The Insulin-like Growth Factor-I Receptor Is Required for EWS/FLI-1 Transformation of Fibroblasts*

Ewing’s family of tumors is characterized by a well described reciprocal translocation, t(11;22)(q24;q12), which produces a fusion protein (EWS/FLI-1) that transforms mouse fibroblasts. The EWS/FLI-1 fusion protein has been shown to act as a potent chimeric transcription factor. Overexpression of insulin-like growth factor-I receptor (IGF-IR) has been implicated in many tumor models as playing a role in cell growth and tumorigenesis. In addition, blockade of the IGF-IR inhibits the growth of Ewing’s family of tumors cells. Therefore, we first studied whether the presence of the IGF-IR is required for transformation by the EWS/FLI-1 fusion protein. To perform this study, we used two previously described fibroblast cell lines, R− and W, derived from an IGF-IR knockout mouse and a wild-type littermate, respectively. Neither W nor R− cells without the fusion protein formed soft agar colonies. However, W clones expressing the fusion message (WF cells) formed soft agar colonies, whereas R− clones expressing the fusion message (R−F cells) did not form soft agar colonies. Because the IGF-IR is required for EWS/FLI-1 transformation, we chose to investigate whether altered signaling occurs from the IGF-IR when the EWS/FLI-1 fusion is present. WF cells demonstrated a greater degree of ligand-stimulated insulin receptor substrate-1 phosphorylation when compared with W cells, suggesting that expression of the EWS/FLI-1 fusion protein alters the IGF-IR signaling pathway.

Ewing's sarcoma family of tumors (ESFT) 1 are characterized by a translocation that occurs in 95% of tumors (1). This translocation joins the EWS gene located on chromosome 22 to an ets family gene, either FLI-1 located on chromosome 11, t(11;22) (2) or ERG located on chromosome 21, t(21;22) (3,4). The fusion protein generated as a result of either of these translocations contains two primary domains. The EWS domain is a potent transcriptional activator, whereas the FLI-1 or ERG domain contains a highly conserved DNA binding domain (5)(6)(7)(8). The EWS/ets translocation and the resulting fusion protein are thought to play a role in both the etiology and growth promotion of ESFT.
The EWS/FLI-1 fusion protein appears to play a role in maintaining cell growth. Growth inhibition has been demonstrated by three methods of inhibiting the EWS/FLI-1 fusion protein: 1) RNA antisense (9), 2) a dominant negative truncated protein with an intact DNA binding domain (10), and 3) antisense oligodeoxynucleotides (11,12). Each of these models demonstrates that when the fusion protein or its transcript is decreased, tumor cell growth is reduced. The chimeric EWS/ FLI-1 fusion protein generated as a result of the translocation transforms mouse fibroblasts, and this transformation requires both the EWS and FLI-1 functional domains to be intact (13). The mechanism whereby the EWS/FLI-1 fusion protein transforms cells remains unknown. The insulin-like growth factor-I receptor (IGF-IR) is an ␣2-␤2 heterodimeric receptor where the ␣-subunit contains the extracellular ligand binding domain and the ␤-subunit contains the transmembrane and intracellular tyrosine kinase domains. The IGF-IR binds IGF-I and IGF-II and to a lesser extent insulin. The receptor transduces its signal via autophosphorylation of tyrosine residues with subsequent tyrosine phosphorylation of downstream targets, including insulin receptor substrate-1 (IRS-1), IRS-2, Shc, and Crk (14). The receptor is critical for many physiologic functions including development, cell growth, transformation, and prevention of apoptosis (15,16). Embryos from mice with a targeted disruption of the IGF-IR were 45% of normal size and nonviable at birth (17). Blockade of ligand-mediated signal transduction through the IGF-IR using a monoclonal antibody inhibits tumor growth in many model systems (18,19). Transformation occurs when the IGF-IR is overexpressed in mouse fibroblasts but only in the presence of ligand (20). Until recently, all oncogenes investigated were found to require the presence of the IGF-IR. However, a constitutively active G-protein mutant G ␣ 13 transforms RϪ cells, which do not express IGF-IR. 2 The oncogenes that require the IGF-IR to transform cells and their IGF-IR-dependent pathways are currently being elucidated.
