Immune correlates of anti-BCMA CAR-T products idecabtagene vicleucel and ciltacabtagene autoleucel in a real-world cohort of patients with multiple myeloma

Abstract

We performed the first in-depth, comparative and prospective biomonitoring of Multiple Myeloma (MM) patients (N = 39) receiving ciltacabtagene autoleucel (cilta-cel) or idecabtagene vicleucel (ide-cel) chimeric antigen receptor T cells (CAR T) in the real-world setting. In cilta-cel patients response rates were higher and atypical neurotoxicities/infections more frequent. Peak CAR T counts were significantly higher in cilta-cel patients, driven by CD4⁺ CAR expansion, correlating with clinical responses. Expansion of cilta-cel cells was associated with higher CAR and CD27 expression while, in contrast to ide-cel, there was no correlation between TIM3 expression and CAR T proliferation. Cilta-cel CAR T expansion was followed by a CAR-specific switch from proliferation-associated genes to genes/surface markers indicating effector/memory function. The longer persistence of cilta-cel CAR T was associated with increased IL-7R expression; in vitro data showed persistent antigen-independent activation and higher metabolic activity of cilta-cel vs. ide-cel CAR T. Among cilta-cel-treated patients experiencing atypical neurotoxicities, central nervous system (CNS)-infiltrating, effector-type CAR T presented a distinct inflammatory phenotypic/cytokine-expression profile. This in-depth biomonitoring report following real-world cilta-cel or ide-cel highlights intrinsic biological differences between BCMA-targeting CAR T products, potentially explaining differences in clinical activity and toxicity. Our findings may guide optimization of cellular immunotherapy strategies in MM.

Fig. 1. Distinct expansion pattern of of cilta-cel vs. ide-cel CAR-T in myeloma patients in the real-world setting.

Time course of CAR T cell numbers expressed as T cells identified by surface expression of the CAR using CAR detection reagent and flow cytometry for (A) total CAR T, (B) CD4⁺ CAR T, (C) CD8⁺ CAR T, and (D) CD4⁺CD8⁺ CAR T and proportions in myeloma patients after lymphodepleting chemotherapy and CAR T cell infusion. Data are expressed as the respective CAR-T subtype per total number of T cells. Blue curves indicate cilta-cell numbers and red curves indicate ide-cell numbers. E Levels of CD4⁺ and CD8⁺ Non-CAR T cells in the patients’ blood at the time of apheresis for cilta-cel (blue; N = 15) vs. ide-cel (red; N = 11) patients. F Peak CD3⁺ CAR T levels, peak CD4⁺ CAR T levels, peak CD8⁺ CAR T levels, peak CD4⁺CD8⁺ CAR T levels, time to peak for CD3⁺ CAR T, time to peak for CD4⁺, and time to peak for CD8⁺CAR T for cilta-cel (blue; N = 23) vs. ide-cel (red; N = 16). G CAR T cell numbers at different timepoints post cell infusion for cilta-cel (blue) vs. ide-cel (red). H Surface expression levels of the respective BCMA CAR as measured by flow cytometry after staining with BCMA CAR detection reagent. The histogram shows expression levels on CD3⁺ CAR T in two exemplary patients (gray=unstained control, red=patient who received ide-cel, blue=patient who received cilta-cel). The bar graph on the right shows surface expression levels of the given CAR at peak in the two groups of patients (blue, N = 23; red=ide-cel, N = 16). Bar graphs indicate median values with 95% confidence intervals (CI). Statistical differences between groups were calculated using a two-sided Mann-Whitney U test. Source data are provided as a Source Data file.

Fig. 2. Different patterns of immune cell reconstitution in myeloma patients post cilta-cel vs. ide-cel CAR-T.

A Dots indicate absolute counts of white blood cells (WBC), neutrophils, lymphocytes and conventional T cells in patients post cilta-cel (blue; N = 23) or ide-cel (red; N = 16) at the time of cell infusion (D0) and the time of the patient’s individual CAR T peak. The dotted line and the solid lines indicate median baseline and peak levels, respectively. B Absolute CAR T cell counts in patients post cilta-cel (blue; N = 23) or ide-cel (red; N = 16) at the time of cell infusion (D0) and the time of the patient’s individual CAR T peak. C Bar graphs on the left show absolute peak levels of conventional T cells and bar graphs on the right show absolute peak levels of CAR T cells in patients post cilta-cel (blue; N = 23) or ide-cel (red; N = 16), respectively. D Orange bar graphs show absolute counts of CAR T across cilta-cel (N = 23) or ide-cel (N = 16) patients who did or did not receive steroids post CAR T infusion. All bar graphs indicate median values with 95% confidence intervals (CI). Statistical differences between groups were calculated using a two-sided Mann-Whitney U test. Source data are provided as a Source Data file.

