Neural cell adhesion molecule (NCAM) mediates cell-cell adhesion and signaling in the nervous system, yet NCAM is also expressed in non-neural tissues, in which its function has in most parts remained elusive. We have previously reported that NCAM stimulates cell-matrix adhesion and neurite outgrowth by activating fibroblast growth factor receptor (FGFR) signaling. Here, we investigated whether the interplay between NCAM and FGFR has any impact on the response of FGFR to its classical ligands, FGFs. To this end, we employed two fibroblast cell lines, NCAM-negative L cells and NCAM-positive NIH-3T3 cells, in which the expression of NCAM was manipulated by means of transfection or RNAi technologies, respectively. The results demonstrate that NCAM expression reduces FGF-stimulated ERK1/2 activation, cell proliferation and cell-matrix adhesion, in both L and NIH-3T3 cells. Furthermore, our data show that NCAM inhibits the binding of FGF to its high-affinity receptor in a competitive manner, providing the mechanisms for the NCAM-mediated suppression of FGF function. In this context, a small peptide that mimics the binding of NCAM to FGFR was sufficient to block FGF-dependent cell proliferation. These findings point to NCAM as being a major regulator of FGF-FGFR interaction, thus introducing a novel type of control mechanism for FGFR activity and opening new therapeutic perspectives for those diseases characterized by aberrant FGFR function.

Introduction

Neural cell adhesion molecule (NCAM) is a cell-surface glycoprotein member of the immunoglobulin superfamily mediating calcium-independent intercellular adhesion. The extracellular portion of NCAM contains five Ig-like domains (Ig1-Ig5) and two fibronectin type III (FNIII) repeats. Alternative splicing yields three main NCAM isoforms: NCAM140 and NCAM180, which contain a transmembrane and a cytoplasmic region, and NCAM120, which is linked to the membrane via a glycosylphosphatidyl inositol anchor (Walmod et al., 2004).

NCAM is widely expressed in the central nervous systems, in which it mediates several neuronal functions by controlling intercellular adhesion, neurite outgrowth, and cell migration, proliferation and survival. These events are triggered by the homophilic interaction of NCAM molecules on adjacent cells as well as by the heterophilic binding of NCAM to other adhesion molecules, extracellular matrix components and cell-surface receptors (Hinsby et al., 2004).

Previous studies have highlighted an interplay between NCAM and fibroblast growth factor receptor (FGFR) in neuronal cells, which underlies NCAM-dependent neurite outgrowth (Walsh and Doherty, 1997). These results were further supported by the demonstration of a physical association between NCAM and FGFR. Indeed, we previously reported the co-immunoprecipitation of the two proteins in non-neuronal cell types (Cavallaro et al., 2001), subsequently confirmed in other cellular systems (Kos and Chin, 2002; Sanchez-Heras et al., 2006). Finally, surface plasmon resonance and NMR studies have mapped the interaction domains in the second FNIII repeat of NCAM and the second and third Ig domains of FGFR1 (Kiselyov et al., 2003). More recently, NCAM has also been reported to bind to FGFR2 (Christensen et al., 2006).

We previously reported the formation of an NCAM-FGFR complex in tumor cells isolated from the Rip1Tag2 transgenic mouse model of pancreatic beta cell carcinogenesis. In those cells, NCAM-mediated stimulation of FGFR signaling leads to the modulation of β1-integrin-mediated cell-matrix adhesion (Cavallaro et al., 2001). Ablation of the expression of NCAM in Rip1Tag2 mice results in the disruption of the tumor tissue architecture, tumor-associated lymphangiogenesis and lymph node metastasis (Cavallaro et al., 2001; Crnic et al., 2004; Perl et al., 1999; Xian et al., 2006). Thus, the crosstalk between adhesion molecules and receptor tyrosine kinases has important functional implications, and its deregulation can play a pathogenic role in a number of diseases, including cancer and neurological disorders (Cavallaro and Christofori, 2004).

Here, we address the issue of whether the NCAM-FGFR crosstalk has any impact on the cellular response to the classical FGFR ligands, the fibroblast growth factors (FGFs). Our results implicate NCAM as a novel regulator of FGF function, in that several FGF-induced processes (including signal transduction and cell proliferation) are repressed by NCAM. This negative effect of NCAM depends on its ability to compete with FGF for binding to FGFR. Overall, our findings introduce a novel type of control mechanism for FGFR activity.

Results

NCAM colocalizes with FGFR1 at the cell surface

We previously reported the co-immunoprecipitation of NCAM and FGFR from lysates of pancreatic beta tumor cells (Cavallaro et al., 2001). This observation was subsequently supported by surface-plasmon-resonance studies showing the binding between recombinant fragments of the two proteins (Kiselyov et al., 2003). However, the demonstration of the physical association between NCAM and FGFR in living cells has remained elusive. To address this issue, in the current study we employed total internal reflection fluorescence (TIRF), a highly sensitive technology for the visualization of protein complexes at the cell surface (Piehler, 2005). Because the studies described below were conducted in fibroblasts, we first determined whether the NCAM-FGFR complex could be detected in this cell type. Thus, the TIRF analysis was performed on NIH-3T3 cells, which endogenously express moderate levels of both NCAM and FGFR1 (supplementary material Fig. S1A). As shown in Fig. 1, a colocalization of NCAM with FGFR1 was clearly detectable, confirming that the two molecules are closely associated in living cells. Furthermore, given that the TIRF analysis is intrinsically restricted to a distance from the coverslip of ∼100 nm (Schneckenburger, 2005), the immunoreactivity of the NCAM-FGFR1 complexes in NIH-3T3 cells was mostly localized at the cell surface.

