Oncogene
SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Archive
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Subscribe
Advertising
naturereprints
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
 
21 March 2002, Volume 21, Number 13, Pages 2009-2019
Table of contents    Previous  Article  Next   [PDF]
Original Paper
Induction of the interleukin-2/15 receptor bold beta-chain by the EWS-WT1 translocation product
Jenise C Wong1, Sean B Lee1, Moshe D Bell2, Paul A Reynolds1, Emilio Fiore4, Ivan Stamenkovic4, Vivi Truong5, Jonathan D Oliner5,a, William L Gerald3 and Daniel A Haber1

1Massachusetts General Hospital Cancer Center and Harvard Medical School, Charlestown, Massachusetts, MA 02129, USA

2Department of Pediatrics, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA

3Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA

4Molecular Pathology Unit, Massachusetts General Hospital, Charlestown, Massachusetts, MA 02129, USA

5Affymetrix, 3380 Central Expressway, Santa Clara, California, CA 95051, USA

Correspondence to: D Haber, Laboratory of Molecular Genetics, Massachusetts General Hospital Cancer Center CNY 7, Building 149, 13th Street, Charlestown, Massachusetts 02129, USA; E-mail: Haber@helix.mgh.harvard.edu

aCurrent address: Amgen, One Amgen Center Drive, MS 5-2-A, Thousand Oaks, California, CA 91320, USA

Abstract

EWS-WT1 is a chimeric transcription factor resulting from fusion of the N-terminal domain of the Ewing sarcoma gene EWS to the three C-terminal zinc fingers of the Wilms tumor suppressor WT1. This translocation underlies desmoplastic small round cell tumor (DSRCT), which is noted for the abundance of reactive stroma surrounding islets of tumor cells, suggestive of paracrine signals contributing to tumor cell proliferation. Hybridization to high-density oligonucleotide microarrays can be used to identify targets of EWS-WT1. Expression of EWS-WT1 from a tetracycline-regulated promoter leads to the induction of growth-associated genes, of which the most remarkable is the beta-chain of the interleukin-2/15 receptor (IL-2/15Rbeta). Potent transcriptional activation by the chimeric protein maps to two bindings sites within the IL-2/15Rbeta promoter. Analysis of primary DSRCT tumor specimens demonstrates high levels of IL-2/15Rbeta within the tumor cells, along with expression of IL-2 and IL-15 by the abundant hyperplastic endothelial cells within the reactive stroma. Activation of this cytokine signaling pathway is consistent with the nuclear localization of its downstream effectors, phosphorylated STAT3 and STAT5. These observations suggest that the transcriptional induction of a cytokine receptor by a tumor-associated translocation product enables a proliferative response of epithelial cancer cells to ligands secreted by the surrounding stroma.

Oncogene (2002) 21, 2009-2019 DOI: 10.1038/sj/onc/1205262

Keywords

EWS-WT1; desmoplastic small round cell tumor; Wilms tumor; interleukin-2/15 receptor; microarrays; tumor-stromal interaction

Introduction

Chromosomal translocations resulting in the generation of novel chimeric proteins underlie specific cancers, where they appear to constitute the initial genetic event triggering malignant transformation (Rabbitts, 1994). The family of primitive tumors defined by translocations involving the EWS gene is remarkable for the specificity with which different translocation partners are correlated with distinct tumor types (de Alava and Gerald, 2000). EWS encodes a putative RNA binding protein of unknown function, with an N-terminal domain (NTD) that mediates potent transcriptional activation when fused to heterologous DNA binding domains (May et al., 1993), presumably activating a limited number of target genes that are sufficient to trigger malignant proliferation. In the prototype tumor, Ewing sarcoma, manic fringe and the transforming growth factor-beta type II receptor have been proposed as targets of EWS-FLI1 (May et al., 1997; Hahm et al., 1999), but a comprehensive screen using expression profiling has not been reported.

Desmoplastic small round cell tumor (DSRCT) is a highly aggressive primitive tumor arising from the serosal surfaces of the abdominal peritoneum and primarily affecting young males (Gerald et al., 1991). It is characterized histologically by solid nests of small neoplastic cells expressing epithelial, muscle, and neural markers, surrounded by a dense reactive stroma, consisting of fibroblasts and hyperplastic blood vessels (hence the description of the tumor as 'desmoplastic'). DSRCT may thus represent an extreme example of the paracrine interactions between tumor and stromal cells that are thought to play an important role in the genesis of epithelial cancers (Wernert, 1997). All cases of DSRCT analysed to date carry a t(11;22)(p13;q12) chromosomal translocation fusing the NTD of EWS to zinc fingers 2-4 of the Wilms tumor suppressor, WT1 (Figure 1a; Ladanyi and Gerald, 1994; Gerald et al., 1995). WT1 was initially identified based on its inactivation in the pediatric kidney cancer Wilms tumor (Lee and Haber, 2001). Its essential role in normal differentiation of the kidney, gonads, and mesothelial structures has been defined in WT1-null mice (Kreidberg et al., 1993). A gain-of-function phenotype for EWS-WT1 is suggested by its coexpression with wild-type WT1 in primary DSRCT specimens and by the transformed properties of NIH3T3 cells expressing the chimeric product (Ladanyi and Gerald, 1994; Kim et al., 1998b).

