Immunophenotypic CAR-T Cells Heterogeneity of Pre- and Post-Infusion of Tisagenlecleucel in Pediatric Patients with BCP-ALL.

Introduction Variability in the CD19 CAR-T cell infusion product may impact the efficacy and toxicity of this therapy. Objectives This study reports an in-depth immunophenotypic characterization of FDA-approved CD19-CAR T cells (tisagenlecleucel, Kymriah®, Novartis), pre and post infusion, among pediatric patients with BCP-ALL. Methods CAR+ cell heterogeneity in infusion products and post-infusion blood samples collected at days 7 and 28 after tisagenlecleucel infusion was evaluated using cytometry by time of flight (CyTOF). The tSNE-CUDA algorithm was implemented to reduce high-parameter data for analysis. Paired Exact Wilcoxon tests with FDR p value adjustment were performed to compare cell subset proportions before and after infusion. Results All pediatric BCP-ALL patients aged 2-16 years who received tisagenlecleucel between September 2022 and March 2023 in Poland (n=9 female to male 3/6) were included in this study.CAR expression ranged from 8.5 to 38.1% of all T cell subsets (mean 19.6%, SD=8.7, n=9), with 95-99% T cell purity in the pre-infusion tisagenlecleucel products. Significant donor variability in T cell subset composition among CAR+ and CAR– cells was apparent (Figure 1) In addition, pre-infusion products exhibited high proportions of CAR-positive T-regs (range: 38-72%) with a notable absence of Naïve T cells, as well as a predominance of central memory and effector memory subsets when compared to post-infusion profiles.Among CAR-positive T cell subpopulations in patient blood, T-regs were neither detected at day 7 nor day 28 post infusion (day 7: p=0.001; day 28: p=0.011). Similarly, the percentage of several phenotypic subsets significantly decreased in post-infusion blood samples, including Th2-like (day 7: p=0.011; day 28: p=0.015), Th17-like (day 7: p=0.007), CD4 total (day 7: p=0.001; day 28: p=0.012), CD4 Central Memory (day 7: p=0.001; day 28: p=0.012) and CD4 Terminal Effector (day 7: p=0,001; day 28: p=0.02) and CD8 Terminal Effector cells (day 7: p=0,001; day 28: p=0.011). There was a concomitant increase in the proportion of CD8 total (p=0.02), CD8 Naïve (p=0.011; day 28: p=0.011) and CD8 Central Memory (p=0.013) T cells between the infusion product and day 7 post infusion (Figure 2). We did not observe significant differences in proportions of CD8 Effector Memory, CD4 Naïve, CD4 Effector Memory, Th1like and CD4-CD8- cells across time. Conclusions We confirm significant heterogeneity of CAR-positive T cell subpopulations in infusion tisagenlecleucel products. After infusion, phenotypic subsets within the CAR-positive T cell compartment changed dramatically, with no detectable of CAR+ Treg cells, a significant decrease of CAR+ CD4 cells, and an increase of CD8 cells within one to four weeks from treatment. Observed heterogeneity of CAR+ T cell subpopulations should be further studied in the context of clinical response and toxicity.