Work

Apheresis Product Characteristics Predict Response to CD22 CAR T-Cell Therapy in Pediatric and Young Adult Patients with B-ALL

Year: 2024

Type: article

Abstract: Emerging data links characteristics of the apheresis product with outcomes following chimeric antigen receptor (CAR) T-cell therapy, but data on predictive markers in the pediatric B-cell acute lymphoblastic leukemia (B-ALL) population are limited. Given the ability to modify manufacturing and pre-collection therapies to optimize CAR T-cell functionality, identifying key biomarkers at the time of apheresis is imperative to guide decision making and improve outcomes. We retrospectively analyzed CD4/CD8-selected apheresis (T-cells) from children and young adults who were treated with CD22 CAR T-cells for B-ALL (NCT02315612). Analysis included flow cytometry for T-cell immunophenotype, activation, and exhaustion markers, Seahorse metabolic profiling, and RNA sequencing. We compared patients who achieved a complete response (CR) with those who had partial response, progressive, or stable disease (NR). Additional analysis compared patient T-cells to those of healthy donors. Results were analyzed using Mann-Whitney tests. Differential gene expression analysis identified genes with >2x fold change. Across 24 patient apheresis samples, 16 (67%) achieved CR and 8 (33%) did not (NR). Demographics, with exception of disease burden, were comparable. As expected, median pre-infusion disease burden was lower in patients with CR: 28% (r: 0.04-98.2%) vs 59% (r: 0.33-97.9) in NR (p=0.18). Analysis by UMAP and FlowSOM distinguished NR from CR by absence of activated, naive CD4 cells, and presence of CD4 effector cells. (Fig 1) Manual gating confirmed that patients with NR had lower CD4/CD8 ratios, fewer naïve (CD45RA+) CD4 T-cells, lower expression of CD28 and CD127 (IL-7R) (activation markers), and higher CD69. (Fig 2) RNA sequencing similarly showed lower IL-7R expression. Among patients with M3 marrow (>25% blasts) (n=15), those with NR (n=6) still had significantly fewer CD127+, CD8+CD28+, and CD4+CD45RA+ cells and more CD69+ cells compared to those with CR. Specifically, having <50% of CD4 cells expressing CD127 fully separated NR from CR in this M3 cohort. Compared to healthy donors, patients had more TIM3, LAG3, and CD39. However, those with CR resembled healthy donors in all markers that distinguished them from NR (Fig 3). Additionally, while healthy donors had higher ATP production and respiratory capacity and less glycolysis compared with B-ALL patients, metabolic characteristics did not distinguish CR from NR. Based on our analysis, starting material characteristics in B-ALL can predict NR, using a flow cytometric assay alone. This information can provide real-time guidance to patients, families, and providers on expectations. Additionally, CD4+CD127+ cells in the starting material appear to be necessary to overcome high disease burden, suggesting the importance of optimized manufacturing to expand this favorable cell subset.

Source: Transplantation and Cellular Therapy

DOI: 10.1016/j.jtct.2023.12.258