Targeted Down-regulation of Caveolin-3 Is Sufficient to Inhibit Myotube Formation in Differentiating C2C12 Myoblasts

Caveolin-3 is the principal structural protein of caveolae membrane domains in striated muscle cells. Caveolin-3 mRNA and protein expression are dramatically induced during the differentiation of C2C12 skeletal myoblasts, coincident with myoblast fusion. In these myotubes, caveolin-3 localizes to the sarcolemma (muscle cell plasma membrane), where it associates with the dystrophin-glycoprotein complex. However, it remains unknown what role caveolin-3 plays in myoblast differentiation and myotube formation. Here, we employ an antisense approach to derive stable C2C12 myoblasts that fail to express the caveolin-3 protein. We show that C2C12 cells harboring caveolin-3 antisense undergo differentiation and express normal amounts of four muscle-specific marker proteins. However, C2C12 cells harboring caveolin-3 antisense fail to undergo myoblast fusion and, therefore, do not form myotubes. Interestingly, treatment with specific p38 mitogen-activated protein kinase inhibitors blocks both myotube formation and caveolin-3 expression, but does not affect the expression of other muscle-specific proteins. In addition, we find that three human rhabdomyosarcoma cell lines do not express caveolin-3 and fail to undergo myoblast fusion. Taken together, these results support the idea that caveolin-3 expression is required for myoblast fusion and myotube formation, and suggest that p38 is an upstream regulator of caveolin-3 expression.

The mammalian caveolin gene family now consists of caveolins-1, -2, and -3 (4,5,(11)(12)(13). Caveolins 1 and 2 are coexpressed and form a hetero-oligomeric complex (14) in many cell types, with particularly high levels in adipocytes, whereas expression of caveolin-3 is muscle-specific and found in both cardiac and skeletal muscle (15). Caveolin-3 is localized to the muscle cell plasma membrane (sarcolemma) where it forms a complex with dystrophin and its associated glycoproteins (15). However, under certain conditions caveolin-3 can be physically separated from the dystrophin complex (16). This indicates that, although caveolin-3 is dystrophin-associated, it is not absolutely required for the biogenesis of the dystrophin complex (16).
Caveolin-3 is most closely related to caveolin-1, based on protein sequence homology; caveolin-1 and caveolin-3 are ϳ65% identical and ϳ85% similar (see Tang et al. (13) for an alignment). However, caveolin-3 mRNA is expressed predominantly in muscle tissue types (skeletal muscle, diaphragm, and heart) (13). Identification of a muscle-specific member of the caveolin gene family has implications for understanding the role of caveolins in different muscle cell types (smooth, cardiac, and skeletal), as previous morphological studies have demonstrated that caveolae are abundant in these cells. A number of studies have highlighted the importance of caveolae and caveolins in the pathogenesis of Duchenne's muscular dystrophy. More specifically, dystrophin has been localized to plasma membrane caveolae in smooth muscle cells using immunoelectron microscopy techniques (17), and skeletal muscle caveolae undergo characteristic changes in their size and distribution in patients with Duchenne's muscular dystrophy, but not in other forms of neuronally based muscular dystrophies examined (18). This indicates that muscle cell caveolae may play an important role in muscle membrane biology.
In collaboration with Minetti and colleagues, we have recently identified an autosomal dominant form of limb-girdle muscular dystrophy (LGMD-1C) 1 in two Italian families that is due to a deficiency in caveolin-3 expression. Analysis of their genomic DNA reveals two distinct mutations in the caveolin-3 gene: (i) a 9-base pair microdeletion that removes the sequence TFT from the caveolin-scaffolding domain, and (ii) a mis-sense mutation that changes a proline to a leucine (Pro 3 Leu) in the transmembrane domain (19). Both mutations lead to a loss of ϳ90 -95% of caveolin-3 protein expression.
