Oncogene (2015) 34, 3483–3492 © 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc

REVIEW Posttranslational modifications of RUNX1 as potential anticancer targets

S Goyama1, G Huang1, M Kurokawa2 and JC Mulloy1

The transcription factor RUNX1 is a master regulator of hematopoiesis. Disruption of RUNX1 activity has been implicated in the development of hematopoietic neoplasms. Recent studies also highlight the importance of RUNX1 in solid tumors both as a tumor promoter and a suppressor. Given its central role in cancer development, RUNX1 is an excellent candidate for targeted therapy. A potential strategy to target RUNX1 is through modulation of its posttranslational modifications (PTMs). Numerous studies have shown that RUNX1 activity is regulated by PTMs, including phosphorylation, acetylation, methylation and ubiquitination. These PTMs regulate RUNX1 activity either positively or negatively by altering RUNX1-mediated transcription, promoting degradation and affecting protein interactions. In this review, we first summarize the available data on the context- and dosage- dependent roles of RUNX1 in various types of neoplasms. We then provide a comprehensive overview of RUNX1 PTMs from biochemical and biologic perspectives. Finally, we discuss how aberrant PTMs of RUNX1 might contribute to tumorigenesis and also strategies to develop anticancer therapies targeting RUNX1 PTMs.

Oncogene (2015) 34, 3483–3492; doi:10.1038/onc.2014.305; published online 29 September 2014

INTRODUCTION recent studies have revealed the important role of RUNX1 in solid 8 RUNX1 is an essential transcription factor for the generation of tumors. Given its central roles for cell fate decisions in definitive hematopoietic stem cells and for hematopoietic hematopoiesis and leukemogenesis, RUNX1 is an attractive target differentiation to myeloid and lymphoid lineages (reviewed in to develop curative therapies for leukemia patients. Such thera- Link et al.1 and Kurokawa2). RUNX1 belongs to a family of pies may also be effective against several RUNX1-dependent solid transcriptional regulators called Runx, which contain a conserved tumors. However, developing therapies for transcription factors Runt domain responsible for sequence-specific DNA binding. such as RUNX1 has been a challenge. Previous attempts to The mammalian RUNX comprise three members: RUNX1, pharmacologically target RUNX1 have focused on the develop- – 9,10 RUNX2 and RUNX3. All RUNX proteins form a complex with ment of inhibitors that block RUNX1 CBFB interaction, but the cofactor CBFB. CBFB enhances the DNA-binding ability of the effects of such drugs have not been proven in clinical trials. In RUNX proteins and protects them from proteasomal degra- this review, we propose RUNX1 PTMs as alternative targets for dation, thereby increasing their function. RUNX1 also has CBFB- developing RUNX1-directed anticancer therapies. independent functions,3,4 and interacts with many other cofactors fi and chromatin modi ers, such as p300, SIN3A and histone RUNX1 IN HEMATOPOIETIC NEOPLASMS deacetylases (HDACs), to both activate and repress a broad range of hematopoietic . As a master regulator for hematopoiesis, RUNX1-mediated hematopoiesis has been the subject of intense RUNX1 function is tightly controlled. Several regulatory mechan- investigation. RUNX1 is involved in lineage commitment during isms exist for fine-tuning RUNX1 activity. These include alter- myeloid, B-cell and T-cell differentiation. In general, loss of RUNX1 function leads to impaired differentiation and subsequent leuke- native splicing, transcriptional control by two different promoters, 1,11 translational control and posttranslational modifications (PTMs).5,6 mia development. On the other hand, several recent studies described the survival role of RUNX1 in sustaining leukemia cell PTMs are a common strategy to regulate protein function in 12–14 diverse biologic processes. RUNX1 protein is modified by multiple growth. We here summarize context- and dosage-dependent PTMs (Figure 1). Many of these targeted residues are conserved in roles of RUNX1 in various types of hematopoietic neoplasms. RUNX2 and/or RUNX3, indicating the shared regulatory mechan- isms of PTMs for all three RUNX proteins (Figure 2; reviewed in Bae Myeloid neoplasms and Lee7). Among these modifications, phosphorylation, acetyla- The role of RUNX1 in myeloid neoplasms is well established. tion and methylation promote transcriptional activity of RUNX1. RUNX1 and its cofactor CBFB are frequent targets of chromosomal Protein stability of RUNX1 is regulated by phosphorylation and translocation in (AML). The chromosomal ubiquitination. These various modifications at multiple sites aberrations t(8;21) and inv(16) create RUNX1-RUNX1T1 (also called interdependently control RUNX1 function. AML1-ETO) and CBFB-MYH11 fusion genes, respectively. The AMLs Perturbation of RUNX1 function often leads to the development with RUNX1-RUNX1T1 or CBFB-MYH11 fusion genes, so-called of hematopoietic neoplasms. Beyond hematopoietic diseases, CBF-AML, are among the most common cytogenetic subtype of

1Division of Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, OH, USA and 2Department of Hematology and Oncology, Graduate School of Medicine, University of Tokyo, Tokyo, Japan. Correspondence: Dr JC Mulloy or Dr S Goyama, Division of Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine, Cincinnati, 45229 OH, USA. E-mail: [email protected] or [email protected] Received 10 July 2014; revised 13 August 2014; accepted 14 August 2014; published online 29 September 2014 PTM of RUNX1 as potential anticancer targets S Goyama et al 3484

Figure 1. PTMs of human RUNX1 (RUNX1b, NP_001001890.1). Numbers indicate positions of amino-acid residues from the N terminus. Runt domain is the DNA- and CBFB-binding domain. Activation domain is important for transcriptional activation. Ac, acetylation; APC, anaphase- promoting complex; K, lysine; Me, methylation; P, phosphorylation; R, arginine; S, serine; SCF, Skp1/Cullin/F-box protein complex; T, threonine; Ub, ubiquitination; Y, tyrosine.

Figure 2. Conservation of modification residues among RUNX proteins. Sequences of RUNX proteins (human RUNX1, RUNX2 and RUNX3) were aligned using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo). Many RUNX1 modification residues are conserved in either RUNX2 or RUNX3 (blue), or in both RUNX2 and RUNX3 (red).

Oncogene (2015) 3483 – 3492 © 2015 Macmillan Publishers Limited PTM of RUNX1 as potential anticancer targets S Goyama et al 3485

Figure 3. Regulation of RUNX1 function in hematopoietic neoplasms. In myeloid neoplasms, RUNX1 activity is attenuated through multiple mechanisms. RUNX1-RUNX1T1, RUNX1-MECOM and CBFB-MYH11 block the function of RUNX1. MLL-fusion proteins downregulate RUNX1 expression. Genetic mutations result in loss of function of RUNX1. Although the reduced RUNX1 activity leads to myeloid maturation block and subsequent disease development, further reduction of RUNX1 activity induces cell-cycle arrest and cell death. In lymphoid neoplasms, block of RUNX1 function by ETV6-RUNX1 contributes to the development of B-ALL, and loss-of-function mutations of RUNX1 are frequently found in T-ALL. RUNX1 function is delicately controlled in t(4;11) leukemia. ALL, acute lymphoid leukemia; CMML, chronic myelomonocytic leukemia; MDS, .

