The complex role of EZH2 in the tumor microenvironment: opportunities and challenges for immunotherapy combinations

Tumor initiation and evolution requires a series of genetic and epigenetic changes which enable uncontrolled cell proliferation while limiting surveillance from specialized components of the immune system capable of eliciting antitumor responses. Epigenetic enzymes including DNA methyltransferases, histone deacetylases, histone acetyltransferases, histone methyltransferases, histone demethylases and chromatin remodeling enzymes have been implicated as critical regulators of the cellular transcriptional states that establish the equilibrium between tumor cells and their immune microenvironment [1–3]. Due to the fundamental role of epigenetic regulation in controlling cellular plasticity in both tumor cells and immune cells, increasing evidence has emerged around the role of epigenetic regulators in mediating multiple pathways of immune evasion and resistance toward immune checkpoint therapies [4,5]. Based on these observations. ongoing clinical trials are currently exploring therapeutic targeting of these classes of enzymes as opportunities to reverse epigenetic resistance mechanisms to immune checkpoint therapies including anti-PD-1, anti-PD-L1 and anti-CTLA-4 [6].

Approaches targeting epigenetic regulators for immune therapy combinations require consideration of the diverse functional role of these enzymes and protein complexes in the heterogeneous cell lineages that comprise the innate and adaptive immune system. Ideally, epigenetic drug combinations with immune checkpoint inhibitors should enhance the immunogenicity of tumors, sustain effector functions of T cells and reprogram suppressive immune cell types in the tumor microenvironment. Paradoxically, many epigenetic targets being considered for immune therapy combinations have critical roles in immune cell differentiation, maturation, and function required for antitumor activity. Therefore, it is important to define the broader impact of inhibiting these targets in multiple cell types within the tumor microenvironment.

In this review, we highlight emerging data supporting utilization of inhibitors of the histone methyltransferase EZH2 in clinical immunotherapy applications. We will discuss increasing evidence for the role of dysregulation of EZH2 function as a key mediator of tumor-intrinsic programs driving immune-suppression. Additionally, we describe the role of EZH2 in immune cell populations interfacing with the tumor microenvironment and how epigenetic changes within these cell types may also contribute to suppressive phenotypes observed in tumors. Finally, we will discuss the complexity of effects anticipated from inhibiting EZH2 in both tumor and immune cells and how these may both productively and negatively impact the clinical application of EZH2 inhibitors in cancer therapy.

Role of EZH2 in epigenetic regulation

Epigenetic regulation refers to a physiological process of stable alternations in gene expression patterns without changing underlying DNA sequence. DNA methylation, histone post-translational modifications including phosphorylation, acetylation, ubiquitination, methylation at enhancers and promoters and the activity of chromatin remodeling enzymes are believed to play a key role in the control of gene expression in response to developmental and environmental changes [7–9]. The histone methyltransferase EZH2 is a key regulator of cellular epigenetic states. As the core enzymatic subunit of PRC2 consisting of EZH2, EED, SUZ12, and RbAp46/48, EZH2 mediates H3K27me3 to influence chromatin compaction that promotes transcriptional silencing at targeted genes [10,11]. A paralog of EZH2, EZH1, has also been isolated in PRC2 complexes and appears to show a distinct expression pattern compared with EZH2. While EZH2 is highly expressed in proliferating cells, EZH1 shows abundant expression primarily in nondividing cells. As PRC2 complexes with EZH1 possess low methyltransferase activity compared with the PRC2 complex containing EZH2, nonenzymatic functions have also been ascribed to EZH1-containing PRC2 complexes [12,13].

EZH2 and PRC2 members were initially identified in Drosophila where they play a critical role during embryonic development, regulating accurate gene expression to facilitate various biological functions, including cell proliferation, lineage commitment, X-chromosome inactivation and stem cell self-renewal. EZH2 mediated H3K27me3 to represses transcription of numerous groups of developmental genes like the HOX clusters and genes encoding the homeodomain-containing transcriptional factors. In adult tissues, EZH2 functions in maintaining homeostasis of various tissues through control of gene expression programs in tissue-specific stem cells. Through regulation of developmental gene programs, EZH2 function is also critical in regulating early and late stage of hematopoiesis.

