R-Cadherin Expression Inhibits Myogenesis and Induces Myoblast Transformation via Rac1 GTPase

Cadherins are homophilic cell-cell adhesion molecules essential for the organization of cells into tissues during embryonic development. They are also involved in cell growth, migration, and differentiation. Cadherins are highly conserved transmembrane glycoproteins that mediate homotypic cell-cell adhesion through their extracellular domain. The cadherin cytoplasmic domains provide filamentous actin (F-actin) cytoskeleton attachment points via association of catenins and other cytoskeletal-associated proteins (1). These adhesive receptors regulate diverse functions, beyond the basic adhesive process, including intracellular signaling events (2). In particular, cadherins activate Rho family GTPases and participate in receptor tyrosine kinase signaling (3). Developing skeletal muscles express N-cadherin, M-cadherin, and R-cadherin, and cadherin-dependent adhesion is required for a variety of cellular events. N-cadherin–dependent intercellular adhesion has a major role in the cell cycle exit and in the induction of the skeletal muscle differentiation program (4–7). N-cadherin regulates RhoA and Rac1, two Rho GTPases whose activity needs to be tightly controlled to allow myogenesis induction (8). Indeed, RhoA is activated in an N-cadherin–dependent manner, which allows serum response factor–dependent genes expression and myogenesis induction. In contrast, N-cadherin activation decreases Rac1, which is known to inhibit myogenesis by preventing the withdrawal of myoblasts from the cell cycle (9, 10). During embryonic development, M-cadherin is expressed later than N-cadherin. Its expression increases at the onset of secondary myogenesis, and it has been implicated in terminal myoblast differentiation, mainly during myoblast fusion (11). R-cadherin is detected in somites and limbs, and it is localized at the cell-cell contacts of primary myoblasts during primary myogenesis, as well as at the primary myotube/secondary myoblast boundaries during secondary myogenesis (12). R-cadherin expression is lost in adult muscles and is found expressed in rhabdomyosarcoma (RMS), a highly malignant soft-tissue tumor committed to the myogenic lineage, but arrested before terminal differentiation (13). RMS are divided into two major histologic subtypes, alveolar and embryonal (14); both express R-cadherin. R-cadherin expression in RMS is accompanied by a decrease in the expression of N-cadherin and M-cadherin, which is reminiscent of the cadherin switching observed in tumorigenesis, particularly in metastasis (15, 16). Indeed, cadherin switching is often observed in the epithelium-to-mesenchyme transition during normal embryonic development and when cancer cells invade adjacent tissues (17–19). R-cadherin was first found to be expressed in the chick embryonic retina and, later, was also detected, besides skeletal muscle, in heart, kidney, thymus, and lung during development (12, 20–22). R-cadherin mRNA has been detected in bladder and prostate carcinomas, and its expression in various cell lines down-regulates expression of other cadherins (23–25). Conversely, R-cadherin is silenced by promoter methylation in gastrointestinal tumors (26). This suggests that the cellular context is important for R-cadherin–induced downstream effects.

Here, we show that R-cadherin expression in C2C12 cell line results in myogenesis inhibition and myoblast transformation. On the contrary, inhibition of R-cadherin expression by short interfering RNA (siRNA) decreases the transformation potential of RMS cells. Furthermore, we show that R-cadherin–mediated adhesion increases Rac1 activity and that R-cadherin–induced myoblast transformation is dependent upon Rac1 activation. Finally, we evidence a cadherin switch induced by R-cadherin with down-regulation of the endogenous expression of M-cadherin and delocalization of M-cadherin and N-cadherin processes mediated by Rac1.

Cell culture and immunocytochemistry. C2C12 mouse myoblasts and the embryonic RMS cell line RD were grown as described (13).

Cells growing onto 35-mm dishes were fixed in 3.7% formaldehyde in PBS, followed by a 5-min permeabilization in 0.1% Triton X-100 in PBS, and then incubated in PBS containing 0.1% bovine serum albumin (BSA). Myogenin, troponin T, and N-cadherin expression were detected as described (8, 10). Monoclonal anti-myc (Invitrogen) and anti–cyclin D1 (BD PharMingen) antibodies were used, respectively, at 1:1000 and 1:500. Primary antibodies were revealed with either an Alexa Fluor 546–conjugated or an Alexa Fluor 488–conjugated goat anti-mouse or goat anti-rabbit antibodies (Molecular Probes, Interchim). Cells were stained for F-actin using TRITC-conjugated phalloidin (Sigma). Cells were analyzed as described (8).