AML, being detected in 15–20% of adult de novo AML cases.15 transplantation model for MLL-fusion leukemia, Runx1 deficiency RUNX1-MECOM (also called AML1-EVI1) is another fusion led to compensatory upregulation of Runx2, and combined generated by the t(3;21) translocation that retains the N-terminal deletion of Runx1/Cbfb inhibited leukemogenesis.13 Others also half of RUNX1.16 RUNX1-MECOM is found in therapy-related showed the important role of RUNX1 in CBF-AMLs. Knock-in mice myeloid neoplasms, chronic myeloid leukemia with accelerated or expressing a mutant CBFB-MYH11 lacking the RUNX1 high-affinity blastic phase and, rarely, in de novo AML. These fusion genes binding domain developed leukemia quickly despite the ineffi- disrupt the functions of RUNX1/CBFB complex. Furthermore, cient suppression of RUNX1 function by this mutant leukemia RUNX1 is mutated frequently in various types of myeloid neo- oncogene.37 Similarly, a C terminally truncated form of RUNX1- 17 plasms. Familial platelet disorder with predisposition to AML is RUNX1T1, RUNX1-RUNX1T1-9a, possesses weaker repression 18,19 caused by germline mutations in RUNX1. Somatic RUNX1 activity for RUNX1-mediated transcription but has greater mutations have been found in 15% of cytogenetically normal leukemogenic potential in a mouse bone marrow transplantation 20–22 23–26 AML, 6–11% of myelodysplastic syndromes, 10% of model.38 Furthermore, RUNX1 knockdown in Kasumi-1 and ME-1 27 chronic myelomonocytic leukemia and 20% of systemic masto- cells, which harbor RUNX1-RUNX1T1 and CBFB-MYH11, respec- 28 cytosis. These mutations commonly involve the Runt domain or tively, attenuated the cell-cycle mitotic checkpoint and induced transcription activation domain, include frameshift and nonsense 12 29,30 apoptosis. mutations and lead to loss of transcriptional activity. Thus, RUNX1 activity is attenuated in many myeloid neoplasms Functionally, RUNX1 mutants lose the ability to promote early through multiple mechanisms. However, a continued low level of hematopoietic development31 and to induce myeloid matura- 13 RUNX activity is required to preserve the proliferative ability of tion. In addition to these genetic alterations, RUNX1 expression AML cells (Figure 3). is suppressed at the protein level in MLL-fusion leukemia.32 MLL- fusion proteins such as MLL-AF9, MLL-ENL and MLL-AF4 trigger loss of RUNX1/CBFB protein expression, an effect mediated by the Lymphoid neoplasms MLL CXXC domain and flanking region.33 Several mouse models The t(12;21) chromosomal translocation forming a fusion protein have also demonstrated the tumor suppressor role of RUNX1 in ETV6-RUNX1 (also called TEL-AML1) is observed in 25% of myeloid leukemia. Runx1-deficient cells showed increased sus- pediatric B-cell precursor acute lymphoblastic leukemia ceptibility to AML development in collaboration with MLL-ENL, (B-ALL).39,40 RUNX1 supports the development of B-cell-specified NRASG12S and EVI5.34–36 progenitors and the transition through the pre-B-cell stage.41,42 However, we recently showed that RUNX function is necessary ETV6-RUNX1 is thought to antagonize the function of wild-type to sustain the leukemogenic cell phenotype in CBF and MLL- RUNX1 by inducing multimerization and recruitment of transcrip- fusion leukemias. Suppression of RUNX1 function in human tional corepressors, thereby promoting ALL development.43 RUNX1-RUNX1T1- and MLL-AF9-expressing cord blood cells Paradoxically, however, wild-type RUNX1 is not only preserved induced cell-cycle arrest and apoptosis in these cells. In a mouse but is frequently amplified in ETV6-RUNX1 B-ALL patients,44,45

© 2015 Macmillan Publishers Limited Oncogene (2015) 3483 – 3492 PTM of RUNX1 as potential anticancer targets S Goyama et al 3486 suggesting that ETV6-RUNX1 B-ALL requires some RUNX1 activity in several epithelial cancers. RUNX1 is highly expressed in various for leukemia development. epithelial tumors including skin and head/neck squamous cell The t(4;11) translocation is a major cause of infant B-ALL. A carcinomas (SCCs).60 Genetic deletion of Runx1 in mouse inhibits recent paper showed the important role of RUNX1 in t(4;11) tumor formation in a murine model of chemically induced skin leukemia harboring MLL-AF4 and the reciprocal AF4-MLL-fusion SCC. Runx1 also supports the tumorigenesis of oral SCC induced proteins. MLL-AF4 transcriptionally upregulates RUNX1, and AF4- by oncogenic KrasG12D. In line with these observations, RUNX1 MLL collaborates with RUNX1 protein to activate target genes knockdown in human skin and head/neck SCC cells induces such as MEF2C.14 Interestingly, we recently showed that various growth arrest. Mechanistically, p21 (CIP1/WAF1) is upregulated in MLL-fusion proteins, including MLL-AF4, ‘downregulate’ protein Runx1-depleted mouse keratinocytes, and p21 depletion rescues expression of RUNX1/CBFB to enhance stem cell self-renewal.33 cell proliferation of Runx1-depleted cells in vitro.61 Additionally, These seemingly inconsistent observations likely indicate that Runx1 prevents expression of SOCS3 and SOCS4 transcription, RUNX1 activity is strictly regulated to promote optimal growth of t which leads to the activation of STAT3.60 The combination of p21 (4;11) leukemia cells. repression and STAT3 activation appears to underlie RUNX1- RUNX1 has essential roles in T-cell development,41,46 and recent mediated SCC formation. studies identified RUNX1 mutations in 15% of T-cell acute 47,48 lymphoblastic leukemia (T-ALL). As in the myeloid neoplasms, Other cancers these mutations are frequently found in the Runt or transcription Deletions of RUNX1 have been found in 15% of esophageal activation domain and are predicted to be deleterious. Experi- adenocarcinoma, and reintroduction of RUNX1 into the RUNX1- mental analyses also revealed a tumor suppressor role of RUNX1 in deleted esophageal adenocarcinoma cell line substantially inhib- T-cell transformation induced by TLX1/TLX3 or NOTCH1.49,50 On ited anchorage-independent growth.62,63 A mouse study showed the other hand, a recent study showed that a cyclin-dependent that RUNX1-deficient mice develop significantly more tumors in kinase 7 (CDK7) inhibitor THZ1 efficiently inhibits the growth of the colon and small intestine.64 Thus, RUNX1 is likely to be a tumor T-ALL cell lines presumably through RUNX1 downregulation, suppressor in these tumors. Furthermore, single-nucleotide indicating a possible growth-promoting role of RUNX1 in T-ALL.51 polymorphism studies have implicated a possible role for Runx1 The Runx1 gene is also a target for insertional mutagenesis in dysfunction in prostate, colon and rectal cancers.65,66 Functional T-cell lymphomas of mice carrying an MYC oncogene.52 consequences of these single-nucleotide polymorphisms need to Thus, impaired RUNX1-mediated B/T-cell differentiation caused be clarified experimentally. by chromosomal translocation or mutation underlies many lymphoid neoplasms. However, similar to myeloid neoplasms, it appears that most lymphoid leukemia cells need a certain level of PTMS OF RUNX1 RUNX1 activity for efficient growth and survival (Figure 3). RUNX1 can be a tumor suppressor or a tumor promoter depending on the context and its expression level. Modulating RUNX1 IN SOLID TUMORS RUNX1 activity, either by activation or inhibition, could offer a therapeutic opportunity in cancer. It has been shown that RUNX1 A number of studies have shown the role for RUNX genes in non- activity is modulated through multiple PTMs, including phosphor- hematopoietic cancers.11 In particular, RUNX3 has been described ylation, acetylation, methylation and ubiquitination. Therefore, as a tumor suppressor in epithelial cancers.53 Although RUNX1 has RUNX1 PTMs may be therapeutic targets that could impact RUNX1 been considered a hematopoietic gene, recent evidence has activity and thereby inhibit tumorigenesis. Table 1 summarizes the revealed a pivotal role for RUNX1 in solid tumors. RUNX1 effects of individual PTMs on RUNX1 function. mutations and single-nucleotide polymorphisms are associated with several types of cancers as described below. Furthermore, profiles of metastatic adenocarcinomas (lung, Phosphorylation breast, prostate, colorectal, uterus and ovary cancers) revealed Phosphorylation is often the first wave of PTMs in response to that RUNX1 is one of nine downregulated genes whose expression cellular stimuli. Phosphorylation-driven transcriptional activation pattern predicts metastasis.54 Thus, RUNX1 is likely to be involved and degradation of RUNX1 is one of the best-characterized in the development of various tissues and tumors. mechanisms to control RUNX1 activity. It was first reported that extracellular signal-regulated kinase (ERK), which is activated by Breast cancer hematopoietic cytokines and growth factors (e.g. interleukin-3, thrombopoietin, epidermal growth factor), phosphorylates RUNX1 RUNX1 is highly expressed in human breast epithelial cells at serine (S) 249 and S266 (In this review, we use amino-acid (reviewed in Chimge and Frenkel55), and recent studies identified numbering according to human RUNX1b (NP_001001890.1).), somatic mutations of RUNX1 and CBFB in 4% and 2% of breast thereby enhancing the transcriptional activity and transforming cancers, respectively.56,57 Both RUNX1 and CBFB mutations appear potential of RUNX in fibroblast cells.67 Another group reported to result in loss-of-function mutants. Experimentally, Ras-mediated that phorbol 12-myristate 13-acetate stimulation induces ERK- transformation of mammary epithelial cells (MCF10A) was mediated phosphorylation of five serine and threonine (T) associated with loss of RUNX1.58 Another report showed that residues located in the C-terminal region of RUNX1 (S249, S266, RUNX1 knockdown causes hyperproliferation and abnormal S276, S435 and T273).68 The ERK-dependent phosphorylation at morphogenesis in MCF10A cells. Interestingly, RUNX1 down- S249 and S266 disrupts the interaction of RUNX1 with the regulation is strongly associated with compensatory upregulation transcriptional corepressor SIN3A to enhance RUNX1-mediated of FOXO1, and RUNX1-depleted cells require normal FOXO func- transcription.69 RUNX1 phosphorylation also promotes its degra- tion for proliferation.59 Thus, the RUNX1/CBFB complex acts as a dation through the ubiquitin–proteasome pathway (described in tumor suppressor in breast cancers. The functional interaction detail below). between RUNX1 and FOXOs in tumorigenesis warrants further In addition, RUNX1 is phosphorylated by homeodomain- studies. interacting protein kinase 2 (HIPK2). HIPK2 phosphorylates RUNX1 at S249 and S276, and the phosphorylated RUNX1 then induces Skin and head/neck cancers phosphorylation of p300 and subsequent transcriptional activa- In contrast to the tumor suppressor role for RUNX1 in breast tion.70 Another group reported that RUNX1/CBFB/DNA complex cancers, RUNX1 appears to have survival/growth-promoting roles promotes the HIPK2-mediated phosphorylation of RUNX1 at S249,

