Mutant DNA methylation regulators endow hematopoietic stem cells with the preleukemic stem cell property, a requisite of leukemia initiation and relapse

Abstract

Genetic mutations are considered to drive the development of acute myeloid leukemia (AML). With the rapid progress in sequencing technologies, many newly reported genes that are recurrently mutated in AML have been found to govern the initiation and relapse of AML. These findings suggest the need to distinguish the driver mutations, especially the most primitive single mutation, from the subsequent passenger mutations. Recent research on DNA methyltransferase 3A (DNMT3A) mutations provides the first proof-of-principle investigation on the identification of preleukemic stem cells (pre-LSCs) in AML patients. Although DNMT3A mutations alone may only transform hematopoietic stem cells into pre-LSCs without causing the full-blown leukemia, the function of this driver mutation appear to persist from AML initiation up to relapse. Therefore, identifying and targeting preleukemic mutations, such as DNMT3A mutations, in AML is a promising strategy for treatment and reduction of relapse risk.

Keywords

preleukemic stem cell / acute myeloid leukemia / relapse / DNMT3A

Introduction

Acute myeloid leukemia (AML) is a result of a series of genetic mutations and epigenetic dysregulations [ 1]. In the past three decades, numerous works investigated the crosstalk among the mutated genes of transcription factors and signal transduction molecules related to AML based on the classical two-hit model. However, with the rapid development of sequencing technologies, additional newly reported mutated genes in AML have highlighted their pivotal roles. Many of these mutations are closely linked to epigenetic dysregulations, such as DNA methylation and histone modifications [ 2− 5]. These novel classes of mutant disease alleles directly expanded the canonical gene mutational classification of myeloid malignancies from class I and II to an additional class III [ 6, 7]. Class I mutations are widely accepted to confer proliferation advantages, whereas class II mutations are known to block cell differentiation. However, class III mutations are hardly endowed with unique features. Considering that perturbations of epigenetic modifiers may result in various alterations, including aberrant cell proliferation and differentiation [ 8], we believe that strictly separating each kind of mutation from one another is difficult to attain. Interestingly, class III mutations are more often involved in the preleukemic transformation [ 9]. According to the cancer-stem-cell evolution model, preleukemic stem cells serve as the root of AML pathogenesis [ 10]. The related driver mutations not only play a key role in AML initiation but also survive chemotherapy and consequently trigger AML relapse [ 11, 12]. Therefore, identifying the driver mutations from subsequent mutations may enable the targeting of perturbations in these AML initiating events and potentially minimize the risk of relapse. Herein, we briefly summarize the roles of preleukemic mutations in the initiation and relapse of AML and focus on the notable preleukemic mutation of DNMT3A.

Preleukemic stem cells in AML initiation and relapse

Normal hematopoiesis is hierarchically organized, beginning with the multipotent hematopoietic stem and progenitor cells (HSPCs) [ 13− 16]. Several xenograft mouse models have demonstrated that in acute leukemias, hematopoietic stem cells (HSCs), common myeloid progenitors (CMPs), or granulocyte monocyte progenitors (GMPs) are believed to be the initial malignant cells-of-origin, which subsequently give rise to the pre-LSCs or the leukemic stem cells (LSCs) [ 17− 22]. Pre-LSCs refer to a small subsection of cells directly derived from HSCs that bear a sole mutation and share many properties with normal HSCs, such as self-renewal and differentiation [ 17]. Therefore, pre-LSCs are capable of maintaining and replenishing all blood cell types but unable to induce full-blown leukemia. On the contrary, LSCs are functionally defined by their capacity to produce fully penetrant, short-onset leukemia in murine transplantation models [ 17, 20, 23, 24]. The existence of LSCs has been demonstrated for over more than 20 years following the reports that only a small sub-fraction of leukemic cells from AML models are capable of propagating the disease in mouse xenografts [ 23− 26]. However, the identification of pre-LSCs has not been reported until recent years by studying human blood samples [ 12, 27, 28]. For many years, only a logical postulation that a more naïve form of malignant stem cells than the LSCs might exist [ 29]. The progression from a normal HSC to a fully malignant LSC demands the HSC to overcome the homeostatic control of their numbers and survival. In this regard, an evolutionary selection of the advantageous cells maintaining the first genetic lesion must have occurred, which are thought to be the pre-LSCs. Distinguishing pre-LSCs from normal HSCs is difficult because pre-LSCs could retain most HSC properties. This finding may explain why the first proof of principle for the identification of pre-LSCs has not been reported until recently after the hypothesis on pre-LSCs had been proposed for many years [ 27]. However, pre-LSCs also holds a genetic lesion; the single mutation confers the pre-LSCs with selective advantages over the nonleukemic stem cells, allowing a more powerful clonal expansion and producing more progenies (Fig. 1).