Dysregulation of EZH2 in cancer

Epigenetic dysregulation of tumors, evidenced by changes in epigenetic modifications at gene promoters and enhancers, promotes transcriptional changes that shape oncogenic gene expression programs required for tumor cell growth and survival. Changes in epigenetic regulation in tumors have been directly linked to molecular alterations of chromatin regulating enzymes themselves due to mutation, overexpression, translocation or misrecruitment.

EZH2 dysfunction has been identified as an important oncogenic mechanism in both hematological and solid tumors [14,15]. The best studied example of EZH2 alterations occur in B-cell lymphomas, follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL), in which recurrent heterozygous gain-of-function somatic mutations in EZH2 catalytic SET domain have been identified in 30% of germinal center (GC) B-cell-like subtype DLBCLs and 27% of FLs. These mutations enhance the efficiency of EZH2 activity toward H3K27 trimethylation resulting in more pronounced repression of EZH2 target genes. Mechanistically, these mutations behave as direct drivers of oncogenesis. Mice engineered to express mutant EZH2^Y641 in GC B cells develop GC hyperplasia and accumulate high levels of H3K27me3. EZH2 gain-of-function mutations confer a gene expression signature featuring repression of genes involved in B-cell terminal differentiation and proliferation checkpoints. This observation is consistent with the critical role of EZH2 function during normal B-cell GC maturation. Within B-cell lineages, EZH2 was shown to be highly expressed in progenitor cells however decreased in expression in differentiated and mature B cells [16]. Upon activation, B cells were found to massively upregulate EZH2 expression during GC formation and immunoglobulin affinity maturation [17]. EZH2 prevents terminal differentiation of GC B cells and sustains GC B-cell survival [18–20]. Thus, EZH2 gain-of-function mutations likely serve to maintain expression of B-cell GC maturation programs which sustain cell proliferation and survival.

In addition to EZH2 activating mutations, overexpression of EZH2 has been discovered in over 80% B-cell lymphomas including Burkitt lymphomas, FLs [21,22]. Similarly, overexpression of EZH2 has also been identified in mantle cell lymphoma in which the transcription factor c-MYC contributes to EZH2 upregulation via repression of the EZH2 targeting miR-26a microRNA [23,24]. A correlation between elevated EZH2 expression and various myeloid malignancies has been described in patients. For instance, EZH2 is commonly overexpressed in myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) along with PRC1 members RING1 and BMI1, driving abnormal epigenetic repression of the tumor suppressor p15 (INK4B) [25]. In another study, c-MYC dysfunction has been linked to the enhancement of EZH2 expression in AML [26]. Moreover, the inhibition of EZH2 in a murine AML model prevented cancer progression by driving myeloid cell differentiation [27]. In contrast to the oncogenic function of EZH2 in cancers described above, potential tumor suppressor roles of EZH2 have also been observed in certain hematological cancers highlighting its highly lineage-specific function. Inactivation of EZH2 due to genomic aberrations including allelic deletion and frameshift mutations were found in a subset of patients with MDS, myeloproliferative neoplasms and MDS/myeloproliferative neoplasm overlap disorder [28,29]. Loss-of-function mutations or deletion of EZH2 and other subunits (e.g., SUZ12 and EED) of PRC2 have also been identified in human T-cell acute lymphoblastic leukemia [30,31]. Altogether, EZH2 exhibits dual roles as an oncogene and tumor suppressor in a context and lineage-specific manner leading to diverse hematological malignancies.

EZH2 overexpression has been observed in several solid tumors including prostate cancer, endometrial cancer, small-cell-lung cancer (SCLC), breast cancer and melanoma. In these tumors EZH2 overexpression is associated with a more aggressive and advanced form of each disease [10,32–37]. In prostate cancer, EZH2 has been shown to regulate the proper chromatin binding of androgen receptor in castration-resistant disease. Notably, EZH2 can also act independently of PRC2 and/or its histone methyltransferase activities in breast and prostate cancer [25,38–41].

Dysregulation of EZH2 function can occur through mutations in epigenetic enzyme complexes that antagonize the effects of H3K27 methylation. Recurrent cancer mutations in the H3K27 histone demethylase UTX (ubiquitously transcribed tetratricopeptide repeat gene on X chromosome) and the chromatin remodeling complexes switch/sucrose non-fermentable (SWI/SNF) are associated with altered EZH2 functions. This is best exemplified in malignant rhabdoid tumors where loss of SNF5 (INI1) results in the altered genomic occupancy of the repressive chromatin mark deposited by PRC2 at the H3K27me3 residues, leading to the repression of lineage-specific targets [42–45].

