Single-cell dissection of CAR-T and blinatumomab-elicited T cell states uncovers therapy-specific functional decay patterns

Abstract

Abstract Introduction Chimeric antigen receptor T-cells targeting CD19 (CD19 CAR-T) and blinatumomab represent two leading immunotherapeutic approaches for relapsed/refractory (r/r) B-cell acute lymphoblastic leukemia (B-ALL). Even though both therapies target CD19, outcomes vary significantly. Our prior clinical observations demonstrated superior efficacy of CD19 CAR-T therapy compared to blinatumomab in this patient population. Here, we focus on the divergent functional states of CAR-T and blinatumomab-activated T cells following tumor engagement, aiming to elucidate the mechanistic basis for their differential therapeutic efficacy. Methods To characterize the post-killing features of CAR-T cells and blinatumomab-elicited T cells, we performed single-cell RNA sequencing (scRNA-seq) on pure CAR-T cells (CD3+CAR+, isolated from patients after CAR-T infusion) and endogenous T cells (CD3+, obtained from patients following blinatumomab treatment), both collected from fresh peripheral blood two weeks post-infusion. Additionally, we performed a direct in vitro comparison of blinatumomab and CAR-T cell therapies under controlled laboratory conditions and assessed the phenotypic profiles of CAR-T cells and blinatumomab-activated T cells using flow cytometry. Transduction was performed on primary T cells from r/r B-ALL patients using lentiviral vectors encoding for CD19 CAR constructs. The same patients' untransduced T cells (UNT) were used for blinatumomab therapy. Results For scRNA-seq, we sequenced a total of 102,548 cells from 9 patients, comprising: (i) a CAR-T group (n=3) who achieved complete remission (CR) after CD19 CAR-T infusion, and (ii) a BiTE group (n=6) treated with blinatumomab, of which 3 attained CR and 3 showed no response (NR). The clustering result showed that T cells in the two groups were distributed across several distinct clusters, and the proportion of each subset was different. During peak expansion at 2 weeks post-infusion, effector T cells predominated in both groups, though with marked differences: CAR-T samples contained 89.9% effector cells versus 61.3% in BiTE-treated samples, suggesting more focused effector polarization in CAR-T populations. Gene expression analysis demonstrated that BiTE_NR samples showed significant enrichment of T cell dysfunction markers (CD38, PDCD1, TIGIT, LAG3), while CAR-T samples exhibited elevated expression of cytotoxic effector molecules (GZMK, GNLY, CD27, JUN, NFKBIA). Protein-level characterization (antibody-derived tags, ADT) identified TIGIT+CD69+ T cell expansion in the BiTE group, particularly in BiTE non-responders. Transcriptomic profiling of these cells via GSEA revealed significant suppression of hypoxia (NES=-1.98, FDR<0.0001) and mTORC1 signaling (NES=-2.16, FDR<0.0001) in CR versus NR patients, suggesting impaired metabolic adaptation may underlie treatment resistance. For in vitro functional comparison, CAR-T cells and UNT cells (activated with 50 ng/mL blinatumomab) were co-cultured with Nalm6 cells at an effector-to-target (E:T) ratio of 1:1 for 24 hours, followed by flow cytometric analysis. While baseline expression of CD62L, PD-1, TIM-3 and LAG-3 showed no significant differences between the two groups prior to co-culture, post-stimulation analysis revealed distinct phenotypic profiles: CAR-T cells maintained significantly higher CD62L expression (p=0.0106) and demonstrated lower levels of exhaustion markers PD-1 (p=0.0045) and TIM-3 (p=0.0451) compared to blinatumomab-activated UNT cells. Specific lysis assays showed superior cytotoxic activity of CAR-T cells compared to blinatumomab-elicited T cells against Nalm6 targets during 72-hour co-culture at low E:T ratios (1:40), suggesting enhanced tumor-killing capacity under sustained antigen exposure (CAR-T: 80.98±1.76% vs BiTE: 50.92±2.02%, p<0.0001). Conclusion Our data suggest that CAR-T cells retain robust effector function and sustained cytotoxicity following tumor engagement, whereas blinatumomab-elicited T cells display heightened exhaustion markers and reduced proliferative capacity post-killing. The differential exhaustion profiles and functional divergence may contribute to the more durable responses seen in CAR-T-treated patients compared to those receiving blinatumomab.