Transcriptional regulation of IGF-I receptor gene expression ... - Nature

ã

Oncogene (2002) 21, 1890 ± 1898 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Transcriptional regulation of IGF-I receptor gene expression by novel isoforms of the EWS-WT1 fusion protein Ina Finkeltov1, Scott Kuhn3,4, Tova Glaser1, Gila Idelman1, John J Wright2, Charles T Roberts Jr3 and Haim Werner*,1 1

Department of Clinical Biochemistry, Sackler School of Medicine, Tel Aviv University, Ramat Aviv, 69978 Israel; 2Medicine Branch, Division of Clinical Science, National Cancer Institute, NIH, Bethesda, Maryland MD 20889, USA; 3Department of Pediatrics, Oregon Health Sciences University, Portland, Oregon OR 97201, USA

The EWS family of genes is involved in numerous chromosomal translocations that are characteristic of a variety of sarcomas. A recently described member of this group is desmoplastic small round cell tumor (DSRCT), which is characterized by a recurrent t(11;22)(p13;q12) translocation that fuses the 5' exons of the EWS gene to the 3' exons of the WT1 gene. The originally described chimera comprises exons 1 ± 7 of EWS and exons 8 ± 10 of WT1. We have previously reported that the WT1 protein represses the expression of the IGF-I receptor gene, whereas the EWS(1 ± 7)-WT1(8 ± 10) fusion protein activates IGF-I receptor gene expression. It has recently become apparent that EWS-WT1 chimeras produced in DSCRT are heterogeneous as a result of fusions of dierent regions of the EWS gene to the WT1 gene. We have recently characterized additional EWS-WT1 translocations that involve the juxtaposition of EWS exons 7 or 8 to WT1 exon 8, and an EWS-WT1 chimera that lacks EWS exon 6. The chimeric transcription factors encoded by these various translocations dier in their DNA-binding characteristics and their ability to transactivate the IGF-I receptor promoter. These data suggest that the molecular pathology of DSRCT is more complex than previously appreciated, and that this diversity may provide the foundation for predictive genotype-phenotype correlations in the future. Oncogene (2002) 21, 1890 ± 1898. DOI: 10.1038/sj/ onc/1205042 Keywords: DSRCT; IGF-I receptor; EWS; WT1; translocation Introduction The insulin-like growth factor-I receptor (IGF-I-R) is a membrane-bound heterotetramer that mediates the trophic and dierentiative actions of IGF-I and IGFII (Cohick and Clemmons, 1993; LeRoith et al., 1995;

*Correspondence: H Werner; E-mail: [email protected] 4 Current address: Agritope, Inc., Beaverton, OR 97008, USA Received 19 March 2001; revised 30 September 2001; accepted 9 October 2001

Werner et al., 1994a). In addition to its central role in normal growth processes, the IGF-I-R plays a pivotal role in malignant transformation (Baserga et al., 1994; Grimberg and Cohen, 2000; Werner and LeRoith, 1996). The IGF-I-R is highly expressed in most tumors and cancer cell lines, where it functions as a potent antiapoptotic agent, conferring enhanced survival to malignant cells (Resnico et al., 1995; Werner and LeRoith, 1997). Transcription of the IGF-I-R gene is negatively regulated by a number of tumor suppressors, including p53, BRCA1 and WT1 (Maor et al., 2000; Ohlsson et al., 1998; Werner et al., 1995, 1996a). Inhibitory control of the IGF-I-R gene, with ensuing reduction in the levels of IGF cell-surface binding sites, has been postulated to keep the receptor in its `nonmitogenic' mode, thus preventing progression through the cell cycle (Resnico et al., 1995). The product of the WT1 Wilm's tumor suppressor gene is a nuclear protein of 52 ± 54 kDa that contains four zinc ®ngers of the C2H2 class in its C terminus (Rauscher, 1993). This domain binds target DNAs that contain versions of the consensus sequence GCGGGGGCG. We have previously shown that the WT1 protein binds to multiple sites in the IGF-I-R promoter and suppresses the activity of cotransfected promoter fragments, as well as the endogenous IGF-I-R gene (Tajinda et al., 1999; Werner et al., 1994b, 1996b). The ®ndings that the IGF-I-R gene is highly expressed in Wilms' tumors (which in many cases are associated with mutations in the WT1 gene), and that the growth of Wilms' tumor heterotransplants in nude mice is inhibited by anti-IGF-I-R antibodies, suggest that the IGF-I-R gene constitutes a bona ®de molecular target of WT1 action (Gansler et al., 1989; Werner et al., 1993). Tumor-speci®c chromosomal translocations that disrupt the molecular architecture of transcription factors have emerged as a common theme in oncogenesis (Rabbitts, 1994). As a result of these rearrangements, chimeric proteins are generated that are composed of modules derived from unrelated genes. The EWS gene is a member of the TET gene family (Plougastel et al., 1993). The products of this gene family are RNA-binding proteins of unknown physiological function that contain potential transcriptional activation domains in their N termini (Ohno et al.,