EWS/FLI-1 fusion protein functions as an aberrant transcription factor, but as yet there is no mechanism for how this protein transforms cells. The IGF-IR is expressed in ESFT cell lines; inhibition of IGF-IR signaling, by blocking ligand stimulation, has reduced the growth of these cells (21,22). Because investigators have shown that other transforming proteins, including SV-40 large T antigen (23), requires the presence of IGF-IR to transform cells, we sought to determine whether the presence of these receptors was required for the transforming activity of the EWS/FLI-1 fusion protein. To perform this study, we used two previously described cell lines, RϪ and W, derived from an IGF-IR knockout mouse and a wild-type littermate, respectively (24). Both cell lines were transfected, individual clones were selected, and clones were evaluated for anchorage-independent growth. We show here that the IGF-IR is required for the EWS/FLI-1 fusion protein to transform mouse fibroblasts. We also demonstrate that the presence of the EWS/FLI-1 fusion protein leads to altered signal transduction through the IGF-IR.

EXPERIMENTAL PROCEDURES
Cell Lines and Materials-W and RϪ cell lines were a generous gift of R. Baserga (Philadelphia, PA), whereas Rϩ cells are transfected RϪ cells that overexpress wild-type IGF-IR. All cells were grown in Dulbecco's modified Eagles medium ϩ 10% fetal bovine serum. ESFT cell lines TC-32, CHP-100, and RD-ES have previously been shown to contain the t(11;22) translocation (25,26) and were grown in RPMI 1640 medium ϩ 10% fetal bovine serum. All cells were maintained in humidified, 6% CO 2 atmosphere at 37°C. EWS/FLI-1 type I cDNA was a gift of C. Denny (Los Angeles, CA). Lyophilized ␣IRϪ3 (IGF-IR antibody) was purchased from Oncogene Science Inc. (Cambridge, MA), and the IgG 1 isotype-matched antibody MOPC 21 was purchased from Sigma. IGF-IR ␤-subunit antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Phosphotyrosine antibodies 4G10 and PY-20 were purchased from Upstate Biotechnology Inc. (Lake Placid, NY) and Transduction Laboratories (Lexington, KY), respectively. IRS-1 antibodies, both pleckstrin and carboxyl domains, were purchased from Upstate Biotechnology Inc. Grb2 antibody was purchased from Transduction Laboratories. IGF-I was obtained from Ci77ba-Geigy.
Characterization of W and RϪ Cells-Immunoblot analysis for the ␤-subunit of the IGF-IR was performed on W, RϪ, and Rϩ cells stimulated with 20 nM IGF-I. Lysates were prepared in Tris/sodium chloride buffer with Nonidet P-40, vanadate, and protease inhibitors (12), and 500 g were immunoprecipitated with ␣IRϪ3. Separation on a polyacrylamide gel was followed by transfer to nitrocellulose and blotting with anti-phosphotyrosine (4G10), followed by chemiluminescent detection, described below. Solution hybridization assays were performed as previously reported (27, 28) using a [ 32 P]UTP probe from exon 3 from the IGF-IR and 40 g of total RNA from both W and RϪ cells. Soft agar assays were performed with 5000 cells/well in 0.3% agarose on top of a 0.6% feeder layer. Colonies greater than 0.12 mm were scored and counted as previously reported (29,30).
