The adhesion molecule NCAM promotes ovarian cancer progression via FGFR signalling
2011; Springer Nature; Volume: 3; Issue: 8 Linguagem: Inglês
10.1002/emmm.201100152
ISSN1757-4684
AutoresSilvia Zecchini, Lorenzo Bombardelli, Alessandra Decio, M. Bianchi, Giovanni Mazzarol, F. Sanguineti, Giovanni Aletti, Luigi Maddaluno, Vladimir Berezin, Elisabeth Bock, Chiara Casadio, Giuseppe Viale, Nicoletta Colombo, Raffaella Giavazzi, Ugo Cavallaro,
Tópico(s)Wnt/β-catenin signaling in development and cancer
ResumoResearch Article8 July 2011Open Access The adhesion molecule NCAM promotes ovarian cancer progression via FGFR signalling Silvia Zecchini Silvia Zecchini IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy These authors contributed equally to this work. Search for more papers by this author Lorenzo Bombardelli Lorenzo Bombardelli IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy These authors contributed equally to this work. Search for more papers by this author Alessandra Decio Alessandra Decio Department of Oncology, Mario Negri Institute for Pharmacological Research, Milano, Italy Search for more papers by this author Marco Bianchi Marco Bianchi IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Search for more papers by this author Giovanni Mazzarol Giovanni Mazzarol European Institute of Oncology, Milano, Italy Search for more papers by this author Fabio Sanguineti Fabio Sanguineti European Institute of Oncology, Milano, Italy Search for more papers by this author Giovanni Aletti Giovanni Aletti European Institute of Oncology, Milano, Italy Search for more papers by this author Luigi Maddaluno Luigi Maddaluno IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Search for more papers by this author Vladimir Berezin Vladimir Berezin Protein Laboratory, Department of Neuroscience and Pharmacology, Panum Institute, University of Copenhagen, Copenhagen, Denmark Search for more papers by this author Elisabeth Bock Elisabeth Bock Protein Laboratory, Department of Neuroscience and Pharmacology, Panum Institute, University of Copenhagen, Copenhagen, Denmark Search for more papers by this author Chiara Casadio Chiara Casadio European Institute of Oncology, Milano, Italy Search for more papers by this author Giuseppe Viale Giuseppe Viale European Institute of Oncology, Milano, Italy University of Milano School of Medicine, Milano, Italy Search for more papers by this author Nicoletta Colombo Nicoletta Colombo European Institute of Oncology, Milano, Italy Search for more papers by this author Raffaella Giavazzi Raffaella Giavazzi Department of Oncology, Mario Negri Institute for Pharmacological Research, Milano, Italy Search for more papers by this author Ugo Cavallaro Corresponding Author Ugo Cavallaro [email protected] IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Present address: European Institute of Oncology, Molecular Medicine Program, Milano, Italy Search for more papers by this author Silvia Zecchini Silvia Zecchini IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy These authors contributed equally to this work. Search for more papers by this author Lorenzo Bombardelli Lorenzo Bombardelli IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy These authors contributed equally to this work. Search for more papers by this author Alessandra Decio Alessandra Decio Department of Oncology, Mario Negri Institute for Pharmacological Research, Milano, Italy Search for more papers by this author Marco Bianchi Marco Bianchi IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Search for more papers by this author Giovanni Mazzarol Giovanni Mazzarol European Institute of Oncology, Milano, Italy Search for more papers by this author Fabio Sanguineti Fabio Sanguineti European Institute of Oncology, Milano, Italy Search for more papers by this author Giovanni Aletti Giovanni Aletti European Institute of Oncology, Milano, Italy Search for more papers by this author Luigi Maddaluno Luigi Maddaluno IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Search for more papers by this author Vladimir Berezin Vladimir Berezin Protein Laboratory, Department of Neuroscience and Pharmacology, Panum Institute, University of Copenhagen, Copenhagen, Denmark Search for more papers by this