In addition to that in the pre-LSCs, the first mutation might also occur in more differentiated myeloid progenitor cells, giving rise to the “preleukemic myeloid progenitors” (pre-LMP) [ 30, 31]. Notably, this conclusion was only confirmed in knock-in mouse models, and no direct evidence has proven that human AML can be driven by pre-LMP. These preleukemic cells are not only pivotal in AML initiation, but recent findings have also implicated their core role in AML relapse [ 12, 32− 34]. Models of AML relapse are closely related to the pathways of AML initiation and evolution [ 29]. Typically, three possible scenarios for AML evolution might provide clues to the relapse of diseases (Fig. 2). First, the primitive mutation directly hits the HSCs, thereby producing the pre-LSCs, followed by the second hit in the pre-LSCs and thus generating the LSCs. In this scenario, both the initial and secondary mutations are at the stem-cell level; hence, the LSCs share many characteristics with the normal HSCs and consequently become resistant to drug therapy [ 29, 35− 37]. Second, the HSCs are hit by the initial mutation, but afterward differentiate into pre-LMPs, in which the secondary mutations occur. In this category, the LSCs originate from more mature pre-LMPs, and thus are more vulnerable to chemotherapy. Unlike the pre-LMPs, such LSCs cease to serve as reservoirs for relapse [ 30, 31, 38]. In the third scenario, initial mutation hits the downstream progenitor cells but not the HSCs. This process also produces pre-LMPs and LSCs, but all the mutations miss the stem-cell level, thus rendering the blasts, as well as the pre-LMPs and CMPs, susceptible to the therapeutic agents [ 29− 31, 38]. In this case, AML derived from a myeloid progenitor origin might be completely treated.

In summary, we discussed the predominant roles of preleukemic condition in the initiation and relapse of AML. As already mentioned, inspection of DNMT3A-mutated HSCs was the first proof-of-principle identification of pre-LSCs in AML patients. Our group also conducted several works investigating this epigenetic modifier. As such, we further discuss the DNMT3A mutations in AML as typical paradigms of preleukemic mutations.

DNA methylation in leukemia

Epigenetic modifications are employed in various manners, such as DNA methylation, histone methylation, and chromatin remodeling. DNA methylation is a highly important epigenetic regulatory activity of gene expression occurring merely at the CpG residues in eukaryotic cells [ 39]. CpG residues usually cluster together and form the so-called “CpG island,” rather than sporadically scatter throughout the genome [ 7]. A landscape analysis of a DNA methylome suggested that in addition to CpG islands, CpG residues also occur on island “shores,” defined as those situated less than 2 kb from an island, and “shelves,” which are located 2−4 kb from an island [ 39, 40]. Approximately 60% of human genes contain CpG islands in their gene promoter regions. DNA methylation of the promoter region is more commonly believed to downregulate gene expression [ 41− 43].

DNA methylation is dynamically altered in both solid tumors and leukemia; it can affect cells in three main ways, namely, hypomethylation, hypermethylation, and loss of imprinting [ 44]. Recent studies implied that DNA demethylation mechanisms can potentially alter the DNA methylation pattern in leukemias [ 45]. Ten-eleven translocation (TET) proteins, especially TET2, play a vital role in lowering DNA methylation levels by sequentially converting 5-methyl-cytosine (5-mC) to 5-hydroxymethyl-cytosine (5-hmC), 5-formylcytosine (5fC), and 5-carboxycytosine (5caC) [ 46]. On the other hand, isocitrate dehydrogenase (IDH) indirectly influences DNA demethylation by negatively regulating the function of TET proteins after IDH mutations [ 47]. Global hypomethylation has been most frequently reported in different cancers; in these cases, the inhibitory effects of DNA methylation on promoters are postulated to be impaired, thus promoting the transcription of oncogenes and genes of cell replication [ 48− 50]. However, several newly reported works indicated that global methylation of DNA promoter regions remained almost intact despite the mutation or ablation of DNMT3A [ 51, 52]. Notably, gene bodies are markedly hypomethylated or hypermethylated in response to the aberrantly functioning DNA methyltransferase [ 51, 53]. These counterintuitive observations remind us that gene bodies are associated with transcribed regions that possess higher methylation levels than promoter or intergenic sequences [ 54]. In particular, the genes expressed mid-level exhibit the highest amount of gene-body methylation; this finding may explain why the methylation levels of gene bodies were more strikingly varied than that of the promoters in AML. Even so, what determines the hypo- or hypermethylation of gene bodies remains to be defined, and the consequent alterations of cell fate based must be explored.