Initial studies suggested that EWS-WT1 might encode a transcriptional activator of target promoters normally repressed by the wild-type WT1 product (Ladanyi and Gerald, 1994; Karnieli et al., 1996). However, more recent studies have indicated that WT1 itself may function as a transcriptional activator, and that endogenous target genes regulated by these two proteins are likely to differ (Reddy and Licht, 1996; Lee et al., 1999; Mayo et al., 1999). The EWS-WT1 chimera invariably loses the first of the four zinc fingers of WT1, suggesting that an altered DNA binding domain may be essential for cellular transformation in DSRCT. In addition, the distinct transactivation domain conferred by the NTD of EWS is likely to lead to interactions with other components of the transcriptional apparatus that may be recruited to target promoters, thereby contributing to different targets for these related transcription factors (Bertolotti et al., 1998; Petermann et al., 1998). We have previously demonstrated that the platelet-derived growth factor-A (PDGFA) gene, encoding a potent chemoattractant for fibroblasts and endothelial cells, is directly induced by EWS-WT1, but not by the wild-type WT1 product (Lee et al., 1997). To perform an unbiased search for potential physiological targets of EWS-WT1, we screened high-density oligonucleotide microarrays for cellular transcripts whose levels were altered following induction of this chimeric product. We report here that the IL-2/15Rbeta gene is a direct transcriptional target of EWS-WT1, and that its aberrant expression in tumor cells of DSRCT is correlated with activation of the downstream transcription factors STAT3 and STAT5. IL-2 and IL-15, both ligands of the IL-2/15Rbeta-chain, are produced by endothelial cells within the reactive stroma in DSRCT. These observations suggest that IL-2/15 receptor signaling may contribute to the tumor-stromal interactions that are characteristic of DSRCT, and suggest implications for the role of this cytokine-dependent proliferation pathway in other cancers.

Results

Induction of IL-2/15Rbeta by EWS-WT1

To identify endogenous transcripts induced by EWS-WT1, we used U2OS osteosarcoma cells in which expression of the chimera is driven by a tetracycline-repressible promoter (Lee et al., 1997). These cells express the (-KTS) isoform of EWS-WT1, which mediates potent transcriptional activation in promoter-reporter assays. This is in contrast to the alternative splice form with three additional amino acids (lysine-threonine-serine or KTS) between zinc fingers 3 and 4 of WT1 that abrogates DNA binding and transcriptional activation of known reporter constructs (Rauscher et al., 1990). Poly(A)+ RNA was isolated from subconfluent cells grown either in the presence of tetracycline or for 12 h following drug withdrawal, and used to interrogate oligonucleotide arrays representing 6800 known genes and expressed sequence tags (ESTs) (Lockhart et al., 1996). Of 16 known genes demonstrating greater than twofold induction by microarray analysis, five were confirmed to be induced by Northern blotting (Table 1). Tenfold induction was observed for IL-2/15Rbeta, the cytokine receptor chain shared by the heterotrimeric receptors for IL-2 and IL-15 (Figure 1b) and for PDGFA, a previously identified EWS-WT1 target gene (Lee et al., 1997). Lower level induction was observed for adrenomedullin, encoding a vasoactive peptide with postulated autocrine growth properties in cancer-derived cell lines (Miller et al., 1996), and the fibroblast growth factor receptor 4 gene. Of note, amphiregulin, the predominant target identified using microarrays for wild-type WT1(-KTS) (Lee et al., 1999) was only moderately induced by EWS-WT1(-KTS). Consistent with the distinct endogenous targets recognized by these related transcription factors, IL-2/15Rbeta mRNA was not induced following expression of wild-type WT1(-KTS) (Figure 1b). In addition, cells with inducible expression of the alternative isoform EWS-WT1(+KTS) did not demonstrate induction of IL-2/15Rbeta, confirming that this effect is dependent upon an intact DNA binding domain (data not shown). Given the magnitude of the IL-2/15Rbeta induction by EWS-WT1(-KTS) and the potential biological significance of this target gene, we searched for evidence of its expression in primary DSRCT tumor specimens.

Expression of IL-2/15Rbeta in DSRCT

We first demonstrated IL-2/15Rbeta mRNA expression by Northern blotting in DSRCT specimens. Expression of the predicted 4 kb transcript was abundant in these tumors, containing the EWS-WT1 chimeric transcript (Figure 1c). No IL-2/15Rbeta mRNA expression was detected in examples of Wilms tumors expressing wild-type WT1, or Ewing sarcoma, containing the EWS-FLI1 chimeric fusion. To confirm the cell type within DSRCT that expressed IL-2/15Rbeta, we undertook immunohistochemical analysis of primary tumor specimens. Expression of IL-2/15Rbeta was observed primarily in the islets of tumor cells within all DSRCT specimens tested (representative staining in Figure 2b; quantitation and results for multiple tumors in Table 2). Rare infiltrating lymphoid aggregates within the stroma also stained for IL-2/15Rbeta, but the reactive fibroblasts and endothelial cells that comprise the desmoplastic component of the tumor were negative or weakly reactive. DSRCT tumor cells in all specimens also expressed IL-2/15Rgammac, the second receptor chain shared by heterotrimeric receptors for IL-2 and IL-15, and IL-2Ralpha and IL-15Ralpha, the two chains that confer ligand specificity (Minami et al., 1993; Giri et al., 1995; Anderson et al., 1995) (Figure 2c-e; Table 2). The presence of these components on the cell surface makes it likely that at least a fraction of IL-2/15Rbeta associates with the other subunits within a heterotrimeric receptor complex.

Transcriptional activation of the IL-2/15Rbeta promoter by EWS-WT1

The strong induction of the IL-2/15Rbeta transcript following inducible expression of EWS-WT1(-KTS) suggested that this cytokine receptor chain may be a direct transcriptional target of the chimera. During lymphocyte activation, induction of IL-2Ralpha or IL-15Ralpha is thought to be the primary regulated component of the IL-2 or IL-15 receptor complexes, although induction of beta-chain expression is also noted following antigen stimulation (Minami et al., 1993; Waldmann and Tagaya, 1999; Fehniger and Caligiuri, 2001). The IL-2/15Rbeta promoter contains three positive regulatory regions, consisting of phorbol myristate acetate (PMA)-inducible enhancer-like regions, and binding sites for known transcription factors, including the Ets family proteins Ets-1 and GA-binding protein (GABP), Sp1, and early growth response protein 1 (EGR1) (Figure 3a) (Lin et al., 1993; Lin and Leonard, 1997). Insertion of the full-length promoter fragment (nucleotides -393 to +44) into the promoterless luciferase reporter plasmid pGL3, led to an average 20-fold activation following cotransfection with cytomegalovirus (CMV) promoter-driven EWS-WT1(-KTS), into NIH3T3 cells (Figure 3b). Minimal activation was observed following cotransfection with wild-type WT1(-KTS) (1.5-fold) or EWS-WT1(+KTS) (3.6-fold).