author Elisabeth Bock Elisabeth Bock Protein Laboratory, Department of Neuroscience and Pharmacology, Panum Institute, University of Copenhagen, Copenhagen, Denmark Search for more papers by this author Chiara Casadio Chiara Casadio European Institute of Oncology, Milano, Italy Search for more papers by this author Giuseppe Viale Giuseppe Viale European Institute of Oncology, Milano, Italy University of Milano School of Medicine, Milano, Italy Search for more papers by this author Nicoletta Colombo Nicoletta Colombo European Institute of Oncology, Milano, Italy Search for more papers by this author Raffaella Giavazzi Raffaella Giavazzi Department of Oncology, Mario Negri Institute for Pharmacological Research, Milano, Italy Search for more papers by this author Ugo Cavallaro Corresponding Author Ugo Cavallaro [email protected] IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy Present address: European Institute of Oncology, Molecular Medicine Program, Milano, Italy Search for more papers by this author Author Information Silvia Zecchini1, Lorenzo Bombardelli1, Alessandra Decio2, Marco Bianchi1, Giovanni Mazzarol3, Fabio Sanguineti3, Giovanni Aletti3, Luigi Maddaluno1, Vladimir Berezin4, Elisabeth Bock4, Chiara Casadio3, Giuseppe Viale3,5, Nicoletta Colombo3, Raffaella Giavazzi2 and Ugo Cavallaro *,1,6 1IFOM – The FIRC Institute of Molecular Oncology, Milano, Italy 2Department of Oncology, Mario Negri Institute for Pharmacological Research, Milano, Italy 3European Institute of Oncology, Milano, Italy 4Protein Laboratory, Department of Neuroscience and Pharmacology, Panum Institute, University of Copenhagen, Copenhagen, Denmark 5University of Milano School of Medicine, Milano, Italy 6Present address: European Institute of Oncology, Molecular Medicine Program, Milano, Italy *Tel. +39 02 9347 5165; Fax +39 02 9347 9305 EMBO Mol Med (2011)3:480-494https://doi.org/10.1002/emmm.201100152 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Epithelial ovarian carcinoma (EOC) is an aggressive neoplasm, which mainly disseminates to organs of the peritoneal cavity, an event mediated by molecular mechanisms that remain elusive. Here, we investigated the expression and functional role of neural cell adhesion molecule (NCAM), a cell surface glycoprotein involved in brain development and plasticity, in EOC. NCAM is absent from normal ovarian epithelium but becomes highly expressed in a subset of human EOC, in which NCAM expression is associated with high tumour grade, suggesting a causal role in cancer aggressiveness. We demonstrate that NCAM stimulates EOC cell migration and invasion in vitro and promotes metastatic dissemination in mice. This pro-malignant function of NCAM is mediated by its interaction with fibroblast growth factor receptor (FGFR). Indeed, not only FGFR signalling is required for NCAM-induced EOC cell motility, but targeting the NCAM/FGFR interplay with a monoclonal antibody abolishes the metastatic dissemination of EOC in mice. Our results point to NCAM-mediated stimulation of FGFR as a novel mechanism underlying EOC malignancy and indicate that this interplay may represent a valuable therapeutic target. The paper explained PROBLEM: Epithelial ovarian carcinoma (EOC) represents an outstanding challenge for clinical oncologists, due to the difficulty of early diagnosis and to the high rate of relapse that follows the dissemination of cancer cells to the peritoneal cavity. Recurrent tumours often develop chemoresistance, accounting for the poor 5-year survival rate of EOC patients. Therefore, a better understanding of the biological mechanisms and molecular players involved in EOC malignancy is essential to design innovative therapeutic strategies and improve the prognosis for patients. RESULTS: We have shown that the NCAM is not expressed in normal ovarian epithelium, while high levels were detected in a subset of cancer samples from EOC patients, where NCAM expression was associated with the tumour grade. To understand whether this expression pattern underlies a functional contribution of NCAM to EOC malignancy, we combined loss and gain-of-function approaches in EOC-derived cells. Our results revealed that NCAM confers a migratory and invasive phenotype to cancer cells. This function