Epigenetic regulation in tumor immunity

In addition to the direct impact on tumor cell growth and survival, epigenetic changes in cancer cells can elicit a variety of mechanisms mediating escape from recognition and elimination by the immune system [2,46–49]. These mechanism include loss of expression of tumor-associated antigens or neoantigens, impairment of cell surface antigen presentation, changes in expression of immunosuppressive molecules and proinflammatory cytokines and aberrant expression of checkpoint pathway proteins such as PD-L1 [50–53].

Epigenetic changes have also been observed in the many cell types that comprise the tumor microenvironment including tumor-infiltrating immune cells. Chronic antigen stimulation of tumor-associated CD8⁺ T cells results in an ‘exhausted’ phenotype characterized by rearrangement in chromatin accessibility conferring altered transcriptional programs [47,48,54,55]. Changes in the epigenetic landscape have also been described in additional tumor-infiltrating effector immune cell populations including dendritic cells and natural killer (NK) cells where transcriptional changes confer gradual loss of their mature antigen-presenting function and cytotoxicity toward cancer cells [56–58]. Epigenetic regulators also functionally impact transcriptional programs driving increased differentiation and proliferation/survival of immunosuppressive cells such as myeloid-derived suppressor cells (MDSCs), tumor-associated macrophages and Tregs in the tumor microenvironment [59–64]. Within the tumor microenvironment, signaling molecules tied to chronic inflammation, hypoxia and altered metabolism have been associated with functional changes in the activity of epigenetic regulators influencing the production of cytokines, chemokines, growth factors and adhesion molecules [65,66].

EZH2 functions as an essential mediator of tumor immunity

Epigenetic changes conferred by EZH2 dysregulation have been identified as a major mechanism of transcriptional repression of genes that confer immunogenicity to tumor cells. This is best exemplified by the role of EZH2 in transcriptional repression of genes involved in antigen presenting pathways in cancer [67,68]. In a recent functional genomics screen conducted in K562 leukemia cells, inhibition of EZH2 and other PRC2 genes were identified as modulators of the IFN-γ response leading to MHC-I upregulation [69]. These results were further substantiated by Burr et al. who demonstrated a pivotal role for EZH2-mediated H3K27me3 repressive marks in maintenance of MHC-I silencing in MHC-I-deficient cancers. Interestingly, in neuroendocrine tumors including SCLC and neuroblastoma, EZH2 function appears to hijack a conserved developmental program in neural progenitor cells to confer immune privilege via MHC-I down-regulation. Genetic inhibition of EZH2 in a mouse genetically-engineered model of SCLC led to MHC-I up-regulation and immune-mediated tumor regression [70].

In uveal melanoma, a malignant tumor of the eye, EZH2-mediated repression of the MHC class II master control transcription factor, CIITA, prevents expression of MHC-II molecules after IFN-γ induction. These observations were extended into additional tumor types in a recent study, by Ennishi and colleagues, who noted that EZH2 activating mutations were enriched in 30% of DLBCL patients which lack MHC-II cell surface expression [64]. Moreover, 77% of EZH2 mutated cases lost either MHC-I and/or MHC-II expression in DLBCL, suggesting epigenetic repression occurs in both class I and II presentation pathways. Experimentally, introduction of either EZH2^Y641N or EZH2^Y641F mutations into B-cell lymphoma increases H3K27me3 at the promoters of NLRC5 and CIITA which could be reversed by EZH2 inhibitor treatment. These results indicate EZH2 inhibition could be a therapeutically useful approach for restoration of MHC expression.

