Host Methyltransferases and Demethylases: Potential New Epigenetic Targets for HIV Cure Strategies and Beyond

Abstract

A successful HIV cure strategy may require reversing HIV latency to purge hidden viral reservoirs or enhancing HIV latency to permanently silence HIV transcription. Epigenetic modifying agents show promise as antilatency therapeutics in vitro and ex vivo, but also affect other steps in the viral life cycle. In this review, we summarize what we know about cellular DNA and protein methyltransferases (PMTs) as well as demethylases involved in HIV infection. We describe the biology and function of DNA methyltransferases, and their controversial role in HIV infection. We further explain the biology of PMTs and their effects on lysine and arginine methylation of histone and nonhistone proteins. We end with a focus on protein demethylases, their unique modes of action and their emerging influence on HIV infection. An outlook on the use of methylation-modifying agents in investigational HIV cure strategies is provided.

Introduction

Epigenetic modifications of nucleic acids, histones, and nonhistone proteins are important regulators of gene transcription and expression and are targets in clinical efforts to reverse HIV latency. One such modification is methylation, which involves the transfer of a methyl group with S-adenosyl-L-methionine (SAM) as a methyl donor.¹ This modification is a reversible process regulated by methyltransferases (MTs) and demethylases. Both methyltransferases and demethylases are highly specific for the position of their substrate and extent of methylation. Methylation of nucleic acids, histones, and nonhistone proteins regulates gene transcription and expression by altering chromatin packaging, promoting chromatin accessibility, creating DNA–protein interactions, and generating interaction interfaces for the assembly of macromolecular complexes.^1,2

Epigenetic regulation of DNA, histone, and nonhistone proteins plays an important role in HIV transcription and is a target in preclinical efforts to reverse HIV latency, a major hurdle to curing HIV infection. HIV latency is established by integrating viral DNA into the host chromatin.³ Once integrated, proviral cDNA is organized into higher-order chromatin and becomes subject to regulation by host chromatin-modifying enzymes, including deacetylases, methyltransferases, and demethylases. Hypoacetylation of histone proteins by histone deacetylases (HDACs) helps to maintain HIV-1 latency by repressing viral transcription.⁴ Critical cofactors that further contribute to latency include the positive transcription elongation factor b (P-TEFb), NF-κB, and the virally encoded transactivator of transcription (Tat).^5–7

Tat binds to an RNA stem-loop structure called TAR at the 5′ end of all nascent viral transcripts and recruits P-TEFb, which together with other elongation factors, form a “super-elongation complex” at the elongating RNA polymerase II.^8–10 Tat itself is subject to reversible methylation by SETD7/SET7/9/KMT7 at lysine 51 in its basic RNA-binding domain¹¹ (Fig. 1). This methylation event positively supports the transcriptional activity of Tat¹¹ and is reversed by the demethylase activity of LSD1/KDM1A.¹² Tat is also methylated by the histone lysine methyltransferase SETDB1 at lysine 51,¹³ by SETD7 at lysine 71,¹⁴ and by arginine methyltransferase PRMT6 at R52 and R53¹⁵ (Fig. 1).

FIG. 1.

Specificity of protein methyltransferases and demethylase LSD1 that target HIV-1-Tat at K51, R52, R53, and K71.

Latent HIV-1 proviruses are primarily found integrated into actively transcribed genes, but characteristically display heterochromatic features.^16,17 The long terminal repeat (LTR) of latent proviruses accumulates HDACs that result in high levels of deacetylated histones.^18,19 Besides histone deacetylation, a growing list of MTs and demethylases indicates that methylation of DNA, histones, and nonhistone proteins is essential for HIV-1 transcription. In this review, we summarize the DNA and protein methyltransferases (PMTs) and demethylases involved in HIV infection, and their implications for HIV cure strategies.

Methyltransferases

DNA methyltransferases

DNA methylation is an epigenetic mark that mediates transcriptional activation or repression of genes.²⁰ DNA is methylated within the cytosine–phosphate–guanine (CpG) and non-CpG dinucleotide sites by the addition of a methyl group to the fifth carbon atom of the cytosine residues using SAM as a methyl donor.²¹ This process is catalyzed by specific DNA methyltransferase (DNMT) enzymes. In mammals, five structurally and functionally distinct DNMTs have been characterized: DNMT1, DNMT2, DNMT3A, DNMT3B, and DNMT3L.²¹ DNMT3L does not possess catalytic activity, and DNMT2 mainly methylates RNA. DNMT1, DNMT3A, and DNMT3B mediate two different DNA methylation mechanisms: (1) de novo methylation which creates new marks on DNA and (2) maintenance of previously methylated genomic sites during DNA replication.²¹ At gene promoters, the presence of DNA methylation is linked to chromatin silencing.^22,23 The repressive effect of DNA methylation on gene expression is mediated by two mechanisms: (1) the interference of DNA methylation with the recognition of transcription factor-binding sites, which results in impairment of gene activation,²⁴ and (2) the recognition of DNA methylation by specific Methyl-CpG-binding proteins that recruit corepressor protein complexes and thereby mediate silencing.²⁵

The role of DNA methylation in HIV-1 infection is controversial. Kauder et al. identified methyl-CpG-binding domain protein 2 as a regulator of HIV-1 latency and reported the hypermethylation of two CpG islands surrounding the HIV-1 transcriptional start site in latently infected Jurkat cells (J-Lat) and in primary CD4⁺ T cells.²⁶ Furthermore, methyl-CpG-binding domain protein 2, which is present at one of these CpG islands during latency and absent upon inhibition of cytosine methylation with 5-aza-2′deoxycytidine, possibly recruits transcriptional repressors to methylated DNA, and thereby contributes to HIV-1 latency.²⁶ A separate study showed high levels of CpG methylation of the HIV-1 5′ LTR in an in vitro model of latency and in latently infected T cells isolated from HIV-1-infected patients in a model in which CpG methylation acts as a late event during establishment of HIV-1 latency not required for initial provirus silencing.²⁷

Recently, Maricato et al. reported that HIV-1 infection increases the methylation level of cellular DNA in peripheral blood mononuclear cells (PBMC).²⁸ A separate study found overall low levels of 5′ LTR DNA methylation in resting CD4⁺ T cells isolated from a group of HIV-infected individuals under antiretroviral treatment for up to 3 years.²⁹ However, in some long-term treated individuals, the authors detected proviral molecules with a high density of 5′ LTR CpG methylation. They proposed that transient stimulation of cells harboring latent proviruses contributes, at least in part, to the methylation of the HIV-1 promoter over time.²⁹ Collectively, these four studies support a model in which DNA methylation contributes to HIV-1 transcriptional gene silencing, likely by inhibiting transcriptional initiation of the integrated provirus and inhibiting viral postintegration reactivation from latency.