DNMT3A mutations in AML

DNMT3A mutations were first reported in AML by Ley and colleagues and our group [ 55, 56]. These mutations are widely reported in 17%−34% of AML cases with normal karyotype (CN-AML) [ 57]. Almost 96.3% of the DNMT3A mutations are heterozygous. Moreover, 89% of the CN-AML patients harbored a single site mutation of DNMT3A, whereas the occurrence of two mutations in one allele was most frequently observed in T-cell acute lymphoblastic leukemia samples [ 57, 58]. The patients with DNMT3A mutations are usually older than average, possess lower hemoglobin levels, and exhibit higher white blood cell counts. Typically, these patients are also more likely to suffer from monocytic or myelomonocytic leukemia (FAB M4/5) than other patients [ 59].

DNMT3A mutations usually contain missense, nonsense, or frame-shift alterations. Among these mutations, the missenses at codon R882 are the most frequently detected, which account for nearly 60% of the DNMT3A mutations [ 60]. To study the function of R882 mutations, several studies conducted both in vitro and in vivo experiments. Yan and colleagues found that R882H mutations reduced methylation activity twofold [ 56]. Russler-Germain and coworkers co-expressed the wild-type and R882H-mutated DNMT3A in vitro and found that the enzymatic function of the wild-type DNMT3A was profoundly impaired by its mutated partner, probably via inhibition of tetramerization [ 61]. However, in vivo, DNMT3A mutations do not markedly alter global methylation levels [ 7, 26]. Similarly, Challen and colleagues established a DNMT3A-knock-out (KO) mouse model and neither found a whole-genome altered methylation pattern nor observed leukemia in serially transplanted mice [ 52]. Recently, by detecting various hematologic malignancies after one year, Mayle and colleagues demonstrated that in the absence of serial transplantation and with longer in vivo maintenance, the DNMT3A-KO mouse model could recapitulate the DNMT3A mutations [ 62]. Notably, our group established a retrovirally transduced DNMT3A-R882H mutant mouse model. At 12 months after bone marrow transplantation, all mice developed chronic myelomonocytic leukemia with thrombocytosis [ 51]. This result indicates that the DNMT3A-ablated mouse model might not thoroughly mimic the DNMT3A mutant model, and a DNMT3A knock-in mouse model may be fairly helpful in better investigating the biological effects of the DNMT3A mutation. Overall, the discrepancy between the in vitro and in vivo experiments indicates that DNMT3A mutations may trigger AML initiation through many unknown mechanisms rather than solely by the previously hypothesized loss-of-function. Intriguingly, our group discovered that DNMT3A mutations could specifically result in the upregulation of HOX genes (homeobox genes) and Meis1 (Meis homeobox 1), which are master regulators of HSC self-renewal [ 51, 56]. Although this finding might provide a reason why pre-LSCs bearing DNMT3A mutations exhibited enhanced self-renewal as mentioned earlier, why DNMT3A would particularly target a subset of genes rather than induce global alterations consistent with its role as a global DNA methylation modifier remains a mystery. Recently, the newly reported structural insight into the DNMT3A interaction with histone H3 may aid in tackling the conundrum of DNMT3A mutations by extending the investigations to the histone epigenetic apparatus [ 63]. Furthermore, mutations of TET and IDH were frequently reported by various groups, re-emphasizing the connection between aberrant DNA methylation levels and leukemogenesis [ 64, 65].

DNMT3A and other mutations in pre-LSCs

Recently, Shlush and colleagues have identified DNMT3A mutations as preleukemic mutations in acute leukemias [ 27]. To identify the original preleukemic mutations in HSCs, the authors compared the distribution of DNMT3A mutations with that of another supposed “founder” mutation, nucleophosmin-1 (NPM1), from the HSCs to the blast cells. The results indicated that although the NPM1 mutation can initiate AML in mouse models [ 66], it seems to be a subsequent alteration in most cases of AML that carry a driving DNMT3A mutation [ 67]. Interestingly, DNMT3A mutations are not the only verified preleukemic mutations; the TET methylcytosine dioxygenase 2 (TET2), and IDH 1 and 2 [ 9, 12, 68]. Notably, DNMT3A, TET2, and IDH1/2 are closely linked epigenetic modifiers in terms of DNA methylation and demethylation [ 64] (Fig. 3).

Particularly, IDH1/2 mutations can result in the abnormal accumulation of 2-hydroxyglutarate, which in turn impedes the DNA demethylation processes mediated by TET proteins [ 65, 69, 70]. The mutations of TET2 and IDH1/2 may affect the same pathway and hence exist mutually exclusively [ 71− 73]. The reason why epigenetic mutations are preferentially chosen as preleukemic mutations remains unknown, but this phenomenon might support the postulated model that primitive epigenetic abnormalities must accumulate in self-renewing HSCs unless a mutation confers self-renewal ability to a downstream cell, given that the mutations occurring in non-self-renewing cells are lost [ 74]. Another explanation might be that early mutations in epigenetic genes that affect hematopoietic self-renewal and/or differentiation are not sufficient to induce leukemogenesis on their own, thus making them the most primitive mutations in the preleukemic phase [ 75].

