Cell adhesion-mediated signal transduction during rumor invasion and metastasis

For a tumor cell to become invasive and thus metastatic it is not only obligatory to lose cell adhesion but also to activate genetic or epigenetic programs that actively induce invasion into surrounding tissue and dissemination via blood vessels to form metastasis. Recent evidence suggests that loss of E-cadherin-mediated cell adhesion may be functionally coupled to such genetic programs. In particular, b -catenin besides its important role in E-cadherin- mediated cell adhesion is directly involved in Wnt-mediated signal transduction, raising the intriguing possibility that changes in cell adhesion may modulate gene expression and thus cell fate. However, the molecular signals that repress E-cadherin-mediated cell adhesion during tumor progression are not known. Moreover, the downstream effector genes that are induced by changes in cell adhesion remain unidentified. We propose to utilize a transgenic mouse model of pancreatic b cell carcinogenesis in vivo and tumor cell lines in vitro to establish the molecular hierarchy responsible for the induction of tumor invasion and metastasis, i.e. the communication pathway between extracellular signals, changes in cell adhesion, and changes in gene expression. In particular, we will assess whether members of the transforming growth factor beta (TGF b ) family are able to modulate E-cadherin-mediated cell adhesion and/or the Wnt-signalling pathway, since TGF b s have been implicated in the transition from benign to malignant tumors. Furthermore, we plan to investigate potential interconnections between E-cadherin-mediated cell adhesion, the Wnt-signalling pathway, and gene expression. Finally, we will attempt to identify and characterize the target genes that are activated by this pathway during tumor progression. These experiments are aimed at a better understanding of how tumor cells control their malignant phenotype. We hope that these experiments may also contribute to the identification of suitable targets for the development of tumor therapy.

N-CAM is a member of a large family of Ca2+-independent adhesion molecules containing varying numbers of immunoglobulin (Ig)-like domains and fibronectin type III repeats. Alternative splicing of the single copy gene gives rise to various N-CAM isoforms that based on their C-terminal domains are grouped into three major classes, N-CAM120, N-CAM140 and N-CAM180. All isoforms carry five Ig-like domains and two fibronectin type III domains in their extracellular parts. N-CAM is a multivalent adhesion molecule that mediates homotypic and heterotypic cell-cell adhesion via a homophilic binding mechanism, and it is also involved in heterophilic interactions with L1, heparan sulfate proteoglycans and collagens I-IV and IX. In many cancers, including Wilms` tumor, colon carcinoma, melanoma, Ewing sarcoma and neuroblastoma, expression of neural cell adhesion molecule (N-CAM) changes from the GPI-linked 120kD isoform (N-CAM120), which is usually found in adult tissues, to the transmembrane 140kD and 180kD isoforms (N-CAM140 and N- CAM180), which are predominantly expressed during embryonic development. Moreover, reduced overall levels of N-CAM expression correlate with increased malignancy and poor prognosis in pancreatic and colorectal cancers. Previously, our laboratory has employed a transgenic mouse model of pancreatic ß cell carcinogenesis (Rip1Tag2) to determine the functional role of N-CAM in tumor progression. In Rip1Tag2 transgenic mice, SV40 T antigen is expressed under the control of the rat insulin promoter resulting in the development of ß cell tumors in the pancreatic islets of Langerhans. Metastases are usually not found in these mice, however, when Rip1Tag2 mice were crossed with N-CAM knockout mice, a dramatic increase in tumor metastases, predominantly to local lymph nodes, was observed. Notably, loss of N-CAM function did not affect the transition from benign, non-invasive adenoma to invasive, malignant carcinoma, a process that is critically dependent on the loss of E-cadherin-mediated cell-cell adhesion. Hence, loss of N-CAM may directly modulate the metastatic dissemination of ß tumor cells, however, the underlying molecular mechanisms have remained elusive. The histological examination of N-CAM knockout vs. wildtype Rip1Tag2 tumors revealed that the absence of N- CAM results in a dramatic disaggregation of the tumor tissue. Large hemorrhagic cavities appear in N-CAM-/- tumors, often containing clusters of tumor cells. Such a phenotype suggested that cell adhesion was altered in N- CAM-deficient tumors and, therefore, we isolated ß tumor cell lines to investigate their adhesive properties. No difference