Histone modifiers at the crossroads of oncolytic and oncogenic viruses

Abstract

Cancer is a disease caused by loss of regulatory processes that control the cell cycle, resulting in increased proliferation. The loss of control can deregulate both tumor suppressors and oncogenes. Apart from cell intrinsic gene mutations and environmental factors, infection by cancer-causing viruses also induces changes that lead to malignant transformation. This can be caused by both expression of oncogenic viral proteins and also by changes in cellular genes and proteins that affect the epigenome. Thus, these epigenetic modifiers are good therapeutic targets, and several epigenetic inhibitors are approved for the treatment of different cancers. In addition to small molecule drugs, biological therapies, such as antibodies and viral therapies, are also increasingly being used to treat cancer. An HSV-1-derived oncolytic virus is currently approved by the US FDA and the European Medicines Agency. Similarly, an adenovirus-based therapeutic is approved for use in China for some cancer types. Because viruses can affect cellular epigenetics, the interaction of epigenome-targeting drugs with oncogenic and oncolytic viruses is a highly significant area of investigation. Here, we will review the current knowledge about the impact of using epigenetic drugs in tumors positive for oncogenic viruses or as therapeutic combinations with oncolytic viruses.

Graphical abstract

This review examines inhibition of histone modifications in the context of cancer therapy and how this may affect oncogenic and oncolytic viruses. Because viruses can alter cellular genetics and epigenetics, the interaction of epigenetic inhibitors with oncogenic and oncolytic viruses is a highly significant area of research.

Introduction

Cancer is a complex product of mutational changes in the genetic code, carcinogenic effects of gene products, environmental factors, oncogenic viruses, and dysregulation of epigenetic modifiers, which all can result in malignant transformation. Of these, the latter two can dramatically affect gene expression by overtaking cellular signaling and growth checkpoints even in the absence of genetic mutations. The epigenetic modifiers that affect neoplastic growth also have a tremendous impact on viral growth and replication. Here, we will summarize the impact of various epigenetic modifiers on the etiology of cancer-causing as well as cancer-destroying viruses.

Epigenetic modifications of the genome include DNA methylation and histone modifications, which are critical components in the regulation of gene expression.¹ The dynamic process of epigenetic modification is carried out by epigenetic modulators, modifiers, and mediators.² The epigenetic modulators function in upstream signaling pathways from the modifiers and mediators and provide an initial signal for chromatin modification.² Some known modulators include IDH1/2, KRAS, and TP53, and are frequently mutated in cancers and contribute to the dysregulated epigenetic state of neoplasia.² For example, accumulation of 2-hydroxyglutarate due to mutant isocitrate dehydrogenase (IDH) results in hypermethylation of DNA and histones, contributing to the disruption of chromatin structure and aberrant gene expression.^3,4 Epigenetic mediators include transcription factors, such as SOX2, NANOG, and OCT4, whose activity is involved with pluripotency and cell survival.^2,5, 6, 7 These transcription factors are also embryonic cell markers that are repressed in somatic tissues by hypermethylation, and the deletion of DNA methyltransferase 3 results in resistance to differentiation.⁸ As such, they are frequently associated with oncogenesis and maintenance of cancer cell stemness.^2,5, 6, 7

The epigenetic modifiers directly alter the epigenome by changing the chemical markers on histone proteins or DNA.² In this review, we have chosen to focus on histone modifications specifically. Chromatin modification is tightly regulated by maintaining a balance between modifiers that have been characterized into “writers” and “erasers” based on their ability to add or remove modifications on DNA and histones (Table 1). The expression of epigenetic modifiers is frequently dysregulated in cancer and can lead to unchecked cell proliferation and survival.⁹ Drugs that inhibit these epigenetic modifiers are used as anti-cancer therapeutics and have demonstrated synergistic effects when combined with other Food and Drug Administration (FDA)-approved cancer treatments, such as cisplatin, bortezomib, gemcitabine, decitabine, rituximab, and doxorubicin.10, 11, 12, 13, 14 The current status of combining epigenetic inhibitors with different viruses and the predicted prognostic values of using these in combination with viruses for cancer treatment are examined in the subsequent sections.

Table 1.

List of histone-modifying enzymes discussed in this review

Enzyme class	Function
Histone acetyltransferase (HAT)	Transfers acetyl groups to lysine residues in histones. The result is chromatin expansion and increased access to DNA.1
Histone deacetylase (HDAC)	Catalyzes removal of acetyl groups from histones, leading to condensed chromatin and transcriptional repression1
Histone methyltransferase (KMT/PRMT)	Facilitates transfer of methyl groups to arginine and lysine residues on histones15
Histone demethylase	Removes methyl groups from histones and other proteins (i.e., p53)15

Chromatin-modifying enzymes and their canonical functions.

