Oncogene (2006) 25, 6067–6078 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE siRNA-mediated AML1/MTG8 depletion affects differentiation and proliferation-associated expression in t(8;21)-positive cell lines and primary AML blasts

J Dunne1,7, C Cullmann2,7,8, M Ritter3, N Martinez Soria2, B Drescher4, S Debernardi1, S Skoulakis1, O Hartmann5, M Krause6, J Krauter4, A Neubauer3, BD Young1 and O Heidenreich2

1Cancer Research UK Medical Oncology Laboratory, Barts and the London School of Medicine, London, UK; 2Department of Molecular Biology, Interfaculty Institute of Cell Biology, Faculty of Biology, Eberhard Karls University Tuebingen, Germany; 3Department of Haematology, Oncology and Immunology, University Hospital Marburg, Germany; 4Department of Haematology, Haemostaseology and Oncology, Hannover Medical School, Hannover, Germany; 5Institute for Medical Biometry and Epidemiology, Philipps University Marburg, Germany and 6Institute for Molecular Biology and Tumour Research, Philipps University Marburg, Germany

The chromosomal translocation t(8;21) is associated with Introduction 10–15% of all cases of acute myeloid leukaemia (AML). The resultant fusion AML1/MTG8 interferes with In acute myeloid leukaemia (AML), chromosomal haematopoietic and is an important aberrations prevalently affect com- regulator of leukaemogenesis. We studied the effects of plexes controlling normal haematopoiesis. The hetero- small interfering RNA (siRNA)-mediated AML1/MTG8 dimeric core-binding factor complex (CBF) consisting of depletion on global gene expression in t(8;21)-positive AML1 (RUNX1, CBFa2, PEBP2aB) and CBFb is most leukaemic cell lines and in primary AML blasts using frequently targeted by such genetic lesions. AML1 is an cDNA arrays, oligonucleotide arrays and real-time reverse essential haematopoietic transcription factor. Homo- transcription–polymerase chain reaction (RT–PCR). Sup- zygous deletion of Aml1 is embryonic lethal because pression of AML1/MTG8 results in the increased expres- of a lackof definitive haematopoiesis (Okuda et al., sion of associated with myeloid differentiation, such 1996; Wang et al., 1996). as AZU1, BPI, CTSG, LYZ and RNASE2 as well as of The chromosomal translocation t(8;21)(q22;q22), antiproliferative genes such as IGFBP7, MS4A3 and SLA which is observed in 10–15% of all AML cases, fuses both in blasts and in cell lines. Furthermore, expression the AML1 gene on 21 with the MTG8 levels of several genes affiliated with drug resistance or (RUNX1T1, CBFA2T1, ETO, CDR) gene on chromo- indicative of poor prognosis AML (BAALC, CD34, PRG2, some 8, yielding a conserved chimaeric transcript TSPAN7) are affected by AML1/MTG8 depletion. In (Erickson et al., 1992; Nisson et al., 1992; Miyoshi conclusion, siRNA-mediated suppression of AML1/MTG8 et al., 1993). In the resulting AML1/MTG8 fusion cause very similar changes in gene expression pattern in protein, the N-terminal part of AML1, containing the t(8;21)-positive cell lines and in primary AML blasts. DNA-binding domain, the RUNT domain, is fused Furthermore, the results suggest that the specific targeting to the almost complete MTG8 sequence. MTG8 is part of AML1/MTG8 function may be a promising approach of a repressor complex also containing PLZF (Melnick for complementing existing treatment strategies. et al., 2000), and interacts via N-CoR and SIN3A with Oncogene (2006) 25, 6067–6078. doi:10.1038/sj.onc.1209638; histone deacetylases (HDACs) 1, 2 and 3 (Gelmetti published online 1 May 2006 et al., 1998; Lutterbach et al., 1998; Wang et al., 1998; Amann et al., 2001; Hildebrand et al., 2001). In contrast Keywords: RNA interference; siRNA; gene expression to the closely related genes MTGR1 and MTG16 profiling; acute myeloid leukaemia; AML1/MTG8 (ETO2), MTG8 is not expressed in myeloid progenitor cells (Miyoshi et al., 1993; Morohoshi et al., 2000). AML1/MTG8 interferes with haematopoietic trans- Correspondence: Dr O Heidenreich, Department of Molecular cription by recruiting HDACs to AML1-responsive Biology, Interfaculty Institute of Cell Biology, Faculty of Biology, promoters leading to histone deacetylation and trans- Eberhard Karls University Tuebingen, Auf der Morgenstelle 15, 72076 Tuebingen, Germany. criptional silencing of genes such as CSF1R (c-fms), ARF E-mail: [email protected] CSF2RA (GM-CSFRa)orp14 (Frank et al., 1995; 7These authors equally contributed to the work. Linggi et al., 2002; Follows et al., 2003). Notably, intro- 8Current address: Research Group Molecular Therapy of Virus- duction of AML1/MTG8 into one Aml1 allele resulted Associated Cancers, German Cancer Research Center, Heidelberg, Germany. in a phenotype very similar to the one obtained by Received 22 December 2005; revised 15 March 2006; accepted 23 March homozygous inactivation of Aml1, again suggesting that 2006; published online 1 May 2006 AML1/MTG8 acts as a transdominant repressor of Identification of AML1/MTG8-affected genes J Dunne et al 6068 AML1 functions (Yergeau et al., 1997). In addition, AML1/MTG8 binds and inhibits a variety of trans- cription factors relevant for myeloid differentiation (Westendorf et al., 1998; Jakubowiak et al., 2000; Pabst et al., 2001; Puccetti et al., 2002; Vangala et al., 2003; Zhang et al., 2004). The phenotypic consequences of AML1/MTG8 expression depend very much on the cellular context. For instance, ectopic AML1/MTG8 expression inhibits myeloid differentiation, but also induces apoptosis in leukaemic cell lines such as U937 (Burel et al., 2001). However, AML1/MTG8 supports the expansion of haematopoietic progenitor cells both ex vivo and in vivo by enhancing their self-renewal and, at least in some studies, by interfering with their maturation (Hiebert et al., 1996; Rhoades et al., 2000; Mulloy et al., 2002; Tonks et al., 2003; Baesecke et al., 2005). Nevertheless, AML1/MTG8 alone is not sufficient to induce leukae- mia in murine model systems. Furthermore, AML1/ MTG8 expression was found in cord blood at a 100-fold higher frequency than the riskto develop AML, and in AML patients in long-term remission (Kusec et al., 1994; Miyamoto et al., 2000; Mori et al., 2002). We examined the functions of endogenously ex- pressed AML1/MTG8 in terminally developed leukae- mic cells using small interfering RNAs (siRNAs). siRNA-mediated transient depletion of AML1/MTG8 restored myeloid differentiation, inhibited leukaemic clonogenicity and proliferation, and induced a senes- cence-associated cell cycle arrest (Heidenreich et al., 2003; Martinez et al., 2004). Thus, despite its inability to drive leukaemogenesis by its own, AML1/MTG8 seems to be essential for maintaining the leukaemic phenotype. In the current study, we used siRNAs to directly examine AML1/MTG8-dependent changes in gene expression in both t(8;21)-positive leukaemic cell lines and in primary blasts. We show that siRNA-mediated depletion of AML1/MTG8 results in an expression pattern suggesting the onset of myeloid differentiation and inhibition of leukaemic proliferation. Figure 1 siRNA-mediated depletion of AML1/MTG8 suppres- sion in Kasumi-1 cells. (a) AML1/MTG8 suppression in Kasumi-1 Results cells. Total protein was isolated 4 days after siRNA electropora- tions. AML1/MTG8 was detected with an antibody targeting the C-terminus of MTG8. Length standard is shown on the right. To examine AML1/MTG8-dependent changes in gene served as a loading control. (b) Transfection with AML1/MTG8 expression, we transfected t(8;21)-positive leukaemic siRNA induces p14ARF expression. Total RNA was isolated 4 days Kasumi-1 cells either with AML1/MTG8 siRNA after siRNA electroporations. Error bars indicate standard deviations. (c) AML1/MTG8 depletion raises C/EBPa transcript (siAGF1) or with a mismatch control siRNA (siAGF6). levels. Total RNA was isolated 4 days after siRNA electropora- The siRNA-mediated AML1/MTG8 depletion was tions. Error bars indicate standard deviations. (d) AML1/MTG8 validated by immunoblot analysis. In comparison to suppression increases the number of Sudan black-positive Kasumi- the mismatch control siAGF6, siAGF1 caused a 1 cells. Cells were stained with Sudan black4 days after siRNA substantial decrease in AML1/MTG8 protein without transfection. Original magnification was 100-fold. siAGF1 and siAM, AML1/MTG8 siRNAs; siAGF6, mismatch control siRNA; affecting the amount of mutated p53, which served siGL2, sequence-unrelated control siRNA. as a loading control (Figure 1a) (Banker et al., 1998). siRNA-mediated depletion of AML1/MTG8 caused a twofold increased expression of p14ARF (Figure 1b), an increase in Sudan black-positive cells (Figure 1d) and a established target gene of AML1/MTG8 (Linggi et al., decrease in CD34 surface expression (Martinez et al., 2002). Furthermore, transcript levels of the indirect 2004), thus indicating myeloid differentiation. target gene C/EBPa were also raised twofold (Figure 1c) Four days after a single siRNA treatment, total RNA (Westendorf et al., 1998; Pabst et al., 2001). Moreover, was isolated and subjected to global gene expression AML1/MTG8 depletion was associated with a twofold analysis using a human Unigene chip containing 4608

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6069 different cDNAs. We performed two independent protein (GFP) mRNA. Total RNA was isolated 4 days experiments with two arrays per experiment. Selection after transfection as in the previous set of experiments, criteria were an at least twofold changed expression and subjected to oligonucleotide array analysis with two in combination with a ratio between average change arrays used for RNA from siAGF1-treated cells, and and the corresponding standard deviation of at least 2. one array for each of the other samples. Genes Applying these criteria, 40 genes were identified with considered to show substantially changed expression changed expression levels upon AML1/MTG8 depletion upon AML1/MTG8 depletion had to fulfil the following (Table 1). Twenty-eight genes were up regulated and 12 requirements: expression levels had to change at least genes were downregulated. twofold when compared to each of the controls. In a second set of experiments, we employed oligo- Furthermore, the expression changes had to be larger nucleotide arrays to identify genes affected by the than the difference of expression levels between the two AML1/MTG8 status. In this case, Kasumi-1 cells were chips used for AML1/MTG8-suppressed cells. Forty-six either mocktransfected or transfected with AML1/ genes fulfilled these criteria, with 36 genes showing MTG8 siRNA, with a mismatch control siRNA, or with increased expression levels, and 10 genes showing a sequence-unrelated siRNA targeting green fluorescent decreased expression levels upon AML1/MTG8 deple- tion including MTG8 (RUNX1T1, twofold down). As MTG8 itself is not expressed in Kasumi-1 cells (Miyoshi et al., 1993), and since the AML1/MTG8 fusion Table 1 Genes affected by AML1/MTG8 depletion in Kasumi-1 cells: transcript should hybridize with the MTG8 probes on cDNA array analysis the array, this change reflects the siRNA-mediated Gene Cellular process siAGF1/siAGF6 reduction of AML1/MTG8 transcript levels (Table 2). ADAM12 Morphology, migration 0.41 AML1/MTG8 depletion affected several transcription ALDH3B1 Metabolism 2.1 factor genes (BTG1, ELK3, ID1, FOSL2, IFI16, ALOX5AP Immune response 2.3 MEF2C, NFE2), genes involved in signal transduction ARL15 Unknown 2.4 (ATP2B4, CXCR4, DEPDC6, DUSP6, IGFBP7, ATG16L2 Unknown 2.1 LAPTM5, MS4A3, PSTPIP, RASGRP2, SLA), and AZU1 Immune response 3.9 BAALC Unknown 0.32 genes associated with cell morphology, adhesion and BACH Metabolism 2.4 migration (ADAM12, CD34, CKLSF3, EDIL3, GCA, BAT2D1 Unknown 2.5 ICAM3, LCP1, MTSS1, MYO1F, S100A8, SELPLG). BPI Immune response 20 Furthermore, expression levels of several genes linked BTG1 Transcription 2.3 CD24 Immune response 2.7 with drug resistance or poor clinical prognosis (AZU1, CD302 Unknown 2.2 BAALC, CD34, PRG2, RNASE2, TSPAN7) changed CKLFSF3 Morphology, migration 2.0 upon AML1/MTG8 suppression. Fourteen genes CSF3R Signal transduction 2.1 (AZU1, BPI, BTG1, CD24, CTSG, DUSP6, ELK3, CTSG Immune response 7.2 ICAM3, IFI16 IGFBP7, LAPTM5, LYZ, NKG7, SLA) DUSP6 Signal transduction 0.39 ECM1 Morphology, migration 0.45 affected by AML1/MTG8 depletion were identified by EDIL3 Morphology, migration 0.39 both cDNA array and oligonucleotide array analysis ELK3 Transcription 0.47 (Figure 2). Moreover, 35 out of 45 genes showed an at FLJ12935 Unknown 3.0 least 1.3-fold change in each array analysis indicating a FOSL2 Transcription 0.49 GCA Vesicle formation 2.0 substantial extent of agreement (data not shown). ICAM3 Morphology, migration 2.8 Importantly, transcript levels of genes associated with IFI16 Transcription 0.48 interferon responses such as OAS genes, interferon IGFBP7 Signal transduction 3.6 genes or STAT1 were not affected by AML1/MTG8 IL6R Signal transduction 2.0 depletion (data not shown). None of the siRNAs used ITM2C Unknown 2.5 LAPTM5 Signal transduction 2.9 interfered with the expression of wild-type AML1 (data LCP1 Morphology, migration 2.4 not shown), further demonstrating the specificity of the LYZ Immune response 2.4 AML1/MTG8 siRNA approach. MTSS1 Morphology, migration 0.29 To validate observed changes in gene expression, NIP30 Unknown 2.4 NKG7 Unknown 2.7 we repeated the siRNA treatment of Kasumi-1 cells OS9 Unknown 2.1 followed by RNA isolation and subsequent real- SAMSN1 Unknown 0.43 time reverse transcription–polymerase chain reaction SLA Signal transduction 3.0 (RT–PCR) analysis. We examined the transcript levels SLC2A3 Metabolism 0.32 of 38 genes suggested by gene expression profiling to SPCS3 Unknown 2.1 TSPAN7 Development 0.38 be affected by AML1/MTG8 depletion. Changes in transcript levels were confirmed for 32 genes (Figure 3). Abbreviations: AML, acute myeloid leukaemia; si, small interfering. Twelve of those confirmed genes (AZU1, BPI, CD24, Changes in expression levels are shown as ratios between cells treated CTSG, DUSP6, ELK3, ICAM3, IGFBP7, LAPTM5, with active siRNA (siAGF1) and cells treated with control siRNA LYZ, NKG7, SLA) showed substantial AML1/MTG8- (siAGF6). Genes also found to be affected by oligonucleotide array analysis of AML1-MTG8-depleted Kasumi-1 cells are shown in bold, dependent changes in expression in all three analyses genes changed both in AML1/MTG8-depleted Kasumi-1 cells and (cDNA array analysis, oligonucleotide array analysis, AML blasts are underlined. real-time PCR analysis).

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6070 Table 2 Genes affected by AML1/MTG8 depletion in Kasumi-1 cells: oligonucleotide array analysis Gene Cellular process siAGF1 versus

siAGF6 siGFP Mock

ATP2B4 Signal transduction 2.5 2.7 2.1 AZU1 Immune response 2.4 2.1 2.3 BPI Immune response 19 20 18 BTG1 Transcription 2.0–2.2 2.0–2.1 2.5–2.7 C11orf21 Unknown 2.7 2.4 2.4 CAT Metabolism 2.1 2.3 3.0 CD24 Immune response 2.2–8.2 2.4–6.6 2.4–12 CD34 Morphology, migration 0.37 0.31 0.30 CEACAM6 Signal transduction 2.9–10 3.6–8.8 4.7–11 CKB Signal transduction 0.48 0.38 0.39 CST7 Immune response 9.3 7.5 6.5 CTSG Immune response 9.4 9.1 3.1 CXCR4 Morphology, migration 2.5–4–5 2.7–3.7 3.2–4.6 DEPDC6 Signal transduction 0.39 0.33 0.39 DUSP6 Signal transduction 0.43 0.37 0.23 Figure 2 Genes affected by AML1/MTG8 depletion. Venn ELK3 Transcription 0.48 0.36 0.46 diagram showing the overlaps among AML1/MTG8-depleted FNBP1 Morphology, migration 2.2 2.0 2.8 Kasumi-1 cells identified by cDNA and oligonucleotide array HSPA6 Metabolism 2.6 2.7 2.3 analysis, and of AML1/MTG8 siRNA-treated t(8;21)-positive ICAM3 Morphology, migration 2.7 2.3 2.7 leukaemic blasts. ID1 Transcription 0.38 0.32 0.45 IFI16 Transcription 0.49 0.40 0.48 IGFBP7 Signal transduction 4.6–5.1 4.2–5.0 2.8–3.6 LAPTM5 Signal transduction 3.0 3.5 5.7 transcript levels after siRNA-mediated AML1/MTG8 LST1 Signal transduction 2.7 3.4 2.6 depletion. One of the genes showing no significant LYZ Immune response 2.0 2.0 2.7 change in expression upon AML1/MTG8 depletion was MEF2C Transcription 0.35–0.39 0.28–0.38 0.34–0.36 ID1, which could also not be confirmed in Kasumi-1 by MS4A3 Signal transduction 11 15 41 RT–PCR. The expression levels of five genes (BAALC, MYO1F Morphology, migration 2.0 2.5 2.3 NFE2 Transcription 4.2 3.7 3.7 CD24, DUSP6, GCA, IL6R) changed less than twofold NKG7 Unknown 5.3 5.4 6.2 after siRNA treatment of SKNO-1 cells, but showed the NUCB2 Unknown 2.3 2.1 2.2 same tendency as in siRNA-treated Kasumi-1 cells. PLAC8 Unknown 4.2 4.0 2.8 We also examined global expression patterns in PRG1 Unknown 4.5–4.6 4.1–4.6 4.3–8.0 PRG2 Immune response 32 38 61 t(8;21)-positive AML blasts treated with either AML1/ PSTPIP1 Signal transduction 2.2 2.2 2.0 MTG8 siRNA or mismatch siRNA. RT–PCR analysis PTGER4 Immune response 2.6 2.7 2.1 confirmed a twofold decrease in AML1/MTG8 trans- RANBP2 Transport 0.44 0.30 0.47 cript levels upon treatment with siAGF1 (data not RASGRP2 Signal transduction 3.4 2.7–3.1 2.8–3.3 shown). Thirty-seven genes showed an at least three- RNASE2 Metabolism 3.9 4.4 4.4 RNASE3 Metabolism 7.8 6.9 6.3 fold increased transcript level upon AML1/MTG8 RUNX1T1 Transcription 0.49 0.42 0.46 depletion (Table 3). Electroporation with AML1/ S100A8 Immune response 4.2 4.7 4.5 MTG8 siRNA increased the expression of 12 genes SAMSN1 Unknown 2.1 2.7 2.0 (AZU1, CD24, CST7, GCA, IGFBP7, MS4A3, NKG7, SELPLG Morphology, migration 2.4 3.1 2.5 SLA Signal transduction 2.0–2.5 2.2–2.8 2.3 NUCB2, PLAC8, PRG2, RNASE2, S100A8) both in TRGC2 Immune response 2.2–2.9 3.0–3.9 2.6–3.7 Kasumi-1 cells and in AML blasts. Moreover, siRNA treatment of blasts obtained from a second patient Abbreviations: AML, acute myeloid leukaemia; GFP, green fluor- suffering from t(8;21)-positive AML reduced AML1/ escent protein; si, small interfering. Changes in expression levels are MTG8 mRNA levels twofold in comparison to AML1 shown as ratios between cells treated with active siRNA (siAGF1) and (RUNX1, data not shown) and increased transcript cells treated with control siRNAs (siAGF6, siGFP), or mock- transfected cells. Genes also found to be affected by cDNA array levels of BPI, CTSG and IGFBP7 4.6-fold, threefold and analysis of AML1-MTG8-depleted Kasumi-1 cells are shown in bold, 2.4-fold, respectively (Figure 5). Thus, in addition to the genes changed both in AML1/MTG8-depleted Kasumi-1 cells and genes identified by array analysis, BPI and CTSG may AML blasts are underlined. also be affected by AML1/MTG8 suppression in primary AML blasts. Moreover, the substantial overlap of AML1/MTG8-regulated genes strongly suggests that Next, we asked whether the AML1/MTG8-dependent siRNA-mediated AML1/MTG8 depletion has similar changes in transcript levels are common to t(8;21) consequences for both t(8;21)-positive cell lines and leukaemic cells, or if they are specific for Kasumi-1 cells. primary leukaemic blasts. We chose 14 different genes and examined the corres- As siRNAs may affect gene expression of unintended ponding transcript levels in siRNA-treated t(8;21)- targets with rather limited homology, we examined positive SKNO-1 cells by real-time PCR (Figure 4). possible off-target effects of AML1/MTG8 siRNA. Six of these genes (BPI, CTSG, IGFBP7, LAPTM5, Therefore, we transfected the t(8;21)-negative leukaemic NKG7, SLA) exhibited at least twofold changes in cell line U937 either with AML1/MTG8 siRNA or with

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6071

Figure 3 Real-time RT–PCR analysis of AML1/MTG8-dependent gene expression in Kasumi-1 cells. (a) Induced AML1/MTG8 target genes. (b) Repressed AML1/MTG8 target genes. Expression levels were determined at least in triplicates and were normalized to GAPDH expression. Relative expression levels (DDCt) were calculated relatively to expression levels in cells treated with control siRNA and are shown on a log 2 scale. Error bars indicate standard deviations. mismatch control siRNA (Figure 6). Notably, none of preceding the start of transcription. Next, we examined the examined genes showed a substantially changed intracellular binding of AML1/MTG8 by chromatin expression level upon transfection with AML1/MTG8 immunoprecipitation (ChIP) assays. To isolate speci- siRNA, indicating that the observed changes in siRNA- fically DNA sequences bound by AML1/MTG8, a treated t(8;21)-positive cells are a consequence of tandem affinity purification (TAP)-tagged version of AML1/MTG8 depletion, and are not caused by off- this fusion protein was ectopically expressed in U937 target effects of the AML1/MTG8 siRNA. cells. As a control, we used non-transfected U937. The BPI and ELA2 were the only genes affected by c-FOS promoter served as a further negative control. AML1/MTG8 expression with an established functional ChIP showed that not only BPI but also IGFBP7 and AML1 binding site in their promoters (Lennartsson SLA promoter sequences were bound by TAP-tagged et al., 2003; Lausen et al., 2006). To identify possible AML1/MTG8 (Figure 7). In contrast, c-FOS promoter AML1 (and therefore AML1/MTG8) binding sites sequences could not be precipitated, indicating the in the promoters of other genes influenced by AML1/ absence of any in vivo interaction with AML1/MTG8. MTG8 depletion, we inspected selected genes for Furthermore, none of the sequences could be precipi- occurrence of the TGYGGT consensus sequence in tated from lysates derived from control-transfected regions upstream of the start of transcription (Meyers U937 cells. In conclusion, these data suggest that et al., 1993). For instance, the IGFBP7 gene contains six AML1/MTG8 binds directly to upstream sequences of possible binding sites and the SLA gene four possible the BPI, IGFBP7 and SLA genes. This finding in binding sites for AML1/MTG8 within the 2 kb region combination with the observed increase of expression

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6072

Figure 4 AML1/MTG8-dependent gene expression in SKNO-1 cells. (a) siRNA-mediated AML1/MTG8 depletion in SKNO-1 cells. Total cell lysates were isolated 4 days after siRNA electroporations. AML1/MTG8 was detected with an antibody targeting the runt domain of AML1/MTG8. Length standard is shown on the left. Tubulin served as a loading control. (b) Induced AML1/MTG8 target genes. (c) Repressed AML1/MTG8 target genes. Expression levels were calculated as indicated in Figure 3. Relative expression levels are shown on a log 2 scale. Error bars indicate standard deviations.

upon AML1/MTG8 depletion implies that at least these entiation during the transition from the myeloblastic three genes may be direct target genes for AML1/ to the promyelocytic stage (Weiss and Olsson, 1987; MTG8. Zimmer et al., 1992; Garwicz et al., 2005). An initiation of myeloid differentiation is further indicated by increased expression levels of genes such as CD24, Discussion GCA, NKG7 (GIG1), NFE2, PRG1, PRG2, RNASE2, RNASE3 or S100A8 (MRP8). We addressed the role of AML1/MTG8 in both t(8;21)- In addition to initiating myeloid differentiation, positive cell lines and in patient blast cells by inhibiting AML1/MTG8 depletion affects leukaemic clonogenicity its endogenous expression with siRNAs followed by and proliferation, and irreversibly arrests Kasumi-1 gene expression profiling. In total, we identified 76 genes cells in the G1 phase of the cell cycle (Martinez et al., with changes in expression levels upon AML1/MTG8 2004). Several genes with increased expression levels depletion. upon AML1/MTG8 siRNA treatment (BTG1, IGFPB7, We have previously shown that AML1/MTG8 MS4A4, NFE2, SLA) have previously been shown siRNAs facilitate tumour growth factor b/vitamin D3- to inhibit cellular proliferation. For instance, IGFBP7 induced differentiation in the t(8;21)-positive cell lines (IGFBP-rP1, MAC25) binds to insulin and insulin- Kasumi-1 and SKNO-1 (Heidenreich et al., 2003). In like growth factor 1, thereby inhibiting proliferation agreement with this observation, suppression of AML1/ (Oh et al., 1996; Yamanaka et al., 1997). Interestingly, MTG8 results in a gene expression pattern indicative of IGFBP7 expression is downregulated in several types myeloid differentiation. For instance, AML1/MTG8 of cancer, and restoration of its expression induces depletion inhibits the expression of CD34, a marker for cellular senescence or apoptosis (Swisshelm et al., 1995; haematopoietic stem and early progenitor cells. Further- Lopez-Bermejo et al., 2000; Landberg et al., 2001; more, expression of several genes associated with the Mutaguchi et al., 2003). The transmembrane protein formation of azurophil granules (AZU1, BPI, CTSG, MS4A3 (HTm4) inhibits G1–S cell cycle transition in ELA2, LYZ, PRTN3) is raised upon AML1/MTG8 haematopoietic cells by binding to cyclin-dependent suppression suggesting the onset of granulocytic differ- kinase (CDK)-associated phosphatase–CDK2 complexes

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6073 Table 3 Oligonucleotide array analysis of siRNA-treated t(8;21) AML blasts Gene Cellular process siAGF1/siAGF6

ALDOC Metabolism 3.0 ANXA3 Signal transduction 7.0 ARHGDIA Morphology, migration 3.2 AZU1 Immune response 4.9 BNIP3 Signal transduction 4.6 SPFH1 Unknown 3.0 CALR Transcription 3.0 CD24 Immune response 3.0–11 CEACAM8 Immune response 8.6 CLC Immune response 5.7 CSPG2 Development 3.0–4.0 CST7 Immune response 4.3 CSTA Development 3.0 ELA2 Immune response 3.5 FBXL5 Transport 3.0 GCA Vesicle formation 3.5 HGF Signal transduction 4.0 HPR, HP Immune response 3.0–3.2 IGFBP7 Signal transduction 6.1 KBTBD11 Unknown 3.0 MNDA Transcription 4.0 MS4A3 Signal transduction 3.7 NKG7 Unknown 3.7 NUCB2 Unknown 4.9 NXF3 Transport 3.0 P4HA1 Unknown 3.0 PLAC8 Unknown 9.2 PRG2 Immune response 5.3 PRTN3 Morphology, migration 7.0 PTGS2 Immune response 3.5 RETN Signal transduction 5.7 RNASE2 Metabolism 3.2–4.9 S100A8 Immune response 4.9 SERPINB2 Signal transduction 3.0 Figure 5 AML1/MTG8-dependent gene expression in primary SLC2A5 Metabolism 4.0 AML blasts. Expression levels were normalized to GAPDH SPP1 Immune response 6.5 expression. Expression levels were calculated as indicated in SSR3 Transport 3.2 Figure 3. Relative expression levels are shown on a log 2 scale. THBS1 Morphology, migration 3.0–3.7 Error bars indicate standard deviations.

Abbreviations: AML, acute myeloid leukaemia; GFP, green fluor- escent protein; si, small interfering. Changes in expression levels are tion of AML1/MTG8 siRNA, but not of control shown as ratios between cells treated with active siRNA (siAGF1) and cells treated with control siRNA (siAGF6). Genes found to be affected siRNAs. This similarity between cell lines and primary by cDNA and oligonucleotide array analysis of AML1-MTG8- patient material supports the relevance of the cell culture depleted Kasumi-1 cells are shown in bold, genes changed both in model for AML with t(8;21). Furthermore, it strongly AML1/MTG8-depleted Kasumi-1 cells and AML blasts are under- suggests that directly and specifically targeting AML1/ lined. MTG8 may be of benefit for patients suffering from t(8;21)-associated AML. The effects of AML1/MTG8 on global gene expres- (Donato et al., 2002). The Src-like adaptor protein SLA sion have also been examined by its ectopic expression in inhibits NIH3T3 proliferation by blocking the Src t(8;21)-negative leukaemic cell lines. In two studies, binding site of platelet-derived growth factor AML1/MTG8 expression was induced in the mono- (Roche et al., 1998), and impairs erythropoietin (Epo)- blastic cell line U937 (Alcalay et al., 2003; Fliegauf et al., induced erythroblast survival by binding to Epo 2004). Interestingly, ectopic AML1/MTG8 expression receptor (Lebigot et al., 2003). In conclusion, AML1/ inhibited the expression of four genes (LCP1, LYZ, MTG8 may support leukaemic proliferation by directly NFE2, SLA), which are induced upon siRNA-mediated or indirectly downmodulating the expression of anti- AML1/MTG8 depletion. Recently, Cammenga and proliferative genes. Nimer made the results of a gene array study in the Importantly, we observed similar changes in expres- NCBI GEO database available (Barrett et al., 2005), sion pattern upon AML1/MTG8 depletion in both where they examined the consequences of ectopic Kasumi-1 and in t(8;21)-positive AML blasts. Both cell AML1/MTG8 expression in human CD34 þ cells isola- systems react to AML1/MTG8 suppression by increased ted from peripheral blood (Accession no. GSE2049) expression of differentiation-associated and antiproli- (Mulloy et al., 2005). Interestingly, 34 genes found by us ferative genes. Furthermore, both cell types share 12 to be affected in Kasumi-1 cells by AML1/MTGT8 genes with increased transcript levels after the applica- siRNAs were also detected in their analysis (Table 4).

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6074 Table 4 Concordance of gene expression changes upon depletion and ectopic expression of AML1/MTG8 Gene AML1/MTG8-de- AML1/MTG8-positive Accord pleted progenitors

Kasumi-1 AML blasts

ALDOC NS I I No ALOX5AP I NS D Yes ARHGDIA NS I I No ATP2B4 I NS D Yes AZU1 I I D Yes BAT2D1 INSNSNo BPI I NS D Yes BTG1 I NS D Yes CALR NS I NS No CKB D NS I Yes CLC NS I D Yes CSF3R I NS D Yes CSPG2 NS I D Yes CST7 I I D Yes CSTA NS I D Yes CTSG I NS D Yes CXCR4 I NS D Yes DUSP6 D NS I Yes ELA2 NS I D Yes Figure 6 Effects of AML1/MTG8 siRNAs on gene expression in FNBP1 I NS D Yes U937 cells. Total RNA was isolated 4 days after siRNA GCA I I D Yes electroporations. Relative expression levels in AML1/MTG8 HSPA6 I NS D Yes siRNA-treated cells compared to mismatch siRNA-treated cells ICAM3 I NS D Yes are shown. Expression levels of the indicated genes were examined ID1 D NS I Yes by real-time RT–PCR and were normalized to GAPDH. Error bars IGFBP7 I I D Yes indicate standard deviations. LAPTM5 I NS D Yes LCP1 I NS D Yes LST1 I NS D Yes MNDA NS I D Yes MS4A3 I I D Yes MYO1F I NS D Yes NFE2 I NS D Yes NKG7 I I D Yes NUCB2 I I D Yes P4H1 NS I I No PRG1 I NS D Yes PRTN NS I D Yes RANBP2 D NS I Yes RASGRP2 I NS D Yes RNASE2 I I D Yes RNASE3 I NS D Yes S100A8 I I D Yes SELPLG I NS D Yes Figure 7 Binding of AML1/MTG8 to target genes. ChIP analysis SLA I NS D Yes showing the binding of AML1/MTG8 to upstream regions of BPI, SLC2A3 D NS I Yes SLA and IGFBP7 in AML1/MTG8-expressing U937 or untrans- SPP1 NS I D Yes fected control U937 cells. Input (10% of total) is shown in the bottom row. Abbreviations: AML, acute myeloid leukaemia; D, decreased; I, increased; NS, not substantially changed. Array data for AML1/ MTG8-depleted cells are shown in the second and third column, data for AML1/MTG8-transduced CD34+ cells from peripheral blood are In each case, ectopic expression of AML1/MTG8 in shown in the fourth column (GEO Accession no. GSE2049, CD34 þ cells showed the opposite effect on expression Cammenga and Nimer 2005). levels when compared to the consequences of siRNA- mediated AML1/MTG8 suppression. Moreover, for 16 out of 20 genes also detected in AML1/MTG8-depleted In this context, several genes are repressed by either of AML blasts opposite changes in expression levels were these fusion products when ectopically expressed in noted. This remarkable agreement between these two U937 cells (Alcalay et al., 2003). We identified two studies strongly suggests that at least this subset of genes of these genes, NFE2 and CSF3R, to be repressed is directly or indirectly affected by AML1/MTG8. by AML1/MTG8 in Kasumi-1 cells. Moreover, MLL, Different leukaemic fusion such as AML1/ whose gene is frequently targeted by chromosomal MTG8 or PML-RARa interfere with common biologic rearrangements, has been found to associate with six processes such as differentiation and proliferation. (BPI, CD24, CST7, GCA, NKG7, S100A8) out of

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6075 14 genes affected by AML1/MTG8 depletion in both of t(8;21)-positive cells with AML1/MTG8 siRNA t(8,21)-positive leukaemic cell lines and AML blasts also raised the expression levels of these genes. As (Guenther et al., 2005). Thus, AML1/MTG8-mediated APL and monocytic AMLs represent more differen- regulation of these genes might, at least in part, depend tiated subtypes of AML than t(8;21)-associated AMLs, on MLL. Moreover, these genes may also be affected these concordant changes in expression can be explained by, for instance, MLL-AF9 or MLL-ENL, indicating an by the onset of myeloid differentiation upon AML1/ intersection of genes and associated cellular processes as MTG8 depletion. common targets for these aberrant transcriptional High expression of BAALC as well as low expres- regulators. sion of either AZU1, PRG2 or RNASE2 correlate with Comparison of AML1/MTG8-affected genes with a poor treatment outcome of AML (Baldus et al., gene expression profiles obtained for AML patient 2003a, b; Bullinger et al., 2004; Valk et al., 2004). The material identifies two groups of genes common to both leukaemic cell lines Kasumi-1 and SKNO-1 have been approaches (Table 5). One group consists of genes with isolated from patients suffering from, and finally expression levels typical for CBF leukaemias, including succumbing to, leukaemic relapse (Asou et al., 1991; t(8;21)-positive AMLs. Application of AML1/MTG8 Matozaki et al., 1995). Interestingly, transfection of siRNA reduced the expression of those genes (BAALC, these leukaemic cell lines with AML1/MTG8 siRNAs CBFA2T1, CD34, DUSP6) with an elevated expression inhibited BAALC expression and enhanced the expres- in leukaemic cells, and increased transcript levels of sion of AZU1, PRG2 and RNASE2 (Table 5). Further- those (ATP2B4, LAPTM5, LCP1) with diminished more, expression of CD34 and TSPAN7 (TM4SF2), expression in blast cells (Yagi et al., 2003; Bullinger which correlates with drug resistance (Bullinger et al., et al., 2004; Ross et al., 2004; Valk et al., 2004). A 2004; Heuser et al., 2005), is diminished in AML1/ second group (CALR, CST7, HGF, IGFBP7, MEF2C, MTG8-depleted Kasumi-1 cells (Table 5). These find- NKG7, PRG1 and RNASE3) contains genes with ings suggest that AML1/MTG8 suppression or inhibi- increased expression levels in t(15;17)-positive acute tion of its function may enhance drug sensitivity of promyelocytic leukaemia (APL) or in mainly monocytic t(8;21)-positive leukaemic cells, and might improve the AMLs with 11q23 abnormalities (Table 5). Treatment outcome of t(8;21)-positive leukaemia patients. In summary, we show that siRNA-mediated suppres- sion of AML1/MTG8 results in a gene expression Table 5 Comparison of AML1/MTG8-dependent and leukaemia- pattern indicative of myeloid differentiation paired with subtype specific gene expression a compromised leukaemic proliferation in both leukae- mic cell lines and primary AML blasts. These results Gene Expression in leukaemia AML1/MTG8-depleted allude to a central role for this fusion gene in these two Kasumi-1 AML blasts processes, and suggest that AML1/MTG8 is essential for the maintenance of t(8;21)-positive AML. t(8;21) BAALC I D D RUNX1T1 I D NS CD34 I D NS Materials and methods LAPTM5 D I NS LCP1 D I NS Cell culture Kasumi-1 cells (Asou et al., 1991) and U937 cells (Sundstrom t(15;17) and Nilsson, 1976) were cultured in RPMI1640 containing CALR I NS I CST7 I I I 10% fetal calf serum (FCS). SKNO-1 cells (Matozaki et al., HGF I NS I 1995) were maintained in RPMI1640 supplemented with 20% FCS and 7 ng/ml granulocyte–macrophage colony-stimulating 11q23 factor (GM-CSF). Primary t(8;21)-positive AML blasts were IGFBP7 I I I cultured in Iscove’s modified Dulbecco’s medium (Invitrogen, MEF2C I I NS Karlsruhe, Germany) containing 20% Stemspan BIT 9500 NKG7 I I I plus 20 ng/ml interleukin (IL)-3, IL-6, granulocyte colony- PRG1 I I NS stimulating factor, GM-CSF, 50 ng/ml stem cell factor RNASE3 I I NS (Preprotech, London, UK). CBF leukaemias ATP2B4 D I NS Patients DUSP6 I D NS The leukaemia sample used for the oligonucleotide array analysis (Table 3) was obtained from bone marrow of a Poor prognosis/drug resistance 52-year-old male patient with AML FAB M2 in St Bartho- AZU1 D I I lomew’s Hospital. Mononuclear cells were purified from white BAALC I D NS blood cells and neutrophils by Ficoll–Hypaque density- CD34 I D NS PRG2 D I I gradient centrifugation. The sample was frozen in liquid RNASE2 D I I nitrogen with 10% dimethyl sulphoxide in FCS. The TSPAN7 I D NS karyotype was 46, XY, del(7)(q32q36), t(8;21)(q22;q22). Bone marrow blasts used for RT–PCR detection (Figure 5) were Abbreviations: AML, acute myeloid leukaemia; D, decreased; I, isolated from a 67-year-old male patient with AML FAB M2 increased; NS, not substantially changed. subtype by Ficoll–Isopaque density-gradient centrifugation.

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6076 Cytogenetic analysis showed the karyotype 46 XY, t(8;21) ACG CGG AAU ACU UCG AdTdT-30; antisense, 50-UCG (q22;q22). The presence of AML/MTG8 fusion transcripts AAG UAU UCC GCG UAC GdTdT-30), unrelated sequence was confirmed by RT–PCR. Both patients gave informed control siRNA siGFP (sense, 5´ -GCA AGC UGA CCC UGA consent. AGU UCA U-30; antisense, 50-GAA CUU CAG GGU CAG CUU GCC G-30). siRNA transfections Kasumi-1 cells and AML blasts were transfected with the RNA and protein isolation indicated siRNA concentrations using a Fischer EPI 3500 If not indicated otherwise, total RNA was isolated 4 days after electroporator (Fischer, Heidelberg, Germany) as described siRNA electroporation using RNeasy columns (Qiagen, previously (Martinez et al., 2004). The following siRNAs were Hilden, Germany). Total protein was precipitated by adding used: AML1/MTG8 siRNAs siAGF1 (sense, 50-CCU CGA two volumes of acetone to the flow through of the RNeasy AAU CGU ACU GAG AAG-30; antisense, 50-UCU CAG columns and resuspended in urea buffer (9 M urea, 4% (w/w) UAC GAU UUC GAG GUU-30) and siAM (same target site 3-[(3-cholamidopropyl)-dimethylammonio]-propansulphonat as siAGF1, contains dT 30-overhangs; sense, 50-CCU CGA (CHAPS), 1% (w/w) dithiothreitol). AAU CGU ACU GAG AdTdT-30; antisense, 50-UCU CAG UAC GAU UUC GAG GdTdT-30), mismatch control cDNA microarrays siAGF6 (sense, 50-CCU CGA AUU CGU UCU GAG cDNA microarrays containing 4608 different cDNAs spotted AAG-30; antisense, 50-UCU CAG AAC GAA UUC GAG twice were generated using an Affymetrix 417 Arrayer GUU-30), unrelated sequence control siGL2 (sense, 50-CGU (Affymetrix, St Clara, CA, USA). Total RNA (50 mg) were

Table 6 Primer sequences for real-time RT–PCR cDNA Forward primer Reverse primer

ADAM12 GGACCATTCAGACAATCCCCT ATTCATCCCGAAATTGTGGC ALDH3B1 TTTTGGAGGAGTGGGTGCC TGGGAGAAGGTGTCGAAGGA ALOX5AP CCTTTGAGCGGGTCTACACTG CCAGAGCACAGCGAGGAAA AML1/MTG8 AATCACAGTGGATGGGCCC TGCGTCTTCACATCCACAGG AZU1 AACCTGAACGACCTGATGCTG ATCGTCACGCTGCTGGTGA BAALC TGTCCCCACTGGCATTACTCA TGCAGGATCACTGCCAACC BPI TGCAGCAGCCACATCAACA GTTGGATCAGCCCACCGAC BTG1 GCCGTGTCCTTCATCTCCAA GGTAACCCGATCCCTTGCAT CD24 CAGGCCGGAGTGCAGTG TCGCTTGAGCCCGGG CD302 GTACTGACCATGGAGCGGACA TCAAACCACTTGAAATCTGCA CD34 TCCAGAAACGGCCATTCAG CCCACCTAGCCGAGTCACAA CSF3R CCAGTCCTTCTCCGCCATC TGATACAGACTGGCGGGCTC CST7 CCAACCACACCTTGAAGCAGA GGGTCAGTGACAACGGAGAAC CTSG TCCTGGTGCGAGAAGACTTTG GGTGTTTTCCCGTCTCTGGA DUSP6 AGCTCAAGGACGAGGGCTG GGAGAACTCGGCTTGGAACTT EDIL3 GGATCATGAAGCGCTCGGTA TTCACATGGATTGGGATCACA FOSL2 GGGTTTGCCAAACGCCTAAT AGCTGGCTAGGACACCCGT GAPDH GAAGGTGAAGGTCGGAGTC GAAGATGGTGATGGGATTTC GCA CTACCCGGGATACGGAGGAG CGAGGAGTATAGCTGGGCCTG HPRT TGACACTGGCAAAACAATGCA AGCTTGCGACCTTGACCATC HSPA6 GAGTGGCTGCCAAAAACTCG CCCTAAGGCTTTCCTCTTGCA ICAM3 TCCCGTTCTTCGTCAACGTAA AGCCTCAATGTCCATCACCAC ID1 GGGCAAGAGGAATTACGTGC TTGTTCTCCCTCAGATCCGG IFI16 ACTTCACCTGCACCCTCCAC CAGGAGAGGGAGGCTGAGTCT IGFBP7 GAAGTAACTGGCTGGGTGCTG GCTGATGCTGAAGCCTGTCC IL6R AAAGGCTGTGCTCTTGGTGAG GAATACTGGCACGGCTCCTG ITM2C GCCAAGAACTGCAATGCCA CGCAGATGAGCGTCTCCAC LAPTM5 CAAGTGCATGAACTCGGTGG TTCCTCGTAGGACGGCAGG LCP1 TGATCCTGATTGTCGGCATG CAATGCCATCTCCAACAGCA LYZ TGCTGCAAGATAACATCGCTG CCATGCTCTAATGCCTTGTGG MS4A3 CCAAGCCATAAACAACCCCA TTCTGGTCCCGTCTCACTGC MTSS1 ATGGCCACCAACACACGTG CCTCATGCACATCCTGGTGA NFE2 CCAAGGTGTGTTCAAAGAGGC GGAGCCGAGTCAGGGAAGAC NIP30 AGGTGAGGTCGTCTCGCCT CATCCCCTCCATCCATAATGA NKG7 CTGATTGCTTTGAGCACCGA CCTGATATGATGTCCCCATGC PLAC8 TTCCTGCTCGGAACCTTGTT CACACATGCCTGTCTGCCAG PRG2 TGGATTGGAGGCAGGATCA TGCCGTCAACCCACTGAAA PSTPIP1 GCTGCAGTTCAAAGATGCCTT CTTTGCACATCTTCCTGCCA RNASE2 CCCCTGAACCCCAGAACAA ACCATGTTTCCCAGTCTCCG S100A8 CGAGCTGGAGAAAGCCTTGA GACGTCTGCACCCTTTTTCC SAMSN1 TCAAAGCCAGTGACTCCATGG TCGAAAGCTGTCCCGGTTAC SLA GGCTCACCTTCCAGTGCCT TTCTGACGCAAGGTGACAGG SLC2A3 GGTCTGAAGAGCTATGGCCG AACCGCTGGAGGATCTGCTT SPCS3 ACTGTTCGCCTTCTCGCTGA GTGGTGATGAAGCAGCCGA TSPAN7 TCCCAGTTAATTGGCATGCTG TCATACTGATTGGCCGTGATG

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6077 used to synthesize Cy3- and Cy5-labeled cDNAs according to Chromatin immunoprecipitation published protocols (http://brownlab.stanford.edu). Micro- U937 cells were electroporated with pNTAPRC1 encoding array experiments and analyses including standardization a TAP-tagged AML1/MTG8 under the control of a cyto- and quality control were performed as described (Berwanger megalovirus promoter (Philippar et al., 2004). ChIP assays et al., 2002). were performed using the ChIP assay kit according to the manufacturer (Upstate, Lake Placid, NY, USA) with the following modifications. TAP-AML1/MTG8 complexes were Oligonucleotide arrays isolated using immunoglobuline G sepharose 6 fast flow beads RNA isolation, labelling and hybridization onto HG- (Amersham Biosciences, Freiburg, Germany). Eluted DNA U133_Plus2 Gene chip arrays (Affymetrix, St Clara, CA, fragments were amplified by PCR, separated by agarose gel USA) were performed as described (Debernardi et al., 2003). electrophoresis and visualized by ethidium bromide staining. Bioconductor was used to calculate the level of expression for The primers for BPI (forward, 50-CCA CCC ACC TTC AAG each gene (Gentleman et al., 2004). The data were normalized GGA GT-30; reverse, 50-TCC ATT GTG GTT AGG GCC C-30), by quantile normalization as implemented in the affymetrix c-fos (forward, 50-CTC TCA GAG TCG CGG TCT GAC-30; package of Bioconductor. All the statistical analysis was done reverse, 50-CGA CCC TTC CCC CCA TAT AA-30), IGFBP7 with R (http://www.R-project.org). Empirical Bayesian meth- (forward, 50-CAG GAG TTT GAG ACC AGC CTG-30; reverse, odology, as implemented in the EBarrays package of 50-CTC CCT GTG TCG CCT AGT CTG-30)andSLA Bioconductor, was used and leads to shrinkage estimates of (forward, 50-TTT GCG AGC CTG GTC ATT G-30; reverse, fold changes. Genes were identified as differentially expressed 50-TTG AGG GAG GTC TGC ACT GC-30) were designed with with the assumption of the Log-Normal model of gene the Primer Express software (Applied Biosystems, Foster City, expression. CA, USA).

Real-time RT–PCR and immunoblotting Acknowledgements Real-time RT–PCR analyses were performed as described (Martinez et al., 2004). Primer sequences are shown in Table 6. We thankTracy Chaplin for her expert processing of the Immunoblotting was performed as described (Martinez et al., Affymetrix chips, Gael Molloy for helpful advice and Kerstin 2004). The following antibodies were used: AML1/MTG8, Go¨ rlich for excellent technical assistance. This workwas ETO C-20 (sc-9737; Santa Cruz Biotechnology, Heidelberg, supported by grants from the Deutsche Krebshilfe (10-2104- Germany; 1:1000); p53 Bp53-12 (sc-263; Santa Cruz Bio- He2) and the Wilhelm-Sander-Stiftung (2003.169.1) to JK and technology, Heidelberg, Germany; 1:5000) and tubulin Ab-4 OH and by the Kompetenznetz ‘Akute und chronische (1 mg/l, MS-719-P0; NeoMarkers, Fremont, CA, USA). Leuka¨ mien’ (BMBF 01GI9974) to AN.

References

Alcalay M, Meani N, Gelmetti V, Fantozzi A, Fagioli M, Follows GA, Tagoh H, Lefevre P, Hodge D, Morgan GJ, Orleth A et al. (2003). J Clin Invest 112: 1751–1761. Bonifer C. (2003). EMBO J 22: 2798–2809. Amann JM, Nip J, Strom DK, Lutterbach B, Harada H, FrankR, Zhang J, Uchida H, Meyers S, Hiebert SW, Nimer Lenny N et al. (2001). Mol Cell Biol 21: 6470–6483. SD. (1995). Oncogene 11: 2667–2674. Asou H, Tashiro S, Hamamoto K, Otsuji A, Kita K, Kamada Garwicz D, Lennartsson A, Jacobsen SE, Gullberg U, N. (1991). Blood 77: 2031–2036. LindmarkA. (2005). Haematologica 90: 38–44. Baesecke J, Schwieger M, Griesinger F, Schiedlmeier B, Gelmetti V, Zhang J, Fanelli M, Minucci S, Pelicci PG, Lazar Wulf G, Trumper L et al. (2005). Leukemia Lymphoma 46: MA. (1998). Mol Cell Biol 18: 7185–7191. 265–272. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Baldus CD, Tanner SM, Kusewitt DF, Liyanarachchi S, Choi Dudoit S et al. (2004). Genome Biol 5: R80. C, Caligiuri MA et al. (2003a). Exp Hematol 31: 1051–1056. Guenther MG, Jenner RG, Chevalier B, Nakamura T, Croce Baldus CD, Tanner SM, Ruppert AS, Whitman SP, Archer CM, Canaani E et al. (2005). Proc Natl Acad Sci USA 102: KJ, Marcucci G et al. (2003b). Blood 102: 1613–1618. 8603–8608. Banker DE, Radich J, Becker A, Kerkof K, Norwood T, Heidenreich O, Krauter J, Riehle H, Hadwiger P, John M, Willman C et al. (1998). Clin Cancer Res 4: 3051–3062. Heil G et al. (2003). Blood 101: 3157–3163. Barrett T, SuzekTO, Troup DB, Wilhite SE, Ngau WC, Heuser M, Wingen LU, Steinemann D, von Neuhoff GC, Ledoux P et al. (2005). Nucleic Acids Res 33: D562–D566. Tauscher M, Bullinger L et al. (2005). Haematologica 90: Berwanger B, Hartmann O, Bergmann E, Bernard S, Nielsen 1484–1492. D, Krause M et al. (2002). Cancer Cell 2: 377–386. Hiebert SW, Downing JR, Lenny N, Meyers S. (1996). Curr Bullinger L, Dohner K, Bair E, Frohling S, SchlenkRF, Top Microbiol Immunol 211: 253–258. Tibshirani R et al. (2004). N Engl J Med 350: 1605–1616. Hildebrand D, Tiefenbach J, Heinzel T, Grez M, Maurer AB. Burel SA, Harakawa N, Zhou L, Pabst T, Tenen DG, Zhang (2001). J Biol Chem 276: 9889–9895. DE. (2001). Mol Cell Biol 21: 5577–5590. Jakubowiak A, Pouponnot C, Berguido F, Frank R, Mao S, Debernardi S, Lillington DM, Chaplin T, Tomlinson S, Amess Massague J et al. (2000). J Biol Chem 275: 40282–40287. J, Rohatiner A et al. (2003). Genes Cancer 37: Kusec R, Laczika K, Knobl P, Friedl J, Greinix H, Kahls P 149–158. et al. (1994). Leukemia 8: 735–739. Donato JL, Ko J, KutokJL, Cheng T, ShirakawaT, Mao XQ Landberg G, Ostlund H, Nielsen NH, Roos G, Emdin S, et al. (2002). J Clin Invest 109: 51–58. Burger AM et al. (2001). Oncogene 20: 3497–3505. Erickson P, Gao J, Chang KS, Look T, Whisenant E, Lausen J, Liu S, Fliegauf M, Lubbert M, Werner MH. (2006). Raimondi S et al. (1992). Blood 80: 1825–1831. Oncogene 25: 1349–1357. Fliegauf M, StockM, Berg T, Lubbert M. (2004). Oncogene Lebigot I, Gardellin P, Lefebvre L, Beug H, Ghysdael J, 23: 9070–9081. Quang CT. (2003). Blood 102: 4555–4562.

Oncogene Identification of AML1/MTG8-affected genes J Dunne et al 6078 Lennartsson A, Pieters K, UllmarkT, Vidovic K, Gullberg U. Pabst T, Mueller BU, Harakawa N, Schoch C, Haferlach T, (2003). Biochem Biophys Res Commun 311: 853–863. Behre G et al. (2001). Nat Med 7: 444–451. Linggi B, Muller-Tidow C, van de Locht L, Hu M, Nip J, Philippar U, Schratt G, Dieterich C, Muller JM, Galgoczy P, Serve H et al. (2002). Nat Med 8: 743–750. Engel FB et al. (2004). Mol Cell 16: 867–880. Lopez-Bermejo A, Buckway CK, Devi GR, Hwa V, Plymate Puccetti E, Obradovic D, Beissert T, Bianchini A, Washburn SR, Oh Y et al. (2000). Endocrinology 141: 4072–4080. B, Chiaradonna F et al. (2002). Cancer Res 62: 7050–7058. Lutterbach B, Westendorf JJ, Linggi B, Patten A, Moniwa M, Rhoades KL, Hetherington CJ, Harakawa N, Yergeau DA, Davie JR et al. (1998). Mol Cell Biol 18: 7176–7184. Zhou L, Liu LQ et al. (2000). Blood 96: 2108–2115. Martinez N, Drescher B, Riehle H, Cullmann C, Vornlocher Roche S, Alonso G, Kazlauskas A, Dixit VM, Courtneidge HP, Ganser A et al. (2004). BMC Cancer 4: 44. SA, Pandey A. (1998). Curr Biol 8: 975–978. Matozaki S, Nakagawa T, Kawaguchi R, Aozaki R, Ross ME, Mahfouz R, Onciu M, Liu HC, Zhou X, Song G Tsutsumi M, Murayama T et al. (1995). Br J Haematol 89: et al. (2004). Blood 104: 3679–3687. 805–811. Sundstrom C, Nilsson K. (1976). Int J Cancer 17: 565–577. MelnickAM, Westendorf JJ, Polinger A, Carlile GW, Arai S, Swisshelm K, Ryan K, Tsuchiya K, Sager R. (1995). Proc Natl Ball HJ et al. (2000). Mol Cell Biol 20: 2075–2086. Acad Sci USA 92: 4472–4476. Meyers S, Downing JR, Hiebert SW. (1993). Mol Cell Biol 13: Tonks A, Pearn L, Tonks AJ, Pearce L, Hoy T, Phillips S et al. 6336–6345. (2003). Blood 101: 624–632. Miyamoto T, Weissman IL, Akashi K. (2000). Proc Natl Acad ValkPJ, VerhaakRG, Beijen MA, ErpelinckCA, Barjesteh Sci USA 97: 7521–7526. van Waalwijkvan Doorn-Khosrovani S, Boer JM et al. Miyoshi H, Kozu T, Shimizu K, Enomoto K, Maseki N, (2004). N Engl J Med 350: 1617–1628. Kaneko Y et al. (1993). EMBO J 12: 2715–2721. Vangala RK, Heiss-Neumann MS, Rangatia JS, Singh SM, Mori H, Colman SM, Xiao Z, Ford AM, Healy LE, Schoch C, Tenen DG et al. (2003). Blood 101: 270–277. Donaldson C et al. (2002). Proc Natl Acad Sci USA 99: Wang J, Hoshino T, Redner RL, Kajigaya S, Liu JM. (1998). 8242–8247. Proc Natl Acad Sci USA 95: 10860–10865. Morohoshi F, Mitani S, Mitsuhashi N, Kitabayashi I, Wang Q, Stacy T, Binder M, Marin-Padilla M, Sharpe AH, Takahashi E, Suzuki M et al. (2000). Gene 241: 287–295. SpeckNA. (1996). Proc Natl Acad Sci USA 93: 3444–3449. Mulloy JC, Cammenga J, MacKenzie KL, Berguido FJ, Weiss J, Olsson I. (1987). Blood 69: 652–659. Moore MA, Nimer SD. (2002). Blood 99: 15–23. Westendorf JJ, Yamamoto CM, Lenny N, Downing JR, Mulloy JC, Jankovic V, Wunderlich M, Delwel R, Cammenga Selsted ME, Hiebert SW. (1998). Mol Cell Biol 18: 322–333. J, Krejci O et al. (2005). Proc Natl Acad Sci USA 102: Yagi T, Morimoto A, Eguchi M, Hibi S, Sako M, Ishii E et al. 4016–4021. (2003). Blood 102: 1849–1856. Mutaguchi K, Yasumoto H, Mita K, Matsubara A, Shiina H, Yamanaka Y, Wilson EM, Rosenfeld RG, Oh Y. (1997). Igawa M et al. (2003). Cancer Res 63: 7717–7723. J Biol Chem 272: 30729–30734. Nisson PE, Watkins PC, Sacchi N. (1992). Cancer Genet Yergeau DA, Hetherington CJ, Wang Q, Zhang P, Sharpe Cytogenet 63: 81–88. AH, Binder M et al. (1997). Nat Genet 15: 303–306. Oh Y, Nagalla SR, Yamanaka Y, Kim HS, Wilson E, Zhang J, Kalkum M, Yamamura S, Chait BT, Roeder RG. Rosenfeld RG. (1996). J Biol Chem 271: 30322–30325. (2004). Science 305: 1286–1289. Okuda T, van Deursen J, Hiebert SW, Grosveld G, Downing Zimmer M, Medcalf RL, FinkTM, Mattmann C, Lichter P, JR. (1996). Cell 84: 321–330. Jenne DE. (1992). Proc Natl Acad Sci USA 89: 8215–8219.

Oncogene