Pre-existing cytotoxic capacity of CD8+ and CD4+ T cells is a hallmark of richter transformation and favors response to anti-PD-1 therapy Free

Immune checkpoint blockade (ICB) therapy has shown promising activity in Richter Transformation (RT), yet the tumor microenvironmental (TME) determinants of response and resistance to this therapy remain incompletely defined. Our previous single-cell RNA sequencing (scRNA-seq)-based analysis of splenocytes from CLL and RT mouse models identified enrichment in CD8⁺ effector/exhausted and CD4⁺ cytotoxic T cells, together with CXCL9/10⁺ pro-inflammatory macrophages, as a characteristic of most RT cases. The deeper characterization and the functional activities of these candidate cell populations, however, have not yet been elucidated.

Consistent with human disease, we identified overall higher tumor mutational burden (measured by whole-genome sequencing of sorted B-cell splenocytes) from our mouse models of RT (n=11) compared to CLL (n=4) (p=0.04, Wilcoxon rank sum test), accompanied by computational prediction of neoantigens. To investigate the potential of CD8⁺ and CD4⁺ PD1⁺ T cells to become activated, we measured ex vivo intracellular staining of effector molecules by flow cytometry after PMA/Ionomycin stimulation. This revealed an increase in granzyme B (GZMB)⁺, tumor necrosis factor alpha (TNFα)⁺, interferon gamma (IFNγ)⁺, GZMB⁺IFNγ⁺, and triple positive (GZMB⁺ TNFα⁺IFNγ⁺) CD8⁺ and CD4⁺ PD1⁺ T cells in RT compared to CLL (CD8⁺: p_adj<0.024, Wilcoxon rank-sum test + FDR), indicating a higher cytotoxic potential of RT-T cells. To more directly assess T-cell cytotoxic capacity, we performed co-culture assays using flow-cytometrically sorted CLL (n=4) and RT (n=4) tumor cells with autologous CD4⁺ or CD8⁺ T cells, and measured tumor cell viability following 24 hours of in vitro co-culture in presence of IL-2. Of note, RT tumor cells displayed reduced viability when co-cultured with their corresponding CD4⁺ and CD8⁺ T cells (p<0.007, t-test), while CLL instead showed increased viability in the presence of T cells (p<0.016, t-test).

To further identify determinants of response to ICB, we exposed 4 genetically distinct RT mouse models to anti-PD-1 monotherapy (or isotype control treatment, n=5-6 mice/group) for 3 weeks, yielding 2 responders (R), one mixed responder (MR) and one non-responder (NR). scRNA-seq combined with non-negative matrix factorization revealed reduced T-cell exhaustion in R mouse models after treatment, manifesting as reduced expression of Ifng (p<0.004, t-test) and signatures of exhaustion and inflammatory signaling (p<0.014, t-test) compared to baseline. These results were validated by flow cytometry, in which a reduction of CD8⁺ and exhausted T cells after ICB treatment (p<0.015, t-test) were detected compared to isotype-treated groups. Notably, ex vivo killing assays showed an improved capacity of CD4⁺ and CD8⁺ T cells from both R lines to kill tumor cells in vitro compared to T cells from NR or MR models (p<0.0001, ANOVA). Immunofluorescence staining of the spleens from the R vs. NR mouse models revealed the former had a more conserved spleen structure with separated red and white pulp, while the latter had intermixed red and white pulp and only weak remnants of T-cell zones visible. Consistently, we observed an enrichment of migratory cell markers in cytotoxic T cells (p<0.046, t-test), and a more diverse T-cell clonal repertoire in R compared to NR (p_adj = 0.025, ANOVA), suggesting preserved migratory T-cell capacity in R models. To confirm the relevance of our results to human RT, we examined pre-treatment lymph node (LN) biopsies from 4 patients [2R, 1 partial responder (PR), and 1 NR] enrolled in the phase II RT1 clinical trial using the anti-PD-1 agent tislelizumab combined with the BTK inhibitor zanubrutinib by spatial transcriptomics (Visium HD, 10X genomics). This analysis revealed a relatively lower ratio of tumor cell burden to aggregates of immune cells (primarily T cells and macrophages) in R compared to NR cases.Overall, these findings indicate that preserved spleen structure and residual nodal immune aggregates along with more functional cytotoxic T cells are important for response to ICB in RT. We are currently validating this effect in additional human samples from the RT1 clinical trial using multiplex immunofluorescence staining, and investigating molecular pathways associated with these differences between responders and non-responders. We foresee that these insights into determinants of response in RT may also be relevant for other lymphomas and for other RT immunotherapies.