Several oncogenic RNA and DNA viruses encode for oncogenes and/or have the power to disable tumor suppressors, resulting in cellular transformation. Conversely, oncolytic virotherapy is an emerging cancer therapy that relies on virus-mediated destruction of neoplastic cells and the subsequent anti-tumor immunity. At the time of writing this review, Imylgic, an oncolytic herpesvirus, is the only oncolytic agent approved in the USA and European Union (EU) for use against melanoma. Similarly, the oncolytic adenovirus, Oncorine, is marketed in China for nasopharyngeal carcinoma, and the Japan Ministry of Health, Labor, and Welfare has granted conditional and time-limited approval to Delytact, an oncolytic herpesvirus for the treatment of patients with malignant glioma. Although these viruses are being investigated in patients for safety and efficacy, the interaction of these agents with other cancer drugs is a highly significant area of research.^16,17 Because epigenetic modifiers can affect gene expression of both cellular and viral genes in infected cells, it is critical to understand the direct interaction between these agents and their combined effect on tumors that are positive or negative for oncogenic viruses.

Histone deacetylases and cancer

Histone deacetylases (HDACs) are a class of histone-modifying enzymes that facilitate the removal of acetyl groups from lysine.¹⁰ HDACs canonically function as repressors of gene expression, because the removal of the acetyl group allows DNA to bind to histones, resulting in a condensed chromatin conformation.^10,18 HDACs have been characterized into four classes: class I, located in cellular nuclei (HDACs 1, 2, 3, and 8); class II, which can translocate between the nucleus and cytoplasm (HDACs 4, 5, 6, 7, 9, and 10); class III, localized to the nucleus, cytoplasm, or mitochondria (sirtuin family: SIRT1–7); and class IV, located in the cytoplasm and nucleus (HDAC 11).18, 19, 20 Classes I, II, and IV are considered “classic” HDACs because they are Zn²⁺-dependent whereas class III are NAD⁺-dependent.^10,18 Dysregulated expression of HDACs in cancer is often associated with poor prognosis, and the aberrant expression can have a wide range of effects, including repression of tumor suppressor genes and pro-oncogenic cell signaling.²¹ In silico analysis of each of the HDACs 1 through 11 from 55 primary study cohorts of lung, liver, ovarian, gastric, and breast cancer patients revealed that 34 of the 55 studies reported a hazard ratio (HR) exceeding 1 (shortened overall survival) (Table 2). Therefore, HDAC inhibition is a very attractive drug target for cancer therapy. Inhibition of HDACs results in the overaccumulation of acetylation marks, leading to DNA damage, apoptosis, and cell-cycle arrest.22, 23, 24, 25 HDAC inhibitors (HDACis) thus represent the most successful class of epigenetic inhibitors so far, with several being FDA approved for hematological malignancies as early as 2006.^22,24,26, 27, 28 HDACs and HDACis have also been shown to have non-canonical interactions with other proteins.²⁹ HDACis can affect inflammatory genes through both the actions on histones and their ability to prevent the deacetylation of immune signaling proteins, such as mitogen-activated protein kinase (MAPK) phosphatase 1, which results in anti-inflammatory signaling.³⁰ A major class of HDACis are the hydroxamic acids, because most function as pan-HDACis or target HDAC classes I and II. In 1990, trichostatin A (TSA) was the first naturally occurring compound discovered to possess HDACi activity.¹⁹ Although TSA is limited to laboratory experiments due to toxicity, analogs of this compound, such as vorinostat, have been successful in treating cancer patients.²⁴

Table 2.

Predicted hazard ratio of different HDACs in each of the five cancer subtypes studied in the meta data analysis

HDAC	HR <1	HR >1
1	Gastric	Liver, ovarian, breast, lung
2	Gastric, liver	Breast, ovarian, lung
3	Breast, ovarian, lung	Gastric, liver
4	Breast, lung	Liver, ovarian, gastric
5	Breast, lung	Gastric, liver, ovarian
6	Breast, ovarian, liver	Gastric, lung
7	Breast	Gastric, liver, ovarian, lung
8	Breast, liver	Lung, gastric, ovarian
9	Breast, lung, liver	Gastric, ovarian
10	Breast	Gastric, lung, ovarian, liver
11	Breast	Gastric, lung, ovarian liver

Predicted HR of different HDACs in cancer. Median expression level was used as the threshold for separating high from low HDAC expression.

