Leukemia (2014) 28,44–49 & 2014 Macmillan Publishers Limited All rights reserved 0887-6924/14 www.nature.com/leu

CONCISE REVIEW EZH2 in normal and malignant hematopoiesis

K Lund1, PD Adams2 and M Copland3

The methyltransferase Enhancer of Zeste Homologue 2 (EZH2), a component of the polycomb group complex, is vital for stem cell development, including hematopoiesis. Its primary function, to deposit the histone mark , promotes transcriptional repression. The activity of EZH2 influences cell fate regulation, namely the balance between self-renewal and differentiation. The contribution of aberrant EZH2 expression to tumorigenesis by directing cells toward a cancer stem cell (CSC) state is increasingly recognized. However, its role in hematological malignancies is complex. Point mutations, resulting in gain-of- function, and inactivating mutations, reported in lymphoma and leukemia, respectively, suggest that EZH2 may serve a dual purpose as an oncogene and tumor-suppressor gene. The reduction of CSC self-renewal via EZH2 inhibition offers a potentially attractive therapeutic approach to counter the aberrant activation found in lymphoma and leukemia. The discovery of small molecules that specifically inhibit EZH2 raises the exciting possibility of exploiting the oncogenic addiction of tumor cells toward this protein. However, interference with the tumor-suppressor role of wild-type EZH2 must be avoided. This review examines the role of EZH2 in normal and malignant hematopoiesis and recent developments in harnessing the therapeutic potential of EZH2 inhibition.

Leukemia (2014) 28, 44–49; doi:10.1038/leu.2013.288 Keywords: EZH2; hematopoiesis; cancer stem cell; lymphoma; EZH2 inhibitors

INTRODUCTION EZH2 AND THE POLYCOMB GROUP COMPLEX Strictly, is defined as a collection of heritable EZH2, a polycomb group (PcG) protein, functions in concert with mechanisms capable of altering genome function other than by other proteins (EED, SUZ12 and RBBP4), which together form the direct alteration of the DNA sequence. Here we use ‘epigenetic’ multi-subunit polycomb repressive complex (PRC)-2. PRC2 acts via more loosely to refer to elements of structure that a number of mechanisms to initiate polycomb-mediated gene control genome function, regardless of whether the control is repression (Figure 1). The mechanisms include direct inhibition of heritable. Epigenetic or chromatin-based regulation contributes to transcriptional machinery via RNA polymerase II and chromatin the control of in normal development and cancer compaction, which inhibits the access and action of states. factors.9 EZH2 has also been suggested to facilitate DNA Histone proteins, which are responsible for packaging and , by acting as a recruitment platform for DNA ordering DNA into the nucleosomal subunits that comprise methyltransferases.10 Importantly, following trimethylation by chromatin fibers, are frequently covalently modified at their the suppressor of variegation 3–9, enhancer of zeste and N-terminal tails. These histone ‘marks’ are defined by the histone, trithorax (SET) domain of EZH2, the histone mark H3K27me3 the modified amino acid, its position in the polypeptide chain and functions as a docking site to recruit a second polycomb complex the nature of the modification, for example, methylation or PRC1, which acts to maintain gene repression via ubiquitination of (for example, trimethylation on 27 of H2AK119.2,11 is abbreviated to H3K27me3). These marks are thought to control DNA transcription and other DNA functions, and have become known as the ‘histone code’ or perhaps, more accurately, ‘histone EZH2 FUNCTION IN STEM CELLS language’.1 EZH2 contributes to the regulation of cell fate decisions, Changes to the structure and/or function result orchestrating gene expression to control the balance between from the activity of histone-modifying enzymes. One of the best self-renewal and differentiation.12 EZH2 has its predominant role described is the histone methyltransferase Enhancer of Zeste in embryonic development and becomes downregulated in some Homologue 2 (EZH2), responsible for trimethylating lysine 27 of adult differentiated tissues.13 However, PcG proteins also have an histone H3 to generate H3K27me3.2 This article will focus on the important role in maintaining the multi-potency of adult stem effect of mutations and other modes of dysregulation of EZH2 in cells in various contexts, including hematopoietic tissues.14 In hematological malignancy. EZH2 is one of several chromatin hematopoietic stem cell (HSC) models, at times of replicative regulatory proteins that have prompted great clinical and stress, or as cells age, the balance shifts toward differentiation with scientific interest as they offer the possibility of new therapeutic potential to ‘exhaust’ the HSC pool.15 However, Ezh2 has been targets in cancer.3–8 shown to stabilize the chromatin structure and maintain

1Department of Epigenetics of Cancer and Aging, Institute of Cancer Sciences, University of Glasgow, Cancer Research UK Beatson Labs, Glasgow, Scotland, UK; 2Department of Epigenetics of Cancer and Aging, Institute of Cancer Sciences, University of Glasgow, Beatson Institute for Cancer Research, Glasgow, Scotland, UK and 3Paul O’Gorman Leukaemia Research Centre, Institute of Cancer Sciences, College of Medical, Veterinary and Life Sciences, University of Glasgow, Gartnavel General Hospital, 1053 Great Western Road, Glasgow, Scotland. Correspondence: Professor M Copland, Paul O’Gorman Leukaemia Research Centre, College of Medical, Veterinary and Life Sciences, University of Glasgow, Gartnavel General Hospital, 1053 Great Western Road, Glasgow, G12 0ZD, Scotland, UK. E-mail: [email protected] Received 7 May 2013; revised 1 September 2013; accepted 20 September 2013; accepted article preview online 7 October 2013; advance online publication, 1 November 2013 EZH2 in normal and malignant hematopoiesis K Lund et al 45


PC1 BMI1 H3K27 H2AK119

Me Ub MeC

Figure 1. The interactions and effects of EZH2 in regulation of transcriptional repression. Polycomb complex 2 (PRC2) exerts methyltransferase activity to H3K27 via the SET domain of the EZH2 subunit. This results in transcriptional repression via various mechanisms, including the recruitment of DNA methyltransferases (DNMT) and PRC1, which ubiquitinates H2AK119.

