Th1-like Treg cells are dressed to suppress anti-tumor immunity

Abstract

Mechanisms of Th1-like Treg suppression are unknown in cancer. Two studies in Immunity by Ayala et al. and Zagorulya et al. demonstrate that Th1-like Treg cells interact with type 1 dendritic cells in tumors and draining lymph nodes to potently suppress anti-tumor immunity.

Over the past 20 years, a great deal has been learned about regulatory T (Treg) cell-mediated function in cancer.¹ Among their immunosuppressive mechanisms, Treg cells can interact with dendritic cells (DCs), impairing their costimulatory function and inhibiting CD8⁺ T cell priming.² Still, gaps remain in our understanding of how Treg cell suppression can dominate in cytokine conditions that otherwise favor a CD8⁺ T cell response. Based on studies in autoimmune disease models, the field now appreciates that Treg cells can adopt features of conventional T helper (Th) cell subsets.^3,4 Similar to their CD4⁺ T helper (Th1) counterparts, Th1-like Treg cells depend on IL-12 and interferon gamma (IFN-γ) for their generation, express the transcription factor Tbet, and maintain high levels of the chemokine receptor CXCR3.³ These features tailor Treg cells for accumulation in sites of IFN-γ and CXCL9-driven inflammation, and the suppression of type-1 autoimmune diseases.⁵ Although Treg cells expressing Tbet and CXCR3 are present in human cancers,^6,7 the field has largely focused on the function of Th2-like Treg cells that express CCR4 and CCR8.¹ Studies by Zagorulya et al. ⁸ and Ayala et al.⁹ show how the acquisition of Th1-like features by Treg cells favors their interactions with cross-presenting type-1 dendritic cells (DC1s), making them ideally suited to suppress anti-tumor CD8⁺ T cell responses.

While exploring unexpected differences in T cell responses in different tumor-draining lymph node (TDLN) compartments, Zagorulya et al. uncovered the presence of a Th1-like Treg population that restrains regional CD8⁺ T cell responses to lung cancer.⁸ Using a Kras/p53 mutant lung cancer cell line expressing the model antigen OVA, the authors sought to explain an interesting prior observation that lung (mediastinal) TDLNs uniquely promote a dysfunctional phenotype in tumor-specific CD8⁺ T cells. This defect was associated with lower expression of costimulatory molecules and production of IL-12 by the DC1 compartment in the lung but not the flank TDLN. Importantly, in vivo depletion of Treg cells preferentially restored DC1 and CD8⁺ T cell function in lung TDLNs, and ex vivo coculture confirmed that differences were attributed to MHC II-dependent Treg cell engagement, which could be restored by exogenous costimulatory signals and IL-12. Surprisingly, Treg cells were in similar abundance in lung and flank TDLNs, and single-cell RNA sequencing (scRNA-seq) analysis did not point to meaningful clonal differences, nor overt differences in transcriptional clustering of Treg cells in the different locations. However, closer examination of an activated Treg cell transcriptional cluster revealed higher expression of genes associated with Th1 transcriptional program—most notably Tbx21 (Tbet) and Cxcr3—in lung TDLNs. Flow cytometry confirmed higher expression of these two markers, as well as PD-1 and CTLA-4, validating the presence of an activated Th1 like Treg cell population that was uniquely generated in response to tumor growth in the lung.

While the presence of Th1-like Treg cells explained the outsized suppressive effect on DC1s in the lung relative to flank TDLN, the regional cause of this phenomenon remained mysterious. Interestingly, the difference in Treg cell suppression was lost in Batf3^−/− mice, indicating a reciprocal requirement for regional DC1s in generating Th1-like Treg cells. Moreover, Tbet expression was decreased on Treg cells in tumor-bearing mice that lacked the IFN-γ receptor, and in vivo neutralization of IFN-γ both reduced Tbet expression in Treg cells and restored a functional phenotype to CD8⁺ T cells. These findings established a paradoxical mechanism whereby IFN-γ suppresses anti-tumor immunity through the induction of Th1-like Treg cells. Moreover, this finding provided a key clue into locational differences, as higher levels of IFN-γ were present in lung relative to flank TDLNs. Recognizing the lung as a mucosal tissue site, the authors hypothesized that these discrepancies could relate to differences in the host microbiome. Indeed, levels of IFN-γ in lung and flank TDLNs were normalized in germ-free mice, suggesting that regional microbiomes differentially skew Treg populations in the setting of cancer. Moreover, Th1-like Treg cells could represent a similar barrier in humans, as analysis of a melanoma scRNA-seq dataset revealed a higher score of IFN-γ response genes and Tbx21 levels in Treg cells from immune checkpoint inhibitor non-responding patients.