Viruses and histone deacetylases

The HDAC/co-repressor of REST (CoREST)/RE-1 silencing transcription factor (REST)/lysine-specific demethylase 1 (LSD1) transcription repressor complex is considered an important anti-viral mechanism utilized by cells to silence viral genes.^31,32 Thus, viruses have also evolved mechanisms to counter this cellular mechanism. For example, herpes simplex virus 1 (HSV-1) encodes immediate–early (α) infected cell protein 0 (ICP0) which functions to silence this repressor complex in non-neuronal cells by displacing the HDAC component.^31,32 The silencing of this repressor complex allows for virus circumvention of host defenses, resulting in virus replication and spread.^31,32 Interestingly, HSV-1 has also been shown to hijack the HDAC/CoREST/REST/LSD1 repressor complex to silence itself and to establish latency in neuronal cells.³¹ Therefore, the disruption of this complex through HDACis like TSA or sodium butyrate can successfully reactivate latent HSV-1 in vitro and in vivo.^31,33, 34, 35 Similarly, activation of Kaposi's sarcoma-associated herpesvirus (KSHV) by inhibitors of class III HDACs has been noted and led to the identification of SIRT1 as a regulator of the KSHV life cycle.³⁶ These results suggest that cancer patients treated with hydroxamic acid HDACis could experience reactivation of latent herpesvirus infections.

Interestingly, several other oncogenic viruses encode for viral proteins that aid in hijacking cellular HDAC machinery to augment viral latency as well. For example, increased HDAC activity has been noted in cancer tissues of patients with hepatitis B virus (HBV)-associated hepatic cellular carcinomas.³⁷ Both HBV and hepatitis C virus (HCV), which are associated with hepatic cancers, have been noted to increase HDAC activity; this is largely considered to be a turning point in tumor cell proliferation, angiogenesis, and invasion.^38,39 HDACs are also important for regulating transcription of HPV E6 and E7 oncoproteins, which in turn can dissociate HDACs to promote hypoxia-inducible factor 1 (HIF-1) expression. Thus, HDACis, such as vorinostat, can block productive HPV genome replication and inhibit growth of cancers that depend on HPV for survival.^40,41 The combination of HDAC inhibition with proteasome inhibition has also been shown to induce potent anti-tumor immunity against HPV-E6/E7-expressing tumor cells.⁴² Similarly, HDACs appear to play a key role in maintaining latency of oncogenic viruses like KSHV and Epstein-Barr virus (EBV), a common cause of Hodgkin's lymphoma. In a recent trial, vorinostat in combination with everolimus or sirolimus was shown to synergistically inhibit relapsed refractory Hodgkin's lymphoma in patients.⁴³

Oncolytic viruses and histone deacetylase inhibition

HDACis have been utilized to increase viral replication and augment oncolytic HSV-1 virotherapy. For example, TSA treatment of oral squamous cell carcinoma (SCC) resulted in the activation of nuclear factor κB (NF)-κB, which augmented the replication of a gamma(1)34.5-deficient oncolytic HSV-1 (oHSV).⁴⁴ This allowed for enhanced killing of the oral SCC cells. A similar effect was observed for glioblastoma and colorectal cancer cells treated with oHSV.⁴⁵ TSA has, thus far, been found to be the most effective in enhancing oHSV replication compared with other HDACis.⁴⁶ Interestingly, HSV-1 infection of T cells induced the expression of the selectin ligand, sialyl Lewis X (sLeX), known to be important for leukocyte interaction with endothelial cells for tissue infiltration,⁴⁷ and treatment with TSA augments the transcription machinery needed to induce sLeX expression.²⁹ This suggests that TSA might enhance the ability of infected leukocytes to infiltrate tissue. While it has not yet been fully explored, these studies collectively imply that hydroxamic acid HDACi treatment might yield an improved immune response in the wake of oncolysis that could result in increased therapeutic benefit.

Furthermore, TSA has been shown to increase the expression of the adenovirus entry receptor (coxsackievirus and adenovirus receptor) on small cell lung cancer (SCLC) cells, required for entry of oncolytic adenoviruses.⁴⁸ Consistent with this, TSA augmented the anti-tumor benefit of the oncolytic adenovirus H101 in esophageal SCC models in vitro and in vivo.⁴⁹ Additionally, TSA treatment enhanced oncolytic adenovirus-mediated cell killing in cisplatin-resistant ovarian cancer cells.⁵⁰ Similarly, TSA treatment has been found to significantly increase the spread and replication of an oncolytic vaccinia virus (oVV) in several infection-resistant cancer cell lines.⁵¹ Treatment with both TSA and oVV in mice bearing metastatic lung cancer prolonged survival and decreased the number of metastases in a syngeneic model.⁵¹ Importantly, the tumor specificity of oVV replication in normal tissue and in a human xenograft model were not affected.⁵¹ Based on the interactions with both wild-type (WT) viruses and oncolytic viruses, hydroxamic acid HDACis, such as TSA, play a critically important role in repression of host anti-viral mechanisms and have great implications for existing viral-based therapies.