Notably, preleukemic mutations not only initiate AML; they also persist and survive chemotherapy in the evolution of the disease [ 28]. Shlush and colleagues found that DNMT3A-mutant HSCs, together with their progeny cells, persisted in peripheral blood despite AML remission following chemotherapy, suggesting that at least some of these preleukemic ancestral cells were resistant to treatment [ 27, 67]. Recent studies also found DNMT3A and TET2 as the most frequently involved somatic mutations in clonal expansions [ 33, 34]. The authors tested that during “clonal hematopoiesis,” which refers to HSC population dynamics preceding visible hematologic cancers, stem cells carrying a subset of driver mutations are capable of becoming resistant to chemotherapy. Subsequently, these cells acquire novel mutations, triggering a relapse. This hypothesis suggests that a span of many years is necessary for a clonal expansion to occur from a single hematopietic stem cell carrying driver mutations and additional numerous years before further evolution of the disease from preleukemic HSCs (Fig. 4). Interestingly, for clonal hematopoiesis, the DNMT3A R882 mutations most preferentially occur in patients not younger than 90 years, although it could also be found in individuals as young as 25 years [ 76]. This finding implies that DNMT3A mutations per se cause minimal effect on life span.

Summary and perspective

Destroying all LSCs, as well as pre-LSCs, is an ultimate goal in leukemia treatment. Therefore, understanding the genetic and epigenetic transformations during leukemic reprogramming should be given the highest priority to reach this end. Experimental evidence revealed that epigenetic processes may control leukemic reprogramming. Preleukemic mutations like DNMT3A, TET2, and IDH1/2 appear to tightly persist in AML patients from disease initiation to relapse. Therefore, targeting these preleukemic mutations in AML may be a promising strategy for treating leukemia and reducing relapse risk.

However, many issues remain. The target genes and pathways that DNMT3A mutations influence still need to be defined. Why TET2 functions as a DNA demethylation modifier opposite to DNMT3A but when mutated along with DNMT3A could equivalently induce pre-LSCs in AML is yet to be investigated. If DNMT3A and TET2 mutations target the same clusters of genes or pathways, they are expected to be equal in level in the primed stage. However, they do not exist mutually exclusively with each other, unlike TET2 and IDH1/2 mutations, hence entailing further study. Why epigenetic mutations are preferred to occur in the preleukemic stage must also be explored. We have hypothesized that pre-LSCs might sustain their clonal expansion advantages over normal HSCs through a greater number of rounds of self-renewal and a much longer survival that would allow subsequent mutations to occur [ 77, 78]. Therefore, the foremost mutations may function to protect the mutated HSCs from apoptosis, autophagy, and ultimate death. These various anti-death effects are likely difficult to be achieved by mutated genes specifically involved in proliferation or differentiation. Hence, the upper epigenetic mutations more probably hold a more extensive control of these variations, which causes class III mutations to frequently occur at the preleukemic stage. However, this theory still hardly illustrates why pre-LSCs are able to survive chemotherapy, which is a dramatically significant issue in AML treatment. Why mutations of histone modifiers were seldom found in pre-LSCs, considering that the anomaly of DNA methylation or histone modification could analogously promote epigenetic aberration in leukemogenesis remains to be studied [ 79].

Moreover, only a small proportion of leukemias are discussed in this review, which excluded those with cytogenetic abnormalities. Leukemias triggered by mixed-lineage leukemia translocations, t(9;22), and inv16 other than preleukemic mutations such as DNMT3A may not be effectively treated by targeting preleukemic lesions but by directly relieving the leukemic burden exerted by those cytogenetic abnormalities. The cell-of-origin in AML relapse may present another issue. The production of relapsed clones from the further evolution of a dominant clone or subclone outgrowth rather than the further evolution of a driving clone from a preleukemic HSC is a possibility worth investigating. In this regard, the clues and targets indicating cancer heterogeneity and evolution must be determined. Further studies are necessary to determine all the genetic and epigenetic crosstalks in leukemic reprogramming. Combining chemotherapy with promising regenerative medicine could facilitate better treatment for leukemias.

References

[1] Chen SJ, Shen Y, Chen Z. A panoramic view of acute myeloid leukemia. Nat Genet 2013; 45(6): 586–587

[2] Miller CA, Wilson RK, Ley TJ. Genomic landscapes and clonality of de novo AML. N Engl J Med 2013; 369(15): 1473

[3] Cancer Genome Atlas Research Network. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med 2013; 368(22): 2059–2074

[4] Steensma DP. The beginning of the end of the beginning in cancer genomics. N Engl J Med 2013; 368(22): 2138–2140

[5] Marchesi V. Genetics: the AML mutational landscape. Nat Rev Clin Oncol 2013; 10(6): 305