Vector Construction, Transfection, and Clone Selection-A 1400-bp HindIII/XbaI fragment containing the EWS/FLI-1 fusion cDNA was transferred to a shuttle vector, pSP72 (Promega, Madison, WI), that was modified to contain NotI restriction sites at each end of the multiple cloning region. The NotI fragment was bidirectionally cloned into the pCI.Neo vector (Promega), and orientation was determined with restriction mapping. In vitro translation of the EWS/FLI-1 fusion cDNA yielded a full-length protein as described previously (11). Plasmids were purified using ion exchange chromatography (Qiagen, Chatsworth, CA) according to instructions from the manufacturer. Plasmids were introduced into W and RϪ cells by electroporation in serum-free Dulbecco's modified Eagles medium in a 0.4-cm gap chamber with 350 V at 1600 microfarad. Cells were then transferred to prewarmed medium. 24 h after electroporation, 400 g/ml G418 was added. Clones were isolated with cloning rings when macroscopic and grown in the continued presence of G418. Clones were frozen at regular intervals, and experiments were performed on similar passage clones.
Evaluation of Clones for EWS/FLI-1 Fusion Message-Clones were initially screened using DNA prepared from 10 5 cells in K Buffer consisting of 50 mM KCl, 10 mM Tris, pH 8, 2.5 mM MgCl 2 , 0.5% Tween 20, and 100 g/ml proteinase K. DNA was amplified using a sense primer from within the distal cytomegalovirus promoter, CI.758 (5Ј-GCTTTAT-TGCGGTAGTTTATCACAG-3Ј) and an antisense primer complimentary to the 3Ј end of the cDNA, FLI-C (5Ј-CCACGCGGATCCAGCT-TCTAGTAGTAGCTGCCTAAGTG-3Ј). PCR was performed using a kit according to instructions from the manufacturer (Perkin-Elmer, Branchberg, NJ) for 40 cycles. Cycle parameters were: annealing at 65°C for 2 min, extension at 72°C for 3 min, and denaturing at 95°C for 1 min. PCR products were separated on 2% agarose gels. Those clones containing the 2000-bp sequence including the distal one-third of the cytomegalovirus promoter and the entire cDNA were further evaluated for fusion message expression using reverse transcriptase PCR. RNA was extracted using RNAzol (Tel-Test Inc., Friendswood, TX) according to instructions from the manufacturer. 2 g of RNA was then digested with DNase I (Boehringer Mannheim) at 37°C for 1 h and then separated into two parts. Reverse transcriptase was added to one part and water to the other. cDNA synthesis was then performed according to instructions from the manufacturer using random hexamer priming (Perkin-Elmer). Expression of EWS/FLI-1 fusion message was shown using primers EWS-1 (5Ј-ATGGCGTCCACGGATTACAGTACC-3Ј) and FLI-A (5Ј-GACTGCTGGTCGGGCCCAGG-3Ј). RNA integrity was confirmed using mouse ␤-actin primers A (5Ј-GCTGCGCTGGTCGTCG-3Ј) and B (5Ј-ATCCTGTCAGCAATGCCTGG-3Ј). PCR cycle parameters were as above, and PCR products were separated on 2% agarose gels. Nomenclature for the clones is: W cells or RϪ cells that express the fusion message are designated WF, or RϪF respectively, followed by a clone number (i.e. WF3, RϪF17), whereas clones that contain the control vector but lack the fusion protein insert are designated by W or RϪ followed by clone number (i.e. W5, RϪ6).
Cell Growth Assays-Doubling times of wild-type and clonal cell lines were determined by plating equal numbers of cells in duplicate 96-well plates and evaluating cell growth daily for 7 days. Cultures were analyzed for viable cells using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma). Absorbance was analyzed at 540 nM, and cell number was determined using a standardized curve created with the software Kaleidagraph (Synergy Software, Reading, PA). The doubling time was calculated from the maximal slope of the cell growth curve. Clonally derived cell lines were evaluated in soft agar as described above.