The Rac1 inhibitor NSC23766 (Calbiochem) was used at 100 μm and added 30 h before cell harvesting in proliferation medium.

Establishment of stable cell lines. C2C12 were transfected with either pEGFPN1 plasmid (Clontech), in which the R-cadherin cDNA was subcloned, or infected by recombinant retroviruses produced by Phoenix cells transfected with the LZRS-MS-Rcad-2Xmyc plasmid or the LZRS-MS-R(AAA)cad-2Xmyc plasmid (kindly provided by Dr. M. J. Weelock). Stably transfected cells were selected in 1 mg/mL G418 and sorted by fluorescence-activated cell sorting for green fluorescent protein (GFP) expression.

To produce stable RD cell lines in which R-cadherin expression is inhibited, siRNA constructs were made by cloning a double-strand DNA corresponding to the R-cadherin sequence (13) in the retroviral vector pSIREN-RetroQ according to the manufacturer's protocol (BD Biosciences). A second R-cadherin small hairpin RNA was also used. For this, annealed double-strand oligonucleotides GATCCGTCAGAACGTGAAATGCAAAGTTCAAGAGAGACTGTGTACTGCTGAACTCTTTTTTACGCGTG (top) and AATTCACGCGTAAAAAAGAGTTCAGCAGTACACAGTCTCTCTTGAACTTTGCATTTCACGTTCTGACG (bottom) were inserted into the RNAi-Ready pSIREN-RetroQ and the RNAi-Ready pSIREN-RetroQ-ZsGreen vectors (Clontech, BD Biosciences). As a control, we made RD cell lines stably expressing a siRNA for the unrelated gene luciferase. Stable transfectants were selected in medium containing puromycin (2.5 μg/mL), and different clones were isolated by limited dilution.

Immunoblotting and immunoprecipitation. Cell extracts were prepared as described (8). Equal amounts of protein (20–40 μg) were resolved on SDS-PAGE and transferred to polyvinylidene difluoride membranes. Membranes were blocked with 5% milk protein or 1% BSA when an antiphosphorylation antibody was used. Membranes were incubated with monoclonal anti–N-cadherin (1:250), anti–β-catenin (1:500), anti–γ-catenin (1:200), anti–p120 catenin (1:200), anti–cyclin D1 (1:500; all from BD Transduction Laboratories),anti–α-tubulin (1:100, P. Mangeat, France), anti–α-actin (1:500, Sigma), and anti–phosphorylated c-Jun (Ser⁶¹, 1:1,000, Cell Signaling) antibodies, rat monoclonal anti–R-cadherin antibody MRCD5 (1:100, a gift from M. Takeichi, Japan), polyclonal anti–M-cadherin (13), and anti–α-catenin (1:200, Sigma) antibodies. Membranes were processed as described (8).

For immunoprecipitation, C2C12 Rcad/GFP or C2C12 Rcad/2myc cells were lysed as described previously (8). Lysates were immunoprecipitated using a polyclonal anti-GFP (1:250, Torrey Pines) or a monoclonal anti-myc antibody (Invitrogen, 1:1000) followed by sequential incubation with protein A–sepharose or protein G–sepharose (Amersham). The immunoprecipitates were analyzed by immunoblotting.

Bromodeoxyuridine incorporation. When appropriate density was reached, cells were incubated with bromodeoxyuridine (BrdUrd, Boehringer) and processed as described (27).

Luciferase assay. C2C12 cells plated in 35-mm dishes (20,000 cells for subconfluent condition; 50,000 cells for confluent conditions) were cotransfected using the JetPEI method with 0.6 μg of construct containing the cyclin D1 promoter fused to the luciferase gene (28) and 40 ng of pRL cytomegalovirus vector (Renilla luciferase cytomegalovirus). For differentiation medium (DM) conditions, 250,000 cells were cotransfected as soon as they were attached to the plate. Cells were induced to differentiate 4 h after transfection. Forty-eight hours after transfection, luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega).