In contrast, Blazkova et al. reported very low levels of methylated CpG dinucleotides within the promoter/enhancer region of latent HIV proviruses found in resting CD4⁺ T cells isolated from aviremic infected individuals.³⁰ The different results can possibly be explained by the use of different primers and the length of antiretroviral treatment of patients.

So far, only DNMT1, DNMT2, and DNMT3B have been implicated in HIV infection.^31,32 DNMT1, which is the most abundant DNMT in the cell and transcribed mostly during the S phase of the cell cycle, is primarily responsible for maintaining methylation through transfer of methyl groups to the hemimethylated DNA strands after DNA replication, but also has de novo DNMT activity.³³ In addition, DNMT1 influences transcriptional regulation by interacting with HDAC2 and novel corepressors to form a complex at replication foci.³⁴ Furthermore, DNMT1 represses transcription from E2F-responsive promoters by recruitment of transcription factors retinoblastoma protein Rb and E2F1, as well as HDAC1.³⁵ Recently, DNMT2 was shown to relocalize from the nucleus to stress granules and to methylate HIV-1 RNA in a sequence-independent manner, thereby providing posttranscriptional stability to the HIV-1 RNA.³²

DNMT3B, which is highly expressed in early embryonic stem (ES) cells, cannot differentiate between unmethylated and hemimethylated CpG sites, and cannot copy or contribute to the maintenance of a specific pattern of methylation.³³ However, DNMT3B has an important role in de novo methylation of unmodified cytosine residues.³³

Several groups reported that infection of CD4⁺ T cells with HIV-1 increases expression and activity of DNMTs.^31,36–38 One study showed that acute infection with wild-type and integration-defective HIV-1 increases cellular DNMT1 expression and activity, resulting in hypermethylation and reduced expression of the tumor suppressor gene, p16^INK4A, in lymphoid cell lines.³⁷ Youngblood and Reich reported that HIV-1 early gene expression caused DNMT1 induction, a process prevented by treatment with the phytoalexin resveratrol thought to interfere with transcription factor AP1 activity.³⁸ The HIV-1 responsive element was mapped to nt–1634 to–1214 in the DNMT1 promoter, which overlaps with AP1 sites.³⁸ An independent study evaluating the effect of HIV-1 infection on cellular vitamin D receptor expression showed that infection enhanced expression of HDAC1, DNMT1, and DNMT3B.³¹ Recently, Pion et al. reported that HIV-1 infection downregulates expression of master transcription factor Foxp3 in regulatory T cells, which was associated with an increase in the expression of DNMT3B and higher methylation of CpG sites in the FOXP3 locus.³⁹

Others linked the increase in DNMT expression to an overall increase in methylated genomic DNA in HIV-infected cells and the de novo methylation of a single CpG dinucleotide in the gamma interferon (IFN-γ) gene promoter to the downregulation of IFN-γ production in HIV-infected CD4⁺ T cells.^36,37 In an attempt to exploit these findings therapeutically, Martinez-Colom et al. designed a chimeric protein (IN3b) that linked the N-terminal domain of the HIV integrase enzyme with the C-terminal domain of DNMT3B to induce long-term silencing of HIV gene expression in host cells.⁴⁰ However, reduced DNMT1 and DNMT3A activity was reported in primary oral epithelial cells isolated from aviremic treated HIV⁺ individuals,⁴¹ suggesting that DNA methylation is deregulated, yet not always upregulated, in HIV infection.

A recent study aimed to investigate the influence of antiretroviral therapy on methylation markers found in PBMCs isolated from blood from a group of HIV-infected, ART-treated patients that the percentage of 5-methylcytosine was inversely correlated with proviral DNA and active replication, whereas DNMT1 and DNMT3A were independently correlated with active viral replication.⁴² Differential DNA methylation associated with HIV infection was also reported in a recent large-scale epigenome-wide study, where HIV-infected and uninfected patients from the Veteran Aging Cohort Study (VACS) were profiled for CpG sites in DNA extracted from the blood.⁴³ The authors identified 20 epigenome-wide significant CpGs for HIV-1 infection, including two CpGs in the promoter of the NLR family, CARD domain containing gene 5 (NLRC5), a key regulator of major histocompatibility complex class I gene expression, which showed significantly lower methylation in HIV-infected subjects than in uninfected subjects and which was negatively correlated with viral load in the HIV-infected samples.⁴³

Protein methyltransferases

PMTs are enzymes that covalently transfer methyl groups from the cofactor SAM to specific amino acid residues within nuclear and cytoplasmic proteins. Based on their structure, SAM-dependent MTs are divided into three classes: PMTs with seven-strand twisted β-sheet structure,⁴⁴ SET (SuVar3-9, enhancer of Zeste, Trithorax) domain lysine MTs,⁴⁵ and membrane-associated MTs.⁴⁶ These enzymes mediate protein methylation through O-methylation or N-methylation. Methylations of glutamic acid and aspartic acid are a type of O-methylation, and N-methylation involves methylation of lysine, arginine, histidine, alanine, proline, glutamine, phenylalanine, asparagine, and methionine. Among these, methylation of lysines and arginines is most common.⁴⁷ Lysines can be mono-, di-, or trimethylated, but arginines can be monomethylated or symmetrically or asymmetrically dimethylated. Protein lysine and arginine methylation results either in gene activation or repression, depending on the amino acid residue that becomes methylated and the state of methylation.

Lysine methyltransferases (KMT)

More than 50 human KMTs have been found that methylate histones and nonhistone substrates. Based on their catalytic domain, KMTs are grouped into two different families: (1) DOT1-like (DOT1L) that does not possess a SET domain, and (2) the SET domain-containing MTs (reviewed in Ref.⁴⁸). SET domain-containing MTs can be further divided into four families: SET1, SET2, SUV39, and RIZ.⁴⁸ Except for family members EZH1 and EZH2, the SET1 family is characterized by the SET domain, followed by a post-SET domain. SET2 class members are NSD1-3, SETD2, and the SMYD family, all of which have a SET domain positioned between a post-SET and an AWS domain. The SUV39 family members SUV39H1, SUV39H2, G9a, GLP, SETDB1, and SETDB2 all contain a pre-SET domain. Finally, the RIZ family members BLIMP1, PFM1, and RIZ1 are characterized by a SET domain at the amino terminus. In addition, several SET domain-containing methyltransferases include SET7/9, SET8, SUV4-20H1, and SUV4-20H2, which do not belong to any of these families. Figure 2 shows the KMT that have been implicated in HIV infection (Fig. 2).