NCAM represses FGF-induced signaling and cell proliferation

To study the role of NCAM in the regulation of FGF function, we reduced the expression of NCAM in NIH-3T3 cells by means of RNA interference (RNAi). The expression of two different siRNAs against mouse NCAM (siNCAM1 and siNCAM3) induced a strong reduction of NCAM levels, which lasted at least 96 hours, in NIH-3T3 cells, whereas a control siRNA targeting human NCAM did not affect NCAM expression in these cells (Fig. 2A and supplementary material Fig. S1A). Importantly, the downregulation of NCAM had no effect on the expression of FGFR1 (supplementary material Fig. S1A). NCAM-deficient and control NIH-3T3 cells were stimulated with FGF2 and the activation of the extracellular signal-regulated kinases 1 and 2 (ERK1/2) cascade, a classical signaling pathway elicited by FGF (and other growth factors), was assessed. Interestingly, the knockdown of NCAM led to a remarkable increase in the sensitivity of NIH-3T3 cells to FGF2 stimulation, because ERK1/2 activation was much stronger in cells transfected with either siNCAM1 or siNCAM3 than in control cells, both at 5 and 10 ng/ml FGF2 (Fig. 2A). Given that the ERK1/2 cascade often underlies growth-factor-induced cell proliferation, we determined whether NCAM also modulated the mitogenic activity of FGF2 on NIH-3T3 cells. Under basal conditions, control and siNCAM-expressing NIH-3T3 cells exhibited a very similar proliferative rate (Fig. 2B), indicating that NCAM per se had no effect on cell proliferation. As expected, FGF2 stimulated the proliferation of control NIH-3T3 cells. However, such a proliferative response was markedly enhanced upon knockdown of NCAM. In particular, the abrogation of NCAM accelerated FGF-induced cell proliferation, with a stronger response at 48 and 72 hours, whereas the proliferative rate of control and NCAM-deficient cells was again very similar at 96 hours (Fig. 2B). The data shown in Fig. 2B refer to cells transfected with siNCAM1; very similar results were obtained with siNCAM3 (not shown). In both control and siNCAM-expressing cells, the mitogenic activity of FGF2 was abolished by the MEK inhibitor PD98059 (supplementary material Fig. S1B), indicating that FGF2 stimulated cell proliferation via the ERK1/2 pathway. Hence, the abrogation of NCAM potentiated the signaling and proliferation elicited by FGF2, implicating NCAM as a negative regulator of FGF function.

To determine whether NCAM is sufficient to inhibit FGF activity, we selected L cells, another fibroblast cell line that does not express NCAM (Cavallaro et al., 2001). L cells were stably transfected with an empty vector (L-mock), with the full-length 140-kDa isoform of NCAM (L-NCAM) or with a mutant version that lacked the second FNIII repeat (L-ΔFN2), in which the FGFR-binding motif is located (Kiselyov et al., 2003). In L-ΔFN2 cells, mutant NCAM was expressed at a level comparable to full-length NCAM in L-NCAM cells and retained the localization at the cell surface (supplementary material Fig. S2), as well as the ability to induce cell-cell adhesion (not shown), indicating that the deletion did not alter the folding or the functional properties of the protein. Serum-starved L-mock cells showed a very low level of basal ERK1/2 activation, which was strongly enhanced by FGF2. By contrast, FGF2 failed to increase the moderate, constitutive activation of ERK1/2 that was detected in L-NCAM cells (Fig. 3A), indicating that NCAM repressed FGF2 signaling. Notably, ERK1/2 showed no constitutive activation in L cells expressing ΔFN2-NCAM and in these cells FGF2 stimulated an ERK1/2 activation comparable to control cells (Fig. 3A). The absence of constitutive ERK1/2 activation in L-ΔFN2 cells, as opposed to L-NCAM cells, indicated that NCAM stimulated FGFR signaling by the interaction of its FN2 domain with FGFR, confirming and extending previous observations in neurons and other cell types (Williams et al., 1994; Cavallaro et al., 2001; Kiselyov et al., 2003; Petersen et al., 2006). A detailed molecular and biochemical analysis of NCAM-mediated activation of FGFR in L cells will be reported elsewhere (C.F. and U.C., unpublished).

Fig. 1.

NCAM and FGFR1 colocalize at the cell surface. (A,B) NIH-3T3 cells were co-stained with antibodies against NCAM (red) and FGFR1 (green), followed by fluorochrome-conjugated secondary antibodies. After fixation, cells were subjected to TIRF analysis as described in the Materials and Methods. (B) Magnification of the boxed area in A. Arrows indicate the widespread, tight association of NCAM with FGFR1. Bars, 10 μm (A) and 0.2 μm (B).

Fig. 1.

NCAM and FGFR1 colocalize at the cell surface. (A,B) NIH-3T3 cells were co-stained with antibodies against NCAM (red) and FGFR1 (green), followed by fluorochrome-conjugated secondary antibodies. After fixation, cells were subjected to TIRF analysis as described in the Materials and Methods. (B) Magnification of the boxed area in A. Arrows indicate the widespread, tight association of NCAM with FGFR1. Bars, 10 μm (A) and 0.2 μm (B).

Fig. 2.

NCAM is required for the regulation of the cellular response to FGF. (A) At 48 hours after transfection either with a control siRNA, or with siNCAM1 or siNCAM3, NIH-3T3 cells were left untreated or were stimulated with 5 or 10 ng/ml FGF2 for 10 minutes. After the treatment, cells were lysed, subjected to SDS-PAGE and immunoblotting for phospho-ERK1/2, followed by stripping and immunoblotting for total ERK1/2 and NCAM. After longer exposure, the activation of ERK1/2 with 5 ng/ml FGF2 became detectable also in control cells (not shown). (B) Cells transfected with either a control siRNA or with siNCAM1 were left untreated or were treated with FGF2 for the indicated time periods. Cell proliferation was assessed as described in the Materials and Methods. *P<0.005, relative to untreated control cells.

Fig. 2.

NCAM is required for the regulation of the cellular response to FGF. (A) At 48 hours after transfection either with a control siRNA, or with siNCAM1 or siNCAM3, NIH-3T3 cells were left untreated or were stimulated with 5 or 10 ng/ml FGF2 for 10 minutes. After the treatment, cells were lysed, subjected to SDS-PAGE and immunoblotting for phospho-ERK1/2, followed by stripping and immunoblotting for total ERK1/2 and NCAM. After longer exposure, the activation of ERK1/2 with 5 ng/ml FGF2 became detectable also in control cells (not shown). (B) Cells transfected with either a control siRNA or with siNCAM1 were left untreated or were treated with FGF2 for the indicated time periods. Cell proliferation was assessed as described in the Materials and Methods. *P<0.005, relative to untreated control cells.