EWS-WT1 regulation of the IGF-I receptor gene I Finkeltov et al

1994). The EWS gene is involved in a number of translocation events that are hallmarks of a variety of speci®c cancers (Ladanyi, 1995). Ewing's sarcoma is associated with a t(11 : 22)(q24;q12) translocation that fuses the 5' exons of the EWS gene on chromosome 22 to the 3' exons of the FLI-1 gene (Zucman et al., 1993b; Delattre et al., 1992; Jeon et al., 1995). Fusion of the EWS and ATF-1 genes is characteristic of soft tissue clear cell sarcoma or malignant melanoma of soft parts (Zucman et al., 1993a). EWS-CHN gene fusions have been reported in human myxoid chondrosarcoma (Clark et al., 1996; Gill et al., 1995). Although these fusions exhibit heterogeneity with respect to the position of the chromosomal breakpoints (Kuroda et al., 1995; Zucman et al., 1992), a common feature is the fusion of the N-terminal transcription activation domain of EWS to either the full-length sequence or the C-terminal DNA-binding domain of any of a number of transcription factors. Antisense and antibody inhibition of the chimeric EWS-FLI-1 and EWS-ATF1 products reduces the tumorigenicity and viability of Ewing's and clear cell carcinoma cells, respectively (Bosilevac et al., 1999; Ouchida et al., 1995), implicating the altered transcriptional activity of these chimeric proteins in tumor development. Desmoplastic small round cell tumor (DSRCT), a very aggressive primitive tumor of children and young adults, particularly males (Gerald et al., 1991; Gonzalez-Crussi et al., 1990), is a recently described malignancy that is also characterized by a recurrent translocation involving the EWS gene. The distinctive hallmark of DSRCT is the t(11;22)(p13;q12) translocation (Sawyer et al., 1992) that fuses the N-terminal (activation) domain of EWS to the C-terminal (DNAbinding) domain of WT1, including the three terminal zinc ®ngers (Brodie et al., 1995; Gerald et al., 1995; Ladanyi and Gerald, 1994; Rauscher et al., 1994). Consistent with the postulate that the EWS-WT1 fusion protein may modulate transcription of target genes containing WT1 binding motifs, we have previously shown that EWS(1 ± 7)-WT1(8 ± 10) can recognize and transactivate the IGF-I-R promoter in transient transfection assays (Karnieli et al., 1996). More recently, we and others have shown that the t(11;22)(p13;q12) translocation is also a heterogeneous event that may involve several alternative breakpoints (Antonescu et al., 1998; Chan et al., 1999; Liu et al., 2000; Shimizu et al., 1998). Speci®cally, EWS-WT1 chimeras have been described in which exons 7, 8, 9 or 10 of the EWS gene are fused to exon 8 of the WT1 gene. We have also identi®ed an additional EWS exon 7-WT1 exon 8 fusion that lacks EWS exon 6 (Liu et al., 2000). As a result of the alternative splicing of exon 9 that occurs in the WT1 gene (Haber et al., 1991), each of these fusion proteins may include or lack a Lys-Thr-Ser (KTS) insert located between zinc ®ngers 3 and 4 of the WT1 portion of the chimera. This diversity suggests that the etiology of DSRCT is much more complex than originally suspected. In light of the important role of the IGF-I-R in tumorigenesis, we have investigated the potential dierential regula-

tion of the IGF-I-R promoter by various EWS-WT1 isoforms.

1891

Results Reverse transcription-polymerase chain reaction (RT ± PCR) of DSRCT-derived RNA samples, followed by direct sequencing of PCR products, allowed recently one of us to identify a number of novel breakpoint regions in fusion transcripts (Liu et al., 2000; Figure 1). In addition to the previously described and most frequently detected chimeric gene (present in 12 out of 14 tumors) resulting from the in-frame fusion of exons 1 ± 7 of EWS to exons 8 ± 10 of WT1 (7/8), a novel isoform in which EWS exon 8 is fused to WT1 exon 8 (8/8) was identi®ed in one tumor. A 6-bp heterologous sequence inserted at the junction and a 4-bp deletion of the 5' end of WT1 exon 8 forms an in-frame fusion of the EWS and WT1 ORFs. An additional novel variant resulting from the 7/8 fusion, but lacking EWS exon 6 (7/8D6), was isolated from four tumors. Since WT1 exon 9 is subject to alternative splicing, DSRCT samples express isoforms that either include or lack the KTS insert. To verify that the expression constructs to be utilized in functional assays of transcriptional activity and in DNA-binding experiments encoded proteins of the expected size, synthetic RNAs were generated by in vitro transcription of pcDNA3 constructs and used to program rabbit reticulocyte lysates for in vitro translation in a coupled system. As shown in Figure 2, the apparent molecular weight of the 35S-methioninelabeled EWS-WT1(7/8+KTS) protein was *56 ± 58 kDa, the size of the EWS-WT1(8/8+KTS) protein was *60 ± 62 kDa, and the size of the EWS-WT1(7/8 D67KTS) was *48 ± 50 kDa. Although the sizes of the in vitro-translated proteins were slightly larger than expected, similar discrepancies have been reported by others (Kim et al., 1998). To establish whether WT1 binding sites located in the 5'-¯anking region of the IGF-I-R gene are involved in EWS-WT1 protein binding, electrophoretic mobility shift assays (EMSA) were performed using a labeled promoter fragment extending from 7331 to 740. This DNA region was previously shown to bind EWSWT1(7/8 7KTS)-derived recombinant proteins containing the 21-amino acid fragment of EWS located Nterminal to the fusion point and most of the C-terminal domain of WT1, including zinc ®ngers 2 ± 4 (Karnieli et al., 1996). To assess the potential contribution of the dierent N-terminal EWS-derived domains (i.e., exons 1 ± 7, exons 1 ± 7 D6, or exons 1 ± 8) on DNA binding, full-length in vitro-translated proteins were employed. As shown in Figure 3, incubation of the labeled DNA fragment with the full-length EWS-WT1(7/8 D67KTS) fusion protein generated one retarded band. Incubation with the EWS/WT1(7/87KTS) protein resulted in the formation of three retarded bands, whereas the 7/8 isoform containing the KTS insert (EWS/WT1(7/ 8+KTS)) did not generate any retarded band. Oncogene


1892

Figure 1 Schematic representation of EWS, WT1, and EWS-WT1 proteins. The EWS gene is located at 22q12 and includes 17 exons. The EWS gene product is a 656-amino acid protein that includes an N-terminal activation domain (AD, dotted) and an RNA-binding domain (RBD, black). Exon 6, which is missing in the EWS-WT1(7/8 D67KTS) fusion protein, is boxed. The WT1 gene, located at 11p13, includes 10 exons and encodes a 429-amino acid protein. The WT1 protein includes an N-terminal repression domain (RD, dashed) and a DNA-binding domain (DBD) comprised of four zinc-®nger motifs. Alternative splicing of exon 9 results in WT1 molecules containing or lacking a Lys-Thr-Ser (KTS) insert between zinc ®ngers 3 and 4. DSRCT is characterized by a recurrent t(11;22)(p12;q12) translocation that fuses either exons 1 ± 7 or 1 ± 8 of the EWS gene to exons 8 ± 10 of the WT1 gene. As a result of the ®rst rearrangement, EWS-WT1(7/8+KTS) fusion proteins including or lacking the KTS insert are generated. An additional fusion protein that results from the same translocation lacks EWS exon 6 (EWS-WT1(7/8 D67KTS)). As a result of the second event, EWS-WT1(8/8+KTS) fusion proteins are generated. Arrowheads indicate breakpoints in the EWS and WT1 genes. Exon numbers of the EWS and WT1 genes are shown above the protein diagrams

Similarly, no binding was seen with the EWS/WT1(8/ 8+KTS) chimeric proteins. Band formation was abrogated when the binding reactions were performed in the presence of a *200-fold molar excess of unlabeled probe (data not shown). In addition, potential DNA binding of the various chimeric proteins was assessed using a labeled genomic DNA fragment extending from nucleotide 740 to +115 (including the IGF-I-R `initiator' element), and a double-stranded synthetic oligonucleotide extending from nucleotide 7382 to 7354 (including three Sp1 binding sites). No speci®c binding was observed to either of these DNA fragments (data not shown). We then evaluated the ability of the various EWSWT1 isoforms to regulate the activity of the IGF-I-R promoter in transient transfection assays. For this purpose, osteosarcoma-derived Saos-2 cells and malignant rhabdoid tumor-derived G401 cells were cotransOncogene

fected with a series of expression vectors encoding EWS-WT1(7/8+KTS), EWS-WT1(8/8+KTS), and EWS-WT1(7/8 D67KTS), together with a reporter construct [p(7476/+640)LUC] containing most of the proximal region of the IGF-I-R promoter upstream of a luciferase reporter gene. This promoter fragment has a high G-C content (475%), and, using DNase I footprinting assays with the expressed zinc-®nger domain of WT1, we have previously mapped 12 WT1 binding sites in this region (Werner et al., 1994b). As shown in Figure 4a, cotransfection of Saos-2 cells with 2.5 mg of the EWS-WT1(7/8 D67KTS) isoform resulted in the strongest transactivation (617+172% of control values). Consistent with our previous report (Karnieli et al., 1996), the activity of the EWS-WT1(7/ 8+KTS) protein was signi®cantly lower than that of the 7KTS isoform (EWS-WT1(7/87KTS)) (172+ 65% vs 481+147% stimulation). Interestingly, both


1893

Figure 2 In vitro transcription/translation analysis of EWS-WT1 expression vectors. T7 RNA polymerase-driven in vitro transcription reactions were performed using the TNT1 T7 Quick Coupled Transcription/Translation System (Promega). Each reaction included 1 mg of each of the following expression vectors: empty pcDNA3, EWS-WT1(7/8 D67KTS), EWS-WT1(7/8+KTS), and EWS-WT1(8/8+KTS). In vitro translation reactions were performed in the presence of 35S-labeled methionine using rabbit reticulocyte lysates. One ml of each translation reaction (out of a total volume of 50 ml) were electrophoresed through 8% SDS ± PAGE gels and exposed for 20 h to Kodak X-Omat ®lm. The sizes of prestained molecular weight markers (Sigma) are indicated

EWS-WT1(8/8+KTS) isoforms stimulated promoter activity to a similar extent (326+50% vs 326+84%). As expected, cotransfection of a WT1 expression vector consistently suppressed promoter activity (data not shown). On the other hand, only the EWS-WT1(7/8 D67KTS) isoform exhibited transactivational activity in G401 cells (329+12% of controls) (Figure 4b). To more precisely map the promoter region responsible for mediating the eect of EWS-WT1, cotransfections were performed using the reporter plasmids p(7188/+640)LUC and p(740/+640)LUC (Figure 5a), along with the EWS-WT1(7/8 D67KTS) expression vector. Construct p(7188/+640)LUC lacks four WT1 consensus sites (located at 7262/7254, 7250/7242, 7220/7212, and 7196/7188) that we have previously shown in DNaseI footprinting assays to bind a recombinant WT1 zinc-®nger domain with relatively high anity (Werner et al., 1994b). Construct p(740/+640)LUC lacks, in addition, a WT1 ciselement at position 7163/7155 that binds WT1 with medium anity. The transactivation activity of the EWS-WT1(7/8 D67KTS) protein on the p(7188/ +640)LUC construct in Saos-2 cells was signi®cantly reduced in comparison to the p(7476/+640)LUC

Figure 3 Dierential binding of EWS-WT1 isoforms to the IGF-I-R promoter. The promoter region extending from 7331 to 740 (containing ®ve WT1 binding sites) was end-labeled with g-32P-ATP in the presence of T4 polynucleotide kinase, and employed in gel-shift assays with equal amounts of the dierent in vitro-translated EWS-WT1 isoforms (or with control pcDNA3 reaction product). Binding reactions were performed as indicated in the Materials and methods section. Arrows denote the positions of three DNA-protein complexes that result from the binding of the EWS-WT1(7/87KTS) protein to the IGF-I-R promoter. The dotted arrow shows the position of a DNA ± EWSWT1(7/8 D67KTS) complex

plasmid (199+28% vs 617+172% of controls, respectively). No further reduction was seen with the p(740/ +640)LUC construct (221+42% of control) (Figure 5b). Likewise, the stimulatory eect of EWS-WT1(7/8 D67KTS) in G401 cells was completely abrogated when the reporter plasmids p(7188/+640)LUC and p(740/+640)LUC were transfected (Figure 5c). To examine the physiological relevance of the transactivation eect of EWS-WT1 chimeras on the IGF-I-R promoter, we measured the levels of endogenous IGF-I-R protein following transient transfections using Western blot analysis. Saos-2 cells were transfected with 6 mg of the EWS-WT1(7/8 D67KTS) expression vector, and after 24, 48 and 72 h, cells were lysed as described in the Materials and methods section. Cell lysates (50 mg protein) were electrophoresed through 10% SDS ± PAGE gels, blotted onto nitrocellulose membranes, and incubated with an antiIGF-I-R b-subunit antibody. Western analysis of lysates of transfected cells demonstrated a 1.7-fold increase in the levels of the IGF-I-R at 72 h after transfection (Figure 6). Finally, triple cotransfections were performed using pCB6+-derived expression vectors encoding EWSWT1(7/87KTS) and WT1, together with the Oncogene


1894

Figure 4 Stimulation of IGF-I-R promoter activity by EWSWT1 isoforms. Five micrograms of the p(7476/+640)LUC reporter plasmid (shown in Figure 5A) were cotransfected into Saos-2 cells (A) with 2.5 mg of each of the EWS-WT1 expression vectors (or empty pcDNA3) and 2.5 mg of the pCMVb plasmid using the calcium phosphate method. G401 cells (B) were cotransfected with 0.6 mg of reporter plasmid, 0.2 mg of expression vector, and 0.2 mg of the pCMVb plasmid using the Fugene-6 Reagent (Roche Molecular Biochemicals). After 40 h, cells were harvested and the levels of luciferase and bgalactosidase activities were measured. Luciferase values, normalized for b-galactosidase, are expressed as a percentage of the luciferase activity of the empty pcDNA3 expression vector. Experiments were performed between three and ®ve times (Saos-2), and two times (G401), each time in duplicate. Bars are mean+s.e.m. Where not shown, the s.e.m. bars are smaller than the symbol size Oncogene

Figure 5 Mapping of EWS-WT1 target regions of the IGF-I-R promoter. (A) Schematic representation of reporter constructs. Plasmids p(7476/+640)LUC, p(7188/+640)LUC, and p(740/ +640)LUC contain, respectively, 476, 188, or 40 bp of 5'-¯anking region (open bar), and 640 bp of 5'-untranslated region (closed bar), linked to a luciferase cDNA (LUC). The arrow indicates the transcription initiation site. Triangles denote the location of WT1 binding sites (7262/7254, 7250/7242, 7220/7212, and 7196/ 7188) that were footprinted by the puri®ed zinc-®nger domain of WT1 with relatively high anity. The site at 7163/7155 (denoted by a circle) was shown to bind WT1 with medium anity. The luciferase cDNA is not shown to scale. Saos-2 (B) and G401 (C) cells were cotransfected with the indicated reporter plasmid, along with the EWS-WT1(7/8 D67KTS) and bgalactosidase expression plasmids, as described in the legend to Figure 3 and in the Materials and methods section. The luciferase values, normalized for b-galactosidase (closed bars), are expressed as a percentage of the activity seen in the absence of expression vector (open bars). Results of transfections in Saos-2 (B) are mean+s.e.m. (n=3 experiments, performed each one in duplicate). (C) (G401) shows the results of a representative experiment, performed in duplicate. Note the dierences in scale in the y axis between panels B and C

p(7476/+640)LUC reporter plasmid. The rationale for these experiments was the fact that tumor cells containing the t(11;22)(p13;q12) translocation still retain one normal allele encoding wild-type WT1. Because WT1 was previously shown to bind to consensus cis elements in the IGF-I-R promoter with


1895

Figure 6 Eect of EWS-WT1 expression on endogenous IGF-IR gene expression. Saos-2 cells were transfected with 6 mg of the EWS-WT1 (7/8 D67KTS) expression vector (or empty pcDNA3) using the Fugene 6 Reagent. Cells were collected after 24, 48, and 72 h as indicated in the Materials and methods section. Equal amounts of protein (50 mg) were separated by 10% SDSpolyacrylamide gel electrophoresis, transferred to nitrocellulose ®lters and blotted with an anti-human IGF-I-R antibody

an anity similar to that displayed by the fusion protein, it was important to assess potential functional interactions between WT1 and EWS-WT1 in transcriptional control of the IGF-I-R gene. As shown in Figure 7, cotransfection of EWS-WT1(7/87KTS) attenuated the repression of the IGF-I-R promoter seen with WT1 alone. Discussion DSRCT is a recently described primitive tumor that occurs most frequently in adolescent males and is mainly localized in the abdomen. As is the case with Wilms' tumors, pathological analyses of DSRCT samples revealed the presence of epithelial, mesenchymal, and neural-derived cells. The distinctive hallmark of DSRCT is a recurring t(11;22)(p13;q12) chromosomal translocation that fuses the N-terminal domain of the EWS gene in frame to the C-terminal, DNAbinding domain of WT1. The chimeric protein that results from this rearrangement, EWS-WT1, is an example of a growing family of aberrant transcription factors harboring domains encoded by unrelated genes. In general, these disrupted transcription factors display gain-of-function activity that abrogates the biological role of each one of the parental genes, resulting in an oncogenic role for the chimeric protein. In the particular case of EWS-WT1, the translocation event seems to abolish the RNA-binding activity of EWS as well as the transcriptional repression activity of WT1. Furthermore, the EWS-WT1(7/87KTS) protein exhibits oncogenic activity as determined by formation of transformed foci in cell monolayers, anchorage-independent growth, and tumor formation in nude mice. The EWS-WT1(7/8+KTS) protein, on the other hand, exhibited no transforming potential (Kim et al., 1998). The involvement of the IGF system in the etiology of a number of solid tumors, including Wilms' tumor, rhabdomyosarcoma, and others, has been well established. The main ligand in most of these malignancies appears to be IGF-II, whose mitogenic and proliferative activity is mediated by the IGF-I-R (El-Badry et al., 1990). The expression of the IGF-I-R gene is

Figure 7 Functional interactions between EWS-WT1 and WT1 proteins in regulation of IGF-I-R promoter activity. Cotransfections were performed in Saos-2 cells using 5 mg of the p(7476/ +640)LUC plasmid and 0.2 mg pCMVb, together with 2.1 mg of pCB6+DNA, 0.1 mg of the pCB6+(WT1-KTS), 2.0 mg of pCB6+(EWS-WT1(7/87KTS)), or both expression vectors. Total DNA amount transfected was kept constant with pCB6+. The values of luciferase normalized for b-galactosidase are expressed as a percentage of the activity seen with the pCB6+ control DNA. The results of a typical experiment repeated at least four times are presented

negatively controlled by a number of tumor suppressors, including WT1. In the present study, we have expanded our previous observations showing that the IGF-I-R gene is a molecular target for the EWSWT1(7/8) protein, and that the transactivating eect of the fusion protein involves binding to WT1-related sites in the IGF-I-R promoter (Karnieli et al., 1996). Speci®cally, the recent ®nding that alternative breakpoints in the EWS and WT1 genes are involved in the t(11;22)(p13;q12) translocation suggested that the molecular basis underlying DSRCT pathology may, in fact, be more complex than originally appreciated, in that variant translocation products may display dierent transactivating activities. The results of this study are in accord with this concept by demonstrating that EWS-WT1 isoforms stimulated IGF-I-R promoter activity to dierent extents and, additionally, exhibited distinct DNA-binding properties. One of us has recently reported that EWSWT1(7KTS) isoforms signi®cantly activated an arti®cial promoter containing three tandem copies of the early growth response (EGR) binding site (EBS), as well as the activity of the EGR-1 promoter (Liu et al., 2000). These activities correlated with the speci®c binding of the 7KTS forms of EWS-WT1 to labeled EBS oligonucleotides. The results of the present study extend the spectrum of potential targets of EWS-WT1 isoforms to a gene encoding a growth factor receptor Oncogene


1896

Oncogene

whose involvement in the etiology of a large number of pediatric tumors is well established. In addition, our ®ndings with the complex IGF-I-R promoter (which contains 12 WT1/EBS binding sites in both the 5'¯anking and 5'-untranslated regions, in comparison to only one site in the EGR-1 promoter) dier from those obtained by Liu et al. (2000) and therefore suggest that EWS-WT1 isoforms activate target promoters of dierent complexities via distinct mechanisms. The EWS-WT1(7/8 D67KTS) chimera exhibited the strongest transactivation activity (*6.2-fold stimulation above control), and this eect was associated with its speci®c binding to a promoter fragment extending from nucleotides 7331 to 740 that includes four highanity WT1 sites (7262/7254, 7250/7242, 7220/ 7212, and 7196/7188) and one medium-anity site (7163/7155). Because the stimulatory eect of EWSWT1 (7/8 D67KTS) was abrogated when a promoter fragment lacking the four high-anity sites was employed, and since EMSA data demonstrated the formation of a single retarded band, we infer that only one site in the cluster of high-anity WT1 elements was responsible for this eect. Furthermore, the fact that no further reduction in transcriptional activity was seen with the p(740/+640) LUC construct in comparison to p(7188/+640)LUC suggests that the medium-anity WT1 site at position 7163/7155 had no major eect. A strong stimulatory eect was also displayed by the EWS-WT1 (7/87KTS) chimera (*4.8-fold stimulation above control), but not by the 7/8 isoform that includes the KTS insert. These dierences in the results of functional assays correlated with marked dierences in DNA-binding patterns. Thus, whereas the 7KTS isoform generated three retarded bands in EMSA, no discernible binding was displayed by the+KTS isoform. Furthermore, we have previously shown that, unlike the exon 6-deleted isoform, the exon 6-containing chimera retained its stimulatory eect towards the p(7188/+640)LUC, but not towards the p(740/ +640)LUC region (Karnieli et al., 1996). These results suggest that the medium-anity site at position 7163/ 7155 is important for the action of the EWS-WT1(7/ 87KTS) protein. Finally, both EWS-WT1(8/8) variants had a relatively lower transactivating eect (*3.3-fold stimulation), regardless of the presence or absence of the KTS insert. Accordingly, we were unable to detect any speci®c DNA binding by either 8/8 isoform. It is possible that the mechanism of action underlying the transactivating eect of particular EWSWT1 isoforms involves protein ± protein interactions with cell-type speci®c DNA-binding or non-binding transcription factors. This possibility is underscored by the dierent transactivating activities displayed by the various isoforms in Saos-2 and G401 cells. An alternative mechanism of action may involve binding to another region of the IGF-IR promoter. Taken together, the results of this study show that the lack of the EWS exon 6-encoded sequence results in speci®c binding to a unique high-anity site (in comparison to the exon 6-containing isoform that