Oncogene (2015) 3483 – 3492 © 2015 Macmillan Publishers Limited PTM of RUNX1 as potential anticancer targets S Goyama et al 3487

Table 1. RUNX1 posttranslational modifications

Site Modifying Biochemical and cellular functions Knock-in mice/primary cultures enzyme with Runx1( − / − ) cells

Phosphorylation (serine/threonine) S249, S266 ERK Promotes transcriptional activity, transforming Knock-in mice carring S249/S266 mutation potential (NIH3T3 cells), degradation and are phenotypicaly normal releases RUNX1 from Sin3A S249, S266, T273, ERK Promotes transcriptional activity An alanine mutant at S249/S266/T273/S276 S276, S435 has impaired T-cell differentiation activity. An alanine mutant at S249/S266/T273/S276/S435 has no function S249, S276, (T273) HIPK2 Promotes transcriptional activity and p300 Knock-in mice carring S249/S276 mutation phosphorylation are phenotypically normal S249, S276 CDK1, CDK2, CDK6 Promotes Cdc20-mediated degradation S21, S276, S397 CDK1, CDK6 Promotes transcriptional activity, reduces N/A interaction with HDACs, promotes proliferation of bone marrow cells and CBFB-MYH11-expressing Ba/F3 cells

Phosphorylation (tyrosine) Y254, Y258, Y260, SFKs Alters protein interaction with GATA1 and N/A Y376, Y378, Y379, Shp2 SWI/SNF complex, inhibits Runx1-mediated Mk Y386 and other Ys (dephosphorylation) matulation and and T-cell differentiation

Acetylation K24, K43 p300 Promotes DNA-binding ability, transcriptional An arginine mutant at K24/K43 has the ability activity and transforming potential (NIH3T3 to rescue the defective hematopoiesis of cells) Runx1-deficient P-Sp cells

Methylation R206, R210 and PRMT1 Releases RUNX1 from Sin3A and promotes N/A other Rs transcriptional activity R223 PRMT4 Increases interaction with DPF2, inhibits N/A miR-223 expression and inhibits myeloid differentiation of cord blood CD34+ cells

Ubiquitination K24, K43, K83, K90, APC, SCF, Promotes degradation An arginine mutant at K24/K43 has the ability K125, K144, K167, CHIP to rescue the defective hematopoiesis of K182, K188 Runx1-deficient P-Sp cells Abbreviations: APC, anaphase-promoting complex; CDK, cyclin-dependent kinase; CHIP, C terminus of HSC70 interacting protein; ERK, extracellular signal- regulated kinase; HDAC, histone deacetylase; HIPK, homeodomain-interacting protein kinase; K, lysine; Mk, megakaryocyte; N/A, not assessed; PRMT, protein arginine N-methyltransferase; P-Sp, para-aortic splanchnopleural; R, arginine; S, serine; SCF, Skp1/cullin/F-box protein complex; SFK, Src family kinase; T, threonine; Y, tyrosine.

S276 and T273, which in turn induces p300 phosphorylation.71 RUNX1 activity in megakaryocyte maturation and CD8 T-cell Thus, HIPK2-mediated RUNX1 phosphorylation activates p300, differentiation.76 leads to local histone acetylation, and upregulates target gene Taken together, RUNX1 phosphorylation at various serine/ expression. threonine residues has opposite effects on RUNX1 function; it Several CDKs also induce RUNX1 phosphorylation. RUNX1 is increases the transcriptional activity of RUNX1 while decreasing its phosphorylated by CDK1/2/6 at S249 and S276, and the protein stability, suggesting the presence of negative feedback phosphorylation promotes the anaphase-promoting complex regulation to keep RUNX1 activation transient. RUNX1 function is (APC)-mediated degradation of RUNX1.72 Others demonstrated also regulated by tyrosine phosphorylation, presumably through that CDK1 and CDK6 phosphorylate RUNX1 at S21, S276 and S397. altered protein–protein interactions. Phosphorylation at these three sites increases transactivation potency of RUNX1 by reducing interaction with HDAC1 and Acetylation HDAC3, and promotes proliferation of bone marrow progenitors Acetylation is another major PTM for not only histones but also 73,74 and CBFB-MYH11-expressing Ba/F3 cells. Furthermore, PIM1 nuclear proteins. It was shown that RUNX1 is acetylated at lysine kinase was shown to induce RUNX1 phosphorylation and enhance (K) 24 and K43 by p300. The acetylation significantly augments the its transactivation activity.75 DNA-binding activity, moderately enhances transcriptional activity In addition to serine/threonine residues, multiple tyrosine (Y) and promotes transforming capacity of RUNX1 in fibroblast cells.77 residues of RUNX1, including Y254, Y258, Y260, Y376, Y379, Y380 The fusion protein RUNX1-RUNX1T1 is also acetylated by p300 at and Y387, are phosphorylated by Src family kinases and are K24 and K43 in leukemia cells isolated from patients.78 In addi- dephosphorylated by SHP2. The tyrosine phosphorylation alters tion to p300, other transcriptional coactivators, MOZ and CBP, key protein–protein interactions of RUNX1, such as interactions can acetylate RUNX1 in vitro.79 However, whether the MOZ/ with GATA1 and SWI/SNF complex, and negatively regulates CBP-mediated acetylation actually occurs in cells is not clear.

