Mir-139-5P Controls Translation in Myeloid Leukemia Through EIF4G2

ORIGINAL ARTICLE miR-139-5p controls translation in myeloid leukemia through EIF4G2

S Emmrich1,9, F Engeland1,9, M El-Khatib1, K Henke1, A Obulkasim2, J Schöning1, JE Katsman-Kuipers2, C Michel Zwaan2, A Pich3, J Stary4, A Baruchel5, V de Haas6, D Reinhardt7, M Fornerod2,8, MM van den Heuvel-Eibrink2 and JH Klusmann1

MicroRNAs (miRNAs) are crucial components of homeostatic and developmental gene regulation. In turn, dysregulation of miRNA expression is a common feature of different types of cancer, which can be harnessed therapeutically. Here we identify miR-139-5p suppression across several cytogenetically deﬁned acute myeloid leukemia (AML) subgroups. The promoter of mir-139 was transcriptionally silenced and could be reactivated by histone deacetylase inhibitors in a dose-dependent manner. Restoration of mir-139 expression in cell lines representing the major AML subgroups (t[8;21], inv[16], mixed lineage leukemia-rearranged and complex karyotype AML) caused cell cycle arrest and apoptosis in vitro and in xenograft mouse models in vivo. During normal hematopoiesis, mir-139 is exclusively expressed in terminally differentiated neutrophils and macrophages. Ectopic expression of mir-139 repressed proliferation of normal CD34+-hematopoietic stem and progenitor cells and perturbed myelomonocytic in vitro differentiation. Mechanistically, mir-139 exerts its effects by repressing the translation initiation factor EIF4G2, thereby reducing overall protein synthesis while speciﬁcally inducing the translation of cell cycle inhibitor p27Kip1. Knockdown of EIF4G2 recapitulated the effects of mir-139, whereas restoring EIF4G2 expression rescued the mir-139 phenotype. Moreover, elevated miR-139-5p expression is associated with a favorable outcome in a cohort of 165 pediatric patients with AML. Thus, mir-139 acts as a global tumor suppressor-miR in AML by controlling protein translation. As AML cells are dependent on high protein synthesis rates controlling the expression of mir-139 constitutes a novel path for the treatment of AML.

Oncogene (2016) 35, 1822–1831; doi:10.1038/onc.2015.247; published online 13 July 2015

INTRODUCTION of mir-139 is associated with a favorable prognosis in hepatocellular Hematopoietic stem and progenitor cells (HSPCs), which acquire carcinoma and deletion of DLEU2/mir-15a/16-1 cluster, a gate- mutations conferring a maturation arrest and clonal proliferation, keeper of B-cell proliferation, leads to chronic lymphocytic represent the origin of acute myeloid leukemia (AML).1 Over the leukemia.9,10 In AML, downregulation of mir-181 indicated poor last decades, improved understanding of the underlying mole- prognosis in cytogenetically abnormal karyotypes.11 cular basis and the characterization of cytogenetically defined Therefore, it is not surprising that several recent reports subgroups led to the development of effective therapeutic demonstrated deregulation of mRNA translation to promote 12,13 strategies and to an improved outcome.2 Taking this into account, cellular transformation and a malignant phenotype. Translation the new World Health Organization classification scheme for regulation is primarily achieved during initiation, which requires myeloid neoplasms classifies the disease according to cytogenetic the protein complex known as eukaryotic initiation factor 4 F criteria including the chromosomal translocations t(8;21)(q22;q22) (EIF4F), consisting of three proteins: cap-binding protein EIF4E, (RUNX1-RUNX1T1), inv(16)(p13q22) (CBF-MYH11), t(15;17)(q22; scaffolding protein EIF4G and ATP-dependent RNA helicase q21) (PML-RARα), rearrangements of the mixed lineage leukemia EIF4A.14 The aberrant expression of EIF4F components, such as – (MLL) gene, and molecular changes in WT1, NPM1, CEBPA and EIF4E and EIF4G, has been linked to leukemia progression.15 17 DNMT3A.2,3 A series of recent studies has demonstrated the Therefore, targeting translation initiation emerged as a new importance of microRNAs (miRNAs) for blood homeostasis and the paradigm for AML therapy.18 This can be achieved by disrupting emergence of pediatric AML.4–6 MiRNAs are posttranscriptional the initiation complex using the 4EGI-1 compound, anti-EIF4E regulators, which control gene expression by modifying the antisense oligonucleotides, the antiviral drug ribavirin or by mRNA stability or translational initiation/elongation.7 Deregulation second-generation mammalian target of rapamycin inhibitors of miRNA-expression has been shown for solid tumors and (TORKinibs). In mammalian cells, the eukaryotic translation hematopoietic malignancies, indicating the importance of post- initiation factor 4, gamma (EIF4G) is present as three homologs, transcriptional deregulation for malignant transformation or EIF4G1, EIF4G2 (alias: death associated protein 5, DAP5; p97) and maintenance of malignant growth.8 For example, overexpression EIF4G3.19–21 The homologs are biochemically similar, but have

1Department of Pediatric Hematology and Oncology, Hannover Medical School, Hannover, Germany; 2Department of Pediatric Oncology/Hematology, Erasmus MC/Sophia Children’s Hospital, Rotterdam, The Netherlands; 3Institute for Toxicology, Hannover Medical School, Hannover, Germany; 4Department of Pediatric Hematology/Oncology, Charles University and University Hospital Motol, Prague, Czech Republic; 5Department of Hematology, Hopitaux universitaires Saint-Louis, St Louis Hospital, Paris, France; 6Dutch Childhood Oncology Group (DCOG), Stichting Kinderoncologie Nederland (SKION), Hague, The Netherlands; 7Clinic for Pediatrics III, University Hospital Essen, Essen, Germany and 8Department of Biochemistry, Erasmus MC/Sophia Children’s Hospital, Rotterdam, The Netherlands. Correspondence: Dr JH Klusmann, Department of Pediatric Hematology and Oncology, Hannover Medical School, Carl-Neuberg-Street 1, Hannover D-30625, Germany. E-mail: [email protected] 9These authors equally contributed to this work. Received 5 December 2014; revised 21 May 2015; accepted 22 May 2015; published online 13 July 2015 mir-139 controls translation S Emmrich et al 1823 overlapping as well as divergent functional roles.21,22 Although In conclusion, mir-139 reduces proliferation, colony formation, elevated levels of EIF4G1 were shown to correlate with a cell viability and cell cycle activity in a broad range of AML cell malignant cell transformation,15,16,23 little is known about EIF4G2 lines representing the major AML subgroups. expression and its function in translation initiation. Here we describe the function of the epigenetically regulated Restoration of mir-139 expression has an anti-leukemic effect microRNA-139-5p that suppresses leukemic growth in AML cell in vivo lines in vitro and in vivo by targeting EIF4G2. By disrupting the To evaluate the anti-leukemic effect of mir-139 in vivo,we EIF4F initiation complex, mir-139 interferes globally with transla- performed xenotransplantation experiments using KASUMI-1 cells tional initiation and thereby reduces overall protein synthesis. with (8;21). As the t(8;21) cell context tolerates the highest amount of endogenous miR-139-5p (Figure 1a), using this cell line would RESULTS exclude artificial effects introduced by ectopic miRNA expression in an aberrant cellular context. In a competitive transplantation MiR-139-5p is suppressed in pediatric AML assay, NRG mice were intrafemorally (i.f.) transplanted with mir- fi Our recent miRNA expression pro ling in a cohort of pediatric AML 139-transduced (GFP+) or miR-NSC(ctrl)-transduced (mCherry+) cases followed by a gain-of-function screening indicated a putative 4 cells mixed together in a ratio of 1:1 (Figure 2a). The ratio shifted role of mir-139 as miRNA tumor suppressor. MiR-139-5p was to 1:17 (GFP+:mCherry+) in the bone marrow of the recipients after highly expressed in human terminally differentiated, post-mitotic 28 days, pointing toward a strong growth disadvantage conferred macrophages and neutrophils (22% and 11% of RNU48, respec- – by mir-139. In a second competitive transplantation setting, mice tively), but was low to absent in pediatric AML (0.01 0.24% of were transplanted with mir-139-transduced (GFP+) or miR-NSC RNU48) and normal HSPCs (0.18% of RNU48; Figure 1a). (ctrl)-transduced (GFP+) cells mixed together in a ratio of 1:1. A similar expression pattern can be found in normal mouse Quantification of the genomic copy number of vector-derived hematopoietic cells (Supplementary Figure 1a). In patients with miR-NSC- or mir-139 expression cassettes by quantitative PCR MLL-rearrangements, t(7;12), t(15;17) or normal karyotype miR-139- 27 days post transplantation again showed a 17.5-fold shift of the 5p was mainly absent. In cases with the t(8;21) and inv(16) AML mir-139:miR-NSC ratio at the favor of miR-NSC (Figure 2b; for (generally referred to as core-binding factor leukemias, CBF AML) as + quantitative PCR validation, see Supplementary Figures 2a and b). well as in normal CD34 -HSPCs, low levels of miR-139-5p were In a third assay, mice were intrafemorally transplanted with either detected (Figure 1a). CBF AMLs generally represent an AML 24 miR-NSC (ctrl) or mir-139-transduced KASUMI-1 cells and sub- subgroup with a favorable outcome. Comparing the expression jected to survival analysis (Figure 2c). mir-139 conferred a survival levels of both mature miRNAs strands confirmed miR-139-5p to be benefit to the transplanted mice (PMantel–Cox = 0.022) with an the guide and miR-139-3p to be the passenger strand increase of the median survival from 22.5 to 107 days. (Supplementary Figure 1b). The expression data suggested that Hence, mir-139 confers a growth disadvantage to transduced miR-139-5p is upregulated in terminally differentiated myelomono- KASUMI-1 cells in vivo leading to an increased overall survival of cytic cells to control proliferation, whereas it is suppressed in xenotransplanted recipients. normal and malignant stem and progenitor cells. Thus, we hypothesized that elevating the expression of mir-139 expression in leukemic cells would exert tumor-suppressive properties. mir-139 expression in human HSPCs suppresses growth and differentiation along the granulocytic and monocytic lineages Enforced mir-139 expression suppresses proliferation and colony To test if the effects of mir-139 are restricted to leukemic cells or if formation in AML mir-139 controls proliferation during normal hematopoiesis, we transduced and characterized human CD34+-HSPCs. During To elucidate the function of mir-139 as a tumor suppressor, we used 4,25 monocytic in vitro differentiation, mir-139 reduced the cell a previously described lentiviral expression system to ectopically number by 3.5-fold (Figure 3a). Moreover, monocytic differentia- express mir-139 in KASUMI-1 and SKNO-1 cells carrying the t(8;21) + + tion was impaired, as quantified by the fraction of CD11b /CD14 translocation and ME-1 cells with inv(16) rearrangement. Further, we monocytes in flow cytometry and the smaller cell size detected transduced THP-1 with MLL-rearrangement (MLL-AF9), OCI/AML3 by forward scatter/side scatter properties (Figure 3b and with a complex karyotype and HL-60 as a model for promyelocytic Supplementary Figure 3a). For neutrophilic in vitro differentiation, leukemia. Similar to the primary patient samples, miR-139-5p and the number of mir-139-transduced cells was fourfold lower on day -3p were largely suppressed throughout all cell lines (Figure 1a and 15 (Figure 3c) and significantly fewer CD13+/CD15+ granulocytes Supplementary Figure 1b) and forced expression resulted in were detected in culture (Figure 3d). Accordingly, in the forward miR-139-5p expression levels between 15 and 500% of RNU48 scatter/side scatter, the typical population of large highly (Supplementary Figure 1c), which is within a broad physiological granulated neutrophils was lost upon mir-139 expression spectrum as compared with more abundant endogenous (Supplementary Figure 3b). In the methylcellulose CFU assay, the miRNAs,26,27 such as miR-126 (380–26 100% of RNU48 in primary numbers of granulocytic and erythroid burst-forming units were AML samples of our data set4). Cells transduced with a lentivirus significantly reduced, whereas the other CFU types were also carrying the miR-30-backbone with a non-silencing guide (miR- negatively affected by mir-139 (Figure 3e). The collagen-based NSC) served as a control. assay confirmed reduction of myeloid and erythroid colony- Total cell counts and a luminescent cell viability assay (CellTiter-Glo) forming cells (Figure 3f). revealed anti-proliferative properties of mir-139 in all cell lines Thus, the anti-proliferative effect of mir-139 is not purely (Figures 1b and c). The colony-forming capacity was similarly restricted to leukemic cells, but can be extended to myelomono- reduced, as confirmed by methylcellulose colony-forming unit cytic and erythroid progenitors. (CFU) assays (Figure 1d and Supplementary Figure 1d). Annexin V-apoptosis assays revealed induction of apoptosis throughout all cell lines by mir-139 (Figure 1e), which was also evident by Activation of miR-139-5p expression by histone deacetylase an increase of cells in the subG1 fraction in the 5-brom-2- (HDAC) inhibitor treatment desoxyuridin/7-aminoactinomycin-D staining (Supplementary High miR-139-5p expression in mature myelomonocytic cells and Figures 1e and f). At the same time, the percentage of actively low to absent expression in leukemic myelomonocytic blasts cycling cells (S-phase) decreased upon mir-139 expression in points toward transcriptional repression in malignant hemato- HL-60 and THP-1 cells (Figure 1f and Supplementary Figure 1e). poietic cells. HDACs remove acetyl groups from lysine moieties of

30 AML patient samples AML cell lines 20 10

0.4

0.2

miR - 139-5p (% of RNU 48) 0.0

ME-1 HL-60 THP-1 SKNO-1 CN [n=37] Kasumi-1 OCI/AML3 MLL [n=35] t(7;12) [n=4] HSPC [n=3] t(8;21) [n=20] Other [n=41] inv (16) [n=19] t(15;17) [n=12] monocytes [n=4] neutrophils [n=8] macrophages [n=3]

G2/M S

ctrl miR-139 ctrl miR-139 G1 100 15 200 80 100 ** ) ) 6 80

6 150 60 60 10 100 40 50 40 5 **

LU (x10 ** 50 20 Annexin V ** + cells (%)

cells (x10 20 **

KASUMI-1 *

0 0 populations (%) 0 CFUs / 1000 cells 0 0

100 10 200 80 100

) ** ) 6 80 8

6 150 60 60 6 100 40 50 40 4 ** **

LU (x10 50 20 Annexin V SKNO-1 + cells (%)

cells (x10 20 ** 2 ** 0 ** 0 0 populations (%) 0

0 CFUs / 1000 cells

15 6 200 60 100 * ) ) 6

6 150 10 4 40 100 ** 50

ME-1 5 ** 2 ** 20

LU (x10 ** 50 Annexin V + cells (%) cells (x10 * 0 0 0 0 populations (%) 0 CFUs / 2500 cells

150 20 200 40 100 ) ) 6

6 15 150 30 ** 100 10 100 20 50 50 HL-60

LU (x10 5 50 10 Annexin V + cells (%) cells (x10 ** ** ** ** **

** populations (%) 0 0 0 0

0 CFUs / 1000 cells

250 15 400 60 100 ** ) ) 6 200 300 6 150 10 40 200 50 100 5 ** 20 ** 100 ** LU (x10 Annexin V + cells (%) cells (x10 50 ** ** OCI/AML3 ** populations (%)

0 0 CFUs / 1000 cells 0 0 0

15 15 250 50 100

) ** ) 6 200 40 10 6 10 150 30 ** 50 100 20 5 ** 5 THP-1 **

LU (x10 ** Annexin V + cells (%) cells (x10 ** 50 10

0 0 0 0 populations (%) 0 CFUs / 1000 cells 0 5 10 15 036912 ctrl miR-139 days of culture days of culture Figure 1. mir-139 suppresses proliferation in AML cells. (a) Expression of miR-139-5p (mean ± s.e.m.) as measured by quantitative reverse transcription–PCR (TaqMan). Expression is significantly higher in t(8;21) AML (Po0.01) and inv(16) AML (Po0.0001) when compared with all other AML patients except t(7;12) translocations (t(8;21) P = not significant; inv(16) Po0.01). HSPC, purified cord blood CD34+ cells; macrophages, in vitro differentiated macrophages; neutrophils/monocytes, fluorescence-activated cell sorted peripheral blood CD11b+/CD14+ monocytes and CD66b+ neutrophils. Growth curves displaying total cell counts (b) or CellTiter-Glo measurements (c) of mir-139- or miR-NSC (control (ctrl))-transduced cells. LU, luminescence units. (d) Quantitative analysis of CFU assays of transduced cell lines. Apoptosis detection by Annexin V/7-AAD-staining (e) and cell cycle analysis (f) were measured on day 4 post transduction. For cell cycle analyses, only the cells in G1, S and G2/M were displayed (set to 100%). Data are presented as mean ± s.d. of replicates of three to four independent experiments (*Po0.05; **Po0.01).

miR-139- miR-139- monocytic differentiation neutrophilic differentiation + 1:1 ctrl 1:1 ctrl-mCherry GFP+ GFP+ 20 25 ctrl ctrl 20 ) )

6 15 miR-139 6 miR-139 15 10 10 cells (x10 cells (x10 5 5 i.f. transplantaion i.f. transplantaion 0 0 NRG-mice NRG-mice 0 5 10 15 0 5 10 15 days in culture days in culture

67.9±2% 50.6±1% 4 weeks 4 weeks ctrl

33.9±3% 21.9±2% GFP+/mCherry+ qPCR for vector integration

10 100 miR-139 + 1 + 10 * / ctrl CD15 +

** CD11b 0.1 1 CD14 CD13 / mCherry +

0.01 miR-139 0.1 210 150

GFP CFU-M CFU-GM 0.01 CFU-G BFU-E cells 140 0.001 cells 100 3 CFU-GM 3 day 0 day 4 day 28 day 27 70 BFU-E 50 CFU/10

CFU-GEMM CFU/10 1.0 ctrl 0 0

miR-139 ctrl ctrl miR-139 miR-139

0.5 Figure 3. mir-139 blocks in vitro cell growth and differentiation of HSPCs. Growth curves depict total cell counts of mir-139 or control (ctrl)-transduced CB-HSPCs during monocyctic (a) and neutrophil

Fraction survival Fraction (c) in vitro differentiation. Scatter plots show percentage of CD11b+/CD14+ cells during monocytic (b) and CD13+/CD15+ cells 0.0 during neutrophil (d) in vitro differentiation. Number of CFUs in 0 50 100 150 methylcellulose-based (e) and collagen-based (f) CFU assays of days elapsed mir-139 and miR-NSC(ctrl)-transduced HSPCs. (a–f) Data are presented as mean ± s.d. of replicates from two independent Figure 2. mir-139 exerts an anti-leukemic effect in vivo. For all experiments. transplantation experiments, KASUMI-1 cells were injected i.f., schemes illustrate the experimental setup of each competition + + Supplementary Figure 3c). In SKNO1 cells, HDAC inhibitor assay. (a) Ratio of mCherry miR-NSC(ctrl)-transduced vs GFP − 3 = treatment increased the miR-139-5p level from 1.2 × 10 %to mir-139-transduced cells in NRG mice (n 3) 28 days post 5 transplantation. (b) Log-transformed ratio of mir-139 vs miR-NSC 25% of RNU48 (2.1 × 10 -fold), as compared with 270% upon (ctrl)-amplicon Ct’s from bone marrow leukocyte genomic DNA on forced lentiviral expression (Supplementary Figure 3d). Recently, day 4 (n = 4) or day 27 (n = 20) post transplantation. For quantitative Shen and co-workers obtained comparable results with SAHA PCR validation, see Supplementary Figure 2. (c) Survival curve and TSA.28 We could not observe a similar response upon of miR-NSC(ctrl)- (n = 4) vs miR-139-transduced (n = 5) cells (Log-rank 5-acacytidine treatment, arguing against the possibility of test P = 0.022). (a, b) Data are presented with mean promoter hypermethylation as the underlying mechanism of o o (green line). *P 0.05; **P 0.01. ctrl, control. mir-139 repression (data not shown). The miR-139 gene is located on human chromosome 11 in an histones, yielding condensed and transcriptionally silent chroma- intron of the host gene PDE2A. Recent evidence suggested tin. Accordingly, when we treated the t(8;21)-positive KASUMI-1 independent transcriptional regulation of PDE2A and miR-139. and SKNO-1 cells or t(9;11)-positive THP-1 cells with HDAC class I A promoter of mir-139 has been found, residing ~ 2 kb upstream and class II inhibitors vorinostat (SAHA), valproic acid (VPA) from the published transcription start site.28,29 Treatment of or trichostatin A (TSA), we observed induction of miR-139-5p KASUMI-1 and SKNO-1 cells with SAHA, VPA and TSA led to expression with increasing concentrations (Figures 4a–f and a signiﬁcant enrichment of acetylated Histone H3 (H3ac) at the

© 2016 Macmillan Publishers Limited Oncogene (2016) 1822 – 1831 mir-139 controls translation S Emmrich et al 1826 mir-139-promoter compared with vehicle-treated (MOCK) cells, as Kasumi-1 SKNO-1 quantified by chromatin immunoprecipitation (Figure 4g). A silent 20 150 locus on the X chromosome (IgX) showed no or low variation of ** H3ac (Figure 4g). 15 ** 100 In summary, HDAC inhibitor treatment of AML cells activates 10 mir-139 expression in a dose-dependent manner through ** acetylation of Histone H3 at the miR-139 promoter. ** 50 5 mir-139 exerts its effects by targeting EIF4G2 0 0 Next, we sought to identify the molecular target genes of 1 3 1 3 0.3 10 0.3 10 mir-139 responsible for its phenotype. To this aim, we DMSO DMSO performed microarray global mRNA expression profiling of SAHA [µM] SAHA [µM] miR-NSC(ctrl) or mir-139-transduced KASUMI-1 and SKNO-1 cells, which revealed EIF4G2, DPY30 and BTG3 to be among the top 10 downregulated transcripts harboring predicted miR-139-5p- binding sites (Supplementary Figure 4a). Downregulation of all 25 40 ** three genes was validated by quantitative reverse transcription– 20 30 PCR (Figure 5a and Supplementary Figure 4b). The global ** fi 15 ** mRNA expression pro ling was complemented by stable isotope 20 labeling of amino acids in culture (SILAC) followed by liquid 10 * chromatography-mass spectrometry using KASUMI-1 cells.30 This 5 10 experimental setup allowed the identification of mir-139 targets on the protein level. A total of 1837 proteins were cross-identified 0 0 in both label-switched fractions (mir-139-heavy:miR-NSC(ctrl)-light (miR-139/RNU48)/DMSO 1 2 4 1 2 4 0.5 0.5 and mir-139-light:miR-NSC(ctrl)-heavy; Figure 5b). Of those, 32 DMSO DMSO proteins were downregulated in both fractions with a ratio VPA [mM] VPA [mM] mir-139:miR-NSC(ctrl) of less than 0.75, whereas 38 proteins were upregulated (Supplementary Table 1). Remarkably, EIF4G2—the second most downregulated miR-139-5p target in the global mRNA expression profiling—was one of two (second protein: 25 20 PSAT1) downregulated proteins, where the transcript contained ** ** 20 predicted miR-139-5p-binding sites (Figure 5b and Supplementary 15 Figure 4c. PSAT1 mRNA was undetectable by quantitative reverse 15 transcription–PCR (data not shown)). Consistently, western blot- 10 fl 10 * ting and intracellular labeling followed by ow cytometric 5 quantitation revealed robust repression of EIF4G2 protein 5 (Figure 5c and Supplementary Figures 4d and e). Luciferase 0 0 reporter assays confirmed direct targeting of two binding sites (in (miR-139/RNU48)/DMSO (miR-139/RNU48)/DMSO the 5’UTR and 3’UTR) by miR-139-5p (Figure 5d). Moreover, we 50 50 200 400 200 400 found significant inverse correlation of EIF4G2-5p and miR-139-5p DMSO DMSO expression in AML patients carrying the t(8;21) translocation TSA [nM] TSA [nM] (ρ = − 0.56, P = 0.01). To validate EIF4G2 as a functionally active mir-139-5p target, we used two shRNAs for lentiviral knockdown. Western blotting fi 50 IgX determined an intermediate and strong knockdown ef ciency ** for sh-EIF4G2 #1 and sh-EIF4G2 #2, respectively (Supplementary miR-139 promoter 40 Figures 4f and g). For shEIF4G2 #2, the knockdown was higher than for mir-139. EIF4G2-knockdown was at least in part sufficient to recapitulate the effect of mir-139 on proliferation 30 ** and viability of KASUMI-1 cells grown in liquid culture (Figures 5e and f). Furthermore, both shRNAs led to a significant reduction 20 of colony-forming capacity (Figure 5g). Consistent with the * knockdown efficiency, sh-EIF4G2 #2 was more potent than 10 sh-EIF4G2 #1. Most importantly, expression of EIF4G2 protein from a transcript 0 without binding sites for miR-139-5p (for western blot see H3Ac enrichment over IgG/ MOCK SAHA VPA TSA Supplementary Figure 4h) completely rescued the anti-proliferative and pro-apoptotic effect of mir-139 (Figures 5h–j). Figure 4. HDAC inhibitors activate miR-139-5p expression. miR-139- Collectively, we showed that miR-139-5p binds to the EIF4G2 5p expression upon SAHA (a), VPA (c) and TSA (e) treatment of KASUMI-1 cells or upon SAHA (b), VPA (d) and TSA (f) treatment of transcript to reduce mRNA and protein levels. shRNA-mediated SKNO-1 cells in relation to dimethyl sulfoxide (DMSO, control (ctr)). EIF4G2 knockdown recapitulates the effect of mir-139 on AML cells, Statistics were calculated against DMSO. (g) chromatin immuno- whereas mir-139-5p-resistant EIF4G2 rescues the mir-139 phenotype. precipitation for acetylated Histone H3 at the promoter of miR-139 μ and a negative control locus (IgX) in SAHA (1 M), VPA (2 mM) or TSA mir-139 reduces protein synthesis in AML (400 nM) or vehicle (MOCK; red dotted line) treated cells. All HDAC inhibitor treatment was performed for 72 h. (a–g) Data are The EIF4G protein homologs function as scaffolds to link EIF4A presented as mean ± s.d. of replicates from three independent RNA helicase and EIF4E cap-binding protein to form the experiments (*Po0.05; **Po0.01). translation initiation complex EIF4F.31 It has been shown by

2.0

EIF4G2BTG3 DPY30 CDK6 1.2 1.0 1.5 0.8 * EIF4G2 100 ** 70 0.5 0.4 0.0 0.5 1.5 2.0 55 (GOI/B2M) /ctrl 0.5 ACTB 40 RLU (miR -139 / ctrl) 0.0 0.0 Kasumi-1 SKNO-1 ** ** ** ** (miR-139) / light (ctrl) HEAVY EIF4G2 0.0 light (miR-139) / HEAVY (ctrl) pMiR report EIF4G2EIF4G2 5'UTR 3'UTR )

6 ctrl 1.0 150 30 miR-139 ** ** sh-EIF4G2 #1 100 20 ** ** ** sh-EIF4G2 #2 0.5 **

** cells ** 50 ** 10 ** RLU / ctrl ** ** ** ** **

KASUMI-1 cells (x10 ** ** 0 0.0 ** ** CFUs / 1000 Kasumi-1 0 0 5 10 15 0 3 6 9 12 15 ctrl days of culture days of culture miR-139

sh-EIF4G2sh-EIF4G2 #1 #2

) ctrl 6 100 150 ns 40 ns ** miR-139 80 ns EIF4G2 + ctrl 30 100 ns 60 EIF4G2 + miR-139 ns 20 ns 40 cells ** 50 20 10 Annexin V + cells(%)

KASUMI-1 cells (x10 ** 0 ** CFUs / 1000 Kasumi-1 0 0 0 5 10 15 ctrl ctrl days of culture miR-139 miR-139

EIF4G2 + ctrl EIF4G2 + ctrl

EIF4G2 + miR - 139 EIF4G2 + miR - 139

ACTB ACTB 40 40

35 40 CDKN1B CDK6

Kasumi-1 SKNO-1 Kasumi-1 Figure 5. EIF4G2 is a miR-139-5p target gene. (a) Quantitative PCR validation of indicated genes (GOI) in KASUMI-1 cells transduced with mir-139. Data are shown as fold change compared with miR-NSC(ctrl) cells, presented as mean ± s.d. of replicates of two independent experiments. (b) Liquid chromatography-mass spectrometry protein quantification (SILAC; label switch) in KASUMI-1 cells. EIF4G2 is highlighted. (c) Western blot of EIF4G2 in mir-139- and miR-NSC(ctrl)-transduced KASUMI-1 and SKNO-1 cells. (d) Luciferase reporter assay with indicated EIF4G2 transcript fragments, harboring a miR-139-5p-binding site, in 293T cells. Cell counts (e) and luminescence cell viability assay (f), CFU assays (g) of KASUMI-1 cells transduced with indicated constructs. Cell counts (h) and CFU assays (i) of miR-139- and miR-NSC (ctrl)-transduced KASUMI-1 cells compared with EIF4G2-transduced KASUMI-1 cells co-transduced with either control (ctrl) or miRNA construct. (j) Apoptosis detection by Annexin V/7-AAD-staining on day 4 post miRNA transduction. (k-l) Western blot of (k) p27 (CDKN1B) and (l) CDK6 in mir-139- and ctrl-transduced KASUMI-1 and SKNO-1 cells. Arrowheads denote respective ladder bands. Data are presented as mean ± s.d. of replicates of three to four independent experiments (*Po0.05; **Po0.01). NS, not significant. metabolic labeling experiments that inhibition of EIF4G2 reduced synthesis by premature translation termination, served as control. protein synthesis in axons.32 Therefore, we investigated whether Flow cytometry analysis demonstrated a strong decrease of the mir-139 controls protein synthesis via its main target EIF4G2. MFI in the puromycin-treated cells compared with MOCK cells To this aim, we pulse-labeled newly translated peptides using (Supplementary Figure 5a). Next, we subjected six mir-139- homopropargylglycine and azide-fluorophore conjugates.33 HL-60 responsive cell lines to metabolic labeling. Five out of six cell cells treated with puromycin (1 μg/ml), which inhibits protein lines displayed a profound inhibition of protein synthesis by

F AU)MFI (A.U.) 15 MFI (A.U.) 200 F AU)MFI (A.U.) MFI (A.U.) 15 MFI (A.U.) 150 20 300 10 150 10 100 200 100 10 5 5 50 100 * 50 ns ** 0 0 0 0 0 0 10 10 10 10 10 10 10 10 10 10 10 10

THP-1 OCI/AML3 HL-60 600 800 60 20 25

400 600 15 20 40 400 MFI (A.U.) 300 15 400 10 ** 200 10 * 20 200 200 * 5 100 5 Counts 0 0 0 0 ** 0 0 10 10 10 10 10 10 10 10 10 10 10 10 HPG

1.0 1.0 miR-139 low miR-139 low miR-139 high miR-139 high p value = 0.001 p value = 0.112 0.8 0.8

0.6 0.6

0.4 0.4 Probability Probability

0.2 0.2

0.0 0.0

01234 5 012345 Time to event (years) Time to death (years) Figure 6. mir-139 reduces translation and is a prognostic factor in childhood AML. (a) Flow cytometry histograms of Alexa Fluor 594 homopropargylglycine (HPG)-labeled cell lines transduced with mir-139 or miR-NSC(ctrl). Graphs to the right of each plot depict mean ﬂuorescence intensities (MFI). Data are presented as replicates (dots) with mean (line) of three independent experiments (*Po0.05, **Po0.01; NS, not signiﬁcant). (b) Five-year event-free survival in high-expressing miR-139-5p (green) and low-expressing miR-139-5p (black) AMLs. (c) Overall survival in high-expressing miR-139-5p (green) and low-expressing miR-139-5p (black) AMLs; P values according to log rank tests.

mir-139 (Figure 6a and Supplementary Table 2). Only in mir-139 is a prognostic factor in pediatric AML ME-1 cells with a doubling time of 5 days, no reduction was mir-139 generally exerts tumor-suppressive functions and pro- observed.34 Strikingly, co-expression of EIF4G2 variant without longed the survival in a xenograft mouse model. Therefore, we miR-139-5p-binding sites prevented mir-139-mediated reduction speculated that increased expression of mir-139 might be of protein synthesis (Supplementary Figure 5b). associated with favorable prognosis in pediatric AML patients. In Previously, it was shown that EIF4G2 knockdown increased the a cohort of 165 patients treated according to protocols from protein levels of cyclin-dependent kinase inhibitor CDKN1B the Dutch Childhood Oncology Group (DCOG), the AML (p27Kip1).35 The encoded protein binds to and prevents the ‘Berlin-Frankfurt-Münster’ Study Group (AML-BFM-SG), the Czech activation of cyclin E-CDK2 or cyclin D-CDK4 complexes, and thus Pediatric Hematology Group (CPH) and the St Louis Hospital controls the cell cycle progression at G1. Accordingly, we detected in Paris, France, we stratified patients according to their higher p27 levels upon mir-139 overexpression (Figure 5k, and miR-139-5p expression (Supplementary Figure 6). The bimodal Supplementary Figure 5c). Inversely, cell cycle activator CDK6, miR-139-5p expression allowed setting a clear cutoff. Although containing five binding sites for miR-139-5p in the 3′UTR increased expression of miR-139-5p was not significantly (Supplementary Figure 5d), was repressed by mir-139 on the (Plog-rank = 0.122; Figure 6b) associated with longer 5-year event- mRNA and protein level in KASUMI-1 cells (Figure 5l and free survival, we found a significant association with a better Supplementary Figure Se). Thus, mir-139 controls the cellular fate overall survival in univariate analysis (Plog-rank = 0.001; Figure 6c). by globally and specifically (for example, p27Kip1 and CDK6) Thus, expression of tumor-suppressive miR-139-5p may serve as a regulating protein expression. prognostic marker in pediatric AML.

Oncogene (2016) 1822 – 1831 © 2016 Macmillan Publishers Limited mir-139 controls translation S Emmrich et al 1829 DISCUSSION phosphorylating (inhibiting) the 4E-BP1 protein that prevents 51 Here we describe a novel function of miR-139-5p as tumor the assembly of EIF4F. Thus, restoring miR-139-5p or miR-520-3p suppressor in AML. mir-139 is suppressed across several different expression to repress translation initiation and thereby leukemic cytogenetically defined pediatric AML subgroups and expression growth may constitute an attractive alternative. However, as with can be restored by HDAC inhibition. High or low miR-139-5p normal chemotherapy, also normal HSPCs ceased proliferating expression could stratify our patient cohort in a good and poor and differentiating upon mir-139 induction. During the revision of prognostic subgroup, respectively. We identified a novel mode of this manuscript, miR-139-3p was found to be upregulated in + 52 action, which miRNAs can exploit to regulate cell fate decision in CD34 HSPCs from patients with Fanconia anemia. Retroviral AML. mir-139 disrupts the translation initiation complex by overexpression of the miR-139 gene, notably involving both targeting EIF4G2 to control the overall protein synthesis rate and mature miRNA strands, repressed proliferation and led to thereby cell cycle and viability. As AML cells are specifically apoptosis in 32D mouse cells, whereas strand-specific miR-139- dependent on high protein synthesis rates controlling the 3p inhibition rescued the colony formation defect of a FA patient. expression of miR-139-5p constitutes a novel path for the Thus, the level of mir-139 expression has to be carefully titered, to treatment of AML. find a therapeutic window, which exploits the vulnerability of There is ample evidence of miRNAs acting as tumor suppressors cancer cells while sparing normal cells. in leukemia by altering the translation of specific proteins, Although we defined EIF4G2 as the major target of mir-139 in resulting in cellular context-dependent tumor-suppressive effects. AML, which can abrogate the effect of mir-139 when restored, a This is primarily exemplified by mir-9 acting as an oncogene in strong EIF4G2 knockdown did not fully mimic the anti-proliferative MLL-transformed cells and as a differentiation inducer inhibiting effect of mir-139. This suggests that other targets may proliferation in t(8;21) cells.4,36 mir-143 has been reported to additionally contribute to the mir-139 phenotype. In fact, we repress the oncogenic MLL-AF4 fusion protein.37 In contrast, our identified and validated DPY30, BTG3 and CDK6 as miR-139-5p data from different models in vitro and in vivo support the notion targets (Figures 5a and l). DPY30 as a part of the MLL-complex is that mir-139 commonly acts as a growth inhibitory molecule. involved in H3K4 trimethylation, an epigenetic marker for 53 Accordingly, mir-139 is silenced in a number of solid tumors transcriptional activation. The tumor-suppressor BTG3 maintains including hepatocellular carcinomas, gastric and metastatic breast genomic stability, and is downregulated in breast cancer, ovarian 29,54–56 cancers, laryngeal squamous cell carcinoma or glioma.29,38–40 carcinoma and non-small-cell lung cancer. However, loss of 57 Different mir-139 targets or pathways were proposed, which BTG3 leads to acute cellular senescence in normal cells. might explain the reduction of metastasis and tumor progression By interaction with cyclins, CDK6 activates the cell cycle at early G1. In the hematopoietic system, CDK6 seems to be required for HSC by mir-139, such as ROCK1, NOTCH1 (Notch-signaling), TCF4 58 (Wnt-pathway), CXCR4 or anti-apoptotic MCL1.10,38,41–43 activation and engraftment of leukemic stem cells. Interestingly, Here, we described a mechanism, which might broadly explain MLL fusions rely on CDK6, whereas its abrogation induced myeloid 59 the cell cycle inhibitory and pro-apoptotic effect of mir-139 in a differentiation in those blasts. Together with targeting of DPY30, wide range of neoplastic cells. By integrating proteomics data, CDK6 repression by mir-139 can be regarded as an additive layer global gene expression profiling and miRNA-target prediction above the tumor suppressive main target EIF4G2. algorithms, we show that mir-139 globally controls the cellular In conclusion, the identification of mir-139 as a global tumor protein synthesis rate by targeting EIF4G2, an essential component suppressor in AML mediated by blocking the translation factor of the translation initiation complex consisting of EIF4E, EIF4G and EIF4G2 opens the avenue for future therapeutic options. Our fi EIF4A. EIF4G2 has initially been suggested to specifically promote ndings show that miRNAs can not only regulate the translation of IRES-dependent translation.44 Later studies demonstrated that particular targets, but also more broadly the global cellular protein EIF4G2 is a translational activator in both unstressed cells and synthesis rate. tumor cells.45,46 The aberrant expression of EIF4F components 15–17 EIF4E and EIF4G has been linked to leukemia progression. MATERIALS AND METHODS When overexpressed, EIF4G2 mediates an increased protein synthesis rate.35,47 Inversely, EIF4G2 knockdown inhibits cell Primary cells and cell lines + proliferation and coincides with increased CDKN1B protein levels, CD34 HSPCs were obtained from cord blood of healthy donors with 35 informed consent according to the Declaration of Helsinki and isolated an inhibitor of cell cycle progression at the G1–S transition. In fact, structural analysis of the conserved middle domain revealed using CD34-immunomagnetic microbeads (Miltenyi, Bergisch Gladbach, fi 48 Germany). HSPCs were cultured in RPMI (Lonza, Basel, Switzerland) with EIF4A recruitment capacity, albeit with lower af nity than EIF4G. 10% fetal calf serum and 1% penicillin–streptomycin adding FLT3L, IL6, In addition, EIF4G2 is known to maintain cell survival during granulocyte macrophage colony-stimulating-factor and macrophage mitosis by promoting cap-independent translation of Bcl-2 and colony-stimulating-factor (all Preprotech, Hamburg, Germany) for mono- 49 CDK1. These data position EIF4G2 as a central regulator of cytic and SCF, IL3, granulocyte macrophage colony-stimulating-factor und translation, but also as a leverage for cell cycle integrity under granulocyte colony-stimulating-factor (all Preprotech) for granulocytic normal conditions, which itself is under the control of miR-139-5p. differentiation. Primary CD11b+/CD14+ monocytes and CD66b+ neutrophils Recently, miR-520-3p was also shown to repress diffuse large were obtained from healthy donors after HetaSep (Stemcell Technologies, B-cell development by targeting EIF4G2 to result in decreased Cologne, Germany) erythrocyte separation and sorting. VPA (SIGMA Life global translation and cell proliferation.46 As deregulation of Science, Taufkirchen, Germany), vorinostat (SAHA; Biomol GmbH, Ham- burg, Germany) and TSA (Applichem GmbH, Darmstadt, Germany) were mRNA translation has been shown to promote cellular transfor- ’ 12,13 dissolved according to the manufacturer s instructions and used at the mation and neoplastic growth, targeting translation initiation indicated concentrations.60 Cell lines OCI/AML3, HL-60, THP-1, ME-1, emerged as a new paradigm for AML therapy and to overcome SKNO-1 and KASUMI-1 were obtained from the German National Resource resistance to anti-BRAF and anti-MEK therapies in melanoma.18,31 Center for Biological Material (DSMZ, Braunschweig, Germany) and Consequently, drugs disrupting the initiation complex, such as cultured as recommended. 4EGI, anti-EIF4E antisense oligonucleotides or the antiviral drug 18 ribavirin are currently clinically or preclinically tested. For AML, Xenograft mouse model this may be of particular interest as constitutive PI3K/mammalian NOD.Cg-Rag1tm1Mom Il2rgtm1Wjl/SzJ (NRG) mice (Jackson Laboratory, Bar target of rapamycin pathway activation is a common feature of Harbor, Maine, USA) were maintained in a specific pathogen-free 50 70% of cases with AML. Mammalian target of rapamycin is a environment. All experimental procedures using these mice were general activator of cap-dependent translation by performed in accordance to protocols approved by the local authorities

© 2016 Macmillan Publishers Limited Oncogene (2016) 1822 – 1831 mir-139 controls translation S Emmrich et al 1830 (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsi- 15 Avdulov S, Li S, Michalek V, Burrichter D, Peterson M, Perlman DM et al. Activation cherheit). For in vivo assays, 5 × 106 KASUMI-1 cells were injected i.f. per of translation complex eIF4F is essential for the genesis and maintenance of the animal. Recipients were randomly assigned. After 28 days, mice were killed malignant phenotype in human mammary epithelial cells. Cancer Cell 2004; 5: and leukemic cells were isolated from femurs and analyzed as described.4 553–563. For survival analysis, 1 × 107 transduced KASUMI-1 cells were injected i.f. 16 Fukuchi-Shimogori T, Ishii I, Kashiwagi K, Mashiba H, Ekimoto H, Igarashi K. per animal and moribund mice were killed. Malignant transformation by overproduction of translation initiation factor eIF4G. Cancer Res 1997; 57: 5041–5044. 17 Hariri F, Arguello M, Volpon L, Culjkovic-Kraljacic B, Nielsen TH, Hiscott J et al. Statistical analyses The eukaryotic translation initiation factor eIF4E is a direct transcriptional target of Statistical analyses of experiments were performed using the two-tailed NF-kappaB and is aberrantly regulated in acute myeloid leukemia. Leukemia 2013; Student’s t-test for two groups and analysis of variance for three or more 27: 2047–2055. groups. Analyses were performed with GraphPad Prism 6 (GraphPad 18 Tamburini J, Green AS, Chapuis N, Bardet V, Lacombe C, Mayeux P et al. Targeting Software Inc., La Jolla, CA, USA). All used tests were two-tailed, and a translation in acute myeloid leukemia: a new paradigm for therapy? Cell Cycle P value of less than 0.05 was considered significant. 2009; 8: 3893–3899. 19 Gradi A, Imataka H, Svitkin YV, Rom E, Raught B, Morino S et al. A novel functional human eukaryotic translation initiation factor 4G. Mol Cell Biol 1998; 18: CONFLICT OF INTEREST 334–342. 20 Imataka H, Gradi A, Sonenberg N. A newly identified N-terminal amino acid fl The authors declare no con ict of interest. sequence of human eIF4G binds poly(A)-binding protein and functions in poly(A)- dependent translation. EMBO J 1998; 17:7480–7489. ACKNOWLEDGEMENTS 21 Qin H, Raught B, Sonenberg N, Goldstein EG, Edelman AM. Phosphorylation screening identifies translational initiation factor 4GII as an intracellular target The authors thank Drs K Weber, B Fehse and D Heckl for providing plasmids; Dr M of Ca(2+)/calmodulin-dependent protein kinase I. J Biol Chem 2003; 278: Ballmaier for sorting. SE, FE and KH were supported by the Hannover Biomedical 48570–48579. Research School. JEK-K was supported by the Children Cancer Free Foundation (KIKA, 22 Caron S, Charon M, Cramer E, Sonenberg N, Dusanter-Fourt I. Selective project 49). JHK is a fellow of the Emmy Noether-Programme from the German modification of eukaryotic initiation factor 4F (eIF4F) at the onset of cell Research Foundation (DFG; KL-2374/2-1). A.O. was supported by KIKA project 109. This differentiation: recruitment of eIF4GII and long-lasting phosphorylation of eIF4E. 24 – work was supported by grants to CMZ and MMvdH-E from the Children Cancer Free Mol Cell Biol 2004; : 4920 4928. Foundation (KIKA, projects 49 and 109) and to JHK from the DFG (KL-2374/2-1). 23 Tu L, Liu Z, He X, He Y, Yang H, Jiang Q et al. Over-expression of eukaryotic translation initiation factor 4gamma 1 correlates with tumor progression and poor prognosis in nasopharyngeal carcinoma. Mol Cancer 2010; 9: REFERENCES 78 4598-9-78. 24 Solh M, Yohe S, Weisdorf D, Ustun C. Core-binding factor acute myeloid leukemia: 1 Sanders MA, Valk PJ. The evolving molecular genetic landscape in acute myeloid Heterogeneity, monitoring, and therapy. Am J Hematol 2014; 89: 1121–1131. 20 – leukaemia. Curr Opin Hematol 2013; :79 85. 25 Weber K, Bartsch U, Stocking C, Fehse B. A multicolor panel of novel lentiviral 2 Creutzig U, van den Heuvel-Eibrink MM, Gibson B, Dworzak MN, Adachi S, "gene ontology" (LeGO) vectors for functional gene analysis. Mol Ther 2008; 16: de Bont E et al. Diagnosis and management of acute myeloid leukemia in children 698–706. and adolescents: recommendations from an international expert panel. Blood 26 Cheng J, Guo S, Chen S, Mastriano SJ, Liu C, D'Alessio AC et al. An extensive 120 – 2012, 10/18 : 3187 3205. network of TET2-targeting MicroRNAs regulates malignant hematopoiesis. Cell 3 Vardiman JW, Thiele J, Arber DA, Brunning RD, Borowitz MJ, Porwit A et al. Rep 2013; 5: 471–481. fi The 2008 revision of the World Health Organization (WHO) classi cation of 27 Masaki S, Ohtsuka R, Abe Y, Muta K, Umemura T. Expression patterns of micro- myeloid neoplasms and acute leukemia: rationale and important changes. Blood RNAs 155 and 451 during normal human erythropoiesis. Biochem Biophys Res 114 – 2009; : 937 951. Commun 2007; 364: 509–514. 4 Emmrich S, Katsman-Kuipers JE, Henke K, Khatib ME, Jammal R, Engeland F et al. 28 Shen K, Mao R, Ma L, Li Y, Qiu Y, Cui D et al. Post-transcriptional regulation of the 28 miR-9 is a tumor suppressor in pediatric AML with t(8;21). Leukemia 2014; : tumor suppressor miR-139-5p and a network of miR-139-5p-mediated mRNA – 1022 1032. interactions in colorectal cancer. FEBS J 2014; 281:3609–3624. 5 Wang Z, Hong Z, Gao F, Feng W. Upregulation of microRNA-375 is associated with 29 Bao W, Fu HJ, Xie QS, Wang L, Zhang R, Guo ZY et al. HER2 interacts with CD44 to 383 poor prognosis in pediatric acute myeloid leukemia. Mol Cell Biochem 2013; : up-regulate CXCR4 via epigenetic silencing of microRNA-139 in gastric – 59 65. cancer cells. Gastroenterology 2011; 141:2076–2087.e6. 6 Lee DW, Futami M, Carroll M, Feng Y, Wang Z, Fernandez M et al. Loss of SHIP-1 30 Vinther J, Hedegaard MM, Gardner PP, Andersen JS, Arctander P. Identification of protein expression in high-risk myelodysplastic syndromes is associated with miRNA targets with stable isotope labeling by amino acids in cell culture. Nucleic miR-210 and miR-155. Oncogene 2012; 31:4085–4094. Acids Res 2006; 34: e107. 7 Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004; 31 Boussemart L, Malka-Mahieu H, Girault I, Allard D, Hemmingsson O, Tomasic G 116:281–297. et al. eIF4F is a nexus of resistance to anti-BRAF and anti-MEK cancer therapies. 8 Berindan-Neagoe I, Monroig Pdel C, Pasculli B, Calin GA. MicroRNAome Nature 2014; 513:105–109. genome: a treasure for cancer diagnosis and therapy. CA Cancer J Clin 2014; 64: 32 Kar AN, MacGibeny MA, Gervasi NM, Gioio AE, Kaplan BB. Intra-axonal synthesis of 311–336. eukaryotic translation initiation factors regulates local protein synthesis and axon 9 Klein U, Lia M, Crespo M, Siegel R, Shen Q, Mo T et al. The DLEU2/miR-15a/16-1 growth in rat sympathetic neurons. J Neurosci 2013; 33: 7165–7174. cluster controls B cell proliferation and its deletion leads to chronic lymphocytic 33 Breinbauer R, Kohn M. Azide-alkyne coupling: a powerful reaction for bio- leukemia. Cancer Cell 2010; 17:28–40. conjugate chemistry. Chembiochem 2003; 4:1147–1149. 10 Wong CC, Wong CM, Tung EK, Au SL, Lee JM, Poon RT et al. The microRNA 34 Yanagisawa K, Horiuchi T, Fujita S. Establishment and characterization miR-139 suppresses metastasis and progression of hepatocellular carcinoma by of a new human leukemia cell line derived from M4E0. Blood 1991; 78: down-regulating Rho-kinase 2. Gastroenterology 2011; 140:322–331. 451–457. 11 Li Z, Huang H, Li Y, Jiang X, Chen P, Arnovitz S et al. Up-regulation of 35 Lee SH, McCormick F. p97/DAP5 is a ribosome-associated factor that facilitates a HOXA-PBX3 homeobox-gene signature following down-regulation of miR-181 is protein synthesis and cell proliferation by modulating the synthesis of cell cycle associated with adverse prognosis in patients with cytogenetically abnormal AML. proteins. EMBO J 2006; 25: 4008–4019. Blood 2012; 119: 2314–2324. 36 Chen P, Price C, Li Z, Li Y, Cao D, Wiley A et al. miR-9 is an essential oncogenic 12 Clemens MJ. Targets and mechanisms for the regulation of translation in malig- microRNA specifically overexpressed in mixed lineage leukemia-rearranged nant transformation. Oncogene 2004; 23: 3180–3188. leukemia. Proc Natl Acad Sci USA 2013; 110: 11511–11516. 13 Lazaris-Karatzas A, Montine KS, Sonenberg N. Malignant transformation by a 37 Dou L, Zheng D, Li J, Li Y, Gao L, Wang L et al. Methylation-mediated repression of eukaryotic initiation factor subunit that binds to mRNA 5' cap. Nature 1990; 345: microRNA-143 enhances MLL-AF4 oncogene expression. Oncogene 2012; 31: 544–547. 507–517. 14 Hinnebusch AG, Lorsch JR. The mechanism of eukaryotic translation initiation: 38 Luo HN, Wang ZH, Sheng Y, Zhang Q, Yan J, Hou J et al. MiR-139 targets CXCR4 new insights and challenges. Cold Spring Harb Perspect Biol 2012; 4: doi:10.1101/ and inhibits the proliferation and metastasis of laryngeal squamous cshperspect.a011544. carcinoma cells. Med Oncol 2014; 31: 789.

Oncogene (2016) 1822 – 1831 © 2016 Macmillan Publishers Limited mir-139 controls translation S Emmrich et al 1831 39 Krishnan K, Steptoe AL, Martin HC, Pattabiraman DR, Nones K, Waddell N et al. 50 Martelli AM, Nyakern M, Tabellini G, Bortul R, Tazzari PL, Evangelisti C et al. miR-139-5p is a regulator of metastatic pathways in breast cancer. RNA 2013; 19: Phosphoinositide 3-kinase/Akt signaling pathway and its therapeutical implica- 1767–1780. tions for human acute myeloid leukemia. Leukemia 2006; 20:911–928. 40 Fan Q, He M, Deng X, Wu WK, Zhao L, Tang J et al. Derepression of c-Fos caused 51 Holcik M, Sonenberg N. Translational control in stress and apoptosis. Nat Rev Mol by microRNA-139 down-regulation contributes to the metastasis of human Cell Biol 2005; 6:318–327. hepatocellular carcinoma. Cell Biochem Funct 2013; 31:319–324. 52 Alemdehy MF, Haanstra JR, de Looper HW, van Strien PM, 41 Gu W, Li X, Wang J. miR-139 regulates the proliferation and invasion of hepato- Verhagen-Oldenampsen J, Caljouw Y et al. Interstrand cross-link induced cellular carcinoma through the WNT/TCF-4 pathway. Oncol Rep 2014; 31: miR139-3p and miR199a-3p have opposite roles in hematopoietic cell expansion 397–404. and leukemic transformation. Blood 2015. 42 Li RY, Chen LC, Zhang HY, Du WZ, Feng Y, Wang HB et al. MiR-139 inhibits Mcl-1 53 Jiang H, Shukla A, Wang X, Chen WY, Bernstein BE, Roeder RG. Role for Dpy-30 in expression and potentiates TMZ-induced apoptosis in glioma. CNS Neurosci Ther ES cell-fate speciﬁcation by regulation of H3K4 methylation within bivalent 2013; 19: 477–483. domains. Cell 2011; 144: 513–525. 43 Zhang L, Dong Y, Zhu N, Tsoi H, Zhao Z, Wu CW et al. microRNA-139-5p exerts 54 Cheng YC, Lin TY, Shieh SY. Candidate tumor suppressor BTG3 maintains genomic tumor suppressor function by targeting NOTCH1 in colorectal cancer. Mol Cancer stability by promoting Lys63-linked ubiquitination and activation of the check- 2014; 13: 124–4598-13-124. point kinase CHK1. Proc Natl Acad Sci USA 2013; 110: 5993–5998. 44 Henis-Korenblit S, Shani G, Sines T, Marash L, Shohat G, Kimchi A. The 55 Yu J, Zhang Y, Qi Z, Kurtycz D, Vacano G, Patterson D. Methylation-mediated caspase-cleaved DAP5 protein supports internal ribosome entry site-mediated downregulation of the B-cell translocation gene 3 (BTG3) in breast cancer cells. translation of death proteins. Proc Natl Acad Sci USA 2002; 99: 5400–5405. Gene Expr 2008; 14: 173–182. 45 Nousch M, Reed V, Bryson-Richardson RJ, Currie PD, Preiss T. The eIF4G-homolog 56 Deng B, Zhao Y, Gou W, Chen S, Mao X, Takano Y et al. Decreased expression of p97 can activate translation independent of caspase cleavage. RNA 2007; 13: BTG3 was linked to carcinogenesis, aggressiveness, and prognosis of ovarian 374–384. carcinoma. Tumour Biol 2013; 34: 2617–2624. 46 Mazan-Mamczarz K, Zhao XF, Dai B, Steinhardt JJ, Peroutka RJ, Berk KL et al. 57 Lin TY, Cheng YC, Yang HC, Lin WC, Wang CC, Lai PL et al. Loss of the candidate Down-regulation of eIF4GII by miR-520c-3p represses diffuse large B cell tumor suppressor BTG3 triggers acute cellular senescence via the ERK-JMJD3-p16 lymphoma development. PLoS Genet 2014; 10: e1004105. (INK4a) signaling axis. Oncogene 2012; 31: 3287–3297. 47 Srivastava T, Fortin DA, Nygaard S, Kaech S, Sonenberg N, Edelman AM et al. 58 Scheicher R, Hoelbl-Kovacic A, Bellutti F, Tigan AS, Prchal-Murphy M, Heller G et al. Regulation of neuronal mRNA translation by CaM-kinase I phosphorylation of CDK6 as a key regulator of hematopoietic and leukemic stem cell activation. Blood eIF4GII. J Neurosci 2012; 32: 5620–5630. 2014; 125:90–101. 48 Virgili G, Frank F, Feoktistova K, Sawicki M, Sonenberg N, Fraser CS et al. Structural 59 Placke T, Faber K, Nonami A, Putwain SL, Salih HR, Heidel FH et al. analysis of the DAP5 MIF4G domain and its interaction with eIF4A. Structure 2013; Requirement for CDK6 in MLL-rearranged acute myeloid leukemia. Blood 2014; 21: 517–527. 124:13–23. 49 Marash L, Liberman N, Henis-Korenblit S, Sivan G, Reem E, Elroy-Stein O et al. 60 Stankov MV, El Khatib M, Kumar TB, Heitmann K, Panayotova-Dimitrova D, DAP5 promotes cap-independent translation of Bcl-2 and CDK1 to facilitate cell Schoening J et al. Histone deacetylase inhibitors induce apoptosis in myeloid survival during mitosis. Mol Cell 2008; 30:447–459. leukemia by suppressing autophagy. Leukemia 2014 03; 28:577–588.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)