Immunoprecipitation and Phosphoprotein Evaluation-Near-confluent monolayers of cells were washed and placed in serum-free medium for 12 h prior to stimulation with IGF-I. IGF-I (20 nM) was added for 3 min at 37°C. Lysates were prepared as previously reported (31), and protein concentration was determined for each lysate according to manufacturers instruction (BCA, Pierce) and either resolved directly on an 8% polyacrylamide gel or immunoprecipitated from fresh lysates. Immunoprecipitated proteins were removed from protein A-Sepharose beads and resolved in a polyacrylamide gel. Proteins were transferred to nitrocellulose using a semi-dry transfer apparatus. Membranes were blocked in 5% nonfat milk (Carnation, Glendale, CA) and blotted with antibodies described above. Antibodies were prepared in Tris-sodium-EDTA-Tween 20 buffer with either 1% bovine serum albumin (Miles Inc., Kankakee, IL) or 2.5% nonfat dry milk. Chemiluminescent detection was performed following incubation with secondary anti-mouse or anti-rabbit antibodies conjugated to horseradish peroxidase (Amersham Corp.) with Renaissance (Dupont) according to instructions from the manufacturer and exposure to X-Omat film (Eastman Kodak Co.). Blots were stripped with ␤-mercaptoethanol buffer (2% sodium dodecyl sulfate, 100 mM ␤-mercaptoethanol, 63 mM Tris, pH 6.8) at 70 C, reblocked, and then probed with various antibodies as described above. Films were evaluated for band density by digitizing films using a CCD camera into a Macintosh computer, checking the threshold on all bands, and analyzing with NIH Image version 1.61 software. Phosphorylation levels were corrected based upon IGF-IR or IRS-1 protein levels on each immunoblot. The density of the phosphotyrosine band was divided by the density of the protein band, either IGF-IR or IRS-1. Fold stimulation was calculated by dividing the corrected stimulated phosphotyrosine band by the unstimulated phosphotyrosine band. The equation appears in Table II. Statistics were performed using Microsoft Excell version 5.0 software.

Characterization of W and RϪ Cells-
The RϪ and W cell lines were obtained from a mouse with a targeted disruption of IGF-IR and a wild-type littermate, respectively, whereas Rϩ cells overexpress the IGF-IR. IGF-IR could not be detected on the surface of W nor RϪ cells by fluorescent automated cell sorting analysis, whereas the Rϩ cells demonstrated the presence of receptor (data not shown). To demonstrate that the RϪ cells did lack the IGF-IR and confirm expression in the W cells, we used a solution hybridization assay. Only the RNA from W cells hybridizes to and protects a 300-bp fragment of the probe (Fig. 1, lane 3), whereas RϪ cell RNA does not contain IGF-IR message (Fig. 1, lane 4). An internal 18S ribosomal probe indicates the presence of RNA in both hybridizations.
Transfection of EWS/FLI-1 into W and RϪ Cells-The pCI.Neo.EWS/FLI-1 vector was electroporated into W and RϪ cells. DNA from G418 selected clones was screened using PCR with the 5Ј primer located in the cytomegalovirus promoter, CI.758, and the distal primer at the 3Ј end of the cDNA, FLI-C. Fig. 2 demonstrates the isolation of six typical W clones, five of which were stably transfected with the EWS/FLI-1 fusion cDNA based on a 2000-bp PCR product (lanes B-F). This screening did not identify clones based upon expression of the fusion message, so following initial screening, clones were expanded for extraction of RNA. Fig. 3 shows an example of six clones evaluated for expression of the fusion message using reverse transcriptase PCR. RNA was extracted from the clones, and following this, residual DNA was eliminated by DNase digestion. Reverse transcription took place in parallel tubes, one of which contained no reverse transcriptase and acted as a control to avoid amplification of genomic (incorporated plasmid) DNA sequences. The top panel of Fig. 3  Functional Effect of EWS/FLI-1 Expression-Liquid culture growth based on doubling time did not distinguish W or RϪ clones that expressed the EWS/FLI-1 fusion protein (data not shown). Soft agar cloning was used to assay for anchorageindependent growth, a marker of transformation. Clones were tested in soft agar assays utilizing 5000 cells/35-mm 2 dish. A colony was defined as a cluster of cells larger than 0.12 m in diameter after 14 days in culture. A summary of the soft agar experiments is shown in Table I. We found that all five of our clones that express both the IGF-IR and the EWS/FLI-1 fusion protein produce colonies. Neither W (expressing IGF-IRs but lacking EWS/FLI-1), RϪF (lacking IGF-IR, but expressing EWS/FLI-1), nor RϪ (lacking both IGF-IR and EWS/FLI-1) produced colonies.
