MyoD transcription factor induces myogenesis by inhibiting Twist-1 through miR-206

Skeletal muscle tissue is derived from the differentiation of myoblasts into myotubes through a process known as myogenesis. Myogenesis is a complex and tightly regulated procedure. During myogenesis, the myoblasts exit the cell cycle through the orchestrated expression of specific genes that are important for muscle cell differentiation, and this is followed by their fusion to form multi-nuclei fibres known as myotubes (Sabourin and Rudnicki, 2000). Myogenesis occurs during different stages of human development. Specifically, myogenesis takes place during embryogenesis in order to form muscle tissue and occurs in adults to replace damaged or lost muscle (Wang and Conboy, 2010). During embryogenesis, the myogenic progenitor cells arise from the somites, which are sequentially combined blocks of mesoderm that form along the anterior–posterior axis of the developing embryo (Parker et al., 2003). Somites differentiate along the dorsal–ventral axis to give rise to the dorsally located epithelial dermomyotome and the ventrally located mesenchymal sclerotome (Parker et al., 2003). The dermomyotome gives rise to dermis and musculature, whereas the sclerotome forms the bone and cartilage. Myogenic precursors restrained in the epithelium of the dermomyotome express Pax3, Pax7 and low levels of the myogenic determination factor Myf-5. During the late stages of embryogenesis, a specific population of myogenic stem cells, called satellite cells, arise in order to provide most of the myonuclei to adult muscles during the postnatal growth of muscle tissue (Parker et al., 2003). At the molecular level, myogenesis is regulated by the basic helix-loop-helix (bHLH) myogenic regulatory factors (MRFs) Myf-5, MyoD, Myf-6 and myogenin. Genetic studies in mice have demonstrated that the genes encoding for the four MRFs are expressed in pairs. Myf-5 and MyoD are the first MRFs required for the commitment of progenitor cells to the myogenic lineage. Myf-6 and myogenin are the second group of MRFs, which are responsible for the downstream regulation of muscle differentiation (Relaix, 2006). MyoD is also expressed in proliferating myoblasts throughout the cell cycle in the absence of differentiation (Rudnicki et al., 1993).

MicroRNAs (miRNAs) are small non-coding regulatory RNA molecules, and their function is to negatively regulate gene expression at the post-transcriptional level (Bartel, 2004). miRNAs are incorporated with the RNA-induced silencing complex (RISC) and bind to the 3′ untranslated region (3′-UTR) of its target mRNA, causing either inhibition of protein translation or mRNA cleavage (MacFarlane and Murphy, 2010).

Many miRNAs are found to be expressed in muscle tissue. The most studied miRNAs in muscle are miR-1, miR-133a, miR-133b and miR-206, which are also called myomiRs owing to their muscle-specific expression. miR-1 and miR-133a are highly expressed in both skeletal and cardiac muscle, whereas miR-206 and miR-133b are specifically expressed in skeletal muscle (Chen et al., 2006; Kim et al., 2006). The expression of the four myomiRs is induced during muscle cell differentiation, indicating that they have a crucial role in regulating the process (Chen et al., 2006; Kim et al., 2006). Their expression is regulated by the MRFs MyoD and myogenin, as well as serum response factor (SRF) and myocyte enhancer factor 2 (MEF2) (Chen et al., 2006; Rao et al., 2006; Liu et al., 2007). MyomiR levels are found to be increased during the late stages of human foetal muscle development, and increases in their expression levels are proportional to the capacity of myoblasts to form myotubes (Koutsoulidou et al., 2011b).