is mediated by NCAM's ability to interact with and activate fibroblast growth factor receptor (FGFR), a receptor tyrosine kinase that has previously been implicated in EOC progression. Furthermore, by employing mouse models of EOC development, we observed that the NCAM/FGFR interplay promotes both local invasion and metastatic dissemination of EOC to organs of the peritoneal cavity. Notably, a mAb that disrupts the interaction of NCAM with FGFR repressed the pro-metastatic activity of NCAM. Our findings implicated NCAM-mediated stimulation of FGFR function as a novel mechanisms underlying EOC aggressiveness. IMPACT: Our study outlines the causal role of NCAM in EOC invasion and metastasis, a role mediated by the peculiar ability of this adhesion molecule to bind and activate FGFR. While we have recently described the outcome of the NCAM/FGFR interaction at the molecular and sub-cellular level, this is the first evidence of the contribution of such a signalling axis to cancer aggressiveness. On one hand, our data highlight novel molecular mechanisms involved in EOC malignancy, allowing for a deeper understanding of the biology of this elusive disease. On the other hand, we provide proof-of-principle evidence that the NCAM/FGFR interplay represents a therapeutic target that can be efficiently inactivated, resulting in repression of EOC metastasis. This supports the rationale for combining NCAM/FGFR-targeted therapy with different treatments as an innovative approach to design more effective anti-EOC strategies. INTRODUCTION Epithelial ovarian carcinoma (EOC) is the fourth most common cause of cancer-related death among women in the western world and is the most lethal tumour type among gynecological malignancies. The difficulty of early diagnosis and the rapid dissemination represent outstanding clinical challenges associated with EOC and contribute to the frequent failure of current treatment protocols. Even upon extensive surgical debulking and chemotherapy, relapse is a common event and, in most cases, recurrent EOC is resistant to any treatment (Lengyel, 2010). EOC is commonly thought to develop from the ovarian surface epithelium (OSE), a single layer of poorly differentiated epithelial cells that lines the ovary, although the fallopian tube epithelium and the peritoneal mesothelium are also considered as potential sites of EOC origin (Lengyel, 2010). While various genetic lesions involved in the pathogenesis of EOC have been identified, our understanding of the biology of this neoplasm is still incomplete. This is partially due to the peculiar molecular and clinico-pathological properties of EOC, which prevents from transferring to ovarian carcinogenesis many of the concepts established in other epithelial tumours. For example, unlike most solid tumours, the metastatic dissemination of EOC very rarely occurs through the blood circulation. In most cases, EOC metastases arise from small clusters of cancer cells that are shed from the ovary and adhere to the surface of the peritoneal cavity and abdominal organs (Bast et al, 2009). Cell adhesion molecules mediating either cell–cell interactions or cell–matrix adhesion have emerged as key players throughout the natural history of EOC development, in that they have been implicated both in cancer cell survival upon detachment from the primary tumour and in the subsequent adhesion to and invasion of metastatic sites (Naora & Montell, 2005; Sundfeldt, 2003). Besides their structural function, adhesion molecules also play an important role in signalling, either directly or by modulating the activity of signalling receptors. While the deregulation of these functional properties has been implicated in cancer progression (Cavallaro & Christofori, 2004), its specific contribution to EOC development remains to be defined. Neural cell adhesion molecule (NCAM) is a cell-surface glycoprotein with an extracellular portion composed of five Ig domains and two fibronectin type-III repeats. NCAM function has been extensively characterized in the nervous system, where it regulates intercellular adhesion, neurite outgrowth and neuronal migration. These activities are mediated both by homophilic interactions and by heterophilic binding of NCAM to a number of different