Article Text
PDF
Abstract
With the advances in cancer immunity regulation and immunotherapy, the effects of histone modifications on establishing antitumor immunological ability are constantly being uncovered. Developing combination therapies involving epigenetic drugs (epi-drugs) and immune checkpoint blockades or chimeric antigen receptor-T cell therapies are promising to improve the benefits of immunotherapy. Histone H3 lysine 4 trimethylation (H3K4me3) is a pivotal epigenetic modification in cancer immunity regulation, deeply involved in modulating tumor immunogenicity, reshaping tumor immune microenvironment, and regulating immune cell functions. However, how to integrate these theoretical foundations to create novel H3K4 trimethylation-based therapeutic strategies and optimize available therapies remains uncertain. In this review, we delineate the mechanisms by which H3K4me3 and its modifiers regulate antitumor immunity, and explore the therapeutic potential of the H3K4me3-related agents combined with immunotherapies. Understanding the role of H3K4me3 in cancer immunity will be instrumental in developing novel epigenetic therapies and advancing immunotherapy-based combination regimens.
- Immunotherapy
- Biomarkers, Tumor
- Therapies, Investigational
- Review
- Drug Therapy, Combination
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Background
Epigenetic mechanisms, including DNA methylation, histone modification, and diverse post-transcriptional regulations, extensively regulate cell development, proliferation, and differentiation, without changes in genetic sequences.1 Especially, histone protein modifications have the capacity of instructing transcriptional regulation by modulating DNA methylation, chromatin spatial conformation, and chromatin accessibility.2 3 Nucleosomes comprise four histone proteins, H2A, H2B, H3, H4, and DNA chains wrapping around the octamer. Different modifying groups are added to different tail domains of histones, coordinating to regulate gene expression.4 The dysregulation of these epigenetic modifications can lead to downstream pathological processes and recently has been defined as a novel hallmark of cancer.5
Histone methylation is the most well-documented histone modification pattern. Unlike histone acetylation marks which are always associated with active transcription,6 histone methylation has more complicated regulatory effects on transcription activation. H3K4, H3K9, H3K27, H3K36, H3K79, and H4K20 are essential methylated sites that can be modified into monomethylation, dimethylation, and trimethylation states,7 and these diversified histone methylations compose a sophisticated regulatory network to engage in cancer processes.8 Especially, H3K4 trimethylation is a near-universal histone modification in developmental biology and cancer processes.9–11 H3K4me3 is recognized as a transcription-activated mark, usually depositing at promoter regions and transcription start sites.12 Many studies have demonstrated that lysine methyltransferases (KMTs) and lysine demethylases (KDMs) that specifically catalyze H3K4 trimethylation extensively engage in cancer immunity regulation.13 14
Although immunotherapies have achieved dramatic and sustained responses in a subset of patients with cancer, optimized combination therapy strategies are necessary to further improve patients’ responses given the high non-response and resistance frequency of monotherapy. Epigenetic drugs (epi-drugs) targeting epigenetic modifiers, correct unbalanced epigenetic signatures of cancer cells to confine their malignant phenotypes by targeting epigenetic modifiers. Epi-drugs are extensively involved in modulating antitumor immunity, so they are expected to be promising candidates for combination therapies with immunotherapies.15 Preclinical and clinical trials have shown the efficacy of DNA methyltransferase inhibitors (DNMTi) and histone deacetylase inhibitors (HDACi) in improving patients’ benefits from immunotherapies.16–20 Preclinical trials with H3K4me3 targeting also display positive effects, such as LSD1 ablation reversing the resistance of melanoma mice to anti-PD-1 treatment by promoting cancer cell immunogenicity and T cell infiltration.21 Although encouraging preclinical results affirm the feasibility of combination regimens with H3K4me3 targeting in immunotherapies; however, the available clinical value and scope of application are vague. More evidence is needed to resolve the inexcitability and additional toxicity of combination therapy and to assist patient stratification.
To provide an overall cancer immunity regulatory network connected by H3K4me3 modification, here we first introduce the modifiers responsible for H3K4me3 and the mechanism by which H3K4me3 regulates transcription. We give an elucidated description of how H3K4me3 participates in cancer immunity establishment. We highlight the rationality and feasibility of therapeutic combination with H3K4me3-targeted therapy and immunotherapies. These understandings will help to develop novel therapeutic regimens using H3K4-me3-associated epigenetic regulation.
H3K4me3-related modifiers
A variety of KMTs and KDMs regulate H3K4 methylation, and their functional changes can affect the local or genome-wide H3K4me3 distribution, then elicit a series of downstream molecular signaling transmissions.
Writers
H3K4me3 is a conserved histone methylation pattern from yeast to humans. In yeast, H3K4 monomethylation, dimethylation, and trimethylation is modified by Set1 which forms a protein complex named complex of proteins associated with SET1 (COMPASS) with other subunits.22 23 Six COMPASS family members exist in human cells: SET1A and SET1B form two COMPASS complexes, and mixed lineage leukemia 1 (MLL1), MLL2, MLL3, and MLL4 form four COMPASS-like complexes.24 25 These complexes have compositional and functional similarities.26 The SET1A complex and SET1B complex all have five non-catalytic components, including CXXC finger protein 1 (CFP1), RBBP5, ASH2, WDR5, WDR82, and DPY30, as well as other complex-specific subunits. Some subunits are critical for the function execution of the methyltransferase complex.27 The H3K4 methylation effect of KMTs has substrate specificity and functional complementarity. SET1A and SET1B are responsible for H3K4 dimethylation and trimethylation of the global genome and broad H3K4 methylation, especially in the promoter regions of active transcription genes.28 MLL1 and MLL2 promote the H3K4me2 and H3K4me3 deposition within promoters, whereas MLL3 and MLL4 contribute to H3K4 monomethylation within enhancers.29
In addition, SMYD2 and PRDM9 are involved in catalyzing H3K4 methylation.30 31 H3K4-targeted KMTs (SETD1A, SETD1B, MLL1, MLL2, MLL3, MLL4, SMYD2, PRDM9) exhibit distinct mutation types, frequencies, and co-mutation rates in different cancer types (figure 1A and C,D), which may be related to the vulnerability of cancers to different KMT inhibitors. Most H3K4me3-related KMT/KDM genes comutate with each other in each patient from the TCGA pan-cancer cohort32 (figure 1D). This is contrary to the feature of mutually exclusive mutations in tumor driver genes.33 Consequently, the degree of oncogenic addiction of most cancers to dysregulation of H3K4me3 modification is likely low.