An important molecule involved in myogenesis is Twist-1, which belongs to the family of bHLH transcription factors. Twist was initially identified in Drosophila (Thisse et al., 1987). Twist isoforms have been identified in other species, including human and mouse (Wolf et al., 1991; Wang et al., 1997). Twist-1 forms either homodimers or heterodimers with different bHLH protein partners that bind to specific DNA sequences called E-boxes (CANNTG) located in the promoters of target genes (Castanon et al., 2001). These complexes can act as repressors or activators of the target genes (O'Rourke and Tam, 2002). Twist-1 plays a crucial role during the development of the embryo. It is expressed during embryonic development, and it has been found to be involved in developmental processes such as myogenesis, neurogenesis, cardiogenesis, cranial tube morphogenesis and mesoderm formation (Miraoui and Marie, 2010). Twist-1 has also been found to be involved in the process of epithelial-to-mesenchymal transition, which plays an essential role in cancer metastasis (Yang et al., 2004). Regarding myogenesis, in Drosophila, Twist has been shown to promote myogenesis. However, in mouse C2C12 myoblasts and in human embryonic stem cell (HESC)-derived embryoid bodies, Twist-1 is found to inhibit muscle cell differentiation (Hebrok et al., 1994; Rohwedel et al., 1995; Cao et al., 2008; Koutsoulidou et al., 2011a). Interestingly, overexpression of Twist-1 reverses the process of muscle cell differentiation (Hjiantoniou et al., 2008; Mastroyiannopoulos et al., 2013). Twist-1 is transcribed primarily in the early somites, and its expression is downregulated when the myogenic factors MyoD and Myf-5 are upregulated in order to localise the newly formed somites to diverse compartments of the embryo (Sassoon, 1993; Hebrok et al., 1994; Castanon and Baylies, 2002). The downregulation of Twist-1 when the myogenic factors are expressed in early stage of development suggests that Twist-1 inhibits myogenesis and that it is involved in a process that prevents premature muscle cell differentiation (Castanon and Baylies, 2002). Overexpression of Twist-1 in the myogenic mouse cell line C2C12 reversibly represses muscle differentiation with an associated decrease in transcript levels of Myf-5 and myogenin (Hebrok et al., 1994). Moreover, it has been shown that overexpression of Twist-1 in the C2C12 cell line can reverse muscle cell differentiation in the presence of growth factors by binding and downregulating myogenin, as well as reverse cellular morphology in the absence of growth factors (Mastroyiannopoulos et al., 2013). Human Twist-1 is expressed endogenously at high levels in human foetal myoblasts, and its expression levels decrease during the late stages of development (Koutsoulidou et al., 2011a). The differentiation capacity of the myoblast increases during development. This shows that there is an inversely proportional relationship between the differentiation capacity of myoblasts and expression of Twist-1, and this indicates that Twist-1 is involved in the regulation of muscle development (Koutsoulidou et al., 2011a).

To date, not much information exists regarding the regulation of Twist-1 expression in muscle. In cancer, Twist-1 has been shown to be regulated by a series of miRNAs, such as miR-543, miR-720 and miR-181a (Liu et al., 2013; Bing et al., 2014; Li et al., 2014). There is no information, however, regarding Twist-1 regulation through miRNAs in muscle. Our investigation focused on the identification of miRNAs that could bind to and regulate Twist-1 expression during myogenesis. Our results show that miR-206 is a negative regulator of Twist-1 and promotes muscle cell differentiation. Moreover, we reveal that MyoD induces muscle cell differentiation by inhibiting expression of Twist-1 through miR-206, both in normal and myotonic dystrophy type 1 (DM1) muscle cells.

The initial aim of this work was to identify miRNAs that regulate the expression of the Twist-1 gene. The 3′-UTR of human Twist-1 was screened for potential miRNA-binding sites using miRanda software. The criteria that were used in order to choose the candidate miRNA were (1) the conservation of the miRNA-binding sites between species, (2) specific miRNA expression in muscle tissue and (3) opposing expression profiles of specific miRNAs and the Twist-1 protein. Among the possible candidate miRNAs, one of the known myomiRs, miR-206, was identified. The human Twist-1 transcript was predicted to contain one canonical miRNA response element (MRE) for miR-206 (Fig. 1A). The binding site of miR-206 to the human Twist-1 3′-UTR was more conserved between closely related species compared to divergent species, indicating possible evolutionary importance (Fig. 1B). To investigate whether the candidate miRNA binds to the 3′-UTR of human Twist-1, the full human Twist-1 3′-UTR was cloned downstream of the luciferase gene and assayed in HeLa cells (Fig. 1C). The luciferase assay indicated that miR-206 repressed luciferase activity efficiently (50%) (Fig. 1D). Introduction of mutations at the Twist-1-binding site of miR-206, and transfections with a miR-206 mutant variant, did not alter the luciferase activity levels, indicating miRNA–mRNA binding specificity (Fig. 1D).

It has been previously shown that overexpression of Twist-1 results in the inhibition of muscle cell differentiation in adult cells (Hebrok et al., 1994; Rohwedel et al., 1995). Before proceeding to evaluating Twist-1 as a possible target of miR-206, we overexpressed Twist-1 through an adenoviral vector in human myoblast cells isolated from a newborn; these cells have the ability to differentiate and to express low levels of Twist-1. High levels of exogenous Twist-1 resulted in a decrease in the differentiation capacity of cells when stained for myosin heavy chain (MyHC), a late muscle cell differentiation marker (supplementary material Fig. S1A). Moreover, reducing the endogenous levels of Twist-1 caused an increase in the differentiation capacity of the cells when stained for MyHC (supplementary material Fig. S1B).

As a next step, human myoblasts isolated from a 14-week-old foetus, which express high levels of Twist-1 and have low differentiation capacity, were used to determine the cellular effects of miR-206 acting on Twist-1 (Koutsoulidou et al., 2011a). The levels of miR-206 were increased following overexpression of miR-206 (supplementary material Fig. S2A) and, as a result, the levels of Twist-1 decreased (Fig. 2A, supplementary material Fig. S2B). The protein levels of muscle differentiation markers – troponin and muscle actin – were also increased in myoblasts that overexpressed miR-206 compared to those of control cells (Fig. 2A). The same cells that had been transfected with miR-206 were then induced to differentiate to form myotubes. Myoblasts that overexpressed miR-206 showed an increased capacity to differentiate in vitro compared to control cells (Fig. 2B). This was also shown by the increased fusion index (Fig. 2C).

