Beta-catenin is essential and sufficient for skeletal myogenesis in P19 cells.

Wnt1 and Wnt3a are signaling factors known to play a role in the induction of myogenesis in the myotome of the differentiating somite. Both factors may transduce their signal by a conserved pathway that leads to transcriptional regulation by beta-catenin/Lef1. beta-Catenin and Lef1 are found in the myotome prior to MyoD expression. We have utilized the P19 cell system to study the mechanisms by which Wnt3a may activate MyoD expression and subsequent skeletal muscle development. We have isolated P19 cell lines that stably express either Wnt3a or activated beta-catenin and found that aggregation of these cells results in the induction of myogenesis compared with control cells. Pax3, Gli2, Mox1, and Six1 were expressed during Wnt3a and beta-catenin-induced differentiation prior to MyoD expression. Furthermore, we have shown that the nuclear function of beta-catenin was essential for skeletal myogenesis in P19 cells by overexpression of a dominant negative beta-catenin/engrailed chimera. Primitive streak factors were present, but expression of Pax3, Mox1, Gli2, and Six1 was lost in these cells, indicating that nuclear beta-catenin is essential for specification of mesodermal precursors to the myogenic lineage. Therefore, Wnt signaling, acting via beta-catenin, is necessary and sufficient for skeletal myogenesis in P19 cells.

The myogenic regulatory factors (MRFs) 1 are a family of muscle-specific transcription factors, including MyoD, Myf-5, myogenin, and MRF4 that control skeletal muscle differentiation (1). The MRFs induce the differentiation of skeletal muscle, first formed in the myotome of the somite. This process is controlled by signals such as Wnts and Sonic hedgehog (SHH) (2)(3)(4) from adjacent tissues, including the dorsal neural tube and the floor plate/notochord (5,6).
Wnt1 and Wnt3a are expressed in the dorsal neural tube (7). Both factors induce myogenesis in co-culture experiments with unsegmented paraxial mesoderm and somites in the chick (2,3) and mouse (8). Furthermore, mice null for both factors (9) and mice treated with Frzb1, a Wnt antagonist (10), both had myotomal defects suggesting a crucial in vivo role for the Wnts in skeletal myogenesis.
Wnts regulate the expression of several factors during early embryogenesis, including Brachyury T, a marker of early mesoderm (11,12). Wnts also activate or maintain the expression of transcription factors found in the dermomyotome, including Gli2 (13) and Pax3 (4,14,15). Gli factors play an essential role in Myf-5 expression (16). Pax3 is a transcription factor known to act upstream of MyoD during skeletal muscle development (17)(18)(19). Overexpression of a dominant negative Pax3 in P19 cells shows Pax3 is essential for Six1 and MRF expression (19). Six1 is a myotomal transcription factor that interacts with co-factors to mediate muscle-specific gene expression (20).
Wnts can effect gene expression by various pathways, including the canonical ␤-catenin pathway, a Ca 2ϩ -sensitive kinase pathway (21), or the JNK signaling cascade (22)(23)(24)(25). Wnt3a is thought to act via the canonical pathway, which ultimately leads to the accumulation of ␤-catenin in the nucleus (26). ␤-Catenin binds to the Lef/TCF family of transcription factors to regulate gene expression (27,28). A combination of Wnt and SHH signals regulates the expression of both ␤-catenin and Lef1 in the myotome prior to MyoD expression in the chick (29). Studies suggest that ␤-catenin regulates the expression of both Pax 3 (15) as well as Gli2 and Gli3 (13). Furthermore, somites were disrupted and dorsal markers were down-regulated in embryos injected with a dominant negative ␤-catenin/engrailed chimera in Xenopus (30). Although these studies implicate ␤-catenin in skeletal muscle development, a direct role has not been shown.
P19 cells are pluripotent embryonal carcinoma cells that differentiate into several cell types, including cardiac and skeletal myocytes upon cellular aggregation in the presence of Me 2 SO (31,32). Stable cell lines that express transcription factors involved in myogenesis, including MyoD, myogenin, MEF2C, and Pax3, have been shown to induce skeletal muscle development upon cellular aggregation, in the absence of Me 2 SO (19,33,34). The resulting myocytes are biochemically and physiologically comparable to embryonic skeletal myocytes (35). The temporal expression of mesodermal factors induced during skeletal muscle differentiation in P19 cells is similar to the pattern of expression seen in the embryo (36). Primitive streak factors, including Wnt3a are expressed early after cellular aggregation and prior to factors found in the somite, such as Pax3, Six1, Gli2, and Mox1. Mox1 is a homeobox gene expressed in the dermomyotome during somite differentiation (37). Subsequently, MRF expression is initiated, accompanied by markers of differentiated skeletal muscle. In addition, P19 cells are induced to form skeletal muscle when co-cultured with pieces of dorsal neural tube (38). Therefore, P19 cells are a * This work was supported in part by the Neuromuscular Research Partnership, an initiative of ALS Canada, Muscular Dystrophy Association of Canada, and Canadian Institutes of Health Research. 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  useful system in which to study the induction of skeletal myogenesis by Wnts expressed in the neural tube.
In this study, we have shown that the expression of Wnt3a or an activated ␤-catenin is sufficient to induce skeletal myogenesis in aggregated P19 cells. Both factors induced the expression of somite factors, including Pax3, Mox1, Gli2, and Six1. By the overexpression of a dominant negative ␤-catenin/engrailed chimera, we have shown that the nuclear function of ␤-catenin is essential for the expression of these factors leading to skeletal myogenesis. These studies represent the first observation that ␤-catenin is both sufficient and necessary for MRF expression leading to skeletal muscle development.
Cell Culture Transfections and Differentiation-P19 embryonal carcinoma cells were cultured as described previously (41) in ␣-minimum essential media (Canadian Invitrogen, Burlington, Ontario, Canada) supplemented with 5% cosmic calf/5% fetal bovine serum (Cansera, Rexdale, Ontario, Canada and HyClone, Logan, UT, respectively). The serum supported the differentiation of skeletal muscle in the presence of Me 2 SO (42). PGK-Wnt3a was stably transfected into P19 cells by the calcium phosphate method (43). In 60-mm culture plates, 7.5 ϫ 10 5 cells in 5 ml of media were exposed to DNA precipitate for 9 h. The DNA precipitates were composed of 8 g of PGK-Wnt3a or PGK vector alone for controls, plus 1 g of PGK-puromycin, 1 g of PGK-LacZ, and 2.5 g of B17 (44). PGK-␤-catenin* and CMV-␤-catenin/engrailed were stably transfected utilizing the FuGENE6 transfection reagent as per the manufacturer's instructions (Roche Molecular Biochemicals, Laval, Quebec, Canada). A mixture of 0.8 g of PGK-␤-catenin* or PGK vector alone for controls, plus 0.2 g of PGK-puromycin, 0.3 g of PGK-LacZ, 0.7 g of B17, and FuGENE6 reagent was added to 2.5 ϫ 10 5 cells in 35-mm tissue culture dishes. A mixture of 2.04 g of CMV-␤-catenin/ engrailed or vector alone for controls, plus 0.1 g of PGK-puromycin, 0.17 g of PGK-LacZ, 0.77 g of B17, and FuGENE6 reagent was added to 2.5 ϫ 10 5 cells in 35-mm tissue culture dishes. In each case, transfection efficiency was analyzed by ␤-galactosidase assays as described previously (45). Cells were then plated into 150-mm tissue culture dishes and selected for puromycin resistance for 7-9 days from which stable cells lines were isolated and termed P19[Wnt3a] (36), P19[␤catenin*], or P19[␤-cat/EnR] cells, respectively.
Differentiation of both P19[Wnt3a] and P19[␤-catenin*] cell lines was carried out by aggregating 5.0 ϫ 10 5 cells for 4 days in Petri dishes, followed by plating on day 4 into tissue culture dishes and maintenance in culture until day 9. P19[␤-cat/EnR] cells were aggregated in the presence of 0.8% Me 2 SO, plated on day 4, and harvested on day 9. In each case, experiments were carried at least twice utilizing at least two cell lines.
Northern Blot Analysis-Total RNA was extracted by the lithium chloride/urea method on the day of differentiation indicated and examined as described previously (19). Briefly, 6 -12 g of total RNA was separated on a 1% agarose, formaldehyde gel, and transferred overnight onto Hybond-N (Amersham Biosciences, Inc., Baie d'Urfe, Quebec, Canada) by capillary blotting. RNA was cross-linked and hybridized overnight at 42°C to DNA probes labeled with [␣-32 P]dCTP. Blots were washed 5 ϫ 5 min in low stringency wash at 42°C (2ϫ SSC and 0.1% SDS) and 1 ϫ 15 min in high stringency wash at 65°C (0.1ϫ SSC and 0.1% SDS).
Fragments utilized for cardiac ␣-actin, Wnt3a, Wnt5b, Brachyury T, Pax3, Mox1, Gli2, Six1, and MyoD probes were described previously (19). The Xenopus ␤-catenin probe utilized was a ClaI fragment encoding the armadillo repeat domain of ␤-catenin. Rat ␤-catenin probe, utilized to detect endogenous ␤-catenin, was a SphI-SstI fragment of cDNA that was kindly provided by W. Rushlow. The Lef1 probe was an NdeI fragment of the coding region of mouse Lef1, provided kindly by J. McDermott. A 750-bp EcoRI fragment of rabbit 18 S cDNA was utilized as a probe to standardize loading. Northern blots were visualized by autoradiography.
Immunofluorescence-For detection of myosin heavy chain (MyHC) expression, aggregates from each differentiation were plated onto 35-mm tissue culture dishes containing gelatin-coated coverslips on day 4. On day 9, cells were washed with PBS, fixed with Ϫ20°C methanol for 5 min at room temperature and rehydrated with PBS. Cells were stained for myosin heavy chain (MyHC) utilizing MF20 antibody (46). To detect localization of ␤-catenin protein, P19[␤-catenin*] and control cells were fixed with Ϫ20°C methanol and reacted with ␤-catenin antibody, kindly provided by K. Knudsen, at a 1:10 dilution in PBS/5% sheep serum. Cells were fixed with Ϫ20°C acetone and reacted with HA anti-body (Roche Molecular Biochemicals, Laval, Quebec, Canada), at a 1:100 dilution in PBS/5% sheep serum. Nuclei were detected by Hoechst stain as described previously (36). Immunofluorescence was visualized on a Zeiss Axioskop microscope. Images were captured on a Sony 3CCD camera and processed utilizing Axiovision software.
To detect nuclear localization of endogenous ␤-catenin during Wnt3a-induced skeletal myogenesis, P19[Wnt3a] and control cells were aggregated in the absence of Me 2 SO. Aggregates in suspension (day 2-4) were fixed in 2.5% paraformaldehyde/PBS, extracted in Ϫ20°C methanol, and washed in PBS. Aggregates were plated onto coverslips on days 5, 6, and 9 and fixed in Ϫ20°C methanol. All aggregates were stained with ␤-catenin antibody at a 1:50 dilution in PBS/5% sheep serum and with propidium iodide to detect nuclei. Fluorescence was visualized utilizing a Bio-Rad MRC-600 confocal laser scanning microscope.

Wnt3a Induces Skeletal Myogenesis in Aggregated P19
Cells-We have shown previously that Wnt3a is expressed early during Me 2 SO-induced skeletal myogenesis in P19 embryonal carcinoma cells (36). To study the ability of Wnt3a to regulate MRF expression and subsequent skeletal muscle development, P19[Wnt3a] cells were aggregated in the absence of Me 2 SO. These cells differentiated into skeletal muscle as shown by immunofluorescence utilizing an antibody against the muscle-specific marker MyHC (Fig. 1). Bipolar MyHCpositive myocytes, which represented between 5 and 10% of total cells, were found in the P19[Wnt3a] cells ( Fig. 1D) but not in the control cell lines (Fig. 1B). Therefore, Wnt3a can induce skeletal myogenesis in aggregated P19 cells. The ability of Wnt factors to induce skeletal muscle in P19 cells agrees with explant and overexpression studies in the mouse and chick (2)(3)(4)8).
Wnt3a Induces the Expression of Several Early Mesoderm and Somite Factors-To determine the cascade of factors activated during Wnt3a-induced myogenesis, a time course was analyzed (Fig. 2). Wnt3a was expressed at high levels throughout the time course as expected in P19[Wnt3a] cells and not in control cells ( Fig. 2A). One or 2 days of aggregation was sufficient to induce the expression of primitive streak factors and early mesoderm markers such as Brachyury T and Wnt5b in P19 control cells (Fig. 2, B and C). However, P19 control cells do not continue to differentiate and expression of these factors diminishes rapidly. In contrast, P19[Wnt3a] cells express high levels of these early factors throughout the majority of the differentiation program (Fig. 2, B and C), indicating an abundance of early mesoderm present in these cultures. Both ␤-catenin and Lef1, which are downstream components of the Wnt signaling pathway, are expressed throughout the time course in both control and P19[Wnt3a] cells (Fig. 2, D and E). The presence of ␤-catenin and Lef1 in aggregated control cells indicates that, in the absence of Wnt3a signaling, ␤-catenin and Lef1 expression is not sufficient to induce myogenesis.
The progression of differentiation was marked by the activation of transcription factors normally found in the early somite, including Mox1, Gli2, Pax3, and Six1, which were present on days 5 and 6 in P19[Wnt3a] cells but not in control cell lines (Fig. 2, F-I). Skeletal muscle differentiation was confirmed by the appearance of MyoD beginning on day 8 of the time course in P19[Wnt3a] cells (Fig. 2J). This temporal pattern of expression suggests that Wnt3a induces skeletal myogenesis by activating the expression of transcription factors upstream of the MRFs.
␤-Catenin Is Sufficient to Induce Myogenesis in Aggregated P19 Cells-Wnt3a is thought to transduce its signal via the canonical ␤-catenin/Lef/TCF pathway. Lef1 and ␤-catenin were expressed fairly constitutively in P19 cells aggregated in the absence of Me 2 SO (Fig. 2E), indicating that Lef1 is available to interact with ␤-catenin to activate downstream genes. Although abundant ␤-catenin staining was present in the cytoplasm and membrane of the cells, we could not reliably detect endogenous ␤-catenin in the nucleus during differentiation of P19[Wnt3a] cells (data not shown). It is likely that our staining methods were not sensitive enough to detect endogenous levels of nuclear ␤-catenin above the intense cytoplasmic and membrane staining at the cell surface in aggregated P19 cells.
Furthermore, the nuclear expression is likely restricted to a small proportion of the aggregate and a narrow window of time and, therefore, is not easily detected in the heterogeneous population of aggregated cells. Others (40,(47)(48)(49) have found that, although nuclear staining of ␤-catenin could not be detected, overexpression of ␤-catenin could mimic the effects of Wnt signaling and induce the expression of putative target genes.
We therefore isolated P19 cell lines expressing an activated ␤-catenin, termed ␤-catenin*, to determine if Wnt3a is acting via ␤-catenin and not a Ca 2ϩ -dependent pathway or the JNK kinase pathway, to stimulate myogenesis. ␤-Catenin* consists of the internal armadillo repeat region, which was shown to be necessary and sufficient for axis duplication in Xenopus (40) and sufficient to emulate Wnt1 activity when ectopically expressed in chick embryos (15). To determine the subcellular localization, ␤-catenin* was detected with an antibody against the HA tag in P19[␤-catenin*] cells. ␤-Catenin* was localized primarily in the nucleus with some staining in the cytoplasm, as compared with Hoechst stain (Fig. 3, compare A and B) as shown previously (40). Conversely, endogenous ␤-catenin, which was detected by anti-␤-catenin antibody, was found predominantly in the cytoplasm and at the cell surface of control cells (Fig. 3, compare G and H). A comparison of ␤-catenin   (Fig. 3H), indicating the difficulty of identifying changes in subcellular localization in a small proportion of the ␤-catenin protein population.
P19[␤-catenin*] cells (Fig. 4I, panel B) were aggregated for 4 days and shown to differentiate into skeletal muscle compared with control cells (Fig. 4I, panel D) on day 9, as shown by the presence of MyHC-positive bipolar skeletal myocytes (2-5% of total cells). The expression of factors activated by ␤-catenin*induced myogenesis was studied by Northern blot analysis. Exogenous ␤-catenin* was expressed at high levels compared with control cell lines (Fig. 4II, panel E). Endogenous ␤-catenin and Lef1 were expressed in both control and P19[␤-catenin*] cells and were slightly up-regulated in P19[␤-catenin*] cells (Fig. 4II, panels F and G). Gli2, Mox1, Pax3, and Six1 expression was activated by the presence of ␤-catenin* (Fig. 4II,  panels H-K). Interestingly, both Pax3 and Gli2 are expressed on day 0 of differentiation, indicating that these factors do not require cellular aggregation for activation by ␤-catenin*. The observation that ␤-catenin* triggers the expression of tran-scription factors similar to those induced by Wnt3a suggests that Wnt3a is acting via ␤-catenin to induce myogenesis in P19 cells.
␤-Catenin/Engrailed Inhibits Me 2 SO-induced Skeletal Myogenesis-To determine whether ␤-catenin is essential for skeletal myogenesis, we isolated cell lines expressing a ␤-catenin/ engrailed chimera (30). This negative regulator was created by replacing the activation domain of ␤-catenin with the repression domain of engrailed, which is known to interact with transcriptional repressors to actively block transcription (50,51). The chimera was shown to specifically repress the transcriptional function of ␤-catenin (30). Control cells and cell lines expressing the chimera, termed P19[␤-cat/EnR] cells, were differentiated in the presence of 0.8% Me 2 SO. Under these conditions, control cells differentiate efficiently along the skeletal muscle pathway, as shown by the appearance of MyHC-positive myocytes on day 9 by immunofluorescence ( Fig.  5B). Skeletal myocytes were not detected in P19[␤-cat/EnR] cells (Fig. 5D).
A time course of myogenesis was analyzed to identify which transcripts were affected by the expression of the inhibitory chimera. The expression of the engrailed fusion was shown to be high throughout the time course in P19[␤-cat/EnR] cells but not in control cells (Fig. 6A). The inhibition of skeletal muscle development was confirmed by a loss of MyoD expression compared with control cells (Fig. 6K). Early primitive streak and mesodermal markers such as Brachyury T, Wnt5b, and Wnt3a were reduced compared with control cells (Fig. 6, B-D). Endogenous ␤-catenin and Lef1 expression was not diminished by the presence of ␤-catenin/engrailed (Fig. 6, E and F). The expression of somite factors detected during both Wnt3a-and ␤-catenin-induced myogenesis, including Mox1, Gli2, Pax3, and Six1, was lost in P19[␤-cat/EnR] cell lines but not in control cells that differentiated efficiently into skeletal muscle (Fig. 6, G-J).
These findings indicate that ␤-catenin is essential for the specification of cells to the myogenic lineage by controlling the expression of Mox1, Gli2, Pax3, and Six1. DISCUSSION In this study, we have shown that Wnt3a and its nuclear mediator, ␤-catenin, are both sufficient to induce the skeletal muscle program when expressed in aggregated P19 cells. The

␤-Catenin Induces Skeletal Myogenesis
expression of transcription factors, including Pax3, Mox1, Gli2, and Six1, is activated in Wnt3a-and ␤-catenin*-induced myogenesis. Furthermore, we have shown that ␤-catenin is essential for the expression of these transcription factors and the progression of skeletal muscle differentiation in P19 cells by expressing a ␤-catenin/engrailed chimera. This data implies a model in which Wnt3a works via ␤-catenin to induce the expression of transcription factors present in the early somite of the developing embryo (Fig. 7). These somite factors may, in turn, activate MRF expression and skeletal muscle differentiation in aggregated P19[Wnt3a] cells. This model is consistent with the order of expression of factors during P19 cell differentiation (36) and murine embryonic myogenesis (5,52). In summary, we have shown that the Wnt mediator ␤-catenin is both necessary and sufficient for skeletal myogenesis.
Although the ability of Wnt3a to induce skeletal myogenesis supports the findings of several others (2)(3)(4)8), we are the first to show that ␤-catenin is sufficient to induce myogenesis. Studies in the chick have shown that ectopic expression of Wnt1 or activated ␤-catenin can enhance Pax3 but not MyoD expression (15). Another study used a similar approach with Wnt1 in the chick and found that both Pax3 and MyoD expression was up-regulated (4). These authors suggest that differences in the position of injection or in levels of Wnt1 may account for the ability of Wnt1 to activate MyoD expression in one study but not the other. Our ability to observe induction of myogenesis with ␤-catenin is most likely due to the lack of background skeletal muscle development in P19 control cells. Studies in the embryo generally observe an enhancement of myogenesis over the background of endogenous developmental programs occurring during embryogenesis. Furthermore, the presence of inhibitory factors in the embryo may hinder the ability of ectopic ␤-catenin to enhance myogenesis.
Our observation that Wnt and ␤-catenin can activate Gli2 expression supports data in the chick where LiCl treatment resulted in Gli2 up-regulation (13). Furthermore, our data agree with previous studies showing that Wnt signals can activate Pax3 expression (4,14,15). We extend these studies to include the activation of Six1 and Mox1. Pax3 can activate the expression of Six1 and Mox1 but not Gli2 (19). Therefore, it is possible that the activation of Gli2 by Wnt3a is independent of Pax3 activity. In fact, recent studies have shown that Gli2 may bind and activate the Myf-5 epaxial enhancer, providing a direct mechanism for MRF activation (16). Although our model of Wnt activating the expression of transcription factors upstream of MyoD (Fig. 7) fits the temporal pattern of expression of factors during Wnt3a-induced myogenesis, we cannot rule out the possibility that ␤-catenin/Lef1 is also able to activate MRF expression directly.
The expression of Pax3 and Gli2 but not Mox1 or Six1 on day 0 in P19[␤-catenin*] cells suggests that ␤-catenin may directly activate the expression of Pax3 and Gli2. Alternatively, ␤-catenin may indirectly activate Pax3 and Gli2 in the absence of secondary events mediated by cellular aggregation. Experiments involving the Pax3 and Gli2 promoters may provide more insight into the mechanism by which ␤-catenin/Lef1 leads to the activation of these factors. The expression of Pax3 and Gli2 on day 0 in P19[␤-catenin*] cells did not accelerate myogenic differentiation compared with Wnt-induced myogenesis on day 9. This is similar to the finding that overexpression of Pax3 in P19 cells did not accelerate myogenic differentiation (19).
Although several studies have shown the ability of Wnts to induce skeletal myogenesis, we are the first to show that ␤-catenin is sufficient to activate the expression of Mox1, Six1, and the myogenic program. The observation that ␤-catenin is sufficient to induce skeletal muscle development suggests that Wnt signaling acts through the canonical pathway and not alternate Wnt pathways that include Ca 2ϩ -dependent signaling pathways and the JNK kinase pathway. However, we did observe less skeletal muscle development in P19[␤-catenin*] cells compared with P19[Wnt3a] cells. Although this observation may be cell line-dependent, we cannot rule out the possibility that the other pathways play a role in the ability of Wnt3a to induce more robust skeletal muscle differentiation in P19 cells.
␤-Catenin is known to function in both cell-cell adhesion, by associating with the cadherin complex at adherens junctions, and as a transcriptional regulator, by associating with Lef/TCF factors in the nucleus (53). Subcellular localization of exogenous ␤-catenin* in P19[␤-catenin*] cells suggests that the factor is acting in the nucleus and not at adherens junctions to induce the myogenic program. Furthermore, the ability of ␤-catenin/engrailed to inhibit myogenesis supports that ␤-catenin is acting in the nucleus, because engrailed is known to act as a repressor of transcription (50,51). The ␤-catenin/engrailed chimera was shown to associate with members of the cadherin complex and function normally at the adherens junctions in cell-cell adhesion, making ␤-catenin/engrailed a specific repressor of ␤-catenin's function in the nucleus (30). Recently, myo-  Wnt3a functions via ␤-catenin to activate the expression of Mox1, Pax, and Gli2, specifying these cells as muscle precursors. Whether this activation is direct or indirect is unknown at present. Pax3 was shown previously to induce Six1 expression and skeletal myogenesis (19). The roles of Mox1 and Gli2 in the commitment of muscle precursors in P19 cells are currently unknown (dashed arrows).
blast differentiation was shown to be dependent on the association of ␤-catenin with adherens junctions (54). This finding was consistent with a role for cadherins in skeletal muscle differentiation (55,56). Together these studies imply a dual role for ␤-catenin in myogenesis. First, ␤-catenin would function in the nucleus to activate transcription during commitment (this study), and, second, it would function in complexes with cadherins during differentiation of committed myoblasts (54). However, we cannot rule out a role for ␤-catenin/cadherinmediated junctions in commitment, and previous studies have shown a role for Wnt signaling in the regulation of MyoD and myogenin activity (36).
Schmidt et al. (29) put forth a model in which Wnt and SHH co-operate to induce ␤-catenin/Lef1 to active MRF expression in the somite. The mesodermal cells then become competent to receive the Wnt signal in the absence of SHH, and myogenesis is induced. We were unable to observe a specific induction of endogenous ␤-catenin expression during myogenesis in P19 cells. This may be due to the decrease in structural architecture found in the P19 cell aggregates compared with embryos as well as to the high levels of ubiquitous ␤-catenin expression found in P19 cells. In addition, we did not detect SHH during differentiation by Northern blot analysis (data not shown), although we cannot exclude the possibility that low levels of SHH are sufficient to play a role in skeletal muscle induction in the P19 cell system. Finally, aggregation may render the cells competent for myogenic induction by the Wnt/␤-catenin pathway. The role of SHH during skeletal muscle development of P19 cells is currently being investigated.
Our observation that ␤-catenin/engrailed inhibits myogenesis suggests that the Wnt/␤-catenin pathway is essential for mammalian skeletal myogenesis. Our findings are similar to those in Xenopus where mesoderm was present in embryos expressing the ␤-catenin/engrailed chimera, whereas somites did not form normally and the expression of dorsal markers was reduced (30). Expression of the Wnt antagonist Frzb1 in developing murine embryos suggested that Wnts are essential for skeletal myogenesis, although somites did form and factors such as Pax3 and Mox1 were not lost (10). We observed a total loss of markers expressed in the dorsal somite, including Pax3, Mox1, as well as Gli2 and Six1. The difference in the stage at which differentiation was blocked may be attributed to redundancy of Wnt signals in the embryo and an incomplete inhibition of embryonic Wnt activity by Frzb1 (57). In contrast, the ␤-catenin/engrailed fusion protein should block all Wnts that signal through the canonical pathway.
The observation that Brachyury T was down-regulated in P19[␤-cat/EnR] cells is consistent with a positive relationship between Brachyury T and Wnt3a (11,12). Although these studies showed that ␤-catenin directly activates the Brachyury T promoter, a recent study indicated that Wnt regulates but does not initiate Brachyury T expression in the embryo (58) suggesting that other factors play a crucial role in activating Brachyury T. In agreement with these findings, expression of the engrailed inhibitory domain results in a decrease but not loss of expression of Brachyury T, Wnt3a, and Wnt5b, similar to the disruption of mesoderm in Xenopus embryos treated with the ␤-catenin/engrailed chimera (30). In contrast, mice lacking ␤-catenin exhibited defects during gastrulation, with a total loss of Brachyury T expression and defective anterior-posterior axis formation (59,60). The earlier and more severe phenotype may implicate an early embryonic role for non-nuclear ␤-catenin.
In conclusion, we have shown that ␤-catenin, the downstream mediator of Wnt3a signaling, is both sufficient and essential for skeletal myogenesis in P19 cells. The finding that a dominant negative ␤-catenin inhibits the expression of several transcription factors present in the developing somite indicates that the Wnt signaling pathway is necessary for the specification of mesodermal cells to muscle precursors. Therefore, P19 cells represent an ideal system in which to study the roles of factors involved in specifying muscle precursors to the myogenic program.