Removing the B (cell) STING to improve cancer immunotherapy.

Abstract

Cancer immunotherapies focus on harnessing the host's own immune system to identify and eliminate tumors. One current strategy is to boost the immune system by activating the pattern recognition receptor, Stimulator of Interferon Genes (STING). STING is well known to induce a potent type I interferon (IFN) response as part of the host innate immune response to infection,1 which has been exploited to great effect in antitumor immunity in mouse models.2, 3 However, the development of tumor resistance to STING agonist monotherapy has hampered current efforts to utilize STING activation to treat cancer successfully in humans. Why STING agonists have not yet been clinically successful as originally touted has remained a mystery. In their recent Nature research article, Li et al. shed new light on the use of STING agonists in cancer therapy by identifying a potential mechanism through which STING activation in vivo can actually dampen the antitumor response rather than boost it.4 Immunotherapies enhance our immune system's capacity to recognize and eliminate cancer cells. However, the selection of low immunogenic tumor cells over time can lead to the formation of “cold” tumors, which can no longer be recognized by the immune system. In addition, tumors can remodel their surrounding tissue and extracellular matrix to establish an immunosuppressive milieu that promotes immune evasion.5 Often, the immunosuppressive tumor microenvironment (TME) recruits regulatory B and T cells (Breg and Treg cells, respectively) and myeloid-derived suppressor cells (MDSCs), which suppress the immune system and promote tumor progression. This can be achieved through several mechanisms, including the production of anti-inflammatory molecules, overexpression of checkpoint molecules (e.g. PD-L1) and co-inhibitory molecules, which act to prevent infiltration, proliferation and dampen the effector functions of key immune cells, such as T cells and natural killer (NK) cells.5 Currently, the agreed consensus on the role of STING activation in antitumor immunity is that STING activation in dendritic cells (DCs) increases DC maturation. This subsequently promotes the cross-presentation of tumor antigens to cytotoxic CD8+ T cells and, as a result, increases CD8+ T cell-mediated cytotoxic killing of tumor cells.2, 6, 7 Recent studies have further identified that NK cells can mediate potent tumor elimination in a STING-dependent manner in the absence of CD8+ T cells.8, 9 Since there is limited understanding of how the cGAS-STING pathway primes NK cells-dependent antitumor immunity, most studies have focused on potential T cell-dependent tumor resistance to STING agonist monotherapy, including STING activation leading to T cell death,10 enhanced PD-L1 expression in tumor cells11 and augmented expression of immunosuppressive molecules12 that suppress CD8+ T cell activity. This interesting new work by Li et al.4 identifies that STING activation within immature B cells leads to the development of a subset of Breg cells that dampen NK cell proliferation and cytotoxic activity.4 Systemic administration of the natural STING agonist, cGAMP, into wildtype mice bearing pancreatic ductal adenocarcinoma (PDAC) tumor increased the percentage of B cells among CD45+ intratumoral cells. This was regardless of the number of total CD45+ cells or tumor size, which was surprisingly unaffected by cGAMP treatment, indicating that systemic administration of cGAMP alone in the PDAC model is insufficient to induce antitumor activity. Although the role of T cells in STING antitumor immunity has been extensively examined, there has been limited focus on the role of other lymphocytes, particularly B cells, in fighting cancer. STING activation within B cells has previously been shown to synergistically function with B cell receptor signaling to activate B cells and, thus, humoral immunity.13 However, the role of humoral immunity in fighting cancers has been highly controversial. Here, deletion of Tmem173 (the murine gene encoding STING) reduced the tumor burden when PDAC-bearing mice were treated with cGAMP compared with wildtype mice. While this illustrates that STING activation within B cells is associated with an increased tumor burden, further investigation is required to determine the effect of Tmem173 deletion on B cell development and function in the absence of STING activation as well as B cell tumor infiltration. STING activation has been shown previously to induce several downstream responses including type I IFN response, NF-κB- and MAP kinase-dependent proinflammatory cytokine production, autophagy, cell senescence and various programmed-cell death pathways.1 To understand which aspect of STING signaling within B cells abrogates cGAMP-induced antitumor immunity, Li et al.4 conducted a RNAseq analysis comparing differentially expressed genes between untreated and cGAMP-treated wildtype B cells. Interestingly, interleukin (IL)-10, the cytokine encoded by gene Il10, and Ebi3, a subunit of IL-35 cytokine encoded by gene Ebi3, were upregulated in wildtype B cells following cGAMP treatment. Further investigation revealed that p35 mRNA (another subunit of IL-35; also shared with IL-12) was constitutively expressed and was not altered by cGAMP treatment. This was reflected in the increase of IL-35+ and IL-10+ splenic B cells from wildtype mice treated with cGAMP but not in