HDAC class III member SIRT1 is a known vesicular stomatitis virus (VSV) restriction factor, and its silencing can augment VSV-mediated oncolysis. Incidentally, treatment of prostate cancer cells with vorinostat downregulated SIRT1 and augmented VSVΔM51 replication and oncolysis.⁵² Treatment of tumor-bearing mice with valproic acid (VPA), an FDA-approved HDACi, resulted in a significant reduction in the recruitment and activity of NK cells in vivo. This suppression of innate immune activity allowed for increased oncolytic virus effects and enhanced therapeutic benefit.⁵³ Interestingly, HDAC inhibition induced the expression of junctional adhesion molecule 1 (JAM-1), the entry receptor for the oncolytic reovirus, Reolysin, thus permitting more efficient viral entry and spread. This allowed for increased anti-tumor efficacy in animals bearing myeloma treated with both Reolysin and HDACis.⁵⁴

Furthermore, HDAC inhibition has been shown to affect immunotherapy like immune checkpoint blockade (ICB). In one study, pretreatment of B16 melanoma cells with the HDACi VPA or AR42 resulted in enhanced efficacy of anti-programmed cell death protein 1 (PD-1) and anti-cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4).⁵⁵ The HDACi combination with anti-PD-1 in particular was associated with proinflammatory cytokines (CCL2, CCL5, CXCL2, and CXCL9), enhanced leukocyte infiltration to the tumors, and increased activation of immune cells, such as T cells and macrophages.⁵⁵ Additionally, anti-PD-1 has been employed as a monotherapy to treat oncogenic-virus-driven tumors, such as human papillomavirus (HPV)-positive oropharyngeal cancers, with limited success.⁵⁶ This study shows that combining a stimulator of interferon genes (STING) agonist with ICB in a metastatic mouse model resulted in tumor regression and enhanced survival in 71% of the treated animals.⁵⁶ This could be an interesting avenue of investigation if an oncolytic virus was used in place of a STING agonist. In the context of treating oncogenic-virus-driven tumors with oncolytic viruses and HDACis, the evidence above would suggest that this could be a potent combination to stimulate the body's immune system. The evidence also lends itself to a possible combination of HDACi, oncolytic virus, and ICB to eliminate tumors. Altogether, it appears that HDAC inhibition is a promising treatment for oncogenic-virus-mediated malignancies and can be exploited to augment oncolytic virotherapy for tumors and enhance anti-tumor immunity (Figure 1).

Figure 1.

HDAC inhibition effect on oncogenic and oncolytic viruses

Interaction of HDACi with oncogenic (broken lines) and oncolytic viruses (solid lines) suggests that combining oncolytic viruses for tumors positive or negative for oncogenic viruses to treat cancers should have increased therapeutic advantage.

Histone acetyltransferases and viral gene expression

Acetyltransferases utilize the acetyl group on coenzyme A (CoA) and transfer it to the amine groups on proteins. N-terminal acetylation transfers the acetyl group from CoA to the N-terminal amine group, while lysine acetyltransferases transfer acetyl groups to ε-amino of lysine. Histones are predominantly acetylated on their lysine residue. Acetylation of histones affects their surface charge and affects changes in nucleosome stability and chromatin assembly, thus affecting DNA transcription.⁵⁷ Histone acetyltransferases (HATs) are divided into two types: type B HATs, which exist primarily in the cytosol and target newly synthesized histones, or type A HATs, which are nuclear proteins that can dictate changes to nucleosome assembly. HAT functionality is frequently dysregulated in malignant cells and used to drive cell proliferation and stemness. Thus, use of HAT inhibitors (HATis) has been investigated as an anti-cancer therapy.⁵⁷

The impact of histone acetylation on herpetic viral reactivation from latency has been studied by using inhibitors for both acetylation and deacetylation. HDACis, such as TSA, MS271 (HDAC1 inhibitor), and MC1568 (HDAC4 inhibitor), induced the reactivation of latent HSV-1 via activation of acetyltransferases p300 and pCAF.⁵⁸ Furthermore, addition of a p300/CBP HATi blocked reactivation from latency in the same organ culture model, showing the importance of histone acetylation in HSV-1 reactivation.⁵⁸ HSV-1 tegument protein VP22 has been shown to bind to host template-activating factor I (TAF-I), a chromatin remodeling protein that is a part of an acetyltransferase inhibitor complex.⁵⁹ Enforced expression of TAF-I in cells interferes with HSV-1 replication, implying that histone acetylation supports HSV-1 replication.⁵⁹ EBV reactivation is an important mechanism for EBV-driven malignancies, such as the aforementioned Hodgkin's lymphoma. EBV reactivation depends on the expression of BamH1Z, an immediate–early viral gene that is regulated by epigenetic mechanisms. Among these mechanisms, histone acetylation appears to correlate with reactivation of the virus, and inhibition of histone acetylation has been shown to inhibit EBV-induced lymphocyte transformation.⁶⁰ While histone acetylation can promote HSV-1 replication, acetylation of p53 promotes anti-viral activities.⁶¹ Adenovirus-encoded E1A can attenuate p53 acetylation by directly binding and blocking the HAT domains of p300/CBP in infected cells.⁶² In addition, HAT inhibition is implicated in enhancement of the anti-tumor immune response by inhibiting intrinsic and interferon (IFN)-γ-induced programmed death-ligand 1 (PD-L1) expression in a prostate cancer model.⁶³ The effects of HAT inhibition in conjunction with oncolytic viruses have not been deeply investigated; however, the literature appears to suggest that HAT inhibition would be expected to reduce oHSV-1 replication at the very least. While the effect of HATi would be expected to reduce p53 acetylation-mediated anti-viral effects, it would also be predicted to reduce adenovirus entry, and the effect of HATi on oncolytic adenovirus has not been investigated, to our knowledge.