The mutation landscape of H3K4-related methyltransferase and demethylase genes in the TCGA pan-cancer cohort.32 (A,B) The summational mutation frequencies and types of (A) H3K4 methyltransferases (SETD1A, SETD1B, MLL1, MLL2, MLL3, MLL4, SMYD2, PRDM9) and (B) H3K4 demethylases (LSD1, LSD2, KDM5A, KDM5B, KDM5C, KDM5D) in various cancer types. The cancer abbreviations are as follows. Gene mutation data analysis and visualization are conducted by cbioportal (https://www.cbioportal.org/).165 166 (C) Heatmap for the mutation status of each KDM/KMT gene in different cancers. The color levels represent the percentage of patients with each kind of gene mutation in the corresponding cancer type. The gene mutation data is downloaded from cbioportal (https://www.cbioportal.org/) and visualized by the R package ‘pheatmap’. (D) The comutation of KDM/KMT genes in pan-cancer. The color lump represents the values of log2Odds Ratio. And all comutation combinations presented in the plot (log2OR>0) are statistically significant. BIDC, breast invasive ductal carcinoma; BSCC, bladder squamous cell carcinoma; CHOL, cholangio carcinoma; CLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; COAD, colon adenocarcinoma; DLBC, diffuse large B-cell lymphoma; DNMB, desmoplastic/nodular medulloblastoma; ESCA, esophageal carcinoma; FL, follicular lymphoma; GBM, glioblastoma; HNSC, head and neck squamous cell carcinoma; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; KIRP, papillary renal papillary cell carcinoma; LC/A MB, large cell/anaplastic medulloblastoma; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; MB, medulloblastoma; OS, osteosarcoma; OV, serous ovarian cancer; PA, pilocytic astrocytoma; PAAD, pancreatic adenocarcinoma; pNET, pancreatic neuroendocrine tumor; PRAD, prostate adenocarcinoma; PTC, papillary thyroid cancer; SKCM, skin cutaneous melanoma; STAD, stomach adenocarcinoma; UCEC, uterine corpus endometrial carcinoma; USC, uterine serous carcinoma/uterine papillary serous carcinoma.
Erasers
Histone lysine demethylases are divided into two subgroups: the flavin adenine dinucleotide (FAD)-dependent LSD1 and LSD2 (named KDM1 subfamily), and members from JMJD subfamily (KDM2-KDM7) characterized by conserved JmjC domain. LSD1 and LSD2 remove methyl from H3K4me1 or H3K4me2 in a FAD-dependent oxidative reaction.34 35 KDM5A-D from the JMJD subfamily mediate the H3K4me2 and H3K4me3 demethylation.36 These H3K4-targeted demethylases participate in cancer initiation, progression, and therapy resistance through changing oncogene or tumor suppressor gene expression,37 38 and their mutation signatures in the pan-cancer cohort also indicate the potential epigenetic defects of different cancers (figure 1B–D).
Readers
Reader proteins recognize specific histone modification through binding pockets and induce the downstream biological outcomes.39 Some specialized reader proteins can identify H3K4me3, such as TAF3 and SGF2 from the inhibitor of growth (ING) family.40 However, except for the effects of ING4 protein on innate immune response intensity,41 no clarified report about whether H3K4me3-related readers play roles in the cancer-immunity cycle. Of interest, some writers and erasers can also ‘read’ the histone substrates modified by the plant homeodomain (PHD) of subunits that preferentially recognize different H3K4 methylation states.42 43
The role of H3K4me3 in transcriptional regulation
H3K4me3 is recognized as a euchromatin modification mark. Active genes have more sharp and narrow peaks of H3K4me3 deposition within the transcriptional start sites.9 44 Mechanistically, H3K4 methyltransferases bind to the RNA polymerase II with serine 5 phosphorylated and initiate H3K4 methylation, which may account for its characteristic distribution pattern.45 Another newly elucidated distribution patterns, broad H3K4me3 domains referring to the wide distribution of H3K4me3 marks over cell-type-specific gene loci, are essential for the identification and function of cells.46 47 For example, ubiquitous tumor-suppressor genes, such as TP53 and PTEN, are marked by broad H3K4me3 modification in normal cells, whereas the shortening of broad H3K4me3 mark emerges in cancer cells. In contrast, cancer-type-specific tumor suppressors are only characterized by broad H3K4me3 in specific cell types.48 Additionally, broad H3K4me3 plays a role in autoimmune disease by regulating the expression of immune-responsive genes49 and is implicated in T cell receptor signaling.48 These findings prompt us to further explore the role of broad H3K4me3 in cancer immunity.