membrane proteins or components of the extracellular matrix (Hinsby et al, 2004). Among the heterophilic partners of NCAM, the fibroblast growth factor receptor (FGFR) has attracted the attention of many investigators due to its functional implications. The four members of the FGFR family (FGFR1–4) consist of receptor tyrosine kinases that are classically activated upon binding of cognate growth factors, the FGFs. FGFR stimulation results in the recruitment and activation of specific effectors that, in turn, trigger a set of signalling pathways (Beenken & Mohammadi, 2009). The functional interaction between NCAM and FGFR was originally reported in neurons, where it was implicated in neurite outgrowth (Williams et al, 1994). Thereafter, we and others have provided extensive evidence of a physical association between the two proteins on different, non-neural cell types (Cavallaro et al, 2001; Francavilla et al, 2007; Kos & Chin, 2002; Sanchez-Heras et al, 2006). These data were further confirmed and extended by surface plasmon and magnetic resonance studies, in which the direct binding of NCAM to FGFR1 and FGFR2 was demonstrated (Christensen et al, 2006; Francavilla et al, 2009; Kiselyov et al, 2003). This approach also revealed that the FGFR-binding motifs are located in the two fibronectin type-III repeats of the NCAM ectodomain (Hansen et al, 2010). We and others have demonstrated that the interaction of NCAM with FGFR results in the stimulation of FGFR signalling in various cell types (Cavallaro et al, 2001; Francavilla et al, 2009; Kiselyov et al, 2003; Williams et al, 1994). Nevertheless, the biological significance of the NCAM/FGFR interplay, especially in non-neural contexts, remains to be established. Aberrant FGFR activity has been implicated in the onset and/or progression of various cancer types, and several FGFR inhibitors are currently in early phases of clinical development (reviewed in Turner & Grose, 2010). The deregulation of FGFR signalling in cancer can result from different mechanisms, including genomic alterations that lead to ligand-independent signalling as well as uncontrolled receptor stimulation by FGFs (Turner & Grose, 2010). All four members of the FGFR family and various FGFs have been found in EOC tissue (Birrer et al, 2007; Cole et al, 2010; Valve et al, 2000), suggesting that dysregulated FGFR signalling contributes to ovarian carcinogenesis (Castellano et al, 2006; De Cecco et al, 2004; Valve et al, 2000) and, therefore, it may represent a suitable therapeutic target (Ivan & Matei, 2010). Based on the ability of NCAM to modulate FGFR function and on the proposed role of FGFR activity in ovarian cancer, we hypothesized that the NCAM/FGFR signalling axis is causally involved in EOC development. Screening of tumour biopsies revealed that NCAM is expressed in a significant proportion of EOC samples, but not in normal tissue, where its levels are increased at the invasive front and correlate with tumour grade. Furthermore, we show that the NCAM/FGFR interplay induces EOC invasion and peritoneal dissemination, and that interfering with this interaction represents a promising strategy to inhibit EOC progression. Our findings, therefore, provide novel insights into molecular mechanisms involved in EOC aggressiveness and offer the possibility to explore innovative therapeutic approaches for EOC treatment. RESULTS The expression of NCAM in human EOC The expression of NCAM was investigated in a panel of surgically resected specimens, including 20 normal ovaries, 4 benign lesions (cystadenoma), 252 primary EOC and 206 metastatic lesions (see Materials and Methods Section for details on selection and analysis of patients). Immunohistochemical staining with anti-NCAM antibody showed no signal in the surface epithelium of normal ovaries (Fig 1A) or in pre-neoplastic lesions (not shown). In contrast, 60 (23.8%) primary EOC (Fig 1B) and 71 (34.5%) metastases (Fig 1D) were clearly positive for NCAM, thus indicating that NCAM expression occurs specifically in transformed ovarian epithelial cells. Interestingly, increased levels of NCAM were frequently observed at the invasive front of the tumour lesions (Fig 1B and C). This led to the hypothesis that NCAM promotes EOC invasion, an