appears to bind to at least three sites). On the other hand, the presence of EWS exon 8 had a negative impact on the DNA-binding activity and transactivating eect of 8/8 isoforms, probably as a result of steric interference of exon 8-encoded sequences with promoter binding. In addition, the ®nding that the EWSWT1(7/8 D67KTS) isoform induced an increase in levels of endogenous IGF-I-R protein strongly suggests that the IGF-IR gene is a physiological target of EWSWT1 chimeric proteins. In the pathophysiological context of DSRCT, EWSWT1 fusion proteins are presumably functioning in the presence of wild-type WT1 encoded by the untranslocated allele. The results of triple cotransfection experiments suggest that EWS-WT1 chimeras can attenuate the repression of the IGF-I-R gene exerted by intact WT1. This interference with normal WT1 function, along with WT1 haploinsuciency and potential gain-of-function eects of the EWS-WT1 chimeras, may all contribute to the etiology of DSRCT. In conclusion, our results demonstrate that the IGFI-R gene is a target for a novel family of EWS-WT1 fusion proteins, and that the binding activities of EWSWT1 isoforms are not only determined by the WT1 component of the chimera, as usually believed, but may also be aected to a large extent by the particular EWS-derived exons that are retained in the translocation product. The contribution of this variability in EWS-WT1 translocation products to heterogeneity in DSRCT remains to be established.

Materials and methods Construction of EWS-WT1 expression vectors EWS-WT1/pcDNA3 expression plasmids were constructed by excising EcoRI ± XbaI fragments from previously described pAMP vectors (Liu et al., 2000) and ligating these fragments into the corresponding EcoRI ± XbaI sites in the multiple cloning site of pcDNA3 (Invitrogen). With the exception of the 8/8+KTS fusion, this positioned the gene in the sense orientation under the in¯uence of the CMV promoter. The 8/8+KTS cDNA was originally cloned in the opposite orientation in the pAMP vector, so a HindIII ± SmaI fragment was isolated and ligated into the HindIII and EcoRV sites of pcDNA3. All pcDNA3 constructs were sequenced with both T7 and Sp6 primers to ensure their proper orientation. Cell cultures and transfections The human osteogenic sarcoma-derived cell line Saos-2 and the human malignant rhabdoid tumor cell line G401 were obtained from the American Type Culture Collection (Rockville, MD, USA). Saos-2 cells were grown in Dulbecco's modi®ed Eagle's medium and G401 cells were cultured in McCoy's 5A medium. Media were supplemented with 10% fetal bovine serum, 2 mM glutamine, and 50 mg/ml gentamicin sulfate. Both cell lines were previously shown to express signi®cant amounts of IGF-I-R mRNA and to support transcription of IGF-I-R promoter-driven reporter constructs


(Ohlsson et al., 1998; Werner et al., 1995). Furthermore, the activity of IGF-I-R promoter constructs in G401 cells was shown to be strongly suppressed by cotransfection of a WT1 expression plasmid (Werner et al., 1995). The p(7476/+640)LUC, p(7188/+640)LUC, and p(740/ +640)LUC IGF-I-R promoter-pOLUC constructs used in transient transfection assays have been described previously (Karnieli et al., 1996). Cells were seeded in 6-well plates and transfected using the calcium phosphate method (Saos-2) or the Fugene-6 Reagent (Roche Molecular Biochemicals, Indianapolis, IN, USA) (G401). Saos-2 cells received 5 mg of reporter plasmid and 2.5 mg of expression vector (or empty pcDNA3), along with 2.5 mg of a b-galactosidase expression plasmid (pCMVb, Clontech, Palo Alto, CA, USA). G401 cells were transfected with 0.6 mg of reporter plasmid, 0.2 mg of expression vector, and 0.2 mg of pCMVb. Cells were harvested 40 h after transfection, and luciferase and b-galactosidase activities were measured as previously described (Werner et al., 1992). Promoter activities were expressed as luciferase values normalized for b-galactosidase activity. Triple cotransfections were performed using the p(7476/ +640)LUC reporter vector together with EWS-WT1 (7/ 87KTS) and WT1(7KTS) (pCMVhWT) expression vectors in the pCB6+plasmid (Werner et al., 1993). The total DNA amount transfected was kept constant with empty pCB6+ plasmid. The pCB6+ derived expression vectors were provided by Dr Frank J Rauscher III (The Wistar Institute, PA, USA). In vitro transcription and translation reactions Coupled in vitro transcription/translation of the various EWS-WT1 fusion proteins was performed using the TNT1 T7 Quick-Coupled Transcription/Translation System (Promega, Madison, WI, USA). Brie¯y, T7 RNA polymerasedriven in vitro transcription reactions were followed by in vitro translations in the presence of 35S methionine (or unlabeled methionine for EMSA) using rabbit reticulocyte lysates. In vitro translation products were electrophoresed through 8% SDS ± PAGE and exposed to Kodak X-Omat ®lm. Western immunoblotting Transiently transfected cells were harvested with ice-cold PBS containing 5 mM EDTA and lysed in a buer composed of 150 mM NaCl, 20 mM HEPES, pH 7.5, 1% Triton X-100, 2 mM EDTA, 2 mM EGTA, 1 mM PMSF, 2 mg/ml aprotinin, 1 mM leupeptin, 1 mM pyrophosphate, 1 mM vanadate and

1 mM DTT. Protein content of the lysates was determined using the Bradford reagent (Bio-Rad, Hercules, CA, USA) using BSA as a standard. Samples were subjected to 10% SDS ± PAGE, followed by electrophoretic transfer of the proteins to nitrocellulose membranes. Membranes were blocked with 3% BSA in T-TBS (20 mM Tris-HCl, pH 7.5, 135 mM NaCl and 0.1% Tween-20). Blots were incubated with a rabbit polyclonal anti-human IGF-I-R b-subunit antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 0.2 mg/ml), washed extensively with T-TBS, and incubated with an HRP-conjugated secondary antibody. Proteins were detected using the SuperSignal1 West Pico Chemiluminescent Substrate (Pierce, Rockford, IL, USA).

1897

Electrophoretic mobility shift assays A DNA fragment extending from nucleotides 7331 to 740 of the IGF-I-R 5'-¯anking region was isolated by digestion of an IGF-I-R genomic clone with XmaI and PmlI. Following digestion, the 7331/740 fragment was gel puri®ed and endlabeled with g-32P-ATP using T4 polynucleotide kinase. Binding reactions were performed in a total volume of 10 ml for 30 min at 48C in 20 mM HEPES, pH 7.5, 70 mM KCl, 12% glycerol, 0.05% Nonidet P-40, 100 mM ZnSO4, 50 mM dithiothreitol, 5 mM MgCl2, 1 mg/ml bovine serum albumin, 0.1 mg/ml poly(dI:dC), 75 000 d.p.m. (0.15 ± 0.75 ng) of the labeled fragment, and 5 ml of the in vitro translated EWS-WT1 proteins. Changes in mobility were assessed by electrophoresis through native 5% polyacrylamide gels that were run at 120 V for 2 h in 0.56TBE. After ®xation in 10% acetic acid, gels were autoradiographed at 7708C. Abbreviations IGF-I-R, insulin-like growth factor-I receptor; DSRCT, desmoplastic small round cell tumor; EWS, Ewing's Sarcoma; WT, Wilms' tumor; RT ± PCR, reverse transcription polymerase chain reaction.

Acknowledgments This work was performed in partial ful®llment of the requirements for the MSc degree by Ina Finkeltov, in the Sackler Faculty of Medicine, Tel Aviv University. HW is the recipient of a Guastalla Fellowship from the Rashi Foundation, Israel. This work was supported by a grant from the Israel Cancer Association to H Werner and by NIH grant DK50810 to CT Roberts.

References Antonescu CR, Gerald WL, Magid MS and Ladanyi M. (1998). Diag. Mol. Pathol., 7, 24 ± 28. Baserga R, Sell C, Porcu P and Rubini M. (1994). Cell Prolif., 27, 63 ± 71. Bosilevac JM, Olsen RJ, Bridge JA and Hinrichs SH. (1999). J. Biol. Chem., 274, 34811 ± 34818. Brodie SG, Stocker SJ, Wardlaw JC, Duncan MH, McConnell TS, Feddersen RM and Williams TM. (1995). Human Pathol., 26, 1370 ± 1374. Chan AS, MacNeill S, Thorner P, Squire J and Zielenska M. (1999). Ped. Dev. Pathol., 2, 188 ± 192.

Clark J, Benjamin H, Gill S, Sidhar S, Goodwin G, Crew J, Gusterson BA, Shipley J and Cooper CS. (1996). Oncogene, 12, 229 ± 235. Cohick WS and Clemmons DR. (1993). Ann. Rev. Physiol., 55, 131 ± 153. Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M, Kovar H, Joubert I, de Jong P, Rouleau G, Aurias A and Thomas G. (1992). Nature, 359, 162 ± 165. El-Badry OM, Minniti C, Kohn EC, Houghton PJ, Daughaday WH and Helman LJ. (1990). Cell Growth Dier., 1, 325 ± 331.