Plasmid construction and production of the chimeric R-cadherin fused to Fc fragment protein. To construct the cDNA coding for the entire mouse R-cadherin ectodomain linked to the human Fc fragment (Fc), a fragment from nucleotides 1 to 2,156 was amplified by PCR reaction using oligonucleotides flanked by BamHI sites and subcloned into the TOPO vector (Invitrogen). The resulting plasmid was digested with BamHI restriction enzyme, and the full-length ectodomain was subcloned into the BamHI site of the pREP10-Fc plasmid (kindly provided by RM Mège). The production of the chimeric protein corresponding to the ectodomain of R-cadherin fused to Fc fragment (R-cad-Fc) was performed as previously described (8).

Time lapse imaging. C2C12 Rcad/GFP and C2C12 cells were analyzed to calculate their migration (29).

Statistical analysis. For experiments with n ≥ 30, t test was used to assess the statistical differences between two experimental conditions (Figs. 3B and 5B). For experiments where n < 30 (Figs. 2A,, C, D, 3A, C, and 5B), the nonparametric Mann–Whitney U test was used to assess the statistical differences between two experimental conditions. A linear regression was used to analyze the repeated data (Figs. 3A,, D and 5C). The tumor progress is significantly different with regard to the two groups (P = 0.001).

GTPase activity assays. C2C12 myoblasts or RD cells were processed as described to measure Rac1 activity (9).

Mice and xenografting. Female C3H HEN/HSD mice and athymic mice were purchased from Harlan Laboratories and used at 6 to 8 wk of age. All mice were maintained at the animal facility of Centre de Recherche en Cancérologie de Montpellier, CRLC Val d'Aurelle-Paul Lamarque. Experimental procedures and handling were performed in a laminar flow hood for nude mice. For xenografts, 2 × 10⁶ C2C12 cells or 5 × 10⁶ RD cells were injected s.c. into the leg or in the thigh muscle. Tumors were detected by palpation and measured periodically with calipers, and tumor volume was deduced using the formula D1 × D2 × D3/D2, where D1 is the length, D2 is the width, and D3 is the depth of the tumor.

R-cadherin expression inhibits myogenesis induction. To establish the role of R-cadherin in skeletal muscle cells, we made stable mouse C2C12 myoblasts expressing either R-cadherin fused to a GFP-tag (Rcad/GFP) or to a 2myc tag (Rcad/2myc). C2C12 myoblasts expressing GFP alone were used as control. The characterization of these clones is shown in Supplementary Fig. S1.

To gain further insight into the role of R-cadherin in myogenesis, we quantified the differentiation rate of parental, GFP-expressing, and Rcad/GFP-expressing myoblasts. Confluent C2C12 myoblasts were induced to differentiate by replacing the growth medium (GM) with DM and analyzed for myotube formation and expression of muscle-specific proteins. Two days after serum withdrawal (D2), myotubes were present both in parental (data not shown) and GFP-expressing C2C12 cells (Fig. 1A,, c). Numerous myotubes were visible 3 to 4 days after DM addition (d and e). In a sharp contrast, no myotubes were visible in Rcad/GFP-expressing myoblasts (Fig. 1A , bottom) even when differentiation conditions were maintained up to 7 days (data not shown), indicating that myoblast-to-myotube transition is efficiently blocked and not simply delayed. Similar data were observed in a separate Rcad/GFP-expressing clone and in three independent Rcad/2myc-expressing clones demonstrating that these effects were not due to clonal variations or influence of the tag (data not shown).

We next examined whether R-cadherin could affect the expression of myogenin and troponin-T, two muscle-specific proteins. As expected, myogenin and troponin T were detected after 1 and 2 days in DM, respectively, in parental and GFP-expressing C2C12 myoblasts, whereas C2C12 Rcad/GFP cells faintly expressed these proteins (Fig. 1B and C). Taken together, these data suggest that R-cadherin expression blocks myogenesis induction.