© 2015 Macmillan Publishers Limited Oncogene (2015) 3483 – 3492 PTM of RUNX1 as potential anticancer targets S Goyama et al 3488 The deacetylation process for RUNX1 has not been defined. deficient mutations at either S249/S266 or S249/S276 were Although RUNX1 is known to interact with HDACs, whether phenotypically normal.89 In line with this, we showed that HDACs deacetylate specific lysine residues of RUNX1 is currently RUNX1-2A, carrying two alanine (phospho-deficient) mutations unknown. at S249/S266, retained normal RUNX1 function to induce early hematopoietic development and T-cell maturation in primary fi 90 – Methylation culture assays using murine Runx1-de cient cells. RUNX1 4A, carrying four mutations at S249, S266, T273 and S276, showed Protein methylation is a type of reversible PTM that has an impaired T-cell maturation activity but still retained the ability to important role in many cellular processes. Several lines of rescue the defective early hematopoiesis of Runx1-deficient para- evidence suggest that arginine (R) methylation of RUNX1 is aortic splanchnopleural (P-Sp) cells. Of note, RUNX1-5A, carrying important for its transcriptional activity. PRMT1, an arginine five mutations at S249, S266, T273, S276 and S435, completely lost methyltransferase targeting histone H4R3, methylates RUNX1 at its hematopoietic activity in the murine primary culture assays.90 multiple residues. The PRMT1-dependent methylation of RUNX1 at We also confirmed the phosphorylation-dependent regulation of R206 and R210 abrogates its association with SIN3A, and enhances RUNX1 activity using a human cord blood cell culture assay. transcriptional activity of RUNX1. The arginine-methylated RUNX1 RUNX1 and RUNX1-2A induced myeloid maturation and inhibited efficiently binds to the promoters of RUNX1 target genes, such as the growth of cord blood cells, whereas RUNX1-5A completely lost CD41 and PU.1, and upregulates their expression.80 PRMT1 also these abilities. RUNX1–4A showed an intermediate effect to methylates RUNX1 and RUNX1-RUNX1T1 at R142 within the induce myeloid maturation in cord blood cells.91 These results ‘SGRGK’ sequence that is found at the N-terminal tail of histone indicate that multiple phosphorylation sites are involved in the H4.80,81 Another arginine methyltransferase PRMT4 methylates regulation of RUNX1 activity. RUNX1 at R223. Interestingly, PRMT4-mediated RUNX1 methyla- The primary culture assay using Runx1-deficient cells also tion promotes the assembly of a DPF2-containing corepressive showed that mutant RUNX1 with two lysine-to-arginine mutations complex. The methyl–RUNX1–DPF2 complex inhibits miR-223 at K24/K43 had the ability to rescue the defective hematopoiesis expression and myeloid differentiation of human cord blood of Runx1-deficient P-Sp cells.31 Therefore, acetylation and CD34+ cells.82 PRMT6 was also shown to interact with RUNX1 to ubiquitination at K24/K43 appear to be dispensable for RUNX1 mediate histone H3 arginine-2 dimethylation.83 However, whether function to promote early hematopoiesis. The physiologic PRMT6 promotes the methylation of RUNX1 itself is not known. relevance of other RUNX1 PTMs remains largely unknown. Thus, multiple arginine methyltransferases are involved in the control of RUNX1 methylation and RUNX1-mediated transcription. Several groups showed the interaction between RUNX1 and a CROSSTALK AMONG MULTIPLE RUNX1 PTMS lysine methyltransferase SUV39H1.84,85 Although this interaction It has become increasingly apparent that PTMs work in concert, affects RUNX1-mediated gene regulation, lysine methylation of and the crosstalk among various modifications determines the RUNX1 in cells has not been demonstrated. final biologic output. This interdependent nature likely explains why RUNX1 proteins substituted at a few sites of modification Ubiquitination show only subtle effects on physiologic functions of RUNX1. As is RUNX1 is an unstable protein and is subjected to proteolytic described above, RUNX1 is phosphorylated at multiple serine/ degradation mediated by the ubiquitin–proteasome pathway. threonine sites. Interestingly, each phosphorylation may affect Multiple lysine residues (K24, K43, K83, K90, K125, K144, K167, phosphorylation of neighboring serine/threonine residues. A K182 and K188) are likely targets of ubiquitination in RUNX1.86 In report showed that mutations of S266/T273 to alanine in RUNX1 line with this, a recent study showed that alternative splicing resulted in increased phosphorylation at S249/S276, whereas isoforms of mouse Runx1 have various protein stabilities depend- serine-to-alanine mutations at S276/S303 prevented phosphoryla- ing on the numbers of lysine residues in them.87 The lysine tion at S266/T273, suggesting a complex interaction among these residues of RUNX1 cluster within or around the Runt domain, phosphorylation sites.72 through which CBFB interacts with RUNX1, and heterodimeriza- Crosstalk can exist between different types of modifications. It tion with CBFB protects RUNX1 from degradation.86 Similar to has been shown by a number of groups that serine/threonine CBFB, a histone methyltransferase MLL binds to RUNX1 through phosphorylation of RUNX1 promotes protein degradation through the Runt domain and prevents RUNX1 degradation by reducing its the ubiquitin–proteasome pathway.69,72,92 Mutant RUNX1 protein polyubiquitination.32 Conversely, MLL-fusion proteins downregu- with phospho-deficient mutations is more stable than wild-type late RUNX1/CBFB expression, although the downregulation RUNX1, and is resistant to degradation mediated by Cdc20-APC appears to be caused partly by ubiquitination-independent and Skp2-SCF ubiquitin ligase complexes.72 Interestingly, RUNX1 mechanisms.33 Several E3-ubiquitin ligases were also shown phosphorylation also activates p300 function,71 and p300 was to induce ubiquitination-dependent degradation of RUNX1. shown to acetylate RUNX1.77 It is therefore likely that phosphor- APC and SCF complex can degrade RUNX1 in a cell-cycle and ylation triggers both ubiquitination and acetylation at lysine phosphorylation-dependent manner.72 Another report identified residues in RUNX1, thereby fine-tuning RUNX1 function. Given STUB1 (also called CHIP) as an E3-ubiquitin ligase that promotes that K24/K43 are common targets of acetylation and ubiquitina- RUNX1 degradation.88 However, none of these reports determined tion, a direct competition between these two modifications could the actual ubiquitination acceptor sites in RUNX1 by these E3- affect RUNX1 activity and stability. In addition, both serine/ ubiquitin ligases. Furthermore, how these E3 ligases are recruited threonine phosphorylation and arginine methylation of RUNX1 to RUNX1 remains to be elucidated. promote dissociation with SIN3A,69,80 suggesting a functional interplay between these modifications. Finally, SHP2 may act as a molecular switch to determine the phosphorylation status of PHYSIOLOGIC RELEVANCE OF RUNX1 PTMS RUNX1.93 SHP2 was shown to dephosphorylate tyrosine residues As most studies described above were performed in vitro using of RUNX1,76 and is known to activate the ERK pathway. Given that cell-based and biochemical assays, the relevance of the obtained ERK is a major RUNX1 kinase at serine/threonine residues,67 SHP2 results to physiologic functions of RUNX1 is an important issue. is likely to promote serine/threonine phosphorylation while Indeed, the predicted RUNX1 regulation by phosphorylation was inhibiting tyrosine phosphorylation of RUNX1. Thus, multiple challenged by a study using knock-in mice. Unexpectedly, mice PTMs interact with each other and create functional networks to expressing mutant RUNX1 protein that harbored phospho- regulate RUNX1 activity (Figure 4).

Oncogene (2015) 3483 – 3492 © 2015 Macmillan Publishers Limited PTM of RUNX1 as potential anticancer targets S Goyama et al 3489

Figure 4. Crosstalk among multiple modifications of RUNX1. Phosphorylation, acetylation and methylation promote transcriptional activity of RUNX1, whereas phosphorylation and ubiquitination promote degradation of RUNX1. Serine/threonine phosphorylation induces ubiqutination and potentially also induces acetylation through p300 activation. Both serine/threonine phosphorylation and arginine methylation induce dissociation of SIN3A. Shp2 may act as a molecular switch to regulate phosphorylation status of RUNX1 by inhibiting tyrosine phosphorylation and promoting serine/threonine phosphorylation through ERK activation. Ac, acetylation; Me, methylation; P, phosphorylation; S/T, serine/threonine; Ub, ubiquitination; Y, tyrosine.

Table 2. Missense mutations of RUNX1 at modification sites in AML and T-ALL patients

Papers Disease Number of missense mutation of RUNX1 Number of PTM site mutation of RUNX1 Mutations

Tang et al.22 AML 29 2 K83Q, K83R Migas et al.95 AML/MDS/MPN 11 0 — Gaidzik et al.94 AML 19 0 — Greif et al.20 AML 9 0 — Grossmann et al.96 T-ALL 6 0 — Zhang et al.48 T-ALL 2 0 — Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; MPN, myeloproliferative neoplasm; PTM, posttranslational modification.