The ability of NCAM to interfere with FGF signaling in L cells was not restricted to ERK1/2 activation, because FGF2-induced tyrosine phosphorylation of p52SHC, a well-known FGFR substrate (Mohammadi et al., 1996), was also inhibited by full-length NCAM, but not by ΔFN2-NCAM (supplementary material Fig. S3). Therefore, NCAM expression is sufficient to repress FGF2 signaling, and its interaction with FGFR is a prerequisite for this inhibitory effect.

The results obtained in NIH-3T3 cells highlighted the role of NCAM in controlling FGF-dependent cell proliferation. To determine whether the ectopic expression of NCAM per se recapitulated this negative regulation, L cells transfected with full-length or ΔFN2-NCAM were stimulated with FGF2 for increasing time periods and counted to assess cell proliferation. As shown in Fig. 3B, FGF2 stimulated the growth of L-mock cells, an effect that required ERK1/2 activation, because it was inhibited by PD98059 (supplementary material Fig. S1C). Notably, the mitogenic activity of FGF2 on L cells was totally abolished by the expression of NCAM but not by ΔFN2-NCAM (Fig. 3B). By contrast, the expression of NCAM alone had no effect on the basal proliferation of L cells (supplementary material Fig. S4A). These results also indicate that the constitutive activation of ERK1/2 induced by the interaction of NCAM with FGFR in L-NCAM cells (see text above and Fig. 3A) does not promote cell proliferation. Indeed, NCAM has been reported to elicit FGFR-mediated cellular events that include neurite outgrowth but not cell proliferation, pointing to a dichotomy between the signaling cascades activated by NCAM-FGFR and FGF-FGFR (Cavallaro et al., 2001; Anderson et al., 2005) (C.F. and U.C., unpublished). In addition, NCAM did not promote apoptosis of L cells (supplementary material Fig. S4B), confirming its specific role in the negative regulation of FGF-stimulated cell proliferation. Finally, the proliferative response to FGF2 was completely rescued in L-ΔFN2 cells (Fig. 3B), indicating that the association of NCAM with FGFR is required for the repression of FGF2 function.

Fig. 3.

NCAM is sufficient for the regulation of the cellular response to FGF. (A) L cells stably transfected with empty vector (L-mock), full-length NCAM or ΔFN2 were stimulated for 10 minutes with FGF2, followed by cell lysis and immunoblotting for phospho-ERK1/2 and then for total ERK1/2. (B) L cells stably transfected with empty vector, full-length NCAM or ΔFN2 were left untreated or were treated with FGF2 for the indicated time periods. Cell proliferation was assessed as described in the Materials and Methods. The growth curves of untreated L-NCAM and L-ΔFN2 cells, which were very similar to that of untreated L-mock cells (see also supplementary material Fig. S4A), have been omitted for simplicity. *P<0.005, relative to untreated L-mock cells.

Fig. 3.

NCAM is sufficient for the regulation of the cellular response to FGF. (A) L cells stably transfected with empty vector (L-mock), full-length NCAM or ΔFN2 were stimulated for 10 minutes with FGF2, followed by cell lysis and immunoblotting for phospho-ERK1/2 and then for total ERK1/2. (B) L cells stably transfected with empty vector, full-length NCAM or ΔFN2 were left untreated or were treated with FGF2 for the indicated time periods. Cell proliferation was assessed as described in the Materials and Methods. The growth curves of untreated L-NCAM and L-ΔFN2 cells, which were very similar to that of untreated L-mock cells (see also supplementary material Fig. S4A), have been omitted for simplicity. *P<0.005, relative to untreated L-mock cells.

We previously reported that FGF induces cell-matrix adhesion (Cavallaro et al., 2001). Here, we determined whether this property of FGF was also regulated by NCAM. Indeed, the stimulation of cell-matrix adhesion by FGF2 was readily inhibited by full-length but not ΔFN2-NCAM (supplementary material Fig. S5), implicating NCAM in the control of several cellular events induced by FGF.

In summary, our studies on NIH-3T3 and L cells imply that NCAM regulates FGF-induced ERK1/2 activation and cell proliferation, and that this function is strictly dependent on the NCAM-FGFR interaction.

NCAM prevents the binding of FGF to FGFR

To gain insights into the molecular mechanisms underlying the inhibitory effect of NCAM on FGF function, we determined whether NCAM altered the expression and/or surface exposure of FGFR in L cells. However, immunoblotting of total cell extracts and cell-surface biotinylation studies revealed that neither the total amount of FGFR1, the only FGFR detectable in L cells (not shown), nor its surface exposure were affected by the expression of NCAM in L cells (supplementary material Fig. S6A,B). Next, we tested whether NCAM affects the interaction of FGF with its cell-surface receptors. FGFs bind to two classes of receptors: low-affinity receptor heparan sulfate proteoglycans (HSPGs) and high-affinity FGFRs (Eswarakumar et al., 2005). Binding assays with 125I-FGF2 showed that NCAM does not affect the interaction of FGF with HSPGs (Fig. 4A). Indeed, the non-linear regression analysis of the data revealed no major changes either in the affinity of 125I-FGF2 for HPSG in both L-mock and L-NCAM cells (Kd ∼60 and 68 nM, respectively), or in the number of low-affinity binding sites (∼6.4×106 and 6.2×106, respectively). By contrast, the binding of FGF to FGFR was dramatically reduced in NCAM-expressing L cells as compared with mock-transfected cells (Fig. 4B). This reduction was reflected both by an increase in the Kd from ∼0.36 to 0.76 nM and by a decrease of available high-affinity binding sites from ∼5.8×104 to 1.8×104. The abrogation of FGF binding in NCAM-expressing cells was unlikely to depend on a ligand sequestration mechanism, because NCAM does not interact with FGF2 (supplementary material Fig. S7).