R-cadherin expression inhibits myoblast cell cycle exit. Because myogenic differentiation and cell cycle progression are mutually exclusive events (30), we asked whether the failure of R-cadherin expressing C2C12 myoblasts to differentiate could be ascribed to a high proliferative rate retained in the differentiation conditions. No obvious differences in the proliferation rate of control and Rcad/GFP-expressing cells were observed in subconfluent cells maintained in GM. In contrast, confluent Rcad/GFP-expressing C2C12 myoblasts proliferated significantly more than control cells cultured at similar density (Fig. 2A). This effect was even sharper when cells were induced to differentiate demonstrating that R-cadherin acts positively on cell cycle progression, bypassing not only cell contact inhibition but also cell cycle exit induced by growth factor deprivation. Cyclins are key factors for cell cycle progression. Cyclin D1 plays a critical role by regulating the transit through the G₁ phase of the cell cycle, and its expression markedly decreases during the C2C12 myoblast differentiation process (31). Cyclin D1 was normally expressed both in proliferating control and Rcad/GFP-expressing C2C12 myoblasts and was reduced in control cells induced to differentiate (Fig. 2B). In contrast, cyclin D1 expression was maintained in C2C12 Rcad/GFP cells cultured in DM. The maintenance of cyclin D1 expression in Rcad/GFP-expressing cells after DM addition was confirmed also by immunocytochemistry (Fig. 2C). To test whether sustained cyclin D1 expression was due to transcriptional regulation of the promoter, we performed luciferase assays in which parental and Rcad/GFP-expressing C2C12 were transfected with the cyclin D1 promoter fused to the luciferase reporter gene. In proliferative conditions, the level of cyclin D1 activation was comparable between control and C2C12 Rcad/GFP cells (Fig. 2D). However, upon differentiation, this level of activity was maintained only in R-cadherin–expressing C2C12 myoblasts, whereas in control cells the activity of the promoter diminished. Next, we examined the level of cyclin A expression, a cyclin that acts during S phase, which is rather linked to proliferation (32). Cyclin A was similarly maintained at a high level in C2C12 Rcad/GFP-expressing myoblasts upon differentiation conditions while down-regulated in control cells (Fig. 2B).

R-cadherin expression induces myoblast transformation and increases cell motility. Because R-cadherin induces the loss of cell contact inhibition and expression of cyclin D1, two characteristics of tumor cells (33), we asked whether R-cadherin might be sufficient to induce myoblast transformation. Therefore, we performed soft agar assay to analyze whether R-cadherin expression might affect anchorage-independent cell growth (Fig. 3A,, left). C2C12 Rcad/GFP cells were able to form colonies in soft agar, whereas parental C2C12 did not. To estimate the ability of R-cadherin to induce myoblast transformation in vivo, we transplanted either s.c. or in thigh muscles GFP-expressing and Rcad/GFP-expressing C2C12 cells in mice. We monitored the appearance of tumors by palpation, and we measured the size of the tumors formed. Two months after injection, the mice that received two different C2C12 Rcad/GFP clones had measurable tumors, whereas mice injected with control C212 GFP cells did not (Fig. 3A,, right). To assess if R-cadherin expression on C2C12 cells could influence their migration as it was already reported for breast tumor cells (23), C2C12 cells that stably expressed GFP alone or R-cadherin–GFP were seeded at a low density and random cell movement was recorded by phase-contrast microscopy for 4 h (Fig. 3B). Rcad/GFP-expressing cells showed a faster migrating rate compared with GFP-expressing cells indicating that R-cadherin increases C2C12 myoblast migration.

To gain further insight into the role of R-cadherin in muscle cell transformation in the pathologic context of RMS, we generated a stable embryonal RMS cell line RD, in which the expression of R-cadherin was inactivated by RNA interference (RDsiRcad; Fig. 3C,, left). Transformation assays showed that the number and the size of colonies were reduced in RDsiRcad cells compared with control RDsiLuc cells (Fig. 3C,, right). Similar data were obtained using a second R-cadherin short hairpin RNA (data not shown). Importantly, in vivo experiments showed that RDsiRcad cells grew slower than control RD cells in nude mice as shown by monitoring the size of the tumors (Fig. 3D). Taken together, these data show that (a) R-cadherin expression induces myoblast cell transformation ex vivo and in vivo and that (b) inhibition of the R-cadherin expression in the RD cell line significantly decreases its transforming activity.

R-cadherin expression activates Rac1 GTPase. To examine whether the R-cadherin–dependent cell-cell contacts might control the Rho GTPase activity, we used the organization of the F-actin cytoskeleton as a functional read-out (Fig. 4A). An increase of lamellipodia was detected in C2C12 Rcad/GFP cells (d) when compared with control C2C12 GFP myoblasts (b). In addition, lamellipodia were detected in C2C12 Rcad/GFP cells plated onto dishes coated with R-cad-Fc ligand when compared with C2C12 Rcad/GFP cells plated onto dishes coated with anti–Fc antibody (Fc; compare f and e and see Supplementary Online Videos). This result suggests that Rac1 could be activated by the forced R-cadherin expression in C2C12 myoblasts and, thus, led us to assess Rac1 activity using pull-down assays. Rcad/GFP-expressing myoblasts showed an increase in the level of activated Rac1 relative to control GFP-expressing cells (Fig. 4B,, left). To further confirm that R-cadherin engagement activates Rac1, isolated Rcad/GFP-expressing C2C12 cells were plated on dishes coated either with an Fc fragment (Fc) or the R-cad-Fc ligand, which allow us to mimic R-cadherin–mediated adhesion. Two hours after plating on R-cad-Fc, we observed a marked Rac1 activation (Fig. 4B,, right). We then wanted to know whether the c-Jun–NH₂ kinase (JNK) pathway, a downstream target of Rac1, had been activated in the R-cadherin–expressing C2C12 cells. The phosphorylation status of the c-Jun transcription factor was analyzed by immunoblotting parental C2C12 and C2C12 Rcad/GFP cellular extracts at different times of differentiation. Both parental and R-cadherin–expressing C2C12 myoblasts cultured in GM showed a marked phosphorylation of c-Jun (Fig. 4C). Upon DM addition, c-Jun phosphorylation decreased in control cells, whereas it was noticeably maintained in C2C12 Rcad/GFP myoblasts, demonstrating that R-cadherin activates the JNK pathway in myoblasts.

In parallel, we used extracts from the RMS cell line RD, in which R-cadherin expression was inactivated by RNA interference. Significantly, the down-regulation of R-cadherin expression in the RMS cell line RD was accompanied by a lower Rac1 activity level, indicating that R-cadherin is partly accountable for the Rac1 activity in RD (Fig. 4D).

Rac1 inhibition decreases R-cadherin–dependent signaling. We next investigated the effect of Rac1 inhibition on R-cadherin–dependent pathways. First, we analyzed the effect of dominant-negative Rac1 (Rac1N17) on R-cadherin–induced cyclin D1 expression. C2C12 Rcad/GFP myoblasts were transfected with empty pRFP or RFP-tagged Rac1N17 and induced to differentiate by addition of DM. Cells were fixed 2 days thereafter and analyzed for cyclin D1 expression. Cyclin D1 was detected in <20% of cells expressing Rac1N17 (Fig. 5A,, d–f and B) whereas it was expressed in almost 50% of the parental cells (data not shown) and of the cells transfected with empty pRFP (Fig. 5A,, a–c and B). A comparable inhibition of R-cadherin–induced cyclin D1 expression was obtained using the specific Rac inhibitor NSC23766 (data not shown; ref. 34). Moreover, mice injected with Rac1N17-infected C2C12 Rcad/GFP myoblasts developed significantly smaller tumors than mice injected with C2C12 Rcad/GFP cells (Fig. 5C).

Taken together, these findings emphasize that the R-cadherin–induced Rac1 activation is critical for R-cadherin–induced cyclin D1 expression and for myoblast transformation in vivo.

R-cadherin expression down-regulates N-cadherin and M-cadherin. Cadherin switching has been implicated in tumorigenesis (15, 16). We thus analyzed whether R-cadherin influences expression of the two main cadherins, i.e., N-cadherin and M-cadherin, present in C2C12 myoblasts. C2C12 Rcad/GFP and C2C12 Rcad/2myc cells expressed significantly less M-cadherin (Fig. 6A). Both N-cadherin and M-cadherin accumulated less at cell contacts in R-cadherin–expressing myoblasts (Fig. 6B, compare a with d and b with f). Taken together, these data show that (a) R-cadherin decreases the expression of M-cadherin and (b) N-cadherin and M-cadherin only slightly accumulate at the cell contacts in R-cadherin–expressing myoblasts.

Cadherin switching induced by R-cadherin has been linked to competition for p120^ctn, a mechanism in which p120^ctn plays the role of a rheostat to regulate the levels of cadherin in the cell (35). To test this hypothesis in our system, we generated stable mouse C2C12 myoblasts expressing the R-cadherin mutant (Rcad-AAA) that do not interact with p120^ctn (35). Ten clones were established that behave similarly, shown are the data obtained with one of them. Figure 6C shows that expression of Rcad-AAA in C2C12 cells did not down-regulate M-cadherin in C2C12 cells. Moreover C2C12-expressing Rcad-AAA cells expressed normal levels of troponin T and myogenin upon addition of DM and fused normally compared with parental myoblasts (data not shown). In contrast, M-cadherin, as well as myogenin and troponin T levels, were significantly decreased in C2C12 cells expressing R-cadherin. We show that p120^ctn level did not change when the cells expressed wild-type or mutant R-cadherin. Finally, expression of Rcad-AAA did not increase Rac1 activity in contrast to wild-type R-cadherin (right).