These results indicate that dramatic reductions in caveolin-3 can produce a disease phenotype in humans. However, it remains unknown whether caveolin-3 expression is required to generate or maintain the differentiated state of muscle cells. To address this issue, we used an antisense approach to essentially ablate caveolin-3 expression in C2C12 cells, a well established murine skeletal myoblast cell line. Our results indicate that drastic down-regulation of caveolin-3 (to undetectable levels) prevents or inhibits myotube formation, but does not affect the expression of a panel of muscle-specific marker proteins. Thus, a deficiency in caveolin-3 expression seen in LGMD-1C patients could potentially slow the process of myotube formation in vivo and partially explain the pathogenic phenotype of this human genetic disorder.
Construction of the Caveolin-3 Antisense Vector-The full-length untagged cDNA encoding rat caveolin-3 (13) was inserted in the antisense orientation into an expression vector that was driven by the ␤-actin promoter (pCAGGS, gift of Dr. Armin Rehn, Ploegh Laboratory, Harvard Medical School, MA). The pCAGGS construct was co-transfected with a plasmid containing hygromycin resistance (pCB7).
Establishment of Stable C2C12 Cell Lines Harboring Caveolin-3 Antisense-C2C12 cells were transfected with caveolin-3 antisense vector using a modified calcium phosphate precipitation protocol. Resistant clones were selected using hygromycin B (200 g/ml). Individual clones were isolated using cloning rings. Lysates from differentiated C2C12 were prepared and assayed for reductions in the expression of caveolin-3 by immunoblotting. C2C12 cells were also transfected with empty vector alone as a critical control.
p38 Inhibitor Treatment-C2C12 cells were treated for the indicated period of time with 10 M amounts of SB203580, SB202190, or SB202474 (an inactive control compound) (Calbiochem, Inc.). Similar results were obtained at a concentration of 5 M.
Phase Microscopy-C2C12, RD, A673, and Hs729 cells were grown in plastic tissue culture dishes and photographed using an inverted Nikon microscope.

Targeted Down-regulation of Caveolin-3 Protein Expression in C2C12 Cells That Harbor
Caveolin-3 Antisense-In order to selectively down-regulate the expression of the caveolin-3 protein, we engineered an expression vector containing the untagged full-length caveolin-3 cDNA in the antisense orientation. For this purpose, we used the well established murine C2C12 skeletal myoblast cell line. Cultured C2C12 cells offer a convenient system to study skeletal myoblast differentiation. These cells can be induced to differentiate from myoblasts into myotubes bearing an embryonic phenotype in low mitogen medium over a period of 2 days. Briefly, proliferating C2C12 cells are cultured in high mitogen medium (DMEM containing 15% fetal bovine serum and 1% chicken embryo extract) and induced to differentiate at confluence in low mitogen medium (DMEM containing 3% horse serum). Overt differentiation is indicated by the assembly of multi-nucleated syncytia, which commences ϳ36 -48 h after the cells are switched to low mitogen media. In addition, we have previously shown that both mRNA and protein levels of caveolin-3 are dramatically induced during the course of differentiation of C2C12 cells from myoblasts to myotubes (13,15).
These caveolin-3 antisense constructs were first tested in transient transfection assays with C2C12 cells and were found to significantly reduce the expression levels of endogenous caveolin-3 during myoblast differentiation, as compared with mock-transfected or vector alone controls (data not shown). Given the preliminary success of this approach in transient transfections, we decided to derive stable cell lines that harbor this caveolin-3 antisense construct.
Three C2C12 cell lines harboring caveolin-3 antisense were derived, and they all behaved similarly. As a consequence, one clone was selected for in depth analysis. Fig. 1 shows a Western blot analysis of the expression of caveolin-3 in C2C12 cells harboring caveolin-3 antisense and untransfected control cells. Note that caveolin-3 levels are ef- It should be noted that, although caveolin-1 expression was detectable in C2C12 cells with caveolin-1 pAb N-20, no caveolin-1 expression was detected with caveolin-1 mAb 2297. This observation may reflect differences in the relative sensitivity of these antibody probes and the fact that C2C12 cells express only very low levels of caveolin-1. These results are consistent with our previous observations that adult skeletal muscle fibers in vivo only express caveolin-3, but do not express caveolin-1 or caveolin-2 (15,19). fectively reduced during the differentiation process. In addition, caveolin-1 and caveolin-2 levels were not affected by the expression of caveolin-3 antisense, demonstrating that the expression of caveolin-3 antisense selectively down-regulates the expression of caveolin-3. Importantly, C2C12 cells harboring vector alone did not show any changes in the levels of caveolin-3 expression (see below).