Histone methyltransferases and viral gene expression

Histone methyltransferases consist of lysine methyltransferases (KMTs) and the protein arginine methyltransferases (PRMTs). These families of histone writers function to transfer up to three methyl groups to the lysine or arginine residues, respectively.⁶⁴ These methyl groups affect the ability of histone readers to bind to and recruit effector proteins. Therefore, the effects of histone methylation frequently appear context-dependent due to recruitment of different effector proteins.⁶⁴

Herpesviruses and lysine methyltransferases

EZH2/1 are the only known H3K27 methyltransferases and catalytic components of the polycomb recessive complex 2 (PRC2).^65,66 Repressive histone marks induced by PRC2 are maintained by PRC1, which functions to inhibit the RNA polymerase II preinitiation complex.⁶⁷ Upon infection with WT HSV-1, the viral genome exhibits repressive histone methylation signatures.⁶⁶ It has been previously shown that WT HSV-1 interacts with host histone KMTs EZH2 and SUV39H to control gene expression during latent infection and reactivation of lytic infection.⁶⁵ Bmi1, a subcomponent of PRC1, was found to be enriched on HSV-1 genes,⁶⁵ and histone markers H3K9me3 and H3K27me3 were associated with latency.^66,68 Knowing that KMTs are involved in HSV-1 latency, Arbuckle et al. (2017) investigated the effect of inhibition of EZH2 and EZH1 on WT HSV infection. Surprisingly, these KMT inhibitors (KMTis) were found to suppress HSV-1 replication. In vivo, this inhibition resulted in the induction of inflammatory, stress, and cellular defense responses that enhanced recruitment of immune cells. Thus, EZH2/1 inhibitors induced a cellular anti-viral state and, consistent with this, could suppress Cytomegalovirus (CMV), Adenovirus, and Zika virus infections.⁶⁶ Interestingly, EZH2 expression is activated by HPV16 E7 and is also reported to be higher in EBV-positive nasopharyngeal carcinoma.⁶⁹ Thus, EZH2 inhibition preferentially inhibited the growth of HPV-positive cancers in vitro.^70,71 In addition, EZH2 inhibition has been shown to rescue resistance to ICB in a melanoma model.⁷² While KMT inhibition is predicted to slow down tumors driven by oncogenic viruses and augment anti-tumor immune responses, the inhibition of viral replication would predict a likely antagonistic relationship with oncolytic viral therapy (Figure 2).

Figure 2.

Predicted opposing effects of histone methyltransferase inhibition on oncogenic and oncolytic viruses

(A) KMTi treatment of HPV⁺ tumors can have a therapeutic advantage; the cellular antiviral state would predict an antagonistic effect on oncolytic virus replication. The broken line indicates author inference based on effects on WT virus clearance. (B) Predicted effects PRMT inhibition on oncogenic virus-induced cancer transformation. The broken line indicates author inference based on known effects of PRMTi on HBV replication; actual inhibition not tested to our knowledge.

Viruses and protein arginine methyltransferases

PRMTs are enzymes that catalyze the methylation of arginine on protein residues, and their deregulation has been associated with cancers.⁷³ PRMTs have been implicated in maintenance of viral infections, including herpesviruses, HIV, HBV, and human T cell leukemia virus type 1 (HTLV-1). These are all significant human pathogens that affect millions of people worldwide and result in increased risk of certain cancers as comorbidities.74, 75, 76