issue that we have investigated at the cellular level as well as in mouse models (see below). We also asked whether NCAM expression correlated with a series of clinicopathological parameters and found a statistically significant association with higher tumour grade (Table 1). NCAM expression was also more frequent in tumours at advanced International Federation of Gynecology and Obstetrics (FIGO) stages as compared to early stages, although the correlation did not reach statistical significance (Table 1). Taken together, these data point to NCAM as a novel biomarker of EOC associated with clinicopathological features of cancer aggressiveness. Figure 1. Expression of NCAM in EOC. Tissue sections were stained with anti-NCAM antibody. A.. Normal ovary. The arrow indicates the NCAM-negative OSE. Scale bar, 100 µm. B,C.. Primary EOC. The arrows indicate the invasive edges of the tumour lesions. T, tumour; S, stroma. Scale bars, 100 µm. D.. Lymph node metastasis of EOC. M, metastasis; LN, lymph node. Scale bar, 50 µm. Download figure Download PowerPoint Table 1. Clinicopathological features and NCAM expression in EOC patients Ovarian tissues p-Value NCAM-positive NCAM-negative Normal OSE 0 (0%) 20 (100%) Cystadenoma 0 (0%) 4 (100%) 0.0003 EOC 60 (23.8%) 192 (76.2%) (OSE vs. EOC) Metastases 71 (34.5%) 135 (65.5%) Histotype Serous 45 (26.6%) 124 (73.4%) NS Endometrioid 8 (18.6%) 35 (81.4%) Mixed 7 (28%) 18 (72%) Clear cell 0 (0%) 8 (100%) Mucinous 0 (0%) 5 (100%) Transitional 0 (0%) 1 (100%) Grade G1 0 (0%) 12 (100%) G2 19 (27.9%) 49 (72.1%) 0.033 G3 41 (24.4%) 127 (75.6%) (G1 vs. G2/G3) FIGO Stage Low (I–II) 6 (16.2%) 31 (83.8%) NS High (III–IV) 54 (25.1%) 161 (74.9%) Nodal status Negative 42 (24.6%) 129 (75.4%) NS Positive 18 (30%) 63 (77.8%) NS, not significant. We also stained a subset of primary EOC samples for FGFR1 and detected the receptor in 75 out of 77 samples (97%), in agreement with previous studies that reported widespread expression of FGFR1, FGFR2 and FGFR4 in EOC (Steele et al, 2001; Valve et al, 2000). NCAM expression was found in 28 of the 75 FGFR1-positive samples (Table S1 and Fig S1 of Supporting Information). Furthermore, we found a highly significant correlation between the expression of NCAM and that of the individual FGFR genes in a published cDNA microarray dataset of 255 EOC samples (Crijns et al, 2009; Fig S2 of Supporting Information). Therefore, the presence of NCAM in EOC tissue is accompanied by the expression of FGFRs, thus supporting the functional role of the NCAM–FGFR interaction in EOC as described below. NCAM-dependent FGFR signalling is required for EOC cell migration and invasion To investigate whether NCAM is involved in the malignant phenotype of EOC cells, we utilized the MOVCAR cell line, originally isolated from the cancer tissue of MISIIR-TAg transgenic mice, a genetic model of ovarian carcinoma (Connolly et al, 2003). This cell line expresses high levels of NCAM, which is properly localized at the cell surface (Fig S3A of Supporting Information). MOVCAR cells were retrovirally transduced with a short-hairpin RNA that targets the murine NCAM mRNA (MOVCAR-shNCAM cells), resulting in efficient knockdown of NCAM expression (Fig S3A and B of Supporting Information). To rule out any off-target effect of the NCAM shRNA, MOVCAR-shNCAM cells were reconstituted with human NCAM, which is not affected by the shRNA (Fig S3A and B of Supporting Information). The loss of NCAM had no effect on MOVCAR cell proliferation (Fig S3C of Supporting Information). In contrast, the migratory activity of MOVCAR-shNCAM was twofold lower as compared to cells transduced with a control shRNA (Fig 2A). This effect was specifically due to NCAM gene silencing, since the expression of human NCAM restored the migratory potential of MOVCAR-shNCAM cells (Fig 2A). Figure 2. NCAM and its interplay with FGFR are required for EOC cell migration and invasion. MOVCAR cells were transduced with either a control short-hairpin RNA (shControl) or a short-harpin RNA against mouse NCAM, either expressed alone (shNCAM) or in combination with human NCAM (shNCAM/reconstituted). A.. Migration assay in the presence of either vehicle ('control'), PD173074 or AG1478. Values are expressed in arbitrary units