DISEASE RELEVANCE OF RUNX1 PTMS of RUNX1. The loss of C-terminal modification sites could alter the To examine whether dysregulation of RUNX1 PTM could have a role of the remaining PTM residues. Of note, recent findings causative role in the development of human disease, we suggest that acetylation at K24/K43 is essential for the leukemo- investigated recent large-scale sequencing studies to determine genic activity of RUNX1-RUNX1T1, whereas it is dispensable for the the frequency of missense mutations of RUNX1 at known hematopoietic activity of RUNX1. RUNX1-RUNX1T1 is acetylated fi by p300, and inhibiting p300 activity abrogates RUNX1-RUNX1T1 modi cation sites (shown in Figure 1) in myeloid and T-cell 78 leukemia.20,22,48,94–96 As shown in Table 2, RUNX1 mutations at the acetylation to impair its leukemogenicity. In contrast, an acetylation-deficient RUNX1 mutant exhibits normal hematopoie- PTM sites are extremely rare, except for K83, which is one of the tic activity.31 Arginine methylation is also likely to have different putative ubiquitination acceptor sites. The mutation at K83 was roles in RUNX1 and RUNX1-RUNX1T1: it is important for binding of also reported previously in patients with relapsed acute promye- 80 SIN3A to RUNX1, but not to RUNX1-RUNX1T1. CBFB-MYH11 was locytic leukemia and familial platelet disorder with predisposition 97,98 shown to prevent phosphorylation of RUNX1 and p300 by to AML. However, the K83 mutations may contribute to sequestering HIPK2 to the cytoplasm in transient transfection leukemogenesis not through the loss of ubiquitination, but rather 71 98 assays. Interestingly, however, another report showed that by disrupting DNA binding of RUNX1. The infrequency of RUNX1 RUNX1 is actually phosphorylated in leukemia cells derived from mutations at PTM sites probably indicates functional redundancy CBFB-MYH11 knock-in mouse and a patient with inv(16).37 Again, among multiple PTMs. Because of the compensation by other this discrepancy probably indicates tight regulation of RUNX1 modifications, loss of a few modifications will not affect RUNX1 phosphorylation and activity in AML cells. Furthermore, MLL- activity, and will not perturb hematopoietic development. fusion proteins induce abnormal RUNX1 ubiquitination, and lead Although PTM sites of RUNX1 are not frequent targets of to the downregulation of RUNX1 expression through both mutation in patients, RUNX1 seems to be modified aberrantly in ubiquitination-dependent and -independent mechanisms.32,33 hematopoietic neoplasms. RUNX1-RUNX1T1 lacks several phos- Thus, leukemia-associated fusion proteins have effects on RUNX1 phorylation and methylation sites present in the C-terminal region modifications, which may contribute to leukemogenesis.

© 2015 Macmillan Publishers Limited Oncogene (2015) 3483 – 3492 PTM of RUNX1 as potential anticancer targets S Goyama et al 3490 Disease-associated mutations of RUNX1 could also affect the Small molecules to prevent interaction between RUNX1 and RUNX1 PTMs. It was shown that DNA and CBFB binding of RUNX1 modifying enzymes may specifically modulate RUNX1 function provoke HIPK2-mediated phosphorylation of RUNX1 and p300.71 without perturbing enzymatic activities. Because most N-terminal RUNX1 mutants are defective for either DNA or CBFB binding, they are likely to be less phosphorylated and acetylated in cells. Similar to RUNX1-RUNX1T1, several phos- CONCLUDING REMARKS phorylation and methylation sites are lost in C-terminal truncation RUNX1 has a key role in various types of cancers both as a tumor types of RUNX1 mutants, which will definitely change the PTM suppressor or a tumor promoter in a dosage-dependent manner. status. The status and functions of RUNX1 PTMs in normal and Cancer-associated aberration in RUNX1-mediated transcription is a malignant hematopoiesis needs to be clarified in future studies. promising drug target. Given that RUNX1 activity is tightly and reversibly regulated by multiple PTMs, PTM-targeted drugs have potential in treating RUNX1-related neoplasms. FUTURE CHALLENGE: THERAPEUTIC TARGETING OF RUNX1 Early studies defined a putative role for individual modifications PTMS of RUNX1, under the assumption that each PTM had specific Although transcription factors such as RUNX1 were traditionally definable functions. It is now clear that PTMs work interdepen- considered as undruggable, targeting them is becoming a realistic dently, affecting one another on a number of levels, to fine tune 99 option with recent technological advances. For RUNX1, several RUNX1 activity. Owing to these compensatory mechanisms inhibitors that block RUNX1–CBFB interaction have been devel- between the different modifications, the results of in vivo genetic oped and showed growth-inhibitory effects on CBF leukemia cells studies often contradict those previously obtained in biochemical 9,10,13 in vitro and in mouse leukemia models. However, the clinical assays. In the future, proteomic survey by mass spectrometry will effects of these inhibitors have not been proven. Furthermore, continue to identify new modification sites in RUNX1. We should recent reports have shown the CBFB-independent function of focus on combinational effects of multiple modifications, rather fi 4 3 RUNX1 in zebra sh hematopoiesis and DNA repair. We believe than the specific role of each modification. It is also important to RUNX1 PTMs are alternative targets to modulate RUNX1 activity. A determine biologic consequences of individual and multiple simple strategy to suppress RUNX1 function is through enhancing modifications ideally using a mouse knock-in strategy. Primary RUNX1 ubiquitination and protein degradation. This strategy will culture assays using Runx1-deficient cells will be an easy and be effective for leukemias that require some RUNX1 activity for useful in vitro counterpart for this purpose.31,104 In addition, more their survival as well as for RUNX1-dependent skin and head/neck knowledge about upstream signaling pathways controlling RUNX1 cancers. Agents to induce RUNX1 degradation should also target PTMs (reviewed in Bae and Lee7) will help us to understand how the oncogenic RUNX1-RUNX1T1 and mutant RUNX1 proteins. fi fi RUNX1 function is ne tuned in diverse biologic processes. Such methods have shown great ef cacy for the similar Despite the tremendous progress that has been achieved, we oncoprotein PML-RARA (promyelocytic leukemia-retinoic acid are still at an early stage in understanding the complex receptor alpha) that causes acute promyelocytic leukemia. It was fi crossregulation of RUNX1 PTMs. We will gain additional insight shown that As2O3 binds to the PML moiety and speci cally to solve the crosstalk puzzle of PTMs by comparing modifications induces a SUMO-dependent, ubiquitin-mediated degradation of of wild-type RUNX1 and that of disease-related RUNX1 mutants. PML-RARA.100,101 Strikingly, acute promyelocytic leukemia patients Such efforts will ultimately lead to the development of novel treated with the combination of As O and all-trans retinoic acid 2 3 anticancer therapies targeting RUNX1 PTMs or possibly more now reach 90% cure rate.102 Experimentally, As O -induced PML 2 3 broadly to include all RUNX family members. degradation increased the efficacy of antileukemic therapy in a mouse model for chronic myeloid leukemia promoted by BCR- ABL.103 Thus, the proteolysis-based strategy for inhibiting RUNX1 CONFLICT OF INTEREST function may have promise in treating RUNX1-related cancers. The authors declare no conflict of interest. Conversely, enhancement of RUNX1 activity by disrupting the interaction between RUNX1 and the ubiquitin ligases that induce RUNX1 degradation is an alternative strategy that is likely to REFERENCES perturb cellular homeostasis. Agents with such ability may be 1 Link KA, Chou FS, Mulloy JC. Core binding factor at the crossroads: determining effective in treating RUNX1-low cancers including MLL-fusion the fate of the HSC. J Cell Physiol 2010; 222:50–56. 33 leukemia. More detailed information on the regulation of RUNX1 2 Kurokawa M. AML1/Runx1 as a versatile regulator of hematopoiesis: regulation protein stability will be necessary to develop such ubiquitination- of its function and a role in adult hematopoiesis. Int J Hematol 2006; 84: based therapies. 136–142. Altered PMTs in oncogenic RUNX1 fusions/mutants may provide 3 Wang CQ, Krishnan V, Tay LS, Chin DW, Koh CP, Chooi JY et al. Disruption of therapeutic opportunities. As described above, K24/K43 acetyla- Runx1 and Runx3 leads to bone marrow failure and leukemia predisposition due tion is essential for leukemogenic activity of RUNX1-RUNX1T1 but to transcriptional and DNA repair defects. Cell Rep 2014; 8:767–782. not for hematopoietic activity of wild-type RUNX1. This differential 4 Bresciani E, Carrington B, Wincovitch S, Jones M, Gore AV, Weinstein BM et al. CBFbeta and RUNX1 are required at 2 different steps during the development of dependency on the K24/K43 acetylation implies that lysine hematopoietic stem cells in zebrafish. Blood 2014; 124:70–78. acetyltransferases represent a potential therapeutic target to 5 Levanon D, Groner Y. Structure and regulated expression of mammalian inhibit selectively the growth of RUNX1-RUNX1T1 leukemia cells RUNX genes. Oncogene 2004; 23: 4211–4219. 78 while preserving normal hematopoiesis. Phosphorylation is also 6 Wang L, Huang G, Zhao X, Hatlen MA, Vu L, Liu F et al. Post-translational an interesting target to modulate RUNX1 activity in cancer cells. It modifications of Runx1 regulate its activity in the cell. Blood Cells Mol Dis 2009; seems that RUNX1 phosphorylation is delicately controlled in inv 43:30–34. (16) leukemia37,71 and potentially in other cancers. Therefore, 7 Bae SC, Lee YH. Phosphorylation, acetylation and ubiquitination: the molecular modulating the phosphorylation status could promote substantial basis of RUNX regulation. Gene 2006; 366:58–66. anticancer effects. Thus, RUNX1 PTMs are promising drug targets 8 Scheitz CJ, Tumbar T. New insights into the role of Runx1 in epithelial stem cell biology and pathology. J Cell Biochem 2013; 114:985–993. for anticancer therapeutics. Given the high conservation of key 9 Gorczynski MJ, Grembecka J, Zhou Y, Kong Y, Roudaia L, Douvas MG et al. PTM residues among RUNX proteins (Figure 2), therapies targeting Allosteric inhibition of the protein-protein interaction between the leukemia- RUNX1 PTMs should also have effects on RUNX2/RUNX3 and the associated proteins Runx1 and CBFbeta. Chem Biol 2007; 14: 1186–1197. corresponding cancers. Challenges remain to develop specific 10 Cunningham L, Finckbeiner S, Hyde RK, Southall N, Marugan J, Yedavalli VR et al. drugs targeting only RUNX1 PTMs with little off-target effects. Identification of benzodiazepine Ro5-3335 as an inhibitor of CBF leukemia