Effect of EWS/FLI-1 on IGF-IR Signaling-We first determined whether the EWS/FLI-1 fusion protein induced the synthesis or direct activation of IGF-IR. Cells were incubated in serum-free medium overnight to synchronize the cells and achieve basal phosphorylation of the IGF-IR. Identical cultures were lysed either before or after a 3-min exposure to IGF-I to stimulate the IGF-IR. Equal amounts of protein were immunoprecipitated with an IGF-IR antibody, followed by polyacrylamide gel electrophoresis, and blotting with anti-phosphotyrosine antibodies is shown in Fig. 4 (upper panel). The blot was then stripped and probed with IGF-IR ␤-subunit antibody (Fig.  4, lower panel). Densitometry was performed on all phosphotyrosine and IGF-IR blots. The fold increases in phosphorylation, after correction for IGF-IR levels, were averaged for all WF and W clones (Table II). WF clones absolute IGF-IR autophosphorylation did not differ significantly from W clones, with an average 9.6-fold stimulation compared with 6.2-fold stimulation, respectively (Student's t test, p Ͻ 0.25). Total protein lysates without immunoprecipitation were evaluated by immunoblot for IGF-IR levels and showed equal amounts of IGF-IR in both WF and W clones (data not shown).
Because neither IGF-IR protein levels nor activity was significantly altered by the EWS/FLI-1 message, we sought to determine if IGF-I-stimulated signaling distal to the IGF-IR illustrates the primer location (bold type) for this experiment and also for fusion message expression screening (Fig. 3).

FIG. 3. Expression of fusion message.
Selected clones were expanded and total RNA was extracted. The RNA was digested with DNase I and transcribed to cDNA, followed by PCR amplification (top panel) with primers EWS-1 (S) and FLI-A (AS). (see Fig. 2 for primer map) A control reaction for each specimen that lacked reverse transcriptase was similarly amplified (middle panel). The bottom panel represents the PCR product amplified from the cDNA using ␤-actin primers to ensure adequate RNA quality. All clones demonstrate intact RNA, whereas only clones WF3, WF6, RϪF2, RϪF17, and WF12 express the fusion message. Clones were evaluated for fusion message expression two times. differed between WF and W clones. Paired IGF-I stimulated/ unstimulated lysates were prepared as described above. Fig. 5  (upper panel) shows a phosphotyrosine immunoblot from 100 g of total protein, with alternate lanes being stimulated with IGF-I. The most striking difference between WF and W clones was the presence of an accentuated band in the stimulated WF lanes occurring at approximately 180 kDa. We confirmed this band to be the 180-kDa protein IRS-1 by stripping the membrane followed by specific blotting with anti-IRS-1 (Fig. 5,  lower panel). Each pair of samples are shown here to have equal amounts of IRS-1 protein, indicating that the phosphorylation difference represents increased protein phosphorylation rather than higher protein levels. Densitometry revealed a trend of increased absolute IRS-1 phosphorylation in WF clones (data not shown), so that we chose to focus our study on IRS-1 phosphorylation.