H3K4me3 activates gene transcription in many ways. Studies have found that the interaction between H3K4me3 and the PHD finger domain of transcription initiation factor TAF3 can recruit TFIID to active genes and regulate the preinitiation complex’s assembly in a TATA box cooperation manner or independent manner.50 In addition, H3K4me3 is coupled with nucleosome remodeling factor (NURF) to maintain Hox expression during development through NURF’s PHD finger.51 H3K4me3 can also interplay with other histone modifications to coregulate gene transcription. For example, bivalent chromatin is characterized by the co-occurrence of H3K4me3 and H3K27me3 at histone tails.52 The simultaneous presence of active and suppressive histone modification marks largely maintains the poised transcription states of lineage-specific genes in immune cells and embryonic stem cells. The alternation of bivalency triggers the shift of gene transcription between activation, poise, and repression states in response to stimulators.53 In addition, H3K4me3 can cooperate with H4K16 acetylation to active homeotic gene transcription.54 H3K4me3 is also associated with DNA methylation regulation. 5-Methylcytosine (5mC) is the most common modification type of DNA. The methylation commonly occurs in cytosine which is followed by a guanine, and numerous clusters of the 5’-to-3’ CpG pairs forming CpG islands (CGIs) which preferentially distribute within the promoter regions to regulate transcription activity.55 H3K4me3 marks are highly coupled to unmethylated CGIs.55 Mechanistically, methylated H3K4 abrogates the accessibility of DNMT3A/DNMT3B to the chromatin and prevents gene silencing mediated by DNA methylation.56
The role of H3K4me3 in the cancer-immunity cycle
H3K4me3 modulates cancer immunogenicity
The immunogenicity of cancer cells largely determines the initiation and intensity of antitumor immunity, including cancer antigens and major histocompatibility complex class Ⅰ (MHC-Ⅰ) expression. More neoantigen generation increases the opportunity for immune cells to recognize tumor cells and upregulated major MHC-Ⅰ promotes the neoantigen presentation.57 H3K4me3 plays an important role in cancer immunogenicity modulation by regulating MHC-Ⅰ and neoantigen expression (figure 2A).
Overview of H3K4me3-related antitumor immune regulatory mechanisms in the TME. (A) H3K4me3-related modifiers regulate H3K4me3 deposition to control gene transcription and initiate downstream signaling. (Step 1) Aberrantly activated tumor antigen genes like cancer/testis antigens promote neoantigen expressions to enhance cancer immunogenicity and the cancer-killing effects of cytotoxic T cells. (Step 2) Ectopic deposition of H3K4me3 at endogenous retrovirus regions causes dsRNA generation, sensed by RIG-I and MDA5 to activate IRF3 and IRF7 and initiate type Ⅰ and type Ⅲ IFN expression. (Step 3) Increased expression of cancer-derived IFN molecules regulates IFN-stimulated gene expression in cancer cells in an autocrine manner and targets IFN receptors distributed on the cell surface of TME immune cells like dendritic cells to modulate their antitumor functions in a paracrine manner. (Step 4) H3K4me3 regulates the expression of various chemokines in cancer cells, facilitating immune cell infiltration in TME. (B) H3K4me3 modifiers regulate several immune cells in TME. The red lines represent the resultant regulation mediated by modifiers is immunostimulatory, while the green lines represent immunosuppressive effects. dsRNA, double-strand RNA; IRF, interferon regulatory factor; MDSC, myeloid-derived suppressor cell; TAM, tumor-associated macrophage.
H3K4me3 regulates the expression of cancer antigens
The production and release of cancer antigens is the first step of the cancer-immunity cycle.58 The dysregulation of histone modification can reactivate those genes that are silent in normal cells during cellular malignant transformation. For example, cancer/testis antigens (CTAs) are only expressed in male germ cells, whose derepression in cancer cells regularly results from DNA demethylation and active histone modification, such as histone acetylation, methylation, especially out-of-control H3K4me3 formation, and transcription factor regulation.59 60 Chromatin Immunoprecipitation (ChIP) experiments showed that H3K4me3 marks are enriched at the promoters of NY-ESO-1, MAGE-A1, and MAGE-A3 in H1299 cells, accompanied by high expression of these antigens.59 The low H3K4me2 abundance depositing within the MAGE-A11 promoters is also associated with MAGE-A11 silence.61 So far, tumor vaccines or CAR-T cells targeting CTAs have shown favorable therapeutic effects in some patients with cancer.62 The combined use of epigenetic therapies that promote H3K4me3 activation can further improve CTA-targeted therapy efficacy by increasing CTA expression in cancers.
The aberrant endogenous retroviruses (ERVs) transcription induced by uncontrolled histone modification is another cancer neoantigen source. ERV is a type of transposon, and approximately 9% of the human or mouse genome is identified as ERVs.63–65 Some aberrant transcription-activated ERVs can produce immunogenic proteins presented on the cancer cell surface by MHC-I molecules as neoantigens.66 In addition, ERVs can act as alternative promoters, enhancers, or cryptic transcription initiation sites to control oncogenes' expression and produce aberrant protein-coding mRNAs in cancers,67 68 which is termed ‘onco-exaptation’.69 70 However, whether H3K4me3 regulates ERVs transcription to generate neoantigens remains uncertain.
The production of cancer neoantigen on the internal and external stresses is critical for cancer immune activation and immunotherapy response.71 Studies demonstrated that some DNMTi and HDACi can play immunomodulatory roles in cancer by mediating neoantigen generation and presentation,72 73 which provides the rationality for their collaboration with immunotherapy to improve treatment response. Whether histone methylation-targeted treatment also engages in cancer neoantigen modulation needs further investigation.
H3K4me3 regulates antigen presentation
Antigen stimulation is a critical initiator for T-cell activation. Cancer cells express MHC-I, which presents tumor antigens to antigen-specific CD8+T cells and activate their killing ability. Loss of MHC-I in cancer cells leads to cancer immune evasion.
Epigenetic mechanisms are involved in regulating MHC molecules’ temporal and spatial expression in cancer cells. Studies found that the presence of bivalent chromatin at promoters maintained by polycomb repressive complex-2 (PRC2) restrains the MHC-I antigen processing pathway, which supports the evasion of cancer cells from T cell-mediated killing effects.74 The PRC2 complex inhibitor EED226 enhances antitumor effects in nasopharyngeal carcinoma by upregulating MHC-Ⅰ gene expression.75 Therefore, it is implied that interventions facilitating more H3K4me3 depositions can upregulate MHC-Ⅰ expression. The epigenetic regulation of MHC molecules driven by H3K4me3 is also found in T cells76 and B cells.77 CⅡTA is essential for the transcriptional activity of the MHC-Ⅱ promoter, and the expression levels of CⅡTA are highly correlated with active histone methylation characteristics within the promoter region.76 Interferon-γ (IFN-γ) and HDACi can induce MHC-Ⅱ expression by enhancing histone acetylation and H3K4me3 enrichment at the MHC-Ⅱ promoter.78 The overexpression of subunits from COMPASS/MLL complexes increases MHC-Ⅱ and CⅡTA expression through catalyzing H3K4 trimethylation at their promoters.79 Collectively, enhancing H3K4me3 enrichment to MHC gene loci is a feasible way to confront immune evasion caused by low MHC expression.