The inhibitory effect of NCAM on the receptor binding of 125I-FGF2 might be caused by steric hindrance. To evaluate this possibility, we used an NCAM-derived peptide, called FGL, previously reported to mimic the interaction of the FNIII repeats of NCAM with FGFR (Kiselyov et al., 2003). Because of its small size, FGL was not expected to cause any significant steric hindrance and, therefore, was tested for its ability to recapitulate the negative effect of full-length NCAM on FGF2 binding to FGFR. Indeed, FGL, but not its mutated version FGLmut, inhibited the binding of 125I-FGF2 to FGFR in a dose-dependent manner (Fig. 4C). Taken together, these results show that the physical association of NCAM with FGFR prevents FGF2 from interacting with its high-affinity receptor.

FGL-mediated inhibition of FGF binding to its receptor also resulted in the repression of FGF-induced cell proliferation (Fig. 5A), indicating that the regulatory role of NCAM on the FGF-FGFR interaction is biologically relevant. Because these findings raised the possibility that FGL inhibited the binding of FGF2 to its receptor in a competitive manner, we verified this hypothesis by an approach previously employed in a similar experimental setting (Presta et al., 1991). Briefly, we measured the proliferation of L cells incubated with increasing concentrations of FGL in the presence of increasing concentrations of FGF2. The inhibitory activity of FGL decreased with increasing concentrations of FGF2 (Fig. 5B), indicating a competitive mechanism by analogy to other negative regulators of FGF function (Presta et al., 1991). Finally, the inhibitory function of FGL on FGF2 activity was further confirmed by the repression of FGF2-induced activation of ERK1/2 (Fig. 5C).

Fig. 4.

NCAM prevents the binding of FGF to FGFR. (A,B) L cells stably transfected with an empty vector (L-mock) or with NCAM were subjected to the 125I-FGF2 binding assay. Specific binding to low-affinity HSPGs (A) and to high-affinity FGFR (B) was determined as described in the Materials and Methods. (C) Non-transfected L cells were subjected to binding assay with 20 ng/ml 125I-FGF2 in the presence of either FGL or FGLmut at the indicated concentrations. Specific binding to high-affinity FGFR was determined as described in the Materials and Methods, and is represented as the percentage of 125I-FGF2 binding in the absence of peptides. *P<0.005, relative to cells incubated with FGLmut.

Fig. 4.

NCAM prevents the binding of FGF to FGFR. (A,B) L cells stably transfected with an empty vector (L-mock) or with NCAM were subjected to the 125I-FGF2 binding assay. Specific binding to low-affinity HSPGs (A) and to high-affinity FGFR (B) was determined as described in the Materials and Methods. (C) Non-transfected L cells were subjected to binding assay with 20 ng/ml 125I-FGF2 in the presence of either FGL or FGLmut at the indicated concentrations. Specific binding to high-affinity FGFR was determined as described in the Materials and Methods, and is represented as the percentage of 125I-FGF2 binding in the absence of peptides. *P<0.005, relative to cells incubated with FGLmut.

Fig. 5.

The FGL peptide inhibits FGF-induced cell proliferation in a competitive manner. (A) L cells were left untreated or were treated for 0-96 hours with 20 ng/ml FGF2, either without peptides or in the presence of FGL or FGLmut at 50 μg/ml. Cell proliferation was assessed as described in the Materials and Methods. *P<0.005, relative to FGF2-treated cells. (B) L cells were stimulated for 48 hours with FGF2 at 10, 20, 50 or 100 ng/ml in the presence of 10, 20, 50 or 100 μg/ml FGL. For each curve (i.e. for each FGF2 concentration), the data are presented as the percentage of mitogenic activity of FGF2 in the presence of FGL in respect to the mitogenic activity of FGF2 alone. (C) L cells were left untreated or were stimulated for 10 minutes with FGF2, either in the absence or presence of 50 μg/ml FGL, followed by cell lysis and immunoblotting for phospho-ERK1/2 and then for total ERK1/2.

Fig. 5.

The FGL peptide inhibits FGF-induced cell proliferation in a competitive manner. (A) L cells were left untreated or were treated for 0-96 hours with 20 ng/ml FGF2, either without peptides or in the presence of FGL or FGLmut at 50 μg/ml. Cell proliferation was assessed as described in the Materials and Methods. *P<0.005, relative to FGF2-treated cells. (B) L cells were stimulated for 48 hours with FGF2 at 10, 20, 50 or 100 ng/ml in the presence of 10, 20, 50 or 100 μg/ml FGL. For each curve (i.e. for each FGF2 concentration), the data are presented as the percentage of mitogenic activity of FGF2 in the presence of FGL in respect to the mitogenic activity of FGF2 alone. (C) L cells were left untreated or were stimulated for 10 minutes with FGF2, either in the absence or presence of 50 μg/ml FGL, followed by cell lysis and immunoblotting for phospho-ERK1/2 and then for total ERK1/2.

In summary, NCAM interferes with the binding of FGFR to FGF in a competitive fashion, thus exerting a negative regulation on the cellular response to FGF.

Discussion

Although NCAM has long been known to control various functions in the nervous systems (Hinsby et al., 2004), its role in non-neuronal tissues has remained elusive. We have previously shown that the loss of NCAM in pancreatic beta cell tumors resulted in tissue disaggregation and tumor-cell detachment, highlighting a role for NCAM in regulating tissue architecture (Cavallaro et al., 2001). Further studies clarified that NCAM is able to induce β1-integrin-mediated cell-matrix adhesion via FGFR signaling. These findings led us to the identification of a novel signaling complex in which NCAM associates with FGFR (Cavallaro et al., 2001). Such a complex was subsequently described also in Jurkat T cells (Kos and Chin, 2002). Finally, the association of NCAM with FGFR was shown to occur via the two membrane-proximal FNIII repeats of NCAM (Kiselyov et al., 2003), whereas NCAM homophilic interactions involve the membrane-distal Ig1-2-3 domains (Soroka et al., 2003). In previous studies, the endogenous NCAM-FGFR complex was detected by co-immunoprecipitation analyses (Cavallaro et al., 2001; Kos and Chin, 2002). Alternatively, either NCAM-transfected cells (Sanchez-Heras et al., 2006) or recombinant fragments of the two proteins (Kiselyov et al., 2003) were employed to document the association between NCAM and FGFR. We have extended these observations to endogenous NCAM and FGFR in NIH-3T3 cells, in which the TIRF analysis supported the notion that the two proteins form a complex at the cell surface. The results of the TIRF analysis on single cells point to a cis interaction between NCAM and FGFR, as suggested by previous colocalization studies on transfected cells (Cavallaro et al., 2001; Sanchez-Heras et al., 2006).

