Abstract Background: CD20xCD3 bispecific antibodies (BsAb) are a transformative therapy for B-cell non-Hodgkin lymphomas (B-NHL), effective even post–CAR T cell therapy. While durable remissions occur, relapse is frequent and resistance mechanisms remain poorly defined, with CD20 loss being the only validated biomarker. BsAb activate the CD3 epsilon (CD3ε) subunit of the T cell receptor, expressed across cytotoxic and immunosuppressive CD4+and CD8+ T cell subsets. Given the heterogeneity of intratumoral T cells in B-NHL, we investigated the impact of distinct T cell subsets on BsAb-mediated anti-lymphoma activity. Methods: We used two orthogonal approaches to determine how T cell abundance influences BsAb responses. First, we analyzed tumor samples from 22 patients enrolled in a multi-arm phase 1/2 study of Epcoritamab with standard therapies for DLBCL/FL (NCT04663347), using flow cytometry and, where available, single-cell surface protein, transcriptome, and TCR profiling (10x Genomics). Flow data were available for all patients (9 DLBCL, 13 FL); scRNA/TCR-seq for 7 on-treatment biopsies (2 DLBCL, 5 FL). We performed in vitro cytotoxicity assays using polarized CD4⁺ and CD8⁺ T cells co-cultured with luciferase-expressing RAJI B cells ± BsAb. In parallel, we used an Ezh2Y641F/Bcl2-driven B-NHL cell line (tFL-p6) which was injected via tail vein in mice and treated with a murine CD20xCD3 BsAb ± prior CD4⁺ or CD8⁺ T cell depletion. Results: Flow cytometry of pre-treatment biopsies revealed that patients who subsequently experienced disease progression (PD, n=2) demonstrated reduced CD8⁺ T cell infiltration (% of T-cells: 6.8 vs. 20.8%, p=0.035) and CD8:CD4 ratios (4.1 vs. 11.6, p=0.032) versus patients who remained progression-free (n=8). This difference persisted in on-treatment biopsies. Single-cell profiling of 14,278 T cells from 7 patients with available on-treatment biopsies (4 CR, 2 PR, 1 PD) showed distinct changes in CD4⁺ and CD8⁺ dynamics by response: CR patients exhibited polyclonal expansion of cytotoxic CD4⁺ and CD8⁺ T cells (expressing CST7, CCL4, NKG7). In contrast, the PD patient demonstrated clonal expansion of CD4⁺ Tfh and Treg cells, with these populations showing increased medium/large clones (≥5 cells) and exhibited minimal CD8⁺ T cell expansion. PR patients showed limited Tfh/Treg expansion and moderate CD8⁺ cytotoxic/exhausted signatures with partial clonal expansion. These data suggest that intratumoral CD8⁺ T cell abundance is associated with durable responses, whereas enrichment of CD4+ T cells–particularly the Tfh subset–was observed in the non-responders. To functionally validate the contribution of CD4⁺ and CD8⁺ subsets to BsAb-mediated cytotoxicity, human T cells were polarized under defined conditions for 5 days and co-cultured with luciferase-expressing RAJI B cells for 24 hours in vitro in the presence or absence of BsAb. CD8⁺ T cells induced robust target cell killing (50.5% lysis), whereas CD4⁺ Tregs and Tfh cells mediated significantly less killing (14.8 and 15.3%, respectively; both p0.01 vs. CD8⁺). In mice bearing tFL-P6 B-NHL, BsAb treatment significantly prolonged survival versus IgG controls (median survival: not reached vs. 10 days; p0.05). On day 4, BsAb-treated mice showed increased proportions of CD8⁺ T cells (50.9 vs. 35.8%, p=0.04) and decreased CD4⁺ T cells (29.1 vs. 46.8%, p0.001) in BsAb-treated mice compared to controls (n=6 vs. n=6). BsAb treatment also increased accumulation of granzyme B⁺ CD8⁺ T cells (7.2 vs. 1.0%, p=0.02) and CD44⁺CD62L⁻ effector memory CD8⁺ T cells (47.7 vs. 15.9%, p0.01). Preliminary experiments using in vivo depletion of CD8⁺ or CD4⁺ T cells prior to BsAb administration demonstrated that CD8⁺ T cells were essential for BsAb-mediated anti-tumor activity, while CD4⁺ T cell depletion had no significant effect on survival. Conclusion: Our data identify CD8⁺ T cells as critical mediators of BsAb-induced anti-lymphoma activity. On-treatment intratumoral CD4:CD8 ratio and T cell subset shifts were associated with clinical response and may hold potential to inform BsAb-based treatment strategies. While cytotoxic CD4⁺ T cells may contribute in some patients, accumulation of CD4⁺ Tfh and Treg populations overall appears to limit therapeutic efficacy. These data provide a mechanistic rationale for selectively suppressing CD4⁺ subsets and/or preferentially enhancing CD8⁺ T cell recruitment and activation.
Rahman et al. (Mon,) studied this question.