Figure 2. Regulation of EZH2 in normal and cancer states. HDAC, long-term self-renewal potential of HSCs by switching off ; miR, microRNA; PI3K, phosphatidylinositol 15,16 3-kinase; PTEN, phosphatase and tensin homolog. (a) Homeostatic pro-differentiation genes and promoting a shift toward relationship between c-MYC, EZH2 and the PI3K pathway in proliferation by increasing expression of genes, which facilitate 17 controlled cell growth. (b) In c-MYC driven lymphomagenesis, progression through the cell cycle. c-MYC and EZH2 contribute to each other’s upregulation via the activity of miR-26a and miR-494. EZH2 forms a co- complex with HDAC3 and c-MYC to repress miR-29, which drive cancerous EZH2 AND CANCER STEM CELLS (CSCs) cell growth. The existence of the CSC, first described in the context of acute myeloid leukemia (AML),18 remains controversial, but is recognized as a viable explanation for disease propagation. The CSC hypothesis postulates that tumor tissue comprises a turn suppresses Akt. The inactivation of Akt causes an increase in heterogeneous population of cells and only a small Ezh2-mediated gene repression, including negative autoregulation subpopulation, the CSCs, has the ability to self-renew and give of c-Myc (Figure 2a).28 rise to malignant progeny. In the leukemia stem cell, EZH2 may In certain cancer states, an alteration in the relationship potentially contribute to tumorigenesis by suppressing between EZH2 and c-MYC has been observed.29 In c-MYC driven differentiation genes, thereby directing ‘healthy’ cells toward a lymphomagenesis, c-MYC and EZH2 contribute to each other’s stem cell state.19–22 This makes aberrantly expressed EZH2 an upregulation via the activity of microRNAs. c-MYC represses EZH2 attractive therapeutic target; for example, in breast and pancreatic via miR-26a and EZH2 suppresses the c-MYC-targeting miR-494 to cancer, its inhibition has the potential to deplete the CSC pool.23 maintain high levels of both proteins (Figure 2b). EZH2 forms a PcG target genes identified in embryonic stem cells (ESCs) are co-repressor complex with HDAC3 and c-MYC to repress miR-29, in more likely to be DNA hypermethylated in cancers, as shown in turn facilitating the expression of miR-29 target genes, which drive colon and embryonic carcinoma cell lines.24–26 Perhaps aggressive mantle cell lymphoma growth. Aberrant microRNA accounting for this, is evidence suggesting that EZH2 recruits expression has been shown to regulate EZH2 in other cancer DNA methyltransferases to the promoter region of PcG target contexts; for example, dysregulation of miR-101 and miR-144 is genes resulting in the silencing of PcG targets, including tumor- linked to overexpression of EZH2 in bladder and prostate cancer, suppressor genes.10 This indicates strong cross-talk between respectively.30,31 histone and DNA methylation resulting in a ‘double-lock’ repressive effect. There is further evidence to indicate that this combined epigenetic repressive effect becomes established in response to cancer-promoting DNA damage generated by EZH2 IN HEMATOLOGICAL CANCERS reactive oxygen species.27 By implication, aberrant EZH2 activity, EZH2 ‘gain-of-function’ ESC gene and/or chromatin signatures and reactive oxygen The occurrence of mutations within chromatin regulators in species (or other stresses) conspire to lock adult somatic cells cancer is linked to inappropriate gene expression and genomic into an aberrant CSC state. instability. Overexpression of wild-type (WT) EZH2 has been described in several non-hematological cancers, including prostate and breast cancer32,33 and evidence demonstrate that REGULATORS OF EZH2 IN CANCER this increased expression promotes tumor progression and EZH2 is controlled by a complex array of mechanisms, which may metastasis in these cancers.34 become dysregulated in different cancer types. For example, in In hematological malignancies, overexpression of mutant the healthy state, a co-dependent relationship exists between EZH2 has been recognized in a wide selection of B- and T-cell Ezh2 and c-Myc, which maintains a homeostatic balance, thereby lymphoproliferative disorders.35–38 One landmark study described promoting normal growth; c-Myc indirectly enhances Ezh2 activity a large cohort of diffuse large B-cell lymphoma (DLBCL) and via the stimulation of the PI3K pathway regulator Pten which in follicular lymphoma cases with EZH2 variants. ‘Gain-of-function’