Electrophoretic mobility shift assays (EMSA) were used to define the specific DNA recognition site for EWS-WT1(-KTS), using overlapping, radiolabeled double-stranded probes spanning the IL-2/15Rbeta promoter fragment, and the bacterially-synthesized DNA binding domain of the chimera (zinc fingers 2-4). Two binding sites were identified: the first between nucleotides -300 and -271 (E-WRE1) and the second between nucleotides -169 and -140 (E-WRE2) (Figure 3c). The relative intensity of the shifted complexes suggests that the zinc finger domain binds site E-WRE1 with higher affinity in vitro than it does E-WRE2. As expected, binding to both probes was abrogated by insertion of the alternatively spliced KTS sequence between zinc fingers 3 and 4 (Figure 3d). Of note, the wild-type WT1(-KTS) binding domain (zinc fingers 1-4) bound to E-WRE1, but only weakly to fragment E-WRE2 (Figure 3c). In vitro, WT1(-KTS) is known to bind to both GC-rich and TC-rich repeats, but a high affinity site identified using a PCR-based selection assay, WTE: 5'-GCGTGGGAGT-3', is closely related to that present in the promoter of the WT1-target gene amphiregulin (Lee et al., 1999; Nakagama et al., 1995). The first zinc finger of WT1 (which is lacking in EWS-WT1) is thought to bind the most 3' nucleotide of this 10-nucleotide sequence, with each of the remaining zinc fingers binding to three nucleotides in a 3' to 5' orientation. Competition experiments using unlabeled WTE were used to show the specificity of EWS-WT1(-KTS) binding to both E-WRE1 and E-WRE2 (Figure 3d). We therefore sought to define the EWS-WT1(-KTS) binding sequences within these two responsive sites of the IL-2/15Rbeta promoter.

Inspection of the E-WRE1 and E-WRE2 sequences identified one potential binding site in each promoter fragment that matched the WTE sequence. Extensive mutagenesis of these and flanking nucleotides was used to identify the precise EWS-WT1(-KTS) recognition site. A 9-nucleotide sequence within E-WRE1, 5'-GCGTGGGGG-3' (nucleotides -290 to -282), was found to be essential for EWS-WT1(-KTS) binding (Figure 4a), as was a comparable element within E-WRE2, (in the antisense orientation) 5'-CGCTGGGG-3' (nucleotides -161 to -153) (data not shown). Comparison of these two sequences with the WT1(-KTS) recognition motif WTE showed the E-WRE1 element to differ in only one position (G8), while the E-WRE2 element differs at four positions (C1, G2, C3, and G8) (Figure 4b). Both sites are located within the previously identified enhancer-like regions, with the E-WRE2 site overlapping the Sp1 and EGR1 binding sites. Abrogation of binding by the zinc finger domain was achieved by substituting both G3 and G5 to T within E-WRE1, and by replacing C1, C3, and G5 with T within E-WRE2 (Figure 4c). These disrupted DNA binding sites were introduced, either alone or in combination, into the full-length IL2-15Rbeta promoter reporter (Figure 4d). Cotransfection of these reporters along with CMV-driven EWS-WT1(-KTS) into NIH3T3 cells demonstrated that both E-WRE1 and E-WRE2 contribute to activation of the promoter by the chimera (Figure 4e). Disruption of binding to E-WRE1 reduced EWS-WT1(-KTS)-mediated activation of the IL-2/15Rbeta promoter by twofold. Disruption of binding to E-WRE2 decreased EWS-WT1-dependent transactivation to minimal levels, comparable to those present when both E-WRE1 and E-WRE2 are mutated. Based on these results and the mutational analysis described above, we propose an EWS-WT1(-KTS) binding consensus sequence: 5'-(G/C)(C/G)(G/C)TGGGGG-3'.

Evidence for IL-2/15 receptor signaling in DSRCT

The direct induction of IL-2/15Rbeta mRNA expression by EWS-WT1 suggests that this cytokine signaling pathway may contribute to tumorigenesis in DSRCT. Together with the other receptor components that appear to be expressed constitutively in DSRCT, IL-2/15Rbeta forms high affinity receptors for both IL-2 and IL-15. We therefore analysed primary DSRCT specimens for expression of these ligands by immunohistochemistry. Eleven of 14 tumor specimens demonstrated expression of IL-2 within reactive stromal cells, especially the perithelial and endothelial cells of blood vessels (representative staining in Figure 5a,b; quantitation and results for multiple tumors in Table 2). IL-2 expression was not detectable within DSRCT tumor cells themselves. High levels were clearly evident within lymphoid aggregates present in some tumor samples, serving as an internal control. IL-15, unlike IL-2, is known to be expressed at low levels in multiple epithelial tissues (Grabstein et al., 1994), raising the possibility that it may be the primary contributor to signaling through IL-2/15Rbeta in DSRCT. Indeed, IL-15 was abundantly expressed in vascular cells of the stroma in nearly all specimens tested (Figure 5c; Table 2). Thus, immunohistochemical studies of primary DSRCT specimens suggest that both IL-2 and IL-15 are secreted by reactive stromal cells in DSRCT, potentially inducing a proliferative signal through elements of the IL-2 and IL-15 receptors expressed within the adjacent primary tumor cells.

The signaling pathways activated by the IL-2 and IL-15 receptor have been defined primarily in T lymphocytes, where ligand binding triggers receptor dimerization, cross-phosphorylation of the Janus kinases Jak1 and Jak3 (bound to the beta- and gammac-chains, respectively), and the phosphorylation and nuclear translocation of the STAT3 and STAT5 transcription factors (Taniguchi, 1995; Fehniger and Caligiuri, 2001). Nuclear localization of phosphorylated STAT3 and STAT5 is therefore indicative of activation of this cytokine signaling pathway. Primary DSRCT specimens were stained for Jak1, Jak3, STAT5, and STAT3, all of which were abundantly expressed within tumor cells (Figure 6a-c,e; Table 2). Remarkably, staining with an antibody that specifically recognizes the phosphorylated tyrosine 694 residue in activated STAT5 demonstrated strong, exclusively nuclear localization within the tumor cells of DSRCT (Figure 6d; Table 2). Similar observations were made using antibody against phosphorylated tyrosine 705 in STAT3 (Figure 6f; Table 2). These observations are consistent with activation of cytokine-dependent signaling within primary tumor cells of DSRCT.

Discussion

The identification of target genes activated by EWS-WT1 in DSRCT is essential to understanding the initiation of tumorigenesis by this chimeric transcription factor. Here we have used inducible expression constructs, together with oligonucleotide microarray-based expression profile analysis, to identify IL-2/15Rbeta as a physiological target of EWS-WT1. The IL-2/15Rbeta promoter is directly regulated by EWS-WT1(-KTS), and primary DSRCT specimens demonstrate high levels of IL-2/15Rbeta expression, as well as activation of the downstream Jak/STAT signaling pathway. The secretion of IL-2 and IL-15 by the stroma, surrounding nests of tumor cells, suggests a paracrine interaction between these characteristic components of DSRCT. These observations also provide insight into other cancers reported to express components of this cytokine-dependent proliferation pathway.

Activation of IL-2/15Rbeta by EWS-WT1

Studies of proteins encoded by both wild-type WT1 and the EWS-WT1 chimeric fusion have underscored the poor correlation between promoters containing responsive elements in reporter assays and regulation of native genes by these transcription factors (Reddy and Licht, 1996). The use of expression profiling following inducible expression of these genes allows unbiased screening of a large fraction of cellular transcripts in vivo. Oligonucleotide microarrays appear well suited for this approach, although duplicate experiments and validation using Northern blotting are essential to distinguish the rare genuine target genes from background variation. As demonstrated for IL-2/15Rbeta, identification of an activated target gene using a heterologous cell type can be correlated in vivo with primary tumor specimens. Furthermore, analysis of the target promoter elements required for in vivo activation provides insight into the constraints that may specify physiological target genes. The native IL-2/15Rbeta transcript is dramatically induced by EWS-WT1(-KTS), but not by WT1(-KTS), a difference that may result from their distinct DNA binding specificity, as well as from protein interactions mediated by the EWS-derived transactivation domain of the chimera. The chimeric DNA binding domain differs from that of WT1 by the absence of one zinc finger, which mediates binding to only one essential nucleotide. In vitro, both DNA binding domains bind to the E-WRE1 element within the IL-2/15Rbeta promoter, which is closely related to the optimal WT1(-KTS) binding consensus (Nakagama et al., 1995; Kim et al., 1998a). EWS-WT1(-KTS) binds with less affinity to the more divergent E-WRE2 sequence, and WT1(-KTS) binds extremely weakly to this element. Remarkably, E-WRE2 is critical for EWS-WT1(-KTS)-mediated transcriptional activation of the IL-2/15Rbeta promoter, an effect that may result from its context within the IL-2/15Rbeta promoter, where it overlaps binding sites for Sp1 and EGR1, and is in close proximity to the Ets site. Further studies will be required to determine whether these transcriptional regulators themselves modulate expression of the native IL-2/15Rbeta gene, and whether they interact in vivo with EWS-WT1.

Potential roles of IL-2 and IL-15 receptors in DSRCT and in epithelial cancers

The functional consequences of IL-2 and IL-15 receptor activation have been studied extensively in lymphocytes, where signaling is involved in cellular activation and proliferation (Taniguchi, 1995; Fehniger and Caligiuri, 2001). In addition, the IL-15 receptor is also expressed in multiple epithelial cell types, and may play a broader role in cellular proliferation in addition to its role in immune cell growth and development (Fehniger and Caligiuri, 2001). IL-2/15Rbeta, induced by EWS-WT1(-KTS), contributes to the formation of both receptors, and both private alpha-chains are detectable in DSRCT tumors, preventing us from distinguishing between the potential roles of these two cytokines in mediating proliferation signals in DSRCT. Expression of IL-2/15Rbeta has been reported in non-lymphoid cells, including normal human fibroblasts, keratinocytes, and intestinal epithelial cells, as well as cell lines derived from a variety of cancers, such as sarcomas, melanomas, and squamous cell carcinomas (Azzarone et al., 1996). The ligand-specific IL-15Ralpha-chain is also expressed at significant levels in normal T cells, brain, intestine, heart, lung, muscle, and liver (Giri et al., 1995; Anderson et al., 1995; Dubois et al., 1999) as well as in leukemias, lymphomas, myelomas, melanomas, and rhabdomyosarcomas (Doucet et al., 1997; Barzegar et al., 1998; De Giovanni et al., 1998; Tinhofer et al., 2000). The functional consequences of IL-2 and IL-15 receptor expression in non-lymphoid cancers have been difficult to interpret, and appear to depend on both cellular context and cytokine dose (Lin et al., 1995; Suminami et al., 1998; Yamada et al., 1998; Döbbeling et al., 1998; Hjorth-Hansen et al., 1999). A role for STAT activation in epithelial cancers has also been proposed, based on in vitro transformation assays and in vivo expression patterns, but the underlying mechanism has not been precisely defined (Frank, 1999; Bromberg et al., 1999). In addition, expression of SOCS-1, an inhibitor of the Jak/STAT signaling pathway, suppressed growth of hepatocellular carcinoma (HCC) cell lines, and is aberrantly methylated and inactivated in primary HCC samples, providing another link between constitutive activation of Jak/STAT pathway and tumorigenesis in nonhematopoetic cells (Yoshikawa et al., 2001). The absence of any DSRCT cell lines prevents direct testing of the IL-2/15 receptor signaling pathway in this tumor. However, our observations that IL-2/15Rbeta is a direct transcriptional target of the oncogenic EWS-WT1 chimera, support a role for this cytokine signaling pathway in DSRCT.

A potential role for IL-2/15 receptor signaling in DSRCT is consistent with other models of tumor-stromal interactions, in which the secretion by stromal cells of growth factors that enhance tumor cell proliferation is thought to play an important role in epithelial cancers. Expression of the matrix metalloproteinase MMP-9 by stromal cells is thought to release growth factors that allow proliferation of transformed keratinocytes (Coussens et al., 2000). Expression of the chemokine receptors CXCR4 and CXCR7 in breast cancer and malignant melanoma cell lines has been linked to their metastatic potential toward ligand-expressing tissues (Müller et al., 2001). In DSRCT, PDGF-A, the other major transcriptional target of EWS-WT1(-KTS), is a chemoattractant for fibroblasts and endothelial cells that may contribute to the prominent recruitment of stromal cells in this tumor (Lee et al., 1997). IL-2/15Rbeta-dependent signaling may constitute one of the pathways by which stromal cells support the proliferation of DSRCT tumor cells. The discovery of additional EWS-WT1-induced growth factors may allow the culture of DSRCT tumor cells in vitro and the functional characterization of these proliferation pathways. Thus, by virtue of its defined genetic etiology an exaggerated desmoplastic reaction, DSRCT provides a unique model to study the proliferation pathways derived from the interaction between malignant and stromal components of epithelial cancers.

Materials and methods

Inducible EWS-WT1 expression and hybridization to oligonucleotide microarrays

U2OS cells with tetracycline-repressible expression of EWS-WT1(-KTS), WT1(-KTS), and WT1(+KTS) were maintained in tetracycline (1 mug/ml) as described previously (Lee et al., 1997). For expression profile analysis, tetracycline was withdrawn for 12 h, and poly(A)+-selected RNA was isolated, amplified, labeled, and hybridized to oligonucleotide microarrays as described (Lockhart et al., 1996). Criteria used for quantitative analysis of array hybridization included a >twofold change in signal intensity and reproducibility in two experiments, as described previously (Lee et al., 1999). Predicted target genes identified by microarray hybridization were confirmed by Northern blotting. Frozen tumor specimens of DSRCT, Wilms tumor, and Ewing sarcoma were obtained from the Memorial Sloan-Kettering tumor bank.

IL-2/15Rbeta promoter reporter assays

The IL-2/15Rbeta promoter sequences were PCR-amplified from genomic DNA and cloned into the promoterless luciferase vector pGL3 (Promega), from which deletion constructs were generated. Missense mutations were introduced by PCR to change the guanines at positions 3 and 5 of E-WRE1 to thymines (mE-WRE1), and the cytosines and guanine at positions 1, 3, and 5 of E-WRE2 to thymines (mE-WRE2). All constructs were verified by sequencing. Constructs encoding full-length EWS-WT1(-KTS), EWS-WT1(+KTS), and WT1(-KTS) were cloned into a vector downstream of the CMV promoter (pCDNA3). For luciferase reporter assays, 10 mug of CMV-driven expression constructs or empty vector were cotransfected with 2 mug of the promoter reporter into NIH3T3 cells using calcium phosphate DNA precipitation. Equal amounts of CMV-driven constructs were transfected in each experiment, and cotransfection of 1 mug of a human growth hormone reporter was used to allow standardization for transfection efficiency. All experiments were performed in triplicate.

Electrophoretic mobility shift assays

The DNA binding zinc finger domains of the EWS-WT1(-KTS), EWS-WT1(+KTS) and WT1(-KTS) proteins were cloned into the GST-expression vector pGEX3X (Pharmacia). The isolated zinc finger domains were used due to the insolubility of the full-length proteins. GST-proteins were expressed in Escherichia coli strain DH5alpha and purified with glutathione sepharose 4B beads (Pharmacia). To identify the EWS-WT1(-KTS) binding site(s), complementary oligonucleotides spanning the responsive fragment of the IL-2/15Rbeta promoter were annealed, end-labeled and incubated with 100 ng GST-protein in binding buffer for 20 min at 4°C and electrophoresed as described (Lee et al., 1999). For competition experiments, 300-fold molar excess of unlabeled competitor was incubated with 50 ng of GST-protein in binding buffer for 20 min at 4°C, after which labeled probe was added and allowed to incubate for 20 min at 4°C.

Immunohistochemistry

Immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissue sections using an avidin-biotin-peroxidase method as described (Hsu et al., 1981), with minor modifications. Antigen retrieval (Pileri et al., 1997) was achieved by microwave-heating for 5 to 8 min in 0.01 M citric acid buffer, pH 6. Sections were counterstained with hematoxylin. For all antibodies, the same panel of 16 DSRCT and seven related tumors (Wilms tumor, Ewing sarcoma, and undifferentiated sarcomas) was used, as were control tissues known to be immunoreactive and non-reactive for each antibody. Negative controls substituted normal rabbit serum for the primary antibody. Except where noted, all primary antibodies used were rabbit polyclonal IgG directed against human peptides of the named antigen, and were applied for overnight incubation at 4°C. Antibodies included: anti-IL-2 (H133), diluted 1 : 200; IL-2Ralpha (N19), 1 : 400; IL-2/15Rbeta (N20, 1 : 100; IL-2/15Rgammac (N20), 1 : 200; IL-15 (H114), 1 : 50; IL-15Ralpha (H107), 1 : 50; Jak1 (Q19), 1 : 200; Jak3 (C21), 1 : 200; mouse monoclonal anti-STAT5 IgG1 (G2), 1 : 100 (Santa Cruz Biotechnology, Santa Cruz, CA, USA); phospho-STAT5 (tyr 694), 1 : 20 (Zymed Laboratories, South San Francisco, CA, USA); STAT3, 1 : 100; phospho-STAT3 (tyr 705), 1 : 50 for 40 h at 4°C; (New England Biolabs, Beverly, MA, USA).

Acknowledgements

We thank Shiv Pillai and members of the Haber laboratory for helpful discussions and critical review of the manuscript, and Muzaffar Akram for technical assistance. This work was supported by National Institutes of Health grant CA58596 (DA Haber), CA68273 (WL Gerald) and the MGH Medical Discovery Fund (SB Lee).

References

Anderson DM, Kumaki S, Ahdieh M, Bertles J, Tometsko M, Loomis A, Giri J, Copeland NG, Gilbert D, Jenkins NA, Valentine V, Shapiro DN, Morris SW, Park LS, Cosman D. (1995). J. Biol. Chem., 270: 29862-29869. MEDLINE

Azzarone B, Pottin-Clemenceau C, Krief P, Rubinstein E, Jasmin C, Scudeletti M, Indiveri F. (1996). Eur. Cytokine Netw., 7: 27-36. MEDLINE

Barzegar C, Meazza R, Pereno R, Pottin-Clemenceau C, Scudeletti M, Brouty-Boyé D, Doucet C, Taoufik Y, Ritz J, Musselli C, Mishal Z, Jasmin C, Indiveri F, Ferrini S, Azzarone B. (1998). Oncogene, 16: 2503-2512. MEDLINE

Bertolotti A, Melot T, Acker J, Vigneron M, Delattre O, Tora L. (1998). Mol. Cell. Biol., 18: 1489-1497. MEDLINE

Bromberg JF, Wrzeszczynska MH, Devgan G, Zhao Y, Pestell RG, Albanese C, Darnell JrJE. (1999). Cell, 98: 295-303. MEDLINE

Coussens LM, Tinkle CL, Hanahan D, Werb Z. (2000). Cell, 103: 481-490. MEDLINE

de Alava E, Gerald WL. (2000). J. Clin. Oncol., 18: 204-213. MEDLINE

De Giovanni C, Nanni P, Sacchi A, Soddu S, Manni I, D'Orazi G, Bulfone-Paus S, Pohl T, Landuzzi L, Nicoletti G, Frabetti F, Rossi I, Lollini P-L. (1998). Br. J. Cancer, 78: 1541-1546. MEDLINE

Döbbeling U, Dummer R, Laine E, Potoczna N, Qin J-Z, Burg G. (1998). Blood, 92: 252-258. MEDLINE

Doucet C, Meazza R, Pottin-Clemenceau C, Scudeletti L, Brouty-Boye D, Ferrini S, Alileche A, Taoufik Y, Jasmin C, Azzarone B, Indiveri F. (1997). Melanoma Res., 7: suppl 2 S7-S17. MEDLINE

Dubois S, Magrangeas F, Lehours P, Raher S, Bernard J, Boisteau O, Leroy S, Minvielle S, Godard A, Jacques Y. (1999). J. Biol. Chem., 274: 26978-26984. MEDLINE

Fehniger TA, Caligiuri MA. (2001). Blood, 97: 14-32. Article MEDLINE

Frank DA. (1999). Mol. Med., 5: 432-456. MEDLINE

Gerald WL, Miller HK, Battifora H, Miettinen M, Silva EG, Rosai J. (1991). Am. J. Surg. Pathol., 15: 499-513. MEDLINE

Gerald WL, Rosai J, Ladanyi M. (1995). Proc. Natl. Acad. Sci. USA, 92: 1028-1032. MEDLINE

Giri JG, Kumaki S, Ahdieh M, Friend DJ, Loomis A, Shanebeck K, DuBose R, Cosman D, Park LS, Anderson DM. (1995). EMBO J., 14: 3654-3663. MEDLINE

Grabstein KH, Eisenman J, Shanebeck K, Rauch C, Srinivasan S, Fung V, Beers C, Richardson J, Schoenborn MA, Ahdieh M, Johnson L, Alderson MR, Watson JD, Anderson DM, Giri JG. (1994). Science, 264: 965-968. MEDLINE

Hahm K-B, Cho K, Lee C, Im Y-H, Chang J, Choi S-G, Sorensen PHB, Thiele CJ, Kim S-J. (1999). Nat. Genet., 23: 222-227. Article MEDLINE

Hjorth-Hansen H, Waage A, Börset M. (1999). Br. J. Haematol., 106: 28-34. MEDLINE

Hsu S-M, Raine L, Fanger H. (1981). J. Histochem. Cytochem., 29: 577-580. MEDLINE

Karnieli E, Werner H, Rauscher IIIFJ, Benjamin L, LeRoith D. (1996). J. Biol. Chem., 271: 19304-19309. MEDLINE

Kreidberg J, Sariola H, Loring JM, Maeda M, Pelletier J, Housman D, Jaenisch R. (1993). Cell, 74: 679-691. MEDLINE

Kim J, Lee K, Pelletier J. (1998a). Oncogene, 16: 1021-1030. MEDLINE

Kim J, Lee K, Pelletier J. (1998b). Oncogene, 16: 1973-1979. MEDLINE

Ladanyi M, Gerald W. (1994). Cancer Res., 54: 2837-2840. MEDLINE

Lee SB, Kolquist KA, Nichols K, Englert C, Maheswaran S, Ladanyi M, Gerald WL, Haber DA. (1997). Nat. Genet., 17: 309-313. MEDLINE

Lee SB, Haber DA. (2001). Exp. Cell Res., 264: 74-99. Article MEDLINE

Lee SB, Huang K, Palmer R, Truong VB, Herzlinger D, Kolquist KA, Wong J, Paulding C, Yoon SK, Gerald W, Oliner JD, Haber DA. (1999). Cell, 98: 663-673. MEDLINE

Lin J-X, Bhat NK, John S, Queale WS, Leonard WJ. (1993). Mol. Cell. Biol., 13: 6201-6210. MEDLINE

Lin J-X, Leonard WJ. (1997). Mol. Cell. Biol., 17: 3714-3722. MEDLINE

Lin W-C, Yasumura S, Suminami Y, Sung M-W, Nagashima S, Stanson J, Whiteside TL. (1995). J. Immunol., 155: 4805-4816. MEDLINE

Lockhart D, Dong H, Byrne MC, Follettie MT, Gallo MV, Chee MS, Mittman M, Wang C, Kobayashi M, Horton H, Brown EL. (1996). Nat. Biotechnol., 14: 1675-1680. MEDLINE

May WA, Arvand A, Thompson AD, Braun BS, Wright M, Denny CT. (1997). Nat. Genet., 17: 495-497. MEDLINE

May WA, Lessnick SL, Braun BS, Klemsz M, Lewis BC, Lunsford LB, Hromas R, Denny CT. (1993). Mol. Cell. Biol., 13: 7393-7398. MEDLINE

Mayo MW, Wang CY, Drouin SS, Madrid LV, Marshall AF, Reed JC, Weissman BE, Baldwin AS. (1999). EMBO J., 18: 3990-4003. MEDLINE

Miller MJ, Martínez A, Unsworth EJ, Thiele CJ, Moody TW, Elsasser T, Cuttitta F. (1996). J. Biol. Chem., 271: 23345-23351. MEDLINE

Minami Y, Kono T, Miyazaki T, Taniguchi T. (1993). Annu. Rev. Immunol., 11: 245-268. MEDLINE

Müller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W, Wagner SN, Barrera JL, Mohar A, Verástegui E, Zlotnik A. (2001). Nature, 410: 50-56. Article MEDLINE

Nakagama H, Heinrich G, Pelletier J, Housman D. (1995). Mol. Cell. Biol., 15: 1489-1498. MEDLINE

Petermann R, Mossier BM, Aryee DNT, Khazak V, Golemis EA, Kovar H. (1998). Oncogene, 17: 603-610. MEDLINE

Pileri SA, Roncador G, Ceccarelli C, Piccioli M, Briskomatis A, Sabattini E, Ascani S, Santini D, Piccaluga PP, Leone O, Damiani S, Ercolessi C, Sandri F, Pieri F, Leoncini L, Falini B. (1997). J. Pathol., 183: 116-123. Article MEDLINE

Rabbitts TH. (1994). Nature, 372: 143-149. MEDLINE

Rauscher IIIFJ, Morris JF, Tournay OE, Cook DM, Curran T. (1990). Science, 250: 1259-1262. MEDLINE

Reddy J, Licht J. (1996). Biochim. Biophys. Acta., 1287: 1-28. MEDLINE

Suminami Y, Kashii Y, Law JC, Lin W-C, Stanson J, Reichert TE, Rabinowich H, Whiteside TL. (1998). Oncogene, 16: 1309-1317. MEDLINE

Taniguchi T. (1995). Science, 268: 251-255. MEDLINE

Tinhofer I, Marschitz I, Henn T, Egle A, Greil R. (2000). Blood, 95: 610-618. MEDLINE

Waldmann TA, Tagaya T. (1999). Annu. Rev. Immunol., 17: 19-49. MEDLINE

Wernert N. (1997). Virchows Arch., 430: 433-443. Article MEDLINE

Yamada Y, Sugawara K, Hata T, Tsuruta K, Moriuchi R, Maeda T, Atogami S, Murata K, Fujimoto K, Kohno T, Tsukasaki K, Tomonaga M, Hirakata Y-I, Kamihira S. (1998). Blood, 91: 4265-4272. MEDLINE

Yoshikawa H, Matsubara K, Qian G-S, Jackson P, Groopman JD, Manning JE, Harris CC, Herman JG. (2001). Nat. Genet., 28: 29-35. Article MEDLINE

Figures

Figure 1 Induction of IL-2/15Rbeta by EWS-WT1(-KTS) and expression in DSRCT. (a) Schematic representation of EWS, WT1, and the EWS-WT1 fusion protein. The zinc finger domain of WT1 and of the chimera are indicated, along with the transactivation domain of WT1, the N-terminal domain (NTD) and putative RNA binding domains of EWS. KTS refers to an alternative splice that encodes three amino acids (lysine, threonine, and serine) inserted between zinc fingers 3 and 4 of WT1. (b) Northern blot of U2OS cells with tetracycline-repressible expression of EWS-WT1(-KTS), WT1(-KTS), WT1(+KTS), or vector, following growth in the presence or absence (11 h) of tetracycline. Blot was probed IL-2/15Rbeta, WT1 (recognizing both wild-type WT1 and the EWS-WT1 chimera), and GAPDH (loading control). (c) Northern blot of Wilms tumor Ewing sarcoma (lanes 1 and 2), and DSRCT (lanes 3 and 4) primary specimens. Variability in EWS-WT1 hybridization signal reflects both the relative tumor and stromal composition of the DSRCT specimens, as well as RNA integrity (GAPDH)

Figure 2 Localization of the IL-2/15 receptor in DSRCT. (a) Hematoxylin and eosin stained section of DSRCT demonstrating the characteristic histologic features of nests of neoplastic tumor cells (T) intimately associated with a fibroblast-rich, hypervascular, reactive stroma (S). (b) Immunohistochemical detection of IL-2/15Rbeta in tumor cells (brown reaction product) with subcellular localization primarily to cell borders (inset). Staining for (c) IL-2Ralpha, (d) IL-2/15Rgammac, and (e) IL-15Ralpha, also in tumor cells. (f) Negative control with normal rabbit sera substituted for the primary antibody

Figure 3 Direct transcriptional activation of the IL-2/15Rbeta promoter by EWS-WT1. (a) Schematic representation of the IL-2/15Rbeta promoter, including the relative positions of previously defined regulatory elements, and of reporter constructs A, B and C. Nucleotide numbers are derived from Lin and Leonard (1997). PMA-inducible elements (-363 to -251, -170 to -139, and -56 to -34) are represented by black bars; EBS (-41 to -38) denotes the Ets family protein binding site (Lin et al., 1993); Sp1 (-169 to -155) and EGR1 (-158 to -142) binding sites (Lin and Leonard, 1997) are noted. The EWS-WT1 responsive elements (E-WRE) are indicated by double-headed arrows. Construct A: nucleotides -393 to +44, B: nucleotides -240 to +44, and C: nucleotides -69 to +44 of the IL-2/15Rbeta promoter. (b) Activation of the IL-2/15Rbeta promoter by EWS-WT1. Luciferase activity (relative to vector) following cotransfection of NIH3T3 fibroblasts with reporter constructs A, B and C (2 mug) and CMV-driven expression plasmids (10 mug). See Materials and methods for experimental details. (c) EMSA analysis of IL-2/15Rbeta promoter fragments following incubation with the DNA binding domains of EWS-WT1(-KTS) and WT1(-KTS). Probes that span the two regions bound by EWS-WT1 are designated E-WRE1 (-300 to -271) and E-WRE2 (-169 to -140). Free probe, incubated in the absence of protein, is indicated (-). (d) Absence of binding of EWS-WT1(+KTS) to E-WRE1 and E-WRE2 and competition using unlabeled optimal WTE (300-fold molar excess). All lanes for each probe are derived from the same experiment and are approximated as shown for best comparison

Figure 4 Identification of the EWS-WT1(-KTS) binding sites within the IL-2/15Rbeta promoter. (a) Definition of the essential nucleotides for EWS-WT1(-KTS) binding within E-WRE1 by EMSA, using labeled wild-type E-WRE1 or probes in which each nucleotide within and flanking the potential binding site has been mutated. The binding site is shown from a 5' to 3' direction; the essential nucleotides for EWS-WT1(-KTS) binding are numbered from 1-9, and flanking nucleotides are labeled as either -1 or +1. All non-thymine residues were changed to thymines, and adenine was substituted for thymine residues. Equal amounts of probe and protein were used in all reactions. Probes incubated in the presence (P) or absence (-) of protein, are indicated. All lanes are derived from the same experiment and electrophoresed on parallel gels. (b) Alignment of the EWS-WT1(-KTS) binding motifs within E-WRE1 (sense) and E-WRE2 (antisense) with WTE, the optimal in vitro binding sequence for WT1(-KTS) (37). Brackets mark the three residues that are presumably bound by the WT1 zinc fingers 2, 3 and 4. (c) Absence of EWS-WT1(-KTS) binding to the substituted mutant E-WRE1 (mE-WRE1; Gright arrowT at positions 3 and 5) and to the mutant E-WRE2 (mEWRE2; Cright arrowT at position 1 and 3, and Gright arrowT at position 5). Equal amounts of protein and probe were used in all reactions, and all panels are derived from a single representative experiment. (d) Schematic representation of the wild-type IL-2/15Rbeta promoter construct (A), and constructs (D-F) containing either one or both mutant elements shown to abolish binding by EWS-WT1(-KTS). (e) Inactivation of EWS-WT1(-KTS) binding sites within the context of the full IL-2/15Rbeta promoter. Luciferase activity (relative to vector) following transfection of NIH3T3 cells with constructs encoding reporter constructs A, D, E and F (2 mug) and CMV-driven expression plasmids (10 mug). See Materials and methods for experimental details

Figure 5 Localization of IL-2 and IL-15 in DSRCT. (a) Strong immunohistochemical expression of IL-2 in fibroblasts and hyperplastic vessels (small arrow) of reactive stroma (S) and rare infiltrating lymphocytes (large arrowhead), but not in neoplastic cells of DSRCT (T). (b) Higher magnification. (c) Expression of IL-15 in hyperplastic vessels (arrow) of reactive stroma (S) but not in neoplastic cells of DSRCT (T)

Figure 6 Localization of JAK1, JAK3, STAT5, STAT3 and phosphorylated STAT5 and STAT3 in DSRCT. Immunohistochemical detection of (a) JAK1, (b) JAK3, (c) STAT5, and (e) STAT3 in the neoplastic cells of DSRCT and to a lesser extent in reactive stromal cells. Immunoreactivity for STAT5 and STAT3 was present in both cytoplasm and nuclei. Strong reactivity was detected for the phosphorylated forms of (d) STAT5 and (f) STAT3, limited to nuclei (insets: high power magnifications of nuclear localization for phosphorylated STAT5 and STAT3)

Tables

Table 1 EWS-WT1 target genes identified by microarray hybridization and confirmed by Northern blotting

Table 2 Expression of IL-2/15Rbeta and related molecules in DSRCT

Received 28 August 2001; revised 11 December 2001; accepted 18 December 2001
21 March 2002, Volume 21, Number 13, Pages 2009-2019
Table of contents    Previous  Article  Next    [PDF]