(W-059) Facilitating RP2D decisions for CAR T cell therapy clinical studies using a QSP model

Clifton Anderson, n/a – Scientist II, QSP, Simulations Plus, Inc.; Akintunde Bello, n/a – Senior Vice President, Bristol Myers Squibb; Steven Chang, n/a – President, QSP Pittsburgh, Simulations Plus, Inc.; Yanping Chen, n/a – n/a, Bristol Myers Squibb; Chris Davis, n/a – Executive Director, Bristol Myers Squibb; Kimberly Harrington, n/a – Principal Scientist, Bristol Myers Squibb; Hongxiang Hu, n/a – Senior Research Investigator, Bristol Myers Squibb; Allison Kaeding, n/a – Senior Director, Bristol Myers Squibb; Lisa Kelly, n/a – Director, Bristol Myers Squibb; Meng Li, n/a – n/a, Bristol Myers Squibb; Yan Li, n/a – Senior Director, Bristol Myers Squibb; Kevin McCormick, n/a – Scientist II, QSP, Simulations Plus, Inc.; Brian Schmidt, n/a – Executive Director, Bristol Myers Squibb; Safiyyah Ziyad, n/a – Scientific Associate Director, Bristol Myers Squibb

Objectives
The objective of this study is to develop a multiple myeloma (MM) quantitative systems pharmacology (QSP) model that characterizes the disease's pathophysiology and predicts patient responses to therapies of interest. The model’s utility extends to enabling virtual clinical studies and facilitating recommended Phase 2 dosing (RP2D) decisions for clinical studies of chimeric antigen receptor (CAR) T cell therapies in MM.

Methods
We developed a QSP disease platform model of MM, which recapitulates disease states while integrating the dynamic interplay of various cell types, cytokines, and therapeutic interventions [1]. The model was expanded to incorporate BCMA- and GPRC5D-targeted CAR T cell therapies, integrating different mechanisms for CAR T expansion and efficacy. It delineates CAR T cell subset dynamics, the influence of target antigen (Ag) expression, and the formation of CAR:Ag receptor complexes. Furthermore, it considers the effects of lymphodepletion on the availability of mitogenic cytokines. The model also incorporates CAR T therapy-induced cytokine production as a safety signal to indicate stopping criteria during the escalation of CAR T therapy doses.

Results
The developed QSP model mechanistically describes MM pathophysiology and predicts patient outcomes to multiple CAR T cell therapies. Here we present a use case for the GPRC5D-targeted CAR T cell therapy (BMS-986393) where the model facilitated the RP2D dosing decision. The model was informed with pre-clinical data and fit to cellular kinetics data, tumor associated biomarkers, and endpoints from the GPRC5D CAR T trial. The therapy dose response was calibrated by fitting to the pre-infused CAR T drug product, immune cell profiling, CAR T cell PK, biomarker, cytokine, and clinical outcome data from other CAR T therapy trials; initial results were also included from the ongoing GPRC5D trial (CC-95266-MM-001). Cytokine risk was assessed by calibrating in vitro cytokine induction experiments and in vivo cytokine spikes post-infusion. These results include ORR and the best overall response (BOR), transient cytokine induction, and serum free light chain (sFLC) and serum monoclonal (M) protein changes in select cohorts. Clinical outcome data withheld for validation of the model’s prediction included several dose cohorts from the CC-95266-MM-001 trial. The GPRC5D CAR T therapy efficacy was predicted to increase until reaching saturation at a certain CAR T cell therapy dose. Cytokine-related risks continue to increase above this dose, suggesting that the RP2D CAR T cell dose maximizes efficacy while mitigating undesired cytokine induction.

Conclusions
The developed QSP model enables virtual screening of alternative regimens, assessing both efficacy and safety. It calibrates clinically relevant pathophysiology and drug mechanisms to various trial outcomes. Furthermore, it provides a platform for virtual clinical trials for CAR T cell therapies and their combinations.

Citations: [1] Anderson et al. "Optimizing clinical dose and scheduling for multiple myeloma therapies and combinations using a QSP model". ACoP14 November 5-8, 2023, National Harbor, MD.