Previous studies have supported the hypothesis that NCAM and other neuronal adhesion molecules, such as N-cadherin and L1CAM, can activate FGFR signaling in an FGF-independent manner (Williams et al., 1994). However, whether the association of NCAM with FGFR has any impact on the FGF-dependent stimulation of the receptor has remained elusive. Our studies demonstrate that NCAM exerts a negative regulation on FGF-induced signaling activities. This repressive function is dependent on the NCAM-FGFR interaction, resulting in the inhibition of the binding of FGF to FGFR.

The interaction of FGF with cell-surface HSPGs, which function as high-avidity low-affinity receptors for FGF, enhances the binding of FGF to high-affinity FGFR (Mohammadi et al., 2005). Hence, based on the notion that NCAM can also interact with HSPGs (Storms et al., 1996), it is possible that NCAM interferes with FGF activity by perturbing its binding to HSPGs. However, a number of observations from this and other reports do not support this hypothesis. First, NCAM specifically reduces the binding of FGF to high-affinity FGFR, whereas the interaction between FGF and HSPGs is not affected (see Fig. 4). Second, heparin fails to promote FGF sequestration by NCAM (see supplementary material Fig. S7). Third, the HSPG-binding domain of NCAM lies in the second Ig loop (Cole and Akeson, 1989), distant from the second FNIII repeat that is involved in the inhibition of FGF binding to the cell surface. Fourth, it is unlikely that a small molecule such as FGL, which binds to FGFR and prevents FGF binding, is able to bind HSPGs at the same time. Rather, the observations that (a) the FGL peptide alone is sufficient to inhibit the binding of FGF to its receptor, and (b) the FGL peptide shares striking sequence and structure similarities with a loop region of FGF2 (Kiselyov et al., 2003), support the hypothesis that NCAM and FGF compete for the same binding site on FGFR. This, of course, does not rule out the possibility that HSPGs might play a role in the inhibitory function of NCAM on FGF-induced cellular responses, an issue that is out of the scope of this study.

The functional crosstalk between adhesion molecules and receptor tyrosine kinases has been reported in various experimental systems, with adhesion molecules exerting a regulatory function on the growth-factor-dependent activation of the receptor (Walker et al., 2005). With regards to FGFR signaling, N-cadherin potentiates FGF activity on breast cancer cells by stabilizing the exposure of FGFR1 at the cell surface (Suyama et al., 2002), whereas a PECAM1–VE-cadherin (CDH5) complex in endothelial cells regulates FGF-induced leukocyte diapedesis (Halama et al., 2001). E-cadherin, by contrast, counteracts ligand-induced FGFR1 signaling and trafficking (Bryant et al., 2005). However, our studies provide the first demonstration that the competitive inhibition of ligand binding by an adhesion molecule can account for the repression of FGF-dependent activation of FGFR. This implies that other adhesion molecules might use a similar mechanism to regulate the ligand-dependent function of different receptor tyrosine kinases.

Our findings support the previous observation that NCAM represses FGF-stimulated cell proliferation in cells of the central nervous system (Amoureux et al., 2000; Krushel et al., 1998) and extend this regulatory function of NCAM not only to the proliferation of non-neuronal cell types (i.e. fibroblasts) but also to other FGF-induced processes, including cell-matrix adhesion. Different cell types exhibit different responses to FGF, and the control mechanisms underlying this variability are assumed to occur mostly intracellularly (Dailey et al., 2005). Our data implicate NCAM as a novel and important membrane regulator of FGF signaling, adding a further level of complexity to the modulation of FGFR activity. Aberrant expression and/or function of NCAM have been observed in several pathological conditions, ranging from neurological to neoplastic diseases (Cavallaro and Christofori, 2004; Mikkonen et al., 2001; Vawter, 2000). In addition, Ncam-null mice exhibit significant developmental and behavioral defects (Cremer et al., 1997; Cremer et al., 1994; Stork et al., 1997; Stork et al., 1999). Thus far, the pathogenic role of NCAM in these disorders has been attributed to the deregulation of its adhesive properties. However, based on the data presented here, changes in ligand-induced FGFR signaling need to be considered as an additional consequence of altered NCAM functions and investigated as a possible pathogenic factor. For example, excessive FGFR signaling in certain tumors induces cancer-cell proliferation, survival and invasion, as well as angiogenesis and metastasis (Grose and Dickson, 2005), and the loss of NCAM has been reported in various tumor types (Fogar et al., 1997; Huerta et al., 2001; Sasaki et al., 1998; Tezel et al., 2001). Thus, it is tempting to speculate that altered levels of NCAM might be causally implicated in the aberrant FGFR function at least in certain neoplasms.

In summary, we have shown that NCAM modulates the cellular response to FGF stimulation. Future studies should address the pathophysiological implications of these findings. In particular, it remains to be determined whether the regulatory role of NCAM on FGFR activity can be exploited to develop novel targeted therapeutic approaches for diseases caused by aberrant FGFR signaling.

Materials and Methods

Reagents

The following commercial reagents were used: FGF2 (Peprotech, London, UK); heparin and PD98059 (Sigma, St Louis, MO); staurosporine (Biomol, Plymouth Meeting, PA); peroxidase-conjugated streptavidin (Jackson ImmunoResearch Laboratories, West Grove, PA). Antibodies: mouse anti-phospho-Erk1/2 (Thr202/Tyr204), rabbit anti-phospho-Shc (Tyr239/240) and anti-Shc (Cell Signaling Technology, Danvers, MA), rabbit anti-Erk1/2 (Sigma) and anti-FGFR1 (C15; Santa Cruz Biotechnologies), mouse anti-NCAM (clone NCAM-13; BD Biosciences, San Jose, CA). The myc-tagged scFv against FGFR1 was isolated by screening the ETH-2 Gold library (Silacci et al., 2005) against recombinant FGFR1-Fc (R&D Systems, Minneapolis, MN). The monoclonal antibody 9E10 (Santa Cruz Biotechnologies) was used to detect the bound scFv anti-FGFR1 on western blots. The FGL peptide from the second FNIII module of NCAM and its mutated version FGLmut that carries two alanine substitutions, which abolish its binding to FGFR (Kiselyov et al., 2003), were a generous gift from ENKAM Pharmaceuticals (Copenhagen, Denmark).