& 2014 Macmillan Publishers Limited Leukemia (2014) 44 – 49 EZH2 in normal and malignant hematopoiesis K Lund et al 46 point mutations resulting in a switch from tyrosine to histidine at model, Ezh2 loss perturbed leukemic progression by reactivating codon 641 (Tyr641) in the catalytically active SET domain of the Ezh2-repressed genes involved in myeloid differentiation, such as EZH2 protein were found in 7.2% and 21.7% of the follicular Egr1, and appeared to convert AML into a MDS/MPN type lymphoma and DLBCL cases, respectively.37 All of the DLBCL cases disease.44 This indicates that the loss of Ezh2 function in this were germinal center (GC) cell phenotype and all appeared to be model converts a high-grade myeloid leukemia to a less heterozygous for the variant. This study demonstrated that, in aggressive MPN, indicating that the presence of Ezh2 confers apparent contrast to the increased EZH2 mRNA levels found in oncogenic activity in the myeloid lineage. Recent studies have breast and prostate cancer, Tyr641 mutations were associated indicated that in Mll-AF9 leukemia, Ezh2 exerts its oncogenic with marked reduction in EZH2 enzymatic activity in vitro. It was effects by inhibiting differentiation. Specifically, Thiel et al. have postulated that these mutations may change EZH2 target gene shown that the trithorax protein menin upregulates Ezh2, resulting specificity and so alter DNA methylation at PcG targets in these in inhibition of C/ebpa-mediated differentiation in Mll-AF9 tumors.37 leukemia.21 However, in a later study, the Tyr641 EZH2 mutation was shown Thus, the contrasting effects of Ezh2 overexpression and to confer an aberrant functional interaction between mutant and deletion, respectively, in these two mouse models ultimately WT gene products. Sneeringer et al. showed that the malignant results in a similar phenotype (that is, a MDS/MPN overlap phenotype arising in Tyr641-mutated B-cell lymphoma occurs due syndrome). In both cases, Ezh2 demonstrates oncogenic activity in to altered cooperation between the mutated and WT EZH2 with the myeloid lineage, but its end point clearly depends on cell and an overall ‘hyper-trimethylating’ effect on H3K27me3, resulting in genetic context. gene repression.39 Hence, the EZH2 Tyr641 mutation appears to be an unusual ‘gain-of-function’ mutation, after all. Massive upregulation of EZH2 expression is evident in GC B cells EZH2 ‘loss-of-function’ when compared with naive B-cell controls and interestingly the Although the evidence from human disease and mouse models EZH2 transcriptional regulatory profile in GC B cells appears to be described above indicates that gain of EZH2 H3K27 trimethylating highly correlated with that seen in human ESCs (ESCs).22 activity promotes lymphoid and myeloid cancers, there is also Chromatin immunoprecipitation (ChIP) of EZH2 from GC B cells, data to suggest that, conversely, loss of H3K27 trimethylation, naive B-cell controls and human ESCs, followed by microarray both directly via EZH2 inactivating mutations and indirectly analysis (ChIP on chip), identified promoters bound by EZH2 and via ASXL1 mutations,45 may also contribute to malignant demonstrated the repression of key cell-cycle-related tumor- hematopoiesis.46,47 This indicates that the EZH2 protein also has suppressor genes (CDKN1a, CDKN1b) in both GC B cells and capacity to work as a tumor suppressor in certain cellular contexts. human ESC as compared with naive B cells. In keeping with this, The landmark studies examining recurrent somatic inactivating small interfering RNA knockdown of EZH2 in DLBCL, a GC B-cell EZH2 mutations in MDS/MPN overlap disorders were performed malignancy, led to significant G1/S cell cycle arrest. This by two groups.46,47 Ernst et al. uncovered 49 mutations experimental evidence is complimentary to other reports (approximately one-third of these were homozygous, indicative indicating that EZH2 upregulation drives proliferative potential of inactivation of a tumor suppressor) in a total of 614 individuals and self-renewal in other B- and T-cell lymphoproliferative with myeloid malignancy–most commonly involving myelofibrosis disorders.35–38 The role of EZH2 in GC formation, and its or MDS/MPN overlap syndromes (chronic myelomonoctyic contribution toward GC-type DLBCL, has been attributed to its leukemia and atypical chronic myeloid leukemia). Nikoloski et al. activity at bivalent domains.40 Through these sites mutant EZH2 sequenced the EZH2 gene in 126 patients with MDS, revealing alleles or overexpression of WT EZH2 resulted in repression of somatic frameshift, nonsense and missense mutations throughout proliferation checkpoint and differentiation genes, thereby the gene. inducing GC hyperplasia and lymphomagenesis. Expression of Specific interest has focused on the link between mutant EZH2 in GC cells was unable to induce DLBCL alone, but 7 abnormalities, common in myeloid cancers, and the location of collaborated with BCL6/BCL2 to cause the disease phenotype. the EZH2 gene, present at position 7q36.1.46 On screening, Ernst A correlation is also described between EZH2 overexpression et al. established that EZH2 mutations in the second allele were and myeloid malignancy.41 Xu et al. examined a heterogeneous not particularly associated with chromosome 7 losses or 7q myelodysplastic syndrome (MDS)/AML population known to deletion. However, they did report finding homozygous EZH2 harbor DNA methylation of tumor-suppressor genes, such as mutations in 75% of individuals with acquired uniparental disomy. p15INK4B. Patients with p15INK4B gene methylation had Univariate analysis in MDS/MPN indicated that EZH2 mutations statistically higher mean relative expression of EZH2 compared were associated with both a poorer overall and progression-free with non-methylated counterparts. The level of EZH2 expression survival when compared with cases without mutations (overall correlated positively with poor disease outcome as judged by the survival: 39 months vs 13 months (P ¼ 0.0006); progression-free International Prognostic Scoring System.42 survival: 30 months vs17 months (P ¼ 0.044)).46 The presence of In order to further investigate the mechanisms by which Ezh2 homozygous mutations conferred a trend toward poorer survival may contribute to myeloid malignancy, recent publications have when compared with heterozygotes, although the difference was used murine models of leukemia to assess the function of the not statistically significant. A similar prognostic picture was protein. A conditional Ezh2 ‘gain-of-function’ mouse model uncovered in a cohort of myelofibrosis patients where loss-of- showed that induction of WT Ezh2 expression increases the function EZH2 mutations were found to independently predict a repopulation potential of HSCs.43 Serial transplantation of these poor overall survival.48 Ezh2-overexpressing HSCs resulted in mice developing a Following assessment of a 469 patients with myeloid malig- myeloproliferative neoplasm (MPN), consistent with the nancies, Khan et al. report that EZH2 mutations were present in 8% hypothesis that MPNs arise from a mutation in the stem cell of cases. Some of these mutations, such as mutations in splicing pool. In this model, a number of HSC maintenance genes were factor genes, U2AF1 and SRSF2, cause dysfunctional processing of regulated by Ezh2, including the transcriptional regulators Evi-1 pre-mRNA and reduced EZH2 expression.20 Khan et al. also and Ntrk3, which are often aberrantly expressed in hematological describe how the stem cell self-renewal HOX gene family is a malignancies.43 In another model, Mll-AF9 transformed AML cells major downstream target of EZH2 and how a similar picture of from WT and Ezh2-deficient mice have been studied in vitro and increased HOXA9 expression occurs as a consequence of both in vivo.44 Deletion of Ezh2 compromised cell cycle progression, inactivating EZH2 mutations or as a consequence of deletion of although cells still maintained some proliferative capacity. In this the whole EZH2 locus in patients harboring -7/del7q abnormality.