in cell-cell adhesion was observed between N-CAM knockout and wild type cells. In contrast, N-CAM - /- cells showed impaired adhesion to specific components of the extracellular matrix, such as collagen IV and heparansulfate proteoglycans, and to extracellular matrix produced by the tumor cells themselves and by endothelial cells. The adhesion defect was rescued by transfecting knockout cells with N-CAM cDNA. Interestingly, the adhesion of tumor cells to collagen I, fibronectin, and laminin, was not affected by the absence of N-CAM, suggesting that only the activity of specific integrins is modulated by N-CAM. Cultured ß tumor cells formed long cytoplasmic protrusions that, by immunophenotypical analysis, were characterized as bona fide neurites. Neurite outgrowth was also controlled by N-CAM in ß tumor cells, as N-CAM -/- cells were not able to extend neurites, and this ability was restored by the forced expression of N-CAM. In an attempt to characterize the signal tranduction pathways underlying N-CAM-mediated cell-substrate adhesion and neurite outgrowth, we performed a series of co-immunoprecipitation experiments to identify signaling molecules eventually associated with N-CAM in ß tumor cells. These studies revealed that N-CAM physically interacts with a member of the fibroblast growth factor receptor (FGFR) family, namely FGFR-4. In addition, various components of the FGFR signaling pathway, such as PLC, FRS2, pp60c-src , and cortactin, were co- immunoprecipitated with N-CAM. The association between N-CAM and FGFR-4 was confirmed in an independent cellular system consisting of L cells co-expressing the two proteins. This approach also allowed us to establish that the intracellular domain of N-CAM is not required for its interaction with FGFR-4, since the GPI- linked N-CAM120 co-precipitated with FGFR-4 at the same level as the transmembrane isoforms. Notably, the interaction with FGFR-4 was essential for N-CAM function, in that forced expression of dominant negative FGFR- 4, but not FGFR-1, prevented N-CAM-mediated substrate adhesion and neurite outgrowth in ß tumor cells. Moreover, chemical inhibitors that interfere with various signal transduction pathways downstream of FGFR activation also blocked N-CAM-dependent cell-substrate adhesion. In particular, our results showed that pp60c-src , PLC, PI3K, and MAPK pathways are all implicated in N-CAM/FGFR-4-dependent cell-matrix adhesion. We also observed that FGFR-4 is constitutively activated in N-CAM +/+ tumor cells, but not in knockout cells. As mentioned above, our results indicated that specific integrins are involved in N-CAM modulation of cell- substrate adhesion. By utilizing neutralizing antibodies, we identified ß1 integrin as a major target of the N- CAM/FGFR-4 signaling complex. Indeed, blocking ß1 integrin function in ß tumor cells resulted in the inhibition of N-CAM-dependent cell-substrate adhesion and neurite outgrowth. Taken together, our results indicate that several signal transduction pathways are activated by N-CAM-induced FGFR-4 signaling, at least some of them leading to ß1 integrin-mediated matrix adhesion and neurite outgrowth. N-cadherin is another cell adhesion molecule which has been proposed to mediate neurite outgrowth in neurons by interacting with FGFR-1. Since N-cadherin is expressed at high level in ß tumor cells, we investigated whether N- cadherin also associated with the N-CAM/FGFR-4 complex in this cellular system. Indeed, N-cadherin co- immunoprecipitated with N-CAM, an interaction that has not been previously described. N-cadherin also co- precipitated with FGFR-4, but only in the presence of N-CAM, i.e. no association was detected in N-CAM- deficient cells. The growth-associated protein-43 (GAP-43), a molecule directly involved in axon growth and guidance, also associated with N-CAM in ß tumor cells. In addition, similar to N-cadherin, GAP-43 interacted with FGFR-4 and N-cadherin in a N-CAM-dependent manner. Taken together, our observations point to N-CAM as a key player in the assembly of a membrane signaling complex modulating cell-matrix adhesion and neurite outgrowth in ß tumor cells. The results described in this report contribute to our understanding of how cell adhesion molecules transmit intracellular signals that modulate cell-matrix adhesion and integrin function. In addition, as mentioned above, N- CAM downregulation correlates with tumor malignancy in many human cancers and causes metastatic dissemination in Rip1Tag2 transgenic mice. Thus, our characterization of N-CAM signaling might shed light on the molecular mechanisms underlying tumor invasion and metastasis and, therefore, it may help in the development of novel therapeutic strategies against neoplastic diseases.