Cell lines and transfection

Mouse fibroblastic L cells and NIH-3T3 cells were maintained in DMEM, 10% fetal calf serum, L-glutamine and antibiotics. L cells were transfected with pcDNA3.1 alone (Invitrogen) or containing the cDNA for mouse NCAM140 or NCAM140-ΔFN2, using Lipofectamine 2000. Stable transfectants were obtained by selection with 0.8 mg/ml G418 (Invitrogen) and cloning by limiting dilution. NIH-3T3 cells were transfected with the following Stealth siRNA duplexes from Invitrogen using the RNAiMAX reagent and following the manufacturer's instructions: siNCAM1 (5′-GAACUGCAGUUUCCCUGCAGGUAGA-3′; 3′-UCUACCUGCAGGGAAACUGCAGUUC-5′), siNCAM3 (5′-CCAUCAGACACUAUCUGGUCAAGUA-3′; 3′-UACUUGACCAGAUAGUGUCUGAUGG-5′), and a control siRNA targeting human NCAM (5′-UUCACUGCAGUACAGUUGUAGUUCC-3′; 3′-GGAACUACAACUGUACUGCAGUGAA-5′).

Cell stimulation, protein extraction and western blot

Cells were cultured in six-well plates in DMEM with 10% FCS, followed by serum starvation overnight. Cells were stimulated for 10 minutes with 20 ng/ml FGF2 and then lysed in lysis buffer (20 mM Tris/HCl pH 8.0, 160 mM NaCl, 1 mM CaCl2, 10 μg/ml aprotinin, 1% Triton X-100, 1 μg/ml leupeptin, 1 mM PMSF, 10 mM NaF and 1 mM sodium orthovanadate). Following sonication and centrifugation to remove cell debris, the protein concentration of cell lysates was determined using the Bio-Rad DC Protein Assay (Bio-Rad, Hercules, CA). Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes (Protran, Biosciences). Proteins of interest were visualized using specific antibodies, followed by peroxidase-conjugated secondary antibodies and by an enhanced chemiluminescence kit (Amersham Biosciences, Little Chalfort, UK). Each experiment was repeated at least five times.

Immunofluorescence and TIRF analysis

Sub-confluent NIH-3T3 cells cultured on coverslips were co-stained with mouse anti-NCAM and rabbit anti-FGFR1 antibodies for 45 minutes at 37°C, followed by incubation with anti-mouse and anti-rabbit secondary antibodies conjugated to Alexa-Fluor-568 and -488 (Molecular Probes), respectively. Cells were then fixed in PBS, 3% paraformaldehyde, 2% sucrose. TIRF imaging of cells was performed with an Olympus BioSystems TIRF workstation based on CellR Imaging System as described previously (Tosoni et al., 2005).

Cell-proliferation assays

NIH-3T3 or L cells were seeded on 24-well plates at 7×103 cells/well, serum-starved overnight and treated for 1-4 days with 5 or 20 ng/ml FGF2, respectively, or with 20 μg/ml FGL or FGLmut, replenished every 24 hours. At each time point, viable cells were counted using the Trypan blue exclusion method and the ratio with non-stimulated cells at time 0 was determined. Values represent the means ± s.e.m. of the results from at least three independent experiments, each performed in quadruplicate.

To assay for competitive inhibition of FGF-induced cell proliferation, L cells were treated with increasing concentrations of FGF2 in the presence of increasing concentrations of FGL. After 48 hours, cell proliferation was measured as above.

125I-FGF-binding assays

Binding of 125I-FGF2 (Amersham Biosciences) to L cells was measured as described previously (Moscatelli, 1988). Briefly, cells in 24-well plates were washed with binding buffer (DMEM, 25 mM HEPES, 0.05% gelatin) at 4°C. Increasing concentrations of 125I-FGF2 (0.25, 0.5, 1, 2 and 5 nM) were added and the cells were incubated at 4°C for 4 hours. Non-specific binding was determined in the presence of a 500-fold molar excess of unlabelled FGF2. To determine the low-affinity binding of 125I-FGF2 to HSPGs, cells were washed with HEPES buffer, pH 7.4, containing 2 M NaCl and the radioactivity of the washes was measured. Subsequently, the high-affinity binding to FGFR was determined by washing the cells with 20 mM sodium acetate pH 4, 2 M NaCl and measuring the radioactivity in these low-pH washes. The Kd values and the numbers of binding sites were calculated by the non-linear regression analysis of the data, which was performed with the software GraphPad Prism 4 (GraphPad Software, San Diego, CA).

When needed, the binding assays were performed in the presence of either FGL or FGLmut at the indicated concentrations. Values represent the means ± s.d. of representative experiments performed in triplicate. The assays were repeated at least three times.

Acknowledgements

We thank P. P. Di Fiore and E. Dejana (IFOM, Milan) for comments on the manuscript. We gratefully acknowledge ENKAM Pharmaceuticals (Copenhagen, Denmark) for the FGL and FGLmut peptides, and V. Berezin and E. Bock (University of Copenhagen) for advice on their use. We thank F. Lehembre and F. Brawand (University of Basel) for advice on NCAM knockdown and for input and technical support, respectively. We are grateful to M. Presta and D. Leali (University of Brescia) for advice and discussion, and to D. Parazzoli (IFOM, Milan) for the assistance with the TIRF analysis. This work was supported by grants from the Association for International Cancer Research (to U.C.), the Associazione Italiana per la Ricerca sul Cancro (to U.C.), the Krebsliga Beider Basel (to G.C.), the EU-FP6 framework programmes LYMPHANGIOGENOMICS LSHG-CT-2004-503573 and BRECOSM LSHC-CT-2004-503224 (to G.C.) and the Swiss National Science Foundation (to U.C. and G.C.).

