Baseline T-cell fitness and subtype-specific CD19 loss as drivers of blinatumomab resistance in B-ALL

Abstract The introduction of Blinatumomab (BLIN), a bispecific T-cell engager targeting CD19, has marked a major advancement in B-ALL therapy, substantially improving remission rates and altering treatment paradigms. However, a considerable number of patients still fail to respond to BLIN, either due to loss of the target antigen or impaired T-cell fitness and exhaustion. Yet, predictive markers of T-cell functionality are largely missing. We retrospectively analyzed 72 cryopreserved bone marrow samples of 36 adult Ph-negative B-ALL patients (median age 35.5 years) before and after the first cycle of BLIN (day 29) using multiparametric spectral flow cytometry (24 markers), complemented by single-cell RNA sequencing (scRNASeq) in 20 patients. Most patients received BLIN according to GMALL protocols for persistence of measurable residual disease (MRD) following frontline intensive therapy blocks (n = 29), a smaller subset following relapse (n = 7). MRD was determined by IG/TR gene rearrangements. To specifically assess T-cell ability to eliminate CD19+ target cells we used post-treatment B-cell status as a surrogate marker for T-cell functionality, categorizing patients into two groups based on the presence of ≥10 CD19+ events: B-cell persistence (n = 18) and B-cell aplasia (n = 18). Of the patients with B-cell persistence, 12 remained MRD-positive, while six patients achieved MRD-negativity despite sustained presence of normal B cells. In the B-cell aplasia group, half (n= 9) of the patients were MRD-negative, while the other half remained MRD-positive with loss of CD19 expression. Our cohort included 11 different molecular B-ALL subtypes (available for 29/36); CD19 loss was exclusively found in KMT2A- and DUX4-rearranged B-ALL, subtypes known for their multilineage potential. Notably, both -MRD responders and non-responders- of the B-cell aplasia group had a comparable baseline T-cell composition, suggesting immune-mediated pressure may contribute to driving CD19 loss. Patients with B-cell persistence and B-cell aplasia showed distinct T-cell profiles. At baseline, patients that later cleared B cells had higher CD4+ (p = 0.03) and lower CD8+ T-cell percentages (p = 0.035). Under BLIN, CD8+ T cells expanded significantly only in this group (p = 0.003). Furthermore, patients of the B-cell aplasia group had a significantly higher proportion of naïve CD8⁺ T cells at baseline (p = 0.029), and higher frequencies of CD27+ CD4+, CD27+ CD8+ T cells and CD28+ CD8+ T cells (p = 0.021, p = 0.002, p = 0.014, respectively). Of note, these patients also maintained higher levels of naïve CD8+ T cells after BLIN treatment (p = 0.027). In contrast, patients with B-cell persistence showed enrichment for terminally differentiated T cells at baseline (p = 0.011), with CD8+ T cells expressing higher levels of CD95, indicative of apoptosis susceptibility (p = 0.021), and the senescence marker KLRG1 (p = 0.016). The exhaustion marker TIGIT was significantly increased on both CD4+ (p = 0.038) and CD8+ T cells (p < 0.001). ScRNA-Seq of patients with B-cell aplasia (n= 10) and B-cell persistence (n = 10) identified distinctive gene expression signatures and differentially expressed genes in the T cells (CD4+: 179 genes, CD8+: 325 genes, adj. p value <0.05, |avg_log2FC| >0.5). Genes highly expressed at baseline prior to B-cell persistence included exhaustion marker TOX, cytotoxic markers reflecting the enrichment of differentiated cells (FGFBP2, FCRL6, GZMH, GZMB, SLAMF7,PRF1) and interferon signaling genes (IFI44, XAF1, GBP1, EPSTI1). Patients with B-cell aplasia showed signatures enriched for naïve markers (CCR7, MYC, KLF2), as well as T-cell proliferation/survival (EIF5A, IGF1R). Again, the T-cell transcriptome profiles in the B-cell aplasia group were similar for MRD responders and non-responders with CD19 loss. In conclusion, we show striking differences in baseline T-cell composition between patients with effective versus ineffective B-cell clearance in homogeneously analyzed B-ALL samples. Low naïve CD8+ T-cell levels and elevated CD95, KLRG1, and TIGIT expression may predict T-cell dysfunction under BLIN. Loss of CD19 target antigen expression in MRD persisters, generally limited to subgroups with multilineage potential, may occur as a consequence of selective pressure from a functional immune compartment. These findings offer a valuable basis for developing predictive scores based on T-cell fitness to guide clinical treatment decisions.