Leukemia (2014) 44 – 49 & 2014 Macmillan Publishers Limited EZH2 in normal and malignant hematopoiesis K Lund et al 47 They propose that EZH2 ‘loss-of-function’ mutations contribute to deacetylation inhibitors, which are increasingly used in combina- formation of the CSC, via HOXA9-mediated self-renewal of myeloid tion with conventional therapies,52 a number of other novel small- progenitors.20 molecule inhibitors, including EZH2 inhibitors, are in Evidence of EZH2 ‘loss-of-function’ mutations is not limited to development. aberrant myelopoiesis. Ntziachristos et al. identified EZH2 or SUZ12 mutations in 25% of primary T-acute lymphoblastic leukemia (T-ALL) samples.49 Studies reporting the function of EZH2 in EZH2 INHIBITORS human and mouse T-ALL have used various strategies to Given that evidence points to a role for EZH2 in increasing self- demonstrate that loss-of-function PRC2 mutations contribute to renewal in at least some cancers (including some hematological the pathogenesis of T-ALL, which is predominantly driven by malignancies), therapies that inhibit EZH2 may promote exhaus- oncogenic activation of NOTCH1 signaling.49,50 EZH2 silencing is tion of CSC populations.15 Moreover, if EZH2 has non-stem cell- shown to increase the tumorigenic potential and mortality of specific oncogenic functions, for example, promoting cell T-ALL cells transplanted into NOD-SCID mice.49 proliferation or inhibiting differentiation, its inhibition would also A summary of subtype and frequency of EZH2 mutations in be expected to be of therapeutic benefit. Consequently, myeloid and lymphoid malignancies is provided in Figure 3 and considerable effort has been directed toward development of Table 1. The conflicting findings discussed above highlight the EZH2 inhibitors. complexity of epigenetic regulation, illustrating that polycomb 3-Deazaneplanocin (DZNep), the first drug proposed to inhibit proteins, such as EZH2, can have a dual role as either oncogenes EZH2, acts indirectly via competitive inhibition of S-adenosylho- or tumor-suppressor genes, depending on factors such as gene mocysteine hydrolase. Consequently, there is an accumulation of dosage and cell context. These divergent phenotypes suggest that the enzyme substrate adenosylhomocysteine, which in turn the PcG complexes act on a broad range of target genes resulting inhibits methyltransferases, induces degradation of the PRC2 in potentially opposing downstream effects.51 complex and reduces EZH2 levels.4,5 There is evidence to indicate that the drug has a broad inhibitory effect on protein methyltransferases other than EZH2, and is therefore not a well- 4 EPIGENETIC THERAPIES targeted therapy. However, based on murine studies, the use of DZNep may offer the possibility of therapeutic potential as an A key feature of epigenetic mutations in malignancy is their immuno-modulatory drug in the treatment of graft-versus-host relative plasticity, unlike genetic changes to DNA sequence that disease, a common complication of allogeneic stem cell are essentially permanent. This has important clinical relevance, as transplantation for hematological malignancies.53 Histone this feature can make them more amenable to inhibitor therapies methylation modulates inflammatory T-cell responses and the that reverse the epigenetic alterations. In addition to established use of DZNep in a graft-versus-host disease mouse model agents, such as DNA methyltransferease inhibitors and histone activated pro-apoptotic genes in allo-antigen-activated T cells (thereby suppressing graft-versus-host disease), whereas maintaining the anti-leukemic activity of donor T cells. The discovery of EZH2 mutations in lymphoma (particularly DLBCL) has provided a model system/novel target for the development of EZH2 inhibitors.7,8,54,55 In their work describing microRNA function in regulating EZH2, Zhao et al. coupled DZNep first with histone deacetylation inhibitor Vorinostat, and second with bromodomain and extra-terminal (BET) domain inhibitor JQ1.54 The action of this drug combination resulted in disruption of the c-MYC-miR-EZH2-HDAC3 feedback loop (alluded to in Figure 2b), which in turn increased tumor-suppressor miR-29 Figure 3. EZH2 domain structure and positions of missense/point expression, downregulated miR-29 target genes and reduced mutations identified in hematological cancers. CXC, cysteine-rich lymphoma growth. domain; D1, domain I; D2, domain II; GCB, germinal center B cell; MDS/MPN, myelodyplastic/myeloproliferative neoplasms; SET, sup- Qi et al. developed an Ezh2-selective small-molecule inhibitor pressor of variegation 3–9, enhancer of zeste and trithorax domain; EI1, which competitively binds to the S-adenosylmethionine (SAM) T-ALL;T-cell acute lymphoblastic leukemia. Mutational hot spots pocket of the Ezh2 SET domain in both WT and Tyr641 mutated identified at D2 and SET domains, frameshift and nonsense cells.55 This inhibition of histone H3K27me3 led to G1 growth mutations are more equally distributed across the genome (not arrest, apoptosis and differentiation of Ezh2 mutant cells into shown on this figure). memory B cells.