We evaluated the tyrosine phosphorylation of IRS-1 by preparing lysates in the manner described in the legend to Fig. 4 and then immunoprecipitating with a pleckstrin homology domain IRS-1 antibody from equal amounts of protein lysates (Fig. 6). The transferred blot was cut at the 43-kDa standard following blocking. The top panel was blotted for phosphotyrosine residues, whereas the middle panel was blotted for the co-precipitant Grb2. The top panel of Fig. 6 shows phosphorylation levels of IRS-1. The clones that express the fusion message, WF3 and WF36, have lower basal prestimulation levels of IRS-1 phosphorylation, whereas the two clones without the fusion message, W5 and W17, have relatively increased basal phosphorylation. A comparison of band density, with correction for IRS-1 protein levels, was performed, similar to that in Fig.  4. WF clones absolute IGF-I-induced IRS-1 phosphorylation differed significantly from W clones, with an average 6.0-fold stimulation compared with 1.5-fold stimulation, respectively (Table II, Student's t test, p Ͻ 0.05). Grb2, an adapter molecule that binds only phosphorylated IRS-1, will co-precipitate with IRS-1. We found that the levels of Grb2 in both unstimulated and stimulated lysate correspond to the phosphorylation state of IRS-1, i.e. lower Grb2 levels in unstimulated WF3 and WF36 lysates compared with unstimulated W5 and W17 lysates (Fig.  6, middle panel). The bottom panel in Fig. 6 demonstrates IRS-1 protein levels in the phosphotyrosine immunoblot following stripping and reprobing with IRS-1 carboxyl-terminal antibody.

DISCUSSION
Ewing's family of tumors contain an EWS/ets gene translocation in 95% of tumors studied (1,32). The fusion protein generated as a result of the translocation transforms fibroblasts, and inhibition of the fusion protein reduces cell growth (9,12,13). The combination of the ubiquitous nature of the translocation, the transformation properties of the fusion protein, and the growth inhibition by the reduction of fusion protein expression indicates that the EWS/ets translocation plays a significant role in the biology of ESFT.
The IGF-IR plays a critical role in growth and development (17,24,33). Ligand stimulation of the IGF-IR generates a mitogenic signal, and when overexpressed, the IGF-IR can be independently transforming (20). ESFT cells express the IGF-IR, and blockade of ligand binding results in decreased cell growth (21,22). These findings suggest that the IGF-IR signaling pathway is important for the growth of ESFT. We developed a model to evaluate the effect of the EWS/FLI-1 fusion protein upon cells in the presence and the absence of the IGF-IR. RϪ cells lack the IGF-IR, whereas the W cells express the IGF-IR.
We first sought to determine whether the EWS/FLI-1 fusion protein would transform primary mouse fibroblast cells (not all FIG. 4. Evaluation of IGF-IR levels and ligand-nduced activation. Serum-starved cells were stimulated with IGF-I for 3 min, followed by cell lysis. The IGF-IR was isolated by immunoprecipitation using an antibody directed at the ␤-subunit, and complexes were captured with protein A-Sepharose beads. Proteins were resolved in acrylamide and transferred to nitrocellulose. The upper panel was probed with a mix of anti-phosphotyrosine antibodies, 4G10 and PY-20. The blot was stripped and reprobed with an antibody directed at the ␤subunit of IGF-IR (lower panel). IGF-I stimulation is indicated by ϩ. Size markers are noted to the left of the gel. Two independent experiments were performed, and the results are combined in Table II.  To confirm that the IGF-IR pathway was activated in some way by the EWS/FLI-1 fusion protein, we selected a series of WF and W clones and analyzed the effect of ligand stimulation in the presence or the absence of the EWS/FLI-1 fusion protein.
Neither IGF-IR levels nor ligand-induced receptor activation changes as a result of the presence of the EWS/FLI-1 fusion protein. When values were averaged, there was a slight difference between WF and W clones; however, this was not significant. The docking protein IRS-I, known to bind directly to the IGF-IR (14), demonstrated an apparent hypophosphorylation in WF compared with W cells preceding IGF-I stimulation. IRS-1 phosphorylation was then specifically evaluated by immunoprecipitation and phosphotyrosine analysis. We show decreased basal levels of IRS-1 phosphorylation in the unstimulated WF clones, which contain the EWS/FLI-1 fusion protein.
