Population PKPD modeling of dose-response and time course of peripheral lymphocytes after single and repeated administration of the S1P1/5 modulator, BAF312, in healthy volunteers.
E. Pigeolet, O. Luttringer, E. Wallstroem, E. Legangneux and B. Hamren
Novartis Pharma AG, Modeling & Simulation
Objectives: BAF312 is a next generation S1P receptor modulator, selectively targeting S1P1 and S1P5. The drug reduces the peripheral blood lymphocytes count by retaining cells into the lymph nodes. This poster presents a model based estimation of the dose-response of BAF312 on this efficacy biomarker in a population of more than 200 healthy volunteers.
Methods: A sequential population PKPD model was built in NONMEM using data from three healthy volunteer studies. All studies were single center, randomized, double-blind and placebo controlled. The first one was a single ascending dose study with dose levels ranging from 0.1 to 75 mg and the two others were multiple ascending doses studies covering a dose range from 0.25 to 20 mg. A one and two-compartment PK model with different absorption structures were fitted to BAF312 concentration data. An indirect response model with inhibition of lymphocyte input rate in the blood linked to BAF312 concentrations was fitted to the peripheral lymphocyte count data. Time course of drug effect was modeled using a parameterization involving the rate of lymphoyte input (Kin), baseline lymphocyte count (Base), maximum inhibitory effect of the drug on Kin (Imax) and the drug concentration that produces 50% of the maximum response (EC50). No covariate analysis was performed.
Results: PK was adequately described by a 2 compartment model with a combined zero and first order absorption process. The precision of the parameter estimations was very good (less than 6%). The peripheral lymphocyte count time profiles were adequately described by an indirect response model with an inhibiting effect on lymphocyte input rate. The inhibitory effect is non-linear with respect to drug concentration in the plasma and can be described using a sigmoid Imax type model. The prescision of the parameter estimations was very good (less than 7%). The inter-individual variability in baseline lymphocyte count is small (19%) and that for IC50 and Imax is moderate (52 and 58% respectively).
Conclusions: This model based analysis very well characterized the time course of the drug effect for healthy volunteers receiving a wide range of doses. Simulations from the model can guide the development team to select doses to reach a pre-specified peripheral lymphocyte count. Further work is needed to define the potential translation into clinical efficacy.