C2C12 Cells Harboring Caveolin-3 Antisense Express Normal Levels of Muscle-specific Marker Proteins-To investigate whether targeted down-regulation of the caveolin-3 protein overtly affects the differentiation process, we next evaluated the expression of a panel of muscle-specific marker proteins in C2C12 cells harboring caveolin-3 antisense. These markers included both cytoskeletal elements (troponin T and myosin heavy chain) and muscle-specific plasma membrane compo-nents (␤-dystroglycan and dystrophin). Fig. 2 shows the results of this analysis. Interestingly, C2C12 cells harboring caveolin-3 antisense expressed normal levels of troponin T, myosin heavy chain, ␤-dystroglycan, and dystrophin, as compared with untransfected control C2C12 cells. These results clearly indicate that the process of differentiation is not overtly affected by targeted down-regulation of caveolin-3 expression.
C2C12 Cells Harboring Caveolin-3 Antisense Fail to Undergo Myoblast Fusion and Myotube Formation-Myoblast fusion and myotube formation is indicated by the assembly of multinucleated syncytia, which commences ϳ36 -48 h after the C2C12 cells are switched to low mitogen media and can be observed morphologically.
Interestingly, C2C12 cells harboring caveolin-3 antisense failed to undergo myotube formation (Fig. 3A), despite the fact that they undergo differentiation normally and continue to express muscle-specific marker proteins that are a hallmark of normal adult muscle (Fig. 2). In striking contrast, C2C12 cells and compared for their ability to generate fused myotubes. Note that both cell lines differentiate from myoblast to myocytes, but only parental C2C12 cells show fused myotubes. Prolif, cells grown in "proliferation" medium; Diff, cells grown in "differentiation" medium.
harboring vector alone continue to express muscle-specific protein markers, caveolin-3, and undergo myotube formation normally (Fig. 4). These results indicate that targeted down-regulation of the caveolin-3 protein is sufficient to block myoblast fusion and subsequent myotube formation.
One possibility is that caveolin-3 down-regulation may cause a delay, rather than a block in myotube formation. To address this issue, we examined myoblast fusion after 5 days of differentiation. Even under these condition, no myoblast fusion/ myotube formation was observed with C2C12 cells that harbor caveolin-3 antisense (Fig. 3B), as compared with control C2C12 cells. These results are more consistent with the idea that down-regulation of caveolin-3 results in a block in myoblast fusion, rather than a delay.
Transient Activation of the p38 MAP Kinase Pathway Occurs during the Differentiation of C2C12 Myoblasts to Myotubes-As we recently observed that activation of the p38 MAP kinase pathway is a prerequisite for the differentiation of 3T3-L1 fibroblasts to adipocytes (24), we next assessed the activation state of the p38 pathway during differentiation in C2C12 cells. For this purpose, we used immunoblotting with phosphospecific antibody probes that are routinely used to assess p38 activation. We observed that p38 MAP kinase activation occurred early during the differentiation program and was transient, with peak activity on day 2 (Fig. 5). In contrast, the protein levels of total p38 MAP kinase remain relatively constant, as seen using a phospho-independent antibody probe. Thus, these results are in agreement with our previous results with the adipocyte system (24).