[6] Shih AH, Abdel-Wahab O, Patel JP, Levine RL. The role of mutations in epigenetic regulators in myeloid malignancies. Nat Rev Cancer 2012; 12(9): 599–612

[7] Hájková H, Marková J, Haškovec C, Sárová I, Fuchs O, Kostečka A, Cetkovský P, Michalová K, Schwarz J. Decreased DNA methylation in acute myeloid leukemia patients with DNMT3A mutations and prognostic implications of DNA methylation. Leuk Res 2012; 36(9): 1128–1133

[8] Macaluso M, Paggi MG, Giordano A. Genetic and epigenetic alterations as hallmarks of the intricate road to cancer. Oncogene 2003; 22(42): 6472–6478

[9] Chan SM, Majeti R. Role of DNMT3A, TET2, and IDH1/2 mutations in pre-leukemic stem cells in acute myeloid leukemia. Int J Hematol 2013; 98(6): 648–657

[10] Kreso A, Dick JE. Evolution of the cancer stem cell model. Cell Stem Cell 2014; 14(3): 275–291

[11] Ding L, Ley TJ, Larson DE, Miller CA, Koboldt DC, Welch JS, Ritchey JK, Young MA, Lamprecht T, McLellan MD, McMichael JF, Wallis JW, Lu C, Shen D, Harris CC, Dooling DJ, Fulton RS, Fulton LL, Chen K, Schmidt H, Kalicki-Veizer J, Magrini VJ, Cook L, McGrath SD, Vickery TL, Wendl MC, Heath S, Watson MA, Link DC, Tomasson MH, Shannon WD, Payton JE, Kulkarni S, Westervelt P, Walter MJ, Graubert TA, Mardis ER, Wilson RK, DiPersio JF. Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature 2012; 481(7382): 506–510

[12] Jan M, Snyder TM, Corces-Zimmerman MR, Vyas P, Weissman IL, Quake SR, Majeti R. Clonal evolution of preleukemic hematopoietic stem cells precedes human acute myeloid leukemia. Sci Transl Med 2012; 4(149): 149ra118

[13] Phillips RL, Ernst RE, Brunk B, Ivanova N, Mahan MA, Deanehan JK, Moore KA, Overton GC, Lemischka IR. The genetic program of hematopoietic stem cells. Science 2000; 288(5471): 1635–1640

[14] Ivanova NB, Dimos JT, Schaniel C, Hackney JA, Moore KA, Lemischka IR. A stem cell molecular signature. Science 2002; 298(5593): 601–604

[15] Bryder D, Rossi DJ, Weissman IL. Hematopoietic stem cells: the paradigmatic tissue-specific stem cell. Am J Pathol 2006; 169(2): 338–346

[16] Sun J, Ramos A, Chapman B, Johnnidis JB, Le L, Ho YJ, Klein A, Hofmann O, Camargo FD. Clonal dynamics of native haematopoiesis. Nature 2014; 514(7522): 322–327

[17] Pandolfi A, Barreyro L, Steidl U. Concise review: preleukemic stem cells: molecular biology and clinical implications of the precursors to leukemia stem cells. Stem Cells Transl Med 2013; 2(2): 143–150

[18] Warner JK, Wang JC, Hope KJ, Jin L, Dick JE. Concepts of human leukemic development. Oncogene 2004; 23(43): 7164–7177

[19] Dick JE. Acute myeloid leukemia stem cells. Ann N Y Acad Sci 2005; 1044(1): 1–5

[20] Luo L, Han ZC. Leukemia stem cells. Int J Hematol 2006; 84(2): 123–127

[21] Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J, Levine JE, Wang J, Hahn WC, Gilliland DG, Golub TR, Armstrong SA. Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature 2006; 442(7104): 818–822

[22] Krivtsov AV, Armstrong SA. MLL translocations, histone modifications and leukaemia stem-cell development. Nat Rev Cancer 2007; 7(11): 823–833

[23] Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, Caceres-Cortes J, Minden M, Paterson B, Caligiuri MA, Dick JE. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 1994; 367(6464): 645–648

[24] Larochelle A, Vormoor J, Hanenberg H, Wang JC, Bhatia M, Lapidot T, Moritz T, Murdoch B, Xiao XL, Kato I, Williams DA, Dick JE. Identification of primitive human hematopoietic cells capable of repopulating NOD/SCID mouse bone marrow: implications for gene therapy. Nat Med 1996; 2(12): 1329–1337

[25] Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 1997; 3(7): 730–737

[26] Holz-Schietinger C, Matje DM, Reich NO. Mutations in DNA methyltransferase (DNMT3A) observed in acute myeloid leukemia patients disrupt processive methylation. J Biol Chem 2012; 287(37): 30941–30951

[27] Shlush LI, Zandi S, Mitchell A, Chen WC, Brandwein JM, Gupta V, Kennedy JA, Schimmer AD, Schuh AC, Yee KW, McLeod JL, Doedens M, Medeiros JJ, Marke R, Kim HJ, Lee K, McPherson JD, Hudson TJ; HALT Pan-Leukemia Gene Panel Consortium, Brown AM, Yousif F, Trinh QM, Stein LD, Minden MD, Wang JC, Dick JE. Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature 2014; 506(7488): 328–333

[28] Corces-Zimmerman MR, Hong WJ, Weissman IL, Medeiros BC, Majeti R. Preleukemic mutations in human acute myeloid leukemia affect epigenetic regulators and persist in remission. Proc Natl Acad Sci USA 2014; 111(7): 2548–2553

[29] Jan M, Majeti R. Clonal evolution of acute leukemia genomes. Oncogene 2013; 32(2): 135–140

[30] Xue L, Pulikkan JA, Valk PJ, Castilla LH. NrasG12D oncoprotein inhibits apoptosis of preleukemic cells expressing Cbfβ-SMMHC via activation of MEK/ERK axis. Blood 2014; 124(3): 426–436

[31] Kuo YH, Landrette SF, Heilman SA, Perrat PN, Garrett L, Liu PP, Le Beau MM, Kogan SC, Castilla LH. Cbf beta-SMMHC induces distinct abnormal myeloid progenitors able to develop acute myeloid leukemia. Cancer Cell 2006; 9(1): 57–68

[32] Busque L, Patel JP, Figueroa ME, Vasanthakumar A, Provost S, Hamilou Z, Mollica L, Li J, Viale A, Heguy A, Hassimi M, Socci N, Bhatt PK, Gonen M, Mason CE, Melnick A, Godley LA, Brennan CW, Abdel-Wahab O, Levine RL. Recurrent somatic TET2 mutations in normal elderly individuals with clonal hematopoiesis. Nat Genet 2012; 44(11): 1179–1181

[33] Jaiswal S, Fontanillas P, Flannick J, Manning A, Grauman PV, Mar BG, Lindsley RC, Mermel CH, Burtt N, Chavez A, Higgins JM, Moltchanov V, Kuo FC, Kluk MJ, Henderson B, Kinnunen L, Koistinen HA, Ladenvall C, Getz G, Correa A, Banahan BF, Gabriel S, Kathiresan S, Stringham HM, McCarthy MI, Boehnke M, Tuomilehto J, Haiman C, Groop L, Atzmon G, Wilson JG, Neuberg D, Altshuler D, Ebert BL. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med 2014; 371(26): 2488–2498

[34] Genovese G, Kähler AK, Handsaker RE, Lindberg J, Rose SA, Bakhoum SF, Chambert K, Mick E, Neale BM, Fromer M, Purcell SM, Svantesson O, Landén M, Höglund M, Lehmann S, Gabriel SB, Moran JL, Lander ES, Sullivan PF, Sklar P, Grönberg H, Hultman CM, McCarroll SA. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med 2014; 371(26): 2477–2487

[35] Majeti R. Clonal evolution of pre-leukemic hematopoietic stem cells precedes human acute myeloid leukemia. Best Pract Res Clin Haematol 2014; 27(3−4): 229–234

[36] Huntly BJ, Gilliland DG. Leukaemia stem cells and the evolution of cancer-stem-cell research. Nat Rev Cancer 2005; 5(4): 311–321

[37] Wiseman DH, Greystoke BF, Somervaille TCP. The variety of leukemic stem cells in myeloid malignancy. Oncogene 2014; 33(24): 3091–3098

[38] Akashi K, Traver D, Miyamoto T, Weissman IL. A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 2000; 404(6774): 193–197

[39] Suzuki MM, Bird A. DNA methylation landscapes: provocative insights from epigenomics. Nat Rev Genet 2008; 9(6): 465–476

[40] Oki Y, Issa JP. Epigenetic mechanisms in AML — a target for therapy. Cancer Treat Res 2010; 145: 19–40

[41] Goll MG, Bestor TH. Eukaryotic cytosine methyltransferases. Annu Rev Biochem 2005; 74(1): 481–514

[42] Chan SW, Henderson IR, Jacobsen SE. Gardening the genome: DNA methylation in Arabidopsis thaliana. Nat Rev Genet 2005; 6(5): 351–360

[43] Klose RJ, Bird AP. Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci 2006; 31(2): 89–97

[44] Ehrlich M, Lacey M. DNA hypomethylation and hemimethylation in cancer. Adv Exp Med Biol 2013; 754: 31–56

[45] Solary E, Bernard OA, Tefferi A, Fuks F, Vainchenker W. The Ten-Eleven Translocation-2 (TET2) gene in hematopoiesis and hematopoietic diseases. Leukemia 2014; 28(3): 485–496

[46] Huang Y, Rao A. Connections between TET proteins and aberrant DNA modification in cancer. Trends Genet 2014; 30(10): 464–474

[47] Im AP, Sehgal AR, Carroll MP, Smith BD, Tefferi A, Johnson DE, Boyiadzis M. DNMT3A and IDH mutations in acute myeloid leukemia and other myeloid malignancies: associations with prognosis and potential treatment strategies. Leukemia 2014; 28(9): 1774–1783

[48] Ehrlich M. DNA hypomethylation in cancer cells. Epigenomics 2009; 1(2): 239–259

[49] Ehrlich M. DNA methylation in cancer: too much, but also too little. Oncogene 2002; 21(35): 5400–5413

[50] Hon GC, Hawkins RD, Caballero OL, Lo C, Lister R, Pelizzola M, Valsesia A, Ye Z, Kuan S, Edsall LE, Camargo AA, Stevenson BJ, Ecker JR, Bafna V, Strausberg RL, Simpson AJ, Ren B. Global DNA hypomethylation coupled to repressive chromatin domain formation and gene silencing in breast cancer. Genome Res 2012; 22(2): 246–258

[51] Xu J, Wang YY, Dai YJ, Zhang W, Zhang WN, Xiong SM, Gu ZH, Wang KK, Zeng R, Chen Z, Chen SJ. DNMT3A Arg882 mutation drives chronic myelomonocytic leukemia through disturbing gene expression/DNA methylation in hematopoietic cells. Proc Natl Acad Sci USA 2014; 111(7): 2620–2625

[52] Challen GA, Sun D, Jeong M, Luo M, Jelinek J, Berg JS, Bock C, Vasanthakumar A, Gu H, Xi Y, Liang S, Lu Y, Darlington GJ, Meissner A, Issa JP, Godley LA, Li W, Goodell MA. Dnmt3a is essential for hematopoietic stem cell differentiation. Nat Genet 2012; 44(1): 23–31

[53] Lund K, Cole JJ, VanderKraats ND, McBryan T, Pchelintsev NA, Clark W, Copland M, Edwards JR, Adams PD. DNMT inhibitors reverse a specific signature of aberrant promoter DNA methylation and associated gene silencing in AML. Genome Biol 2014; 15(8): 406

[54] Jjingo D, Conley AB, Yi SV, Lunyak VV, Jordan IK. On the presence and role of human gene-body DNA methylation. Oncotarget 2012; 3(4): 462–474

[55] Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE, Kandoth C, Payton JE, Baty J, Welch J, Harris CC, Lichti CF, Townsend RR, Fulton RS, Dooling DJ, Koboldt DC, Schmidt H, Zhang Q, Osborne JR, Lin L, O’Laughlin M, McMichael JF, Delehaunty KD, McGrath SD, Fulton LA, Magrini VJ, Vickery TL, Hundal J, Cook LL, Conyers JJ, Swift GW, Reed JP, Alldredge PA, Wylie T, Walker J, Kalicki J, Watson MA, Heath S, Shannon WD, Varghese N, Nagarajan R, Westervelt P, Tomasson MH, Link DC, Graubert TA, DiPersio JF, Mardis ER, Wilson RK. DNMT3A mutations in acute myeloid leukemia. N Engl J Med 2010; 363(25): 2424–2433

[56] Yan XJ, Xu J, Gu ZH, Pan CM, Lu G, Shen Y, Shi JY, Zhu YM, Tang L, Zhang XW, Liang WX, Mi JQ, Song HD, Li KQ, Chen Z, Chen SJ. Exome sequencing identifies somatic mutations of DNA methyltransferase gene DNMT3A in acute monocytic leukemia. Nat Genet 2011; 43(4): 309–315

[57] Roller A, Grossmann V, Bacher U, Poetzinger F, Weissmann S, Nadarajah N, Boeck L, Kern W, Haferlach C, Schnittger S, Haferlach T, Kohlmann A. Landmark analysis of DNMT3A mutations in hematological malignancies. Leukemia 2013; 27(7): 1573–1578

[58] Couronné L, Bastard C, Bernard OA. TET2 and DNMT3A mutations in human T-cell lymphoma. N Engl J Med 2012; 366(1): 95–96

[59] Patel JP, Gönen M, Figueroa ME, Fernandez H, Sun Z, Racevskis J, Van Vlierberghe P, Dolgalev I, Thomas S, Aminova O, Huberman K, Cheng J, Viale A, Socci ND, Heguy A, Cherry A, Vance G, Higgins RR, Ketterling RP, Gallagher RE, Litzow M, van den Brink MR, Lazarus HM, Rowe JM, Luger S, Ferrando A, Paietta E, Tallman MS, Melnick A, Abdel-Wahab O, Levine RL. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 2012; 366(12): 1079–1089

[60] Holz-Schietinger C, Matje DM, Reich NO. Mutations in DNA methyltransferase (DNMT3A) observed in acute myeloid leukemia patients disrupt processive methylation. J Biol Chem 2012; 287(37): 30941–30951