EZH2 function has also been identified as a mechanism of transcriptional repression of chemokines and cytokines in various cancers. In a mouse MYC-driven prostate cancer model, IFNGR1 was identified as a direct target of EZH2 transcriptional repression. In this model, EZH2 overexpression was found to be regulated in a MYC-dependent manner consistent with observations in metastatic prostate cancer patients. Therapeutic treatment of mouse MYC-driven prostate cancer tumors with EZH2 inhibitors enhanced antitumor effects in response to IFN-γ treatment [71]. Overexpression of EZH2 in tumors has also been linked to the downregulation of Th1 chemokines CXCL9 and CXCL10, which are essential for mediating T cell migration to tumor sites for effective tumor killing [50,72]. In the ID8 ovarian cancer syngeneic tumor model, EZH2 inhibition in combination with DNA methyltransferase inhibitors (5-AZA dC) promoted re-expression of Th1 chemokines resulting in increased tumor-infiltrating T cells and tumor regression when combined with immune checkpoint inhibition. In a separate study, Zingg et al. demonstrated that EZH2 overexpression may serve as an immune escape mechanism during immunotherapy. In the B16-F10 mouse melanoma model treated with CTLA-4 or IL-2 blockade, increased TNF-α production from tumor-infiltrating T cells resulted in increased EZH2 expression causing loss of dominant tumor antigens and silencing of the antigen-processing and -presenting machinery [73]. Genetic or pharmacological inactivation of EZH2 in this model produced a synergistic effect with anti-CTLA-4 and IL-2, suppressing tumor growth. EZH2-mediated transcriptional repression has also been observed at the promoters of many IFN-γ stimulated genes, where EZH2 function antagonizes the chromatin remodeling activity of the SWI/SNF complex [74]. Thus, EZH2 regulation may broadly impact IFN-γ signaling pathways within tumor cells.

Immune recognition of transformed/mutated self-peptides is critical for host protection against cancer [75]. Tumor-associated antigens expressed on their cell surface can be presented by MHC molecules, which when recognized by T cells, mediate tumor rejection. Many recent studies have illustrated the importance of epigenetic regulation of tumor antigen expression and presentation. For example, X-linked tumor-associated antigens (e.g., MAGE-A1 and NY-ESO-1) showed downregulated expression via epigenetic repression, including EZH2 function, in several tumors such as ovary, lung, pancreas cancer, as well as melanoma and myeloma [76–78]. EZH2 was also identified in the epigenetic silencing of a specific type of IFN-γ-inducible antisense 3′-UTR endogenous retroviral elements called SPARCS [79]. Derepression of SPARCS following EZH2 inhibition results in double-stranded RNA generation following IFN-γ exposure and engagement of MAVS and STING innate antitumor immunity pathways.

Taken together these observations highlight tumor immune suppression as an important consequence of EZH2 function in the cancer. It is anticipated that EZH2 inhibition should elicit gene expression changes in tumors by increasing the recognition and recruitment of the immune system to the tumor microenvironment. However, certain tumors with genetic lesions including loss-of-function mutations in JAK1 or JAK2, deletion of the wild-type allele of JAK1 or JAK2 and truncation in the antigen-presenting protein B2M may be ‘hardwired’ for defects in antigen presentation that may not be overcome by EZH2 inhibition [80]. Moreover, transcriptional repression of genes may feature additional types of epigenetic repression such as DNA methylation or H3K9me3 that will not be overcome by EZH2 inhibition alone. Along these lines, the combination of EZH2 inhibitors with DNA methyltransferase inhibitors was shown to enhance reactivation of CXCL9 and CXCL10 in ovarian tumor models [50].

Role of EZH2 in regulation of effector T-cell function

Immune cell development is tightly regulated by epigenetic regulators that maintain gene expression programs governing self-renewal and cell differentiation in stem and progenitor cell lineages. The important role of EZH2 function as an essential regulator in hematopoiesis has been highlighted by multiple mouse genetic knockout models, demonstrating that EZH2 controls the balance between self-renewal and multipotency of hematopoietic stem cells through H3K27me3 suppression of developmental gene programs [81–83].

Consistent with the function of EZH2 in hematopoietic cell development, EZH2 also appears to be an important regulator of the lymphoid and myeloid cell types that comprise the tumor microenvironment. During the T-cell maturation process, EZH2-mediated transcriptional regulation plays a key role in the establishment of epigenetic programs necessary for lineage specification and divergent T-cell function (Figure 1). Tarakhovsky and colleagues first demonstrated the importance of EZH2 function in early T-cell development through adoptive transfer experiments utilizing conditional EZH2 knockout (KO) mice. In these experiments bone marrow derived from EZH2 knockout mice failed to support normal thymocyte development through arrest at the early CD4, CD8 double negative CD44^intCD25^hi stage [3]. Interestingly, this observed defect in T-cell maturation was found to be dependent on a unique cytosolic function of EZH2 as opposed to the canonical role as a histone methyltransferase.