Oncogene (2015) 3483 – 3492 © 2015 Macmillan Publishers Limited PTM of RUNX1 as potential anticancer targets S Goyama et al 3491 through quantitative high throughput screen against RUNX1-CBFbeta interac- 34 Motoda L, Osato M, Yamashita N, Jacob B, Chen LQ, Yanagida M et al. Runx1 tion. Proc Natl Acad Sci USA 2012; 109: 14592–14597. protects hematopoietic stem/progenitor cells from oncogenic insult. Stem Cells 11 Blyth K, Cameron ER, Neil JC. The RUNX genes: gain or loss of function in cancer. 2007; 25: 2976–2986. Nat Rev Cancer 2005; 5:376–387. 35 Jacob B, Osato M, Yamashita N, Wang CQ, Taniuchi I, Littman DR et al. Stem cell 12 Ben-Ami O, Friedman D, Leshkowitz D, Goldenberg D, Orlovsky K, Pencovich N exhaustion due to Runx1 deficiency is prevented by Evi5 activation in leuke- et al. Addiction of t(8;21) and inv(16) acute myeloid leukemia to native RUNX1. mogenesis. Blood 2010; 115: 1610–1620. Cell Rep 2013; 4: 1131–1143. 36 Nishimoto N, Arai S, Ichikawa M, Nakagawa M, Goyama S, Kumano K et al. Loss of 13 Goyama S, Schibler J, Cunningham L, Zhang Y, Rao Y, Nishimoto N et al. Tran- AML1/Runx1 accelerates the development of MLL-ENL leukemia through down- scription factor RUNX1 promotes survival of acute myeloid leukemia cells. J Clin regulation of p19ARF. Blood 2011; 118:2541–2550. Invest 2013; 123: 3876–3888. 37 Kamikubo Y, Zhao L, Wunderlich M, Corpora T, Hyde RK, Paul TA et al. Accel- 14 Wilkinson AC, Ballabio E, Geng H, North P, Tapia M, Kerry J et al. RUNX1 is a key erated leukemogenesis by truncated CBF beta-SMMHC defective in high-affinity target in t(4;11) leukemias that contributes to gene activation through an AF4- binding with RUNX1. Cancer Cell 2010; 17:455–468. MLL complex interaction. Cell Rep 2013; 3:116–127. 38 Yan M, Kanbe E, Peterson LF, Boyapati A, Miao Y, Wang Y et al. A previously 15 Goyama S, Mulloy JC. Molecular pathogenesis of core binding factor leukemia: unidentified alternatively spliced isoform of t(8;21) transcript promotes leuke- current knowledge and future prospects. Int J Hematol 2011; 94: 126–133. mogenesis. Nat Med 2006; 12:945–949. 16 Mitani K, Ogawa S, Tanaka T, Miyoshi H, Kurokawa M, Mano H et al. Generation of 39 Romana SP, Mauchauffe M, Le Coniat M, Chumakov I, Le Paslier D, Berger R et al. the AML1-EVI-1 fusion gene in the t(3;21)(q26;q22) causes blastic crisis in chronic The t(12;21) of acute lymphoblastic leukemia results in a tel-AML1 gene fusion. myelocytic leukemia. EMBO J 1994; 13: 504–510. Blood 1995; 85: 3662–3670. 17 Osato M. Point mutations in the RUNX1/AML1 gene: another actor in RUNX 40 Golub TR, Barker GF, Bohlander SK, Hiebert SW, Ward DC, Bray-Ward P et al. leukemia. Oncogene 2004; 23: 4284–4296. Fusion of the TEL gene on 12p13 to the AML1 gene on 21q22 in acute lym- 18 Song WJ, Sullivan MG, Legare RD, Hutchings S, Tan X, Kufrin D et al. Hap- phoblastic leukemia. Proc Natl Acad Sci USA 1995; 92:4917–4921. loinsufficiency of CBFA2 causes familial thrombocytopenia with propensity to 41 Ichikawa M, Asai T, Saito T, Seo S, Yamazaki I, Yamagata T et al. AML-1 is required develop acute myelogenous leukaemia. Nat Genet 1999; 23:166–175. for megakaryocytic maturation and lymphocytic differentiation, but not for 19 Owen CJ, Toze CL, Koochin A, Forrest DL, Smith CA, Stevens JM et al. Five new maintenance of hematopoietic stem cells in adult hematopoiesis. Nat Med 2004; pedigrees with inherited RUNX1 mutations causing familial platelet disorder 10:299–304. with propensity to myeloid malignancy. Blood 2008; 112: 4639–4645. 42 Niebuhr B, Kriebitzsch N, Fischer M, Behrens K, Gunther T, Alawi M et al. Runx1 is 20 Greif PA, Konstandin NP, Metzeler KH, Herold T, Pasalic Z, Ksienzyk B et al. RUNX1 essential at two stages of early murine B-cell development. Blood 2013; 122: mutations in cytogenetically normal acute myeloid leukemia are associated with 413–423. a poor prognosis and up-regulation of lymphoid genes. Haematologica 2012; 97: 43 Zelent A, Greaves M, Enver T. Role of the TEL-AML1 fusion gene in the molecular 1909–1915. pathogenesis of childhood acute lymphoblastic leukaemia. Oncogene 2004; 23: 21 Mendler JH, Maharry K, Radmacher MD, Mrozek K, Becker H, Metzeler KH et al. 4275–4283. RUNX1 mutations are associated with poor outcome in younger and 44 Ma SK, Wan TS, Cheuk AT, Fung LF, Chan GC, Chan SY et al. Characterization of older patients with cytogenetically normal acute myeloid leukemia and with additional genetic events in childhood acute lymphoblastic leukemia with TEL/ distinct gene and microRNA expression signatures. J Clin Oncol 2012; 30: AML1 gene fusion: a molecular cytogenetics study. Leukemia 2001; 15: 3109–3118. 1442–1447. 22 Tang JL, Hou HA, Chen CY, Liu CY, Chou WC, Tseng MH et al. AML1/RUNX1 45 Peter A, Heiden T, Taube T, Korner G, Seeger K. Interphase FISH on TEL/AML1 mutations in 470 adult patients with de novo acute myeloid leukemia: prog- positive acute lymphoblastic leukemia relapses--analysis of clinical relevance of nostic implication and interaction with other gene alterations. Blood 2009; 114: additional TEL and AML1 copy number changes. Eur J Haematol 2009; 83: 5352–5361. 420–432. 23 Haferlach T, Nagata Y, Grossmann V, Okuno Y, Bacher U, Nagae G et al. Land- 46 Taniuchi I, Osato M, Egawa T, Sunshine MJ, Bae SC, Komori T et al. Differential scape of genetic lesions in 944 patients with myelodysplastic syndromes. Leu- requirements for Runx proteins in CD4 repression and epigenetic silencing kemia 2013; 28: 241–247. during T lymphocyte development. Cell 2002; 111:621–633. 24 Rocquain J, Carbuccia N, Trouplin V, Raynaud S, Murati A, Nezri M et al. Com- 47 Grossmann V, Haferlach C, Weissmann S, Roller A, Schindela S, Poetzinger F et al. bined mutations of ASXL1, CBL, FLT3, IDH1, IDH2, JAK2, KRAS, NPM1, NRAS, The molecular profile of adult T-cell acute lymphoblastic leukemia: mutations in RUNX1, TET2 and WT1 genes in myelodysplastic syndromes and acute myeloid RUNX1 and DNMT3A are associated with poor prognosis in T-ALL. Genes Chro- leukemias. BMC Cancer 2010; 10: 401. mosomes Cancer 2013; 52:410–422. 25 Bejar R, Stevenson K, Abdel-Wahab O, Galili N, Nilsson B, Garcia-Manero G et al. 48 Zhang J, Ding L, Holmfeldt L, Wu G, Heatley SL, Payne-Turner D et al. The genetic Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med basis of early T-cell precursor acute lymphoblastic leukaemia. Nature 2012; 481: 2011; 364: 2496–2506. 157–163. 26 Haferlach T, Nagata Y, Grossmann V, Okuno Y, Bacher U, Nagae G et al. Land- 49 Giambra V, Jenkins CR, Wang H, Lam SH, Shevchuk OO, Nemirovsky O et al. scape of genetic lesions in 944 patients with myelodysplastic syndromes. Leu- NOTCH1 promotes T cell leukemia-initiating activity by RUNX-mediated kemia 2014; 28: 241–247. regulation of PKC-theta and reactive oxygen species. Nat Med 2012; 18: 27 Itzykson R, Kosmider O, Renneville A, Gelsi-Boyer V, Meggendorfer M, Morabito 1693–1698. M et al. Prognostic score including gene mutations in chronic myelomonocytic 50 Della Gatta G, Palomero T, Perez-Garcia A, Ambesi-Impiombato A, Bansal M, leukemia. J Clin Oncol 2013; 31: 2428–2436. Carpenter ZW et al. Reverse engineering of TLX oncogenic transcriptional net- 28 Schwaab J, Schnittger S, Sotlar K, Walz C, Fabarius A, Pfirrmann M et al. Com- works identifies RUNX1 as tumor suppressor in T-ALL. Nat Med 2012; 18: prehensive mutational profiling in advanced systemic mastocytosis. Blood 2013; 436–440. 122: 2460–2466. 51 Kwiatkowski N, Zhang T, Rahl PB, Abraham BJ, Reddy J, Ficarro SB et al. Targeting 29 Harada H, Harada Y, Niimi H, Kyo T, Kimura A, Inaba T. High incidence of somatic transcription regulation in cancer with a covalent CDK7 inhibitor. Nature 2014; mutations in the AML1/RUNX1 gene in myelodysplastic syndrome and low blast 511:616–620. percentage myeloid leukemia with myelodysplasia. Blood 2004; 103: 2316–2324. 52 Wotton S, Stewart M, Blyth K, Vaillant F, Kilbey A, Neil JC et al. Proviral insertion 30 Imai Y, Kurokawa M, Izutsu K, Hangaishi A, Takeuchi K, Maki K et al. Mutations of indicates a dominant oncogenic role for Runx1/AML-1 in T-cell lymphoma. the AML1 gene in myelodysplastic syndrome and their functional implications in Cancer Res 2002; 62: 7181–7185. leukemogenesis. Blood 2000; 96: 3154–3160. 53 Chuang LS, Ito Y. RUNX3 is multifunctional in carcinogenesis of multiple 31 Goyama S, Yamaguchi Y, Imai Y, Kawazu M, Nakagawa M, Asai T et al. The solid tumors. Oncogene 2010; 29: 2605–2615. transcriptionally active form of AML1 is required for hematopoietic rescue of the 54 Ramaswamy S, Ross KN, Lander ES, Golub TR. A molecular signature of metas- AML1-deficient embryonic para-aortic splanchnopleural (P-Sp) region. Blood tasis in primary solid tumors. Nat Genet 2003; 33:49–54. 2004; 104: 3558–3564. 55 Chimge NO, Frenkel B. The RUNX family in breast cancer: relationships with 32 Huang G, Zhao X, Wang L, Elf S, Xu H, Zhao X et al. The ability of MLL to bind estrogen signaling. Oncogene 2013; 32:2121–2130. RUNX1 and methylate H3K4 at PU.1 regulatory regions is impaired by MDS/AML- 56 Ellis MJ, Ding L, Shen D, Luo J, Suman VJ, Wallis JW et al. Whole-genome analysis associated RUNX1/AML1 mutations. Blood 2011; 118: 6544–6552. informs breast cancer response to aromatase inhibition. Nature 2012; 486: 33 Zhao X, Chen A, Yan X, Zhang Y, He F, Hayashi Y et al. Down-regulation of 353–360. RUNX1/CBFbeta by MLL fusion proteins enhances HSC self-renewal. Blood 2014; 57 Cancer Genome Atlas N. Comprehensive molecular portraits of human breast 123: 1729–1738. tumours. Nature 2012; 490:61–70.

