CAR-T Cell Subsets and Immune Repertoire Are Associated with Immune-Related Adverse Events and Efficacy after CD19 CAR T-Cell Therapy in B Cell Lymphoma

Background . CAR T-cell therapy is a potentially curative treatment in relapsed or refractory B-cell lymphoma (BCL). However, most CAR T cell recipients develop immune-related adverse events (irAEs) and around 60% of patients have relapse. We aimed to evaluate the characteristics of CD19-CAR T product(s) expansion during treatment, T cell and immune suppressors contributions to CAR T efficacy and toxicity. Methods. Patients with BCL receiving standard-of-care CAR T-cell therapy were enrolled in an IRB approved study. We evaluated cryopreserved blood samples collected before lymphodepletion (LDC) and at days 2, 8, 15 and 1 month after CAR T cell infusion. Each sample was evaluated by flow cytometry coupled with Uniform Manifold Approximation Projection dimensionality reduction and PhenoGraph clustering. We surveyed the characteristics of CD19CAR+ and CD19CAR- T cell subpopulations, including differentiation (CD45RA, CD62L), activation (CD69, CD25, HLA-DR), proliferation (CD71) and fitness (PD1, Tim3, LAG3) of CD4+ and CD8+ subsets. We also analyzed immune suppressors: polymorphonuclear myeloid-derived suppressor cells (PMN-MDSC: CD11b+ CD33+ HLA-DR-/low IL-4R+ CD15+) and regulatory T cells (Treg: CD4+, CD25high, CD127low). Clinical response at 1 and 6 months, incidence and severity of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) were recorded. We tested putative association between CAR and non-CAR T cell subsets, reported as mean ± standard error, and clinical outcomes using mixed-model analyses with Geisser-Greenhouse correction and Mann-Whitney U test. Results. Forty-two patients (38 large BCL, 4 mantle cell lymphoma) were enrolled; 25 received axi-cel, 13 liso-cel and 4 brexu-cel. Median age was 62.7 (range 27-83) years, 31% were female. We identified 32 immune clusters including 2 CAR T cell subsets detected across all CAR T products: activated/proliferating CD4+ effector memory (EM)-CAR T cell and activated/proliferating CD8+ EM CAR T. These two CAR T cell subsets shared a peak expansion at 8 days post CAR T (CD4+ EM CAR T 4.3±0.9% total T cells; CD8+ EM CAR T 12.7±2.1% of total T cells). Twenty-nine patients (69%) experienced CRS [grade (Gr) 1-2]. Patients who developed CRS displayed a more robust CD4+ EM CAR T cells expansion 8 days post-CAR T (CRS 5.5±1.1% vs no CRS 1.2±0.3%, p<0.01). There was trend to association of CD4+ EM CAR T expansion with response to CAR T therapy at 1 month (CR/PR 4.5±1.1% vs PD 2.2±1.3%, p<0.1). The CAR T product (any grade CRS: axi-cel 76% vs liso-cel 46%, p<0.1) and higher prevalence of late activated CD4+ EM T cells pre-LDC (CRS 3.4±0.9% vs no CRS 1.9±0.4%, p<0.05) were both associated with a heightened risk of CRS. Conversely, patients who had higher immune suppressors prior to or 2-3 days post-CAR T developed no or lower grade CRS (PMN-MDSChigh 0.7±0.2 vs Treghigh 0.4±0.2 vs immune suppressorlow 1.5±1.3 average CRS grade; p<0.01) suggesting immune contexture may mitigate CRS risks. Nineteen patients (36.5%) developed ICANS (3 Gr1, 6 Gr2, 7 Gr3, 2 Gr4, 1 Gr5). Patients who developed ICANS had strong CD8+ EM CAR T cell expansion 8 days post CAR T (ICANS 17.5±3.8% vs no ICANS 7.9±1.3%, p<0.01). There was a trend that CD8+ EM CAR T cell expansion was associated with response to CAR T therapy at 1 month (CR/PR 14.4±2.5% vs PD 4.4±1.7%, p<0.1). Higher prevalence of late activated CD8+ EM T cell pre-LDC was also associated with an increased risk of ICANS (ICANS 3.7±0.8% vs no ICANS 1.7±0.5%, p<0.01). Conversely patients with a higher prevalence of resting (p<0.01) or early activated (p<0.05) CD8+ effector T cells pre-lymphodepletion were more likely to achieve a sustained complete response to CAR T-cell therapy at 6 months. Conclusions. Immune profiling of patients who received CAR T-cell therapy for BCL revealed two distinct CAR T subsets associated with response and irAEs associated with CD19 CAR T cell therapy. Overall, our data show that CD4+ EM CAR T cells expansion is associated with CRS and early response while CD8+ EM CAR T cells have more prominent roles in ICANS and sustained response. Furthermore, T cell fitness pre-LDC may be a key parameter behind the CAR T-cell therapy response/toxicity dichotomy with resting/early activated T cells linked to response and late activated T cells to irAEs.