The chromatin regulatory function of EZH2 also appears critical for the specification, functional regulation and survival of both CD4⁺ and CD8⁺ T cells (Figure 2). In mouse tumor models, tumor-infiltrating CD8⁺ T cells expressing EZH2 exhibited a survival advantage due to up-regulation of BCL-2, which is consistent with the anti-apoptotic role of BCL-2 in human effector T cells [84]. Consistent with this result, EZH2-deficient CD8⁺ T cells were also found to display enhanced T cell apoptosis during immune response [85,86]. Mechanistically, EZH2 repression of the Notch pathway suppressors NUMB and FBXW27 activates the Notch pathway in CD8⁺ T cells, stimulating T-cell cytokine expression and survival via BCL-2 signaling.

The expression level of EZH2 in tumor-associated CD8⁺ T cells can be impacted by glucose restriction in the tumor microenvironment. Low glucose conditions maintain high expression of microRNAs miR-101 and miR-26a, which constrain expression of EZH2 [87]. This finding was consistent with a previous study conducted by Long et al., which identified upregulation of miR-26a in CD8⁺ T cells to be associated with decreased EZH2 expression and impaired effector T-cell survival and function [88].

Using a mouse lymphocytic choriomeningitis virus (LCMV) infection model, Kakaradov et al. identified an important role of EZH2 in CD8⁺ T cell fate specification. EZH2, along with genes encoding other PRC2 components EED and SUZ12, is highly expressed in effector T cells relative to memory T cells upon LCMV challenge, suggesting a role of EZH2 in regulating terminal effector cell differentiation. High expression of EZH2 in terminally differentiated effector cells was associated with gains in H3K27me3 at promoters including memory associated transcription factors TCF7 and EOMES, suggesting that epigenetic silencing may be more crucial for the differentiation of terminal effector cells compared with that of memory cells [89]. Functionally, CD8⁺ T cells adoptively transferred from EZH2^-/-.mice showed reduced expansion compared with control T cells when challenged with LCMV and exhibited an impaired capacity to secrete inflammatory cytokines. It is of interest that EZH2-deficient CD8⁺ T cells were not impaired in their ability to undergo activation or proliferation, however, demonstrated increased propensity to undergo apoptosis after antigen stimulation. In an independent study using a different EZH2 knockout transgenic mouse model featuring a tamoxifen Gzmb promoter-driven Cre, EZH2 was similarly found to be required for the expansion and terminal differentiation of effector CD8⁺ T cells when challenged with LCMV. Compared with the previous study EZH2^-/- CD8⁺ T cells did not appear impacted in their survival, however, appeared skewed in their differentiation, acquiring memory-like states at the expense of effector states [90].

Contrasting results for the role of EZH2 in T cell maturation were observed in a different mouse model employing melanoma-associated antigen gp100-specific CD8⁺ T-cell receptor-transgenic Pmel-1 cells. In this study EZH2^-/- naive CD8⁺ T cells were found to have restrained memory differentiation and increased terminal effector differentiation compared with controls following adoptive transfer into B16 melanoma tumor bearing mouse [91]. Similarly, Goswami et al. did not observe any deleterious effects on effector T cells when studying the in vitro and in vivo effects of the EZH2 inhibitor CPI-1205 in human CD8⁺ T cells. Enhanced cytotoxicity and increased IFN-γ and TNF-α levels were noted in the supernatant of activated human T cells following CPI-1205 treatment, which suggested that these cells have enhanced effector function [92]. Moreover, the authors found in vivo EZH2 inhibitor treatment led to T cell-mediated tumor growth inhibition in mouse syngeneic tumor models MB49 bladder and B16-F10 melanoma either as monotherapy or in combination with anti-CTLA4. These results suggest survival of effector T-cells populations may not be detrimentally impacted by EZH2 inhibitors at least for shorter durations of treatment.

Altogether, these findings highlight the importance of EZH2-mediated epigenetic regulation during CD8⁺ T-cell maturation however they also indicate that the function of EZH2 might have opposite roles during different phases of an immune response. It is clear from these studies that under certain circumstances inhibiting EZH2 may have a negative impact on effector T-cell development, function and survival and thus should be taken into consideration when exploring combination of EZH2 inhibitors with immuno-oncology agents.