© 2015 Macmillan Publishers Limited Oncogene (2015) 3483 – 3492 PTM of RUNX1 as potential anticancer targets S Goyama et al 3492 58 Kadota M, Yang HH, Gomez B, Sato M, Clifford RJ, Meerzaman D et al. Deli- 81 Shia WJ, Okumura AJ, Yan M, Sarkeshik A, Lo MC, Matsuura S et al. PRMT1 neating genetic alterations for tumor progression in the MCF10A series of breast interacts with AML1-ETO to promote its transcriptional activation and progenitor cancer cell lines. PLoS One 2010; 5: e9201. cell proliferative potential. Blood 2012; 119:4953–4962. 59 Wang L, Brugge JS, Janes KA. Intersection of FOXO- and RUNX1-mediated gene 82 Vu LP, Perna F, Wang L, Voza F, Figueroa ME, Tempst P et al. PRMT4 blocks expression programs in single breast epithelial cells during morphogenesis and myeloid differentiation by assembling a methyl-RUNX1-dependent repressor tumor progression. Proc Natl Acad Sci USA 2011; 108:E803–E812. complex. Cell Rep 2013; 5: 1625–1638. 60 Scheitz CJ, Lee TS, McDermitt DJ, Tumbar T. Defining a tissue stem cell-driven 83 Herglotz J, Kuvardina ON, Kolodziej S, Kumar A, Hussong H, Grez M et al. Histone Runx1/Stat3 signalling axis in epithelial cancer. EMBO J 2012; 31: 4124–4139. arginine methylation keeps RUNX1 target genes in an intermediate state. 61 Hoi CS, Lee SE, Lu SY, McDermitt DJ, Osorio KM, Piskun CM et al. Runx1 directly Oncogene 2013; 32:2565–2575. promotes proliferation of hair follicle stem cells and epithelial tumor formation 84 Chakraborty S, Sinha KK, Senyuk V, Nucifora G. SUV39H1 interacts with AML1 in mouse skin. Mol Cell Biol 2010; 30: 2518–2536. and abrogates AML1 transactivity. AML1 is methylated in vivo. Oncogene 2003; 62 Gu J, Ajani JA, Hawk ET, Ye Y, Lee JH, Bhutani MS et al. Genome-wide catalogue 22: 5229–5237. of chromosomal aberrations in barrett's esophagus and esophageal adeno- 85 Reed-Inderbitzin E, Moreno-Miralles I, Vanden-Eynden SK, Xie J, Lutterbach B, carcinoma: a high-density single nucleotide polymorphism array analysis. Cancer Durst-Goodwin KL et al. RUNX1 associates with histone deacetylases and Prev Res 2010; 3: 1176–1186. SUV39H1 to repress transcription. Oncogene 2006; 25: 5777–5786. 63 Dulak AM, Schumacher SE, van Lieshout J, Imamura Y, Fox C, Shim B et al. 86 Huang G, Shigesada K, Ito K, Wee HJ, Yokomizo T, Ito Y. Dimerization with – Gastrointestinal adenocarcinomas of the esophagus, stomach, and colon exhibit PEBP2beta protects RUNX1/AML1 from ubiquitin proteasome-mediated degra- 20 – distinct patterns of genome instability and oncogenesis. Cancer Res 2012; 72: dation. EMBO J 2001; :723 733. 4383–4393. 87 Komeno Y, Yan M, Matsuura S, Lam K, Lo MC, Huang YJ et al. Runx1 exon 64 Fijneman RJ, Anderson RA, Richards E, Liu J, Tijssen M, Meijer GA et al. Runx1 is a 6-related alternative splicing isoforms differentially regulate hematopoiesis 123 – tumor suppressor gene in the mouse gastrointestinal tract. Cancer Sci 2012; 103: in mice. Blood 2014; : 3760 3769. 593–599. 88 Shang Y, Zhao X, Xu X, Xin H, Li X, Zhai Y et al. CHIP functions an E3 ubiquitin 386 – 65 Huang SP, Lan YH, Lu TL, Pao JB, Chang TY, Lee HZ et al. Clinical significance of ligase of Runx1. Biochem Biophys Res Commun 2009; : 242 246. runt-related transcription factor 1 polymorphism in prostate cancer. BJU Int 89 Tachibana M, Tezuka C, Muroi S, Nishimoto S, Katsumoto T, Nakajima A et al. 2011; 107: 486–492. Phosphorylation of Runx1 at Ser249, Ser266, and Ser276 is dispensable for bone 66 Slattery ML, Lundgreen A, Herrick JS, Caan BJ, Potter JD, Wolff RK. Associations marrow hematopoiesis and thymocyte differentiation. Biochem Biophys Res 368 – between genetic variation in RUNX1, RUNX2, RUNX3, MAPK1 and eIF4E and Commun 2008; :536 542. 90 Yoshimi M, Goyama S, Kawazu M, Nakagawa M, Ichikawa M, Imai Y et al. Multiple riskof colon and rectal cancer: additional support for a TGF-beta-signaling phosphorylation sites are important for RUNX1 activity in early hematopoiesis pathway. Carcinogenesis 2011; 32:318–326. and T-cell differentiation. Eur J Immunol 2012; 42: 1044–1050. 67 Tanaka T, Kurokawa M, Ueki K, Tanaka K, Imai Y, Mitani K et al. The extracellular 91 Goyama S, Schibler J, Cunningham L, Zhang Y, Rao Y, Nishimoto N et al. Tran- signal-regulated kinase pathway phosphorylates AML1, an acute myeloid leu- scription factor RUNX1 promotes survival of acute myeloid leukemia cells. J Clin kemia gene product, and potentially regulates its transactivation ability. Mol Cell Invest 2013; 123: 3876–3888. Biol 1996; 16: 3967–3979. 92 Biggs JR, Zhang Y, Peterson LF, Garcia M, Zhang DE, Kraft AS. Phosphorylation of 68 Zhang Y, Biggs JR, Kraft AS. Phorbol ester treatment of K562 cells regulates the AML1/RUNX1 regulates its degradation and nuclear matrix association. Mol transcriptional activity of AML1c through phosphorylation. J Biol Chem 2004; Cancer Res 2005; 3:391–401. 