H3K4me3 regulates interferon-related anticancer innate immunity
IFN signaling and innate immune cells play an indispensable role in antitumor immunity. Interestingly, the activation of genomic ERV expression mediated by histone modification can activate subsequent IFN-mediated innate immunity.65 80 81 The double-stranded RNA (dsRNA) produced from rejuvenescent ERVs transcription is sensed by cytoplasmic pattern recognition receptors like MDA5 and RIG-Ⅰ and further activates the downstream IFN production, transport, and secretion signaling pathway, which is an essential mechanism for innate immunity-based antitumor effects. The phenotype characterized by activated dsRNA-IFN signaling is called ‘viral mimicry’, which simulates the IFN signaling activation triggered by virus infection in mammalian cells.82 83 LSD1 (H3K4me3 demethylase) ablation in cancer cells promotes ERVs expression, and consequently leads to dsRNA stress and type Ⅰ IFN expression, thereby enhancing antitumor immunity. Furthermore, loss of LSD1 improves the response to anti-PD-1 therapy in the immune checkpoint blockade (ICB)-refractory mouse melanoma model, which suggests LSD1 is a novel combined therapeutic target for immunotherapy.21 Moreover, H3K4me3 promotes the spreading of euchromatin from the host gene toward unmethylated ERV insertions, further implying that H3K4me3 is a core type of epigenetic regulation of ERVs.84
Cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) axis senses aberrant cytosolic DNA and triggers innate immune responses. H3K4me3-related modifiers also modulate the activation of the cGAS-STING pathway. KDM5B expression negatively correlates with STING expression level and CD8+T cell infiltration in multiple cancer types. Mechanistically, KDM5B and KDM5C bind to the STING promoter and suppress STING transcription by reducing H3K4 trimethylation, which is consistent with that the KDM5 blockade induces robust IFN response.85
H3K4me3 regulates the fate of immune cells
Immune cells’ functional states are determined by the transcriptional profile of differentiation- and effector-related genes, which are, to a large extent, manipulated by epigenetic mechanisms.83 We focus on the H3K4me3-related epigenetic modulation in various immune cells, including antitumor effector immune cells, such as dendritic cells (DCs), natural killer cells (NK cells), and cytotoxic T cells, as well as immune-suppressive cells, including M2-like macrophages, regulatory T cells (Tregs), and myeloid-derived suppressor cells (MDSCs).
T cells
T cells that dominate adaptive antitumor immunity are classified into three major subsets according to their functional states: cytotoxic T cells, helper T (Th) cells, and memory T cells. T cell evolution is driven by dynamic activation or suppression of differentiation-related genes, and these transcription alternations are commonly regulated by epigenetic mechanisms (figure 3).57 86 CFP1-mediated H3K4 trimethylation within survival-associated gene promoters is critical for thymocyte development.87 Notably, bivalent chromatin engages in early T cell selection. In double-negative thymocytes, the CD8A locus is enriched by the bivalent chromatin. On cytokine stimulations, T cells undergo rapid proliferation and coexpress CD4 and CD8 to become double-positive cells. At this development phase, the CD8A locus was gradually dominated by H3K4me3 modification. After positive selection, H3K27me3 exclusively occupies the locus in CD4 single-positive cells.88 The dysregulation of H3K4me3 modification can impair normal T-lineage development, involved in T-cell malignant transformation including T-cell acute leukemia89 and T-cell lymphoma,90 suggesting novel therapeutic schemes combined with H3K4me3-targeted drugs is a promising strategy for T cell malignancy therapy.
Bivalent chromatin, H3K4me3, and H3K27me3 marks regulate T cell differentiation. CD8A expression is controlled by bivalent histone modification. Increased H3K4me3 deposition at the CD8A locus triggers DN thymocytes to differentiate into DP thymocytes. The upregulated H3K27me3 and absent H3K4me3 marks at the CD8A locus promote DP thymocytes differentiation toward CD4+T cells, and H3K4me3-highly-labeled CCR7 and TCF7 promote DP cell differentiation toward naïve CD8+T cells. H3K4me3 modulates function-determining gene expressions to control the evolution of T cells, including effector/memory T cells differentiation and T cell exhaustion processes. H3K4me3 and H3K27me3 facilitate helper T cell differentiation toward Th1 cells, Th2 cells, Th17 cells, and Tregs through controlling differentiation-inducible gene transcriptions. DN, double-negative; DP, double-positive.
Naïve T cells, effector T cells, and memory T cells have distinct histone methylation landscapes. Herein, the bivalent histone modification extensively regulates the functional differentiation of T cells.91 For example, CCR7 and TCF7 gene loci are hypomethylated and enriched with H3K4me3 marks in naïve T cells.92 After activity stimulation and clone expansion, effector genes like IFNG, TBX21, GZMA, GZMB, PRF1, and KLRG1 are demethylated, along with high H3K4me3 occupations.92 Meanwhile, H3K4me3 marks are lost on the memory T cell-related gene loci, including FOXO1, KLF2, LEF1, TCF7, IL2RA, CD27, TNF, CCR7, and SELL.93–95 T cells can be further subdivided into five subsets, including peripheral CD8+T cells, naïve T cells, memory stem T cells (TSCM), central memory T cells (TCM), and effector memory T cells (TEM). Hierarchical cluster analysis identified 5916 T cell function-associated genes which are correlated with H3K4me3 levels differentially expressed in the five subsets.92 These findings implied the multieffect of H3K4me3 on T-cell differentiation and functional interconversion.
