Continuous Fli-1 Expression Plays an Essential Role in the Proliferation and Survival of F-Mulv-Induced Erythroleukemia and Human Erythroleukemia

Home , Engrailed (gene)

Continuous Fli-1 expression plays an essential role in the proliferation and survival of F-MuLV-induced erythroleukemia and human erythroleukemia

J-W Cui1,2,4, LM Vecchiarelli-Federico1,3,4, Y-J Li1, G-J Wang2 and Y Ben-David1,3

1Department of Molecular and Cellular Biology, Sunnybrook Health Sciences Centre, Toronto, Canada; 2Department of Hematology, First Hospital of Jilin University, Changchun, China and 3Department of Medical Biophysics, University of Toronto, Toronto, Canada

Erythroleukemia induced by Friend Murine Leukemia Virus to its role in the initiation and progression of erythroleukemo- (F-MuLV) serves as a powerful tool for the study of multistage genesis in mice. carcinogenesis and hematological malignancies in mice. Fli-1, a proto-oncogene and member of the Ets family, is activated Earlier studies of genes associated with retroviral-induced through viral integration in F-MuLV-induced erythroleukemia, neoplasias have contributed to our understanding of gene and is the most critical event in the induction of this disease. function, as well as to the processes of differentiation and Fli-1 aberrant regulation is also observed in human mali- transformation. A prime example of this is erythroleukemia gnancies, including Ewing’s sarcoma, which is often linked to induced by Friend Murine Leukemia Virus (F-MuLV). Upon expression of the EWS/Fli-1 fusion oncoprotein. Here we injection of F-MuLV, strain-specific neonates develop a multi- examined the effects of Fli-1 inhibition to further elucidate its role in these pathological occurrences. The constitutive stage disease characterized by erythroblastic expansion leading suppression of Fli-1, through RNA interference (RNAi), inhibits to anemia, splenomegaly, and eventually death within 8 weeks growth and induces death in F-MuLV-induced erythroleukemia of injection. Viral integration at the fli-1 locus drives Fli-1 cells. Expression of a dominant negative protein Engrailed overexpression, and has been identified as the pivotal genetic (En)/Fli-1 reduces proliferation of EWS/Fli-1-transformed NIH- event associated with this disease.2 Fli-1 viral integration is 3T3 cells, and both F-MuLV-induced and human erythroleukemia followed by further changes in the expression of anti-apoptotic cells. F-MuLV-induced erythroleukemia cells also display increased apoptosis, associated with reduced expression of and tumor suppressor genes, as well as the modification of bcl-2, a known fli-1 target gene. Introduction of En/Fli-1 into an erythropoietin (Epo), resulting in Epo-independent proliferation 4 F-MuLV-infected erythroblastic cell line induces differentiation, of erythroblasts and fully leukemic clones. Moreover, Fli-1 has as shown by increased a-globin expression. These results been reported to regulate an array of genetic targets, such as Rb, suggest, for the first time, an essential role for continuous MDM2 and bcl-2, whose roles in malignancy have been well Fli-1 overexpression in the maintenance and survival of the established.5–7 The study of F-MuLV-induced erythroleukemia malignant phenotype in murine and human erythroleukemias. Leukemia (2009) 23, 1311–1319; doi:10.1038/leu.2009.20; has also provided insight into the non-pathological role of Fli-1. published online 12 March 2009 Constitutive activation of fli-1 in erythroblasts alters their Keywords: erythroleukemia; Fli-1; RNAi; Engrailed; Friend responsiveness to Epo, switching the signaling events associated erythroleukemia with terminal differentiation to proliferation.5,8,9 While aberrant Fli-1 expression is associated with viral integration in mice, it is also associated with several human disorders. In Ewing’s sarcoma and primitive neuroectodermal tumors, a translocation between chromosome 11 and 22, Introduction t(11;22) (q24;q12) results in a chimeric EWS/Fli-1 fusion protein detectable in 85% of these tumors. This fusion protein contains Fli-1 is a member of the Ets gene family of transcription factors. the carboxy-terminal Ets domain of Fli-1 with the amino- Similar to other Ets proteins, Fli-1 functions as both a terminal RNA binding domain of EWS.10 The EWS/Fli-1 fusion transcriptional activator and repressor with a highly conserved protein functions as a transcriptional activator causing the DNA-binding sequence called the Ets domain, recognizing a initiation and progression of this disease, and possesses a strong consensus GGA(A/T) core motif. Additionally, deregulated transforming ability in fibroblasts mediated through the Ets expression of Ets proteins is associated with aberrant develop- 11 1 domain. Additionally, a group examining the expression of ment and malignant transformation. Fli-1 is preferentially Fli-1 in several benign and malignant neoplasms has further expressed in vascular endothelial cells and hematopoietic implicated the importance of fli-1 aberrant regulation in tumor tissues, including the thymus and spleen, and has been shown 12 2,3 progression. to play an important role in their development. Given the These observations emphasize a role for aberrant fli-1 expression profile of Fli-1, its aberrant regulation would regulation in the initiation of malignancies such as erythro- be expected to result in vascular and/or hematopoietic leukemia and Ewing’s sarcoma. However, the molecular abnormalities, and indeed, the discovery of fli-1 is attributed mechanisms underlying such transformations remain unclear. Modulating Fli-1 expression, such as by RNA interference Correspondence: Dr Y Ben-David, Department of Molecular and (RNAi) and dominant-negative studies, should provide a better Cellular Biology, Sunnybrook Health Sciences Centre Room S-216, understanding of fli-1 function and help to further elucidate its 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada. role in pathological occurrences. E-mail: [email protected] 4These authors contributed equally to this work. We hypothesized that modulating Fli-1 expression, would Received 17 September 2008; revised 18 November 2008; accepted impact the survival, proliferation and differentiation of erythro- 8 December 2008; published online 12 March 2009 leukemia cells. Here we provide evidence to suggest that Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1312 continuous Fli-1 expression is necessary for the maintenance virus, in the presence of polybrene (10 mg/ml final concentra- and survival of several malignant phenotypes. tion), as described earlier.13

Materials and methods Construction of a translational protein, En/Fli-1 The Drosophila Engrailed (En) repressor domain encoding 298 Cell lines amino acids of the En298 N-terminus was removed from the The murine erythroleukemia cell lines, CB3 and HB22.2, were En/Erg plasmid (a gift from Dr Marina Ina Arnone, Stazione maintained in a-minimum essential medium (Gibco, Grand Zoologica, Naples, Italy) by digestion with the BamHI and BglII Island, NY, USA) supplemented with 10% fetal bovine serum restriction enzymes (New England Biolabs). The En298 fragment (FBS) (Gibco). The human erythroleukemia cell line, HEL, and was then cloned upstream of the BamHI site in the MigR1-Fli-1 the human myelogenous leukemia cell line, K562, were vector. This construct was designated En/Fli-1, in which the ATG maintained in RPMI-1640 medium (Gibco) supplemented with start was inherent from the repressor domain of En. 10% FBS (Gibco). The murine erythroleukemia cell line, HB60-5, was maintained in the same medium supplemented with 0.2 U of Epo and 100 Zg of stem cell factor per ml. To induce differentiation, HB60-5 cells were washed twice with phos- Construction of the Fli-1 reporter gene phate-buffered saline (PBS) (Gibco) and incubated in the Double-stranded DNA fragments were selected from a random presence of 10% FBS and 1 U of Epo per ml. The mouse pool by repeated cycles of Fli-1-specific immunoprecipitation in fibroblast cell line, NIH-3T3, and the human embryonic kidney the presence of nuclear extract and PCR amplification (cyclic 14 cell line, 293T, were cultured in Dulbecco’s Modified Eagle’s amplification and selection of target, CAST) (unpublished Medium (Gibco) supplemented with 10% FBS (Gibco). data). One of the CAST sequences containing two underlined Fli-1 binding sites was selected for this study (forward primer: 50-CAGTGGGAGGACCGGAAGTCGACCGGAAGTTGCGTA-30; 0 RNA interference reverse primer: 5 -TACGCAACTTCCGGTCGACTTCCGGTCCT CCCACTG-30). A DNA fragment containing sequences imme- Pre-designed fli-1 small interfering RNA sequences were diately downstream this binding site was synthesized and purchased from Ambion (Ambion/Applied Biosystems, Foster cloned into the luciferase reporter vector pGL4.28. The resulting City, CA, USA) to design small hairpin RNA (shRNA) oligo- vector was designated as pGL4.28-FB. nucleotides containing a hairpin loop sequence (50-TTCAAGAGA-30) and an RNA Polymerase III terminator sequence consisting of a six-nucleotide poly(T) tract. Additionally, a 50-BamH1 restriction site was inserted in the top strand and a 50-EcoR1 restriction site Cell proliferation assay was inserted in the bottom strand of the shRNA oligonucleotides The growth rate of each erythroleukemic cell line was examined to enable directional cloning into the pSIREN RetroQ ZsGreen in low-serum media, containing 1% FBS. NIH-3T3 cells (5 Â 103 retroviral vector (Clontech, Takara Bio Inc., Mountain View, cells) were seeded in triplicates in 3 cm wells with low-serum CA, USA). A unique XbaI restriction site was also inserted media and cell proliferation was measured by performing immediately downstream of the terminator sequence to confirm Trypan-blue exclusion assay. the presence of shRNA oligonucleotide inserts. Complementary polyacrylamide gel electrophoresis-purified shRNA oligonucleotides, synthesized by Sigma Aldrich/Sigma Genosys Soft agar assays Canada (Oakville, ON, Canada), were annealed in a thermal 3 cycler (95 1C for 30 s, 72 1C for 2 min, 37 1C for 2 min, 25 1C for In each experiment, 5 Â 10 cells of each cell line were 2 min). Additionally, a negative control shRNA-annealed trypsinized and resuspended in 2 ml of complete medium in oligonucleotide, provided by Clontech, was cloned into the 0.3% agarose (Promega, Madison, WI, USA). The agar-cell same retroviral vector. The negative control shRNA contains the mixtures were plated in triplicate on plates containing a 1% same nucleotide composition as the fli-1 shRNA oligonucleo- complete medium agar mixture. After 2–3 weeks, the agar tide, but lacks sequence homology to the genome. The annealed assays were scored for viable colonies. double stranded oligonucleotides were ligated into the pSIREN vector (Clontech), using T4 DNA ligase (New England Biolabs, Pickering, ON, Canada). RNA extraction and reverse transcription PCR Total RNA was extracted using TRIZOL (Gibco), according to the manufacturer’s protocol, and treated with amplification RNAi virus production grade DNaseI (Invitrogen) to avoid amplification of contami- The RNAi-ready retroviral vectors were triple-transfected with nated genomic DNA. Complementary DNA was synthesized Lipofectamine 2000 (Invitrogen, Burlington, ON, Canada) into with SuperScript II Reverse Transcriptase (Invitrogen). PCR 293T cells, following the manufacturer’s protocol (Invitrogen). primers were as follows: Glyceraldehyde-3-phosphate dehydro- The vesicular stomatitis virus G glycoprotein (VSVG)-expressing genase (GAPDH), forward primer: 50-AACTTTGGCATTGTG vector as well as the vector expressing the gag and pol viral GAAGG-30, reverse primer: 50-TGTGAGGGAGATGCTCAGTG-30; packaging signals were kindly provided by Dr Dwayne Barber bcl-2, forward primer: 50-GATTCTGACAGACTCAGGAAGAAAC-30, (University of Toronto, Canada). Viral supernatant was collected reverse primer: 50-TCAGTCATCCACAGGGCGAT-30; a-globin, 48, 72 and 96 h post-transfection and concentrated by forward primer: 50-GATTCTGACAGACTCAGGAAGAAAC-30, ultracentrifugation at 25 000 rpm for 2 h. The virus was reverse primer: 50-CCTTTCCAGGGCTTCAGCTCCATAT-30.Gel resuspended in approximately 1/50 of the original volume. images were scanned using Quantity One software (Bio-Rad, CB3 cells (2.5 Â 106) were infected with 2 ml of concentrated Mississauga, Canada).

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1313 Flow cytometric cell cycle analysis 3,30-diaminobenzidine-glucose oxidase for 15 min. Counter- Single parameter analysis of DNA content was performed on staining was performed with hematoxylin. For immunohisto- cells fixed in 70% ethanol, washed twice with PBS (Gibco) and chemical controls, the primary antibody was omitted. resuspended in staining solution containing 500 ml RNaseA (200 U/ml) (Sigma Aldrich, Oakville, ON, Canada) and 500 ml propidium iodide (50 g/ml) (Molecular Probes, Eugene, OR, Luciferase reporter assay USA). Linear fluorescence signals of propidium iodide (area and 293T cells (106 cells) were plated in six-well plates and width) were assessed on a BD FACSCalibur (BD Biosciences, transfected 24 h later with the indicated amount of plasmid San Jose, CA, USA) with dye excitation by a 488 nm laser light. DNA using Lipofectamine 2000 (Invitrogen), as recommended Data were stored as mode files of at least 5 Â 103 single cell by the manufacturer. The DNA mixture used for transfection events for subsequent off-line analysis using Modfit and WinList includes the indicated amounts of reporter gene constructs software (Verity Software, Topsham, ME, USA). DNA cell cycle (pGL4.28 vector and pGL4.28-FB) and expression plasmids analysis was accomplished using the DIP_N2 and DIP_N3 (MigR1 empty vector, MigR1-Fli-1, MSCV empty vector and algorithms in Modfit. MSCV-EWS/Fli-1). Cell lysates were prepared 48 h after transfection and assayed for luciferase activity by using the luciferase assay system kit (Promega). The results shown are the mean of at Immunoblotting and antibodies least three independent transfection experiments. To account for Cells were lysed with radioimmunoprecipitation assay buffer transfection efficiency, b-gal concentrations were used to 13 (0.5% Nonidet P-40, 50 mM Tris HCl (pH 8.0), 120 mM NaCl, normalize results. 50 mM NaF, 10 g of aprotinin per ml, 100 g of leupeptin per ml, 10 mM phenylmethylsulfonyl fluoride). Lysates were resolved on sodium dodecyl sulfate–10% polyacrylamide gels, and immuno- Results blotted as described elsewhere.15 Antibodies to Fli-1 and b-actin were obtained from Santa Cruz Biotechnology (Santa Fli-1 downregulation in erythroleukemic cells by RNAi Cruz Biotechnology, Santa Cruz, CA, USA). Images were leads to a marked growth inhibition and cell death scanned using Quantity One software (Bio-Rad). Retroviral vectors were cloned to express functional fli-1 shRNAs or negative control shRNAs under the control of U6 RNA Polymerase III-dependent promoter. These shRNA retro- Immunohistochemistry viral expression plasmids were triple-transfected into 293T cells 293T cells were plated at a density of 1 Â 106 cells per poly-L- with plasmids expressing either the gag and pol packaging leucine-coated glass coverslip (Assistent, Germany) in six-well signals or the vesicular stomatitis virus glycoprotein envelope plates. After 24 h, cells were transfected with the MigR1 empty protein. Concentrated viral supernatants were used to infect the control vector, MigR1-Fli-1, or MigR1-En/Fli-1, using Lipofecta- F-MuLV-induced erythroleukemia cell line, CB3, expressing mine 2000 (Invitrogen) following the manufacturer’s protocol. high endogenous levels of Fli-1. Initial microscopic observation Immunocytochemical staining was performed 48 h post- and cell counting with Trypan blue showed that fli-1 shRNA- transfection. Coverslips were fixed with cold (4 1C) 4% treated cells exhibited significantly reduced proliferation rate paraformaldehyde at room temperature for 10–20 min, permea- (Figure 1a) and increased cell death compared with the bilized with absolute methanol at À20 1C for 7 min, washed appropriate controls (Figure 1b). thrice with PBS (Gibco) and blocked in 10% normal horse serum To validate the efficacy of RNAi-mediated Fli-1 downregula- (PAA Laboratories GmbH, Germany). After blocking, cells were tion, as well as to correlate growth inhibition and cell death with incubated with Fli-1 polyclonal antibody (Santa Cruz this downregulation, we analyzed an unsorted population of Biotechnology) overnight at 4 1C, washed thrice with PBS (Gibco), CB3 cells 24 h postinfection. Indeed, western blot analysis incubated with biotinylated goat anti-rabbit IgG antibody and confirmed the efficacy of Fli-1 knockdown (Figure 1c). These avidin–biotinylated peroxidase solution (Santa Cruz Biotechnology), data provide evidence to suggest that continuous Fli-1 over- and washed again. Antibody binding was revealed using expression is essential for the survival and proliferation of

Figure 1 Effects of RNAi-mediated Fli-1 downregulation in erythroleukemic cells. Growth rate kinetics of non-transfected (NT), polybrene control, negative control shRNA (NC) or ﬂi-1 shRNA (F) treated CB3 cells. Cells were stained with Trypan blue and cell numbers were counted each day as indicated. Downregulation of Fli-1 in F-MuLV-induced erythroleukemia cell line, CB3, inhibits (a) proliferation and (b) induces cell death. (c) Fli-1 expression was reduced in ﬂi-1 shRNA (F) treated CB3 cells.

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1314 erythroleukemic cells. However, constitutive suppression of Fli-1 protein, designated En/Fli-1. This vector as well as similar through RNAi is extremely toxic to erythroleukemia cells and retroviral vectors expressing Fli-1 and EWS/Fli-1 were transiently the rapid onset of cell death has made further examination transfected into 293T cells. To determine the efficiency of difficult. Therefore, additional analyses were necessary to transgene expression, western blot analysis was performed and investigate this phenomenon. revealed 48 and 51, 68 and 78 kDa bands corresponding to the two protein isoforms of Fli-1, EWS/Fli-1 and En/Fli-1 proteins, respectively (Figure 2a). Immunohistochemical analysis of 293T En/Fli-1 expression inhibits both Fli-1 and EWS/Fli-1 cells transfected with Fli-1 or En/Fli-1 and stained with a Fli-1- activity specific antibody, revealed identical localization of exogenous To support the phenomenon observed with RNAi-mediated Fli-1 Fli-1 protein (Supplementary Figure 1). downregulation, dominant negative strategies were employed to To investigate the efficacy of En/Fli-1 to inhibit the transacti- modulate Fli-1 expression. This dominant negative approach vating ability of Fli-1, luciferase assays were performed using a inhibits the transactivation activity of Fli-1 through introduction vector with the Fli-1 Ets DNA-binding site cloned upstream of of specific transcriptional suppressors that bind and suppress the the luciferase gene, designated pGL4.28-FB. Figure 2b shows Ets domain of Fli-1. Engrailed (En) is a Drosophila homeodomain that exogenous expression of both Fli-1 and EWS/Fli-1 in 293T protein required for proper segmentation and maintenance of cells increased the luciferase activity of the pGL4.28-FB reporter the posterior compartment identity. The repressor activity of En gene in a dose-dependent fashion, as compared with the empty lies in the sequence containing amino acids 2–298 which are vector, pGL4.28. As expected, EWS/Fli-1 expression displayed distant from the DNA binding homeodomain.16 Accordingly, more robust luciferase activity than Fli-1 (Figure 2b).20–22 translational fusions with the En repressor domain have Co-transfection of the En/Fli-1 expression vector with either efficiently converted plant transcription factors,17 b-catenin18 Fli-1 or EWS/Fli-1 expression vectors decreased luciferase and c-myb,19 into dominant-negative proteins. Thus, we have activity as compared with cells transfected with either Fli-1 or generated a retroviral vector expressing a dominant negative Fli-1 EWS/Fli-1 vectors alone (Figure 2b). Additionally, the luciferase

Figure 2 En/Fli-1 suppresses Fli-1 and EWS/Fli-1 activity and inhibits the transforming ability of EWS/Fli-1. (a) Expression of murine Fli-1 (48/51 kDa), EWS/Fli-1 (68 kDa) and En/Fli-1 (78 kDa) proteins following transient transfection with 1 and 2 mg DNA into 293T cells. 293T and CB3 cells transfected with control vector were used as negative and positive controls, respectively. (b) Fli-1 and EWS/Fli-1 increases the luciferase activity of the reporter gene, pGL4.28-FB, in a dose-dependent fashion in 293T cells. En/Fli-1 inhibits the luciferase activity of both Fli-1 and EWS/ Fli-1. (c) Expression of Fli-1, EWS/Fli-1, and En/Fli-1 in NIH-3T3 cells infected with the corresponding retrovirus. (d) Soft agar growth analysis of EWS/Fli-1-transformed NIH-3T3 cells. 5 Â 103 cells of each transfected group were grown for 12 days in medium containing 0.3% agarose. Colonies containing 420 cells were scored. EWS/Fli-1-expressing cells formed large macroscopic colonies as shown in the upper panel. The lower panel displays the average number of colonies observed in experiments performed in triplicate. (e) En/Fli-1 reduces the growth rate of EWS/Fli-1- transformed NIH-3T3 cells in low-serum media. Cells (2 Â 104) were seeded in triplicate, grown in media supplemented with 1% FBS and counted at 3-day intervals.

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1315 reporter carrying point mutations within the Fli-1 binding site cells expressing En/Fli-1 display a decreased proportion of cells exhibited the same level of luciferase activity as the empty in S phase (Figure 3d) and revealed an apoptotic population, as luciferase vector (data not shown). evidenced by the appearance of a sub-G1 peak (Figure 3d). It is likely that Fli-1 overexpression contributes to the anti-apoptotic phenotype through direct regulation of its target genes. En/Fli-1 represses the transforming ability of EWS/Fli-1 The anti-apoptotic gene, bcl-2, overexpressed in F-MuLV- in transformed NIH-3T3 cells induced erythroleukemias,24 has previously been shown to be a As EWS/Fli-1 has the same Ets binding site as Fli-1, En/Fli-1 was direct Fli-1 target gene, and Fli-1-mediated Bcl-2 upregulation examined for its ability to repress or inhibit EWS/Fli-1-mediated contributes to the enhanced survival of erythroblasts.6 Conse- NIH-3T3 cell transformation. To do so, we generated a stable quently, the direct transcriptional regulation of Fli-1 target NIH-3T3 cell line overexpressing mouse EWS/Fli-1, En/Fli-1 or genes, such as bcl-2, is thought to contribute to the transforming Fli-1 (Figure 2c). It is well known that one of the hallmarks of a ability of Fli-1 in erythroleukemic cells. Therefore, as En inhibits transformed phenotype is the ability of cells to grow in semisolid the ability of Fli-1 to act as a transcriptional repressor or medium. Soft agar assays showed that empty vector transfected activator, we examined bcl-2 expression in En/Fli-1-expressing NIH-3T3 cells grew poorly in this condition as only 15±10 CB3 cells. Indeed, reverse transcription-PCR analysis revealed small colonies per 5000 cells plated were counted (Figure 2d). significantly reduced levels of bcl-2 in these cells as compared In contrast, EWS/Fli-1-expressing NIH-3T3 cells grew efficiently with the empty vector control infected cells (Figure 3e). in soft agar, forming large macroscopic colonies of greater than The role of Fli-1 inactivation was further examined in other 200 cells, indicating anchorage-independent growth (Figure 2d). F-MuLV-induced erythroleukemia cell lines overexpressing One of the EWS/Fli-1-transformed clones (designated EWS/Fli-1) Fli-1, termed HB60-5 and HB22.2 (Figure 3a). As expected, was infected with the En/Fli-1, Fli-1 or empty vector retrovirus to En/Fli-1 expression in both cells lines (Figure 3f) also results in a test their ability to form colonies in soft agar. EWS/Fli-1 cells marked growth inhibition (Figure 3g). Interestingly, this phe- infected with the Fli-1 retrovirus or empty vector control nomenon is rescued over time through the loss of En/Fli-1 displayed no inhibition of growth in soft agar. However, expression (Figure 4). HB60-5 cells, infected with the control or infection of these cells with the En/Fli-1 retrovirus significantly En/Fli-1 retroviral expression vector, were sorted by flow reduced the number and size of colonies (Figure 2d). These data cytometry at 3 and 12 days postinfection. En/Fli-1 protein indicate that the En/Fli-1 repressor protein impairs anchorage- expression and growth rates were examined in each population. dependent growth of transformed NIH-3T3 cells, thereby The expression of the dominant negative fusion protein was lost altering the transforming ability of EWS/Fli-1, suggesting a 12 days postinfection and was associated with a proliferation central role for the Fli-1 Ets binding site in malignant profile identical to that of the control cells (Figure 4). Together, transformation. these results confirm a similar phenomenon observed with RNAi-mediated Fli-1 downregulation, suggesting that continuous Fli-1 overexpression is required for the proliferation and En/Fli-1 reduces the proliferation rate of EWS/Fli-1- survival of murine erythroleukemic cells. transformed NIH-3T3 cells in low-serum media Under normal culture conditions of 10% FBS supplemented En/Fli-1 inhibits proliferation of human erythroleukemic medium, no significant difference could be detected in the cells in culture growth rate of both wild type or EWS/Fli-1-transformed NIH-3T3 Similar to F-MuLV-induced erythroleukemia cell lines, the cells co-transfected with En/Fli-1 retrovirus (data not shown). human erythroleukemia cell line, HEL, also displays aberrant However, under low-serum culture conditions (1% FBS), the fli-1 expression (Figure 3a). The importance of continuous Fli-1 EWS/Fli-1-transformed cells grew rapidly, whereas proliferation overexpression was examined in HEL cells by introduction of in the control cells remained unchanged (Figure 2e). EWS/Fli-1- the En/Fli-1 retrovirus (Figure 5a). Expression of En/Fli-1 in HEL transformed cells co-transfected with En/Fli-1 displayed a cells (Figure 5a), similar to murine erythroleukemic cells significantly reduced proliferation rate, similar to the rate of (Figure 3), resulted in a reduced proliferation rate (Figure 5b). control-transfected cells (Figure 2e). Thus, En/Fli-1 expression Moreover, similar to the murine erythroleukemic cells, the was able to inhibit the proliferation of EWS/Fli-1-transformed expression of En/Fli-1 was lost after 12 days of culture (data not cells in low-serum growth conditions. shown). These data for the first time display the importance of Fli-1 expression in the maintenance of human erythroleukemia.

En/Fli-1 suppresses proliferation of erythroleukemic cells overexpressing Fli-1 En/Fli-1 promotes differentiation of erythroid The aforementioned results suggested the ability of En/Fli-1 to progenitors transformed by Fli-1 inhibit the transactivating activity of Fli-1 and EWS/Fli-1 as well The F-MuLV-induced erythroleukemia cell line, HB60-5, as to suppress the transforming ability of the latter. Therefore, to proliferates in the presence of Epo and stem cell factor, but further establish the critical role of continuous Fli-1 over- undergoes terminal differentiation in the presence of Epo alone.5 expression in malignant transformation, En/Fli-1 was introduced We have revealed earlier that ectopic expression of Fli-1 in into an F-MuLV-induced erythroleukemia cell line termed CB3, these cells blocks erythroid differentiation by switching Epo- which overexpresses Fli-123 (Figure 3a). 3 days postinfection induced differentiation to Epo-induced proliferation.5 To further with either En/Fli-1 or control vector retroviruses, green verify the role of Fli-1 in erythroid differentiation, we introduced fluorescent protein-positive cells were sorted by flow cytometry En/Fli-1 in HB60-5 cells overexpressing Fli-1, designated HB60- prior to immunoblotting and growth rate analysis (Figures 3b 5-Fli-1 (Figure 6a). Additionally, our earlier studies have shown and c). Similar to our earlier RNAi experiments, the proliferation that the expression of a-globin, a differentiation marker for rate of CB3 cells was significantly attenuated after infection with erythroid cells, is elevated in response to Epo in HB60-5 cells. the En/Fli-1 retroviral expression vector (Figure 3c). Accordingly, However, Fli-1 overexpression in these cells inhibits this cell cycle analysis using propidium iodide revealed that CB3 response.5 When both HB60-5-Fli-1 and En/Fli-1-expressing

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1316

Figure 3 En/Fli-1 expression suppresses proliferation of murine erythroleukemia. (a) Fli-1 expression in the indicated cell lines was determined using western blot analysis. Green ﬂuorescent protein (GFP)-positive CB3 cells infected with the empty vector control or En/Fli-1-expressing virus were sorted by ﬂow cytometry and subjected to (b) western blot analysis and (c) Trypan blue exclusion assay. (d) Cell cycle analysis of GFP-positive CB3 cells infected with En/Fli-1. (e) En/Fli-1 reduces Bcl-2 expression in GFP-positive CB3 cells as determined by reverse transcription-PCR. F-MuLV-induced erythroleukemia cell lines, HB60-5 and HB22.2, infected with vector alone or En/Fli-1-expressing vector were subjected to (f) western blot analysis and (g) Trypan blue exclusion assay.