Because Rac1 is activated by R-cadherin, we assessed if Rac1 could also be implicated in cadherin switching mediated by R-cadherin in C2C12 cells. As shown in Fig. 6D, Rac1 inhibition, using the specific Rac inhibitor NSC23766 (34), led to an increase of M-cadherin expression and to a relocalization at the membrane junction of N-cadherin and M-cadherin. No modification of M-cadherin and N-cadherin expression was observed in parental C2C12 myoblasts treated with NSC23766 (Supplementary Fig. S2). This suggests that Rac1 activation contributes to the mechanism of cadherin switching mediated by the expression of R-cadherin in C2C12 cells.

The data presented here provide evidence that R-cadherin is a key player in myogenesis inhibition and myoblast transformation. R-cadherin has been detected in developing skeletal muscles, and it might be implicated in the formation of this tissue because its expression can rescue striated muscle in embryonic stem cells that lack E-cadherin (12). On the other hand, its expression in C2C12 myoblasts inhibits myogenesis induction and causes their transformation. Moreover, we have previously observed R-cadherin expression in RMS cell lines and biopsies, whereas this protein is absent from normal adult tissues (13). This suggests that R-cadherin has to be tightly controlled during skeletal muscle development. In this study, we further show that R-cadherin is a major actor of tumorigenesis and/or tumor progression in muscle cells because (a) its expression is able to drive myoblast transformation ex vivo and in vivo and (b) R-cadherin inhibition in the RMS RD cells decreases their tumorigenic potential in vivo. Our results show that R-cadherin expression in C2C12 myoblast is sufficient to induce their transformation and increase their motility, as it was recently reported in various epithelial tumor cell lines (23). The present study is the first demonstration of such an “oncogenic potential” in normal cells. This suggests that whereas R-cadherin might be important during embryonic development for the proliferation and/or the migration of muscle precursor cells, its deregulation can participate in developmental abnormalities and also in pathologic processes, such as RMS formation.

In addition, we deciphered a novel pathway, in which R-cadherin activates Rac1, an event that induces muscle cell transformation through the expression of cyclin D1. Johnson and colleagues previously reported that forced expression of R-cadherin in breast tumor cells acts positively on cell migration through a Rac1-dependent pathway (23). Thus, by activating Rac1 and, consequently, both the growth and the migration of tumor cells, R-cadherin is likely to function as a potent oncogene in muscle cells. The implication of Rac1 in myoblast transformation has been previously described because a constitutively active form of Rac1 was able to induce rat L6 myoblast cells transformation and constitutive Rac1 activation was found in RMS cells (9). Here we show that (a) inactivation of R-cadherin expression in RMS cells decreases Rac1 activation and, as a consequence, their transformation ability in vivo is reduced and (b) inhibition of Rac1 in R-cadherin–expressing cells decreases their tumorigenic potential in vivo, validating the notion that R-cadherin in the pathologic context of RMS might play a critical role in tumor initiation and progression through a Rac1-dependent mechanism. In contrast, RhoA activity is decreased by R-cadherin expression (data not shown), supporting the existence of a balance between Rac1 and RhoA activities downstream of R-cadherin.

