Paired Analysis from Patients with Myeloma Elucidates Favorable CAR-T Manufacture and Phenotype, Starting from Mobilized Pre-Transplant Apheresis Vs. Non-Mobilized Apheresis after Four Prior Lines

Abstract

Introduction B-Cell Maturation Antigen (BCMA) targeting chimeric antigen receptor T-cells (CAR-T) have demonstrated unprecedented efficacy in patients with relapsed and refractory multiple myeloma, approved after 4 or more prior lines of therapy (PLOT). Given the known T-cell dysfunction that develops over the disease trajectory, autologous T-cells harvested after many PLOT may not be as biologically active as those collected earlier. Estimates suggest 93% of stored stem cell products collected from MM patients are never used (Liang et al, 2021). We sought to demonstrate the feasibility of using stored mobilized apheresis products to generate anti-BCMA CAR-T cells. Given the heterogeneity that can exist between MM patients, we evaluated the differences in starting T-cell phenotype, manufacturing characteristics, and CAR-T phenotype and function side by side using paired samples from the same patient, obtained prior to transplant or prior to commercial CAR-T. Methods We identified a cohort of paired samples from: 1) Mobilized unfractionated apheresis obtained prior to ASCT (preASCT); and 2) Apheresis for CAR-T manufacture after ≥4 PLOT (preCAR). The flow cytometry panel used for phenotyping included antibodies against: CD3, CD4, CD8, CCR7, CD45RO, CD39, TIGIT, PD1, LAG3 and TCF7. Second generation anti-aBCMA.4-1BB.CD3z (aBCMA.BBz) CAR-T cells were manufactured from CD3+ T cells negatively selected from either sample using a Pan T cell isolation kit (Miltenyi). Post manufacture flow cytometry was used to characterize immunophenotype and transduction efficiency. Doubling time was calculated using trypan blue exclusion on D+2 to D+5 after transduction. A BCMA-expressing myeloma cell line, MM.1S, was used to stimulate and evaluate cytotoxicity. Results In 10 patients with paired preASCT and preCAR samples, the median time from ASCT to CAR-T infusion was 4.2 years (range 2.5-12.5 years), all patients had received three drug induction regimens, and 50% had >1 line of combination therapy prior to ASCT. PreASCT samples ranged in collection from Nov 2008 - April 2019. All patients were triple class exposed prior to collection for CAR-T and treated with median 6 PLOT (range 4-13). Comparison of the T-cell composition in both apheresis samples, revealed a higher CD4:CD8 ratio in preASCT samples (p=0.006), and a higher proportion of stem central memory/SCM (CCR7+CD45RO-, p<0.05, Figure 1) in both CD4 and CD8 compartments. Additionally, in preASCT samples, patients with only 1 line of induction, exhibited a trend toward a higher proportion of SCM T-cells relative to more heavily treated patients. PreASCT CD4 (p=0.002) and CD8 (p=0.01) T-cells expressed more TCF7 compared with preCAR, however T-cells with an effector phenotype were proportionally higher in preCAR samples across both CD4 and CD8 compartments (CCR7-CD45RO-, p<0.05, Figure 1). PreCAR T-cells express higher levels of checkpoints (CP) PD1 and TIGIT, with dual positive CP more prevalent in CD4+ T-cells. CAR-T cells manufactured from preASCT samples expanded better (p=0.048) and also had increased CD4:CD8 ratios (p=0.011) as compared to preCAR, without significant differences in transduction efficiency, CAR density or viability. After 24h stimulation with BCMA+ MM.1S, CD8+ CAR-T cell phenotype was different between preASCT and preCAR products, with more SCM in preASCT (p=0.004) and more effector memory/EM (CCR7-CD45RO+) in preCAR (p=0.008); however, the killing capacity after 24h was similar for both CAR-T products at different E:T ratios. The CAR-T cells generated from T-cells preCAR (later in the disease course) expressed higher TIGIT, LAG3 and CD39 after stimulation, and more dual (p=0.027) or triple (p=0.027) positive CP than those from preASCT T-cells ( Figure 2). Conclusion This study suggests CAR-T cells manufactured from banked mobilized unfractionated stem cell product expand better during manufacture, yield an equally cytotoxic, less differentiated, and less exhausted product than those manufactured from later timepoints (preCAR samples). Importantly, we demonstrate two major findings: (1) that a possibly more efficacious anti-BCMA CAR-T product can be manufactured from cryopreserved, mobilized apheresis material regardless of storage time and (2) cryopreserved preASCT samples may expand autologous CAR-T cell therapy access to heavily treated patients with poor T cell quality. CLF & JAM contributed equally