REVIEW The DNA methylation landscape of hematological malignancies: an update Pedro Blecua1 , Laura Martinez-Verbo1 and Manel Esteller1,2,3,4 1 Cancer Epigenetics Group, Josep Carreras Leukaemia Research Institute (IJC), Barcelona, Spain 2 Centro de Investigacion Biomedica en Red Cancer (CIBERONC), Madrid, Spain 3  Catalana de Recerca i Estudis Avancßats (ICREA), Barcelona, Spain Institucio 4 Physiological Sciences Department, School of Medicine and Health Sciences, University of Barcelona, Spain Keywords The rapid advances in high-throughput sequencing technologies have made cancer; epigenetics; hematology; leukemia; it more evident that epigenetic modifications orchestrate a plethora of com- lymphoma; methylation plex biological processes. During the last decade, we have gained significant knowledge about a wide range of epigenetic changes that crucially con- Correspondence M. Esteller, Josep Carreras Leukaemia tribute to some of the most aggressive forms of leukemia, lymphoma, and Research Institute (IJC), Carretera de Can myelodysplastic syndromes. DNA methylation is a key epigenetic player in Ruti, Camı de les Escoles s/n, 08916 the abnormal initiation, development, and progression of these malignan- Badalona, Barcelona, Catalonia, Spain cies, often acting in synergy with other epigenetic alterations. It also con- Tel: +34-93-557 2800 tributes to the acquisition of drug resistance. In this review, we summarize E-mail: mesteller@carrerasresearch.org the role of DNA methylation in hematological malignancies described in the current literature. We discuss in detail the dual role of DNA methyla- (Received 17 May 2020, accepted 4 June 2020, available online 3 July 2020) tion in normal and aberrant hematopoiesis, as well as the involvement of this type of epigenetic change in other aspects of the disease. Finally, we doi:10.1002/1878-0261.12744 present a comprehensive overview of the main clinical implications, includ- ing a discussion of the therapeutic strategies that regulate or reverse aber- rant DNA methylation patterns in hematological malignancies, including their combination with (chemo)immunotherapy. opening chromatin (in the case of histones). The latter 1. Introduction phenomena may often provoke the complete rewiring Epigenetic modifications are very common in our gen- of transcriptional programs. These two processes are omes. The most common consist of chemical additions commonly intertwined. For instance, acetylation of to DNA bases (cytosines, within the CpG dinucleotide histones leads to a more open state of chromatin, a context) or histones (acetylation, methylation, phos- lower level of nucleosome occupancy and DNA phorylation) at the DNA level or post-transcription- hypomethylation patterns, favoring transcription [1]. ally, respectively. Although these modifications are However, epigenetic modifications are not restricted to heritable and do not alter the chemical nature of their DNA methylation or histone modifications. Therefore, substrate, they may also have profound effects on their epigenetics could be more precisely defined as the set function, specifically by acting as ‘genetic switches’, of molecular phenomena that are not accompanied by regulating or turning on or off the expression of genes genetic lesions that nevertheless have important effects (in the case of DNA methylation), or closing or on gene regulation and function and that are heritable. Abbreviations 2HG, 2-hydroxyglutarate; 5mC, 5-methylcytosine; aKG, alpha-ketoglutarate; AZA, azacitidine (5-azacytidine); DAC, decitabine (5-aza-20- deoxycytidine azacytidine); DNMT, DNA methyltransferases; FH, fumarate hydratase; GoF, Gain of function; HDAC, Histone deacetylases; HMA, hypomethylating agent; hmC, 5-hydroxymethylcytosine; HSC, Hematopoietic stem cell; IDH, isocitrate dehydrogenase; lncRNA, long noncoding RNA; LoF, loss of function; miRNA, microRNA; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; SDH, Succinate dehydrogenase; TET, ten-eleven translocation methylcytosine dioxygenase; TME, tumor microenvironment; TSG, tumor suppressor gene. 1616 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. P. Blecua et al. Aberrant DNA methylation in hematological cancers Long noncoding RNAs (lncRNAs) and microRNAs immunosuppressive, T-exhausted phenotype). This has (miRNAs) exemplify this broader definition. The lat- recently enabled researchers to design drugs that ter, a family of single-stranded, ‘looped’ RNAs, tightly exploit the combined use of epigenetic therapies and control and downregulate the expression of many novel immune checkpoint inhibitors (ICIs). genes under normal homeostatic conditions [2]. This work focuses on aberrant DNA methylation pat- Technological advances (whole genome sequencing, terns in blood cancers, as there are already many studies WGS) and the advent of novel and improved sequenc- of the connection between genetic lesions and epigenetic ing technologies (assay for transposase-accessible chro- contributors; these may be found elsewhere [8,9]. We dis- matin using sequencing; and Hi-C) have prompted cuss the main causes and clinical implications of malig- recent efforts to link genetic alterations in epigenetic nant DNA methylation patterns, as well as their cellular effectors to cancer, specifically in hematological malig- regulators. This review is organized as follows: In Sec- nancies [3,4]. These mutations include coding and non- tion 1, we summarize what is currently known about the coding areas of the DNA, as well as chromosomal connection between DNA methylation in normal and rearrangements and the identification of novel chro- aberrant hematopoiesis; in Section 2, we discuss the matin structures. In this way, several authors have intertwining of cellular regulation and DNA methylation, shown how these genetic lesions affect key epigenetic specifically the latter’s interaction with the tumor regulators in leukemia and other hematological neo- microenvironment (TME) and the cell’s metabolism plasms that lead to its appearance and progression. (ME); in Section 3, we consider the link between DNA These include the following: alternative splicing events methylation and other epigenetic effectors, namely (e.g., U2AF1-Ser34 and SF3B1-K700E mutants) [4], lncRNAs, miRNAs, and histone modifications; and aberrant methylation of superenhancers of known finally, in Section 4, we comment in detail on the main tumor suppressor genes (TSGs), such as PAX5and clinical implications of all the above, with special empha- GATA2, and the activation of novel or known onco- sis on hypomethylating agents (HMAs) and combina- gene superenhancers (e.g., TAL1, EVI1, MYC), as tions of drugs, including ICIs. originally reported by Heyn et al. [5], aberrant 3D chromatin topologies (e.g., disruption of the cohesin/ 2. DNA methylation and CTCF complex); and dysfunctional lysine histone hematopoiesis demethylases (e.g., mutant LSD1 and LSD2) [4]. Analogously, DNA methylation anomalies that arise during hematopoiesis may trigger the initiation and 2.1. DNA methylation in normal hematopoiesis progression of hematological cancers. The pattern of DNA methylation at cytosine residues in the CpG Normal homeostatic control of hematopoiesis is a sequences is established during early hematopoietic dynamic and tightly spatiotemporally regulated process development and is heritable [6]. In humans, ~ 70% of by which the various types of terminally differentiated CpG dinucleotides are methylated, even though the mature blood cells are formed [10,11]. This takes place frequency of the CpG dinucleotide (3–8% of all cytosi- in the bone marrow, starting with hematopoietic stem nes) is relatively low in most of the human genome cells (HSCs), a set of progenitor cells with the ability because of CG suppression (spontaneous C-to-T con- to self-renew and differentiate into the various types of version), a process that is tightly controlled by the cell. blood cells, each of which has different functionalities When it fails in this purpose, the cell may undergo and biological properties. Even though the precise pathogenic transformations, through the epigenetic molecular mechanisms by which lineage commitment silencing of crucial TSGs or the unleashed expression occurs are currently the subject of intense debate [12– of pro-oncogenic genes (oncogenes), which therefore 14], it is well established that epigenetics is important promote the onset and/or progression of cancers, for regulating HSCs during every step of its transfor- including hemopoietic malignancies. However, DNA mation, and influences self-renewal, differentiation, methylation aberrations do not occur as single events, and the developmental fates of the various hematopoi- but rather often appear in synergy with other epige- etic progenitor cells (i.e., myeloid and lymphoid lin- netic lesions (e.g., with histone modifications) [7]. This eages, [15], see Fig. 1). Specifically, methylation shows that the epigenetic regulation of cell develop- analyses have shown how the epigenetic map changes ment and fate is a very complex and intricate matter. within the different stages of HSC differentiation Furthermore, DNA methylation abnormalities might [16,17]. DNA methylation marks are very stable and have profound effects on the cytotoxicity of certain are passed on through daughter cells through differen- immune system cellular subtypes (e.g., by inducing the tiation, making it an effective tool for cell lineage Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1617 Aberrant DNA methylation in hematological cancers P. Blecua et al. 1618 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers Fig. 1. DNA methylation and hematopoietic development. (A) Schematic of the ‘classical’ view of hematopoiesis, where starting from a HSC, the whole blood cell population is formed in every subsequent step (binary bifurcation points). Epigenetics plays an important role in regulating both the myeloid and lymphoid lineages. Genes with key roles in HSC self-renewal and pluripotency (termed stemness genes in the picture) are preferentially expressed at early stages of the process. On the contrary, as the different lineages are selected, the respective lineage-specific genes are expressed accordingly. (B) In hematological malignancies, the epigenetic patterns present under homeostatic control become aberrant and the cells may suffer malignant transformations in every stage of the process. We illustrate schematically the main consequences of key effector methylation enzymes’ malfunction and their impact in self-renewal, lineage bias, and differentiation, along with some examples of the resulting up- or downregulated target genes. reconstruction [18,19]. The vast majority of methyla- in otherwise healthy individuals might also predispose tion events actually occur in CpG-poor regions, in them to certain types of hematological cancers [25]. normal and neoplastic tissue [20]. In the former, the The main scope of this review is to provide an over- majority of these sites are additionally located distal to view of these DNA methylation-dysregulated pro- transcription start sites (TSSs), like in (super)enhancers cesses, bearing in mind that they often act [5,21]. It has been shown in murine models that DNA synergistically with other epigenetic aberrations, such methylation patterns are very dynamic, with different as histone marks [26], and the deregulation of both and characteristic patterns for myeloid and lymphoid lncRNA [27] and miRNA expressions [28]. lineages [22]. However, according to a complementary hematopoiesis model suggesting that HSCs might actu- 2.2. DNA methylation in hematological ally have a continuum of differentiation stages [23], malignancies emerging evidence shows cell lineage conversion, often as a mechanism to acquire resistance to therapeutic Epigenetic aberrations are common in hematological agents. Along these lines, Bueno-Costa et al. [24] very malignancies, and the latter are largely driven by recently showed how DNA (hypo)methylation is cru- extensive epigenome remodeling. LoF or misfunction cial for pre-B-acute lymphoblastic leukemia (ALL) cell in key DNA methylation-related enzymes is widely transdifferentiation into functional macrophages, observed in many types of hematopoietic neoplasms, affecting both cis- and trans-acting gene regulatory ele- such as myelodysplastic syndrome (MDS), myelopro- ments of the genome. liferative neoplasms (MPNs), acute myeloid leukemia During early stages of HSC development, genes (AML), T-ALL and diffuse large B-cell lymphoma associated with stemness (self-renewal and multipo- (DLBCL; see