Early Leukapheresis in Patients with B-ALL Yields an Activated, Early Memory T-Cell Phenotype Associated with Response to CAR T-Cell Therapy

Introduction: Among patients with B-cell acute lymphoblastic leukemia (ALL) who are treated with CAR T-cell therapy, 20-30% experience primary non-response.1,2 One proposed mechanism of non-response is sub-optimal apheresis starting material, either by virtue of dysfunctional T-cells or the presence of other inhibitory cells, especially in patients who have received prior courses of chemotherapy. Early collection of apheresis products is one potential solution, but it is not clear whether unfavorable T-cell features are the result of chemotherapy or a response to leukemia itself. As earlier apheresis is now increasingly being performed to enhance T-cell fitness, understanding key characteristics of these products is imperative. Methods: Flow cytometry and Seahorse metabolic profiling were used to compare cryopreserved apheresis products across 3 groups: children, adolescents and young adults (CAYA) with relapsed/refractory B-ALL treated on a CD22 CAR T-cell trial (NCT02315612; n=30), CAYA with B-ALL who underwent early apheresis collection (“Early Collection” or EC=either at end of primary induction due to high risk features or after re-induction for relapse; n=6), and healthy donors (HD; n=8). Apheresis materials from trial patients were CD4/8-selected, whereas EC patient and healthy donor samples were mononuclear cells, with analysis gated on CD3 cells. Using T-cell characteristics previously found to distinguish patients who achieved a complete response (CR) from non-responders (NR) following CD22 CAR T-cells, we evaluated for these features in the HD and EC samples. Specifically, T-cell differentiation phenotypes based on CD45RA/CD62L expression, activation markers (CD127, CD28, CD69, and CD62L), and exhaustion markers (CD39, LAG3, TIM3, CTLA-4) were assessed. Results: The median age for CD22 CAR patients, EC patients and HD was 15, 15, and 29 years, respectively. In the CD22 CAR trial group, 73% had high baseline disease burden (>5% blasts in bone marrow) at the time of collection. In the EC group (n=6), 2 patients were collected after re-induction for first relapse, and 4 were collected post-induction after initial diagnosis. One of the relapsed EC patients had high disease burden at the time of collection, and one had isolated extramedullary disease. Of those collected at first diagnosis, 3 had MRD (minimal residual disease)-positive but low burden disease, and one was MRD negative but had experienced severe methotrexate toxicity during induction. Median absolute lymphocyte count at the time of collection for CD22 CAR patients versus EC patients was 660 cells/mcL versus 1385 cells/mcL. When healthy donors were compared to CD22 CAR patients, they had lower expression of TIM3 (0.4% HD vs 1.4% CD22 CAR; p=0.003), LAG3 (0.4% HD vs 0.9% CD22 CAR; p=0.0004), and CD39 (0.9% HD vs 10.9% CD22 CAR; p<0.0001) and more favorable metabolic features, with higher spare respiratory capacity (408% HD vs 346% CD22 CAR; p=0.04). Phenotype and activation markers in HD all resembled those of patients who achieved CR. In the EC group, there were no differences in surface marker expression compared to CD22 CAR patients who achieved CR. However, when compared with the CD22 NR group, EC samples had more stem central memory cells (33% vs 4% in NR, p=0.003), fewer effector memory CD4 cells (34% vs 70% in NR, p=0.008), and higher expression of the activation markers CD127 (57% EC vs 13% in NR, p=0.0008) and CD62L (36% in EC vs 19% in NR, p=0.0007). Expression of CD28 was also higher in the EC group (73% EC vs 37% NR, p=0.11). Compared with HD samples, the EC samples did not differ in T-cell phenotypes or activation markers, but had higher expression of the exhaustion markers TIM3 (0.9% vs 0.4% in HD, p=0.03) and CD39 (8.4% vs 0.9% HD, p=0.001). EC samples also had slightly more glycolysis (34% vs 31% in HD, p=0.02) and less spare respiratory capacity (306% vs 408% in HD, p=0.03) than HD. Conclusion: Although apheresis samples collected early in the leukemia course have some features of exhaustion that distinguish them from healthy donors, they overall express the favorable T-cell characteristics found to be associated with response to CAR T-cells. This preliminary data supports early collection of apheresis products for B-ALL patients with relapsed or high-risk leukemia to improve chances of successful CAR T-cell therapy if needed in the future. References: Fergusson et al. Front Immunol. 2023 Laetsch et al. J Clin Oncol. 2023