A Multiple‐Model‐Informed Drug‐Development Approach for Optimal Regimen Selection of an Oncolytic Virus in Combination With Pembrolizumab

ABSTRACT

The antitumor efficacy of an intratumoral injection of a genetically engineered oncolytic vaccinia virus carrying human IL‐7 and murine IL‐12 genes (hIL‐7/mIL‐12‐VV) was demonstrated in CT26.WT‐bearing mice. In the CT26.WT‐bearing mouse model, the efficacy of the combination of hIL‐7/mIL‐12‐VV plus the anti‐programmed cell death protein (PD)‐1 antibody was determined to be correlated with the timing of administration: greater efficacy was observed when hIL‐7/mIL‐12‐VV was administered before the anti‐PD‐1 agent instead of simultaneous administration. To identify an optimal dosing regimen for first‐in‐human clinical trials, a multiple model‐informed drug‐development (MIDD) approach was used through development of a quantitative systems pharmacology (QSP) model and an agent‐based model (ABM). All models were built and verified using available literature and preclinical study data. Multiple dosing scenarios were explored using virtual populations by altering the interval between hIL‐7/hIL‐12‐VV and pembrolizumab administration. In contrast with observations from preclinical studies, both the QSP and the ABM models demonstrated no antagonistic effect on the dose‐dependent antitumor efficacy of hIL‐7/hIL‐12‐VV by pembrolizumab in simulations of clinical therapy. Based on the MIDD strategy, it was recommended that the highest dose of hIL‐7/hIL‐12‐VV and pembrolizumab should be administered on the same day, but with pembrolizumab administration following hIL‐7/hIL‐12‐VV administration. Multiple different modeling approaches uniquely supported and informed the first‐in‐human clinical trial design by guiding the optimal dose and regimen selection.