Table 1). LoF mechanisms comprise tency) are open for transcription, whereas genes mutation events and hypermethylation-driven gene related to specific functions and lineages remain (epige- silencing. The resulting aberrant DNA methylation netically) silenced (see Fig. 1A). During successive dif- patterns are often lineage-specific and are commonly ferentiation stages, genes associated with specific accompanied by secondary mutations [29,30]. Further- lineages start to be expressed, while, conversely, the more, coupled LoF of two such enzyme-expressing cells’ epigenetic machinery silences genes associated genes can act synergistically (e.g., TET2 loss simulta- with pluripotency (see Fig. 1A). In hematological neously occurring with DNAMT3A loss in AML) [31]. malignancies, these DNA methylation patterns may Interestingly, TET loss (increased hypermethylation) become aberrant and the cells may acquire malignant or DNMT loss (increased hypomethylation) can both phenotypes during this process. Aberrant methylation lead to hypermethylation and hypomethylation of the usually originates from a loss or gain of function (LoF promoters of target genes, upon careful examination and GoF, respectively) of the different proteins and of individual differentially methylated regions [32]. enzymes involved in the methylation process [DNMTs, Nevertheless, epigenetic abnormal plasticity in hemato- ten-eleven translocation methylcytosine dioxygenase logical cancers does not exclusively arise from defects (TETs), isocitrate dehydrogenase 1/2 (IDH1/2)], often in the aforementioned epigenetic remodelers, and sev- due to their acquisition of clonal somatic mutations eral other factors (or a combination thereof) might (including co-occurrence with secondary mutations in contribute to pathogenesis post-transcriptionally. For other ‘synergy-acting’ genes) or external factors affect- instance, it has been recently shown how hypermethy- ing enzyme function (e.g., substrate limitation; see lation-mediated silencing of the decapping enzyme Fig. 1B). Lastly, as has been pointed out in recent NUDT16 [e.g., a protein that removes the N7-methyl years, the process of clonal hematopoiesis (including guanosine (m7G) cap at the 50 end of gene transcripts] recurring mutations in DNMT3A, TET2, or IDH1/2) mediates c-MYC activation in T-ALL, both in vitro Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1619 Aberrant DNA methylation in hematological cancers P. Blecua et al. Table 1. DNMT, TET, and IDH observed mutations in hematological malignancies and their prognostic value. Gene Mutation Condition Frequency Prognostic value References DNMT1 Missense and nonsense mutations AML Small subset of cases Not studied [149–151] (rare mutations) DNMT3A Missense mutation (amino acid AML 20–60% (hot spot) Adverse prognostic impact [9,150,152 R882H) –154] MDS 10% Adverse prognostic impact [155,156] Frameshift and truncating mutations AML 15–20% Not studied [9] DNMT3B Truncating mutations AML Small subset of cases Not studied [157] (rare mutations) Missense mutation (amino acid ATL Small subset of cases Not studied [158] N442K) (rare mutations) TET1 Missense mutations AML ~ 1% Not studied [40,150] Missense and frameshift mutations T-ALL 14% Not studied [40,155] TET2 Several missense, nonsense, and AML ~ 10% Shorter overall survival (mutated [41] frameshift mutations vs no mutated) Truncating mutations MDS 10–30% Not studied [42,43] Several missense, nonsense, and MPN 10–20% Not studied [41–43] frameshift mutations Several missense, nonsense, and CMML 40–50% Not studied [41] frameshift mutations Several missense, nonsense, and DLBCL 5–10% Not observed [117] frameshift mutations IDH1 Missense mutation (amino acid AML ~ 10% Controversial [48,55] R132H) MDS 2–10% Controversial [55,156] IDH2 Missense mutation (amino acid AML ~ 10% Controversial [48,55] R172K) MDS 2–5% Controversial [55,156] Missense mutation (amino acid AML ~ 10% Controversial [48,55] R140Q) MDS 2–10% Controversial [55,156] and in vivo [33]. Bearing this fact in mind, the main HSC differentiation toward the myeloid lineage, with a DNA methylation remodelers and the resulting impact consequent reactivation of the myeloerythroid-specific of their impairment in blood cancers are considered genes Cd48 and Gata1 [34]. In humans, very few muta- below. tions have been detected in this gene (AML) and occur at very low frequencies (see Table 1). Unlike its DNMT1 counterpart, complete loss of Dnmt3a in 2.2.1. DNA methyltransferases mouse models has been shown to favor self-renewal, DNMT3A, DNMT3B, DNMT3L and DNMT1 are col- through the DNA methylation-induced dysregulation lectively known as ‘epigenetic landscaping’ genes. The of multipotency-related genes, such as Runx1 and first two are de novo methyltransferases, depositing Gata3 (see Fig. 1B). Defective HSC differentiation has methylation marks over an otherwise unmodified also been observed in the same Dnmt3a null mice DNA template. In turn, DNMT1 is a maintenance model, through hypermethylation of genes such as methyltransferase, which acts by propagating already Flk2 and Ikaros [32]. Mutations in this enzyme have existing methyl marks upon DNA replication [9]. been reported in AML and MDS patients. 20–60% DNMT3L has no catalytic domain and is thought to and 10% percent of patients, respectively, harbor the couple with DNMT3A. All three enzymes play a piv- hot spot mutation R882H (see Table 1) and are associ- otal role in mammalian development and hematopoi- ated with an adverse prognosis. Furthermore, mutated esis [9]. In murine models, the activity of the catalytic DNMT3A patients were shown to undergo promoter domain of WT Dnmt1 is essential for HSC self-re- hypomethylation and the subsequent upregulation of newal. Its impairment leads to downregulation of the the leukemogenic HOX co-factor MEIS1 [35]. Analo- self-renewal-associated genes Cd26l and Ski in murine gously, Dnmt3b knock-out mouse models have a simi- models (see Fig. 1B). Likewise, when misfunction of lar, but milder, phenotype than their Dnmt3a null Dnmt1 occurs in these models, there is a clear bias of counterpart. Mutations in DNMT3A have been 1620 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers reported in AML and MDS patients, with a preva- hematopoiesis [45]. TET2 deletion is related to lence of 20–60% (hot spot mutation R882H) and 10% enhanced self-renewal capacity and myeloid bias, as percent of patients, respectively, and are associated evidenced in a murine model-based study that showed with an adverse prognosis (see Table 1). A stronger greater expression of the self-renewal regulators Meis1 phenotype is nevertheless achieved when Dnmt3a and and Evi1, and a lower level of expression of myeloid- Dnmt3b both lose their function. In both cases, specific factors Cebpa, Mpo, and Csf1 [46,47]. DNMT3A/B deficiency results in a global hypomethy- lation pattern. 2.2.3. Isocitrate dehydrogenase 1 and 2 2.2.2. Ten-eleven translocation methylcytosine IDH1 and IDH2 are NADP+-dependent enzymes that dioxygenases catalyze the interconversion of isocitrate and aKG. IDH1 and IDH2 often acquire neomorphic (GoF) Ten-eleven translocation methylcytosine dioxygenase mutations, which are thought to occur early in leuke- family proteins are DNA methylation regulators also mogenesis [48]. This bestows novel catalytic activity on involved in hematopoietic differentiation. It consists of these enzymes, enabling them to convert aKG to its three enzymes, TET1, TET2 and TET3. They are able structural analog 2-hydroxyglutarate (2HG) [49,50]. In to convert 5-methylcytosine (5mC) to 5-hydrox- turn, this activity impairs the function of enzymes that ymethylcytosine (5hmC) through oxidation, 5hmC to require aKG as a substrate, like TET2 (described 5-formylcytosine (5fC), and finally to 5-carboxylcy- above), inhibiting the hydroxylation reaction of 5mC tosine (5caC), which eventually leads to loss of the by TET2 [48]. Missense mutations in IDH1 (amino 5mC mark and DNA demethylation [36–38]. They also acid change R132H) or IDH2 (amino acid change depend on cellular alpha-ketoglutarate (aKG) levels to R172K) have been observed in AML and MDS function properly. TET1 plays an indispensable role in patient cohorts (see Table 1), but their clinical impact hematopoiesis by regulating target genes that mediate remains controversial. TET2 LoF mutations and IDH leukemic transformation [39] and is considered to be a GoF mutations are mutually exclusive in AML [48], tumor suppressor (mutated in 14% of T-ALL patients, suggesting that lesions in these genes may in fact be see Table 1) [40] except in MLL-rearranged leukemias, biologically redundant. IDH1/2 GoF arises in AML where it acts as an oncogene, facilitating leukemogene- through the hypermethylation of several transcription sis by being directly activated by MLL fusion proteins factors that control myeloid differentiation, such as and upregulating the expression of key oncogenic tar- GATA1, GATA2, and EVI1 [48] and are associated get genes such as HOXA9, MEIS1, and PBX3 [39]. with increased HSC self-renewal (see Fig. 1B). Impairment through epigenetic silencing or missense and frameshift mutations (see Table 1) of the TET1 enzyme has also been observed in myeloid malignan- 2.3. Intertwining of cellular regulation and DNA cies, such as MDS, MPN, or AML [41–43], and is methylation associated with lymphoid (B lineage) bias and increased HSC self-renewal (see Fig. 1B). Genes down- The cellular metabolism and the TME are both tightly regulated by TET1 loss include GNA14, SMAD2–4, regulated by aberrant cells in order to survive under and CTNNB1 [40]. These genes are important in HSC different hostile environments, and they often influence homeostasis and transformation. Unlike TET1, which each other. Neoplastic cells are known to increase is often inactivated by epigenetic silencing, TET2 LoF nutrient uptake from their surroundings in order to often occurs by mutational processes. Mutated TET2 maintain their high biosynthesis rate and division may contribute to the initiation of both myeloid and capacity. As these cells prefer to perform aerobic gly- lymphoid malignancies [44]. In AML, 10% of patients colysis, intermediate metabolites accumulate during have been observed to harbor mutations (missense, this process, giving rise to by-products such as 3-phos- frameshift, and nonsense). By comparison, patients phoglycerate, which aids the function of the cancer who are wild-type (WT) for TET2 had a better OS cell’s one-carbon pathway (see Fig. 2B). Also, cancer (see Table 1). However, we have a poor understanding cells generate large quantities of lactate, which is usu- of how epigenetic changes induced by its loss con- ally released into the environment in conjunction with tribute to leukemogenesis. TET2 mutations lead to H+, with the consequent acidification of the environ- enhancer hypermethylation (up to 25% of active ment (lower pH; see Fig. 2D). Below, we describe the enhancer elements) and to a critical deregulation of known connections between metabolism, the TME, enhancer-associated gene expression patterns in and DNA methylation in hematological malignancies. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1621 Aberrant DNA methylation in hematological cancers P. Blecua et al. Fig. 2. Cellular regulation of DNA methylation. DNA methylation influences neoplastic cell’s metabolism and vice versa. (A) Glycolysis is regulated by, among others, the HIF1 pathway. A crucial TSG of this pathway, VHL, has been shown to be epigenetically silenced in hematological malignancies. See text for more details. (B) SAM is the substrate needed by DNMTs in order to methylate the DNA, and it is one of the limiting aspects that favors DNMT impairment in tumors. SAM is then converted to SAH, which usually accumulates and acts as an inhibitor of the process in the normal product-negative regulation of the enzyme function. See text for more details. (C) Although the tumor cell prefers to transform glucose into lactate to rapidly obtain ATP, intermediate metabolites and redox power, TCA cycle intermediates play an important role in methylation as several of them act upon TET demethylases. TET enzymes use a-KG as a substrate to actively demethylate DNA, and, as SAM, it is rather limited in the tumor cell. a-KG can be transformed into 2-HG by mutated forms of IDH1 or IDH2, which acts as a competitor of a-KG and impairs TET function. SDH and FH might be silenced in hematological malignancies, which originates an accumulation of succinate and fumarate, which together with 2-HG act as TET inhibitors in the cell. (D) Due to the increase in nutrient uptake, hypoxic conditions, redox stress, and environment acidification, the tumor cell creates an environment, which enhances tumor survival while it dampens immune cell activation, that is, favoring macrophage M2 polarization or Treg phenotype. 