process. Body size did not significantly impact AUC or C_max, supporting a switch from BSA-based dosing to flat dosing in all ongoing and planned clinical studies. Ability to use a flat dosing paradigm may be a benefit for oral administration of MLN9708 due to its potential to simplify clinical management.

Gupta:Millennium Pharmaceuticals, Inc.: Employment. Off Label Use: Investigational agent in clinical development for the treatment of solid tumors, lymphoma and MM. Saleh:Millennium Pharmaceuticals, Inc.: Student internship. Venkatakrishnan:Millennium Pharmaceuticals, Inc.: Employment.