FIG. 2.

Histone lysine methyltransferases (KMT) and histone methyl marks implicated in HIV infection, showing residue-specific KMTs for H3K4/9/27/36 and H4K20. The majority of KMTs are highly specific for a single histone residue, whereas a few enzymes target multiple residues, as indicated.

Enhancer of zeste homolog 2 (EZH2)

EZH2 belongs to the SET1 family of KMTs and trimethylates histone H3 lysine 27 (H3K27me3).⁴⁹ This subunit of the polycomb repressor complex 2 (PRC2) is an important regulator of chromatin state involved in the maintenance of transcriptional silencing.⁵⁰ EZH2 lacks enzymatic activity as an isolated protein and only methylates lysine residues when in complex with EED and SUZ12.⁵¹ EZH2 is modified through O-GlcNAcylation by O-linked N-acetylglucosamine transferase, resulting in increased stability of EZH2 and higher cellular histone H3 lysine 27 trimethylation levels.⁵²

EZH2 associates with the promoter/enhancer region of HIV-1 proviruses in latently infected Jurkat T cell lines and is found with the corresponding H3K27me3 mark.⁵³ Knockdown of EZH2 with shRNA or treatment with the chemical inhibitor 3-Deazaneplanocin A efficiently reactivates a significant portion of silenced proviruses. Furthermore, knockdown of EZH2 sensitizes latent proviruses to external stimuli, such as T cell receptor stimulation, and interferes with the reversion of reactivated proviruses back to latency, underscoring its restrictive role in HIV-1 latency.⁵³ In addition, Nguyen et al. showed that treatment with newer inhibitors of EZH2 (GSK-343, EPZ-6438) was sufficient to induce the reactivation of latent proviruses in a primary T cell model using primarily latently infected Th17-polarized cells or in resting memory T cells isolated from aviremic HIV-1-infected patients.⁵⁴ In contrast, Tripathy et al. showed that GSK343 reduced trimethylation of HIV provirus-associated histone H3K27 without increasing proviral expression in latently infected resting CD4⁺ T cells.⁵⁵ However, after reduction of H3K27 methylation at the HIV-1 LTR, subsequent exposure to the HDAC inhibitors, suberoylanilide hydroxamic acid (SAHA) or vorinostat resulted in increases in viral gag RNA and p24 antigen production, pointing to the combination of HMT and HDAC inhibitors as a promising reversal strategy in HIV latency.⁵⁵

Euchromatin histone methyltransferases EHMT1 (GLP) and EHMT2 (G9a)

Euchromatin histone methyltransferases (EHMTs) are SUV39 family members. They are evolutionarily conserved proteins that mono- and dimethylate histone 3 at lysine 9 (H3K9me1 and H3K9me2) in euchromatic regions of the genome.⁵⁶ In mammals, there are two EHMT proteins: G9a-like protein GLP is encoded by EHMT1, and G9a is encoded by EHMT2.⁵⁶ EHMTs are characterized by an N-terminal SET and Pre-SET domain and a series of ankyrin repeats that are required for protein–protein interactions, particularly with H3K9me1 and H3K9me2, the marks that EHMT proteins generate.⁵⁷ G9a methylates itself and mediates the interaction with the epigenetic regulator heterochromatin protein 1 (HP1).⁵⁸ G9a heterodimerizes with GLP through its carboxyl terminal SET domain to catalyze methylation.^59,60

In addition to H3K9, both proteins, independently and in the complex, mono- and dimethylate histone H3 lysine 27 (H3K27)⁶¹ and histone H1 isotype 4 (H1.4).^62,63 Additionally, during the G1 phase G9a monomethylates histone H3 lysine 56 (H3K56me1) to regulate DNA replication.⁶⁴ Furthermore, G9a and GLP methylate a number of nonhistone substrates, including CEBPβ, DNMT1, HDAC1, KLF12,⁶⁵ and lysine 373 of p53.⁶⁶

In HIV-1 latency, both G9a and GLP are responsible for transcriptional repression of the HIV-LTR by promoting repressive dimethylation at H3K9.⁶⁷ Imai et al. reported that wild-type, but not mutant, G9a lacking the SET domain significantly inhibited basal and tumor necrosis factor alpha (TNFα) or Tat-induced HIV-1 gene expression in ACH-2 and OM10.1 T cell lines. Treatment with BIX01294, a chemical inhibitor of G9a, reactivated HIV-1 in latently infected cells. When G9a expression was knocked down by small interfering RNAs, HIV-1 replication was augmented from cells transiently transfected with a full-length HIV-1 clone. The authors also used chromatin immunoprecipitation assays to show G9a and H3K9me2 on histones in the vicinity of the HIV-1 LTR.⁶⁷ H3K9 dimethylation was also decreased after treatment with 5′-deoxy-5′-methylthioadenosine, a broad-spectrum histone methyltransferase inhibitor, causing reactivation of latent HIV-1 in C11 cells.⁶⁷

In addition, chemical inhibition of G9a by the compound UNC-0638 reactivated latent proviruses in the Th17 primary T cell model of HIV latency, as well as in resting memory T cells isolated from aviremic HIV-1-infected individuals.⁵⁴ As Ding et al. showed that GLP knockdown also induces HIV-1 LTR expression,⁶⁸ the two EHMT enzymes may be significant in maintaining HIV-1 latency by catalyzing dimethylation of H3K9.⁶⁸

SETD2

The histone-lysine N-methyltransferase SETD2, also referred to as HYPB, trimethylates lysine 36 on histone 3 (H3K36me3), using dimethylated H3K36me2 as substrate.⁶⁹ H3K36me3 is a signature chromatin mark associated with active transcription implicated in coupling transcription with mRNA splicing.⁷⁰ Furthermore, SETD2 has been linked for maintenance of genomic integrity through coordination of homologous recombination repair after double strand breaks, suggesting a role as tumor suppressor.^71–73

In HIV infection, SETD2 is recruited to the RNA polymerase II (RNAPII) elongation complex in HLM107 cells, a HeLa-derived cell line that contains a single integrated, Rev-defective, HIV-1 provirus.⁷⁴ Previously, Yoh et al. reported that Spt6, a transcription elongation factor and histone H3 chaperone, binds serine 2 (Ser2P) of the C-terminal domain of RNAPII and recruits the “interacts-with-Spt6” (Iws1) transcription factor and the REF1/Aly mRNA export adaptor to facilitate mRNA export.⁷⁵ Subsequently, the same authors showed that Iws1 recruits SETD2 to the RNAPII elongation complex, which is required for H3K36 trimethylation across the transcribed region of the HIV-1 gene in HLM107 cells.⁷⁴