as fold changes over the migration or invasion of MOVCAR-shControl cells treated with vehicle alone. Error bars represent SEM. **p < 0.005. B.. Matrigel invasion assays in the presence of either vehicle ('control'), PD173074 or AG1478. Values are expressed as in panel A. **p < 0.005 and *p < 0.05. The difference between untreated (left grey bar) and AG1478-treated shNCAM cells (right grey bar) is not statistically significant. Download figure Download PowerPoint Based on the notion that NCAM binds to FGFR and modulates FGFR activity (see Introduction Section), we asked whether this interplay is involved in the NCAM-dependent migration of MOVCAR cells. The latter express functional FGFRs, as demonstrated by their proliferative response to FGF-2 stimulation (Fig S3C of Supporting Information). To determine the contribution of FGFR signalling to NCAM-dependent cell migration, we took advantage of the FGFR inhibitor PD173074 (Skaper et al, 2000). This compound reduced the migration of control MOVCAR cells to a level comparable to untreated MOVCAR-shNCAM cells (Fig 2A). Moreover, the knockdown of NCAM expression and the inhibition of FGFR activity showed no additive effect on MOVCAR cell migration, suggesting that the two molecules are active in the same pathway. The defect in cell migration of MOVCAR-shNCAM cells was rescued upon ectopic expression of human NCAM. However, the ability of human NCAM to restore cell migration was abolished by PD173074 (Fig 2A), thus confirming that FGFR signalling is required for NCAM-induced ovarian cancer cell migration. It should be noted that FGF-2 had no effect on the migration of either control of NCAM-knockdown MOVCAR cells (not shown), as previously shown for different cell types (Francavilla et al, 2009). Finally, MOVCAR cell migration was also measured in the presence of AG1478, a chemical inhibitor of epidermal growth factor receptor (EGFR). In spite of EGFR expression in MOVCAR cells (data not shown), AG1478 had a negligible effect on NCAM-dependent cell migration (Fig 2A), implying that EGFR activity plays no or very little role downstream of NCAM and supporting the specificity of the NCAM/FGFR interplay. We also employed the Matrigel invasion assay to determine whether NCAM and its functional interaction with FGFR were required for the invasive potential of MOVCAR cells. As shown in Fig 2B, the knockdown of NCAM resulted in the abrogation of MOVCAR cell invasion, which was restored upon reconstitution with human NCAM. As observed for cell migration, PD173074 not only reduced the basal invasiveness of control MOVCAR cells, but also abolished the pro-invasive effect of human NCAM (Fig 2B). Taken together, these results indicate that NCAM is required for both migration and invasion of MOVCAR cells and exerts its function via FGFR activity. NCAM stimulates EOC cell migration via its interaction with FGFR Following the observation that NCAM/FGFR interplay is necessary for EOC cell migration and invasion, we asked whether it is also sufficient. To address this question, we selected two human EOC cell lines, SKOV3 and OVCA-433, which express no endogenous NCAM (Figs S4 and S5 of Supporting Information, panels A and B). Both cell lines express various FGFR family members (our unpublished observations; Chandler et al, 1999; Valve et al, 2000), thus providing a suitable experimental system to investigate the impact of ectopically expressed NCAM on FGFR activity. SKOV3 and OVCA-433 cells were stably transfected with full-length NCAM (Figs S4 and S5 of Supporting Information, panels A and B). SKOV3 cells were also retrovirally transduced with a GFP-encoding vector (Fig S4A and B of Supporting Information) for xenotransplantation purposes (see below). In agreement with the data on NCAM silencing in MOVCAR cells, ectopic expression of NCAM in human EOC cells had no effect on cell proliferation (Fig S4C of Supporting Information and data not shown). Rather, NCAM-expressing SKOV3 and OVCA-433 cells exhibited a remarkable increase in their migratory activity as compared to mock-transfected cells (Fig 3A and Fig S5C of Supporting Information). Previous studies have established that the interaction with FGFR involves the two FNIII repeats of NCAM (referred to as FN1 and