[61] Russler-Germain DA, Spencer DH, Young MA, Lamprecht TL, Miller CA, Fulton R, Meyer MR, Erdmann-Gilmore P, Townsend RR, Wilson RK, Ley TJ. The R882H DNMT3A mutation associated with AML dominantly inhibits wild-type DNMT3A by blocking its ability to form active tetramers. Cancer Cell 2014; 25(4): 442–454

[62] Mayle A, Yang L, Rodriguez B, Zhou T, Chang E, Curry CV, Challen GA, Li W, Wheeler D, Rebel VI, Goodell MA. Dnmt3a loss predisposes murine hematopoietic stem cells to malignant transformation. Blood 2015; 125(4): 629–638

[63] Guo X, Wang L, Li J, Ding Z, Xiao J, Yin X, He S, Shi P, Dong L, Li G, Tian C, Wang J, Cong Y, Xu Y. Structural insight into autoinhibition and histone H3-induced activation of DNMT3A. Nature 2015; 517(7536): 640–644

[64] McKenny AS, Levine RL. Isocitrate dehydrogenase mutations in leukemia. J Clin Invest 2013; 123(9): 3672–3677

[65] Gaidzik VI, Paschka P, Späth D, Habdank M, Köhne CH, Germing U, von Lilienfeld-Toal M, Held G, Horst HA, Haase D, Bentz M, Götze K, Döhner H, Schlenk RF, Bullinger L, Döhner K. TET2 mutations in acute myeloid leukemia (AML): results from a comprehensive genetic and clinical analysis of the AML study group. J Clin Oncol 2012; 30(12): 1350–1357

[66] Rakheja D, Konoplev S, Medeiros LJ, Chen W. IDH mutations in acute myeloid leukemia. Hum Pathol 2012; 43(10): 1541–1551

[67] Scourzic L, Mouly E, Bernard OA. TET proteins and the control of cytosine demethylation in cancer. Genome Med 2015; 7(1): 9

[68] Vassiliou GS, Cooper JL, Rad R, Li J, Rice S, Uren A, Rad L, Ellis P, Andrews R, Banerjee R, Grove C, Wang W, Liu P, Wright P, Arends M, Bradley A. Mutant nucleophosmin and cooperating pathways drive leukemia initiation and progression in mice. Nat Genet 2011; 43(5): 470–475

[69] Li KK, Luo LF, Shen Y, Xu J, Chen Z, Chen SJ. DNA methyltransferases in hematologic malignancies. Semin Hematol 2013; 50(1): 48–60

[70] Ye D, Xiong Y, Guan KL. The mechanisms of IDH mutations in tumorigenesis. Cell Res 2012; 22(7): 1102–1104

[71] Ye D, Ma S, Xiong Y, Guan KL. R-2-hydroxyglutarate as the key effector of IDH mutations promoting oncogenesis. Cancer Cell 2013; 23(3): 274–276

[72] Losman JA, Looper RE, Koivunen P, Lee S, Schneider RK, McMahon C, Cowley GS, Root DE, Ebert BL, Kaelin WG Jr. (R)-2-hydroxyglutarate is sufficient to promote leukemogenesis and its effects are reversible. Science 2013; 339(6127): 1621–1625

[73] Eifert C, Powers RS. From cancer genomes to oncogenic drivers, tumour dependencies and therapeutic targets. Nat Rev Cancer 2012; 12(8): 572–578

[74] Weissman I. Stem cell research: paths to cancer therapies and regenerative medicine. JAMA 2005; 294(11): 1359–1366

[75] Abdel-Wahab O, Levine RL. Mutations in epigenetic modifiers in the pathogenesis and therapy of acute myeloid leukemia. Blood 2013; 121(18): 3563–3572

[76] McKerrell T, Park N, Moreno T, Grove CS, Ponstingl H, Stephens J; Understanding Society Scientific Group, Crawley C, Craig J, Scott MA, Hodkinson C, Baxter J, Rad R, Forsyth DR, Quail MA, Zeggini E, Ouwehand W, Varela I, Vassiliou GS. Leukemia-associated somatic mutations drive distinct patterns of age-related clonal hemopoiesis. Cell Reports 2015; 10(8): 1239–1245

[77] Van Zant G, Liang Y. Concise review: hematopoietic stem cell aging, life span, and transplantation. Stem Cells Transl Med 2012; 1(9): 651–657

[78] Keller G, Snodgrass R. Life span of multipotential hematopoietic stem cells in vivo. J Exp Med 1990; 171(5): 1407–1418

[79] Eriksson A, Lennartsson A, Lehmann S. Epigenetic aberrations in acute myeloid leukemia: early key events during leukemogenesis. Exp Hematol 2015; 43(8): 609–624

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg