Cellular Composition and Clinical Factors Influencing Manufacturing Outcomes for Anti-BCMA CAR T Cell Therapy in a Large Real-World Cohort of Relapsed/Refractory Multiple Myeloma

Background

Leukapheresis is a crucial step for production of chimeric antigen receptor (CAR) T-cells, requiring sufficient numbers of autologous lymphocytes, competent to expand and execute effector functions. Previous therapies and remission status may impact these features. Repeated T-cell activation leads to downregulation of costimulatory factors CD27 and CD28, resulting in senescent, non-proliferating CD3+CD27-CD28- cells. So far, the impact of T cell senescence on production of anti-BCMA CAR T cells is unclear.

Aims

To evaluate factors influencing collection of autologous lymphocytes and subsequent manufacturing of idecabtagene vicleucel (ide-cel) or ciltacabtagene autoleucel (cilta-cel) in patients with refractory or relapsed multiple myeloma (RRMM) with a focus on cellular composition, esp. T-cell senescence, defined as a loss of CD27 and CD28 in a real-world setting.

Methods

We retrospectively analyzed the cellular composition of 84 collections of 61 RRMM patients intended to undergo treatment with ide-cel (N=36) or cilta-cel (N=48). The target was defined as ≥1x10⁹ CD3+ cells.

Cellular composition of collections was analyzed per flow cytometry. Optimal cutoff for CD3+CD27-CD28- yield was determined using receiver operating characteristic (ROC) analysis. Survival analysis was performed using the Kaplan-Meier method and the log-rank test. Progression free survival (PFS) was defined as time from CAR T cell infusion until death from any cause or MM progression, or last follow-up (FU).

Results

CD3+ cells yields were median 31.5x10⁸ (range, 4.9-318), with target reached in 73/84 collections. Ten patients (16%) had ≥1 collection attempt. 62 (74%) collections resulted in successful CAR T production, 3 (3%) production terminations and 19 (23%) out of specification (OOS) products, due to insufficient CAR T cell counts (n=7, 37%), insufficient proliferation (n=7, 37%), insufficient IFN-Gamma potency assay (n=2, 10%), or other reasons (n=3, 16%). There was no difference in proportions of manufacturing failures between ide-cel (9/27) or cilta-cel (13/35), p=0.99.

Patients with OOS showed higher median concentration of CD3+ cells in peripheral blood (pb) at apheresis vs successful productions with median 770/µl (range 82-4748) vs 462/µl (range 71-3616), p=0.002.

Median CD3+ yields in OOS products were higher than in successful productions with 39 x10⁸ (range 8-318) and 29x10⁸ (range 5-176) respectively (p=0.04). Furthermore, OOS productions showed significantly higher median CD8+ yields with 27x10⁸ (range 2-222) vs 15x10⁸ (range 2-121), p=0.01 and a trend towards lower median CD4+ yields with 9x10⁸ (range 4-82) vs 11x10⁸ (range 2-139), p=0.09. Noteworthy the median yields of senescent CD3+CD27-CD28- cells were higher in OOS with median 20x10⁸ (range 0.2-102) vs 2x10⁸ (range 0.1-63), p=0.001 and we detected a trend towards lower yields for CD3+CD27+CD28+ cells with median 10x10⁸ (range 5-136) vs 17x10⁸ (range 2-163), p=0.13. Using a cutoff of ≥17.4x10⁸ CD3+CD27-CD28- cells (AUC=0.72), positive predictive value was 0.52 (95%CI: 0.4-0.86) and negative predictive value 0.84 (95% CI: 0.8-0.97) for manufacturing failure.

Manufacturing failures were not associated with sex, age, adverse cytogenetics or remission status prior to apheresis, but the patients with ≥1 collections had significantly higher number of prior median therapy lines with 9 (range 5-13) vs 7 (range 1-14), p=0.04. The exposure to bispecific antibodies and the detection of replication of cytomegalovirus prior to apheresis were higher in OOS productions, but were not significant, p=0.13 and p=0.4 respectively.

Two cilta-cel and two ide-cel patients were infused with OOS products (insufficient cell number: n=3, insufficient IFN-Gamma potency: n=1), of which 3 achieved CR and 1 SD as best response after CAR T cell infusion. After a median follow-up of 10 months, there was no difference in PFS comparing to patients with first successful production (p=0.8).

Discussion

In our real-world cohort of RRMM patients intended to undergo treatment with anti-BCMA CAR T, higher CD3+ cell counts in pB, higher yields for CD3+ cells and CD8+ cells but importantly higher yields of senescent CD3+CD27-28- showed to be risk factors for OOS. However, no differences in clinical outcomes of patients infused with OOS products were observed. Further analyses are required for optimization of CAR T cell manufacturing in RRMM.