Under physiological conditions, the activation of the immune system needs to keep an equilibrium state to avoid autoimmune attack from hyperactivation or uncontrolled infection/poor tumor surveillance from immunosuppression. The engagement of immune checkpoint molecules plays an essential role in timely blocking immune activities. They include cytotoxic T lymphocyte-associated protein 4, programmed cell death protein 1 (PD-1), lymphocyte activation gene-3, T cell immunoglobulin and mucin-domain containing-3 (TIM-3), and T cell immunoglobulin and ITIM domain (TIGIT), expressed on immune cells and mediates immunosuppressive signalings when bound by their ligands during immune responses.96 97 Especially, these immune checkpoints can lead to cancer immune evasion. Up to now, ICBs that abrogate the immune checkpoint signaling have achieved favorable therapeutic effects in subsets of patients.98 For example, the inhibitors of PD-1 and its ligand programmed death-ligand 1 (PD-L1) which is expressed on cancer cells to activate PD-1 by direct binding, can reverse T cell exhaustion and enhance their antitumor efficiency. Anti-PD-(L)1 drugs have shown effective in melanoma, non-small cell lung cancer, urothelial carcinoma, and so on.98 99 They are recognized as the standard and gold immunotherapy regimens. The transcriptional activation of immune checkpoint molecules can be induced by persistent demethylation and decreasing repressive histone marks on their gene loci, and then the cancer-killing ability of T cells and functions of other effector immune cells will be impaired.100 For example, more H3K4me3 enrichment significantly contributed to the high expression of PD-L1 and TOX2 in tumor tissues of colorectal cancer.101 A study found that KDM5A overexpression upregulated PD-L1 by binding to PTEN’s promoter and inhibiting its expression in cancer cells. Molecular drugs that increase KDM5A abundance cooperated with anti-PD-1 therapy to improve therapeutic response in mouse models of multiple cancers.102 Moreover, in vivo CRISPR mutagenesis screening found that KMT2D deficiency can intensify antitumor immunity and sensitize tumor to ICB therapy through modulation of the transcriptional accessibility of effector gene regions and activation of transposable elements.103
H3K4me3 also modulates chemokine-inducible T-cell trafficking and infiltration in the tumor environment. For example, KDM5B binds to the promoters of CXCL11, CXCL9, and CXCL10 to inhibit their expression through the elimination of H3K4me3 deposition in esophageal squamous cell carcinoma cells. These chemokines are responsible for attracting effector T cells into tumors. In the mice model, KDM5B neutralization can promote CD8+T cell trafficking into tumor beds and improve antitumor immunity.104
Specific transcription factors guide the lineage differentiation of CD4+T cells, and different histone methylation marks preferentially promote specific helper T cells (Th cells) lineage differentiation (figure 3).105 The inactivation of EZH2 and PRC2 enhances Th cell differentiation and the generation of IFN-γ and Th2 cytokine-producing populations through regulating H3K27me3 deposition.106 On the contrary, the IFN-γ locus showed H3K4me2 absence in naïve CD4+T cells, while Th2 cells had slight H3K4me2 distribution within the IFN-γ gene region, and H3K4me2 was enriched on the IFN-γ locus in Th1 cells.107
The IL-6/STAT3 pathway is responsible for initiating the Th17 cells differentiation program. The subunit of SET1 complex, CFP1, binds to IL-6Rα and facilitates adequate occupancy of H3K4me3 at its promoters to maintain sufficient IL-6Rα expression levels and IL-6/STAT3 signaling activity. In addition, STAT3 activation mediated by IL-6 facilitates the H3K4me3 marks to deposit on the Il-17 locus and promotes IL-17 expression for Th17 cell differentiation.108 109 The reduced methyl donor S-adenosine-L-methionine caused by dietary restriction of methionine impairs H3K4me3 formation at the promoters of critical genes involved in the expansion and cytokine production of Th17 cells.110
As the indispensable element to sustain peripheral tolerance, Tregs act as a brake to restrain immune activation in the context of cancer. FOXP3 is critical for Tregs development and maintenance, whose expression and stability are determined by the methylation status of three conserved non-coding sequences (CNS1-CNS3) of the FOXP3 gene.111 A study demonstrated that H3K4me3 is prominent at the FOXP3 gene locus in CD25+CD4+Treg cells compared with CD25-CD4+T cells, suggesting H3K4me3 enrichment rendered the FOXP3 locus into an accessible chromatin structure.112 Beyond H3K4me3, H3K4me1 exhibits a more persistent modification enriched at the CNS3 of FOXP3 locus in naïve CD4+T cells and is maintained inherently in T cell subtypes.113 MLL4 catalyzes H3K4 methylation at the distant unbound enhancers of the regulatory elements for FOXP3 by chromatin looping, thereby enhancing FOXP3 expression and inducing Treg cell differentiation.114 Moreover, bivalent histone modification also maintains the plasticity of Tregs via the H3K4me3 and H3K27me3 co-occupancy within GATA3 and TBX21 gene loci.115
Taken together, in terms of the remarkable correlation between gene expression states and H3K4me3 distribution patterns in T cells, quantification of local histone marks may be a novel predictive marker for evaluating T cell differentiated states and patients’ immunotherapy responses. Given the regulatory capacity of H3K4me3 to cover both tumor-inhibiting and tumor-promoting immune cells, pharmacological interference of H3K4me3 modification signature is likely to induce complicated rather than pure ‘stimulative’ or ‘repressive’ effects on cancer immunity. The relative strength of the two opposed effects is dependent on specific individual contexts, stressing the importance of patient classification and predictive markers in the field of epi-therapeutics.