Specifically, PRMT5 seems to play a crucial role in modulating infections of oncogenic viruses, such as KSHV and EBV. KSHV-encoded ORF59 has been shown to modulate histone arginine methylation via displacement of PRMT5. PRMT5 symmetrically dimethylates histone H4 at arginine 3, and these markers are indicative of condensed chromatin.⁷³ This is important for the virus because KSHV relies on compacted chromatin for latent infection but utilizes ORF59 for reactivation of the lytic cycle, thus necessitating the opening of chromatin. In infected cells, the N-terminal domain of ORF59 binds to the middle domain of PRMT5 which contains the methyltransferase activity to modulate its effects. This competitive binding of ORF59 to PRMT5 instead of its linker, cooperator of PRMT5 (COPR5), inhibits symmetric dimethylation of H4R3 and allows for loosened chromatin and lytic replication.⁷⁷ Much like KSHV, EBV also exploits histone modifications to either maintain or escape from latency. PRMT1 and PRMT5 have been shown to bind to a region of arginine-glycine repeats in the EBV nuclear antigen 1 and 2 (EBNA1/2) which are proteins necessary for mitotic segregation of EBV genomes in latent infection (EBNA1) and activate transcription of latency-specific viral genes (EBNA1 and EBNA2) that are necessary to maintain transformed cells.^78,79 The consequence of the interaction between PRMT5 and EBNA2, specifically, was increased EBNA2 occupancy at promoters of target genes and enhanced transcription, suggesting a critical role of PRMT5 in EBV latent infections.⁷⁸ It has also been demonstrated that EBV infection of germinal center B cells, the precursor cells to Hodgkin's lymphoma, induced upregulation of PRMT5 and PRMT1.⁸⁰ Consistent with this, PRMT5 inhibition was capable of blocking EBV-dependent B cell transformation and infected cell survival while sparing uninfected B cells.⁸¹ This novel inhibitor prevented the silencing of various tumor suppressor genes, thereby preventing immortalization of the EBV-infected cells.⁸¹

Similarly to the aforementioned gamma herpesviruses, host-cell PRMT5 was found to be upregulated in HTLV-1-infected T cells undergoing transformation as well as some leukemia/lymphoma cell lines and adult T cell leukemia/lymphoma (ATLL) patient PBMCs.⁷⁶ HTLV-1 is a retrovirus responsible for ATLL, which manifests after a long period of latency (20–30 years).⁷⁶ Panfil et al. (2016) hypothesized that like EBV, HTLV-1 upregulated PRMT5 to mediate cellular transformation. They found that knockdown of PMRT5 using short hairpin RNA (shRNA) or the novel PRMT5i, identified by Alinari et al. (2015), resulted in reduced cell proliferation and selective toxicity against HTLV-1-positive cells, suggesting that HTLV-1 strongly relies on PRMT5 to mediate transformative effects (Figure 2).⁷⁶ From an immunological standpoint, PRMT5 inhibition appears to augment response to ICB by preventing PD-L1 expression on tumor cells.⁸² Likewise, in a separate study, PRMT5 inhibition increased interferon, chemokine, and major histocompatibility complex (MHC) class I expression in a melanoma model.⁸³ The combined prevention of oncogenic transformation with enhanced anti-tumor immune response due to PRMT5 inhibition certainly warrants future investigation and the addition of an oncolytic virus could have impactful results.

HBV has also been shown to have important interactions with PRMTs, primarily PRMT1 and PRMT5.^84,85 Evidence suggests that both PRMT1/5 function as negative regulators of the HBV life cycle.⁸⁴ PRMT5 has been shown to have both a methyltransferase-dependent function in HBV infection by epigenetically silencing the minichromosome produced by HBV infection as well as a methyltransferase-independent function by interfering with pre-genomic RNA packaging, which further inhibited HBV replication.^84,85 Given that PRMT5 has splice variants (v1–v6), Lubyova et al. (2017) hypothesized that these variants give rise to the multifunctional role of PRMT5 and can affect HBV replication at all stages.⁸⁴ With the recent information on HBV and PRMTs, and knowing that chronic HBV infection is a great risk factor for development of hepatocellular carcinoma (HCC), it is likely that a PRMT inhibitor (PRMTi) would be an antagonistic therapy for HBV-mediated HCC.⁸⁶

At the time of writing this review, we found no studies examining histone methyltransferase inhibitors of either KMTs or PRMTs in conjunction with oncolytic viruses. With numerous oncolytic virotherapy trials currently being investigated for treatment of all kinds of malignancies, it would be significant to know the interactions of these agents with histone methyltransferase inhibitors.

Histone demethylases and viral gene expression

Histone demethylases, as the name implies, function to remove methyl groups from histones, particularly at the lysine residues.⁸⁷ Histone LSD1A, which is also known as KMD1A, was among the first demethylases to be identified.⁸⁸ The subsequent discoveries of the family of Jumonji C (JMJC) enzymes that demethylate histones has helped uncover the tightly regulated and dynamic processes used to maintain the delicate balance between methylation and demethylation.^87,89 Thus histone demethylase inhibitors have become an attractive therapeutic target for some cancers. Reversible and irreversible inhibitors are in development,⁸⁹ with a few, such as tranylcypromine (TCP), in clinical trials for acute myeloid leukemia and myelodysplastic syndrome [NCT02717884, NCT02273102].