Table 1. EZH2 pathogenic mutations in hematological cancers

Disease Mutation types identified Effect Frequency (%) Reference

DLBCL-GCB Missense Gain of function 14–21.7 37, 56 Follicular Lymphoma 7.2–22 MDS/MPN Missense frameshift, nonsense Loss of function 8–12 20, 46 MPN 3–13 MDS 2.5 AML 2 T-ALL Missense, frameshift, nonsense Loss of function 18 49 Abbreviations: AML, acute myeloid leukemia; DLBCL-GCB, diffuse large B-cell lymphoma-germinal center type; EZH2, Enhancer of Zeste Homologue 2; MDS/MPN, myelodysplastic/myeloproliferative neoplasm; T-ALL, T-acute lymphoblastic leukemia.

& 2014 Macmillan Publishers Limited Leukemia (2014) 44 – 49 EZH2 in normal and malignant hematopoiesis K Lund et al 48 Two other recent publications have demonstrated further development and clinical trials adopting the new agents that advances in the therapeutic potential of EZH2 inhibition to treat selectively target oncogenic EZH2 have yet to be undertaken. lymphoma.7,8 Two compounds, EPZ005687 and GSK126, However, careful patient selection will be imperative in future independently identified by high-throughput screening, inhibit trials to ensure that the potentially contrary effects of such drugs EZH2 using a similar mechanism to that described for EI1. Both are appropriately exploited, reducing tumor growth rather than compounds are highly selective for EZH2 (500- to 1000-fold, accelerating it. compared with other methyltransferases), reduce H3K27me3 levels and increase transcriptional activation. Both drugs have proven efficacy in inducing apoptotic killing of lymphoma cell CONFLICT OF INTEREST lines harboring Tyr641 mutations with minimal effect on WT cells. The authors declare no conflict of interest. Although the EPZ005687 compound has been limited to in vitro use,7 the GSK126 molecule has also been assessed in mouse xenograft models, where it resulted in complete inhibition of ACKNOWLEDGEMENTS tumor growth and significantly increased survival of the mice.8 In MC is supported by the Scottish Funding Council (SCD/04) and Leukaemia and line with these therapeutic advances, the development of a Lymphoma Research (Grant ref: 11017). PA is supported by Cancer Research UK clinically applicable screening assay to detect Tyr641 mutations to (Grant ref: C10652/A10250). We thank Professor Tessa Holyoake for critical review of guide future ‘genotype targeted therapy’ has been reported.56 the manuscript. Pre-clinical studies clearly show that inhibition of EZH2 warrants further investigation in myeloid malignancy.3,6,44 AML represents AUTHOR CONTRIBUTIONS an important disease in which to investigate epigenetic therapies, KL wrote and revised the manuscript. MC and PA revised the manuscript. All authors and there is also increasing evidence for the use of these agents in checked the final version. myelofibrosis and MDS. However, evidence for inactivating EZH2 mutations in myeloid disease indicates a tumor suppressor function for EZH2 in the myeloid lineage, meaning that EZH2 REFERENCES inhibitors must be applied with caution. It will be necessary to 1 Strahl BD, Allis CD. The language of covalent histone modifications. Nature 2000; identify genetic or other biomarkers to distinguish between 403: 41–45. 2 Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H, Tempst P et al. Role of groups of patients who will benefit or not from EZH2 inhibitors. histone H3 lysine 27 methylation in Polycomb-group silencing. Science 2002; 298: Further research is also required into ‘loss-of-function’ mutations 1039–1043. reported in myeloid malignancies to identify whether inhibition of 3 Zhou J, Bi C, Cheong LL, Mahara S, Liu SC, Tay KG et al. The histone a downstream target upregulated by the ‘loss-of-function’ methyltransferase inhibitor, DZNep, up-regulates TXNIP, increases ROS mutation, for example, HOXA9, or the development of selective production, and targets leukemia cells in AML. Blood 2011; 118: 2830–2839. inhibitors of histone demethylases might provide a practical and 4 Miranda TB, Cortez CC, Yoo CB, Liang G, Abe M, Kelly TK et al. DZNep is a global effective therapeutic approach.20 inhibitor that reactivates developmental genes not silenced There is mounting evidence from preclinical studies that by DNA methylation. Mol Cancer Ther 2009; 8: 1579–1588. epigenetic therapies target AML stem cells.3,57 Current hypotheses 5 Tan J, Yang X, Zhuang L, Jiang X, Chen W, Lee PL et al. Pharmacologic disruption postulate that relapse in AML is related to a population of CSCs, of Polycomb-repressive complex 2-mediated gene repression selectively induces apoptosis in cancer cells. Genes Dev 2007; 21: 1050–1063. which are resistant to conventional chemotherapy agents, for 58 6 Fiskus W, Wang Y, Sreekumar A, Buckley KM, Shi H, Jillella A et al. example, cytarabine and daunorubicin. However, to date, clinical Combined epigenetic therapy with the histone methyltransferase EZH2 inhibitor trials using epigenetic therapies as a single agent have proved 3-deazaneplanocin A and the histone deacetylase inhibitor panobinostat against 59,60 disappointing. Therefore, more targeted epigenetic therapies human AML cells. Blood 2009; 114: 2733–2743. such as EZH2 inhibition may offer the opportunity of a novel 7 Knutson SK, Wigle TJ, Warholic NM, Sneeringer CJ, Allain CJ, Klaus CR et al. treatment approach that could be considered in combination with A selective inhibitor of EZH2 blocks H3K27 methylation and kills mutant conventional chemotherapy to eradicate AML stem cells. However, lymphoma cells. Nat Chem Biol 2012; 8: 890–896. given the dual oncogenic/tumor-suppressor activity of EZH2, this 8 McCabe MT, Ott HM, Ganji G, Korenchuk S, Thompson C, Van Aller GS et al. would need to be tested carefully to ensure a beneficial, rather than EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations. Nature 2012; 492: 108–112. tumorigenic, effect was achieved. Epigenetic therapy could be 9 Francis NJ, Kingston RE, Woodcock CL. Chromatin compaction by a polycomb considered as a consolidation or maintenance therapy after the group protein complex. Science 2004; 306: 1574–1577. high-dose remission-induction chemotherapy. This strategy, using 10 Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C et al. The Polycomb azacitidine as the epigenetic therapy, has recently been trialed in group protein EZH2 directly controls DNA methylation. Nature 2006; 439: the intensive arm MRC AML16 clinical trial for elderly patients with 871–874. AML and high-risk MDS and the trial will report shortly (http:// 11 Wang H, Wang L, Erdjument-Bromage H, Vidal M, Tempst P, Jones RS et al. Role of www.aml16.bham.ac.uk; Trial Reference ISRCTN 11036523). ubiquitination in Polycomb silencing. Nature 2004; 431: 873–878. 12 Bracken AP, Dietrich N, Pasini D, Hansen KH, Helin K. Genome-wide mapping of Polycomb target genes unravels their roles in cell fate transitions. Genes Dev 2006; CONCLUSION 20: 1123–1136. 13 Chou RH, Yu YL, Hung MC. The roles of EZH2 in cell lineage commitment. Advances in our understanding of how the epigenomic landscape Am J Transl Res 2011; 3: 243–250. can be manipulated will continue as the complex relationships 14 de Haan G, Gerrits A. Epigenetic control of hematopoietic stem cell aging - between epigenetic marks are gradually uncovered. Gene The case of Ezh2. Ann NY Acad Sci 2007; 1106: 233–239. repression resulting from EZH2 activity is a key component of 15 Kamminga LM, Bystrykh LV, de Boer A, Houwer S, Douma J, Weersing E et al. the epigenetic machinery with effects on both DNA and histone The Polycomb group gene Ezh2 prevents hematopoietic stem cell exhaustion. methylation. Further investigation to gain better insight into the Blood 2006; 107: 2170–2179. dependence of cancer growth on EZH2 is warranted, particularly 16 Ezhkova E, Pasolli HA, Parker JS, Stokes N, Su IH, Hannon G et al. Ezh2 orchestrates to unravel the complexity of the protein’s capacity to induce both gene expression for the stepwise differentiation of tissue-specific stem cells. Cell 2009; 136: 1122–1135. pro-oncogenic and tumor-suppressive effects. 17 Bracken AP, Pasini D, Capra M, Prosperini E, Colli E, Helin K. EZH2 is downstream of The evidence outlined in this review points to an interesting the pRB-E2F pathway, essential for proliferation and amplified in cancer. dichotomy regarding the potential of EZH2 as a therapeutic target, Embo J 2003; 22: 5323–5335. inhibition of this protein could potentially be beneficial or 18 Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that detrimental dependent on the cancer context. Further originates from a primitive hematopoietic cell. Nat Med 1997; 3: 730–737.

Leukemia (2014) 44 – 49 & 2014 Macmillan Publishers Limited EZH2 in normal and malignant hematopoiesis K Lund et al 49 19 Dolnik A, Engelmann JC, Scharfenberger-Schmeer M, Mauch J, Kelkenberg-Schade S, 40 Beguelin W, Popovic R, Teater M, Jiang Y, Bunting KL, Rosen M et al. EZH2 is Haldemann B et al. Commonly altered genomic regions in acute myeloid leukemia required for germinal center formation and somatic EZH2 mutations promote are enriched for somatic mutations involved in and lymphoid transformation. Cancer Cell 2013; 23: 677–692. splicing. Blood 2012; 120: e83–e92. 41 Grubach L, Juhl-Christensen C, Rethmeier A, Olesen LH, Aggerholm A, Hokland P 20 Khan SN, Jankowska AM, Mahfouz R, Dunbar AJ, Sugimoto Y, Hosono N et al. et al. Gene expression profiling of Polycomb, Hox and Meis genes in patients with Multiple mechanisms deregulate EZH2 and histone H3 lysine 27 epigenetic acute myeloid leukaemia. Eur J Haematol 2008; 81: 112–122. changes in myeloid malignancies. Leukemia 2013; 27: 1301–1309. 42 Xu F, Li X, Wu L, Zhang Q, Yang R, Yang Y et al. Overexpression of the EZH2, RING1 21 Thiel AT, Feng Z, Pant DK, Chodosh LA, Hua X. The trithorax protein partner menin and BMI1 genes is common in myelodysplastic syndromes: relation to adverse acts in tandem with EZH2 to suppress C/EBPalpha and differentiation in MLL-AF9 epigenetic alteration and poor prognostic scoring. Ann Hematol 2011; 90: leukemia. Haematologica 2013; 98: 918–927. 643–653. 22 Velichutina I, Shaknovich R, Geng H, Johnson NA, Gascoyne RD, Melnick AM et al. 43 Herrera-Merchan A, Arranz L, Ligos JM, de Molina A, Dominguez O, Gonzalez S. EZH2-mediated epigenetic silencing in germinal center B cells contributes to Ectopic expression of the histone methyltransferase Ezh2 in haematopoietic stem proliferation and lymphomagenesis. Blood 2010; 116: 5247–5255. cells causes myeloproliferative disease. Nat Commun 2012; 3: 623. 23 van Vlerken LE, Kiefer CM, Morehouse C, Li Y, Groves C, Wilson SD et al. EZH2 is 44 Tanaka S, Miyagi S, Sashida G, Chiba T, Yuan J, Mochizuki-Kashio M et al. required for breast and pancreatic cancer stem cell maintenance and can be used Ezh2 augments leukemogenecity by reinforcing differentiation blockage in acute as a functional cancer stem cell reporter. Stem Cells Transl Med 2013; 2: 43–52. myeloid leukemia. Blood 2012; 120: 1107–1117. 24 Widschwendter M, Fiegl H, Egle D, Mueller-Holzner E, Spizzo G, Marth C et al. 45 Abdel-Wahab O, Adli M, LaFave LM, Gao J, Hricik T, Shih AH et al. ASXL1 mutations Epigenetic stem cell signature in cancer. Nat Genet 2007; 39: 157–158. promote myeloid transformation through loss of PRC2-mediated gene repression. 25 Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmerman J et al. Cancer Cell 2012; 22: 180–193. Polycomb-mediated methylation on Lys27 of histone H3 pre-marks genes for 46 Ernst T, Chase AJ, Score J, Hidalgo-Curtis CE, Bryant C, Jones AV et al. Inactivating de novo methylation in cancer. Nat Genet 2007; 39: 232–236. mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat 26 Ohm JE, McGarvey KM, Yu X, Cheng L, Schuebel KE, Cope L et al. A stem Genet 2010; 42: 722–726. cell-like chromatin pattern may predispose tumor suppressor genes to DNA 47 Nikoloski G, Langemeijer SM, Kuiper RP, Knops R, Massop M, Tonnissen ER et al. hypermethylation and heritable silencing. Nat Genet 2007; 39: 237–242. Somatic mutations of the histone methyltransferase gene EZH2 in myelodys- 27 O’Hagan HM, Wang W, Sen S, Destefano Shields C, Lee SS, Zhang YW et al. plastic syndromes. Nat Genet 2010; 42: 665–667. Oxidative damage targets complexes containing DNA methyltransferases, SIRT1, 48 Guglielmelli P, Biamonte F, Score J, Hidalgo-Curtis C, Cervantes F, Maffioli M et al. and polycomb members to promoter CpG Islands. Cancer Cell 2011; 20: 606–619. EZH2 mutational status predicts poor survival in myelofibrosis. Blood 2011; 118: 28 Kaur M, Cole MD. MYC acts via the PTEN tumor suppressor to elicit autoregulation 5227–5234. and genome-wide gene repression by activation of the Ezh2 methyltransferase. 49 Ntziachristos P, Tsirigos A, Van Vlierberghe P, Nedjic J, Trimarchi T, Flaherty MS et al. Cancer Res 2013; 73: 695–705. Genetic inactivation of the polycomb repressive complex 2 in T cell acute 29 Zhang X, Zhao X, Fiskus W, Lin J, Lwin T, Rao R et al. Coordinated silencing of lymphoblastic leukemia. Nat Med 2012; 18: 298–301. MYC-mediated miR-29 by HDAC3 and EZH2 as a therapeutic target of histone 50 Simon C, Chagraoui J, Krosl J, Gendron P, Wilhelm B, Lemieux S et al. A key role for modification in aggressive B-Cell lymphomas. Cancer Cell 2012; 22: 506–523. EZH2 and associated genes in mouse and human adult T-cell acute leukemia. 30 Varambally S, Cao Q, Mani RS, Shankar S, Wang X, Ateeq B et al. Genomic loss of Genes Dev 2012; 26: 651–656. microRNA-101 leads to overexpression of histone methyltransferase EZH2 in 51 Mochizuki-Kashio M, Mishima Y, Miyagi S, Negishi M, Saraya A, Konuma T et al. cancer. Science 2008; 322: 1695–1699. Dependency on the polycomb gene Ezh2 distinguishes fetal from adult 31 Guo Y, Ying L, Tian Y, Yang P, Zhu Y, Wang Z et al. miR-144 downregulation hematopoietic stem cells. Blood 2011; 118: 6553–6561. increases bladder cancer cell proliferation by targeting EZH2 and regulating Wnt 52 Gore SD, Baylin S, Sugar E, Carraway H, Miller CB, Carducci M et al. Combined DNA signaling. FEBS J 2013; 280: 4531–4538. methyltransferase and histone deacetylase inhibition in the treatment of myeloid 32 Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Kumar-Sinha C, Sanda MG neoplasms. Cancer Res 2006; 66: 6361–6369. et al. The polycomb group protein EZH2 is involved in progression of prostate cancer. Nature 2002; 419: 624–629. 53 He S, Wang J, Kato K, Xie F, Varambally S, Mineishi S et al. Inhibition of 33 Kleer CG, Cao Q, Varambally S, Shen RL, Ota L, Tomlins SA et al. EZH2 is a marker histone methylation arrests ongoing graft-versus-host disease in mice by of aggressive breast cancer and promotes neoplastic transformation of breast selectively inducing apoptosis of alloreactive effector T cells. Blood 2012; 119: epithelial cells. Proc Natl Acad Sci USA 2003; 100: 11606–11611. 1274–1282. 34 Ren G, Baritaki S, Marathe H, Feng J, Park S, Beach S et al. Polycomb protein EZH2 54 Zhao X, Lwin T, Zhang X, Huang A, Wang J, Marquez VE et al. Disruption of the regulates tumor invasion via the transcriptional repression of the metastasis MYC-miRNA-EZH2 loop to suppress aggressive B-cell lymphoma survival and suppressor RKIP in breast and prostate cancer. Cancer Res 2012; 72: clonogenicity. Leukemia 2013; 27: 2341–2350. 3091–3104. 55 Qi W, Chan H, Teng L, Li L, Chuai S, Zhang R et al. Selective inhibition of Ezh2 by a 35 Raaphorst FM, van Kemenade FJ, Blokzijl T, Fieret E, Hamer KM, Satijn DP et al. small molecule inhibitor blocks tumor cells proliferation. Proc Natl Acad Sci USA Coexpression of BMI-1 and EZH2 polycomb group genes in Reed-Sternberg cells 2012; 109(): 21360–21365. of Hodgkin’s disease. Am J Pathol 2000; 157: 709–715. 56 Ryan RJ, Nitta M, Borger D, Zukerberg LR, Ferry JA, Harris NL et al. EZH2 codon 641 36 Visser HP, Gunster MJ, Kluin-Nelemans HC, Manders EM, Raaphorst FM, Meijer CJ mutations are common in BCL2-rearranged germinal center B cell lymphomas. et al. The Polycomb group protein EZH2 is upregulated in proliferating, cultured PLoS One 2011; 6: e28585. human mantle cell lymphoma. Br J Haematol 2001; 112: 950–958. 57 Zhou L, Ruvolo VR, McQueen T, Chen W, Samudio IJ, Conneely O et al. HDAC 37 Morin RD, Johnson NA, Severson TM, Mungall AJ, An J, Goya R et al. Somatic inhibition by SNDX-275 (Entinostat) restores expression of silenced leukemia- mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas associated transcription factors Nur77 and Nor1 and of key pro-apoptotic proteins of germinal-center origin. Nat Genet 2010; 42: 181–185. in AML. Leukemia 2013; 27: 1358–1368. 38 Sasaki D, Imaizumi Y, Hasegawa H, Osaka A, Tsukasaki K, Choi YL et al. 58 Horton SJ, Huntly BJ. Recent advances in acute myeloid leukemia stem cell Overexpression of Enhancer of zeste homolog 2 with trimethylation of lysine 27 biology. Haematologica 2012; 97: 966–974. on histone H3 in adult T-cell leukemia/lymphoma as a target for epigenetic 59 Ozbalak M, Cetiner M, Bekoz H, Atesoglu EB, Ar C, Salihoglu A et al. Azacitidine has therapy. Haematologica 2011; 96: 712–719. limited activity in ’real life’ patients with MDS and AML: a single centre experi- 39 Sneeringer CJ, Scott MP, Kuntz KW, Knutson SK, Pollock RM, Richon VM et al. ence. Hematol Oncol 2012; 30: 76–81. Coordinated activities of wild-type plus mutant EZH2 drive tumor-associated 60 Craddock C, Quek L, Goardon N, Freeman S, Siddique S, Raghavan M et al. Azacitidine hypertrimethylation of lysine 27 on histone H3 (H3K27) in human B-cell fails to eradicate leukemic stem/progenitor cell populations in patients with acute lymphomas. Proc Natl Acad Sci USA 2010; 107: 20980–20985. myeloid leukemia and myelodysplasia. Leukemia 2013; 27: 1028–1036.

& 2014 Macmillan Publishers Limited Leukemia (2014) 44 – 49