Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of aggressive non-Hodgkin's lymphoma with ~40% of patients experiencing refractory or relapsed disease to R-CHOP immunochemotherapy. These higher risk patients often exhibit gene expression profiles that resemble a late germinal center-derived activated B-cell (ABC) phenotype, where diverse specific mutations in genes along BCR and TLR signaling pathways converge into constitutive activation of the NF-κB pathway. Identification of factors that contribute to the survival of malignant ABC-DLBCL cells is crucial for designing specific interventions. In this setting, previous in vitro studies in DLBCL cell lines have evidenced that IL-10 is a direct target of NF-κB signaling that sustains lymphoma cell survival by an auto-stimulatory loop via STAT3 signaling. However, the action of IL-10 could be more complex within the lymphoma microenvironment, as IL-10 can have both stimulatory and inhibitory effects depending on the cellular source of IL-10, the timing of its secretion, and the type of immune cells that receives signals via IL-10 receptor (IL-10R). To explore in vivo the relevance of lymphoma-derived IL-10, we have generated a novel quintuple transgenic mouse model that we call pBIC10, after crossing our previous ABC-DLBCL-like pBIC mice carrying constitutive NF-κB signaling (Pascual et al., Blood 2019) to mice with a floxed Il10 gene for conditional knock-out of IL-10 specifically in malignant ABC-DLBCL cells. As expected, increased intracellular IL-10 signaling was detected in primary murine DLBCL cells from pBIC mice but not in IL-10 deficient pBIC10 mice, which promoted lymphoma cell survival ex vivo as revealed by perturbation studies of the IL-10/JAK1/STAT3 autocrine loop with blocking monoclonal antibodies or selected inhibitors. However, in vivo genetic deletion of lymphoma-derived IL-10 did not offer improved overall survival but, on the contrary, notably accelerated DLBCL progression in pBIC10 mice (Figure 1A, left). A comprehensive integration of multiparametric flow cytometry and transcriptomic analyses using bulk and single-cell RNA-seq of lymphoma cells and the tumor microenvironment coupled with BCR- and TCR-seq, identified lymphoma-derived IL-10 as a key immunomodulator of the DLBCL microenvironment, including unexpected protective paracrine functions in the progression of DLBCL. By comparing IL-10-proficient (pBIC) and -deficient (pBIC10) murine DLBCL models, we observed a multifaceted role of IL-10 produced by DLBCL cells, enabling immune-effector processes (facilitating immune chemotaxis, antigen presentation, and IFNγ responses) that yielded higher percentages of long-lived stem-like CD8 +TCF-1 +PD-1 lo/int T cells and dendritic cells, while restraining PD-1/LAG3-driven T-cell exhaustion and reducing immunosuppression by CD4 +CD25 +FOXP3 + regulatory T (Treg) cells. Thus, depletion of lymphoma-derived IL-10 decreased T-cell cytotoxicity and increased Treg-driven immunosuppression, which collectively accelerated lymphoma development. Consequently, T-cell targeted immune-checkpoint therapy (with anti-PD-1 antibodies) was effective only in IL-10-proficient DLBCL but not in IL-10-deficient pBIC10 mice (Figure 1A, right). Further comparative studies between pBIC and pBIC10 mice showed that tumor vascularization and expression of B-cell calcium channels, which have been previously associated with sensitivity to rituximab, were counteracted by lymphoma-derived IL-10. Accordingly, anti-CD20-based targeting of lymphoma cells (with rituximab surrogate antibodies) showed best in vivo responses in IL-10-deficient DLBCL mice with respect to pBIC mice (Figure 1A, right). Consistent with this preclinical data, analyses of transcriptional IL-10-associated hallmarks to stratify DLBCL patients (Schmitz et al., NEJM 2018) identified better clinical responses to R-CHOP in patients with pBIC10-like DLBCL phenotype (Figure 1B). Collectively, our study provides new functional and mechanistic insights into the role of lymphoma-derived IL-10 in the pathogenesis of DLBCL and the restraint of an immunosuppressed tumor microenvironment, which holds predictive potential as a biomarker for immunotherapy responses in DLBCL.
Read full abstract