Inhibition of p38 MAP Kinase Blocks Caveolin-3 Expression and Myotube Formation, but Does Not Affect the Expression of a Variety of Muscle-specific Marker
Proteins-We next used a well established and highly selective p38 MAP kinase inhibitor (SB203580) to assess the role of p38 activation in the differentiation of C2C12 cells. Importantly, this inhibitor does not affect the activation of the p42/44 and the stress-activated protein/c-Jun N-terminal kinase MAP kinase pathways (see Ref. 24, and references therein). Using this approach with the adipocyte system, we recently showed that activation of p38 is required for achieving the differentiated adipogenic phenotype and for up-regulation of the caveolin-1 protein product (24). Fig. 6 shows that treatment with the p38 inhibitor (SB203580) selectively blocks the expression of the caveolin-3 protein, but has little or no effect on the expression of other specific markers of the muscle cell plasma membrane or cytoskeletal elements. In addition, treatment with the p38 inhibitor SB203580 blocked myoblast fusion/myotube formation (Fig. 7A). However, addition of the p38 inhibitor SB203580 after 2 days of differentiation did not reverse myoblast fusion. Fig. 8 shows a direct comparison of the effects of caveolin-3 antisense or the p38 inhibitor SB203580 on myotube formation. Note that both treatments effectively block this process.
As additional controls for the effects of the p38 inhibitor SB203580, we also evaluated the effects of a second well characterized p38 inhibitor, SB202190, and a known related inactive control compound, SB202474. Our results indicate that the second p38 inhibitor (SB202190) effectively blocks caveolin-3 expression and myoblast fusion, while SB202474 is inactive as predicted (Figs. 6B and 7B). Virtually identical results were obtained after either 2 days or 5 days of differentiation. cancer cell lines (25)(26)(27)(28)(29), we next examined the expression of the caveolin-3 protein in RD cells.

Down-regulation of Caveolin-3 in Human Rhabdomyosar
Interestingly, RD cells express muscle-specific markers (such as troponin T), but do not express the caveolin-3 protein product (Fig. 9A). In addition, they fail to undergo myoblast fusion (Fig. 9B). Thus, it appears that caveolin-3 expression is down-regulated during skeletal muscle cell transformation.
As a consequence of these observations with RD cells, we analyzed two additional ATCC cell lines derived from human rhabdomyosarcomas (A673 and Hs729) and the results are shown in Fig. 9 (A and B). Note that A673 and Hs729 cells do not express the caveolin-3 protein product and they fail to undergo myoblast fusion. Thus, our results indicate that in all three rhabdomyosarcoma-derived cell lines (RD, A673, and Hs729 cells), caveolin-3 levels are down-regulated and these cell lines fail to undergo myoblast fusion.
FIG. 6. Inhibition of p38 MAP kinase blocks caveolin-3 expression, but does not affect the expression of a variety of musclespecific marker proteins. A, immunoblot analysis of lysates from parental C2C12 cells with specific antibody probes that recognize caveolin-3, ␤-dystroglycan, dystrophin, MHC, and troponin T. C2C12 cells were differentiated for 0, 1, and 2 days in the absence or presence of the p38 inhibitor SB 203580 (10 M). Note that the expression of caveolin-3 is dramatically reduced after treatment with the p38 inhibitor. In striking contrast, p38 inhibitor treatment had little or no effect on the expression of ␤-dystroglycan, dystrophin, MHC, and troponin T. Each lane contains equal amounts of total protein. B, immunoblot analysis of lysates from parental C2C12 cells with a specific antibody probe that recognizes caveolin-3. During this 2 day period, cells were treated with SB203580, SB202190, or SB202474 (each at a concentration of 10 M) or left untreated. Note that the expression of caveolin-3 is dramatically reduced after treatment with either p38 inhibitor (SB203580 or SB202190) as compared with untreated controls, but remains unaffected by treatment with a related inactive control compound, SB202474. Each lane contains equal amounts of total protein.

FIG. 7. Inhibition of p38 MAP kinase blocks myotube formation.