References

Amoureux, M. C., Cunningham, B. A., Edelman, G. M. and Crossin, K. L. (
2000
). N-CAM binding inhibits the proliferation of hippocampal progenitor cells and promotes their differentiation to a neuronal phenotype.
J. Neurosci.
20
,
3631
-3640.
Anderson, A. A., Kendal, C. E., Garcia-Maya, M., Kenny, A. V., Morris-Triggs, S. A., Wu, T., Reynolds, R., Hohenester, E. and Saffell, J. L. (
2005
). A peptide from the first fibronectin domain of NCAM acts as an inverse agonist and stimulates FGF receptor activation, neurite outgrowth and survival.
J. Neurochem.
95
,
570
-583.
Bryant, D. M., Wylie, F. G. and Stow, J. L. (
2005
). Regulation of endocytosis, nuclear translocation, and signaling of fibroblast growth factor receptor 1 by E-cadherin.
Mol. Biol. Cell
16
,
14
-23.
Cavallaro, U. and Christofori, G. (
2004
). Cell adhesion and signalling by cadherins and Ig-CAMs in cancer.
Nat. Rev. Cancer
4
,
118
-132.
Cavallaro, U., Niedermeyer, J., Fuxa, M. and Christofori, G. (
2001
). NCAM modulates tumour-cell adhesion to matrix by inducing FGF-receptor signalling.
Nat. Cell Biol.
3
,
650
-657.
Christensen, C., Lauridsen, J. B., Berezin, V., Bock, E. and Kiselyov, V. V. (
2006
). The neural cell adhesion molecule binds to fibroblast growth factor receptor 2.
FEBS Lett.
580
,
3386
-3390.
Cole, G. J. and Akeson, R. (
1989
). Identification of a heparin binding domain of the neural cell adhesion molecule N-CAM using synthetic peptides.
Neuron
2
,
1157
-1165.
Cremer, H., Lange, R., Christoph, A., Plomann, M., Vopper, G., Roes, J., Brown, R., Baldwin, S., Kraemer, P., Scheff, S. et al. (
1994
). Inactivation of the N-CAM gene in mice results in size reduction of the olfactory bulb and deficits in spatial learning.
Nature
367
,
455
-459.
Cremer, H., Chazal, G., Goridis, C. and Represa, A. (
1997
). NCAM is essential for axonal growth and fasciculation in the hippocampus.
Mol. Cell. Neurosci.
8
,
323
-335.
Crnic, I., Strittmatter, K., Cavallaro, U., Kopfstein, L., Jussila, L., Alitalo, K. and Christofori, G. (
2004
). Loss of neural cell adhesion molecule induces tumor metastasis by up-regulating lymphangiogenesis.
Cancer Res.
64
,
8630
-8638.
Dailey, L., Ambrosetti, D., Mansukhani, A. and Basilico, C. (
2005
). Mechanisms underlying differential responses to FGF signaling.
Cytokine Growth Factor Rev.
16
,
233
-247.
Eswarakumar, V. P., Lax, I. and Schlessinger, J. (
2005
). Cellular signaling by fibroblast growth factor receptors.
Cytokine Growth Factor Rev.
16
,
139
-149.
Fogar, P., Basso, D., Pasquali, C., De Paoli, M., Sperti, C., Roveroni, G., Pedrazzoli, S. and Plebani, M. (
1997
). Neural cell adhesion molecule (N-CAM) in gastrointestinal neoplasias.
Anticancer Res.
17
,
1227
-1230.
Grose, R. and Dickson, C. (
2005
). Fibroblast growth factor signaling in tumorigenesis.
Cytokine Growth Factor Rev.
16
,
179
-186.
Halama, T., Groger, M., Pillinger, M., Staffler, G., Prager, E., Stockinger, H., Holnthoner, W., Lechleitner, S., Wolff, K. and Petzelbauer, P. (
2001
). Platelet endothelial cell adhesion molecule-1 and vascular endothelial cadherin cooperatively regulate fibroblast growth factor-induced modulations of adherens junction functions.
J. Invest. Dermatol.
116
,
110
-117.
Hinsby, A. M., Berezin, V. and Bock, E. (
2004
). Molecular mechanisms of NCAM function.
Front. Biosci.
9
,
2227
-2244.
Huerta, S., Srivatsan, E. S., Venkatesan, N., Peters, J., Moatamed, F., Renner, S. and Livingston, E. H. (
2001
). Alternative mRNA splicing in colon cancer causes loss of expression of neural cell adhesion molecule.
Surgery
130
,
834
-843.
Kiselyov, V. V., Skladchikova, G., Hinsby, A. M., Jensen, P. H., Kulahin, N., Soroka, V., Pedersen, N., Tsetlin, V., Poulsen, F. M., Berezin, V. et al. (
2003
). Structural basis for a direct interaction between FGFR1 and NCAM and evidence for a regulatory role of ATP.
Structure
11
,
691
-701.
Kos, F. J. and Chin, C. S. (
2002
). Costimulation of T cell receptor-triggered IL-2 production by Jurkat T cells via fibroblast growth factor receptor 1 upon its engagement by CD56.
Immunol. Cell Biol.
80
,
364
-369.
Krushel, L. A., Tai, M. H., Cunningham, B. A., Edelman, G. M. and Crossin, K. L. (
1998
). Neural cell adhesion molecule (N-CAM) domains and intracellular signaling pathways involved in the inhibition of astrocyte proliferation.
Proc. Natl. Acad. Sci. USA
95
,
2592
-2596.
Mikkonen, M., Soininen, H., Alafuzof, I. and Miettinen, R. (
2001
). Hippocampal plasticity in Alzheimer's disease.
Rev. Neurosci.
12
,
311
-325.
Mohammadi, M., Dikic, I., Sorokin, A., Burgess, W. H., Jaye, M. and Schlessinger, J. (
1996
). Identification of six novel autophosphorylation sites on fibroblast growth factor receptor 1 and elucidation of their importance in receptor activation and signal transduction.