HB60-5-Fli-1 cells were grown in the presence of Epo, an susceptible to the development of erythroleukemia following increased level of a-globin was detected in En/Fli-1-expressing infection with F-MuLV, however, the latency period is cells (Figure 6b), indicating that En/Fli-1 expression reversed the significantly increased.29 Although these observations implicate inhibitory effect on erythroid differentiation caused by Fli-1 that Fli-1 overexpression is vital for the initiation and progres- overexpression. sion of erythroleukemia, the molecular mechanisms underlying the role of aberrant fli-1 regulation in oncogenesis remains poorly understood. Discussion To assess the role of Fli-1 overexpression in the maintenance and survival of murine erythroleukemia, we downregulated F-MuLV-induced erythroleukemia is a well-established tumor Fli-1 expression using RNAi. RNAi-mediated suppression of Fli-1 model that is aptly suited for studying the mechanisms in F-MuLV-induced erythroleukemia cells results in reduced governing the proliferation, maturation and transformation of proliferation associated with rapid cell death. Owing to the erythroid progenitor cells.25 The activation of fli-1 through toxicity of constitutive Fli-1 suppression and to further investi- retroviral insertion in erythroleukemia induced by F-MuLV is a gate this phenomenon, we generated a dominant negative form critical initial step in the multistage development of this of Fli-1, termed En/Fli-1. In this system, ectopic expression of leukemia.2 F-MuLV studies involving examination of the role this dominant-negative fusion protein mimics a loss-of-function of fli-1 in erythropoiesis have suggested that constitutive Fli-1 allele. This approach has also been successful in inhibiting the expression dramatically increases the self-renewal potential function of the Ets-related gene Erg (F Rizzo and MI Arnone of pro-erythroblasts and blocks erythroid differentiation.5 unpublished data/personal communication), plant transcription Conversely, transgenic mice generated to overexpress Fli-1 do factors,17 as well as b-catenin,18 and c-myb.19 A luciferase assay not develop erythroleukemia, but instead exhibit increased confirmed the ability of En/Fli-1 to act as a transcriptional lymphopoiesis, splenomegaly, accumulation of abnormal B and repressor, capable of inhibiting the transactivating activity of T cells and develop a severe lupus-like autoimmune disease, Fli-1, as well as that of the oncogenic fusion protein EWS/Fli-1. This ultimately resulting in death.26 Mice carrying a targeted deletion oncoprotein, generated by a chromosomal translocation in Ewing’s in the Ets domain at the fli-1 locus display abnormal sarcoma and primitive neuroectodermal tumors, is a potent hematopoiesis and vasculogenesis, and die between embryonic transcriptional activator containing the strong transactivating day 11.5 and 12.5.27,28 The first targeted disruption at the fli-1 domain of EWS and the Ets DNA-binding domain of Fli-1, with a locus, deleting the first ATG start site, resulted in the expression more powerful transforming ability than fli-1.20–22,30 We show of a truncated protein and generated a non-lethal minor that expression of the En/Fli-1 repressor impairs anchorage- phenotype, including a reduction in thymus size and in the dependent growth of EWS/Fli-1-transformed NIH-3T3 cells and number of total thymocytes.29 Interestingly, these mice remain reduces the proliferation rate of these cells in low-serum media.

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1317

Figure 4 The gradual loss of En/Fli-1 expression is associated with increased proliferation of erythroleukemic cells. Erythroleukemic HB60-5 empty vector control and En/Fli-1 cell populations were sorted by ﬂow cytometry at (a) 3 and (b) 12 days postinfection, subjected to western blot analysis, stained with Trypan blue and cell numbers were counted each day, as indicated. (a) En/Fli-1 expression results in a marked growth inhibition. (b) The loss of En/Fli-1 expression is associated with an increased growth rate, similar to that of the control cell population.

permit enhanced or novel interactions with genes not normally responsive to Fli-1 to disrupt normal growth and differentiation.30,31 These results support an earlier study showing that the expression of small interfering RNAs targeting the fusion between EWS and fli-1 slows the proliferation rate of a Ewing’s sarcoma cell line, and the expression of fli-1 small interfering RNAs in these cells further potentiates this effect.32 Moreover, a recent study has shown that mice expressing EWS/Fli-1 develop a rapid onset of myeloid/erythroid leukemia that closely resembles F-MuLV-induced erythroleukemia.33 Furthermore, cell cycle analysis revealed the ability of En/Fli-1 to suppress growth and induce apoptosis in an F-MuLV-induced erythroleukemia cell line. Interestingly, Fli-1 suppression also attenu- ates the growth rate of the human erythroleukemia cell line, HEL. This study, for the first time, provides evidence to suggest that Fli-1 plays an essential role in the proliferation of human erythroleukemia. Fli-1 downregulation, mediated through RNAi or expression of the dominant-negative En/Fli-1 fusion protein, confirmed the essential role of fli-1 in the regulation of cellular proliferation and survival.34,35 We have shown that En/Fli-1 expression is only tolerated for a short period after its introduction into erythroleukemia cells. Additionally, the loss of Fli-1 activity in these cells is associated with a marked growth inhibition and Figure 5 En/Fli-1 inhibits growth of human erythroleukemic cells, apoptosis. Previously, Fli-1 has been shown to activate HEL, in culture. (a) The GFP-positive HEL cells infected with the empty transcription of the anti-apoptotic gene, bcl-2.6 To this respect, vector control or En/Fli-1-expressing virus were sorted by flow the expression of bcl-2 in CB3 cells infected with En/Fli-1 was cytometry and subjected to cell growth analysis. (b) Western blot analysis displaying expression of En/Fli-1 as detected by the Fli-1 much lower than cells infected with the control vector alone. antibody. These data suggest that the fli-1 proto-oncogene contributes to cell survival, at least in part, by the upregulation of bcl-2 in erythroleukemic cells,24 resulting in evasion of the normal apoptotic program. Although En/Fli-1 does not completely inhibit the transforming In addition to inhibition of apoptosis, Fli-1 overexpression in ability of EWS/Fli-1, these experiments imply that the oncogenic erythroid cells is also shown to inhibit erythroid differentiation, potential of EWS/Fli-1 at least, in part, is attributed to the leading to massive proliferation, in response to Epo.5 Indeed, transactivating activity of the Fli-1 Ets DNA-binding domain. ectopic expression of Fli-1 in the HB60-5 erythroleukemia cell However, the strong transactivation domain from EWS may line switches Epo-induced differentiation to Epo-induced

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1318

Figure 6 En/Fli-1 reduces inhibited differentiation in an erythroleukemia cell line infected with Fli-1. Erythroleukemic HB60-5 Fli-1 cells were infected with the empty vector control or En/Fli-1-expressing virus and induced to undergo differentiation with addition of 1U/ml of Epo. (a) Expression of Fli-1 and En/Fli-1 in HB60-5, HB60-5-Fli-1 and HB60-5-Fli-1 infected cells. (b) Reverse transcription-PCR analysis of the expression of a-globin in the control and En/Fli-1-expressing HB60-5-Fli-1 cells at the indicated time points.

proliferation. To verify the role of Fli-1 in differentiation, we 7 Truong AH, Cervi D, Lee J, Ben David Y. Direct transcriptional have shown that the expression of En/Fli-1 in HB60-5 cells regulation of MDM2 by Fli-1. Oncogene 2005; 24: 962–969. overexpressing Fli-1 increases their differentiation status, as 8 Pereira R, Quang CT, Lesault I, Dolznig H, Beug H, Ghysdael J. measured through the expression of a-globin. FLI-1 inhibits differentiation and induces proliferation of primary erythroblasts. Oncogene 1999; 18: 1597–1608. In summary, the experiments described here suggest, for the 9 Zochodne B, Truong AH, Stetler K, Higgins RR, Howard J, Dumont first time, an essential role for continuous Fli-1 overexpression in D et al. Epo regulates erythroid proliferation and differentiation the proliferation and survival of murine and human erythro- through distinct signaling pathways: implication for erythropoiesis leukemias. It appears that Fli-1 has the potential to alter these and Friend virus-induced erythroleukemia. Oncogene 2000; 19: cellular processes, likely through the regulation of several target 2296–2304. genes. The approach used in this study could eventually provide 10 Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M et al. Gene fusion with an ETS DNA-binding domain caused by potential avenues for the elucidation of fli-1 function in chromosome translocation in human tumours. Nature 1992; 359: malignant transformation through identification of additional 162–165. genes regulated by this transcription factor. Furthermore, these 11 May WA, Gishizky ML, Lessnick SL, Lunsford LB, Lewis BC, findings reveal that Fli-1 downregulation may have important Delattre O et al. Ewing sarcoma 11;22 translocation produces a implications in the pathogenesis of diseases, such as erythro- chimeric transcription factor that requires the DNA-binding leukemia, and perhaps provide clues to novel treatment options. domain encoded by FLI1 for transformation. Proc Natl Acad Sci USA 1993; 90: 5752–5756. 12 Mhawech-Fauceglia P, Herrmann F, Penetrante R, Beck A, Sait S, Acknowledgements Block AM et al. Diagnostic utility of FLI-1 monoclonal antibody and dual-color, break-apart probe fluorescence in situ (FISH) analysis in Ewing’s sarcoma/primitive neuroectodermal tumor We would like to thank Dr C Danny for providing us with the EWS- (EWS/PNET). A comparative study with CD99 and FLI-1 polyclonal Fli-1 construct, Dr D Barber for the VSVG and gag/pol expression antibodies. Histopathology 2006; 49: 569–575. vectors, and Dr MI Arnone for the En/Erg expression vector. This 13 Cui JW, Li YJ, Sarkar A, Brown J, Tan YH, Premyslova M et al. work was supported by a grant from the Terry Fox Foundation Retroviral insertional activation of the Fli-3 locus in erythro- through the National Cancer Institute of Canada (NCIC) and the leukemias encoding a cluster of microRNAs that convert Canadian Institute of Health Research (CIHR) to YBD. Epo-induced differentiation to proliferation. Blood 2007; 110: 2631–2640. 14 Funk WD, Wright WE. Cyclic amplification and selection of targets for multicomponent complexes: myogenin interacts with factors References recognizing binding sites for basic helix-loop-helix, nuclear factor 1, myocyte-specific enhancer-binding factor 2, and COMP1 factor. 1 Oikawa T, Yamada T. Molecular biology of the Ets family of Proc Natl Acad Sci USA 1992; 89: 9484–9488. transcription factors. Gene 2003; 303: 11–34. 15 Howard JC, Berger L, Bani MR, Hawley R, Ben-David Y. 2 Ben-David Y, Giddens EB, Bernstein A. Identification and mapping Activation of the erythropoietin gene in the majority of of a common proviral integration site Fli-1 in erythroleukemia cells F-MuLV-induced erythroleukemias results in growth factor induced by Friend murine leukemia virus. Proc Natl Acad Sci USA independence and enhanced tumorigenicity. Oncogene 1996; 1990; 87: 1332–1336. 12: 1405–1415. 3 Klemsz MJ, Maki RAPT, Moore J, Hromas R. Characterization of 16 Wang N, Kudryavtseva E, Ch’en IL, McCormick J, Sugihara TM, the ets oncogene family member, fli-1. J Biol Chem 1994; 268: Ruiz R et al. Expression of an engrailed-LMO4 fusion protein in 5769–5773. mammary epithelial cells inhibits mammary gland development in 4 Lee CR, Cervi D, Truong AH, Li YJ, Sarkar A, Ben David Y. Friend mice. Oncogene 2004; 23: 1507–1513. virus-induced erythroleukemias: a unique and well-defined mouse 17 Markel H, Chandler J, Werr W. Translational fusions with the model for the development of leukemia. Anticancer Res 2003; 23: engrailed repressor domain efficiently convert plant transcription 2159–2166. factors into dominant-negative functions. Nucleic Acids Res 2002; 5 Tamir A, Howard J, Higgins RR, Li YJ, Berger L, Zacksenhaus E 30: 4709–4719. et al. Fli-1, an ets-related transcription factor, regulates 18 Montross WT, Ji H, McCrea PD. A beta-catenin/engrailed erythropoietin-induced erythroid proliferation and differentiation: chimera selectively suppresses Wnt signaling. J Cell Sci 2000; evidence for direct transcriptional repression of the Rb gene during 113: 1759–1770. differentiation. Mol Cell Biol 1999; 19: 4452–4464. 19 Lyon JJ, Watson RJ. Interference of Myb transactivation activity 6 Lesault I, Quang CT, Frampton J, Ghysdael J. Direct regulation of by a conditional dominant negative protein: functional inter- BCL-2 by FLI-1 is involved in the survival of FLI-1-transformed ference in a cytotoxic T-cell line results in G1 arrest. Gene 1996; erythroblasts. EMBO J 2002; 21: 694–703. 182: 123–128.

Leukemia Effects of Fli-1 inhibition on erythroleukemia J-W Cui et al 1319 20 Watson DK, Smyth FE, Thompson DM, Cheng JQ, Testa JR, Papas 28 Spyropoulos DD, Pharr PN, Lavenburg KR, Jackers P, Papas TS, TS et al. The ERGB/fli-1 gene: isolation and characterization of a Ogawa M et al. Hemorrhage, impaired hematopoiesis, and new member of the family of human ETS transcription factors. Cell lethality in mouse embryos carrying a targeted disruption of the Growth Differ 1992; 3: 705–713. Fli1 transcription factor. Mol Cell Biol 2000; 20: 5643–5652. 21 Rao VN, Ohno T, Prasad DD, Bhattacharya G, Reddy ES. Analysis 29 Melet F, Motro B, Rossi DJ, Zhang L, Bernstein A. Generation of a of the DNA-binding and transcriptional activation functions of novel Fli-1 protein by gene targeting leads to a defect in thymus human Fli-1 protein. Oncogene 1993; 8: 2167–2173. development and a delay in friend virus-induced erythroleukemia. 22 Zhang L, Lemarchandel V, Romeo P-H, Ben-David Y, Bernstein A. Mol Cell Biol 1996; 16: 2708–2715. The Fli-1 proto-oncogene involved in erythroleukemia and Ewing’s 30 May WA, Lessnick SL, Braun BS, Klemsz M, Lewis BC, Lunsford LB sarcoma, encodes a transcriptional activator with DNA-binding et al. The Ewing’s sarcoma EWS/FLI-1 fusion gene encodes a more specificities distinct from other Ets family members. Oncogene potent transcriptional activator and is a more powerful transform- 1993; 8: 1621–1630. ing gene than FLI-1. Mol Cell Biol 1993; 13: 7393–7398. 23 Ben-David Y, Giddens EG, Letwin K, Bernstein A. Erythroleukemia 31 May WA, Arvand A, Thompson AD, Braun BS, Wright M, Denny induction by Friend murine leukemia virus: insertional activation CT. EWS/FLI1-induced manic fringe renders NIH 3T3 cells of a new member of the ets gene family, Fli-1, closely linked to tumorigenic. Nat Genet 1997; 17: 495–497. c-ets-1. Genes Dev 1991; 5: 908–918. 32 Dohjima T, Lee NS, Li H, Ohno T, Rossi JJ. Small interfering RNAs 24 Howard JC, Li Q, Chu W, Zochodne B, Kapoor M, Ung Y et al. expressed from a Pol III promoter suppress the EWS/Fli-1 transcript Bcl-2 expression in F-MuLV-induced erythroleukemias: a role for in an Ewing sarcoma cell line. Mol Ther 2003; 7: 811–816. the anti-apoptotic action of Bcl-2 during tumor progression. 33 Torchia EC, Boyd K, Rehg JE, Qu C, Baker SJ. EWS/FLI-1 induces Oncogene 2001; 20: 2291–2300. rapid onset of myeloid/erythroid leukemia in mice. Mol Cell Biol 25 Ben-David Y, Bernstein A. Friend virus-induced erythroleukemia 2007; 27: 7918–7934. and the multistage nature of cancer. Cell 1991; 66: 831–834. 34 Yi H, Fujimura Y, Ouchida M, Prasad DD, Rao VN, Reddy ES. 26 Zhang L, Eddy A, Teng T, Fritzler M, Kluppel M, Melet F et al. An Inhibition of apoptosis by normal and aberrant Fli-1 and erg immunological renal disease in transgenic mice that overexpress proteins involved in human solid tumors and leukemias. Fli-1, a member of the ets family of transcription. Mol Cell Biol Oncogene 1997; 14: 1259–1268. 1995; 15: 6961–6970. 35 Goltzene F, Skalski M, Wolff CM, Meyer D, Mager-Heckel AM, 27 Hart A, Melet F, Grossfeld P, Chien K, Jones C, Tunnacliffe A et al. Darribere T et al. Heterotopic expression of the Xl-Fli transcription Fli-1 is required for murine vascular and megakaryocytic develop- factor during Xenopus embryogenesis: modification of cell ment and is hemizygously deleted in patients with thrombocyto- adhesion and engagement in the apoptotic pathway. Exp Cell penia. Immunity 2000; 13: 167–177. Res 2000; 260: 233–247.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia