(M-017) A Semi-mechanistic Pharmacokinetic-Pharmacodynamic/Toxicodynamic Model to Guide Dose Optimization of Combination Therapies Targeting DNA Damage Response Pathways

Derek Bartlett, PhD – Assistant Professor, Division of Pharmacotherapy and Experimental Therapeutics, UNC Eshelman School of Pharmacy, University of North Carolina at Chapel Hill

Objectives: Poly(ADP-ribose) polymerase inhibitors (PARPi) target the DNA damage response and show synthetic lethality in homologous recombination deficient (HRD) tumors. The combination of PARPi and DNA-damaging chemotherapeutics has been pursued as a strategy to overcome PARPi resistance and to enhance efficacy in homologous recombination proficient (HRP) tumors. While preclinical studies have demonstrated synergistic efficacy of these combinations, their success has been limited clinically by toxicities including myelosuppression. Therefore, we have developed a pharmacokinetic (PK)-pharmacodynamic (PD)/toxicodynamic (TD) model to guide dose optimization of PARPi combination therapies through simultaneous consideration of anti-tumor efficacy and myelotoxicity.

Methods:  The PK-PD/TD model was constructed in MATLAB SimBiology and calibrated using previously published preclinical and clinical data. Mouse- and human-specific PK parameters for PARPi (olaparib, veliparib, and talazoparib) and chemotherapy (temozolomide) were estimated using published plasma and tumor exposure profiles. PK models were linked to a tumor PD model incorporating DNA damage-driven cell death and a bone marrow TD model modified from previously published myelosuppression models. Mechanistic representation of both PARP inhibition and trapping enabled differentiation among PARPi compounds. The integrated PK-PD/TD model framework was then used to simulate monotherapy and combination therapy regimens in virtual mouse and human patients.

Results: The tumor PD model incorporating the reversible exchange between healthy and damaged cell states due to DNA damage and repair provided good concordance with published mouse tumor xenograft efficacy data. PARPi monotherapy showed anti-tumor efficacy in HRD but not HRP tumors, while the combination of PARPi and chemotherapy demonstrated anti-tumor efficacy in both HRD and HRP tumors. An analogous representation of healthy and damaged bone marrow cells captured drug-induced myelotoxicity, and TD-specific parameters were calibrated using in vitro bone marrow toxicity data. Simulations using human PK parameters and in vitro-derived TD parameters showed good agreement with longitudinal neutrophil counts in humans following temozolomide treatment. Various dose levels, intervals, and frequencies were explored through simulations to identify combination regimens that could achieve tumor regression while avoiding severe myelotoxicity. Preliminary simulations confirmed that olaparib and temozolomide given together at standard dosing regimens can result in severe myelotoxicity, while a reduction in temozolomide dose level or frequency can maintain anti-tumor efficacy without inducing severe myelotoxicity.

Conclusions: A semi-mechanistic PK-PD/TD model was developed to facilitate dose optimization for combination therapies involving PARPi and DNA-damaging chemotherapeutics through simultaneous consideration of anti-tumor efficacy and myelotoxicity.

Citations: /