Fig. 3. Treatment with cilta-cel or ide-cel and correlation with clinical characteristics and outcome.

A Bar graphs show absolute and relative dose levels received (blue = cilta-cel, N = 23; red = ide-cel, N = 16). Statistical differences between groups were calculated using a two-sided Mann-Whitney U test. B Depth of response at one and three months post-CAR-T, respectively. C Clinical course of patients receiving cilta-cel (blue) or ide-cel (red). D Progression-free survival (PFS) after treatment with cilta-cel or ide-cel. E Depth of response at 3 months post CAR T across both patient groups (N = 37) in relation to peak expansion levels of total, CD4 + , and CD8 + CAR-T. Bar graphs indicate median values with 95% confidence intervals (CI). Statistical differences between response categories were calculated using a Kruskal-Wallis test (one-way ANOVA on ranks). Source data are provided as a Source Data file.

Fig. 4. Rapid expansion of cilta-cel CAR T is associated with a CAR-specific shift towards effector-memory T cell subtypes.

A We investigated the distribution of certain memory subtypes (CM = of central memory; TSCM = T memory stem cells; EM = effector-memory; EMRA = terminal effector memory) the different CAR T products as defined by the expression of certain memory markers as indicated. All 4 T cell memory subtypes were investigated in both patient groups (blue = cilta-cel, N = 23; red = ide-cel, N = 16) at the time of the individual CAR T peak. Results are shown for (B) CD4⁺, (C) CD8⁺, (D) CD4⁺CD8⁺ CAR T cells and for (E) CD4⁺ non-CAR T from the same patient/timepoint. Bar graphs show percentages of the given memory subtype out of all CD4⁺, CD8⁺, CD4⁺CD8⁺ CAR T cells or CD4⁺ non-CAR T in patients post cilta-cel (blue, N = 23) or ide-cel (red, N = 16), respectively. Bar graphs indicate median values with 95% confidence intervals (CI). Statistical differences between groups were calculated using a two-sided Mann-Whitney U test. Source data are provided as a Source Data file.

Fig. 5. Cilta-cel, but not ide-cel, CAR T show a lack of a short-term correlation between TIM3 expression and proliferation with an upregulation of negative regulatory markers on long-term persisting cells.

To explore potential mechanisms underlying long-term persistence of cilta-cel CAR-T we determined levels of negative regulatory markers LAG3, TIM3, and PD1 on the CAR-T in both patient groups right after infusion into the patient and up to 360 days post-infusion. The highest levels of the negative regulatory markers for both (A) cilta-cel (blue) and (B) ide-cel (red) were observed on CD4⁺ and CD8⁺ CAR-T right after infusion followed by a decrease over the next few weeks. C Long-term persisting cilta-cel CAR-T showed a second “late” increase in all three negative regulatory markers, often referred to as exhaustion markers, at around one year after CAR-T infusion. Figures show mean fluorescence intensity (MFI) of the given marker on the given CAR-T subtype as measured by flow cytometry over time. The black lines indicate median values and statistical differences between groups were calculated using a two-sided Mann-Whitney U test. D Only in the ide-cel group (red), but not in the cilta-cel group (blue), the “early” peak in expression of negative regulatory marker TIM3 predicted a reduced CAR-T in vivo proliferation a few days later. E There was an even stronger effect when all three markers were combined. For correlative analyses a Pearson correlation coefficient was calculated. The black line shows the results of a linear regression analysis. Source data are provided as a Source Data file.

Fig. 6. Cilta-cel patients with atypical neurotoxicity show CNS infiltration by effector memory-type CART cells producing an inflammatory cytokine signature.

A Analysis of BCMA-targeted CAR T cell numbers in the peripheral blood (PB) and the cerebrospinal fluid (CSF) from the same timepoint after CART infusion in three myeloma patients who had developed unusual types of neurotoxicity around 3-4 weeks after cilta-cel infusion. Analyses were performed by flow cytometry following co-staining with anti-CD3 and BCMA CAR detection reagent which represents a fluorescent, full-length, recombinant BCMA protein binding to the CAR expressed on the cell surface. B CAR T cell subpopulations in the PB and CSF were determined following staining with anti-CD4 and anti-CD8 monoclonal antibodies. C Different CAR T memory subpopulations in the PB and CSF were identified by flow cytometry using co-staining with anti-CD45RA and anti-CD62L monoclonal antibodies. D The concentration of 22 cytokines/chemokines were measured in the PB and the CSF from the same sample of all three patients using CodePlex Secretome technology. Results are shown as absolute concentrations in pg/mL. Source data are provided as a Source Data file.