Tumor cells often show a decrease in cell-cell and/or cell-matrix adhesion. An increasing body of evidence now indicates that aberrant cell adhesion is causally involved in tumor progression and metastasis rather than merely being a consequence of it (36). Among these adhesion defects, cadherin switching plays a critical role in the progression of some tumors (15). In particular, the switch from E-cadherin to N-cadherin has been shown to enhance the motility, invasiveness, and metastatic potential of cancer cells (37). Moreover, P-cadherin overexpression promotes N-cadherin down-regulation and consequently facilitates the motility of pancreatic cancer cells (38). These changes in cadherin expression not only modulate tumor cell adhesion, but also affect signal transduction and hence tumor malignancy. Such a cadherin switching was observed in RMS cell lines and biopsies because we detected R-cadherin expression concomitant with a perturbation in N-cadherin and M-cadherin function (13). Here, we show that this cadherin switch is induced by R-cadherin expression in myoblasts, suggesting that R-cadherin deregulation during muscle development might be sufficient to induce the loss of N-cadherin and M-cadherin–mediated cell-cell adhesion and to support tumor transition toward malignancy. In this regard, beside the control of myogenesis induction, N-cadherin–dependent signaling mediates contact inhibition of cell growth through the cyclin-dependent kinase inhibitor p27Kip1 (8, 39, 40). p120^ctn and Rac1 play an important role in the molecular mechanisms by which R-cadherin induces N-cadherin and M-cadherin down-regulation. Use of an R-cadherin mutant that could not bind p120^ctn does not perturb the distribution of N-cadherin and M-cadherin at the membrane of C2C12 cells, as well as their ability to undergo differentiation. Because p120^ctn seems to be limiting in the cell, its binding to exogenously expressed R-cadherin could lead to the destabilization and degradation of N-cadherin and M-cadherin. A similar mechanism has been described recently in epithelial cells where ectopic expression of R-cadherin induces down-regulation of E-cadherin and P-cadherin by competition for p120^ctn (35). We also show that Rac1, which is activated by R-cadherin expression, is implicated in this process because Rac1 inhibition clearly restores the levels of M-cadherin in R-cadherin expressing C2C12 cells. Additional unidentified pathways might be activated downstream of R-cadherin because Rac1 activation alone is not sufficient to induce M-cadherin degradation, whereas it leads to M-cadherin and N-cadherin delocalization from the cell-cell contacts (Supplementary Fig. S3).

R-cadherin has recently emerged as a marker of RMS (13, 41). It remains to be determined whether R-cadherin has been maintained due to the development stage of RMS or whether it has been reexpressed in these cells. Interestingly, mice that express, in postnatal terminally differentiated Myf6-positive myofibers, the fusion protein Pax3/Fkhr, issued from a chromosomal translocation found in alveolar RMS, later develop such tumors (42). Therefore, altered expression or activity of transcription factors, which regulate the R-cadherin promoter, might influence the expression of R-cadherin either in muscle precursor cells or in cells already engaged in the differentiation program. The R-cadherin promoter has recently been characterized and contains at least three Pax DNA-binding consensus sites (26). Moreover, the paired box gene Pax-6 seems to be an important regulator of R-cadherin expression in neurons (43). Interestingly, Pax3 and Pax7 are required for the proper migration and specification of myogenic precursor cells and to prevent terminal differentiation (44). In addition, Pax7 is a key regulator of satellite cell self-renewal, as it promotes their proliferation and inhibits their differentiation (45). Pax3 and Pax7 expression has also been reported in RMS and seems involved in the proliferation and survival of these tumor cells (review in ref. 46). It will be interesting to analyze the potential cross-talk between Pax genes and R-cadherin in skeletal muscle cells or to see whether Pax3 and/or Pax7 might control R-cadherin expression by using the various RMS mouse models available.

Another crucial issue is to explore the molecular mechanisms of the functional differences of cell-cell adhesion induced by N-cadherin and R-cadherin in myoblastic cells. Recent data which show an interaction between cadherins and tyrosine kinase receptors suggest that changes in cadherin expression may not only modulate tumor cell adhesion but also affect signal transduction and, hence, the malignant phenotype. Analyzing the interaction of these cadherins with various tyrosine kinase receptors and in particular c-met, due to its role in fetal myoblasts proliferation and tumorigenesis (47), might help in the understanding the processes of R-cadherin–mediated rhabdomyosarcogenesis.

No potential conflicts of interest were disclosed.

Grant support: Centre National Recherche Scientifique, University of Montpellier 2, Association pour la Recherche contre le Cancer, Ligue Nationale contre le Cancer («Equipe labélisée»), Association Française contre les Myopathies, GEFLUC Montpellier, Ligue Nationale contre le Cancer fellowship (J. Kucharczak), and Institut National de la Sante et de la Recherche Medicale (C. Gauthier-Rouvière, B. Robert, and A. Pèlegrin).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank M. Takeichi for R-cadherin antibody; K. Johnson for R-cadherin-2Xmyc and R(AAA)-cadherin-2Xmyc plasmids; R.A. Hipskind for the cyclin D1 promoter plasmid; C. Duperray, I. Aitarsa, M. Brissac, and the team of Montpellier RIO Imaging and RAM (Réseau des Animaleries Montpelliéraines) for technical assistance; and C. Bascoul-Mollevi for statistical analysis.