3PG, 3- phosphoglycerate; G6P, glucose-6-phosphate; a-KG, a-ketoglutarate. other TSGs, is involved in the HIF1 pathway, which is 2.4. DNA methylation and metabolism one of the pathways controlling the tumor’s glycolytic The alteration of metabolic pathways in hematological activity [51] (see Fig. 2A). Hypermethylation of VHL cancers and other malignancies is an essential step in in multiple myeloma (MM) patients [55] leads to tran- the aberrant growth and proliferation of neoplastic scriptional silencing of the gene and therefore cells, which are supported by deregulation of the epi- decreased HIF-1alpha proteolysis, suggesting a possi- genetic machinery [51]. Moreover, an aberrantly repro- ble mechanism for increasing angiogenesis and altering grammed metabolism also promotes changes in the the bone marrow microenvironment. In the case of the epigenetic landscape. DNA methylation is involved in one-carbon pathway (Fig. 2B), it was recently reported regulating three of the main metabolic pathways of that perturbing methionine/S-adenosylmethionine cancer cells: glycolysis [52], one-carbon and methionine (Met/SAM) metabolism in mixed-lineage leukemia- pathway [53], and the tricarboxylic acid (TCA) cycle AF4 cell lines caused the loss of expression and activ- [54] (see Fig. 2A–C). DNA methylation is indirectly ity of the H3K79 methyltransferase DOT1L. In turn, involved in preferential use of the aerobic glycolysis DOT1L loss reduced overall cellular methylation even in normoxia (the Warburg effect). This is sup- potential (low SAM/SAH ratio; SAH, S-adenosylho- ported by the fact that several tumor suppressors mocysteine; see Fig. 2B) and increased the apoptotic involved in pathways that control glycolysis activation rate in those cell lines [56]. The authors confirmed are often epigenetically silenced by promoter hyperme- their findings by pharmacologically inhibiting Met/ thylation, like the TSG VHL [55]. VHL, along with SAM metabolism in a clinically relevant patient- 1622 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers derived MLL-R leukemia xenograft model, which (DNMTs) and demethylases (TETs) [66] in reshaping resulted in increased survival. They attributed this the immune system and immune suppression of the result, in part, to the repression of the expression levels TME is being increasingly recognized. As an illustra- of DOT1L-regulated leukemia-inducing genes, such as tion of this, TET2-deficient CD8+ tumor-infiltrating MEIS1 [56]. On the other hand, another study showed lymphocytes (TILs) were shown to display greater that inhibition of DOT1L induces apoptosis of antitumor efficiency [67]. However, more research DNMT3A-mutated AML cells in vitro [57], resulting effort is needed in the coming years to fully elucidate in diminished expression of key proleukemic genes, the precise molecular mechanisms involved in this pro- including MEIS1. Given all these findings, it is reason- cess. able to speculate that alteration of the Met/SAM path- way and DOT1L inhibition may act synergistically to 2.6. DNA methylation and regulators of other stop leukemogenesis, particularly in DNMT3A-mu- epigenetic events tated and MLL-rearranged AML. Furthermore, dis- tortion of the TCA pathway is directly involved in As mentioned earlier, DNA methylation aberrations hematological cancers (see Fig. 2C). This occurs in are often accompanied by other epigenetic lesions in several types of myeloid malignancies, like MDS [58], hematological cancers. As an illustration of these intri- where IDH1/2 mutations induce the conversion of cate and complex processes, in the next three sections aKG into the oncometabolite 2-HG and therefore, as we describe the link between DNA methylation and mentioned earlier, the impairment of proper aKG-de- other epigenetic regulators. We focus on DNA methy- pendent DNA demethylase TET protein function. As lation effects coupled with lnRNA activity, miRNAs a consequence, aberrant hypermethylation patterns are deregulation and histone modifications, respectively, often observed in IDH1/2-deficient hematological neo- with a brief mention to (super)enhancers. plasms [58]. Finally, inhibition of succinate dehydroge- nase (SDH) and fumarate hydratase (FH) function in 2.7. Long noncoding RNAs the TCA cycle may lead to aberrant methylation pat- terns, leading to decreased aKG levels, as in the case Long noncoding RNAs and small noncoding RNAs mentioned above, and TET misfunction (Fig. 2C) [59]. (e.g., miRNAs) play a key role in the development and However, the latter phenomenon is, to our knowledge, progression of leukemia [68], and are therefore new poorly studied in hematological malignancies and biomarkers and potential targets for novel therapies [69]. more research in that direction is needed to demon- lncRNAs are nonprotein coding RNAs longer than strate LoF patterns, mainly through hypermethylation, 200 bps. They critically regulate gene expression and are of SDH and FH. highly tissue-specific [70]. Alterations in the expression of lncRNAs are thought to affect the onset and develop- ment of various hematological malignancies by modulat- 2.5. TME and DNA methylation ing critical cellular pathways of HSC development [4,71]. DNA methylation changes, including hematological While several studies have recently highlighted the neoplasms, occur frequently in cancer. These changes involvement of lncRNAs in blood neoplasms [72], to our help shape the TME, which in turn elicits immune tol- knowledge, little is known about the regulation of erance and drug resistance. The interaction between lncRNAs through DNA methylation in these type of DNA methylation and TME regulation is complex cancers, which is the main focus of this review. One such and has very recently attracted the attention of many example is the tumor suppressor lncRNA MEG3 hyper- groups [60]. In brief, neoplastic cells create an environ- methylation, reported in AML [73]. MEG3 downregu- ment that enhances tumor survival while dampening lates DNMT3A via MDM2/RB signaling to suppress immune cell activation. This is achieved by the leukemogenesis [74] (see Fig. 3B). Another study [75] tumor’s increase in nutrient uptake and induction of identified the role of (gene body) hypermethylated anti- low oxygen (hypoxia) and high lactate (environment sense lncRNAs (AS-lncRNAs) in ALL, including the acidification) levels in the TME (see Fig. 2D). Under lncRNAs MEIS1-AS2, MEIS1-AS3, AC092669.1, these conditions, macrophages with the M1 NEBL-AS1, and DLX6-AS1. The resulting repressed immunoactive phenotype are often driven toward the genes are MLL fusion genes, such as MEIS1 (See M2 immunosuppressive phenotype [61–64]. Analo- Fig. 3B). The authors proposed a putative oncogenic gously, the TME may inhibit cytotoxic T-cell expan- role of these genes independent of the MLL fusion part- sion and induce the consequent loss of their anticancer ner. James et al. [76] conducted a systemic study of the response [65]. The role of DNA methyltransferases influence of lncRNAs in B-cell ALL patients at initial Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1623 Aberrant DNA methylation in hematological cancers P. Blecua et al. Fig. 3. Impaired regulatory elements in leukemia and lymphoma. Different miRNA/lncRNA colors (left) match the corresponding up/ downregulated genes (right). Methylation of miRNA/lncRNAs promoters in the figure has all negative impact in the disease, except for the lncRNA MEG3, which has been shown to suppress leukemogenesis (see text). miRNAs might be both TSG and oncogenes, as specified in the text. miRNA genes marked with (*) are also found hypermethylated both in lymphoma and in leukemia (A) miRNAs genes are small noncoding RNA fragments that usually interact with the target mRNA and repress its translation. In leukemia and lymphoma, several miRNAs’ promoter regions have been described to be hypermethylated, which allows the target genes to be expressed. (B) lncRNA are long RNA fragments that by interacting with their target can interfere in several stages of target expression and function. In the figure, we show how hypermethylation of the promoter region of different genes allows that several members of MLL fusion gene family to be translated contributing to leukemia progression. (C) Superenhancers are DNA elements, which by loop formation allow an increase in the production of its targets. In leukemia, enhancers of TSGs are hypermethylated, whereas oncogene’s enhancers are found to be hypomethylated. E, enhancer; Me, 5-methyl cytosine; RISC, RNA-induced silencing complex. diagnosis and at relapse. They focused on the main dis- they found lncRNAs with a novel role in B-cell ALL, ease subtypes and characterized the link between diagno- namely R11-138M12.1 and RP11-624M8.1, that were sis- and relapse-specific lncRNAs based on the significantly hypomethylated at their promoter region differential expression and differential methylation pat- and transcriptionally upregulated in one of the sub- terns between the two conditions and within subtypes. groups. We summarize some of these and other examples They identified previously known (relapse-specific) onco- of novel (de)methylated lncRNAs in Fig. 3B, as well as lncRNAs that were promoter-hypomethylated in at least novel DNA methylation-induced deregulation of (super) one of the subgroups, including RP11-701P16.5 and enhancers (Fig. 3C), such as those corresponding to SLC38A3, as well as TCL6 and LINC00312, which were MYC and RNF43, whose role in hematological malig- observed to be promoter-hypermethylated. Furthermore, nancies is being increasingly recognized [5]. 1624 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers DNA methylation patterns and histone modifications 2.8. microRNAs in hematological neoplasms (and cancer in general) microRNAs are a set of regulatory ncRNAs, about 22 have been reported. In hematological cancers, histone nucleotides long, that usually repress the transcription marks accumulated on promoters are often enriched in of their target genes [77]. They are crucial to physio- H3k27me3 alone or in combination with H3K4me3 logical and pathological processes such as cell differen- (bivalent promoters) [85]. Both of these histone modifi- tiation and inflammation, and to the pathogenesis of cations are negatively correlated with DNA methyla- major hematological malignancies, such as AML [78]. tion and transcription regulation. H3K27me3 is miRNAs may regulate DNMT activity, and on the associated with promoter hypermethylation and gene other hand, hypermethylation of miRNA promoters silencing [86] in HSCs, while H3K4me3 is usually asso- may inhibit their function, which can be either tumor- ciated with hypomethylation of promoters and gene suppressing or oncogenic [72]. An example of miRNA upregulation [87]. However, recent studies suggest that deregulation caused by aberrant methylation is miR- H3K4m3 and DNA hypermethylation also have a syn- 29b tumor suppressor upregulation in bone marrow ergistic function [88]. For the sake of clarity, we specif- cells, which reduces the expression of DNMTs, ically focus here on H3k27me3, whose proper enhancing the activity of demethylating agents (decita- maintenance is known to be critical for the homeosta- bine) and improving the remission rate of AML sis of normal cells, and whose deregulation often leads patients, reactivating the other TSGs [79]. On the to aberrant hematopoiesis [4]. Trimethylation of his- other hand, inhibition of the onco-miRNA-221 could tone H3K27 (H3K27me3) is mechanistically linked to directly reactivate TSGs and, at the same time, upreg- the polycomb group (PcG) proteins through its core ulate the expression of other TSGs by downregulating member, EZH2, which has histone methyltransferase Dnmt1 activity in mice [80] (see Fig. 3A). As men- activity with substrate specificity for H3K27 through tioned above, the promoters of certain miRNAs may its SET domain (as do other core components of the also be targeted by DNMTs, causing their hyperme- PRC2, acting with EZH2, such as EED and SUZ12) thylation (in the case of tumor suppressor activity). [4]. The histone methyl mark imposed by PRC2 is Along these lines, HOXB3 enhanced DNMT3B bind- strongly associated with transcriptional repression [89]. ing to the promoter of the tumor suppressor miR-375, Consistent with these observations, PRC2 