Mol. Cell. Biol.
16
,
977
-989.
Mohammadi, M., Olsen, S. K. and Ibrahimi, O. A. (
2005
). Structural basis for fibroblast growth factor receptor activation.
Cytokine Growth Factor Rev.
16
,
107
-137.
Moscatelli, D. (
1988
). Metabolism of receptor-bound and matrix-bound basic fibroblast growth factor by bovine capillary endothelial cells.
J. Cell Biol.
107
,
753
-759.
Perl, A.-K., Dahl, U., Wilgenbus, P., Cremer, H., Semb, H. and Christofori, G. (
1999
). Reduced expression of neural cell adhesion molecule induces metastatic dissemination of pancreatic [beta] tumor cells.
Nat. Med.
5
,
286
-291.
Petersen, L. G., Storling, J., Heding, P., Li, S., Berezin, V., Saldeen, J., Billestrup, N., Bock, E. and Mandrup-Poulsen, T. (
2006
). IL-1beta-induced pro-apoptotic signalling is facilitated by NCAM/FGF receptor signalling and inhibited by the C3d ligand in the INS-1E rat beta cell line.
Diabetologia
49
,
1864
-1875.
Piehler, J. (
2005
). New methodologies for measuring protein interactions in vivo and in vitro.
Curr. Opin. Struct. Biol.
15
,
4
-14.
Presta, M., Rusnati, M., Urbinati, C., Sommer, A. and Ragnotti, G. (
1991
). Biologically active synthetic fragments of human basic fibroblast growth factor (bFGF): identification of two Asp-Gly-Arg-containing domains involved in the mitogenic activity of bFGF in endothelial cells.
J. Cell. Physiol.
149
,
512
-524.
Sanchez-Heras, E., Howell, F. V., Williams, G. and Doherty, P. (
2006
). The FGF receptor acid box is essential for interactions with N-cadherin and all of the major isoforms of NCAM.
J. Biol. Chem.
281
,
35208
-35216.
Sasaki, H., Yoshida, K., Ikeda, E., Asou, H., Inaba, M., Otani, M. and Kawase, T. (
1998
). Expression of the neural cell adhesion molecule in astrocytic tumors: an inverse correlation with malignancy.
Cancer
82
,
1921
-1931.
Schneckenburger, H. (
2005
). Total internal reflection fluorescence microscopy: technical innovations and novel applications.
Curr. Opin. Biotechnol.
16
,
13
-18.
Silacci, M., Brack, S., Schirru, G., Marlind, J., Ettorre, A., Merlo, A., Viti, F. and Neri, D. (
2005
). Design, construction, and characterization of a large synthetic human antibody phage display library.
Proteomics
5
,
2340
-2350.
Soroka, V., Kolkova, K., Kastrup, J. S., Diederichs, K., Breed, J., Kiselyov, V. V., Poulsen, F. M., Larsen, I. K., Welte, W., Berezin, V. et al. (
2003
). Structure and interactions of NCAM Ig1-2-3 suggest a novel zipper mechanism for homophilic adhesion.
Structure
11
,
1291
-1301.
Stork, O., Welzl, H., Cremer, H. and Schachner, M. (
1997
). Increased intermale aggression and neuroendocrine response in mice deficient for the neural cell adhesion molecule (NCAM).
Eur. J. Neurosci.
9
,
1117
-1125.
Stork, O., Welzl, H., Wotjak, C. T., Hoyer, D., Delling, M., Cremer, H. and Schachner, M. (
1999
). Anxiety and increased 5-HT1A receptor response in NCAM null mutant mice.
J. Neurobiol.
40
,
343
-355.
Storms, S. D., Anvekar, V. M., Adams, L. D. and Murray, B. A. (
1996
). Heterophilic NCAM-mediated cell adhesion to proteoglycans from chick embryonic brain membranes.
Exp. Cell Res.
223
,
385
-394.
Suyama, K., Shapiro, I., Guttman, M. and Hazan, R. B. (
2002
). A signaling pathway leading to metastasis is controlled by N-cadherin and the FGF receptor.
Cancer Cell
2
,
301
-314.
Tezel, E., Kawase, Y., Takeda, S., Oshima, K. and Nakao, A. (
2001
). Expression of neural cell adhesion molecule in pancreatic cancer.
Pancreas
22
,
122
-125.
Tosoni, D., Puri, C., Confalonieri, S., Salcini, A. E., De Camilli, P., Tacchetti, C. and Di Fiore, P. P. (
2005
). TTP specifically regulates the internalization of the transferrin receptor.
Cell
123
,
875
-888.
Vawter, M. P. (
2000
). Dysregulation of the neural cell adhesion molecule and neuropsychiatric disorders.
Eur. J. Pharmacol.
405
,
385
-395.
Walker, J. L., Fournier, A. K. and Assoian, R. K. (
2005
). Regulation of growth factor signaling and cell cycle progression by cell adhesion and adhesion-dependent changes in cellular tension.
Cytokine Growth Factor Rev.
16
,
395
-405.
Walmod, P. S., Kolkova, K., Berezin, V. and Bock, E. (
2004
). Zippers make signals: NCAM-mediated molecular interactions and signal transduction.
Neurochem. Res.
29
,
2015
-2035.
Walsh, F. S. and Doherty, P. (
1997
). Neural cell adhesion molecules of the immunoglobulin superfamily: role in axon growth and guidance.
Annu. Rev. Cell Dev. Biol.
13
,
425
-456.
Williams, E. J., Furness, J., Walsh, F. S. and Doherty, P. (
1994
). Activation of the FGF receptor underlies neurite outgrowth stimulated by L1, N-CAM, and N-cadherin.
Neuron
13
,
583
-594.
Xian, X., Hakansson, J., Stahlberg, A., Lindblom, P., Betsholtz, C., Gerhardt, H. and Semb, H. (
2006
). Pericytes limit tumor cell metastasis.
J. Clin. Invest.
116
,
642
-651.

Supplementary information