In addition to CD8⁺ T cells, the chromatin regulatory function of EZH2 appears critical for specification of naive CD4⁺ helper cells into effector lineages including Th1, Th2 and Th17 and their subsequent proliferation and survival following differentiation (Figure 1). Using a CD4-specific knockout of EZH2, Tumes et al. demonstrated that loss of EZH2 in naive CD4⁺ T cells was associated with differentiation skewing and lineage plasticity in both Th1 and Th2 cells [93]. Loss of EZH2-mediated H3K27me3 at the IFNγ, EOMES and TBX21 loci resulted in upregulation of IFN-γ production. EZH2 deficient CD4⁺ Th cells enhanced IL4, IL5 and IL13 cytokine production through upregulation of GATA3 expression under Th2 polarizing condition both in vitro and in vivo. Similarly, a separate study by Yang et al. identified EZH2-deficient CD4⁺ T cells as producing higher amounts of IFN-γ, IL-13 and IL-17 in Th1, Th2 and Th17 cells, respectively [94]. Although these data indicate that EZH2 loss promotes increased production and activation of CD4⁺ effector cells, Zhang et al. found loss of EZH2 also negatively impacted survival of Th1, Th2 and Th17 cells upon differentiation through reactivation of genes in apoptosis pathways. Consistent with this result, CD4⁺ T-cell-specific knockout of EZH2 in a mouse model of graft-versus-host disease protected mice from disease due to the inability of allogeneic effector CD4⁺ T-cell expansion [95,96]. Overall, these studies highlight the central role of EZH2 in instructing CD4⁺ T cell differentiation into Th1 and Th2 functional states as well as impacting their overall activation and survival. Recent data has highlighted the importance of both antigen-specific CD8⁺ and CD4⁺ T cells as well as MHC class II-restricted antigens for successful antitumor immune responses. As indicated in these genetic studies, inhibition of EZH2 may transiently increase CD4⁺ T-cell effector function, although longer duration of treatment may detrimentally impact the survival of effector CD4⁺ T-cell populations within the tumor microenvironment.

EZH2 function in NK cells

Although the activation program and recognition mechanism of NK cells are distinct from cytotoxic T cells, they share similar features of lytic granule exocytosis in the elimination of cancer cells [97]. Reduced numbers or impaired function of NK cells have displayed a profound impact on overall immune response, which is often associated with cancer progression [98]. Several studies have demonstrated the contribution of EZH2 epigenetic regulation on NK cell lineage commitment and functional maturation (Figure 2). An early study by Nagel et al. described the role of PRC2 in NK cell differentiation through regulation of HOXA9 and HOXA10 [99]. A subsequent study showed that the loss of EZH2 function enhances NK cell lineage commitment partially by promoting the survival of NK cell precursors. Further mechanistic studies revealed that genetic or pharmacological perturbation of EZH2 increases proliferation, activation and cytotoxicity of NK cells concomitant with upregulation of the IL-15 receptor CD12 and NK cell activating receptor NKG2D [100].

More recent papers highlighted the immunosuppressive role of EZH2 in NK cell-mediated eradication of carcinoma cells. Bugide et al. discovered that in a mouse model of hepatocellular carcinoma, EZH2 serves as a transcriptional repressor of NKG2D ligands which enables evasion of tumor cells from immune recognition by NK cells thereby conferring resistance to NK cell-mediated cytotoxicity [101]. Inhibition of EZH2 in hepatocellular carcinoma cells lead to re-expression of NKG2D ligands such as ULBP1, MICA and MICB, enhancing the ability of NK cells to mediate tumor clearance. Similarly, EZH2 inhibition in a muscle invasive bladder cancer model not only limited proliferation of tumor cells in the context of KDM6A and SWI/SNF mutations, but also promoted NK-cell activity [102]. Interestingly, EZH2 inhibitor treatment in this model leads to upregulation of genes associated with activated NK signaling including MIP-1α, ICAM1, ICAM2 and CD86 and increased expression of IFN-γ. Although the exact mechanism of epigenetic regulation of NK cell maturation and cytotoxic activity in the tumor microenvironment is not understood, EZH2 has been presented as an exciting therapeutic target to enhance NK cell-mediated antitumor immunity.