Macrophages
Macrophages are essential components of innate immunity in immunogenic cancer killing. The tumor microenvironment can reprogram the epigenomic characteristics of tumor-infiltrating macrophages.116 H3K4me3-related modifiers’ dysfunction promotes cancer progression by regulating the heterogenization of tumor-associated macrophages (TAMs). Studies demonstrated that the expression of M2-like macrophage markers was handled by the antagonistic interplay between H3K4 and H3K27 methylation.117 Herein, H3K4me3 decorates more than 20 000 promoters in monocyte-derived macrophages.118 WDR5 is a subunit of SET1A/SET1B complexes, which supports M2 polarization of TAMs by catalyzing H3K4me3 formation within the corresponding gene promoter regions in an integrated methionine metabolism (figure 2B).119 Furthermore, WDR5 upregulates proinflammatory cytokine production of macrophages on chemotherapy stimulation, a novel mechanism where chemotherapy promotes tumor progression.120 In contrast, pharmacological inhibition of LSD1 can activate the transcription of M1-like markers in the murine triple-negative breast cancer model.121 In addition, lncRNA can recruit hnRNPL to the CCL2 promoter and then promote H3K4me3 formation. Activated CCL2 expression promotes macrophage trafficking into the tumor and increased lymphatic metastasis.122
Dendritic cells
DCs present tumor-associated antigens to stimulate T cell activation and initiate antitumor immunity. H3K4me3 is widely implicated in shaping the transcriptional profiles of DCs and then regulating their maturation and cancer-killing abilities. The deposition of H3K4me3 enhances the expression of genes related to DCs maturation and modulates integrin-regulated phenotype, strengthening the tumor rejection of DCs.123 Intracellular heat shock protein 70-like protein (HSP70L1) abrogates the human DCs maturation process by inhibiting the recruitment of ASH2L on the CD40, CD86, HLA-DR, and STAT3 promoters, reducing H3K4me3 abundance within these gene promoters.124 H3K4me3 absence at the CCR7 promoter in monocyte-derived DCs also inhibits the migration of DCs into TME in patients with multiple myeloma.125
Some agents can regulate DCs activation and functional differentiation by interacting with H3K4-associated modifiers. In plasmacytoid DCs, diethylhexyl phthalate restrains WDR5 nuclear translocation and inhibits H3K4 trimethylation at the IRF7 gene promoter region, disturbing the anti-infective response.126 Prostaglandin Ⅰ2 and long-acting beta2-adrenoreceptor agonists modulate the cytokine expression and function of DCs by H3K4 methylation-related epigenetic regulation.127 128 Although the mentioned studies are not conducted in the cancer context, they suggest the extensive regulatory effects of non-epigenetic agents on immune cells via epigenetic reprogramming, which may be novel cancer-inhibiting therapies.
Myeloid-derived suppressor cells
MDSCs are immunosuppressive cellular components within the tumor microenvironment. Many H3K4-related modifiers participate in regulating MDSCs infiltration and differentiation. For example, the expression of KDM1A is positively correlated with MDSCs infiltration levels in most cancer types.129 Ectopic expression of WDR5, ASH2L, and MLL1 promoted differentiation of tumor-cocultured polymorphonuclear MDSCs (PMN-MDSCs) into mature neutrophil-like cells as restrained their expansion and function, indicating the value of MLL1 complex as an immunological therapeutic target.130 However, the granulocyte-macrophage colony-stimulating factor and IL-6 secreted by cancer cells can activate STAT3/CEBPβ in PMN-MDSCs, which induces the expression of several microRNAs targeting the transcripts of MLL1 complex subunits, thereby inhibiting their expression.130 LncRNA lnc-chop supported the cancer-promoting functions of MDSCs by facilitating H3K4me3 deposition on the promoter region of immunosuppressive factors, such as arginase-1, NO synthase 2, NADPH oxidase 2, and cyclooxygenase-2, leading to chromatin permission and active transcription.131 Therefore, H3K4me3 performs bidirectional effects on cancer immunity in an MDSC function-related manner.
Natural killer cells
NK cells have the capability of killing cancer cells directly. Notably, compelling evidence shows that adoptive infusion with chimeric antigen receptor NK cell therapy was superior to CAR-T cell therapy in reducing potential toxicity and more allogeneic cell donors,132 so the strategies for optimizing NK cell therapy are of great significance. Many NK activation-related genes are chromosomal bivalency, highlighting the epigenetic modulatory mechanism behind the poised chromatin state and rapid activation of NK cells on stimulation.133 Demethylase KDM5A is required for resting NK cell activation by repressing the expression of the suppressor of cytokine signaling 1 (SOCS1). Loss of KDM5A enhances H3K4me3 deposition within the SOCS1 promoter region and leads chromatin states to be open, which increases SOCS1 expression and subsequently inhibits JAK/STAT4 activation.134 In addition, RUNX1 and RUNX3 are responsible for virus-specific NK cell expansion and memory cell generation. A study found that STAT4 induced H3K4me3 formation at the RUNX gene loci promoters to enhance RUNX expression in activated NK cells.135 The epigenetic reprogramming in NK cell functions is a topic that warrants more investigation to improve available NK cell therapies and develop novel combinatorial therapies.
The opportunities of H3K4me3-related agents in cancer immunotherapy
The role of histone methylation-related agents in cancer treatment
Except for Tazemetostat (EZH2 inhibitor), most KMT/KDM-targeted drugs are primarily examined in phase Ⅰ or Ⅱ clinical trials now. EZH2, DOT1L, and LSD1 are the three well-informed histone methylation-related targets for cancer epigenetic therapy,136 and their small molecular inhibitors have shown favorable responses in preclinical trials, encouraging more exploration in the development of histone methylation drugs. The LSD1 inhibitors seclidemstat, JBI-802, and CC-90011 are being tested in patients with solid tumors or hematological malignancies. But these trials are ongoing and the results are not available.