has been leading to DNA hypermethylation and a lower level of shown to be associated with de novo DNMT activity expression of miR-375 [81] (see Fig. 3A), thereby pre- and, consistent with this, H3K27me3 and PRC2 tar- dicting poorer outcome in AML. In Fig. 3A,C, we gets are known to be positively correlated with DNA summarize recent findings of these and other pre- (hyper)methylation [90,91]. However, PRC2 may func- miRNA promoter regions in hematological cancers, tion as a tumor suppressor or an oncogene in leuke- which are found to be hypermethylated, as well as mogenesis and lymphomagenesis, depending on the their molecular impact on the respective diseases, biological context. LoF mutations in EZH2 are associ- where these have been observed. ated with poor survival in MDS and MPNs [92], high- lighting the role of the PRC2 complex as a tumor suppressor. On the other hand, GoF mutations or 2.9. Association between histone modifications overexpression of EZH2 may often lead to a more and DNA methylation patterns aggressive disease phenotype, repressing tumor sup- It is not currently fully understood how histone modi- pressor CDKN2A expression, which in turn is linked fications affect DNA methylation and vice versa. Aside to HSC proliferation and self-renewal [93], like in the from aberrant DNA promoter hypermethylation or case of MLL-rearranged leukemias [94], suggesting hypomethylation, histone modifications are recognized that the PRC2 complex may act as an oncogene. With as important mechanisms in cancer initiation and pro- respect to the latter, that is, GoF EZH2 mutations, gression, including those of hematological cancers relatively recent research provides evidence of that the [4,82]. These modifications are of three main types— clinical utility and development of small molecule acetylation, methylation, and phosphorylation—all of EZH2 inhibitors provide a means to arrest aberrant which occur post-translationally. As with DNA methy- leukemic/lymphoid transformation in certain hemato- lation patterns, they may activate or repress transcrip- logical malignancies. Future research effort is needed tion. For instance, histone-3/4 acetylation leads to a to evaluate the usefulness of the combined action of more open chromatin state, favoring transcription [83]. EZH2 and DNMT inhibitors in blood cancers, as However, histone 3 methylation may lead to transcrip- recently reported in MM therapy-resistant cell lines tional repression or activation [84]. Links between [95]. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1625 Table 2. (Pre)Clinical status of current DNA methylation-related drugs in combination with other epigenetic therapies or chemoimmunotherapy. BP, blast phase; CR: complete remission; DHFR, dihydrofolate reductase; G-CSF: granulocyte colony-stimulating factor; RR, refractory/relapse; TK, tyrosine kinase. 1626 Current clinical Drug Target Target domains Condition status NTC number Reference AZA + belinostat DNMT/HDAC DNA methylation/histone deacetylation AML Completed NCT00351975 – AZA + entinostat DNMT/HDAC DNA methylation/histone deacetylation AMLMDS Phase II, completed NCT00313586 [161] AZA + entinostat DNMT/HDAC DNA methylation/histone deacetylation AML (old) Phase II, recruiting NCT01305499 – AZA + homoharringtonine DNMT/TET DNA methylation/ AML Phase III, DNA demethylation recruiting NCT04248595 – AZA + pinometostat DNMT/DOTL1 DNA methylation/histone methylation AML recruiting NCT03701295 – AZA + pracinostat DNMT/HDAC DNA methylation/histone methylation AML Phase II, completed NCT01912274 [162] AZA + vorinostat + DNMT/HDAC/CD33 DNA methylation/histone methylation/ AML Phase I/Phase II, NCT00895934 [163] gemtuzumab Leukemic cell marker completed ozogamicin ASTX727 + FT-2102 DNMT/IDH-1 DNA methylation/cell metabolism RR-MDS Phase I/Phase II, NCT04013880 – Aberrant DNA methylation in hematological cancers RR-AML recruiting Decitabine + LBH589 DNMT/HDAC DNA methylation/histone deacetylation AML Phase I/Phase II, NCT00691938 [164] MDS completed Decitabine + vorinostat DNMT/HDAC DNA methylation/histone methylation RR-AML Phase I, completed NCT00479232 [165] AML MDS Decitabine + vorinostat DNMT/HDAC DNA methylation/histone deacetylation RR-AML Phase II, terminated NCT00882206 [166] Decitabine + vorinostat + DNMT/HDAC/ DNA methylation/ RR-AML Phase I, filgrastim + fludarabine + G-CSF/ histone recruiting cytarabine DNA and deacetylation/ RNA Neutrophil growth/ polymerases DNA synthesis NCT03263936 – Decitabine + vorinostat + DNMT/HDAC/ DNA methylation/ RR-AML Phase I, filgrastim + fludarabine + G-CSF/ histone cytarabine + sorafenib DNA and deacetylation/ RNA Neutrophil growth/ polymerases/ DNA synthesis TK terminated NCT02412475 – Decitabine + vorinostat + DNMT/HDAC/ glucocorticoid DNA vincristine + dexamethasone + tubulin/ receptor/ methylation/ mitoxantrone + pegaspargase + topoisomerase histone methotrexate type II/DHFR deacetylation/Cell ALL Phase I/Phase II, terminated NCT01483690 [167] division/DNA synthesis and repair/cell metabolism P. Blecua et al. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Table 2. (Continued). Current clinical Drug Target Target domains Condition status NTC number Reference NTX-301 DNMT1 DNA methylation AML Phase I, not NCT04167917 – MDS yet recruiting Decitabine + nivolumab + DNMT/PD-1/ DNA methylation/ AML Phase I, CDX-1401 DEC-205 Immune MDS active presentation (not recruiting) NCT03358719 – AZA + nivolumab + ipilimumab DNMT/PD-1/ DNA methylation/ RR-AML Phase II, CTLA4 Immune recruiting presentation NCT02397720 – Decitabine + nivolumab + venetoclax DNMT/PD-1/ DNA methylation/ AML Phase I, BCL-2 Immune recruiting presentation/ apoptosis NCT04277442 – Guadecitabine + atezolizumab DNMT/PD-L1 DNA methylation/ AML Phase I/ Immune MDS Phase II, presentation CMML recruiting NCT02935361 – AZA + lenalidomide DNMT/IKAROS DNA methylation/Immune response CR-AML Phase II, completed NCT01301820 [168] Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. AZA + lenalidomide DNMT/IKAROS DNA methylation/Immune response RR-AML Phase II, completed NCT01743859 [169] Aberrant DNA methylation in hematological cancers 1627 Aberrant DNA methylation in hematological cancers P. Blecua et al. whether the degree of hypomethylation achieved with 2.10. Clinical implications drug treatment predicts clinical response in humans As described above, it is increasingly evident that epi- [105]. DAC and AZA both reverse DNA methylation genetic modifications play a major role in aberrant patterns and rewire oncogenic transcriptional pro- hematopoiesis. DNA methylation is one of the main grams, modulating the expression of the genes that processes whose deregulation leads to malignant cell drive neoplastic progression and differentiation, as well transformation and progression in hematological can- as derepressing transcription of TSGs [106], as men- cers. These aberrant modifications in DNA (de)methy- tioned earlier. However, treatment with either agent lation-inducing genes may be reversed and are thus has been reported to cause different gene expression attractive clinical targets. Furthermore, disease-specific profiles in various leukemia cell lines [105]. Mechanisti- DNA methylation signatures have been developed cally, these two drugs act differently. DAC binds to within the last decade in an attempt to stratify patients DNA, then, during the S phase of the cell cycle (DNA and therefore personalize their treatment to achieve replication), binds covalently to the DNMT enzymes improved outcomes. In this last section, we will briefly and inhibits their activity (mainly DNMT1, with describe the main drugs (either FDA-approved or higher sensitivity), triggering their further proteasomal being clinically trialed) that reverse these aberrant degradation. In addition, DAC interferes in the syn- methylation patterns, either alone or in combination thesis of new DNA during the S phase, impairing cell with other targeted therapies and including proliferation and causing apoptosis [107]. Conversely, immunotherapy (Table 2), as well as their clinical rele- AZA preferentially binds the cells’ newly synthesized vance and implications. We will also review the prog- RNA (~ 80–90%, the rest binds to DNA), giving rise nostic methylation signatures developed to date with to mRNA and protein metabolism disruption and respect to the various hematological neoplasms. apoptosis. Around 10–20% of AZA is converted to DAC and binds DNA [108]. Nevertheless, and despite their proven efficacy in several hematopoietic malig- 2.11. DNMT inhibitors nancies, HMAs are not curative and treatment with Patients who are not fit to receive standard chemother- them needs to be continuous [101]. Currently, several apy for AML, or high-risk MDS patients, may be efforts to bypass this and other limitations and achieve treated with HMAs [96]. HMAs are generally consid- more durable responses and patient remission rates, ered effective and safe [97]. Some have already been along with combination therapies, are being investi- approved, and others are currently undergoing clinical gated by the research community to improve the effi- trials. The first DNMT inhibitors were discovered in cacy of these epigenetic remodeling drugs (Table 2). the 1960s [98], and two decades later were introduced One such example is given by administering both into the clinical setting as effective epigenetic modify- DNMT inhibitors and histone deacetylase inhibitors ing agents. In order for these inhibitors not to be cyto- (HDACis) [105], given that DNMTs also attract toxic, they are usually administered at low, HDACs to the CpG loci they act upon, thus further subcytotoxic doses [99]. The two most clinically estab- stabilizing the silencing of the target gene [110]. In lished first-generation HMAs that have gained FDA Table 2, we summarize some ongoing clinical trials approval are the cytosine analogs decitabine (5-aza-20- that combine these two therapeutic agents. However, deoxycytidine azacytidine, DAC) and azacitidine (5- with the development of new and effective HDAC azacytidine, AZA). Briefly, they work by being incor- inhibitors arise, like the novel small molecule HDAC6 porated into the replicating DNA in place of cytosine. inhibitor QTX125 for mantle cell lymphoma [111], it The covalent, irreversible bond DAC-G traps will be possible to test new combination therapies in DNMTs, bringing about their degradation by the pro- the clinical setting. Another interesting example is that teasomal machinery [100]. As this process takes place of the FDA-approved combination of venetoclax during the S phase of the cell cycle, DNMTs are not (BCL-2 inhibitor) and DAC or AZA [112] for treating available and aberrant methylation patterns are no elderly AML patients, which is already the subject of longer reproduced in daughter cells, inducing demethy- two Phase 3 trials (NCT02993523 and NCT03069352). lation and reactivation of previously silenced genes, Finally, second-generation HMAs are being investi- including TSGs [101]. They are the most effective epi- gated. One such drug is guadecitabine (Phase 3 clinical genetic therapies to date: Both drugs provide signifi- trial) [97], which is an HMA that is more resistant to cant clinical benefits, with high overall response rates degradation and, consequently, increases both the and improved overall survival in patients with myeloid response rate and treatment efficacy, as reported by malignancies [102,103,104]. However, it is not clear Griffiths et al. [113]. 1628 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers of angioimmunoblastic T-cell lymphomas (AITLs) 2.12. TET inhibitors [118]. These authors speculated about the possible To our knowledge, no specific TET2 inhibitors have oncogenic cooperation between TET2 and DNMT3A so far had any clinical application. However, TET2 mutations. The effect of AZA in AITL patients with mutations are relatively common in myeloid cancers TET2 mutations was evaluated by Delarue et al. [119], [42] and are present in 20–25% of MDS, 7–23% of who reported an objective response rate to AZA treat- AML, and up to 53% of chronic myelomonocytic leu- ment, including complete remission. Further evidence kemia (CMML) patients [114–116]. TET2 mutations of the clinical utility of TET2 inhibitors was provided have also been observed in lymphomas (12% of by Dominguez et al. [120], who observed growth inhi- DLBCL patients, predominantly in the GCB subtype) bition of TET2-knockdown DLBCL cells after treat- [117]. They occur most frequently in T-cell lym- ment with a histone deacetylase 3 (HDAC3) inhibitor phomas, specifically, they are present in at least 50% in vitro. This highlights once more, as in the case of Fig. 4. Mechanism of action of TET inhibitors. (A) Direct inhibition. A novel compound discovered by Chua et al. [121], Bobcat 339 emerged from a selection of different TET enzyme inhibitors as the most successful on inhibiting the enzyme function by binding its Cl residue on the pocket reserved for Me. The drug blocks both TET1 and TET2 enzymes and does not interact with other methylation enzymes such as DNMTs. (B) Indirect inhibition. JAK/STAT pathway is involved in TET1 transcription, and STAT inhibitor UC-514321 seems to stop aberrant TET1 function found in AML according to Jiang et al. [122]. (C) Indirect inhibition. SP1 appears to be a transcription factor involved in TET1 transcription, and as in the case of UC-514321, blocking TET1 transcription and translation in AML gives promising results avoiding the spread of the malignancy. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1629 Aberrant DNA methylation in hematological cancers P. Blecua et al. HMAs, the clinical utility of combining epigenetic early MDS was significantly associated with transfor- therapies. Even though the study of TET2 inhibitors is mation to AML. Notably, there was no significant currently only at the preclinical stage, several mecha- variation in the promoter methylation status over time nisms of action (MoA) have been proposed. Direct (a period of 1132 days, with a median 284 days inhibition of TET1 or TET2 was suggested by Chua between measurements) [124]. Patient age was not sig- et al. [121], who found that the novel compound Bob- nificantly associated with these observations either. cat 339 best inhibited the enzymatic function of TET1/ Considered as a whole, this and the other signatures 2 by binding to its catalytically active residue. The described in the literature suggest that they may have drug blocked the enzymatic activity of both TET1 and profound clinical implications, as the respective disease TET2, and did not interact with DNMTs (Fig. 4A). subtype or patient group could be treated with the Two other studies demonstrated the indirect inhibition HMAs described above (or in combination with other of TET1. Jiang et al. [122] showed that the JAK/STAT epigenetic drugs, Table 2). Nevertheless, it should be pathway is involved in TET1 transcription in AML, pointed out that even though the response to epige- and by using the STAT inhibitor UC-514321, they netic drugs might be monitored, for instance, by mea- were able to stop aberrant TET1 function in vitro suring signature reversal throughout the course of (Fig. 4B). Finally, Li et al. [123] recently showed that treatment, there are currently no cost-effective tools SP1 is a transcription factor involved in TET1 tran- available in the clinic for that purpose. Some authors scription in AML. Using the potent FDA homohar- have already proposed that pretreatment methylation ringtonine, they managed to block TET1 transcription levels of long interspersed element (LINE-1) in elderly and translation while arresting disease progression AML patients treated with AZA are sufficient to pre- (Fig. 4C). dict the clinical response. In this study, lower baseline levels of LINE-1 methylation were noted in those patients who went on to achieve a complete or partial 2.13. DNA methylation signatures as diagnostic remission [126]. However, the low patient number was and prognostic tools a key limiting factor of this study,larger cohorts need The reversible nature of epigenetic changes, as to be studied in order to validate this relationship. In opposed to genetic mutations, offers the possibility of fact, the reliable assessment of HMA response in the using DNA methylation as an attractive therapeutic clinical setting remains a topic of intense debate target. Therefore, several attempts have been made [105,127]. over the past 20 years or so to find aberrant DNA methylation patterns that might have a diagnostic or 2.14. DNA methylation and the (chemo) prognostic value for several hematological cancers. immunotherapy response These patterns or signatures may consist of a single gene or a combination of them. The highly dynamic Resistance to HMAs is frequent in MDS [128] and and plastic nature of DNA methylation alterations other hematological malignancies, as is, to a lesser means that these aberrations very often change over extent, nonresponsiveness [129,130]. Treatment of time, that is, over the course of a disease [124] and MDS or AML patients with HMAs may induce with age [125]. In this way, some of these patterns are immune reaction alterations in those patients [131], positively selected by cancer cells to favor neoplastic indicative of a putative role of this immune deregula- development, as is certainly the case for both MDS tion in HMA resistance. For instance, it has been and AML [125]. Thus, the signatures found so far are shown that treatment of DLBCL cells in vitro with usually stage- or age-specific, as well as often subtype- low-dose DNMTi sensitized them to standard specific (due to the great heterogeneity of methylation chemotherapy, mainly through promoter demethyla- patterns in these malignancies). As a good example of tion of SMAD1 [132]. Consequently, a Phase I clinical a DNA methylation signature, Aggerholm et al. [124] study was performed combining AZA with chemoim- found four genes whose promoters were often hyper- munotherapy, in which demethylation of SMAD1 methylated in MDS (p15INK4B, HIC1, CDH1, and could be tracked and confirmed pre- and post-treat- ERa). While hypermethylation of three or more of ment, reaffirming the initial study’s identification of these genes occurred more frequently in advanced this gene as a candidate chemosensitization agent MDS, it became clear that promoter hypermethylation [132]. Despite this and other research efforts, the pre- of one or more of these four genes was predictive of cise molecular mechanisms linking resistance/refractory poor prognosis in patients with early-stage MDS. Fur- behavior and the immune landscape reshaping of these thermore, p15INK4B hypermethylation in patients with neoplastic cells are not well understood. A pioneer 1630 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers study by Yang et al. [133] showed how MDS and considered to be ‘epigenetic malignancies’. When DNA AML patient samples and leukemia cell lines treated methylation anomalies arise during hematopoiesis, they with demethylating agents caused upregulation (in- often trigger the initiation and progression of hemato- creased expression) of the immunosuppressive ligands logical cancers. However, DNA methylation aberrations PD-L1, PD-L2, and their (immunosuppressive) check- do not function exclusively, often acting in synergy with point receptor PD-1, including in vitro PD-1 demethy- other epigenetic lesions, showing that the epigenetic reg- lation and subsequent reactivation. The authors ulation of cell development and fate is a very complex proposed that this indicated a possible mechanism for and intricate process. In this review, we have summa- resistance to HMAs. They also noted that MDS/AML rized the current knowledge about the main DNA bone marrow blasts were positive for PD-L1, whereas methylation remodelers (DNMTs, TETs, and IDHs), the stroma/nonblast cellular compartment (i.e., the their relationship with other epigenetic effectors, and TME) was positive for PD-1, leading them to specu- the possible treatments and clinical applications. HMAs late about the sensitivity of these cell lines to ICIs. have emerged as promising agents to treat several Analogously, Srivastava et al. [134] showed that HMA hematological neoplasms, due to their capacity to therapies can reactivate testicular cancer antigens and reverse aberrant methylation patterns. However, induce robust immune recognition by T cells in AML because HMAs inhibit DNMTs, they are highly cyto- patients. Overall, these and other results pointed toxic at high doses, causing an indiscriminate global toward the possible use of ICIs to improve outcomes loss of methylation. Indeed, inclusion of the HMA in leukemia and lymphoma, either alone or in combi- agents AZA and DAC in the RNA or DNA at high nation with HMAs, since HMAs have an affinity for doses produces severe DNA damage and inhibits pro- immunosensitive neoplastic cells, and therefore confer tein synthesis, making these compounds highly toxic to on them a greater sensitivity to ICIs. This link between patients. Therefore, optimal, subcytotoxic doses must immune regulation and HMAs has been explored by be applied. Survival data [103,104] currently suggest several other authors [135,136] and is the subject of that AZA may be clinically superior to DAC, but the current research efforts [137]. For instance, it has been lack of systemic clinical trials comparing the two recently reported that apart from modulating the regimes makes this comparison difficult and the results tumor’s immune responses, HMAs might act directly unclear [140]. However, the observed higher efficacy of on exhausted T cells to reverse the onset of exhaustion AZA might be due precisely to its ability to incorporate and restore their cytotoxicity activity [138]. At the time into the RNA [141]. The development of specific cat- of writing, several clinical trials involving both HMAs alytic inhibitors of individual DNMT enzymes, or tar- and ICIs are already in progress (Table 2). Another geting specific DNMT-containing complexes, is of promising combination therapy is the joint use of fundamental importance to improving the clinical effi- HMAs and immunomodulatory drugs (IMiDs). The cacy of current HMAs [142]. We also mentioned the latter are a set of drugs that have initially been highly emerging evidence supporting the use of combinations success in the treatment of MM. They have antiprolif- of epigenetic drugs, and noted that these combinations erative effects and costimulate T and NK cells, may be important in the treatment of hematological enhancing anti-MM immune activity in vitro [139]. malignancies (Table 2). As mentioned above, clinical While their in vivo effects are not yet clear, two potent trials combining HMAs and ICIs are already underway IMiDs, lenalidomide and pomalidomide, are currently (Table 2). Their merit is supported by the fact that under active investigation, for the treatment not only DNA methylation abnormalities may actually have pro- of MM, but also of other hematological neoplasms. found effects on the cytotoxicity of certain immune sys- tem cellular subtypes (e.g., by inducing the T-exhausted phenotype). While HMAs and HDACis alone may have 3. Conclusions and perspectives proimmunogenic effects, combinations of them have DNA methylation is a highly dynamic process that aids also been shown to increase the infiltration and activa- normal hematopoiesis control. The rapid progress in tion of effector immune cells while reducing the infiltra- high-throughput sequencing technologies (WES, WGS, tion of immunosuppressive cells [143]. Given the EPIC methylation arrays, and WGBS) over the last increasing evidence linking HMA action and DNA decade has enabled researchers to identify many genetic damage, and the frequent co-occurrence of DNMT and epigenetic aberrations in different types of blood mutations and other genetic lesions, we speculate that cancers. These diseases also usually present a low (or synthetically lethal approaches might be an option for very low, in the case of AML, for example) somatic treating certain types of blood cancers. Recent studies mutation rate in their genomes, and are often have demonstrated a synergistic interaction between Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1631 Aberrant DNA methylation in hematological cancers P. Blecua et al. HMAs and PARP inhibitors [144,145]. Valdez et al. 2 Barthel PD (2018) Metazoan microRNAs. Cell Rev [146] have recently shown that the combined use of 173, 20–51. HMA, HDACi, and PARPi resulted in extensive DNA 3 Feinberg AP, Koldobskiy MA & G€ ond€or A (2016) damage, double-strand breaks (DSBs), and apoptosis. Epigenetic modulators, modifiers and mediators in Furthermore, evidence is accumulating that links the cancer aetiology and progression. Nat Rev Genet 17, microsatellite instability (MSI) phenotype of hematolog- 284–299. ical tumors to the disruption of the mismatch repair 4 Ntziachristos P, Abdel-Wahab O & Aifantis I (2016) pathway by promoter hypermethylation of the path- Emerging concepts of epigenetic dysregulation in hematological malignancies. Nat Immunol 17, 1016–1024. way’s own genes [147]. Thus, it is reasonable to specu- 5 Heyn H, Vidal E, Ferreira HJ, Vizoso M, Sayols S, late further that, as observed in solid tumors, the MSI Gomez A, Moran S, Boque-Sastre R, Guil S, phenotype might confer some hematological neoplasms Martinez-Cardus A et al. (2016) Epigenomic analysis with higher rates of ICI responses, through enhance- detects aberrant super-enhancer DNA methylation in ment of neoantigen presentation [148]. In conclusion, human cancer. Genome Biol 17, 11. other epigenetic regulators, such as miRNAs or 6 Tucker MS (1999) The establishment and maintenance lncRNAs, are also attractive putative therapeutic tar- of DNA methylation patterns in mouse somatic cells. gets, since they control several pathways that are dys- Semin Cancer Biol 9, 329–337. regulated in hematological malignancies and are often 7 Kondo YY (2009) Epigenetic cross-talk between DNA linked to DNA methylation in these diseases. Further methylation and histone modifications in human research in this area is likely to yield highly valuable cancers. Med J 50, 455–463. results. 8 Guillamot M, Cimmino L & Aifantis I (2016) The impact of DNA methylation in hematopoietic malignancies. Trends Cancer 2, 70–83. Acknowledgements 9 Langstein J, Milsom M & Lipka DB (2017) Impact of We would like to thank Pere Llinas for his insightful DNA methylation programming on normal and pre- advice regarding DNA methylation and metabolism. leukemic hematopoiesis. Semin Cancer Biol 51, 89–100. We thank CERCA Programme/Generalitat de Catalu- 10 Doulatov S, Notta F, Laurenti E & Dick JE (2012) nya for institutional support. Work at ME laboratory Hematopoiesis: a human perspective. Cell Stem Cell is supported by the Health Department PERIS—pro- 10, 120–136. ject no. SLT/002/16/00374 and AGAUR—project no. 11 Morrison SJ, Uchida N & Weissman IL (1995) The 2017SGR1080 of the Catalan Government (Generalitat biology of hematopoietic stem cells. Annu Rev Cell Dev de Catalunya); Ministerio de Ciencia e Innovaci on Biol 11, 35–71. 12 Busch K, Klapproth K, Barile M, Flossdorf M, (MCI), Agencia Estatal de Investigaci on (AEI) and Holland-Letz T, Schlenner SM, Reth M, H€ ofer T & European Regional Development Fund (ERDF) pro- Rodewald HR (2015) Fundamental properties of ject no. RTI2018-094049-B-I00; the Cellex Foundation; unperturbed haematopoiesis from stem cells in vivo. and “la Caixa” Banking Foundation (LCF/PR/GN18/ Nature 518, 542–546. 51140001). 13 Sawai CM, Babovic S, Upadhaya S, Knapp DJHF, Lavin Y, Lau CM, Goloborodko A, Feng J, Fujisaki Conflict of interest J, Ding L et al. (2016) Hematopoietic stem cells are the major source of multilineage hematopoiesis in ME is a consultant for Ferrer International and adult animals. Immunity 45, 597–609. Quimatryx. The other authors declare that they have 14 Sun J, Ramos A, Chapman B, Johnnidis JB, Le L, Ho no conflict of interest. YJ, Klein A, Hofmann O & Camargo FD (2014) Clonal dynamics of native haematopoiesis. Nature 514, 322–327. Author contributions 15 Yu VWC, Yusuf RZ, Oki T, Wu J, Saez B, Wang X, All authors contributed to writing the manuscript, Cook C, Baryawno N, Ziller MJ, Lee E et al. (2016) read, edited, and approved the final manuscript. Epigenetic memory underlies cell-autonomous heterogeneous behavior of hematopoietic stem cells. Cell 167, 1310–1322.e17. References 16 Cabezas-Wallscheid N, Klimmeck D, Hansson J, 1 Spektor R, Tippens ND, Mimoso CA & Soloqay PD Lipka DB, Reyes A, Wang Q, Weichenhan D, Lier A, (2019) methyl-ATAC-seq measures DNA methylation von Paleske L, Renders S et al. (2014) Identification of at accessible chromatin. Genome Res 29, 969–977. regulatory networks in HSCs and their immediate 1632 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers progeny via integrated proteome, transcriptome, and 28 Cheng J, Guo S, Chen S, Mastriano SJ, Liu C, DNA methylome analysis. Cell Stem Cell 15, 507–522. D’Alessio AC, Hysolli E, Guo Y, Yao H, Megyola 17 Lipka DB, Wang Q, Cabezas-Wallscheid N, Klimmeck CM et al. (2013) An extensive network of TET2- D, Weichenhan D, Herrmann C, Lier A, Brocks D, targeting MicroRNAs regulates malignant von Paleske L, Renders S et al. (2014) Identification of hematopoiesis. Cell Rep 5, 471–481. DNA methylation changes at cis-regulatory elements 29 Celik H, Mallaney C, Kothari A, Ostrander EL, during early steps of HSC differentiation using Eultgen E, Martens A, Miller CA, Hundal J, Klco JM tagmentation-based whole genome bisulfite sequencing. & Challen GA (2015) Enforced differentiation of Cell Cycle 13, 3476–3487. Dnmt3a-null bone marrow leads to failure with c-Kit 18 Bock C, Beerman I, Lien WH, Smith ZD, Gu H, mutations driving leukemic transformation. Blood 125, Boyle P, Gnirke A, Fuchs E, Rossi DJ & Meissner A 619–628. (2012) DNA methylation dynamics during in vivo 30 Mayle A, Yang L, Rodriguez B, Zhou T, Chang E, differentiation of blood and skin stem cells. Mol Cell Curry CV, Challen GA, Li W, Wheeler D, Rebel VI 47, 633–647. et al. (2015) Dnmt3a loss predisposes murine 19 Farlik M, Halbritter F, M€ uller F, Choudry FA, Ebert hematopoietic stem cells to malignant transformation. P, Klughammer J, Farrow S, Santoro A, Ciaurro V, Blood 125, 629–638. Mathur A et al. (2016) DNA methylation dynamics of 31 The Cancer Genome Atlas Research Network (2013) human hematopoietic stem cell differentiation. Stem Genomic and epigenomic landscapes of adult de novo Cell 19, 808–822. acute myeloid leukemia. N Engl J Med 368, 2059– 20 Lister R, Pelizzola M, Dowen RH, Hawkins RD, Hon 2074. G, Tonti-Filippini J, Nery JR, Lee L, Ye Z, Ngo QM 32 Challen GA, Sun D, Jeong M, Luo M, Jelinek J, Berg et al. (2009) Human DNA methylomes at base JS, Bock C, Vasanthakumar A, Gu H, Xi Y et al. resolution show widespread epigenomic differences. (2011) Dnmt3a is essential for hematopoietic stem cell Nature 462, 315–322. differentiation. Nat Genet 44, 23–31. 21 Benetatos L & Vartholomatos G (2018) DNA 33 Anad on C, van Tetering G, Ferreira H, Moutinho C, methylation in acute myeloid leukemia and Martınez-Card us A, Villanueva A, Soler M, Heyn H, myelodysplastic syndromes. Cell Mol Life Sci 75, Moran S, Castro de Moura M et al. (2017) Epigenetic 1999–2009. loss of the RNA decapping enzyme NUDT16 mediates 22 Ji H, Ehrlich LI, Seita J, Murakami P, Doi A, Lindau C-MYC activation in T-cell acute lymphoblastic P, Lee H, Aryee MJ, Irizarry RA, Kim K et al. (2010) leukemia. Leukemia 31, 1622–1625. Comprehensive methylome map of lineage 34 Br€oske AM, Vockentanz L, Kharazi S, Huska MR, commitment from haematopoietic progenitors. Nature Mancini E, Scheller M, Kuhl C, Enns A, Prinz M, 467, 338-342. Jaenisch R et al. (2009) DNA methylation protects 23 Majzner RG & Mackall CL (2018) Tumor antigen hematopoietic stem cell multipotency from escape from CAR T-cell therapy. Cancer Discov 8, myeloerythroid restriction. Nat Genet 41, 1207–1215. 1219–1226. 35 Ferreira H, Heyn H, Vizoso M, Moutinho C, Vidal E, 24 Bueno-Costa A, Pi~ neyro D, Soler M, Javierre BM, Gomez A, Martınez-Card us A, Sim o-Riudalbas L, Raurell-Vila H, Subirana-Granes M, Pasquali L, Moran S, Jost E et al. (2016) DNMT3A mutations Martinez-Climent JA & Esteller M (2020) B-cell mediate the epigenetic reactivation of the leukemogenic leukemia transdifferentiation to macrophage involves factor MEIS1 in acute myeloid leukemia. Oncogene 35, reconfiguration of DNA methylation for long-range 3079–3082. regulation. Leukemia 34, 1158–1162. 36 He YF, Li BZ, Li Z, Liu P, Wang Y, Tang Q, Ding J, 25 Genovese G, K€ahler AK, Handsaker RE, Lindberg J, Jia Y, Chen Z, Li L et al. (2011) Tet-mediated Rose SA, Bakhoum SF, Chambert K, Mick E, Neale formation of 5-carboxylcytosine and its excision by BM, Fromer M et al. (2014) Clonal hematopoiesis and TDG in mammalian DNA. Science 333, 1303–1307. blood-cancer risk inferred from blood DNA sequence. 37 Ito S, Shen L, Dai Q, Wu SC, Collins LB, Swenberg N Engl J Med 371, 2477–2487. JA, He C & Zhang Y (2011) Tet proteins can convert 26 Roadmap Epigenomics Consortium (2015) Integrative 5-methylcytosine to 5-formylcytosine and 5- analysis of 111 reference human epigenomes. Nature carboxylcytosine. Science 333, 1300–1303. 518, 317–330. 38 Tahiliani M, Koh KP, Shen Y, Pastor WA, 27 Luo M, Jeong M, Sun D, Park HJ, Rodriguez BA, Bandukwala H, Brudno Y, Agarwal S, Iyer LM, Liu Xia Z, Yang L, Zhang X, Sheng K, Darlington GJ DR, Aravind L et al. (2009) Conversion of 5- et al. (2015) Long non-coding RNAs control methylcytosine to 5-hydroxymethylcytosine in hematopoietic stem cell function. Cell Stem Cell 16, mammalian DNA by MLL partner TET1. Science 324, 426–438. 930–935. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1633 Aberrant DNA methylation in hematological cancers P. Blecua et al. 39 Huang H, Jiang X, Li Z, Li Y, Song CX, He C, Sun IDH1 mutations produce 2-hydroxyglutarate. Nature M, Chen P, Gurbuxani S, Wang J et al. (2013) TET1 462, 739–744. plays an essential oncogenic role in MLL-rearranged 50 Ward PS, Patel J, Wise DR, Abdel-Wahab O, Bennett leukemia. Proc Natl Acad Sci USA 110, 11994–11999. BD, Coller HA, Cross JR, Fantin VR, Hedvat CV, 40 Cimmino L, Dawlaty MM, Ndiaye-Lobry D, Yap YS, Perl AE et al. (2010) The common feature of Bakogianni S, Yu Y, Bhattacharyya S, Shaknovich R, leukemia-associated IDH1 and IDH2 mutations is a Geng H, Lobry C et al. (2015) TET1 is a tumor neomorphic enzyme activity converting a-ketoglutarate suppressor of hematopoietic malignancy. Nat Immunol to 2 -hydroxyglutarate. Cancer Cell 17, 225–234. 16, 653–662. 51 Crispo F, Condelli V, Lepore S, Notarangelo T, 41 Abdel-Wahab O, Mullally A, Hedvat C, Garcia- Sgambato A, Esposito F, Maddalen F & Landriscina Manero G, Patel J, Wadleigh M, Malinge S, Yao J, M (2019) Metabolic dysregulations and epigenetics: a Kilpivaara O, Bhat R et al. (2009) Genetic bidirectional interplay that drives tumor progression. characterization of TET1, TET2, and TET3 alterations Cells 8, 798. in myeloid malignancies. Blood 114, 144–147. 52 Shanmugam M, McBrayer SK & Rosen ST (2009) 42 Delhommeau F, Dupont S, Della Valle V, James C, Targeting the Warburg effect in hematological Trannoy S, Masse A, Kosmider O, Le Couedic JP, malignancies: from PET to therapy. Curr Opin Oncol Robert F, Alberdi A et al. (2009) Mutation in TET2 in 21, 531–536. myeloid cancers. N Engl J Med 360, 2289–2301. 53 Zeng J, Wu W, Wang H & Li X (2019) Serine and 43 Langemeijer SM, Kuiper RP, Berends M, Knops R, one-carbon metabolism, a bridge that links mTOR Aslanyan MG, Massop M, Stevens-Linders E, van signaling and DNA methylation in cancer. Pharmacol Hoogen P, van Kessel AG, Raymakers RA et al. Res 149, 104352. (2009) Acquired mutations in TET2 are common in 54 Rashkovan & Ferrando (2019) Metabolic dependencies myelodysplastic syndromes. Nat Genet 41, 838–842. and vulnerabilities in leukemia. Genes Dev 33, 1460– 44 Quivoron C, Couronne L, Della Valle V, Lopez CK, 1474. Plo I, Wagner-Ballon O, Do Cruzeiro M, 55 Hatzimichael E, Dranitsaris G, Dasoula A, Benetatos Delhommeau F, Arnulf B, Stern MH et al. (2011) L, Stebbing J, Crook T & Bourantas KL (2009) Von TET2 inactivation results in pleiotropic hematopoietic Hippel-Lindau methylation status in patients with abnormalities in mouse and is a recurrent event during multiple myeloma: a potential predictive factor for the human lymphomagenesis. Cancer Cell 20, 25–38. development of bone disease. Clin Lymphoma 45 Rasmussen KD, Guangshuai J, Johansen JV, Pedersen Myeloma 9, 239–242. MT, Rapin N, Bagger FO, Porse BT, Bernard OA, 56 Barve A, Vega A, Shah PP, Ghare S, Casson L, Christensen J & Helin K (2015) Loss of TET2 in Wunderlich M, Siskind LJ & Beverly LJ (2019) hematopoietic cells leads to DNA hypermethylation of Perturbation of methionine/S-adenosylmethionine active enhancers and induction of leukemogenesis. metabolism as a novel vulnerability in MLL Genes Dev 29, 910–922. rearranged leukemia. Cells 8, 1322. 46 Li Z, Cai X, Cai CL, Wang J, Zhang W, Petersen BE, 57 Rau RE, Rodriguez BA, Luo M, Jeong M, Rosen A, Yang FC & Xu M (2011) Deletion of Tet2 in mice Rogers JH, Campbell CT, Daigle SR, Deng L, Song Y leads to dysregulated hematopoietic stem cells and et al. (2016) DOT1L as a therapeutic target for the subsequent development of myeloid malignancies. treatment of DNMT3A-mutant acute myeloid Blood 118, 4509–4518. leukemia. Blood 128, 971–981. 47 Moran-Crusio K, Reavie L, Shih A, Abdel-Wahab O, 58 Medeiros BC, Fathi AT, DiNardo CD, Pollyea DA, Ndiaye-Lobry D, Lobry C, Figueroa ME, Chan SM & Swords R (2017) Isocitrate dehydrogenase Vasanthakumar A, Patel J, Zhao X et al. (2011) Tet2 mutations in myeloid malignancies. Leukemia 31, 272–281. loss leads to increased hematopoietic stem cell self- 59 Anderson NM, Mucka P, Kern JG & Feng H (2018) renewal and myeloid transformation. Cancer Cell 20, The emerging role and targetability of the TCA cycle 11–24. in cancer metabolism. Protein Cell 9, 216–237. 48 Figueroa ME, Abdel-Wahab O, Lu C, Ward PS, Patel 60 Zhang M, Fujiwara K, Che X, Zheng S & Zheng L J, Shih A, Li Y, Bhagwat N, Vasanthakumar A, (2017) DNA methylation in the tumor Fernandez HF et al. (2010) Leukemic IDH1 and IDH2 microenvironment. J Zhejiang Univ Sci B (Biomed and mutations result in a hypermethylation phenotype, Biotechnol) 18, 365–372. disrupt TET2 function, and impair hematopoietic 61 Colegio OR, Chu N-Q, Szabo AL, Chu T, Rhebergen differentiation. Cancer Cell 18, 553–567. AM, Jairam V, Cyrus N, Brokowski CE, Eisenbarth 49 Dang L., White D. W., Gross S., Bennett B. D., SC, Phillips GM et al. (2014) Functional polarization Bittinger M. A., Driggers E. M., Fantin V. R., Jang of tumour-associated macrophages by tumour-derived HG, Jin S, Keenan MC et al. (2009) Cancer-associated lactic acid. Nature 513, 559–563. 1634 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers 62 Galletti G, Caligaris-Cappio F & Bertilaccio MT in acute lymphoblastic leukemia. Epigenomics 9, (2016) B cells and macrophages pursue a common 635–645. path toward the development and progression of 76 James A, Schroeder MP, Neumann M, Bastian L, chronic lymphocytic leukemia. Leukemia 30, Eckert C, G€ okbuget N, Tanchez JO, Schlee C, 2293–2301. Isaakidis K, Schwartz S et al. (2019) Long non-coding 63 Ke W, Chen C, Song Y, Cai Q, Li J, Tang Y, Han X, RNAs defining major subtypes of B cell precursor Qu W, Chen A, Wang H et al. (2019) Hypoxia acute lymphoblastic leukemia. J Hematol Oncol 12, 8. modifies the polarization of macrophages and their 77 Vasudevan S, Tong Y & Steitz JA (2007) Switching inflammatory microenvironment, and inhibits from repression to activation: microRNAs can up- malignant behavior in cancer cells. Oncol Lett 18, regulate translation. Science 318, 1931–1934. 5871–5878. 78 Esteller M (2011) Non-coding RNAs in human 64 Komohara Y, Niino D, Ohnishi K, Ohshima K & disease. Nat Rev Genet 12, 861–874. Takeya M (2015) Role of tumor-associated 79 Blum W, Garzon R, Klisovic RB, Schwind S, Walker macrophages in hematological malignancies. Pathol Int A, Geyer S, Liu S, Havelange V, Becker H, Schaaf L 65, 170–176. et al. (2010) Clinical response and miR-29b predictive 65 Kouidhi S, Ben Ayed F & Benammar Elgaaied A significance in older AML patients treated with a 10- (2018) Targeting tumor metabolism: a new challenge day schedule of decitabine. Proc Natl Acad Sci USA to improve immunotherapy. Front Immunol 9, 353. 107, 7473–7478. 66 Feng Y, Li X, Cassady K, Zou Z & Zhang L (2019) 80 Deng R, Shen N, Yang Y, Yu H, Xu S, Yang YW, TET2 function in hematopoietic malignancies, immune Liu S, Meguellati K & Yan F (2018) Targeting regulation, and DNA repair. Front Oncol 9, 210–219. epigenetic pathway with gold nanoparticles for acute 67 Lee M, Zeng H, Li J, Han W, Sun D & Huang Y myeloid leukemia therapy. Biomaterials 167, 80–90. (2018) Disruption of TET2 dioxygenase enhances 81 Bi L, Zhou B, Li H, He L, Wang C, Wang Z, Zhu L, antitumor efficiency in CD8+ tumor infiltrating Chen M & Gao S (2018) A novel miR-375-HOXB3- lymphocytes. Blood 132 (Suppl 1), 860. CDCA3/ DNMT3B regulatory circuitry contributes to 68 Garzon R, Volinia S, Papaioannou D, Nicolet D, leukemogenesis in acute myeloid leukemia. BMC Kohlschmidt J, Yan PS, Mr ozek K, Bucci D, Carroll Cancer 18, 182. AJ, Baer MR et al. (2014) Expression and prognostic 82 Chi P, Allis CD & Wang GG (2010) Covalent histone impact of lncRNAs in acute myeloid leukemia. Proc modifications–miswritten, misinterpreted and mis- Natl Acad Sci USA 111, 18679–18684. erased in human cancers. Nat Rev Cancer 10, 457–469. 69 Chen S, Liang H, Yang H, Zhou K, Xu L, Liu J, Lai 83 Roth SY & Allis CD (1996) Histone acetylation and B, Song L, Luo H, Peng J et al. (2017) Long non- chromatin assembly: a single escort, multiple dances? coding RNAs: the novel diagnostic biomarkers for Cell 87, 5–8. leukemia. Environ Toxicol Pharmacol 55, 81–86. 84 Jenuwein T & Allis CD (2001) Translating the histone 70 Ward M, McEwan C, Mills JD & Janitz M (2015) code. Science 293, 1074–100. Conservation and tissue-specific transcription patterns 85 Voigt P, Tee WW & Reinberg D (2013) A double take of long noncoding RNAs. J Hum Transcr 1, 2–9. on bivalent promoters. Genes Dev 27, 1318–1338. 71 Gourvest M, Brousset P & Bousquet M. (2019) Long 86 Schlesinger Y, Straussman R, Keshet I, Farkash S, noncoding RNAs in acute myeloid leukemia: Hecht M, Zimmerman J, Eden E, Yakhini Z et al. functional characterization and clinical relevance. (2007) Polycomb-mediated methylation on Lys27 of Cancers (Basel) 11, 1638. histone H3 pre-marks genes for de novo methylation in 72 Liu Y, Cheng Z, Pang Y, Cui L, Qian T, Quan L, cancer. Nat Genet 39, 232–236. Zhao H, Shi J, Ke X & Fu L (2019) Role of 87 Okitsu CY & Hsieh CL (2007) DNA methylation microRNAs, circRNAs and long noncoding RNAs in dictates histone H3K4 methylation. Mol Cell Biol 27, acute myeloid leukemia. J Hematol Oncol 12, 51. 2746–2757. 73 Zhou Y, Zhang X & Klibanski A (2012) MEG3 88 Scalea S, Maresca C, Catalanotto C, Marino R, noncoding RNA: a tumor suppressor. J Mol Cogoni C, Reale A, Zampieri M & Zardo G (2020) Endocrinol. 48, R45–R53. Modifications of H3K4 methylation levels are 74 Lyu Y, Lou J, Yang Y, Feng J, Hao Y, Huang S, Yin associated with DNA hypermethylation in acute L, Xu J, Huang D, Ma B et al. (2017) Dysfunction of myeloid leukemia. FEBS J. 287, 1155–1175. the WT1-MEG3 signaling promotes AML 89 Di Croce L & Helin K (2013) Transcriptional leukemogenesis via p53-dependent and -independent regulation by Polycomb group proteins. Nat Struct pathways. Leukemia 31, 2543–2551. Mol Biol 20, 1147–1155. 75 Arthur G, Almamun M & Taylor K (2017) 90 Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Hypermethylation of antisense long noncoding RNAs Didelot C, Morey L, Van Eynde A, Bernard D, Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1635 Aberrant DNA methylation in hematological cancers P. Blecua et al. Vanderwinden JM et al. (2006) The Polycomb group the treatment of higher-risk myelodysplastic protein EZH2 directly controls DNA methylation. syndromes: a randomised, open-label, phase III study. Nature 439, 871–874. Lancet Oncology 10(3), 223–32. 91 Wu X, Gong Y, Yue J, Qiang B, Yuan J & Peng X 103 Fenaux P, Mufti GJ, Hellstrom-Lindberg E, Santini V, (2008) Cooperation between EZH2, NSPc1-mediated Gattermann N, Germing U, Sanz G, List AF, Gore S, histone H2A ubiquitination and Dnmt1 in HOX gene Seymour JF et al. (2010) Azacitidine prolongs overall silencing. Nucleic Acids Res 36, 3590–3599. survival compared with conventional care regimens in 92 Ernst T, Chase A, Score J, Hidalgo-Curtis CE, Bryant elderly patients with low bone marrow blast count C, Jones AV, Waghorn K, Zoi K, Ross FM, Reiter A acute myeloid leukemia. J Clin Oncol 28, 562–569. et al. (2010) Inactivating mutations of the histone 104 Kantarjian HM, Thomas XG, Dmoszynska A, methyltransferase gene EZH2 in myeloid disorders. Wierzbowska A, Mazur G, Mayer J, Gau JP, Chou Nat Genet 42, 722–726. WC, Buckstein R, Cermak J et al. (2012) Multicenter, 93 Shi J, Wang E, Zuber J, Rappaport A, Taylor M, randomized, open-label, phase III trial of decitabine Johns C, Lowe SW & Vakoc CR. (2013) The versus patient choice, with physician advice, of either Polycomb complex PRC2 supports aberrant self- supportive care or low-dose cytarabine for the renewal in a mouse model of MLL-AF9;NrasG12D treatment of older patients with newly diagnosed acute acute myeloid leukemia. Oncogene 32, 930–938. myeloid leukemia. J Clin Oncol 30, 2670–2677. 94 Neff T, Sinha AU, Kluck MJ, Zhu N, Khattab MH, 105 Duong VH, Komrokji RS & List AF (2014) Update Stein L, Xie H, Orkin SH & Armstrong SA (2012) on the pharmacotherapy for myelodysplastic Polycomb repressive complex 2 is required for MLL- syndromes. Expert Opin Pharmacother 15, 1811–1825. AF9 leukemia. PNAS 109, 5028–5033. 106 Michalak EM, Burr ML, Bannister AJ&Dawson MA 95 Dimopoulos K, Søgaard HA, Fibiger M-PH, Sj€ o L, (2019) The roles of DNA, RNA and histone Christensen J, Sommer KL, Asmar F, Hermansen methylation in ageing and cancer. Nat Rev Mol Cell NEU, O’Connel C, Gimsing P et al. (2018) Dual Biol 20, 573–589. inhibition of DNMTs and EZH2 can overcome both 107 Nieto M, Demolis P, Behanzin E, Moreau A, Hudson intrinsic and acquired resistance of myeloma cells to I, Flores B, Stemplewski H, Salmonson T, IMiDs in a cereblon-independent manner. Mol Oncol Gisselbrecht C, Bowen D et al. (2016) The European 12, 180–195. medicines agency review of decitabine (Dacogen) for 96 Burnett AK, Milligan D, Prentice AG, Goldstone AH, the treatment of adult patients with acute myeloid McMullin MF, Hills RK & Wheatley K (2007) A leukemia: summary of the scientific assessment of the comparison of low-dose cytarabine and hydroxyurea Committee for Medicinal Products for human use. with or without all-trans retinoic acid for acute Oncologist 21, 692–700. myeloid leukemia and high-risk myelodysplastic 108 Aimiuwu J, Wang H, Chen P, Xie Z, Wang J, Liu S, syndrome in patients not considered fit for intensive Klisovic R, Mims A, Blum W, Marcucci G et al. treatment. Cancer 109, 1114–1124. (2012) RNA-dependent inhibition of ribonucleotide 97 Contieri B, Duarte BKL & Lazarini M (2020) Updates reductase is a major pathway for 5-azacytidine activity on DNA methylation modifiers in acute myeloid in acute myeloid leukemia. Blood 119, 5229–5238. leukemia. Ann Hematol 99, 693–701. 109 Prebet T, Sun Z, Figueroa ME, Ketterling R, Melnick 98 Sorm F & Vesely J (1964) The activity of a new A, Greenberg PL, Herman J, Juckett M, Smith MR, antimetabolite, 5-Azacytidine, against lymphoid Malick L et al. (2014) Prolonged administration of Leukaemia in Ak mice. Neoplasma 11, 123–130. azacytidine with or without entinostat for 99 Momparler RL (2005) Pharmacology of 5-Aza-20 - myelodysplastic syndrome and acute myeloid leukemia deoxycytidine (decitabine). Semin Hematol 42 (3 Suppl with myelodysplasia-related changes: results of the US 2), S9–S16. Leukemia Intergroup trial E1905. J Clin Oncol 32, 100 Christman JK (2002) 5-Azacytidine and 5-aza-20- 1242–1248. deoxycytidine as inhibitors of DNA methylation: 110 Herman JG & Baylin SB (2003) Gene silencing in mechanistic studies and their implications for cancer cancer in association with promoter hypermethylation. therapy. Oncogene 21, 5483–5495. N Engl J Med 349, 2042–2054. 101 Pleyer K & Greil R (2015) Digging deep into “dirty” 111 Perez-Salvia M, Aldaba E, Vara Y, Fabre M, Ferrer drugs – modulation of the methylation machinery. C, Masdeu C, Zubia A, San Sebastian E, Otaegui D, Drug Metab Rev 47, 252–279. Llinas-Arias P et al. (2018) In vitro and in vivo activity 102 Fenaux P, Mufti GJ, Hellstrom-Lindberg E, Santini V, of a new small-molecule inhibitor of HDAC6 in Finelli C, Giagounidis A, Schoch R, Gattermann N, mantle cell lymphoma. Haematologica 103, e537–e540. Sanz G, List A et al. (2009) Efficacy of azacitidine 112 Bogenberger JM, Kornblau SM, Pierceall WE, Lena compared with that of conventional care regimens in R, Chow D, Shi C-X, Mantei J, Ahmann G, Gonzales 1636 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers IM, Choudhary A et al. (2014) BCL-2 family proteins 124 Aggerholm A, Holm MS, Guldberg P, Olesen LH & as 5-Azacytidine-sensitizing targets and determinants Hokland P (2006) Promoter hypermethylation of of response in myeloid malignancies. Leukemia 28, p15INK4B, HIC1, CDH1, and ER is frequent in 1657–1665. myelodysplastic syndrome and predicts poor prognosis 113 Griffiths EA, Choy G, Redkar S, Taverna P, Azab M in early-stage patients. Eur J Haematol 76, 23–32. & Karpf AR (2013) SGI-110: DNA methyltransferase 125 Maegawa S, Gough SM, Watanabe-Okochi N, Lu Y, inhibitor oncolytic. Drugs Future 38, 535–543. Zhang N, Castoro RJ, Estecio MRH, Jelinek J, Liang 114 Kosmider O, Gelsi-Boyer V, Ciudad M, Racoeur C, S, Kitamura T et al. (2014) Age-related epigenetic drift Jooste V, Vey N, Quesnel B, Fenaux P, Bastie J-N, in the pathogenesis of MDS and AML. Genome Res Beyne-Rauzy O et al. (2009) TET2 gene mutation is a 24, 580–591. frequent and adverse event in chronic myelomonocytic 126 Cross M, Bach E, Tran T, Krahl R, Jaekel N, leukemia. Haematologica 94, 1676–1681. Niederwieser D, Junghanss C, Maschmeyer G & Al- 115 Shih AH, Abdel-Wahab O, Patel JP & Levine RL Ali HK (2013) Pretreatment long interspersed element (2012) The role of mutations in epigenetic regulators (LINE)-1 methylation levels, not early in myeloid malignancies. Nat Rev Cancer 12, 599–612. hypomethylation under treatment, predict 116 Yamazaki J, Estecio MR, Lu Y, Long Hai, Malouf hematological response to azacitidine in elderly GG, Graber D, Huo Y, Ramagli L, Liang S, patients with acute myeloid leukemia. Onco Targets Kornblau SM et al. (2013) The epigenome of AML Ther 6, 741–748. stem and progenitor cells. Epigenetics 8, 92–104. 127 Treppendahl MB, Kristensen LS & Gronbaek K 117 Asmar F, Punj V, Christensen J, Pedersen MT, (2014) Predicting response to epigenetic therapy. J Clin Pedersen A, Nielsen AB, Hother C, Ralfkiaer U, Invest 124, 47–55. Brown P, Ralfkiaer E et al. (2013) Genome-wide 128 Jabbour E, Garcia-Manero G, Batty N, Shan J, profiling identifies a DNA methylation signature that O’Brien S, Cortes J, Ravandi F, Issa J-P & Kantarjian associates with TET2 mutations in diffuse large B-cell H (2010) Outcome of patients with myelodysplastic lymphoma. Haematologica 98, 1912–1920. syndrome after failure of decitabine therapy. Cancer 118 Couronne L, Bastard C & Bernard O (2012) A. TET2 116, 3830–3834. and DNMT3A mutations in human T-cell lymphoma. 129 Kadia TM, Jabbour E & Kantarjian H (2011) Failure N Engl J Med 366, 95–96. of hypomethylating agent-based therapy in 119 Delarue R, Dupuis J, Sujobert P, Barbieux S, Marcßais myelodysplastic syndromes. Semin Oncol 38, 682–692. A, Tournilhac O, Lachenal F, Cheminant M, Sarkozy 130 Stahl M, DeVeaux M, Montesinos P, Itzykson R, C, Fataccioli V et al. (2016) Treatment with Ritchie EK, Sekeres MA, Barnard JD, Podoltsev NA, hypomethylating agent 5-azacytidine induces sustained Brunner AM, Komrokji RS et al. (2018) response in angioimmunoblastic T cell lymphomas. Hypomethylating agents in relapsed and refractory Blood 128, 4164. AML: outcomes and their predictors in a large 120 Dominguez PM, Ghamlouch H, Rosikiewicz W, international patient cohort. Blood Adv 2, 923–932. Kumar P, Beguelin W, Fontan L, Rivas MA, 131 Bontkes HJ, Ruben JM, Alhan C, Westers TM, Pawlikowska P, Armand M, Mouly E et al. (2018) Ossenkoppele GJ & van de Loosdrecht AA (2012) TET2 deficiency causes germinal center hyperplasia, Azacitidine differentially affects CD4(pos) T-cell impairs plasma cell differentiation, and promotes B- polarization in vitro and in vivo in high risk cell lymphomagenesis. Cancer Discov 8, 1632–1653. myelodysplastic syndromes. Leuk Res 36, 921–930. 121 Chua GNL, Wassarman KL, Sun H, Alp JA, Jarczyk 132 Clozel T, Yang S, Elstrom RL, Tam W, Martin P, EI, Kuzio NJ, Bennett MJ, Malachowsky BG, Kruse Kormaksson M, Banerjee S, Vasanthakumar A, M & Kennedy AJ (2019) Cytosine-based TET enzyme Culjkovic B, Scott DW et al. (2013) Mechanism-based inhibitors. ACS Med Chem Lett 10, 180–185. epigenetic chemosensitization therapy of diffuse large 122 Jiang X, Chao H, Ferchen K, Nie J, Cui X, Chen CH, B-cell lymphoma. Cancer Discov 3, 1002–1019. Cheng L, Zuo Z, Seibel W, He C et al. (2017) 133 Yang H, Bueso-Ramos C, DiNardo C, Estecio MR, Targeted inhibition of STAT/TET1 axis as a Davanlou M, Geng Q-r, Fang Z, Nguyen M, Pierce S, therapeutic strategy for acute myeloid leukemia. Nat Wei Y et al. (2014) Expression of PD-L1, PD-L2, PD- Commun 8, 1–13. 1 and CTLA4 in myelodysplastic syndromes is 123 Li C, Dong L, Rui S, Bi Y, Qing Y, Deng X, Zhou Y, enhanced by treatment with hypomethylating agents. Hu C, Yu M, Huang H et al. (2020) Leukemia 28, 1280–1288. Homoharringtonine exhibits potent anti-tumor effect 134 Srivastava P, Paluch BE, Matsukazi J, James SR, and modulates DNA epigenome in acute myeloid Collamat-Lai G, Blagitko-Dorfs N, Ford LAnn, leukemia by targeting SP1/TET1/5hmC. Naqash R, L€ ubbert M, Karpf AR et al. (2016) Haematologica 105, 148–160. Induction of cancer testis antigen expression in Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1637 Aberrant DNA methylation in hematological cancers P. Blecua et al. circulating acute myeloid leukemia blasts following 146 Valdez BC, Li Y, Murray D, Liu Y, Nieto Y, hypomethylating agent monotherapy. Oncotarget 7, Champlin RE & Andersson BS (2018) Combination of 12840–12856. a hypomethylating agent and inhibitors of PARP and 135 Li H, Chiappinelli KB, Guzzetta AA, Easwaran H, HDAC traps PARP1 and DNMT1 to chromatin, Yen R-WC, Vatapalli R, Topper MJ, Luo J, Connolly acetylates DNA repair proteins, down-regulates RM, Azad NS et al. (2014) Immune regulation by low NuRD and induces apoptosis in human leukemia and doses of the DNA methyltransferase inhibitor 5- lymphoma cells. Oncotarget 9, 3908–3921. azacitidine in common human epithelial cancers. 147 Bhattacharya P & Patel T (2018) Microsatellite Oncotarget 5, 587–598. instability and promoter hypermethylation of DNA 136 Roulois D, Yau HL & De Carvalho DD (2015) repair genes in hematologic malignancies: a Pharmacological DNA demethylation: implications for forthcoming direction toward diagnostics. Hematology cancer immunotherapy. Oncoimmunology 5, e1090077. 23, 77–82. 137 Yau HL, Ettayedi I & De Carvalho DD (2019) The 148 Mandal R, Samstein RM, Lee KV, Havel JJ, Wang cancer epigenome: exploiting its vulnerabilities for H, Krishna C, Sabio EY, Makarov V, Kuo F, Blecua immunotherapy. Trends Cell Biol 29, 31–43. P et al. (2019) Genetic diversity of tumors with 138 Ghoneim HE, Fan Y, Abdelsamed HA, Dash P, mismatch repair deficiency influences anti-PD-1 Dogra P, Carter R, Awad W, Neale G, Thomas PG & immunotherapy response. Science 364, 485–491. Youngblood B (2017) De novo epigenetic programs 149 Dolnik A, Engelmann JC, Scharfenberger-Schmeer M, inhibit PD- 1 blockade-mediated T cell rejuvenation. Mauch J, Kelkenberg-Schade S, Haldemann B, Fries Cell 170, 142–157.e19. T, Kr€ onke J, K€ uhn MWM, Paschka P et al. (2012) 139 Quach H, Ritchie D, Stewart AK, Neeson P, Harrison Commonly altered genomic regions in acute myeloid S, Smyth MJ & Prince HM et al. (2014) Mechanism of leukemia are enriched for somatic mutations involved action of immunomodulatory drugs (IMiDS) in in chromatin remodeling and splicing. Blood 120, multiple myeloma. Leukemia 24, 22–32. 83–92. 140 Bohl SR, Bullinger L & Ruecker FG (2018) Epigenetic 150 Ley TJ,Miller C, Ding L, Raphael BJ, Mungall AJ, therapy: azacytidine and decitabine in acute myeloid Robertson A, Hoadley K, Triche TJ Jr, Laird PW, leukemia. Expert Rev Hematol 11, 361–371. Baty JD et al. (2013) Genomic and epigenomic 141 Santini V (2009) Azacitidine: activity and efficacy as landscapes of adult de novo acute myeloid leukemia. N an epigenetic treatment of myelodysplastic syndromes. Engl J Med 368, 2059–2074. Expert Rev Hematol 2, 121–127. 151 Peters SL, Hlady RA, Opavska J, Klinkebiel D, 142 Pappalardi MB, Cockerill M, Handler JL, Stowel A, Novakova S, Smith LM, Lewis RE, Karpf AR, Keenan K, Sherk CS, Minthorn EA, McHugh CF, Simpson MA, Wu L et al. (2013) Essential role for Burt C, Wong K et al. (2018) Abstract 2994: discovery Dnmt1 in the prevention and maintenance of MYC- of selective, noncovalent small molecule inhibitors of induced T-cell lymphomas. Mol Cell Biol 33, 4321– DNMT1 as an alternative to traditional DNA 4333. hypomethylating agents. Cancer Res 78(Suppl. 13), 152 Yamashita Y, Yuan J, Suetake I, Suzuki H, Ishikawa 2994. Y, Choi Y l, Ueno T, Soda M, Hamada T, Haruta H 143 Stone ML, Chiappinelli KB, Li H, Murphy LM, et al. (2010) Array-based genomic resequencing of Travers ME, Topper MJ, Mathios D, Lim M, Shih I- human leukemia. Oncogene 29, 3723–3731. M, Wang T-L et al. (2017) Epigenetic therapy 153 Yan XJ, Xu Jie, Gu Z-H, Pan C-M, Lu G, Shen Y, activates type I interferon signaling in murine ovarian Shi J-Y, Zhu Y-M, Tang L, Zhang X-W et al. (2011) cancer to reduce immuno-suppression and tumor Exome sequencing identifies somatic mutations of burden. Proc Natl Acad Sci USA 114, E10981–E10990. DNA methyltransferase gene DNMT3A in acute 144 Orta ML, Hoglund A, Calderon-Monta~ no JM, monocytic leukemia. Nat Genet 43, 309–317. Domınguez I, Burgos-Mor on E, Visnes T, Pastor N, 154 Yang L, Rau R & Goodell MA (2015) DNMT3A in Str€om C, L opez-lazaro M & Helleday T (2014) The haematological malignancies. Nat Rev Cancer 15, PARP inhibitor Olaparib disrupts base excision repair 152–165. of 5-aza-2’-deoxycytidine lesions. Nucleic Acids Res 42, 155 Walter MJ, Ding L, Shen D, Shao J, Grillot M, 9108–9120. McLellan M, Fulton R, Schmidt H, Kalicki-Veizer J, 145 Palii SS, Van Emburgh BO, Sankpal UT, Brown O’Laughlin M et al. (2011) Recurrent DNMT3A KD& Robertson KD (2008) DNA methylation mutations in patients with myelodysplastic syndromes. inhibitor 5-Aza-2’-deoxycytidine induces reversible Leukemia 25, 1153–1158. genome-wide DNA damage that is distinctly 156 Gore AV & Weinstein BM. (2016) DNA methylation influenced by DNA methyltransferases 1 and 3B. Mol in hematopoietic development and disease. Exp Cell Biol 28, 752–771. Hematol 44, 783–790. 1638 Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. P. Blecua et al. Aberrant DNA methylation in hematological cancers 157 Herold T, Metzeler KH, Vosberg S, Hartmann L, relapsed or refractory acute myeloid leukemia: a phase Jurinovic V, Opatz S, Konstandin NP, Schneider S, I/II study. Haematologica 99, 54–59. Zellmeier E, Ksienzyk B et al. (2017) Acute myeloid 164 Uy GL, Duncavage EJ, Chang GS, Jacoby MA, leukemia with del(9q) is characterized by frequent Miller CA, Shao J, Heath S, Elliott K, Reineck T, mutations of NPM1, DNMT3A, WT1 and low expression Fulton RS et al. (2017) Dynamic changes in the clonal of TLE4. Genes Chromosomes Cancer 56, 75–86. structure of MDS and AML in response to epigenetic 158 Yeh CH, Bai XT, Moles R, Ratner L, Waldmann TA, therapy. Leukemia 31, 872–881. Watanabe T & Nicot C (2016) Mutation of epigenetic 165 Kirschbaum M, Gojo I, Goldberg SL, Bredeson C, regulators TET2 and MLL3 in patients with HTLV-I- Kujawski LA, Yang A, Marks P, Frankel P, Sun X, induced acute adult T-cell leukemia. Mol Cancer 15, Tosolini A et al. (2014) A phase 1 clinical trial of 1–7. vorinostat in combination with decitabine in patients 159 De Keersmaecker K, Atak Z, Li N, Vicente C, with acute myeloid leukaemia or myelodysplastic Patchett S, Girardi T, Gianfelici V, Geerdens E, syndrome. Br J Haematol 167, 185–193. Clappier E, Porcu M et al. (2013) Exome sequencing 166 Burke MJ, Lamba JK, Pounds S, Cao X, Ghodke- identifies mutation in CNOT3 and ribosomal genes Puranik Y, Lindgren BR, Weigel BJ, Verneris MR & RPL5 and RPL10 in T-cell acute lymphoblastic Miller J (2014) A therapeutic trial of decitabine and leukemia. Nat Genet 45, 186–190. vorinostat in combination with chemotherapy for 160 DiNardo CD, Jabbour E, Ravandi F, Takahashi K, relapsed/refractory acute lymphoblastic leukemia. Am Daver N, Routbort M, Patel KP, Brandt M, Pierce S, J Hematol 89, 889–895. Kantarjian H et al. (2016) IDH1 and IDH2 mutations 167 Burke MJ,Kostadinov R, Sposto R, Gore L, Kelley in myelodysplastic syndromes and role in disease SM, Rabik C, Trepel JB, Lee M-J, Yuno A, Lee S progression. Leukemia 30, 980–984. et al. (2020) Decitabine and vorinostat with 161 Prebet T, Sun Z, Ketterling RP, Zeidan A, Greenberg chemotherapy in relapsed pediatric acute P, Herman J, Juckett M, Smith MR, Malick L, lymphoblastic leukemia: A TACL pilot study. Clin Paietta E et al. (2016) Azacitidine with or without Cancer Res 26, 2297–2307. entinostat for the treatment of therapy-related myeloid 168 Hunault-Berger M, Maillard N, Himberlin C, Recher neoplasm: further results of the E1905 North C, Schmidt-Tanguy A, Choufi B, Bonmati C, Carre American Leukemia Intergroup study. Br J Haematol M, Couturier M-A, Daguindau E et al. (2017) 172, 384–391. Maintenance therapy with alternating azacitidine and 162 Garcia-Manero G, Abaza Y, Takahashi K, Medeiros lenalidomide in elderly fit patients with poor prognosis BC, Arellano M, Khaled SK, Patnaik M, Odenike O, acute myeloid leukemia: a phase II multicentre FILO Sayar H, Tummala M et al. (2019) Pracinostat plus trial. Blood Cancer 7, e568. azacitidine in older patients with newly diagnosed 169 Stevens B, Winters A, Gutman JA, Fullerton A, acute myeloid leukemia: results of a phase 2 study, Hemenway G, Schatz D, Miltgen N, Wei Q, Abbasi Blood Adv 3, 508–518. T, Vali S et al. (2019) Sequential azacitidine and 163 Walter RB, Medeiros B c, Gardner KM, Orlowski lenalidomide for patients with relapsed and refractory KF, Gallegos L, Scott Bl, Hendrie PC & Estey EH acute myeloid leukemia: clinical results and predictive (2014) Gemtuzumab ozogamicin in combination with modeling using computational analysis. Leuk Res 81, vorinostat and azacitidine in older patients with 43–49. Molecular Oncology 14 (2020) 1616–1639 ª 2020 The Authors. Published by FEBS Press and John Wiley & Sons Ltd. 1639