Knockdown of Iws1 disrupts binding of SETD2, but not Spt6, to the coding region of the integrated HIV-1 provirus, whereas depletion of SETD2 did not affect binding of these factors. These data suggest a mechanism by which Iws1 connects SETD2-mediated H3K36me3 with Spt6-driven nucleosome reassembly. Both are thought to depend upon the histone H3 chaperone activity of Spt6, thereby affecting mRNA export and the histone modification state of actively transcribed HIV-1.⁷⁴

SET domain containing lysine methyltransferase SETD7 (SET7/9, KMT7)

SETD7 was first identified to specifically monomethylate histone 3 lysine 4 (H3K4me1), a marker for transcriptional activation.^76,77 In addition, SETD7 methylates several nonhistone proteins, including p53,⁷⁸ retinoblastoma (RB),⁷⁹ NF-κB subunit p65,⁸⁰ TAF10,⁸¹ ARTD1,⁸² estrogen receptor (ER),⁸³ HIF-1α,⁸⁴ STAT3,⁸⁵ DNMT1,⁸⁶ SOX2,⁸⁷ SUV39H1,⁸⁸ PGC-1α,⁸⁹ β-catenin,⁹⁰ IFITM3,⁹¹ FOXO3,⁹² and YY1.⁹³ Despite the large number of substrates, SETD7's exact in vivo function remains largely undetermined. SETD7 knockout mouse models have no obvious developmental defects and show no apparent deficiencies in DNA damage and p53 responses.^94,95

SETD7 functions as a coactivator of HIV transcription.¹¹ The enzyme associates in vivo with the HIV promoter in reactivated J-Lat T cell models of HIV latency and monomethylates lysines 51 and 71, two highly conserved residues in Tat.^11,14 Knockdown of SETD7 suppresses Tat transactivation of the HIV promoter, but does not affect the transcriptional activity of methylation-deficient Tat (K51A or K71R).^11,14 SETD7 itself binds TAR RNA and forms a complex with Tat and P-TEFb, suggesting a positive role of SETD7-mediated Tat methylation in early steps of the Tat transactivation cycle.¹¹

SET domain, bifurcated 1 (SETDB1)

SETDB1 regulates the trimethylation of histone H3 on lysine 9 (H3K9me3), a specific mark for gene silencing and transcriptional repression.^77–79 In mouse ES cells, SETDB1 recruited to histone 3 by the Krüppel-associated box-associated protein 1 (KAP1/TRIM28/TIF1B), results in H3K9 trimethylation and subsequent silencing of endogenous and introduced retroviruses.⁸⁰ The human SETDB1 protein is organized into six domains: the tandem Tudor domains in the N-terminal region,⁹⁶ the methyl-CpG-binding domain in the middle region,⁷⁸ and the pre-SET, SET, and post-SET domains in the C-terminus.^77,78 Interestingly, to function as an H3K9 MT, SETDB1 needs to undergo posttranslational modifications.⁷⁸ The C-terminal region of SETDB1 is ubiquitinated at lysine 867, which is necessary for full H3K9 MT activity in mammalian cells.⁹⁷

SETDB1 associates with the HIV-Tat protein.¹³ SiRNA knockdown of SETDB1 in cell systems with both transient and integrated LTR reporter genes results in increased transcription of the HIV-LTR in the presence of suboptimal levels of Tat, indicating a repressive role for SETDB1 in HIV transcription.¹³ In vitro methylation assays with Tat peptides and SETDB1 show increased incorporation of methyl groups on lysine 51; however, lysine 50 was also susceptible for methylation.¹³ The association of Tat with histone methyltransferases and the ability for Tat to be differentially methylated at K51 suggest a very sophisticated mechanism of transcriptional regulation of the HIV-1 promoter as monomethylation by SETD7 acts as an activator and trimethylation by SETDB1 presumably as an inhibitory mark in Tat.

SET domain, bifurcated 2 (SETDB2)

SETDB2, also referred to as CLLD8/KMT1F, belongs to the SUV39 subfamily of histone 3 lysine 9 MTs. SETDB2 specifically trimethylates H3K9me3. The H3K9 histone methylation mark is associated with gene silencing, and recent work has linked SETDB2 to antiviral and anti-inflammatory responses through negative regulation of lipopolysaccharide and IFNβ-induced genes in macrophages.^98,99 Furthermore, SETDB2 has been associated with embryonic development and cell division.^100–102

SETDB2 coimmunoprecipitates with the HIV-Tat protein, but the function of this in vitro association remains to be determined.¹³ A separate study evaluating epigenetic modifications in primary blood cells after HIV-1 infection found SETDB2 transcripts highly upregulated in activated as well as resting cells, indicating that HIV-1 infection could be linked to the expression of this gene.²⁸

SET and MYND domain-containing protein 2 (SMYD2)

SMYD2 is a member of the SMYD family of five methyltransferases. SMYD1–5 contain a catalytic SET domain that is split by a zinc finger that contains the myeloid translocation protein-8, Nervy, and DEAF-1 (MYND) motif, followed by a cysteine-rich post-SET domain.¹⁰³ SMYD2 regulates transcription by methylating H3K36 and H3K4, functioning as a repressor or activator, respectively, depending on the presence of heat shock protein 90 (HSP90).^103,104 SMYD2 itself methylates HSP90 and, in muscle cells, controls protein complex formation in the cytoplasm.¹⁰⁵ Furthermore, SMYD2 inhibits p53 function by methylating lysine 370 (K370)¹⁰⁶ as well as K810 and K860 within the RB tumor suppressor (citation). Other nonhistone substrates of SMYD2 include ERα, poly-(ADP-ribose) polymerase 1 (PARP1), BTF3, PDAP1, AHNAK, and AHNAK2.¹⁰⁷

In latently infected T cell lines and in primary CD4⁺ T cells knockdown of SMYD2 or its pharmacological inhibition reactivates latent HIV-1, identifying SMYD2 as a repressor of HIV transcription.¹⁰⁸ This repressive function is associated with the enrichment of monomethylated lysine 20 at histone H4 (H4K20me1) at the latent promoter, a modification catalyzed by SMYD2 in vitro and in cells. The H4K20me1 “reader protein” lethal 3 malignant brain tumor 1 (L3MBTL1), which has chromatin-compacting properties, is recruited to the latent HIV-1 promoter in a SMYD2-dependent manner, supporting the model that the SMYD2-H4K20me1-L3MBTL1 axis promotes latency in HIV infection.¹⁰⁸