Notably, the treatment scope of these epi-drugs mainly encompasses hematological cancers (table 1). How to cover broader cancer types is the major obstacle to boosting epi-drugs development. The values of these drugs in solid tumors are less proven, especially some natures of solid tumors lead to their insensitivity to epigenetic therapy.136 Mechanisms contributing to poor response may lie in (1) histone methylation-related genes generally have fewer mutations in solid tumors, and major chromatin remodeling mutations are unknown significance or loss-of-function, which is not conducive to drug development;136 (2) solid tumors are surrounded by complicated TME. Epi-drugs interfere with each cellular composition within the microenvironment in tumor-promoting or tumor-repressing manners, and the net effect may be insufficient to curb tumor progression;137 and (3) solid tumors originate from more-differentiated cells, which are more challenging to undergo epigenetic reprogramming compared with malignant hematological cells.16 Therefore, the poor outcome of monotherapy renders combination therapy a promising strategy to improve histone methylation-related epi-drugs efficacy in more cancer types.
Overview of the major histone methylation modifiers-targeted drugs and combination therapies with immune checkpoint blockades in clinical trials
Off-target effects of epigenetic therapies are also critical issues to be focused on. First, the universal existence of histone methylation within the cellular genome gives rise to the non-intentional gene targeting by histone methylation-related agents, which may trigger unpredictable changes in gene function.136 For example, H3K4me3 engages in the regulation of both effector T cells and Tregs differentiation, and pharmacological inhibition of specific H3K4me3 modifiers can stimulate cancer immunity in some cancer and genetic contexts, while in other circumstances, the therapeutic strategy may not ensure establish a tumor microenvironment where antitumor effects prevail. Some modifiers modulate different histone modification sites in direct or indirect manners, and this characteristic is another possible cause of off-target effects. LSD1 demethylates the active mark H3K4 methylation and the suppressive mark H3K9 methylation, thus playing as an activator or repressor in different contexts.138 Likewise, H3K4me3 methyltransferase MLL1 moderately mediates methylation on H3 arginine8 as well as H3K9.139 H3K4me3 demethylase KDM5B mediates cancer immunosuppression in an indirect manner where KDM5B recruits SETDB1 (H3K9 methyltransferase) to repress ERVs expression.140 Therefore, the multi-target regulation of these modifiers is likely to trigger off-target effects in practical treatment with their inhibitors or agonists, especially since these targets have opposite downstream responses. In addition, except for nuclear regulation, histone modifiers have been found to play modification roles in cytoplasms.136 141–143 And these non-histone modifications may be another origin of the off-target effects of modifier inhibitors.
Even so, histone methylation-targeted drugs usually show mild and manageable adverse effects (AEs) in clinical trials, and the AEs are mainly myelosuppression and other non-hematologic events, including fatigue, infections, hypokalemia, diarrhea.144 To reduce and overcome the off-target effects of histone modification-targeted therapies, a deeper understanding of the regulatory mechanisms of these modifiers is required, covering the histone and non-histone modifications. Researchers should further refine the dose and cycle of drug administration of less-informed epi-drugs to provide feasible suggestions for their application in clinical trials.
Combination therapies with ICBs
Numerous clinical trials actively investigate the combination strategy of epigenetic therapies and immunotherapies (table 1). DNMTi and HDACi are the mainstream drugs in combination with ICBs, which have shown good outcomes in selective solid tumors.17 18 145 146 Nonetheless, the combinatorial therapy data of other histone methylation-related agents are lacking. Many preclinical studies underpinned the therapeutic potential of H3K4me3-related modifiers combined with ICBs. H3K4 methyltransferases like MLL1, MLL2, MLL3, SETD1A, and SETD1B were expressed at low levels in resting CD3+T cells, whereas their expressions increased following T cell activation. ChIP analysis showed increased H3K4me3 deposition at the PDCD1 promoter region in CD4+ and CD8+T cells to increase PD-1 expression.147 The combination of inhibitors targeting H3K4me3 formation and anti-PD-1 therapy may synergistically retard T cell exhaustion (figure 4A). Meanwhile, a study found that H3K4me3 was enriched in the CD274 promoter in pancreatic cancer cells. Loss of MLL1 would lead to decreased PD-L1 expression by reducing H3K4me3 formation, when combined with anti-PD-L1 antibodies could achieve greater tumor repression.148 Osteopontin (OPN) is a cancer-promoting factor. The WDR5-H3K4me3 axis can upregulate the product of OPN in pancreatic cancer cells and MDSCs. Inhibition of WDR5 significantly represses mouse pancreatic cancer growth and enhances the efficacy of anti-PD-1 immunotherapy.149 Interestingly, the expression of WDR5 is associated with worse histologic grade and higher PD-L1 expression level of muscle-invasive bladder cancer (BCa). The WDR5 catalytic activity inhibitor, OICR-9429, can restrict BCa cell proliferation and metastasis, and simultaneously repress PD-L1-derived immune evasion by inhibiting H3K4 trimethylation.150 This WDR5-H3K4me3-PD-L1 axis also displays regulatory functions of cancer immunity in prostate cancer.151 Bally et al disclosed that LSD1 could be recruited to the Pdcd1 gene locus by Blimp-1 during acute lymphocytic choriomeningitis virus infection and promote H3K4me3 removal to reduce Pd-1 expression in CD8+T cells in a murine melanoma model.152 LSD1 knockdown leads to upregulated H3K4me2 levels within CD47 and CD274 promoter regions and increased corresponding protein expression. The synergistic effect of LSD1 inhibition for ICB therapy is also displayed in cervical cancer.153
Epigenetic therapy combines with immunotherapy. (A) The promising H3K4 methylation-related targets, MLL1, WDR5, LSD1, and KDM5A, enhance the efficacy of ICBs through transcriptional regulation of PD-1 and PD-L1. (B) The using model of epi-drugs in CAR-T therapy. Pretreating CAR-T cells and patients with epi-drugs improves the efficacy of ACT therapy. Epi-drugs can promote CAR-T cell proliferation, upregulate chemokine receptor expression, and inhibit CAR-T cell apoptosis, exhaustion, and differentiation. Meanwhile, patients’ premedication with epi-drugs can enhance the sensitivity of tumors to CAR-T therapy. ACT, adoptive cellular therapy.