FN2; Kiselyov et al, 2003). Therefore, to determine the relative contribution of FGFR binding in NCAM-induced EOC cell migration, SKOV3 and OVCA-433 cells were transfected with a mutant version of NCAM lacking FN2 (ΔFN2), in which the interaction with FGFR is disrupted (Francavilla et al, 2007). Mutant NCAM was expressed at a level comparable to full-length NCAM and retained the localization at the cell surface (Figs S4 and S5 of Supporting Information, panels A and B), as well as the ability to induce cell–cell adhesion (not shown), which is mediated by distal Ig domains (Soroka et al, 2003). Furthermore, deletion of the FN2 domain did not alter folding, localization or the adhesive properties of NCAM. The ability of NCAM to stimulate FGFR signalling has been mostly characterized on FGFR1 (Francavilla et al, 2009; Kiselyov et al, 2003), a member of the FGFR family that is prominently expressed in both SKOV3 and OVCA433 cells (Fig S6B of Supporting Information and data not shown). Therefore, we focused on FGFR1 to define the impact of NCAM expression on FGFR activation. First, we confirmed that full-length NCAM, but not ΔFN2, forms a complex with FGFR1 in EOC cells (Fig S6A of Supporting Information). Accordingly, only full-length NCAM was able to induce autophosphorylation of FGFR1 in SKOV3 cells (Fig S6B of Supporting Information), confirming that also in EOC cells NCAM stimulates FGFR activation through its FN domains. In both SKOV3 and OVCA-433 cells, NCAM-ΔFN2 failed to promote cell migration (Fig 3A and Fig S5C of Supporting Information), thus implicating the interaction with FGFR as a pre-requisite for NCAM-induced migration of EOC cells. To further confirm this notion, SKOV3 cells expressing NCAM or a control vector were either treated with the FGFR inhibitor PD173074 or transfected with a dominant-negative version of FGFR1. In both cases, FGFR inhibition resulted in the abrogation of NCAM-dependent migration (Fig 3A) supporting the results obtained with NCAM-ΔFN2. By analogy to MOVCAR cells (see above), FGF-2 showed no pro-migratory activity on SKOV3 cells, a situation that was not changed by the ectopic expression of NCAM (not shown). This suggests that FGF does not cooperate with NCAM in the stimulation of FGFR-mediated EOC cell migration. Figure 3. NCAM induces EOC cell migration by interacting with FGFR. A.. SKOV3 cells transfected with empty vector (mock), with full-length NCAM or with NCAM-ΔFN2 were subjected to migration assays in the presence or absence of PD173074. A set of cells was co-transfected with dn-FGFR1 as indicated. Values are expressed in arbitrary units as fold changes over the migration of SKOV3-mock cells treated with vehicle. B.. Parental SKOV3 cells were treated with Encamin-C or with a control scrambled peptide, either in the absence or in the presence of PD173074 or AG1478, followed by migration assay. Values are expressed in arbitrary units as fold changes over the migration of cells treated with the control peptide. C.. C SKOV3 cells transfected with empty vector (mock) or with full-length NCAM were subjected to migration assays in the presence of the mAb 123C3, Eric-1 or anti-HA (as an isotype-matched control antibody). Values are expressed in arbitrary units as fold changes over the migration of untreated SKOV3-mock cells. Download figure Download PowerPoint To test whether NCAM stimulation of FGFR function per se is sufficient to induce EOC cell migration, we took advantage of the Encamin-C peptide, derived from the FN1 module of NCAM, which has been recently demonstrated to bind to and activate FGFR1 (Hansen et al, 2008). First, we confirmed the FGFR-activating properties of Encamin-C in EOC cells. Encamin-C-treated SKOV3 cells displayed time-dependent autophosphorylation of FGFR1 (Fig S6C of Supporting Information), while a control, scrambled peptide had no effect (not shown). Encamin-C also enhanced the migratory potential of SKOV3 cells, an effect that was suppressed by PD173074 (Fig 3B), confirming that the peptide acts through FGFR signalling. In contrast, the EGFR inhibitor AG1478 showed no significant effect on Encamin-C-induced migration (Fig 3B), supporting the specificity of the peptide interaction with FGFR. These findings demons
Referência(s)