Inhibition of histone demethylase activity inhibits replication of herpesviruses, which are DNA viruses. For example, treatment of mice with TCP, an inhibitor of LSD1/KDM1A, reduces HSV-1 reactivation and decreased the severity of virus-induced encephalitis.⁹⁰ Similarly, another histone demethylase inhibitor, ML324, also showed strong anti-viral effects against HSV-1- and CMV-infected mice.⁹¹ Likewise, inhibition of LSD1 using TCP was also shown to inhibit viral reactivation/lytic infections by both HSV-1 and varicella-zoster virus (VZV).^92,93 This is thought to occur by promoting epigenetic suppression of viral genomes.⁹⁴ In fact, the combination of ganciclovir, a common anti-herpetic agent, and a histone demethylase inhibitor, OG-L002, is known to decrease viral load of equine herpesvirus (EHV) more than either treatment alone, implying that this strategy might have implications for ganciclovir-resistant herpesvirus infection management in humans.⁹⁵ Pharmacological inhibition of histone demethylases JMJD3 and UTX also functioned to block HSV-1 reactivation from latency, indicating that removal of silencing methylation marks is critical for viral reactivation.^96,97 Similarly, KSHV infection produces polyadenylated nuclear (PAN) RNA, which interacts with UTX and JMJD3 and corresponds to decreased repressive methylation marks at a viral promoter.^96,97 Furthermore, the latency-associated nuclear antigen (LANA) expressed in KSHV-infected cells was shown to form a complex with the histone demethylase KDM3A/JMJD1A for recruitment to the KSHV episome.⁹⁸ Overexpression of a KDM3A mutant that cannot bind to LANA resulted in significantly reduced recruitment of KDM3A to KSHV genes.⁹⁸ These results suggest that the LANA complex with KDM3A plays an important role in KSHV epigenetic regulation. The control of repressive methylation to either mask the viral genome or reactivate a lytic infection is clearly a critical component of KSHV infection as well as some of the other herpesviruses discussed in this review, such as HSV-1.

Contrary to the herpesviruses, inhibiting LSD1 histone demethylase using shRNA or TCP worsened disease in mice infected with influenza A virus, suggesting that this enzyme is needed for host defense against RNA viruses.⁹⁹ The induction of anti-viral immunity in this case suggests a potential for increasing anti-cancer immunity and might be beneficial as a combination strategy with oncolytic viruses. In fact, in one study, depletion of LSD1 led to double-stranded RNA (dsRNA) cellular stress and type 1 interferon activation, which corresponded to increased T cell activation and infiltration in a melanoma model.¹⁰⁰ This was further supported by evidence from The Cancer Genome Atlas (TCGA), showing an inverse relationship between LSD1 expression and T cell infiltration into different types of tumors.¹⁰⁰ Therefore, the combination of histone demethylase inhibitors and oncolytic viruses that are not derived from herpesviruses warrants further exploration, because there would possibly be strong anti-tumor immunity in response to these agents.

Conclusions

Although viruses interact in multitude of ways with cellular epigenetic machinery, studies investigating the interaction of drugs targeting the epigenetic machinery with viral therapy in the context of tumors positive for oncogenic virus are limited. The most well studied class of drugs that modulate cellular epigenetic machinery in conjunction with virotherapy for cancer treatment is HDACis. This class of agents appears to synergize with different viruses by a diverse range of effects, from modulating cellular anti-viral activity to regulating expression of cell surface viral entry receptors. In general, inhibition of cellular HDACs appears to augment viral replication and spread and thus would be predicted to support viral gene and oncolytic therapy for cancer.

Conversely, HAT and histone methyltransferase inhibition have not been as widely studied in the context of viral infections. It seems as though HAT activity can play a key role in the function of viruses, such as HSV-1 and influenza, and inhibition of HATs blocked replication of those viruses. Perhaps this effect could be explored further as a way to identify useful anti-viral treatments. Based on the evidence, however, it would be predicted that HAT inhibition would be antagonistic to oncolytic virus therapy. Histone methyltransferase inhibition appears to interfere with oncogenic virus replication, which could prove important in a clinical setting. Of note would be the use of a KMTi versus PRMT), because it can exert different effects in different viral infections. Particularly, PRMTs appear to be important for KSHV, EBV, and HTLV-1 replication, so further investigation into PRMTi as an anti-viral could yield interesting data. However, the effect of viral inhibition by histone methyltransferase would predict an antagonistic relationship with oncolytic viral therapy in this case as well.

Interestingly, evidence suggests that histone demethylase inhibitors inhibit DNA virus replication and possibly augment RNA virus replication. Histone demethylase inhibitors may be an option for treatment of oncogenic viruses like KSHV but would likely be an antagonistic combination with oncolytic DNA viruses. This evidence does support exploring the combination of oncolytic RNA viruses and histone demethylase inhibitors as a way to enhance viral replication and the subsequent immune response. Given the broad and varied effects of these epigenetic inhibitors on viral gene expression, more study is needed to fully grasp the mechanisms at play (Table 3).

Table 3.

This table details the different epigenetic inhibitors found in publications throughout this review

Drug name	Class	Virus effects	Combination effect on virus	Ref.
Trichostatin A (TSA)	Hydroxamic acid histone deacetylase inhibitor	Augments oncolytic viral replication, enhances viral-mediated oncolysis, induces immune activation, augments leukocyte infiltration	+	44, 45, 46,48, 49, 50, 51
Vorinostat (SAHA)	Hydroxamic acid histone deacetylase inhibitor	Enhanced oncolytic viral proliferation due to reduced interferon response and reduced recruitment of NK cells	+	52
Valproic acid (VPA)	Short chain fatty acid histone deacetylase inhibitor	Enhances oncolytic viral replication and oncolysis of tumor cells	+	53
Curcumin	Histone acetyltransferase inhibitor	Inhibits WT HSV reactivation from latency	−	58
C646	Histone acetyltransferase inhibitor	Attenuates influenza A replication	−	101
GSK126	Lysine methyltransferase inhibitor (EZH2/1)	Suppresses WT HSV-1 infection in vitro and in vivo, inhibits lytic replication, induces cellular anti-viral state	−	66
GSK343	Lysine methyltransferase inhibitor (EZH2/1)	Suppresses WT HSV-1 infection in vitro and in vivo, inhibits lytic replication, induces cellular anti-viral state	−	66
UNC1999	Lysine methyltransferase inhibitor (EZH2/1)	Suppresses WT HSV-1 infection in vitro and in vivo, inhibits lytic replication, induces cellular anti-viral state	−	66
Astemizole	Lysine methyltransferase inhibitor (EZH2/1)	Suppresses WT HSV-1 infection in vitro and in vivo, inhibits lytic replication, induces cellular anti-viral state	−	66
CMP5	Protein arginine methyltransferase inhibitor	Inhibits EBV-dependent B cell transformation and silencing of tumor suppressor genes; selective toxicity against HTLV-1-infected cells, reduces infected cell proliferation	−	76,81
Tranylcypromine (TCP)	Histone demethylase inhibitor (LSD1/KDM1A)	Reduces WT HSV-1 reactivation from latency and viral encephalitis in mice; inhibits WT HSV and VZV lytic replication; results in worse disease outcome with influenza	−	90,93,99
ML324	Histone demethylase inhibitor (JMJD2)	Inhibits viral IE gene expression of WT HSV and CMV, reduces HSV plaque formation, and inhibits HSV reactivation from latency	−	91
OG-L002	Histone demethylase inhibitor (LSD1)	Inhibits WT HSV IE gene expression and viral replication and and suppresses HSV-1 infection in vivo	−	92

Epigenetic inhibitors and the impact on viruses. The type of inhibitor, the effects of the inhibitor on viruses, and whether these combinations appear to be positive or negative for viruses are included.

Although in this review we have focused on virus interactions with epigenetic histone modifiers, it is important to note that a large function of oncolytic viral therapy is as an immune-activating agent. Therefore, treatment of tumors positive for oncogenic viruses with histone acetylation and methylation-modifying drugs should also keep in perspective the impact on the host's immune system and its ability to mount an anti-tumor immune response. For example, HDACi-mediated activation of NK cells and maturation of dendritic cells can augment virotherapy as well as sensitize tumors to immune checkpoint therapy. Numerous studies have shown an increased anti-tumor immune benefit with a combination of oncolytic viruses. Given the impact of HDACi on oncogenic viruses, the results would suggest a highly synergistic therapy when oncolytic viruses are combined with HDACi in the context of HPV/KSHV/EBV-positive malignancies. On the other hand, while histone KMTis like EZHi can suppress HPV-induced malignant transformation, they can also can induce cellular anti-viral immunity, which would be predicted to clear oncolytic virus treatment. However, the impact of this combination of oncolytic virus-induced anti-tumor immunity and efficacy is insufficiently investigated currently. The effect of histone PRMTs on oncogenic viruses is somewhat context-dependent, as described above. Since a detailed analysis of the impact of epigenetic modifiers on anti-tumor immunity has been discussed in several recent reviews, we have focused on the interaction between epigenetic regulation of oncolytic and oncogenic viruses. Future forays of combining these agents should take into consideration the impact of individual agents on virus (oncolytic and oncogenic) and immune effects. These studies will help elaborate new ways to treat cancers positive for oncogenic viruses and exploit host cellular machinery to inhibit pathogenesis, while enhancing therapeutic benefit of oncolytic viruses.