Response to Bcma CAR-T Cells Correlates with Pretreatment Target Antigen Density and Is Improved By Small Molecule Inhibition of Gamma Secretase

Introduction: The adoptive transfer of B-Cell Maturation Antigen (BCMA) chimeric antigen receptor (CAR) T cells is demonstrating early promise in multiple myeloma [MM], however durable responses remain elusive and most studies report >50% of patients relapsing within 18 months of receiving CAR-T cell therapy. The mechanism of relapse is likely the consequence of multiple factors including the variable distribution of BCMA on tumor cells, allowing cells with low antigen density to escape. Initial target density, receptor downregulation and the emergence of antigen loss variants have all been implicated in relapse following CAR-T cells directed against CD22 and CD19. Reduced or absent BCMA expression may similarly be linked to relapse in MM. We have previously demonstrated that BCMA cleavage by the γ-secretase complex reduces ligand density for CAR-T cell recognition, and that a small molecule γ-secretase inhibitor (GSI) markedly increases surface BCMA levels in a dose-dependent fashion while improving CAR-T cell recognition in preclinical models.

Methods and Results: In a phase I first-in-human study (NCT03338972) employing a CAR-T cell construct encoding a fully human BCMA scFv and 4-1BB/CD3z, rapid and deep objective responses at CAR-T cell doses starting at 5 x 10⁷ have been observed. All patients had bone marrow (BM) involvement at baseline (mean 42.5 % CD138⁺ by IHC) and 14/15 had no detectable disease in the BM 28 days after therapy. One patient with comparatively very low BCMA expression (BCMA antibody binding capacity [ABC; QuantiBRITE] = 269; 16.9% of the malignant plasma cells (PCs) BCMA⁺ by flow cytometry) was the only subject with persistent tumor cells in the BM 28 days after therapy. Despite complete BM responses in all remaining patients, late relapses have occurred. Differences in the BCMA expression level on tumor cells prior to CAR-T cells between long term responders and those with relapse are evident. Among the 12 subjects with at least 3 months of follow up, those remaining in remission (median 12 months, range 3-16; data cut off 7/15/19) demonstrated a median pre-treatment BCMA ABC of 1761 (range 781-2922, n=5), in contrast patients with relapse (mean of 7.3 months, range 2-12) had a median pre-treatment BCMA ABC of 920 (range 260-1540, n=7).

Six patients with a pretreatment mean ABC of 919 (range 260-1540) had BM evaluable for BCMA expression at relapse and the mean ABC decreased to 304 (range 121-519). The percent PCs expressing BCMA decreased from 77.5% (range 13 - 99.8) to 30% (range 10.4-60.4). The impact of gamma secretase inhibition on BCMA expression was assessed on BM cells obtained from a patient relapsing after BCMA CAR-T cells. At relapse a 9.5-fold decrease in ABC from baseline was observed. The cells were cultured for 5 hours in the presence of GSI (JSMD194) at a concentration of 1mM, which is readily achievable by oral administration. A significant increase in BCMA antigen expression was observed (ABC=917).

The impact of modulating BCMA expression on tumor cells by concurrently administering an oral GSI with CAR-T cells is being explored in a phase one clinical trial (NCT03502577). In this setting, the GSI has increased BCMA expression when low level residual BCMA was observed following relapse after prior BCMA therapy failure. Two patients have been evaluated for response to an JSMD194 after failing other BCMA targeted agents. One received a prior BCMA CAR-T cell product and after relapse demonstrated a BCMA ABC of 769. Target expression increased in this patient almost nine-fold to 6828 (ABC) after three oral doses of JSMD194. A second patient had a BCMA ABC of 666 after failing a BCMA bispecific T cell engager. BCMA density increased over 14-fold to 9583 after GSI. Comprehensive data from the combination GSI and BCMA CAR-T cell trial are being reported separately.

Conclusion: Pretreatment BCMA target density quantified with a uniform flow cytometry method of measurement and performed on all patients enrolled on a single center BCMA CAR-T cell clinical trial is associated with the durability of response. Further, BCMA expression can be significantly increased following GSI exposure in patients evidencing low BCMA ABC at baseline or when downregulation is the consequence of prior BCMA targeting therapy. The capacity for GSIs to increase BCMA target density and decrease soluble BCMA levels is a promising approach to be exploited in clinical trials.