Oncogene


1898

Oncogene

Gansler T, Furlanetto R, Gramling TS, Robinson KA, Blocker N, Buse MG, Sens DA and Garvin AJ. (1989). Am. J. Pathol., 135, 961 ± 966. Gerald WL, Miller HK, Battifora H, Miettinen M, Silva EG and Rosai J. (1991). Am. J. Surg. Pathol., 15, 499 ± 513. Gerald WL, Rosai J and Ladanyi M. (1995). Proc. Natl. Acad. Sci. USA, 92, 1028 ± 1032. Gill S, McManus AP, Crew AJ, Benjamin H, Sheer D, Gusterson DA, Pinkerton CR, Patel K, Cooper CS and Shipley JM. (1995). Genes Chromos. Cancer, 12, 307 ± 310. Gonzalez-Crussi F, Crawford SE and Sun CC. (1990). Am. J. Surg. Pathol., 14, 633 ± 642. Grimberg A and Cohen P. (2000). J. Cell. Physiol., 183, 1 ± 9. Haber DA, Sohn RL, Buckler AJ, Pelletier J, Call KM and Housman DE. (1991). Proc. Natl. Acad. Sci. USA, 88, 9618 ± 9622. Jeon IS, Davis JN, Braun BS, Sublett JE, Roussel MF, Denny CT and Shapiro DN. (1995). Oncogene, 10, 1229 ± 1234. Karnieli E, Werner H, Rauscher III FJ, Benjamin LE and LeRoith D. (1996). J. Biol. Chem., 271, 19304 ± 19309. Kim J, Lee K and Pelletier J. (1998). Oncogene, 16, 1973 ± 1979. Kuroda M, Ishida T, Horiuchi H, Kida N, Uozaki H, Takeuchi H, Tsuji K, Imamura T, Mori S and Machinami R. (1995). Am. J. Pathol., 147, 1221 ± 1227. Ladanyi M. (1995). Diag. Mol. Pathol., 4, 162 ± 173. Ladanyi M and Gerald W. (1994). Cancer Res., 54, 2837 ± 2840. LeRoith D, Werner H, Beitner-Johnson D and Roberts Jr CT. (1995). Endocrine Rev., 16, 143 ± 163. Liu J, Nau MM, Yeh JC, Allegra CJ, Chu E and Wright JJ. (2000). Clin. Cancer Res., 6, 3522 ± 3529. Maor SB, Abramovitch S, Erdos MR, Brody LC and Werner H. (2000). Mol. Gen. Metab., 69, 130 ± 136. Ohlsson C, Kley N, Werner H and LeRoith D. (1998). Endocrinology, 139, 1101 ± 1107. Ohno T, Ouchida M, Lee L, Gatalica Z, Rao VN and Reddy ES. (1994). Oncogene, 9, 3087 ± 3097. Ouchida M, Ohno T, Fujimura Y, Rao VN and Reddy ES. (1995). Oncogene, 11, 1049 ± 1054. Plougastel B, Zucman J, Peter M, Thomas G and Delattre O. (1993). Genomics, 18, 609 ± 615. Rabbitts TH. (1994). Nature, 372, 143 ± 149. Rauscher III FJ. (1993). FASEB J., 7, 896 ± 903.

Rauscher III FJ, Benjamin LE, Fredericks WJ and Morris JF. (1994). Cold Spring Harbor Symp. Quant. Biol., 59, 137 ± 146. Resnico M, Burgaud J-L, Rotman HL, Abraham D and Baserga R. (1995). Cancer Res., 55, 3739 ± 3741. Sawyer JR, Tryka AF and Lewis JM. (1992). Am. J. Surg. Pathol., 16, 411 ± 416. Shimizu Y, Mitsui T, Kawakami T, Ikegami T, Kanazawa C, Katsuura M, Obata K, Yamagiwa I and Hayasaka K. (1998). Cancer Genet. Cytogenet., 106, 156 ± 158. Tajinda K, Carroll J and Roberts Jr CT. (1999). Endocrinology, 140, 4713 ± 4724. Werner H, Adamo M, Roberts Jr CT and LeRoith D. (1994a). Vitamins and Hormones. Vol. 48. Litwack G (ed.). Academic Press: San Diego, CA, pp. 1 ± 58. Werner H, Bach MA, Stannard B, Roberts Jr CT and LeRoith D. (1992). Mol. Endocrin., 6, 1545 ± 1558. Werner H, Karnieli E, Rauscher III FJ and LeRoith D. (1996a). Proc. Natl. Acad. Sci. USA, 93, 8318 ± 8323. Werner H and LeRoith D. (1996). Adv. Cancer Res., 68, 183 ± 223. Werner H and LeRoith D. (1997). Crit. Rev. Oncogenesis, 8, 71 ± 92. Werner H, Rauscher III FJ, Sukhatme VP, Drummond IA, Roberts Jr CT and LeRoith D. (1994b). J. Biol. Chem., 269, 12577 ± 12582. Werner H, Re GG, Drummond IA, Sukhatme VP, Rauscher III FJ, Sens DA, Garvin AJ, LeRoith D and Roberts Jr CT. (1993). Proc. Natl. Acad. Sci. USA, 90, 5828 ± 5832. Werner H, Roberts Jr CT, Rauscher III FJ and LeRoith D. (1996b). J. Mol. Neurosci., 7, 111 ± 123. Werner H, Shen-Orr Z, Rauscher III FJ, Morris JF, Roberts Jr CT and LeRoith D. (1995). Mol. Cell. Biol., 15, 3516 ± 3522. Zucman J, Delattre O, Desmaze C, Epstein AL, Stenman G, Speleman F, Fletchers CD, Aurias A and Thomas G. (1993a). Nature Genet., 4, 341 ± 345. Zucman J, Delattre O, Desmaze C, Plougastel B, Joubert I, Melot T, Peter M, De Jong P, Rouleau G, Aurias A and Thomas G. (1992). Genes Chromosomes Cancer, 5, 271 ± 277. Zucman J, Melot T, Desmaze C, Ghysdael J, Plougastel B, Peter M, Zucker JM, Triche TJ, Sheer D and Turc-Carel C. (1993b). EMBO J., 12, 4481 ± 4487.