Tmem173−/− mice. Both IL-10 and IL-35 have been described previously as immunosuppressive cytokines and act as a negative regulator of immune responses.14, 15 While IL-10 has been shown to promote antitumor immunity by dampening tumor-induced inflammation in the TME,14 IL-35 expression promotes the development of tumors, such as PDAC, hepatocellular carcinoma and non-small cell lung cancer.16 While IL-35 can be produced by most immune cells, the major producers of IL-35 are Breg and Treg cells. Interestingly, in a positive-feedback loop, IL-35 is shown to drive the development of conventional B cells and IL-10-producing B cells (also termed B10 cells) into Breg cells.16 In line with these data, Li et al.4 discovered that the IL-35+ B cells displayed the markers phenotypical of murine Breg cells (CD19+CD1dhiCD5+CD21hi). Additionally, cGAMP treatment increased the number of total CD19+CD1dhiCD5+CD21hi Breg cells and the percentage of IL-35+ cells among this Breg cell population. Therefore, STING activation within B cells leads to the development of IL-35+ Breg cells. To further understand the role of IL-10 and IL-35 in the TME, Li et al.4 conducted a correlation analysis that demonstrated that IL-10 and IL-35 expression correlated positively with tumor weight and decreased tumor-infiltrating CD8+, CD4+ and NK1.1+ effector cells. This indicates that both IL-10 and IL-35 are important in establishing an immunosuppressive TME. B cell-specific deletion of Ebi3 or p35 alone increased intratumoral CD8+ T cells and NK cells which indicates that IL-35 plays a major role in suppressing antitumor immune responses within the TME. Intriguingly, cGAMP treatment of these mice did not further increase the number of CD8+ T cells but significantly increased NK cells. This demonstrates that systemic administration of cGAMP drives the development of Breg cells, which attenuates NK tumor infiltration in an IL-35-dependent manner. Cotreatment with an anti-IL-35 antibody (anti-IL-35) and cGAMP had significantly reduced the tumor burden and increased survival.4 Anti-IL-35 and cGAMP dual treatment increased the number of intratumoral NK1.1+ and NK1.1+GzmB+ over each of the mono-treatments alone. Thus, Li et al.4 present a model in which systematic or intratumoral administration of cGAMP induces IL-35 expression in B cells, which dampens NK cytotoxic activity and tumor infiltration leading to tumor resistance to cGAMP monotherapy. However, combinational treatment with anti-IL-35 rescues NK cell activity to overcome tumor resistance (Figure 1). To delineate which aspect of STING signaling was required to induce Il10, Ebi3 and p35 expression in B cells, the authors stimulated splenic B cells isolated from wildtype, Tmem173−/−, Irf3−/− or Ifnar1−/− mice with cGAMP ex vivo and analyzed cells using flow cytometry to quantify the specific IL-35+ and IL-10+ B cell populations. Interestingly, the frequency of IL-35+ and IL-10+ B cells were comparable between Ifnar1−/− and wildtype B cells but markedly diminished in Tmem173−/− and Irf3−/− B cells, thus demonstrating a critical role for IRF3. This was confirmed through IRF3 chromatin immunoprecipitation (ChIP) assays, revealing that IRF3 bound directly to the promoters of p35, Ebi3 and Il10 in wildtype but not Tmem173−/− B cells. This correlated with significant increases in IL-10 and IL-35 cytokine levels in PDAC tumor homogenates. These data demonstrate that p35, Ebi3 and Il10 expression is dependent on STING and IRF3 but independent of type I IFN signaling. This interesting finding highlights our limited understanding of IRF3 as a transcription factor beyond the induction of type I IFNs. Hence, future research is required to understand the effects of non-IFN responses downstream of IRF3 activity within different immune cells. While STING activation has been shown to induce potent antitumor immunity, there are reports of a dichotomous role of STING in cancer progression and immunotherapy. Indeed, STING activation can lead to suppression of T cells by inducing T cell death,10 increased PD-L1 expression11 and expression of immunosuppressive molecules.12 In line with this, the new findings from the Ting laboratory highlight an additional pitfall in the administration of STING agonists in cancer immunotherapy. While this study focused on the effect of cGAMP administration on B cells, these findings can potentially be applied to a more general understanding of the impact of tumor-derived cGAMP on modulating B cell activity in the context of developing tumor resistance. The finding that STING-mediated resistance can be overcome with anti-IL-35 treatment is consistent with observations made by others where IL-35 has been shown to increase tumor burden in PDAC mouse models17 and inversely correlate with the survival of patients with NSCLC.17 Overall, this work further highlights the complex nature of STING activation within the TME and across different immune cells for mediating antitumor immunity. It appears that STING must be very carefully targeted if it is to be effectively exploited for successful cancer immunotherapy. RV is supported by an Australian Government Research Training Program (RTP) Scholarship. DDN is supported in part by a Monash University FMNHS Senior Postdoctoral Fellowship. Rajan Venkatraman: Conceptualization; writing – original draft. Dom De Nardo: Conceptualization; writing – original draft; writing – review and editing. We declare no conflicts of interest.