Correlation of T cell fitness of apheresis products with response to CAR T-cell therapy in patients with relapsed or refractory large B-cell lymphoma: A prospective observational study at a single center

Abstract

Abstract Background CD19-directed chimeric antigen receptor (CAR) T-cells have revolutionized the treatment of relapsed/refractory large B-cell lymphoma (r/r LBCL). However, long-term remissions are observed in less than half of patients. Mechanisms of resistance include tumor intrinsic factors, microenvironmental factors, and CAR T-cell dysfunction. A single-cell RNA sequencing analysis revealed that memory and exhausted signatures of CAR T-cells are associated with the prognosis (Nat Med. 2020;26:1878-87). Furthermore, the quality of apheresed T cells, which represent the starting material for CAR T-cell manufacture, has also been noted to be critical (Nat. Med. 2018; 24:563–71). In this context, it would be valuable to identify the cellular features of the apheresis product that predict therapeutic response. Methods We conducted a single-center, prospective observational study of patients with r/r LBCL who were treated with CAR T-cell therapy between January 2023 and December 2024, and performed detailed immune profiling of apheresed T cells using multicolor flow cytometry. To characterize T cells based on their differentiation and exhaustion status, we examined the expression of CD45RA, CCR7, and CD95, which are commonly used to identify naïve and memory T cells (Nat Med. 2011;17:1290-7), as well as TIGIT and PD-1, which have recently been identified as markers for discriminating between functional and exhausted stem-like memory T (TSCM) cells (Nat Immunol. 2020;21:1552-62). We also analyzed the impact of cumulative chemotherapy on T-cell subsets in lymphoma patients treated at our center. Results First, we analyzed the initial 13 consecutive cases treated in 2023, including 4 cases with durable remission and 9 refractory/relapsed cases.T cells from these cases were downsampled and concatenated, giving a total of 144,000 cells. Dimensionality reduction analysis using t-SNE and algorithm-assisted clustering yielded 18 unique T-cell clusters. Three clusters (C1, C2, and C3) were observed more frequently in cases with durable remission than in refractory/relapsed cases (C1: 5.39% vs 1.31%; C2: 17.4% vs 1.51%; C3: 7.22% vs 3.26%). C1 and C2 exhibited characteristics of CD8- or CD4-naïve T (TN) cells (CD45RA+CCR7+CD95-TIGIT-PD-1-), respectively, while C3 exhibited characteristics reminiscent of functional CD4-TSCM cells (CCR7+CD95+TIGIT-PD-1-). We then focused on the TN and TSCM subsets and applied manual gating to identify them in the apheresed samples from 27 patients who had been followed up for at least 6 months, and compared the frequencies of these subsets in cases with durable remission (n = 16) with those in refractory/relapsed cases (n = 11). Both CD4 (TN + TSCM) and CD8 (TN + TSCM) cells were more frequently observed in cases with durable remission (CD4: median (IQR): 14.1% (6.83-20.4%) vs 4.5% (3.7-7.57%), p = 0.01; CD8: median (IQR): 5.0% (2.83-6.5%) vs 0.88% (0.86-1.51%), p = 0.054, Welch's t-test). These results were consistent among patients who achieved CR or PR at CAR T-cell infusion, indicating that these subsets are highly associated with an improved response to CAR T-cell therapy. Finally, we examined the effect of chemotherapy on these subsets within another cohort of lymphoma patients treated at our center. Of the 18 patients treated with bendamustine-based regimens, the median frequency of CD4 (TN + TSCM) decreased from 24.4% of total T cells before treatment to 1.36% after 2 cycles, 1.30% at the end of treatment (EOT), and 0.92% 6 months after EOT. Similarly, the median frequency of CD8 (TN + TSCM) decreased from 3.44% to 0.46%, 0.51%, and 1.0%, respectively. This is in stark contrast to 33 patients treated with CHOP-based regimens: the median CD4 (TN + TSCM) frequency was 26.7% before treatment, 27.6% after 2 cycles, 20.2% at EOT, and 24.5% 6 months after EOT; the median CD8 (TN + TSCM) frequency was 3.07%, 3.04%, 2.68%, and 4.5%, respectively. Given the established finding that bendamustine treatment before apheresis negatively affects outcomes following CAR T-cell therapy, these results suggest that TN and TSCM subsets could serve as useful indicators of immune fitness. Conclusions Our data further support the importance of the T-cell composition of apheresis products, and highlight their potential utility as pretreatment biomarkers. These can be easily identified using a simple flow cytometry analysis and could be employed to guide the management of patients eligible for CAR T-cell therapy.