Preclinical evaluation of the dose-concentration-marker-tumor growth relationship of a new pro-apoptotic compound using population PK-PD modeling
Pierrillas P, Hénin E, Bouzom F, Tod M
[1] Université Claude Bernard, Lyon, France ; EMR 3738 Ciblage thérapeutique en oncologie; [2] Technologie Servier, Orléans, France
Objectives: A new compound (S), in non-clinical settings, has been designed as a pharmacological tool to restore apoptosis functions. Preliminary analyses revealed nonlinearity in the relationship between dose, exposure, biomarker and tumor size. Based on data issued from in vitro experiments and in vivo studies in mice, the aim of this work was to build a semi-mechanistic PK-PD model to characterize the relationships between dose, plasma concentration, caspase activity and tumor size.
Methods: Data from preclinical studies evaluating S, administered either intravenously or per os, in xenograft mice were considered: 218 measures of concentrations (after single and repeated administrations), 54 of caspase activity and 64 measures of tumor size were modeled simultaneously. The pro-apoptotic effect was defined as an enhanced activation of caspase. The effective concentration was linked to plasma concentration, considering the unbound fraction and target affinity, derived from in vitro experiments. Tumor growth inhibition was then linked to caspase activity. Parameters were estimated using NONMEM 7.3 and model development was guided by residual- and simulation-based diagnostics.
Results: PK, caspase kinetics and tumor dynamics were successfully characterized by the proposed PK-PD model: the non-linear plasma pharmacokinetics was best described by a two-compartment disposition model with both saturable absorption and elimination. Auto-induction of elimination was characterized by stimulation of metabolism of S [1]; effective concentrations were expressed from the unbound plasma concentrations, through an interface model, defining a threshold level [2]; inhibition of the pharmacological target was resulting from a receptor occupancy model triggered by the effective concentration; caspase activity was modeled as an indirect effect model, whose production is inhibited by the target; tumor growth dynamics was modeled by a bi-phasic model [3],inhibited by an all-or-nothing effect of caspase. Model evaluation by goodness-of-fit and Visual Predictive Check, were satisfactory.
Conclusions: A semi-mechanistic approach, based on experimental mice data and in vitro parameters provides an interesting tool to quantify the expected antitumor effects and to propose an optimal dosing regimen in mice. In further steps, the pharmacodynamic model will be linked to a physiologically based pharmacokinetic model, to provide an extrapolable model for S to other species.
References:
[1] Hassan, M., et al., A mechanism-based pharmacokinetic-enzyme model for cyclophosphamide autoinduction in breast cancer patients. Br J Clin Pharmacol, 1999. 48(5): p. 669-77.
[2] Meille, C., et al., An interface model for dosage adjustment connects hematotoxicity to pharmacokinetics. J Pharmacokinet Pharmacodyn, 2008. 35(6): p. 619-33.
[3] Koch, G., et al., Modeling of tumor growth and anticancer effects of combination therapy. J Pharmacokinet Pharmacodyn, 2009. 36(2): p. 179-97.