Abstract 843: Development of a quantitative systems pharmacology model for clinical dose and schedule optimization of RG6234, a T-cell engaging antibody targeting GPRC5D in multiple myeloma

Abstract

Abstract Background: RG6234 is a GPRC5DxCD3 T-cell engaging bispecific antibody (TCB) that redirects T cells to target and eliminate cells expressing GPRC5D, including malignant plasma cells. RG6234 has a novel 2:1 (GPRC5D:CD3) configuration that confers bivalent binding to GPRC5D and increased T-cell directed killing compared with other molecular formats. RG6234 is initiated with Cycle 1 step-up dosing to mitigate the risk for severe cytokine release syndrome (CRS). Introduction: We performed an in silico evaluation of the dynamics of soluble B-cell maturation antigen (sBCMA), used as a surrogate for tumor burden and informing probability of response, and of the maximum release of IL-8, used as a surrogate for immune activation and informing the probability of CRS, using a QSP model. The model was calibrated with clinical data from 43 relapsed/refractory Multiple Myeloma (R/R MM) patients from the ongoing IV dose escalation study (NCT04557150). The model was set up to: 1) perform patient-specific calibrations and characterize the population with regard to patients’ sensitivity to tumor killing and immune activation and associated heterogeneity; and 2) simulate different dosing regimens in virtual populations and predict their probability of response and CRS. Here we present the model development and calibration results. Methods: The QSP model is a minimal mechanistic model integrating key elements of the Mechanism of Action (MoA) of RG6234. It describes immune activation by RG6234 and resulting MM cell killing. It comprises a system of two ordinary differential equations and 20 parameters, two of which are fitted to longitudinal clinical data (sBCMA and IL-8). Mechanistic assumptions regarding the MM disease and the MoA of RG6234 are represented in the model and supported by clinical or preclinical evidence. Of note, the immune tolerance observed in Cycle 1, indicated by the progressive decrease of cytokine peak levels despite the increased step-up dose level, is captured in the model by limiting the number of activated and proliferating T cells in the tumor microenvironment. Results: The calibrated model shows an accuracy of 92% in recapitulating Partial Response or better and of 78% in recapitulating CRS occurrence after the first step-up dose, demonstrating its appropriateness to address clinically relevant questions. Patient specific model calibrations show that the treated R/R MM population is more heterogeneous with regard to its sensitivity to RG6234-induced MM cell killing than to immune activation. Conclusions: The mechanistic model is able to simulate RG6234-induced T-cell mediated tumor cell killing and can be utilized to predict response and CRS in virtual populations after IV administrations at different dosing regimens. A model validation is planned with data from the expansion cohort of the study. Citation Format: Cristina C. Santini, Emilie Schindler, Jan Attig, Jan Eckmann, Suresh Vatakuti, Francesco Brizzi, Antoine Soubret, Sara Belli. Development of a quantitative systems pharmacology model for clinical dose and schedule optimization of RG6234, a T-cell engaging antibody targeting GPRC5D in multiple myeloma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 843.