279: 53116–53125. 93 Neel BG, Speck NA. Tyrosyl phosphorylation toggles a Runx1 switch. Genes Dev 69 Imai Y, Kurokawa M, Yamaguchi Y, Izutsu K, Nitta E, Mitani K et al. The cor- 2012; 26:1520–1526. epressor mSin3A regulates phosphorylation-induced activation, intranuclear 94 Gaidzik VI, Bullinger L, Schlenk RF, Zimmermann AS, Rock J, Paschka P et al. location, and stability of AML1. Mol Cell Biol 2004; 24:1033–1043. RUNX1 mutations in acute myeloid leukemia: results from a comprehensive 70 Aikawa Y, Nguyen LA, Isono K, Takakura N, Tagata Y, Schmitz ML et al. Roles of genetic and clinical analysis from the AML study group. J Clin Oncol 2011; 29: HIPK1 and HIPK2 in AML1- and p300-dependent transcription, hematopoiesis 1364–1372. and blood vessel formation. EMBO J 2006; 25:3955–3965. 95 Migas A, Savva N, Mishkova O, Aleinikova OV. AML1/RUNX1 gene point 71 Wee HJ, Voon DC, Bae SC, Ito Y. PEBP2-beta/CBF-beta-dependent phosphor- mutations in childhood myeloid malignancies. Pediatr Blood Cancer 2011; 57: ylation of RUNX1 and p300 by HIPK2: implications for leukemogenesis. Blood 583–587. 112 – 2008; : 3777 3787. 96 Grossmann V, Kern W, Harbich S, Alpermann T, Jeromin S, Schnittger S et al. 72 Biggs JR, Peterson LF, Zhang Y, Kraft AS, Zhang DE. AML1/RUNX1 phosphor- Prognostic relevance of RUNX1 mutations in T-cell acute lymphoblastic ylation by cyclin-dependent kinases regulates the degradation of AML1/RUNX1 leukemia. Haematologica 2011; 96: 1874–1877. 26 – by the anaphase-promoting complex. Mol Cell Biol 2006; : 7420 7429. 97 Osato M, Asou N, Abdalla E, Hoshino K, Yamasaki H, Okubo T et al. Biallelic and 73 Guo H, Friedman AD. Phosphorylation of RUNX1 by cyclin-dependent kinase heterozygous point mutations in the runt domain of the AML1/PEBP2alphaB 286 – reduces direct interaction with HDAC1 and HDAC3. JBiolChem2011; :208 215. gene associated with myeloblastic leukemias. Blood 1999; 93:1817–1824. 74 Zhang L, Fried FB, Guo H, Friedman AD. Cyclin-dependent kinase phosphor- 98 Michaud J, Wu F, Osato M, Cottles GM, Yanagida M, Asou N et al. In vitro analyses ylation of RUNX1/AML1 on 3 sites increases transactivation potency and sti- of known and novel RUNX1/AML1 mutations in dominant familial platelet dis- 111 – mulates cell proliferation. Blood 2008; : 1193 1200. order with predisposition to acute myelogenous leukemia: implications for 75 Aho TL, Sandholm J, Peltola KJ, Ito Y, Koskinen PJ. Pim-1 kinase phosphorylates mechanisms of pathogenesis. Blood 2002; 99:1364–1372. RUNX family transcription factors and enhances their activity. BMC Cell Biol 2006; 99 Yan C, Higgins PJ. Drugging the undruggable: transcription therapy for cancer. 7: 21. Biochim Biophys Acta 2013; 1835:76–85. 76 Huang H, Woo AJ, Waldon Z, Schindler Y, Moran TB, Zhu HH et al. A Src family 100 Tatham MH, Geoffroy MC, Shen L, Plechanovova A, Hattersley N, Jaffray EG et al. kinase-Shp2 axis controls RUNX1 activity in megakaryocyte and T-lymphocyte RNF4 is a poly-SUMO-specific E3 ubiquitin ligase required for arsenic-induced differentiation. Genes Dev 2012; 26: 1587–1601. PML degradation. Nat Cell Biol 2008; 10:538–546. 77 Yamaguchi Y, Kurokawa M, Imai Y, Izutsu K, Asai T, Ichikawa M et al. AML1 is 101 Lallemand-Breitenbach V, Jeanne M, Benhenda S, Nasr R, Lei M, Peres L et al. functionally regulated through p300-mediated acetylation on specific lysine Arsenic degrades PML or PML-RARalpha through a SUMO-triggered RNF4/ubi- residues. J Biol Chem 2004; 279: 15630–15638. quitin-mediated pathway. Nat Cell Biol 2008; 10: 547–555. 78 Wang L, Gural A, Sun XJ, Zhao X, Perna F, Huang G et al. The leukemogenicity of 102 Lo-Coco F, Avvisati G, Vignetti M, Thiede C, Orlando SM, Iacobelli S et al. Retinoic AML1-ETO is dependent on site-specific lysine acetylation. Science 2011; 333: acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med 2013; 765–769. 369:111–121. 79 Kitabayashi I, Aikawa Y, Nguyen LA, Yokoyama A, Ohki M. Activation of AML1- 103 Ito K, Bernardi R, Morotti A, Matsuoka S, Saglio G, Ikeda Y et al. PML targeting mediated transcription by MOZ and inhibition by the MOZ-CBP fusion protein. eradicates quiescent leukaemia-initiating cells. Nature 2008; 453:1072–1078. EMBO J 2001; 20: 7184–7196. 104 Kawazu M, Asai T, Ichikawa M, Yamamoto G, Saito T, Goyama S et al. Functional 80 Zhao X, Jankovic V, Gural A, Huang G, Pardanani A, Menendez S et al. Methy- domains of Runx1 are differentially required for CD4 repression, TCRbeta lation of RUNX1 by PRMT1 abrogates SIN3A binding and potentiates its tran- expression, and CD4/8 double-negative to CD4/8 double-positive transition in scriptional activity. Genes Dev 2008; 22:640–653. thymocyte development. J Immunol 2005; 174: 3526–3533.

Oncogene (2015) 3483 – 3492 © 2015 Macmillan Publishers Limited