A, parental C2C12 cells were differentiated for 4 days in the absence or presence of p38 inhibitor (SB203580). a, no inhibitor addition; b, inhibitor is added at the beginning of differentiation (day 0); c, inhibitor is added after 1 day of differentiation (day 1); d, inhibitor is added after 2 days of differentiation (day 2). Note that myotube formation is compromised if the p38 inhibitor is added at day 0 or day 1 of differentiation. In contrast, if the inhibitor is added after day 2, no effect on myotube formation is observed. B, as in A, except parental C2C12 cells were differentiated for 2 days or 5 days in the absence or presence of SB203580, SB202190, or SB202474, each at a concentration of 10 M. Each compound was added at the beginning of differentiation (day 0). Note that either p38 inhibitor (SB203580 or SB202190) blocks myoblast fusion, while an inactive control compound (SB202474) has no effect on myoblast fusion. Virtually identical results were obtained after either 2 days or 5 days of differentiation.
However, it is important to note that recombinant expression of caveolin-3 in RD cells was not sufficient to drive myoblast fusion and myotube formation (data not shown). These results indicate that although caveolin-3 expression may be required for or greatly facilitates myoblast fusion, caveolin-3 expression is clearly not sufficient to drive myoblast fusion in the context of RD cells. DISCUSSION LGMD-1C is an autosomal dominant form of limb-girdle muscular dystrophy that is genetically caused by mutations within the coding region of the caveolin-3 gene. In collaboration with Minetti and colleagues (19), we have recently identified two different families in Italy with this form of muscular dystrophy. In these patients, the levels of the caveolin-3 protein are reduced by ϳ90 -95% as revealed by immunofluorescence and Western blot analysis. These results indicate that dramatic reductions in caveolin-3 can produce a disease phenotype in humans. However, it remains unknown whether caveolin-3 expression is required to generate or maintain the differentiated phenotype of muscle cells.
Here, we have directly addressed this issue by using an antisense approach to ablate caveolin-3 expression in C2C12 cells. We show that C2C12 cells harboring caveolin-3 antisense undergo differentiation and express normal amounts of four muscle-specific marker proteins. However, C2C12 cells harboring caveolin-3 antisense fail to undergo myoblast fusion and do not form myotubes. Thus, a deficiency in caveolin-3 expression may potentially slow the process of myotube formation in vivo, contributing to the pathogenesis of LGMD-1C.
Using phosphospecific antibody probes, we noted that p38 MAP kinase activation was transiently induced during the early phase of myoblast differentiation. Interestingly, treatment with a specific p38 inhibitor (either SB203580 or SB202190) blocked both myotube formation and caveolin-3 expression, but did not affect the expression of other muscle-specific proteins. These results support the idea that caveolin-3 expression is required for myoblast fusion and myotube formation, and suggest that p38 is an upstream regulator of caveolin-3 expression. These data also suggest that p38 MAP kinase activation and subsequent caveolin-3 expression at the muscle  Recently, we and other laboratories have shown that activation of the p38 MAP kinase pathway occurs during a variety of differentiation processes. Inhibition of p38 activation effectively blocks these differentiation processes. These processes include the nerve growth factor-induced differentiation of PC12 cells into neuron-like cells (30), the conversion of 3T3-L1 fibroblasts to adipocytes (24), and the erythropoietin-mediated induction of red blood cell formation (31). In the case of 3T3-L1 cells, we have also shown that treatment with p38 MAP kinase inhibitors blocks the induction of caveolin-1 protein expression (24). Normally, both the caveolin-1 mRNA and protein levels are induced ϳ10 -25-fold during the process of adipogenesis (14,32). Thus, by analogy with p38-mediated regulation of caveolin-1 expression during adipogenesis, it is not completely unexpected to observe that inhibition of the p38 MAP kinase pathway prevents expression of the caveolin-3 protein and inhibits myotube formation.
Addendum-While our paper was being revised, two other groups reported the effects of p38 inhibition on C2C12 (33) and L8 (34) myoblast differentiation. As we observe here, they also found that p38 activation was required for myotube formation. However, they did not evaluate the effects of p38 inhibition on the expression of caveolin-3 and they did not implicate p38-mediated induction of caveolin-3 in the process of myotube formation. In addition, their results with p38 inhibitors independently support our current results made here using an antisense approach to selectively down-regulate caveolin-3 protein expression.