Suppressor of variegation 3–9 homolog 1 (SUV39H1)

SUV39H1, also known as KMT1A, was the first identified human lysine MT.¹⁰⁹ It specifically catalyzes trimethylation on histone H3 lysine 9 (H3K9me3) using monomethylated H3K9 as substrate.¹⁰⁹ H3K9me3 is a hallmark of facultative and constitutive heterochromatin and is enriched in silenced genes.¹¹⁰ SUV39H1 exerts its function in complexes with other H3K9 MTs, such as G9a, GLP, and SETDB1, by creating a binding site for HP1α, which binds H3K9me2/3 through its chromodomain.¹¹¹ HP1α further recruits SUV4–20H enzymes to heterochromatic regions, which generate H4K20me3, another heterochromatic histone modification.¹¹² SUV39H1 consists of a catalytic SET domain and a chromodomain, a reader domain for trimethylated H3K9.^113–115 The catalytic activity of SUV39H1 is reduced by trimethylated H3K4 confirming an observation in Drosophila that H3K9 methylation-dependent heterochromatin formation is initiated through active removal of H3K4 methylation.^116,117 So far, only histones have been identified as substrates for SUV39H1.

Several studies link SUV39H1 to HIV-1 latency. Marban et al. reported that, in microglial cells, DNA-bound CTIP2 associates with SUV39H1, which increases local H3K9 methylation.¹¹⁸ This results in the recruitment of HP1 proteins to the viral promoter and formation of local heterochromatin, leading to HIV-1 silencing.¹¹⁸ Similarly, du Chene et al. showed that SUV39H1, HP1γ, and H3K9me3 are associated with chromatin-mediated repression of integrated HIV-1 gene expression in several cell systems, including HeLa cells containing a single integrated copy of an HIV-1 LTR reporter gene and peripheral blood mononucleated cells (PBMCs) isolated from infected individuals.¹¹⁹ The importance of H3K9 trimethylation has also been shown in recent work using chaetocin, an SUV39H1/G9a histone MT inhibitor (HMTI).¹²⁰

Chaetocin treatment caused a 25-fold induction of latent HIV-1 expression in Jurkat cells containing a pTY-LAI-luciferase reporter virus, while exhibiting minimal toxicity and T cell activation. Induction of HIV-1 gene expression is associated with loss of H3K9 trimethylation at the viral LTR, and a corresponding increase in H3K9 acetylation, a marker for gene activation.^1,120 A separate study evaluating the therapeutic potential of MT inhibitors in HIV infection found that chaetocin induced HIV-1 transcription in 50% of CD8⁺-depleted PBMCs and in 86% of resting CD4⁺ T cell cultures isolated from aviremic HIV-1-infected individuals.¹²¹ Besides chaetocin, multiple short-chain fatty acids from periodontal pathogens and also therapeutic doses of irradiation reactivate HIV-1 and reduce the presence of SUV39H1 at the HIV-1 promoter.^122,123

B Lymphocyte-induced maturation protein-1 (Blimp-1)

The RIZ family member BLIMP-1, also referred to as PR domain zinc finger protein 1 (PRDM1), is a transcription factor that contains five Krüppel-like zinc fingers that mediate DNA binding, nuclear import, and recruitment of histone-modifying enzymes, as well as a PR/SET domain at the amino terminus (reviewed in Ref.¹²⁴). BLIMP-1 is critical for the differentiation of mature B cells into immunoglobulin-secreting cells and is also expressed in dendritic cells, macrophages, keratinocytes, and T cells.^125–129 In T cells, BLIMP-1 regulates the activation and generation of CD4⁺ and CD8⁺ T cell effector populations.^129–131 BLIMP-1 represses the transcription of several regulatory factors, including Bcl-6, T-bet, IL-2, IFN-γ, and IFN-β, while enhancing the transcription of IL-10.^132–135

Several studies address the role of BLIMP-1 in HIV-1 infection in different cell populations.^136–139 De Masson et al. showed that BLIMP-1 overexpression is associated with low viral transcription levels in central memory CD4⁺ T cells from HIV-1 elite controllers, a rare group of HIV-positive individuals who maintain undetectable viral loads in the absence of any treatment, suggesting that induction of BLIMP-1 may reduce the size of the HIV-1 reservoirs.¹³⁹ Other studies showed that BLIMP-1 expression is increased in chronically infected HIV-1 patients and correlates with enhanced expression of negative regulators of T cell activation, including PD-1, LAG3, and CTLA-4, as well as with T cell exhaustion and apoptosis.^136,137

Seddiki et al. reported higher BLIMP-1 levels in total CD4⁺ T cells from HIV-1 progressors than long-term nonprogressors or healthy controls, implying a positive role in HIV infection.¹³⁸ In support of this, the HIV-1 transactivator Tat was shown to induce BLIMP-1 expression in activated, but not in resting, CD8⁺ T cells.¹⁴⁰ Furthermore, Tat upregulates, in both activated CD4⁺ and CD8⁺ T cells, the release of IL-2,¹⁴⁰ which is an inducer of BLIMP-1,¹⁴¹ supporting a model that in activated T cells BLIMP1 could support HIV infection.

At the molecular level, the HIV-1 LTR includes binding sites for BLIMP-1, suggesting that this factor directly binds proviral DNA.¹⁴² BLIMP-1 was found highly expressed in memory, as compared with naive CD4⁺ T cells, where it represses basal and Tat-mediated HIV-1 transcription.¹⁴³ To do this, BLIMP-1 binds an interferon-stimulated response element within the HIV-1 provirus under latent conditions and is displaced after T cell activation. Reduction of BLIMP-1 in infected primary T cells increases RNA polymerase II processivity, histone acetylation, and basal HIV-1 transcription.¹⁴³ It remains unknown whether and how the SET domain of BLIMP1 is involved. However, it is intriguing that BLIMP1 has these strong repressive activities and we speculate that high BLIMP-1 levels in memory T cells contribute to the establishment and maintenance of latent HIV-1.

Protein arginine N-methyltransferases

In mammals, nine protein arginine methyltransferases (PRMT1-9) transfer methyl groups from SAM to the terminal guanidino nitrogens of arginine residues in target proteins⁴⁴ (Fig. 3). PRMT family members are characterized by four different motifs (I, post-I, II, and III) and a conserved threonine-histidine-tryptophan-loop important for the formation of the AdoMet-binding pocket.¹⁴⁴ Type I PRMTs form asymmetric (ω-N^G,ω-N^G) dimethyl-arginine, whereas type II enzymes give rise to symmetric (ω-N^G,ω-N‘^G) dimethyl-arginine.¹⁴⁵ Type III PRMTs generate monomethyl arginine at terminal nitrogen atoms (ω-NG-methylarginine), which is an intermediate of PRMT type I and II reactions.¹⁴⁶ Methylation of an arginine does not alter the positive charge of the guanidinium side chain, but changes its structure and the affinity between the substrate and its binding partners, resulting in promotion or inhibition of interactions.¹⁴⁷ So far, CARM1, PRMT1, and PRMT6 have been implicated in HIV infection (Fig. 3).

FIG. 3.

Histone arginine methyltransferases (PRMT) and histone methyl marks implicated in HIV infection, showing residue-specific PRMTs for H3 and H4.

Coactivator-associated arginine methyltransferase 1 (CARM1)

CARM1, also called PRMT4, produces mainly asymmetric dimethylated arginines (Rme2a) and methylates histone H3 primarily on two sites, R17 and R26, with an additional methylation site R42 recently identified.^148–151 CARM1 is recruited to promoters upon gene activation and methylates multiple nonhistone proteins involved in gene transcription, including transcriptional coactivators CBP/p300 and SRCs; mediator Med12; RNA pol II; RNA-binding proteins PABP1, HuR, and HuD, as well as the splicing factors CA150, SAP49, and SmB.^152–159

A recent study, investigating how H3K27 acetylation (H3K27ac) regulates HIV-1 transcription, found that acetylation at H3K27 (H3K27ac) increases H3R26 methylation (H3K26me2a) catalyzed by CARM1.¹⁶⁰ H3K27ac is a key epigenetic mark that correlates with gene transcriptional activation.^161,162 It recruits the super elongation complex (SEC), which is essential for HIV-1 LTR-mediated transcription.¹⁶⁰ As H3K27ac stimulates H3R26 methylation, this methylation mark interferes with SEC recruitment, initiating a negative regulatory feedback loop.¹⁶⁰ In line with CARM1's negative effect on SEC recruitment, its inhibition resulted in reactivation of HIV-1 transcription in several HIV latency cell models, including in primary resting CD4⁺ T cells, and acted synergistically with other latency reversing agents, such as the BET inhibitor JQ1 and HDAC inhibitor SAHA.¹⁶⁰

PRMT1

PRMT1 is a type I protein arginine methyltransferase that is primarily responsible for asymmetric dimethylation of H4R3me2a, a mark of transcriptional activation.¹⁶³ In addition, PRMT1 modifies a large number of other proteins, including histone H2, H3, hnRNPs, RNA helicase A, ERα, PIAS1, CITED2, FOXO1, and TAF15 (reviewed in Ref.⁴⁴).

In HIV-1 transcription, PRMT1 participates in NF-κB-dependent gene expression.¹⁶⁴ Hassa et al. showed that, under TNFα stimulation, PRMT1 synergistically coactivates the HIV-1 LTR together with p300/CREB-binding protein, coactivator-associated arginine methyltransferase 1 (CARM1), and poly (ADP-ribose) polymerase 1(PARP1). PRMT1 forms a nuclear aggregate with p65 and PARP1 and is recruited to p65-containing complexes at the HIV promoter.¹⁶⁴ However, the exact mechanism of crosstalk between these coactivators is still unclear. PRMT1 might directly methylate NF-κB, p300, CARM1, or PARP1 or exert its function through dimethylation of H4R3me2a, which has been shown to be essential to maintain “active” chromatin.¹⁶³

PRMT6

The type I protein arginine methyltransferase PRMT6 catalyzes asymmetric dimethylation of H3R2me2a, using monomethylated arginine as a substrate.¹⁶⁵ Dimethylation of H3R2 results in transcriptional repression and is also a negative regulator of H3K4 trimethylation (H3K4me3), as it inhibits the activity of the H3K4 methyltransferase MLL1 and blocks recruitment of the MLL-complex subunit WDR5 to histone H3.^165–167 In addition, PRMT6 dimethylates H3R42me2a¹⁵¹ and H2AR292a,¹⁶⁸ leading to either transcriptional repression or activation. Similar to other PRMTs, PRMT6 typically targets glycine–arginine-rich (GAR) motifs in substrates.^169,170 However, PRMT6 methylates also non-GAR motifs, such as in the HIV-1 Tat protein leading to restriction of viral transcription and replication.¹⁷¹

Interestingly, PRMT6 was also reported to automethylate itself in a non-GAR motif, which results in increased protein stability and activity.¹⁷² In agreement with its primary nuclear localization, PRMT6 methylates HMGA1a, a high-mobility group protein involved in transcriptional regulation of genes.^169,173 In addition, PRMT6 controls gene expression by directly interacting with transcription factors, including NF-κB and G-protein pathway suppressor 2.^174,175 Because those two molecules are directly involved in inflammatory responses, it is possible that PRMT6 also plays a role in inflammation responses.

PRMT6 restricts HIV infection by methylating and impairing the function of various HIV-1 proteins, such as Tat, Rev, and nucleocapsid protein p7 (NC). In the context of HIV-1, PRMT6 was first identified to associate and methylate Tat within cells.¹⁷¹ Overexpression of wild-type PRMT6 decreases Tat transactivation of an HIV-1 LTR reporter plasmid in a dose-dependent manner, and knockdown of PRMT6 increases HIV-1 production in HEK293T cells, indicating that PRMT6 acts as a restriction factor for HIV replication.¹⁷¹ Subsequently, Xie et al. reported that PRMT6 targets R52 and R53 in Tat, which leads to decreased interaction with the Tat transactivation region (TAR) in viral RNAs and also negatively affects complex formation of Tat-TAR with the critical cofactor cyclin T1.¹⁵

Interestingly, PRMT6 increases Tat half-life, which could play a critical role in Tat persisting within the infected cell and the extracellular environment.¹⁷⁶ Furthermore, overexpression of PRMT6 leads to the exclusion of R52/53-methylated Tat from the nucleolus.¹⁷⁷ Experiments using fluorescence recovery after photobleaching indicate that Tat's nucleolar accumulation is largely mediated through binding to nucleolar components, which is prevented by methylation of Tat by PRMT6.¹⁷⁷ PRMT6 also dimethylates Rev at a single arginine in the N-terminal portion of its arginine-rich motif and associates with Rev in vivo.¹⁷⁸ This study also showed that PRMT6 significantly decreases Rev-mediated viral RNA export from the nucleus to the cytoplasm in a dose-dependent manner.¹⁷⁸

Beyond that, PRMT6 dimethylates HIV NC in each of its two basic regions at positions R10 and R32, leading to decreased RNA annealing and diminished initiation of reverse transcription.¹⁷⁹ Thus, PRMT6 is an HIV restriction factor that acts at multiple steps of the viral lifecycle. Recently, inhibitors of PRMT6 have been developed for cancer therapy and could also have potential for treatment of HIV infection.^180,181

Other methyltransferases

Protein-L-isoaspartate (D-aspartate) O-methyltransferase (PCMT1)

PCMT1, also referred to as PIMT1, is an enzyme that catalyzes methyl esterification of L-isoaspartyl (L-isoAsp) and D-isoaspartyl (D-isoAsp) residues to L-aspartic acid (L-Asp).¹⁸² During stress-related conditions, oxidative environmental conditions, and with age, L-Asp and L-Asn residues are nonenzymatically modified through dehydration (L-Asp) or deamidation (L-Asn), resulting in formation and accumulation of L-isoAsp, D-isoAsp, and D-aspartic acid (D-Asp) residues.¹⁸² These changes result in structurally nonfunctional proteins.¹⁸³ In this context, PCMT1 functions as a chaperone or repair enzyme of aged or damaged proteins and facilitates the restoration to aspartate residues.¹⁸⁴

PCMT1 is ubiquitously expressed in all living organisms. Since its expression decreases with age, it has been associated with age-related diseases such as Alzheimer's dementia.¹⁸⁵ Furthermore, PCMT1-deficient mice show acceleration of aging characterized by induced accumulation of L-isoAsp, D-isoAsp, and D-Asp residues in proteins and dysfunction of proteins.¹⁸⁶

PCMT1 methylates histone H4 at aspartic acid 24 (H4D24me).¹⁸⁷ This histone mark is recognized by VprBP (HIV-1 viral protein R [Vpr]-binding protein), also known as DCAF1 (DDB1- and CUL4-associated factor 1), which is a chromodomain-containing protein.¹⁸⁷ In HIV infection, VprBP is recruited by HIV-1 Vpr to hijack the CUL4A-RBX1-DDB1-DCAF1/VprBP complex leading to cell-cycle arrest in G2 phase, and also to protect the viral protein from proteasomal degradation.¹⁸⁸

Trimethylguanosine synthase

Human trimethylguanosine synthase (TGS1) was originally identified as an interaction partner of PRIP (PPAR-interacting protein) and, therefore, named PIMT (PRIP-interacting protein with methyltransferase domain).¹⁸⁹ TGS1 is specific for guanine and catalyzes two successive methyl-transfer reactions from AdoMet to the N2 position of 7-methylguanosine.^190,191 The enzyme adds two methyl groups to RNAPII-transcribed small nuclear RNA (snRNA), small nucleolar RNA (snoRNA), and telomerase RNA, for the conversion of 7-monomethylguanosine caps to 2,2,7-trimethylguanosine caps.^190,191 The 5′-mRNA capping affects pre-mRNA synthesis and splicing, RNA transport to the cytoplasm, and mRNA translation and turnover.

Yedavalli and Jeang showed that, like snRNAs and snoRNAs, some Rev/RRE-dependent HIV-1 RNAs are TMG-capped by TGS1 and proposed a new regulatory mechanism for selective expression.¹⁹² The study showed that TGS1 enhances HIV-1 gene expression and that intracellular HIV-1 RNAs are 7-methylguanosine and trimethylguanosine capped. Furthermore, TGS1 selectively modulates the expression p55 and p24 HIV-1 proteins encoded by Rev/RRE-dependent RNAs. Activation of PBMCs or purified CD4⁺ T cells significantly increased the expression of TGS1, indicating that TGS1 might be a limiting factor in quiescent cells.¹⁹²

Demethylases

Up to date, two classes of lysine-specific KDMs are known: (1) the amine-oxidase type lysine-specific demethylases 1 and 2 (LSD1 and 2; also known as KDM1A and B, respectively), which are both dependent on flavin adenine dinucleotide (FAD) as a cofactor; and (2) the JumonjiC (JMJC) domain-containing histone demethylase, in which the demethylase activity is dependent on Fe(II) and α-ketoglutarate (2-oxoglutarate) (reviewed in Ref.⁴⁸). The latter consists of a group with over 30 members and can be divided, based on the JMJC-domain homology, into seven subfamilies (KDM2–8).^48,193–195

The LSD family members generate an imine intermediate that is hydrolyzed to the demethylated lysine and formaldehyde.¹⁹³ Upon recycling of the cofactor FAD, hydrogen peroxide is formed as a byproduct of demethylation. As these enzymes require a free electron pair on the lysine ɛ-nitrogen atom to initiate demethylation, LSD1 and 2 demethylate only mono- and dimethylated, but not trimethylated, lysines.¹⁹³ In contrast, the JMJC domain-containing demethylases are able to remove methyl groups from all three methyl lysine states, with concomitant production of succinate, carbon dioxide, the demethylated lysine and formaldehyde.^196,197 The target specificity of KDMs is regulated by their participation in different complexes.⁴⁸

So far, only LSD1 and the JMJC protein UTX have been implicated in HIV infection (Fig. 4). No reports link HIV to ten-eleven translocation (TET) enzymes involved in active DNA demethylation.¹⁹⁸ A subset of JMJC proteins also act as arginine demethylases.¹⁹⁹

FIG. 4.

Histone demethylases and histone methyl marks implicated in HIV infection, showing residue-specific KMTs for H3K4, H3K9, and H3K27.

Lysine demethylases

Lysine-specific histone demethylase 1A (KDM1A, LSD1)

LSD1 is the first discovered histone demethylase and belongs to the family of FAD-dependent amine oxidases.¹⁹³ LSD1 contains an N-terminal SWIRM domain, a tower domain and an amine oxidase domain at the C-terminus.^200–202 In the presence of the RCOR1/CoREST complex, LSD1 catalyzes demethylation of mono- and dimethylated H3K4.¹⁹³ In addition, LSD1 interacts with JMJD2C, a histone tridemethylase, and cooperatively removes methyl groups from mono-, di- and trimethylated H3K9.^203,204 LSD1 also demethylates several nonhistone proteins, including K370 in p53, K185 of E2F1, and K1096, in the DNA methylase DNMT1.^205–207

LSD1 is an HIV-Tat K51-specific demethylase, which is required for the activation of HIV-1 transcription in latently infected CD4⁺ T cells.¹² LSD1 and its cofactor CoREST associate with the HIV promoter in vivo and activate Tat transcriptional activity in a K51-dependent manner in T cells.¹² In addition, small hairpin RNAs directed against LSD1 or inhibition of its activity with the monoamine oxidase inhibitor phenelzine suppresses the activation of HIV transcription in latently infected T cells, indicating that the LSD1/CoREST complex, normally known as a transcriptional repressor, acts as a coactivator of HIV transcription in infected T cells by demethylating K51 in Tat.¹²

In contrast, in microglial cells, an important HIV-1 target in the central nervous system, LSD1 represses HIV-1 transcription and viral expression in a synergistic manner with the COUP-TF interacting protein 2 (CTIP2).²⁰⁸ CTIP2 forces heterochromatin formation and HIV-1 gene silencing by recruiting HDAC and HMT activities at the integrated viral promoter.¹¹⁸ Le Douce et al. further showed that recruitment of LSD1 at the HIV-1 proximal promoter is associated with H3K4me3 and H3K9me3 epigenetic marks and that LSD1-induced H3K4 trimethylation is linked to SETD1A recruitment at the integrated provirus.²⁰⁸ The cell-type-specific role of LSD1 in HIV-1 infection remains an interesting subject of investigation.

However, in latently infected T cells LSD1 inhibitors function to suppress reactivation of HIV transcription, in accordance with a proposed coactivator role of LSD1 in this cell type. Interestingly, they also suppress reactivation of latent infections of α-herpesvirus,²⁰⁹ suggesting a broader role of LSD1 in activating viral gene expression.

Lysine-specific demethylase 6A (KDM6A, UTX)

KDM6A, also referred to as UTX, specifically demethylates lysine 27 of histone 3 (H3K27).^210,211 UTX demethylates di- and trimethylated, but not monomethylated, H3K27.²¹² Since H3K27me2 and H3K27me3 are highly correlated with genomic silencing and repression of transcription, removal of these marks by UTX results in gene activation.²¹³ In addition, UTX interacts with protein complexes that are associated with H3K4 methylation, a mark of active transcription.^214,215 Furthermore, UTX binds the SWI/SNF chromatin-remodeling complex by engaging the BRG1 catalytic subdomain, resulting in nucleosome remodeling and gene activation of T-box factors.²¹⁶

At the HIV-1 LTR, the removal of the H3K27me3 mark by UTX is important for the robust induction of many specific genes during Tat-mediated HIV-1 transactivation.²¹⁷ Zhang et al. reported that Tat upregulates the expression of UTX and downregulates the H3K27me3 mark in TZM-bl cells.²¹⁷ Using chromatin immunoprecipitation assays, the authors found that UTX associates with nucleosomes at the 0 and +1 position relative to the start of transcription and downregulates the H3K27me2 and H3K27me3 states.²¹⁷ Also, UTX promoted HIV-1 gene expression by enhancing NF-κB p65 nuclear translocation, suggesting a novel function of UTX in the timely transition from poised to active chromatin in HIV-1 infection.²¹⁷

Concluding remarks

Current antiretroviral therapy suppresses HIV viral replication, requiring life-long adherence to continuously limit viral loads, but does not typically eradicate virus from the host. Successful sterilizing cure of HIV will require elimination of persistent viral reservoirs. One approach is to reactivate proviral genomes in latently infected cells to “purge” viral reservoirs.²¹⁸ Another is to find ways to permanently silence HIV in latent reservoirs so that likelihood of reactivation is diminished. Several epigenetic modifying agents, including inhibitors of HDAC, HMT, and DNMT, show great promise as antilatency therapeutics in vitro and ex vivo, and some, including valproic acid and SAHA/vorinostat, are currently being examined in clinical trials. Although the role of DNA methylation in HIV-1 infection is controversial, DNMT inhibitors are considered as reactivating agents. For example, the DNMT inhibitors Decitabine (5-aza-2′ deoxycytidine, aza-CdR) and its analog azacitidine (5-azacytidine, Vidaza^®), which are approved by the FDA for the treatment of myelodysplastic syndrome, are inducers of latent HIV-1.²¹⁹ In addition, decitabine synergizes with TNF-α and prostratin to significantly increase viral gene expression in several J-Lat cell lines.²⁶

Similarly, PMTs and demethylases are regulators with emerging impacts on HIV infection. Both groups harbor repressors (EZH2,⁵³ G9a,⁶⁷ SETDB1,¹³ SMYD2,¹⁰⁸ SUV39H1,^118,119 CARM1,¹⁶⁰ and PRMT6^15,171) and coactivators (SETD7,^11,14 LSD1,^12,208 UTX²¹⁷) for HIV infection through histone and nonhistone protein modification, most importantly Tat and NF-κB. Several inhibitors, such as BIX01294,⁶⁷ Chaetocin,¹²⁰ UNC-0638,⁵⁴ 3-deazaneplanocin,⁵³ AZ391,¹⁰⁸ GSK-343,^54,55 and EPZ-6438,⁵⁴ have been tested in preclinical model cell lines for latency reactivation as well as in cells derived from patients.⁶⁷ Interestingly, some HMT inhibitors, as discussed above, can enhance proviral reactivation in combination with bromodomain inhibitors, such as JQ1,¹⁰⁸ or PKC agonists (prostratin), or HDAC inhibitors (such as vorinostat/SAHA),⁵⁵ making this class of compounds possible candidates for combinatorial treatments.

Although research into HMT inhibitors and HIV latency reversal is less advanced than HDAC inhibitors, some methyltransferase inhibitors are now in clinical trials for the treatment of malignancies, making future clinical use in HIV infection more feasible. For example, the EZH2 inhibitors GSK126, tazemetostat, and CPI-1205, are currently in phase I clinical trials for the treatment of B cell lymphomas bearing EZH2-activating mutations (reviewed in Ref.⁴⁸). Similarly, LSD1 inhibitors TCP, ORY-1001, GSK2879552, and 4SC-202 are in current trials for the treatment of acute myeloid leukemia (AML), acute leukemia (phase I/IIA), for AML and small cell lung cancer (phase I), and hematological malignancies (phase I), respectively (reviewed in Ref.⁴⁸).

Acknowledgments

The authors thank the members of the Ott laboratory for helpful discussions. They thank Gary Howard and Lisa Loeb Stanga for editorial, John Carrol for graphics, and Veronica Fonseca for administrative assistance. This publication was made possible with the help from the amfAR Institute for HIV Cure Research, with funding from amfAR grant number 109301. They gratefully acknowledge support from the California HIV/AIDS Research Program (award no.: F13-GI-316) to D.B., as well as grant support from the CARE Collaboratory (U19 AI096113) and the NIH (RO1 AI083139 and RO1 DA043142) to M.O.

Author Disclosure Statement

No competing financial interests exist.