Potential for EZH2 inhibitors in reprogramming tumor immunosuppressive cells

Tregs are a unique subset of CD4⁺ T cells defined by expression of the transcription FOXP3 with inhibitory function on effector T cells during immune response. The control of EZH2 in Treg cell function was first postulated by observations of significant EZH2 upregulation in activated Treg cells as well as co-immunoprecipitation experiments which demonstrated that EZH2 forms complexes with FOXP3 in activated mouse Tregs (Figure 1). Furthermore, EZH2 knockout prevented inducible Treg differentiation in vitro through loss of EZH2-mediated H3K27me3 at FOXP3-regulated target genes [103]. Deletion of EZH2 in CD4⁺ T cells in a mouse model of autoimmune colitis reduced FOXP3⁺ Treg numbers in the spleen and mesenteric lymph nodes and also impaired Treg cell function through reduced expression of FOXP3 [94]. Interestingly, induction of antibodies against IFN-γ and IL-4 can rescue FOXP3 expression due to EZH2 deficiency, suggesting that downregulation of FOXP3 upon loss of EZH2 in CD4⁺ T cells is possibly mediated by aberrant cytokine production [86,94]. In addition, EZH2-deficient T cells showed defects in maintenance of the Treg population due to instability of lineage-specific transcriptional program upon CD28 activation [104].

More recently, studies by Wang et al. highlighted the critical role of EZH2 activity in mediating Treg function in anticancer immunity (Figure 2). Selective upregulation of EZH2 was observed in tumor-infiltrating Tregs but not in effector T cells and Treg cells in peripheral blood. Genetic deletion or pharmacological inhibition of EZH2 in vitro or in vivo destabilized FOXP3 expression thus promoting immune-mediated rejection of tumors in mouse syngeneic models [105]. These results were further confirmed by Goswami et al., who recently demonstrated that genetic depletion of EZH2 in Tregs or EZH2 catalytic inhibition using CPI-1205 elicited phenotypic and functional alterations of the Tregs leading to robust antitumor immunity. Collectively, EZH2-mediated epigenetic programs are essential for the lineage commitment as well as functional suppressive activity of Treg cells, providing rationale for exploring combination therapies of EZH2 inhibitors with immune checkpoint inhibitors [92,105]. In patients receiving ipilimumab (anti-CTLA) treatment, increased expression of EZH2 was observed in multiple immune cell types including CD4⁺ effector T cells, FOXP3⁺ Tregs and CD8⁺ T cells. Consistent with this result, T cells derived from CTLA-4^-/- KO mice were found to overexpress EZH2 in a CD28-dependent manner [104,105]. Given the importance of FOXP3⁺ Tregs in tumor immunosuppression and the important role of EZH2 in regulation of FOXP3⁺ Tregs cells, EZH2 overexpression was postulated as a mechanism of resistance to anti-CTLA-4 therapy. To this extent, EZH2 inhibition combined with anti-CTLA-4 resulted in fewer FOXP3-expressing Treg cells and augmented antitumor responses compared with anti-CTLA-4 alone [73,92,105].

MDSCs are a heterogeneous population consisting of granulocytes, macrophages or dendritic cells derived from immature myeloid cells with the capacity to suppress T cell functions [106]. Epigenetic changes influenced by the tumor microenvironment result in abnormal differentiation and function of myeloid cells leading to MDSCs. Huang et al. demonstrated that inhibition of EZH2 promotes the expansion of MDSCs during tumor development and attenuates antitumor immune responses (Figure 2). Treatment of mouse syngeneic tumor models with the EZH2 inhibitor GSK126 resulted in accumulation of CD11b⁺Gr-1⁺MDSCs in the tumor tissue as well as decreased IFN-γ producing CD8⁺ and CD4⁺ T cells [107]. Interestingly when MDSCs were depleted from the tumor microenvironment using chemotherapy agents, GSK126 promoted antitumor immune responses through increased infiltration of functional T cells into the tumor microenvironment. These data highlight the divergent roles of EZH2 function in various immune cell subtypes contributing to antitumor immune responses. Although EZH2 has been linked to the development of MDSCs that might tip the balance against antitumor immunity, the mechanistic understanding of how EZH2 regulates differentiation and migration in this population through DNA/histone methylation needs to be further elucidated.