Moving beyond TDLNs, studies by Ayala et al. in the current issue of Immunity extend our understanding of Th1-like Treg cells to the tumor microenvironment (TME).⁹ Whereas Zagorulya et al. observed minimal Th1-associated CXCR3 expression on Treg cells in LNs, Ayala et al. show prominent populations of CXCR3⁺ Treg cells in the TMEs of three syngeneic mouse tumor models. By developing an elegant strategy whereby random X chromosome inactivation was exploited for diphtheria toxin receptor (DTR) expression only in Treg cells that express CXCR3, the authors used DT to selectively deplete CXCR3⁺ Treg cells while preserving their CXCR3^−/− counterparts. This technique was used to show that CXCR3 expression supports Treg cell accumulation and acquisition of an activated CTLA-4^hi PD-1^hi phenotype in tumors. Importantly, the loss of CXCR3⁺ Treg cells significantly impaired tumor growth across all three tumor models, with closer investigation of the MC38 tumor model revealing clear CD8⁺ T cell dependence. Tumor protection was accompanied by increased numbers of antigen specific CD8⁺ T cells in tumors but not in dLNs, suggesting that CXCR3⁺ Treg cells oppose CD8⁺ T cell access to the TME.

To better understand the role of CXCR3 on Treg cell function within the TME, Ayala et al. next focused on interactions with DCs. Confocal microscopy showed that CXCR3 expression by Treg cells facilitated their contact with DCs in tumors, and that the frequency of CXCR3⁺ Treg cells strongly correlated with CXCL9-expressing DCs. Additionally, CXCL9 production was lost in Batf3^−/− mice, implicating DC1s, which produced much higher levels of CXCL9 than their type-2 DC counterparts in tumors. Interestingly, blockade of either IFN-γ or depletion of CD8⁺ T cells eliminated CXCL9 from the TME, suggesting a key yet unexpected requirement for CD8⁺ T cell-derived IFN-γ in supporting Treg cell function. Moreover, CXCR3⁺ Treg cell depletion enhanced CD8⁺ T cell-DC interactions, indicating that Treg cells in turn interfere with CD8⁺ T cell-DC communication. Interestingly, although CXCR3^−/− Treg cells promoted larger tetramer-specific CD8⁺ T cell responses in tumors, they had no effect on costimulatory molecule expression by DCs. This suggested that CXCR3⁺ contribution to Treg cell function is distinct from what has been reported in other models.^2,10 To better elucidate the suppressive mechanism, the authors transplanted mice with MC38 tumors expressing OVA but lacking MHC-I, and then directly measured cross-presentation by DC1s using anantibody that recognizes the MHC-I:OVA peptide complex. Indeed, in vivo loss of CXCR3⁺ Treg cells significantly improved the cross-presenting capabilities of tumor-derived DC1s. Taken together these studies reveal a delicate balance between Treg cells, DC1s, and CD8⁺ T cells in the TME, and show that CXCR3 expression by Treg cells tips thescales in favor of immune suppression.

In sum, work by Ayala et al. and Zagorulya et al. adds substantially to our understanding of the insidious nature of Treg cells in tumors (Figure 1). By revealing how Treg cells co-opt features of a Th1 response to optimally suppress CD8⁺ T cell responses, cytokines and chemokines that were long considered our weapons in the battle against cancer have emerged as double-edged swords. Prior studies have acknowledged the difficulty in selectively recruiting CD8⁺ T cells to inflamed tumors without attracting Treg cells,¹¹ and the present findings underscore Treg cell CXCR3 expression as a basis for this fundamental challenge. Moreover, the knowledge that local microbial features can promote deleterious Treg cell skewing now must factor into our understanding of variable tumor immunity across different tissue locations. However, Th1-like Treg cells also express high levels of CTLA-4 which can potentially serve as a target for depletion by immune checkpoint inhibitory drugs.¹ Future studies will be needed to determine the antigen specificity of Th1-like Treg cells in cancer, and what factors support their differential function in the TME versus draining lymph nodes. Achieving productive anti-tumor immunity is now more than ever recognized as interfering with a balancing act that was designed to maintain peripheral tolerance.

Figure 1. Th1-like Treg Cells in the tumor-draining lymph node (TDLN) and the tumor microenvironment (TME) suppress anti-tumor immunity.

(Left) Th1-like Treg cells interact with and suppress DC1s, downregulating costimulatory molecule expression and IL-12 production, and reducing CD8⁺ T cell activation. (Right) CXCR3 expression on Treg cells is crucial for their suppression of tumor control. CXCR3 is also important for Treg cell accumulation in the TME where they interact with CXCL9-secreting DC1s and results in decreased DC1 cross-presentation.