Taken together, we conclude that H3K4me3 modifiers are feasible targets for combination immunotherapy. Future exploration should focus on therapeutic strategies for H3K4me3-related targets in combination with other immune checkpoint molecules.
Combination therapy with adoptive cellular therapy (ACT)
CAR-T cells’ differentiation, activation, infiltration, and persistence are similar to the regulatory trajectory of natural T cells in vivo. Since epigenetic therapies can reprogram the transcriptional profiles of cells, treating CAR-T cells and tumor tissues with epi-drugs prior to patient infusion for CAR-T cells is a promising approach to sensitizing cancer cells to CAR-T cell-mediated clearance (figure 4B).154 The detailed mechanisms include reversing the heterogenetic expression of cancer antigens, remodeling the immunosuppressive tumor microenvironment, and optimizing the functional states of CAR-T cells.155–157 Regarding H3K4me3 modification-related targets, LSD1 knockdown exhibits superior antitumor capacity when combined with anti-CD19 CAR-T cells.158 However, it awaits further investigation on how to balance the stimulative and suppressive effects of H3K4me3 on cancer immunity. Whether the functional differentiation of Treg cells,112 Th2 cells,107 and MDSCs129 mediated by H3K4me3 modification will be a significant obstacle to the therapeutic efficacy of H3K4me3-targeting coupled with ACT is unknown. Therefore, more clinical trials are needed to observe the net regulatory effects of epi-drugs on CAR T cells, host TME reshaping, and patients’ long-term response.
Conclusion and future directions
H3K4 trimethylation is a critical epigenetic mechanism in regulating the cancer-immunity cycle. KMTs and KDMs regulate the transcriptional profiles of cancer cells and immune cells, thereby reshaping cancer cells' immunogenicity and modulating the functional evolution of immune cells.
Advances in high-throughput screen technology developments facilitate the detection of individual tumors’ epigenetic vulnerabilities, especially in vivo CRISPR-based screens. The initiation and progression of some tumors can be highly sensitive to pharmacological or genetic modification of specific epigenetic mechanisms. On this basis, researchers can explore the underlying mechanisms that determine the epigenetic phenotype of each patient with cancer and develop novel therapeutic targets. An extension of this strategy is that researchers search for the core epigenetic mechanisms of antitumor immunity regulation to optimize the immunotherapy-based combination regimens. For example, Li et al showed that KDM3A was a potential epigenetic regulator of immunotherapy response in pancreatic ductal adenocarcinoma through a CRISPR epigenetic library screen. KDM3A ablation sensitized the tumor to immunotherapy.159 In the same way, another research group found that ASF1A deletion can synergize with anti-PD-1 immunotherapy to inhibit KRAS-mutation lung adenocarcinoma.160
The CRISPR-based screen also plays a vital role in searching the novel targeted lesion in the genetic context of tumor suppressor inactivation mutations. This therapeutic strategy is called synthetic lethality.161 The frequent loss-of-function mutation of epigenetic genes renders it more tricky to develop target agents than gain-of-function mutations, so finding the second gene products whose pharmacological inhibition coupled with specific epigenetic mutations is lethal for cancer cells by using high-throughput screen technologies is a feasible strategy. For example, the EZH2 inhibitors are used for SMARCB1-deficient malignant rhabdoid tumors and SMARCA4-deficient ovarian small cell carcinomas based on the principle of synthetic lethality.162 We think the integrated utilization of high-throughput technologies is of great significance for finding novel epigenetic therapy regimens.
With the recent discoveries in mind, we propose that some notions about epigenetic regulation of cancer immunity should be renewed. For example, fixed epigenetic modification is widely recognized as the cause of irreversible exhausted CD8+T cells on ICB therapy. However, a study displayed that making T cells rest also elicited wholescale epigenetic reprogramming.163 Therefore, the role of epigenetic regulation in the progression and reversibility of T-cell exhaustion needs more exploration. As far as the topic of this review,164 the deposition of H3K4me3 on gene loci is likely to be the result of activated transcription, and the notion challenges the generally accepted view that H3K4me3 is the initiative factor for activated transcription. H3K4me3 formation is regulated by many methylated catalyzes. Intrinsic and external factors can stimulate the expression of these modifiers, like their epistatic transcriptional factors and other unknown factors. In this case, the fluctuation of H3K4me3 distribution is the consequence. The defective upstream mechanisms of histone modification should be further elucidated to establish an integrated epigenetic network for the regulation of cellular phenotypes.
DNMTis and HDACis have shown favorable therapy efficiency combined with immunotherapies in some cancers.145 Conversely, KMT or KDM inhibitors display more complicated effects due to the diversity of histone methylation. Therefore, how to balance the multifaceted regulation of histone methylation-targeted drugs to boost the tumor-suppressive on-target effects without tumor-promoting off-target effects needs to be further studied, which is the prerequisite to obtaining stable and predictable therapeutic efficacy in clinical treatment.
Taken together, we conclude that H3K4 trimethylation is a critical epigenetic mechanism for regulating anticancer immunity processes. Novel developed H3K4me3-related epi-drugs and the combination strategy with immunotherapies are promising therapeutic regimens for cancers in the future.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
References
Footnotes
Contributors CXL and JH conceptualized the study. CX wrote the original draft. TF, YZ, HT, and ZD revised the manuscript. Visualization was performed by CX and JL. All authors read and approved the final manuscript.
Funding This work was supported by the National Key R&D Program of China (2021YFF1201300, 2020AAA0109505, YS2021YFF120009), the National Natural Science Foundation of China (81972196), and the CAMS Innovation Fund for Medical Sciences (CIFMS) (2021-1-I2M-012).
Competing interests None declared.
Provenance and peer review Commissioned; externally peer reviewed.