Targeting EZH2 in cancer therapy

Unlike genetic mutations, epigenetic aberrations are reversible so targeting the relevant epigenetic factors by small molecules is potentially an efficient approach to ‘fix’ dysregulated gene/chromosome-regulatory systems caused by epigenetic changes in cancer. Along these lines, small molecule inhibitors of EZH2 and related PRC2 complexes, CPI-1205, CPI-0209, DS-3201, EPZ-6438 (tazemetostat), GSK-126 (GSK2816126), MAK683 and PF-06821497, have entered clinical trials targeting EZH2-driven epigenetic alterations in the regulation of cancer cell processes (Figures 3 & 4). These small molecules mainly inhibit EZH2 enzyme function, leading to reactivation of EZH2-target genes. DS-3201 and CPI-0209 are dual inhibitors of EHZ1 and EZH2 reported to have enhanced activity over EZH2-selective inhibitors via their targeting of potential H3K27me3 compensation driven by EZH1 paralog function [108]. Additionally, MAK683, an allosteric PRC2 inhibitor developed by Novartis, which targets the H3K27me3 binding pocket of EED, has also entered clinical studies [109].

Early clinical exploration around EZH2 inhibitors have focused on molecularly targeted patient populations in DLBCL and FL where activating mutations in EZH2 have been linked as drivers of disease. In FL, response rates to tazemetostat have been reported at 71% in patients harboring EZH2 activating mutations [110]. Interestingly, a 33% response rate observed in patients with wild-type EZH2 suggests broader function of EZH2 in the cell-of-origin of this disease. Additional clinical exploration for EZH2 inhibitors has focused largely on solid tumors with dysregulation of wild-type EZH2 function through overexpression (e.g., SCLC and prostate cancer) or mutations in other epigenetic remodeling complexes that antagonize EZH2 function such as BAP1 mutation (e.g., mesothelioma), SNF5/INI1 mutation (e.g., epithelioid sarcoma) [14,111]. In January 2020, tazemetostat was approved by the US FDA for patients with metastatic or locally advanced epithelioid sarcoma, marking the first clinical approval of an EZH2 inhibitor [112].

While clinical data assessing the efficacy of EZH2 inhibitors in additional oncology indications is still emerging, the favorable safety profile of this class of drugs raises the prospects around the utility of EZH2 inhibitors in drug combinations with chemotherapies and other targeted therapies [113]. Indeed, numerous drug combinations are currently being assessed with EZH2 inhibitors to enhance responses in certain cancers where single agent responses have been limited (Figure 3). In a recent Phase II clinical trial in relapsed or refractory follicular lymphoma, 18% of patients dosed with tazemetostat reported grade ≥3 treatment-related adverse advents, most commonly thrombocytopenia and anemia. The lack of broader hematopoietic toxicity observations in early clinical trials with EZH2 inhibitors also highlights the potential for combinations with immune checkpoint inhibitors. Interestingly, a recent report by Gounder et al. highlighted an exceptional patient response to EZH2 inhibitors due to the activation of antitumor immunity, suggesting the potential for tumor extrinsic effects for this class of drugs [114]. In this patient, tazemetostat resulted in a significant increase in tumor-infiltrating CD8⁺ T cells and immune cells expressing checkpoint regulators PD-1 and LAG-3, suggesting EZH2 inhibition can promote a sustained antitumor response. It is of consideration that genetically engineered mouse models have demonstrated that dual inhibition of EZH1 and EZH2 present with additional hematopoietic defects, thus EZH1/EZH2 dual inhibitors, as well as EED inhibitors, may exhibit unique toxicities not previously observed with EZH2-selective inhibitors [115].

Future perspective

Based off the emerging clinical and preclinical data highlighting the combination potential of EZH2 inhibitors with immune checkpoint inhibitors described in this review, clinical trials have been initiated assessing the clinical combination of CPI-1205 with ipilimumab (NCT03525795) and tazemetostat with pembrolizumab (NCT03854474) or atezolizumab (NCT02220842). In these studies, it will be critical to assess whether the immune promoting effects of EZH2 inhibition will overcome the potential for immune suppressing effects anticipated in certain immune cell populations such as effector T cells. Like other epigenetic drugs previously pursued in the clinic, one possibility would be to explore dosing schedules that ‘prime’ the immune system to checkpoint inhibitors but remove drug prior to checkpoint therapy. Further assessment of the immunomodulatory effects of EZH2 inhibitors will help inform the best utilization of this promising class of drugs in the clinic.