Ewing's sarcoma is a member of Ewing's family tumors (EFTs) and the second most common solid bone and soft tissue malignancy of children and young adults. It is associated in 85% of cases with the t(11;22)(q24:q12) chromosomal translocation that generates fusion of the 5′ segment of the EWS gene with the 3′ segment of the ETS family gene FLI-1. The EWS-FLI-1 fusion protein behaves as an aberrant transcriptional activator and is believed to contribute to EFT development. However, EWS-FLI-1 induces growth arrest and apoptosis in normal fibroblasts, and primary cells that are permissive for its putative oncogenic properties have not been discovered, hampering basic understanding of EFT biology. Here, we show that EWS-FLI-1 alone can transform primary bone marrow–derived mesenchymal progenitor cells and generate tumors that display hallmarks of Ewing's sarcoma, including a small round cell phenotype, expression of EFT-associated markers, insulin like growth factor-I dependence, and induction or repression of numerous EWS-FLI-1 target genes. These observations provide the first identification of candidate primary cells from which EFTs originate and suggest that EWS-FLI-1 expression may constitute the initiating event in EFT pathogenesis. (Cancer Res 2005; 65(24): 11459-68)
Sarcomas constitute some of the most aggressive adult and childhood malignancies that are believed to originate from as yet poorly defined mesenchymal stem/progenitor cells (1, 2). Currently, sarcomas are subdivided into two broad subclasses based on the genetic mutations that are associated with their development. One subclass is composed of tumors bearing multiple complex chromosomal translocations, duplications, and deletions with the corresponding loss of tumor suppressor genes and amplification/activation of oncogenes. The second subclass, which includes Ewing's family tumors (EFTs), is composed of tumors associated with specific chromosomal translocations that give rise to fusion genes that are believed to play an active role in transformation (1, 2). Chromosomal translocations associated with EFTs lead to the fusion of the 5′ segment of the EWS gene on chromosome 22 with the 3′ segment of one of the ets family genes on chromosomes 2, 7, 11, 17, and 21, including ERG, t(21;22)(q22;q12); ETV1, t(7;22)(p22;q12); FEV, t(2;22)(q33;q12); and ETV4 (3, 4). However, by far the most common fusion occurs with the FLI-1 gene resulting from the translocation t(11;22)(q24:q12) (5).
The EWS-FLI-1 fusion protein behaves as an aberrant transcriptional regulator (6) whose suppression in EFT cell lines results in decreased cell growth in vitro and tumor-forming capacity in vivo (7). EWS-FLI-1 expression therefore seems required for EFT development, but the mechanisms whereby it induces transformation and/or controls tumor growth and progression remain unknown. Genes whose expression is reported to be induced by EWS-FLI-1 include MYC (8), EAT-2 (9), MMP-3 (10), FRINGE (11), ID2 (12), and CCND1 (13), whereas TGFBR2 (14), CDKN1A (p21/CIP1/WAF1; ref. 15), and p57KIP (16) are among those reported to be repressed. However, induction and repression of potential EWS-FLI-1 target genes are dependent on the cellular background (17), as is the physiologic outcome of EWS-FLI-1 expression.
EWS-FLI-1 elicits enhanced tumorigenicity in immortalized NIH3T3 cells (18). By contrast, it induces p53-dependent growth arrest in primary human fibroblasts (19) and apoptosis in mouse embryonic fibroblasts (MEF; ref. 20). Deletion of p19ARF, p16INK4A, and p53 attenuates apoptosis and allows stable EWS-FLI-1 expression in MEFs, possibly facilitating subsequent transformation (20). To exert its putative transforming properties, EWS-FLI-1 may therefore require a cellular environment with mutations in genes implicated in growth, senescence, and cell cycle control. However, loss of p19ARF, p16INK4A, and p53 occurs in only a minority of EFTs (21), and the chromosomal translocation leading to EWS-FLI-1 expression may be the only genetic event in a significant fraction of Ewing's sarcomas (21). These observations suggest the existence of nonmutated primary cells that are permissive for EWS-FLI-1-mediated transformation. In view of the lack of adequate EFT models, identification of such cells constitutes a key step toward elucidating EFT biology (22).
Although EFTs arise in bone and soft tissues, their expression of neuronal markers has kept open the debate as to their potential mesenchymal or neuroectodermal origin. In an effort to identify candidate primary cells from which EFTs originate, we addressed the effects of EWS-FLI-1 expression on the tumorigenic potential and gene expression profile of cells spanning a broad plasticity spectrum, including primary bone marrow–derived mesenchymal progenitor cells (MPC), embryonic stem cells, and spontaneously immortalized embryonic (STO) fibroblasts. Our results indicate that unlike embryonic stem and STO cells, MPCs display permissiveness for EWS-FLI-1-mediated transformation and generate tumors that display hallmarks of Ewing's sarcoma. Based on these observations, we propose that mesenchymal progenitor cells may constitute a candidate environment from which EFTs originate.
Materials and Methods
Cells and fluorescence-activated cell sorting analysis. The MEF cell line STO and the embryonic stem cell line ES-D3 were obtained from the American Type Culture Collection (ATCC, Rockville, MD). Embryonic stem cells were cultured on gelatin-coated plates supplemented with 10 ng/mL leukemia inhibitory factor (LIF; Sigma, St. Louis, MO). MPCs were isolated from bone marrow of wild-type adult C57BL/6 mice according to described procedures (23) and cultured at low density on fibronectin-coated plates (Sigma) in medium containing 2% dialyzed FCS (Sigma), 10 ng/mL epidermal growth factor (EGF; Sigma), 10 ng/mL PDGF-BB (R&D Systems, Minneapolis, MN), and LIF produced by the CHO LIF720D cell line. MPCs were tested by fluorescence-activated cell sorting for mesenchymal stem cell marker expression before and after infection and selection. Phycoerythrin-conjugated monoclonal antibodies (mAb; BD PharMingen, San Diego, CA) were against CD44, CD45, CD13, CD11b, CD117 (c-kit), CD90 (Thy1), Sca1, and Flk1. SK-N-MC, SK-ES1, and U2OS cell lines (ATCC) were cultured, respectively, in RPMI 1640, McCoy's 5A, and DMEM, all containing 10% fetal bovine serum (FBS).
Differentiation assays. Cells were plated at 1 × 105 in 35-mm plates, in medium without EGF, PDGF-BB, or LIF but with cytokines specific for the different lineages. For adipogenesis, MPCs were allowed to reach 80% confluency and maintained in differentiation medium containing 5 μg/mL insulin, 1 μmol/L dexamethasone, and 0.5 μmol/L 3-isobutyl-1methylxanthine for 14 days. Accumulation of neutral lipids in the cells was detected by Oil Red-O staining.
For osteogenic differentiation, MPCs were maintained at 50% confluence in differentiation medium containing 1 μmol/L dexamethasone, 50 μg/mL ascorbic acid, and 10 mmol/L β-glycerophosphate for 14 days. Osteoblast differentiation was detected by revelation of alkaline phosphatase activity using the BM purple substrate (Roche Diagnostics AG, Rotkreuz, Switzerland) as early as 7 days following stimulation. Neural differentiation was induced by maintaining 50% confluent MPCs in differentiation medium containing 100 ng/mL basic fibroblast growth factor for 14 days and confirmed by immunofluorescence detection of nestin expression (mouse anti-nestin rat 401 antibody, diluted 1:1,000, Developmental Studies Hybridoma Bank and Cy3-conjugated secondary antibody). Nestin expression was observed as early as 48 hours following stimulation.
Cloning and reverse transcription-PCR. A cDNA clone encoding the human EWS-FLI-1 type 2 fusion gene was amplified from the SK-N-MC Ewing sarcoma cell line by reverse transcription-PCR (RT-PCR) using Super Script one-step RT-PCR with the platinum Taq kit (Invitrogen, Carlsbad, CA) under the following conditions: one cycle at 50°C for 30 minutes and 94°C for 2 minutes followed by 35 cycles of 94°C for 45 seconds, 55°C for 45 seconds, and 72°C for 2 minutes and a final extension at 72°C for 10 minutes. The amplified fragment was digested with XhoI and HpaI and inserted into the pMSCV Puro and pMSCV Neo retroviral expression vectors (BD Biosciences Clontech, Palo Alto, CA). cDNA clones encoding the mouse Igf1, Bcl2, c-myc, p53, Rex-1, and Oct-4 genes were amplified by RT-PCR as described above using specific primers. The fusion gene primer sequences were hEWS forward XhoI, CCGCTCGAGCCACCATGGCGTCCACGGATTACAG; hFLI-1 reverse (including a stop codon), CTAGTAGTAGCTGCCTAAGTG. The amplified fragment was inserted into the cloning vector pNOF (GL BioTech, Bremen, Germany). The V5 epitope tag was added at the 3′ end of the EWS-FLI-1 sequence by PCR using the primers: hEWS forward XhoI, CCGCTCGAGCCACCATGGCGTCCACGGATTACAG; hFLI-1V5 reverse (without a stop codon), AGGGTTAGGGATAGGCTTACCTTCGAACCGCGGGCCGTAGTAGCTGCCTAAGTGGA; V5 reverse HpaI (including a stop codon), GTTAACTCACGTAGAATCGAGACCGAGGAGAGGGTTAGGGATAGGCTTACC. The cDNA clones encoding the mouse Igf1, Bcl2, and c-myc genes were amplified by RT-PCR using the primers: mIgf1 forward BglII, AGATCTATGACCGCACCTGCAATAAAG; mIgf1 reverse EcoRI, GAATTCCTAGCCCAGTCTTTTTTCTCT; mBcl2 forward BglII, GGAAGATCTTCCCCACCATGGCGCAAGCCGGGAGAACA; mBcl2 reverse XhoI, CCGCTCGAGCGGTCACTTGTGGCCCAGGTATGC; mc-Myc forward EcoRI, CCGGAATTCCCACCATGCCCCTCAACGTGAACTTC; mc-Myc reverse XhoI, CCGCTCGAGCGGTTATGCACCAGAGTTTCGAAG. The amplified fragment was inserted into the pMSCV Puro and MSCV Neo retroviral expression vectors (BD Biosciences Clontech). Plasmids were sequenced to verify the presence and integrity of cDNAs encoding the human EWS-FLI-1V5 fusion gene and the mouse Igf1, Bcl2, and c-myc genes.
RT-PCRs of p53 and p19 were done using primers: mp53 forward, ATGACTGCCATGGAGGAGTCACAGTC; mp53 reverse, TCAGTCTGAG TCAGGCCCCACTTT; mp19 forward, ATGGGTCGCAGGTTCTTGGTCAC; mp19 reverse, GATTGGCCGCGAAGTTCCAG. RT-PCR of Oct-4 and Rex-1 from equal quantities of mRNA derived from MPCs, primary fibroblasts, and embryonic stem cells was done using the following primers: mOct-4 forward, CCGTGAAGTTGGAGAAGGTG; mOct-4 reverse: TGATTGGCGATGTGATGTAT; mRex-1 forward, CACCATCCGGGATGAAAGTGAGAT; mRex-1 reverse, ACCAGAAAATGTCGCTTTAGTTTC. RT-PCR of β-actin provided an internal loading control.
Retrovirus generation and infection. Expression of hEWS-FLI-1V5 in embryonic stem, MPCs, and STO was achieved using the Retroviral Gene Transfer and Expression (BD Biosciences Clontech), according to the manufacturer's recommendations. Expression of the fusion gene and corresponding protein were tested at each time point in all three cell lines by, respectively, RT-PCR and Western blot analysis using the mouse anti-V5 antibody. Stable expression of mIgf1, mBcl2, and c-Myc in MPCs was achieved using the same retroviral delivery system described above. The infected cells were selected with 1.5 μg/mL puromycin or 500 μg/mL G418 for a maximum of 14 days, and bulk cultures of resistant cells were used in subsequent experiments.
Western blot analysis. Cell lysis, SDS-PAGE, blotting, and immunostaining were done by standard procedures, and protein bands were detected using a chemiluminescent substrate kit (Amersham Biosciences, Amersham, United Kingdom) according to the manufacturer's recommendations. Primary mouse anti-V5 epitope mAb (Invitrogen), mouse anti-ErbB2 mAb (Chemicon, Temecula, CA), mouse anti-mouse p21 mAb (BD Biosciences Clontech), hamster anti-mouse Bcl2 mAb (BD PharMingen), polyclonal rabbit anti-mouse Igf1 receptor a (Cell Signaling, Beverly, MA), and Histone H3 (Abcam, Cambridge, United Kingdom) antibodies were used. Secondary antibodies were horseradish peroxidase (HRP)–conjugated goat anti-mouse (Bio-Rad, Hercules, CA) and mouse anti-rabbit (Sigma) antibodies.
cDNA array hybridization. Total RNA was extracted from each cell line using RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's recommendations. The quality of total RNA was verified by an Agilent RNA 600 nanoassay and by measuring the 260/280 absorbance ratio. Quality-tested total RNA was amplified using the RiboAmp RNA Amplification kit (Arcturus, Mountain View, CA) and processed using a reverse transcription–based method of label incorporation to yield labeled cDNA. For each sample, 5 μg of amplified RNA were used in the cDNA probe synthesis with Cy5-dCTP or Cy3-dCTP (Amersham Biosciences, Amersham, United Kingdom) and random primers. Probes were purified using a Mini Elute PCR purification kit (Qiagen) and concentrated using Centricon YM-30 filters (Amicon, Millipore, Billerica, MA). Expression analysis was done using the NIA-17k clone set (24)5
cDNA array analysis. Following hybridization and washing, microarrays were imaged using a ScanArray 4000 scanner (Perkin-Elmer, Foster City, CA), and scanned slide images were converted to a tagged image file format. Fluorescence ratios for array elements were extracted by using ScanAlyze software,6788 by the within-print-tip group lowess normalization procedure (25). Quality control of slide hybridization was done using variables described on the corresponding web site.9
Statistical analysis of the expression data. For each time point and cell line, five m17k microarrays (among which two were dye swaps) were done comparing hEWS-FLI-1-V5 expressing with empty vector control cells. Expression data for each time point and cell line were analyzed with standard one-sample, one-sided t tests applied to the logarithm of the ratio of the expression levels of the hEWS-FLI-1-V5 and the control sample. The null hypothesis is that the mean of such a logarithm is 0, and the alternate hypothesis is that the mean is >0 or <0 for the one-sided test used to identify induced or repressed clones. The one-sample t test was shown in a recent study to be more effective than the two-sample test due to the correlation between expression levels measured on the same array (26).
The use of a standard statistical test (as opposed to methods based on cutoff on fold change) allowed us to estimate the false discovery rate (FDR) of the lists of induced and repressed genes. This was done using the Benjamini-Hochberg method (27).
Functional analysis of the lists of induced and repressed genes. The annotations of the clones included in the array to Gene Ontology (28) terms were downloaded from the mouse 17k deconvolution file of the Swiss Institute for Experimental Cancer Research.10
Real-time quantitative reverse transcription-PCR. cDNA was obtained using an Moloney murine leukemia virus reverse transcriptase and RNase H minus (Promega, Madison, WI). Typically, 500 ng of template total RNA and 250 ng of random hexamers were used per reaction. Real-time PCR amplification was done using a Taqman Universal PCR mastermix and Assays-On-Demand gene expression products in an ABI Prism 7700 instrument (Applied Biosystems, Foster City, CA). Relative quantitation of target, normalized with an endogenous control (cyclophyllin A), was done using a comparative Ct or a standard curve method (Applied Biosystems). Probes used included mouse Igf1, Igf1 binding protein 3, Igf1 binding protein 5, Alcam, Jam2, Lum, Birc2, Emb, Yap, and TGF-βRII.
In vitro NVP-AEW541 sensitivity assays. For the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, MPC, SK-N-MC, and U2OS cells were incubated in 24-well plates with normal medium, 0.5 mmol/L NVP-AEW541, or solvent (DMSO) for 10 days. Cell sensitivity to the drug was tested with CellTiter AQueous Nonradioactive Cell Proliferation Assay (Promega) according to the manufacturer's recommendations.
Tumorigenicity assays. For each MPC cell population, twelve 5-week-old BALB/cJHanHsd-SCID mice Harlan (Indianapolis, IN) were injected s.c. with 1 × 106 cells stably expressing hEWS-FLI-1-V5, SYT/SSX-1, ERBB2, mIGF-I, mBcl2, or empty vector. Tumor growth was monitored weekly, and mice were euthanized 9, 12, and 15 weeks after injection. Six mice were injected s.c. with 1 ×106 MPCEWS-FLI-1-T1.1 and MPCEWS-FLI-1-T1.2 and sacrificed 4 weeks after injection. The same protocol was applied to MPCEWS-FLI-1-T2.1 and MPCEWS-FLI-1-T2.2 injection, and the animals were sacrificed 3 weeks after injection. All tumors were resected at autopsy and sectioned for histologic analysis. Experimental protocols involving mice were approved by the Etat de Vaud, Service Vétérinaire, authorization no. VD1477.0.
Tumor explants. Tumors were excised, dissociated in PBS supplemented with 10% FBS, and resuspended for 40 minutes at 37°C in an enzymatic cocktail containing 0.1 mg/mL collagenase VIII + 1 mg/mL trypsin in PBS. The supernatants were filtered and centrifuged for 6 minutes at 400 × g, and the resulting cellular pellets were plated in MPC medium without LIF, containing 1.5 μg/mL puromycin or 500 μg/mL G418.
Immunohistochemistry. Paraffin-embedded sections of MPCsEWS-FLI-1- and MPCsBCL2-derived tumors were stained with mouse anti-human CD99 (1:40 dilution; Signet Laboratories, Dedham, MA) and neural-specific enolase (NSE, 1:100 dilution; DAKO, Glostrup, Denmark) mAb, or rabbit anti-mouse Ki67 (1:200 dilution; Novocastra, Newcastle upon Tyne, United Kingdom) and terminal deoxynucleotidyl transferase (TdT, 1:10 dilution; DAKO) polyclonal antibodies. HRP staining was done using biotin-conjugated horse anti-mouse or goat anti-rabbit immunoglobulin (Vector Laboratories, Burlingame CA) and revealed with a DAKO 3,3′-diaminobenzidine kit.
Expression of EWS-FLI-1 in embryonic stem cells, mesenchymal progenitor cells, and STO fibroblasts. The type 2 EWS-FLI-1 fusion gene, which is associated with Ewing's sarcoma bearing the poorest prognosis (29), was RT-PCR amplified from total RNA derived from the SK-N-MC Ewing's sarcoma cell line, and the corresponding cDNA was inserted into the pMSCV retroviral vector containing sequences encoding the v5 tag at the 3′ end. After sequence verification, the construct was transfected into Eco293 packaging cells, and the resulting retrovirus was used to infect embryonic stem cells, MPCs, and STO cells. In all cases, bulk cultures of infected cells were used for subsequent experiments.
MPCs were derived from the bone marrow of wild-type C57BL/6 mice as described in Materials and Methods and characterized by a CD45-negative, CD11b-negative, Flk-negative, CD44-positive, Thy1-positive, CD117 weakly positive, CD13 weakly positive, and Sca1 strongly positive phenotype (Fig. 1A). Although mesenchymal progenitor/stem cells can differ in cell surface receptor expression according to the microenvironment from which they are derived (30) and the mouse strain of origin (31), the phenotype displayed by the MPCs used in the present work is consistent with that described in other studies (30, 32). In addition, these cells display strong Rex-1 and weak Oct-4 expression (Fig. 1B), reflecting their progenitor status (29) and differentiate into adipocytes, osteocytes, and neurons in response to appropriate stimuli (Fig. 1C). Their differentiation potential, therefore, encompasses the mesenchymal and neuronal attributes of EFTs.
Expression of the EWS-FLI-1 fusion protein in each cell type was assessed by RT-PCR and Western blot analysis using the anti-V5 antibody, 24, 48, and 72 hours following infection. In all three cell types, expression of EWS-FLI-1 could be detected by Western blot analysis at each of the time points (Fig. 2A-C). By day 14, however, the fusion protein was no longer detectable in embryonic stem and STO cells, despite the presence of the corresponding transcripts, as assessed by RT-PCR (Fig. 2D). In contrast, expression of the EWS-FLI-1 fusion protein in MPCs persisted several weeks following infection (Fig. 2D; data not shown) and displayed the expected nuclear localization (Fig. 2E).
EWS-FLI-1 expression induces tumorigenicity in mesenchymal progenitor cells. Embryonic stem and STO cells as well as MPCs infected with EWS-FLI-1 (MPCEWS-FLI-1) or empty vector were injected s.c. into severe combined immunodeficient (SCID) mice 5 days following infection, when the fusion protein was detectable in all three cell types. Expression of EWS-FLI-1 in embryonic stem and STO cells had no effect on the intrinsic ability of these cells to form teratomas and slowly growing spindle cell tumors, respectively (Supplementary Data 3A). By contrast, although MPCs infected with empty virus failed to produce tumors, all 12 mice injected with MPCEWS-FLI-1 developed tumors within 6 weeks (Fig. 3A). MPCEWS-FLI-1 tumors (MPCEWS-FLI-1-T1) were composed predominantly of small round cells with dense homogeneous nuclei, characteristic of EFT, occasionally organized into nests surrounded by fibrotic strands and spindle cells (Fig. 3B). Reinjection of the tumor cells into SCID mice resulted in development within 2 to 4 weeks of tumors (MPCEWS-FLI-1-T2) composed of sheets of small round cells, virtually identical to tumors formed by injection of the human EFT cell lines SK-ES1, SK-N-MC, and A673 (Fig. 3B; data not shown). Tumors resulting from an additional round of injection (MPCEWS-FLI-1-T3) emerged more rapidly than their predecessors but displayed the same histologic phenotype (data not shown). Tumor cells were positive for the EFT-associated markers CD99 and NSE (ref. 33; Fig. 3C) and retained EWS-FLI-1 protein expression as determined by Western blot analysis of corresponding lysates (Fig. 3D).
To exclude the possibility that the observed MPCEWS-FLI-1 tumor phenotype was a reflection of intrinsic MPC histogenetic properties irrespective of the transforming agent, MPCs stably expressing the oncogenes ErbB2, SYT-SSX-1, and Bcl-2 (Fig. 2F) were tested for tumorigenicity in SCID mice. Overexpression of ErbB2 is implicated in the formation of a variety of tumor types (34), whereas expression of SYT-SSX-1, the fusion protein associated with synovial sarcoma, can enhance tumor formation by immortalized fibroblasts (35). By suppressing apoptosis, Bcl-2 overexpression is believed to facilitate oncogenic mutations in hematopoietic cells, and chromosomal translocations involving Bcl-2 are associated with the pathogenesis of hematopoietic malignancies (36).
Neither MPCErbB2 nor MPCSYT-SSX-1 cells formed tumors in mice (Fig. 3A), suggesting that the oncogenic effect of EWS-FLI-1 does not reflect indiscriminate MPC susceptibility to transformation. Of the 12 mice injected with MPCBcl-2, only four developed tumors that became detectable >4 months following injection. The late emergence of these tumors in only a fraction of the mice is consistent with the occurrence of oncogenic mutations favored by Bcl-2-promoted survival. MPCBcl-2 tumors were composed of CD99- and NSE-negative spindle cells with large clear nuclei, bearing no resemblance to their MPCEWS-FLI-1 counterparts (Fig. 3B-C). These observations support the notion that MPCEWS-FLI-1 tumor histogenesis is a function of EWS-FLI-1 expression.
To ensure that induction of tumorigenic properties in MPCs was not due to altered EWS-FLI-1 activity associated with the V5 tag or to potential genetic events secondary to in vitro selection, MPCEWS-FLI-1 tumorigenicity experiments were repeated using fresh, independently isolated batches of MPCs, V5-tagged and untagged EWS-FLI-1 constructs, different selection markers, and variable selection periods. Neither the V5 tag, the duration of selection in vitro nor the type of selection marker used affected the observed transformation. Altogether, six sets of tumorigenicity experiments were done in 36 test and 36 control mice using three independent MPC batches. In each set of experiments, all of the mice injected with MPCEWS-FLI-1 developed tumors within 4 to 6 weeks (data not shown), whereas none of the mice bearing vector only–expressing MPCs developed tumors even several months after injection.
Mesenchymal progenitor cells and MPCEWS-FLI-1 tumors maintain p53 function and p19ARF expression. Because loss of p53 function seems to be a prerequisite for the maintenance of EWS-FLI-1 expression in MEFs (20) and because induction of a p53 response precludes primary human fibroblast transformation by EWS-FLI-1 (19), we assessed the p53 status in MPCs and MPCEWS-FLI-1 tumors. Sequence analysis of PCR-amplified p53 from MPC RNA revealed no mutations. Treatment of MPCEWS-FL1-1-T1 and MPCEWS-FLI-1-T2 cells with 5-fluorouracil, a potent activator of p53 (37), revealed induction of the p21 protein and several other p53 targets, as assessed by microarray analysis (Fig. 4A and B), confirming the preservation of a functional p53 pathway.
MPCs displayed weak constitutive expression of p19ARF, as detected by RT-PCR (Fig. 4C). Although the physiologic signals that regulate p19ARF are only partly understood, myc overexpression has been observed to induce p19ARF in MEFs, where its action is believed to provide a p53-dependent safeguard against transformation (38). Similar to its effect in MEFs, overexpression of myc in MPCs induced p19ARF (Fig. 4C), indicating that p19ARF is responsive to oncogenic signals in these cells. Consistent with this observation, p19ARF expression was observed to be induced in MPCEWS-FLI-1-T1 and MPCEWS-FLI-1-T2 cells (Fig. 4D), and both p16INK4A and p19ARF displayed wild-type sequence in MPCs, MPCEWS-FLI-1-T1 and MPCEWS-FLI-1-T2 cells (data not shown). Thus, EWS-FLI-1 can transform MPCs despite the presence of wild-type p16INK4A/p19ARF and intact p53 function.
Immunohistochemical staining of the tumor sections with anti-Ki67 and anti-TdT antibodies confirmed the high proliferative and low apoptotic index of MPCEWS-FLI-1 tumors (Supplementary Data S3B).
Mesenchymal progenitor cells, embryonic stem, and STO cells display different responses to EWS-FLI-1 expression. To address the response of MPCs to EWS-FLI-1 at the molecular level, we compared the gene expression profile of MPCs infected with EWS-FLI-1 containing retrovirus to that of MPCs infected with empty virus 24, 48, and 72 hours after infection. RNA was obtained from two independently EWS-FLI-1-infected and two corresponding empty vector–infected cell batches and expression of 17,000 cDNA clones was compared using the NIA-17k mouse cDNA array (24). Five microarrays were used to assess expression profile changes at each time point. Expression data for each clone present on the microarray were analyzed as described in Materials and Methods.
Twenty-four hours following infection with EWS-FLI-1 retrovirus, 1,046 and 779 genes were, respectively, induced and repressed in MPCs with a FDR as low as 1% (Supplementary Data S1). The induced transcripts included genes annotated to Gene Ontogeny terms of chaperone activity, structural molecule activity, ribosomes, protein biosynthesis, ATP binding, biosynthesis, macromolecule biosynthesis, ribonucleocomplex, and adenyl nucleotide binding (Supplementary Data S1). Interestingly, 72 hours following EWS-FLI-1 expression marked induction of insulin-like growth factor-1 (IGF-I) and IGF binding proteins 3 and 5 (IGFBP-3 and IGFBP-5; Fig. 5A-B) was observed. Neither ErbB2, SYT/SSX, nor Bcl-2 induced IGF-I in MPCs (data not shown), supporting the notion that the observed IGF-I induction was EWS-FLI-1 specific. In contrast to its effect in MPCs, EWS-FLI-1 expression in embryonic stem and STO cells resulted, respectively, in decreased and <2-fold increased IGF-I expression (Fig. 5B), whereas IGFBP-3 and IGFBP-5 expression remained unaltered (data not shown). IGF-I receptor (IGF-IR) expression was observed in all three cell types (Fig. 5C) and was unaffected by EWS-FLI-1 (data not shown).
EWS-FLI-1–induced mesenchymal progenitor cell tumors are insulin-like growth factor-I dependent. Association of IGF-I/IGF-IR signaling with cancer is well established (39). EFTs are among tumors that not only express IGF-I (13) but also require IGF-I/IGF-IR signals for growth and survival (40–42) and are highly sensitive to IGF-IR inhibition (40, 42). Accordingly, >90% of MPCEWS-FLI-1-T1/T2 cells dissociated from surgically removed tumors died in response to IGF-IR inhibition (Fig. 5C-D) by the small molecular weight kinase inhibitor NVP-AEW541, a pyrrolo[2,3-d]pyrimidine derivative that is highly selective for IGF-IR (43). Human EFT SK-N-MC cells displayed similar sensitivity to IGF-IR blockade, whereas MPCEWS-FLI-1 cells harvested before injection into mice and osteosarcoma U2OS cells displayed lower sensitivity and resistance, respectively (Fig. 5C-D). Although IGF-I/IGF-IR signaling may be essential for EFT cell survival, additional pro-oncogenic events are believed to be necessary to generate and promote EFT growth (40–42). Consistent with this view, IGF-I overexpression alone was not sufficient to initiate MPC tumorigenicity, as MPCs displaying a 400-fold increase in IGF-I expression following infection with IGF-I retrovirus (data not shown) failed to form tumors in mice (Fig. 3A).
MPCEWS-FLI-1 tumor–derived cells display altered expression of EWS-FLI-1 target genes that may be implicated in tumor development and progression. To gain insight into candidate mechanisms that might cooperate with IGF-I/IGF-IR signals to promote MPCEWS-FLI-1 tumor formation, we compared the gene expression profile of cells derived from two independent first-round tumors (MPCEWS-FLI-1-T1.1/T1.2) to that of MPCEWS-FLI-1 cells harvested immediately before injection into mice (Supplementary Data S2). MPCEWS-FLI-1-T1.1/T1.2 displayed altered expression of several known EWS-FLI-1 target genes, including induction of MYC (8), MMP-3 (11), ID2 (12), and CCND1 (13) and repression of CDKN1A (p21; ref. 15; Fig. 5E).
Additional potentially relevant genes whose expression was induced in tumor-derived cells were yes-associated protein (YAP), baculoviral IAP containing 2 (BIRC2), EMBIGIN, IGFBP-3; several chemokines, including CXCL12; proteolytic enzymes, including CATHEPSIN C; and the neuronal inducer NECDIN (Supplementary Data S2). Repressed transcripts included LUMICAN and several genes encoding adhesion receptors, most notably α5 and α7 integrins, Jam-2 and Jam-3, and ALCAM (Supplementary Data S2). Real-time PCR comparison of expression of ALCAM, Jam-2, LUMICAN, BIRC2, EMBIGIN, and YAP in MPCEWS-FLI-1 and MPCEWS-FLI-1-T1 confirmed the changes suggested by microarray analysis (Fig. 6).
Comparison of MPCEWS-FLI-1-T1 and MPCEWS-FLI-1-T2 transcripts by cDNA microarray analysis revealed similar gene expression profiles, with nevertheless reduced expression of the EWS-FLI-1 target gene TRGBRII in the latter (Fig. 6).
Our observations constitute the first demonstration of primary cell transformation by EWS-FLI-1, leading to the formation of tumors that display hallmarks of EFT, including a small round cell phenotype, expression of CD99 and NSE, IGF-I dependence, and changes in expression of several known EWS-FLI-1 target genes.
EWS-FLI-1 expression has been observed to block differentiation of p19ARF−/− MPCs along osteogenic and adipogenic lineages (44), but its transforming effect in neither these cells nor their wild-type counterparts had been tested thus far. The observation that EWS-FLI-1 transformed several independently isolated batches of primary MPC derived from wild-type mice and displaying a normal diploid karyotype (data not shown) argues in favor of its expression being the initiating event in MPCEWS-FLI-1 tumor formation. Moreover, our results indicate that neither p19ARF nor p53 constitute a barrier to the transforming potential of EWS-FLI-1 in MPCs, and that similar to the majority of EFTs, MPCEWS-FLI-1 tumors retain wild-type p16INK4A/p19ARF and a functional p53 pathway.
The ability of a single oncogenic event to induce tumor formation may seem to diverge from the widely held view that malignant transformation of cells, particularly those which give rise to solid tumors, requires at least three genetic events. However, the complexity of mutational events necessary for tumor formation may depend, at least in part, on the nature of the cells from which the tumor originates. Thus, the self-renewal capacity intrinsic to progenitor cells, which is advantageous to malignant growth, may reduce the number of mutations necessary to achieve transformation and render the cells more sensitive to the transforming potential of selected oncogenic agents. Consistent with this notion, single chromosomal translocations that generate aberrant transcription factors may suffice to initiate some types of leukemia from hematopoietic progenitor cells (45–47), and expression of the myxoid liposarcoma–associated FUS/TLS-CHOP fusion protein in adipocyte precursor cells is able to induce the corresponding tumors in transgenic mice (48). EFTs may resemble these malignancies in that their initiation may require a single genetic event in the appropriate progenitor cell environment.
The timing of EWS-FLI-1 expression may also be important in EFT pathogenesis. Most EFTs arise in bone at puberty, which coincides with robust activation of IGF-I/IGF-IR signaling triggered by a growth hormone spurt, suggesting that IGF-I may participate in the early stages of EFT development. The EWS-FLI-1-mediated induction of IGF-I in MPCs observed in the present work may recapitulate events that precede EFT formation and help establish conditions that facilitate transformation by EWS-FLI-1 target gene products. Sustained IGF-I overproduction (42) or the recently observed down-regulation of IGFBP-3 (49), which sequesters IGF-I and limits its bioavailability, may provide at least two possible mechanisms for maintaining IGF-I levels required for subsequent EFT survival. Selection of one mechanism over the other may depend on the cellular context (17) and possibly the location and duration of tumor growth.
Several EWS-FLI-1 target genes that are potentially implicated in transformation and tumorigenesis and that may cooperate with IGF-I, including MYC, ID2, CCND1, MMP-3, CDKN1A, and TGFBRII, displayed altered expression in MPCEWS-FLI-1 tumors. Most of these genes have been identified as EWS-FLI-1 targets individually in different cells, consistent with cell type–specific EWS-FLI-1 target gene expression patterns (17). The potential physiologic consequences relative to EFT development of their combined expression changes in a single cell population have therefore not been fully appreciated. Id proteins are general inhibitors of differentiation and stimulators of proliferation and are implicated in the development of neuroectodermal tumors (50). Id2 overexpression in neuroblastoma cells has been shown to mediate cellular transformation and maintenance of the malignant phenotype (50, 51). Ewing's sarcoma is also associated with elevated expression of Id2, which is a target of myc oncoproteins, IGF-I and EWS-FLI-1, raising the possibility that EWS-FLI-1, c-myc, and IGF-I may cooperate in maintaining abnormal Id2 expression in EFT (50). Id proteins are recruited by myc to bypass the tumor suppressor function of the retinoblastoma protein (Rb; refs. 50, 52). Concomitant induction of cyclin D1 and down-regulation of p21CIP1/WAF1, a key inhibitor of cyclin-dependent kinases, may facilitate proliferation by accelerating S-phase entry and decreasing G1 arrest efficacy, respectively. It would therefore seem reasonable to suggest that activation of proliferation checkpoints by the combined action of IGF-I, myc, Id2, cyclin D1, and repression of p21WAF1/CIP1 may provide a basis for EWS-FLI-1-mediated transformation. Products of other known EWS-FLI-1 target genes found to display altered expression in MPCEWS-FLI-1 tumors may subsequently contribute to tumorigenesis. Thus, in agreement with recent observations on EFT cells (14), decreased expression of TGFBRII correlated with increased rapidity of MPCEWS-FLI-1 tumor growth and transition from a histologic phenotype characterized by tumor cell nests surrounded by fibrotic strands to one dominated by sheets of tumor cells.
At least one of the newly identified EWS-FLI-1 target gene products (i.e., IGF-I) may provide clues as to how EWS-FLI-1 might transform cells that maintain functional p19ARF and p53. Induction of myc by EWS-FLI-1 may be expected to up-regulate p19ARF, leading to p53 activation and apoptosis. In MPCs, however, induction of IGF-I precedes that of myc. Given that a major role of IGF-I is to promote cell survival by activating the phosphatidylinositol 3-kinase/Akt pathway, early induction of IGF-I may provide a mechanism to prevent p19ARF- and p53-mediated apoptosis. This notion is supported by observations that in the presence of constitutively active Akt, myc can induce lymphoma formation and immortalize primary cells despite the presence of wild-type p19ARF and p53 (53). IAP and YAP may further promote survival of transformed cells, explaining, at least in part, tumor growth in the face of a functional p19/p53 axis. Induction of MMP-3, MMP-13, and CATHEPSIN C may promote surrounding stromal tissue remodeling, whereas down-regulation of integrins, ALCAM and lumican may augment tumor cell motility, further contributing to the malignant phenotype.
The present model should provide the means to test the physiologic relevance of individual EWS-FLI-1 target genes in MPCEWS-FLI-1 tumor pathogenesis and determine the chronology of molecular events leading from EWS-FLI-1-mediated primary cell transformation to tumor growth and progression. Although extrapolation of mechanisms underlying mouse tumor development to human cancer must be made with caution, the similarity of MPCEWS-FLI-1 tumors to EFTs supports the premise that key aspects of their pathogenesis are likely to be shared.
Taken together, our observations have identified primary bone marrow–derived mesenchymal progenitor cells as a permissive environment for EWS-FLI-1-mediated transformation. Their location in the bone, where the majority of EFTs arise, and their ability to form tumors with molecular, morphologic, and immunohistochemical features of EFTs render MPCs strong candidates for the origin of these tumors. Moreover, our results provide experimental support to the long-suspected notion that EWS-FLI-1 expression in permissive cells can constitute the initiating event in the pathogenesis of Ewing's sarcoma.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Grant support: Swiss National Research Foundation grant 3100-65090.01 and National Centre of Competence in Research in Molecular Oncology. P. Provero is a Lagrange Fellow of the C.R.T. Foundation, Turin.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Richard Iggo (Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland) for the anti-p21 antibody, Louis Guillou for helpful discussions, Carlo Fusco for technical assistance, and Delphine Galaud for histology and immunohistochemistry.