The observed absolute increase in phosphorylation of IRS-1 in response to ligand stimulation of WF cells is therefore due to an altered (lower) basal phosphorylation state. We corroborated our findings with the coprecipitation of Grb2, an SH2 domaincontaining adapter protein that only binds to phosphorylated sites on IRS-1. In these studies, co-precipitated Grb-2 levels were lower in unstimulated WF cells compared with W cells, thus corroborating the decreased basal phosphorylation of IRS-1 in the unstimulated WF clones. When either WF or W cells were stimulated, coprecipitated Grb-2 levels rose in tandem with the rise in IRS-1 phosphorylation.
Our finding of increased ligand-stimulated IRS-1 phosphorylation in WF clones compared with W clones is likely a reflection of the communication between the fusion protein and the IGF-IR pathway. The data indicate that the communication between IGF-IR signaling and the target genes of the EWS/ FLI-1 fusion protein occurs downstream of the IGF-IR, based on IGF-IR protein levels and autophosphorylation activity, which are similar in both W or WF cells. This may represent a unique interaction in the ESFT or may turn out to be characteristic of how novel chimeric transcription factors function in transformation. The precise communication between the fusion protein and the IGF-IR signaling pathway remains to be specifically elucidated.
One hypothesis we are currently testing is that Syp or another protein tyrosine phosphatase may be a downstream target of the EWS/FLI-1 fusion protein. Syp, for example, is a SHPTP2 protein that is directly activated by IGF-IR (34) and as an activated tyrosine phosphatase is mitogenic (35). Additional studies have shown that Syp is required for effective signaling through the insulin receptor (36,37). IRS-1 is recognized as a target of Syp, and mutations in IRS-1 can be resistant to the phosphatase activity of Syp (38). In addition, this protein binds to phosphotyrosine residues within the SH2 domains of Grb2 and other growth regulatory receptors such as the platelet-derived growth factor receptor (39) and the epidermal growth factor receptor (40). Both the activation of Syp by IGF-IR, albeit not the only activator of Syp, and IRS-1 as a target of activated Syp, may explain the difference in basal IRS-1 phosphorylation. Syp may therefore play a key role in the communication between the EWS/FLI-1 protein and the IGF-IR pathway.
It is also possible that although the IGF-IR appears to be required in our system, what may really be required is the activation of a downstream target, like IRS-1. Combinations of transfections into RϪ cells including an oncogene, SV40 large T antigen, plus either IRS-1 or Grb2 (adapter protein downstream of many receptor tyrosine kinases, including the IGF-IR), produce cells capable of anchorage-independent growth (29,30). Transformation that occurs by combining a known transforming oncogene, SV-40 large T Ag, with the overexpression of targets known to be downstream of the IGF-IR indicates that potentially converging pathways may allow for a bypass of the IGF-IR.
We conclude that the EWS/FLI-1 fusion protein requires the IGF-IR to transform cells. Our model shows that clones expressing the EWS/FLI-1 fusion protein had increased IRS-1 phosphorylation compared with those clones without the fusion protein. Based on this, we presume that communication occurs between the EWS/FLI-1 fusion protein and the IGF-IR signaling pathway; however, the mechanism is not yet described. 3 C. Denny, personal communication. Lysates were prepared as in Fig. 4. Immunocomplexes were isolated with an antibody directed against the pleckstrin homology domain of IRS-1. Resolved proteins were transferred to nitrocellulose and probed with anti-phosphotyrosine antibodies, 4G10 and PY-20 (top panel). The middle panel is the lower half of the membrane, which was immunoblotted for Grb2. The lower panel shows the high molecular mass half of the blot, which was stripped and reprobed with anti-IRS-1 antibody (COOH-terminal domain). IGF-I stimulation is indicated by ϩ. Two independent experiments were performed, and the results are combined in Table II. In addition, two other WF clones (WF6 and WF7) were evaluated, with results similar to those shown here.
Further investigation is needed to elucidate alterations in signaling pathways that result from the aberrent EWS-FLI-1 fusion protein during transformation.