Sasp-driven immune evasion defines a novel African ancestry-associated DLBCL subtype with distinctive molecular and immune vulnerabilities

Abstract

Abstract Diffuse large B-cell lymphoma (DLBCL) arises at an earlier age and is associated with inferior survival in African American patients (pts). These disparities reflect socioeconomic and genetic factors. For example, SETD2, and other DNA damage-linked mutations are highly enriched in U.S. pts with African ancestry (AA) but are exceptionally rare in European ancestry (EUR) DLBCLs (Lee et al, 2020). We showed that SETD2 mutations disrupt repair of AICDA-mediated DNA damage in germinal center (GC) B cells, causing DLBCLs with genomic instability (Leung et al, 2022). We now report these mutations are equally enriched in DLBCLs from Malawi vs EUR pts (p<0.001), supporting ancestry dependence. We generated a BCL2;Setd2+/- (SB2) mouse model for AA-DLBCLs and analyzed mice tumors at 6 and 12 months by IHC, IF and flow cytometry. Compared to BCL2-only, SB2 mice showed enriched exhausted CD4+ T cells at 6 months (p<0.0001) and exhausted CD8+ T cells at 12 months (p=0.0009). We developed luciferase+ SB2 lymphoma cell line that engrafts in syngeneic mice and 5 isogenic human SETD2+/- DLBCL cell lines to broaden our AA-DLBCL models. IV injection of SB2 cells killed mice within 5 weeks and displayed abundant CD4/CD8 effector/exhausted (Eff/Exh) T cell infiltration. Both murine and human SETD2+/- DLBCLs showed marked genomic instability vs controls (COMET assay, p<0.001), a known trigger of senescence and inflammation. Senescence is not reported as a characteristic of DLBCLs, however primary SB2 tumors (p=0.02), murine, and human lines (p<0.0001) had high levels of senescence by β-galactosidase (β-gal) staining and C12FDG flow cytometry. RNA-seq, proteomic, and phospho-proteomic studies in SB2 and human SETD2+/- cell lines vs controls showed enrichment for senescence-associated secretory phenotype (SASP) signatures and activation of p38 MAPK, JAK-STAT, TLR4, TNFα, and NFκB (p<0.01). CODEX imaging of a tissue microarray (TMA) with 25 AA- and 97 EUR-DLBCLs showed a significantly enriched inflammatory microenvironment of Eff/Exh CD4+ and CD8+ (both p<0.0001) cells in AA-DLBCL. RNA-seq showed upregulation of SASP associated kinases in AA vs Eur DLBCL pts. To test if primary AA-DLBCLs also featured abundant senescent cells, we performed GLB1 (β-gal) IHC in TMAs from three large independent AA-DLBCL cohorts, including one from Malawi, vs two EUR cohorts. Each AA-DLBCL cohort showed striking SASP enrichment (>20% GLB1+ cells; p<0.0005), confirmed with an independent functional senescence marker (SentraGor). SASP+ tumor cell abundance correlated with exhausted CD4+ (p=0.02) and CD8+ (p=0.001), suggesting a link between SASP and immune modulation. To test this, we sorted top-quartile SASP+ and SASP– SB2 cells. Both proliferated similarly, indicating SASP does not impair DLBCL growth. Upon transplantation, only SASP+ cells formed tumors in immunocompetent C57BL/6 mice (p=0.02), while both SASP+ and SASP- engrafted in RAG1KO immunodeficient mice, showing SASP is required to evade immune surveillance. SASP+ cells secreted CD4-activating cytokines (IL20, IL12, IL16) and canonical SASP factors (IL1β, IL6, IL10, TNFα). We hypothesized that CD4 exhaustion is critical for SASP-driven immune evasion. Accordingly, in vivo depletion of CD4+ T cells impaired lymphoma progression, whereas CD8+ depletion accelerated it. Though checkpoint inhibitors (CPI) generally fail in DLBCL, we reasoned the SASP phenotype might confer susceptibility. Indeed, SB2 DLBCLs but not GCB DLBCLs were highly sensitive to murine-optimized PD1 inhibitors (p=0.04), with prolonged complete remission. We tested whether SETD2+/- DLBCLs would show sensitivity to SETD2 inhibitor due to defective DNA repair and found that SETD2i selectively depleted SB2 SASP⁺ cells by inducing overwhelming DNA damage (p=0.01), delayed tumor growth (p=0.002), and restored CD4/CD8 functionality (p=0.01 and p=0.02). Combinatorial CPI + SETD2i studies are underway. In summary, we identified a novel DLBCL subtype almost entirely restricted to AA individuals, driven by a novel SASP-like program that promotes lymphomagenesis via CD4 exhaustion. This subtype is readily diagnosed by IHC or flow cytometry using validated senescence markers. Recognizing these “ancestry variant” cases is critical, as they may respond to CPI and senolytic therapies as an ancestry-informed precision immunotherapy strategy for this population historically underrepresented in clinical trials and often unable to access state-of-the-art care.