Human muscle myoblasts isolated from a newborn, which express low levels of Twist-1 and have a high differentiation capacity, were then used for transfections with a miR-206 inhibitor (amiR-206) (supplementary material Fig. S2C). Following inhibition of miR-206, the levels of human Twist-1 protein increased, whereas the levels of the markers of differentiation, muscle actin and troponin, were decreased (Fig. 3A, supplementary material Fig. S2D). The differentiation capacity of the cells that had been transfected with amiR-206 was decreased compared to control cells (Fig. 3B,C).

These results support the findings of earlier experiments that Twist-1 is a target for miR-206 in order to induce muscle cell differentiation.

One of the main positive regulators of miR-206 is the transcription factor MyoD (Rosenberg et al., 2006). MyoD was found to regulate miR-206 expression through direct binding to its promoter. In an attempt to further elucidate the mechanism of Twist-mediated regulation of myogenesis, experiments were performed in order to investigate the possible implication of MyoD. As a first step, 14-week-old foetus myoblasts were transduced with an adenovirus expressing the MyoD transcription factor (AdM) (supplementary material Fig. S3A). MyoD overexpression increased the capacity of the cells to differentiate, as reflected by the myogenesis markers (muscle actin and troponin). At the same time, MyoD overexpression caused an increase in the levels of miR-206 (supplementary material Fig. S3B).

As a next step, the possibility of MyoD directly binding to the miR-206 promoter and its subsequent regulation of expression in 14-week-old human myoblasts were investigated. The miR-206 promoter contains two E-boxes that are highly conserved among species (Fig. 4A,B). A chromatin immunoprecipitation assay showed that endogenous MyoD binds to both E-boxes located on the miR-206 promoter in 14-week-old human myoblasts (Fig. 4C). Luciferase assays were next performed in order to investigate the effect of MyoD on miR-206 expression. Overexpression of MyoD in HeLa cells that had been transfected with a plasmid containing the promoter of miR-206 upstream of a luciferase gene showed that MyoD induces the expression of miR-206 (Fig. 4D,E). The direct binding of MyoD onto the human miR-206 promoter was further investigated using mutagenesis experiments. Single mutations were introduced into each of the E-boxes located on the promoter of miR-206. MyoD overexpression in HeLa cells that had been transfected with the mutated plasmids showed a dramatic reduction in luciferase activity. Introduction of mutations into both E-boxes located on the promoter of human miR-206 and overexpression of MyoD supressed luciferase activity (Fig. 4E).

The effect of MyoD overexpression on Twist-1 expression was next investigated. The levels of human Twist-1 protein were found to be significantly decreased in muscle cells overexpressing MyoD compared to control muscle cells (Fig. 5A). In order to investigate whether MyoD induces muscle cell differentiation through inhibition of Twist-1, we overexpressed MyoD and at the same time overexpressed Twist-1 using adenoviruses in 14-week-old human myoblasts. By overexpressing both MyoD and Twist-1, the inhibition of Twist-1 caused by the overexpression of MyoD was abolished (Fig. 5B). Following the co-expression of the two transcription factors MyoD and Twist-1, the cellular differentiation capacity was reduced compared to that upon overexpression of MyoD only. Specifically, the double-transduced cells had a similar capacity to differentiate to that of the control cells (Fig. 5C,D).

Based on our results, MyoD promotes muscle cell differentiation, induces the expression of miR-206 and inhibits Twist-1 expression. In order to investigate whether MyoD inhibits Twist-1 expression through miR-206 induction, MyoD was overexpressed in 14-week-old human myoblasts, and miR-206 was inhibited through expression of antagomiR. Analyses of the proteins showed that the levels of Twist-1 were increased compared to that in the cells that overexpressed either MyoD or miR-206 (Fig. 6A). Furthermore, the levels of Twist-1 protein in double-transfected cells were similar to those in the control cells (Fig. 6A). The levels of the myogenesis markers muscle actin and troponin were elevated when either MyoD or miR-206 was overexpressed (Fig. 6A). Moreover, double-transfected cells showed a decrease in their capacity to differentiate in vitro compared to that upon the overexpression of either MyoD or miR-206 alone, and a similar differentiation capacity to the control cells (Fig. 6B). The differences in muscle cell differentiation were also confirmed by fusion index analysis (Fig. 6C).

These results indicate, therefore, that MyoD can induce muscle differentiation by inhibiting Twist-1 through miR-206.

To further evaluate these findings, experiments were performed in primary mouse cells. This is possible because (1) the binding site of miR-206 on the Twist-1 3′-UTR is conserved between human and mouse (Fig. 1B), (2) the expression profile of miR-206 in mouse is inversely proportional to the expression of Twist-1 (Kim et al., 2006; Koutsoulidou et al., 2011b), (3) Twist-1 overexpression in mouse has the same negative effect on muscle cell differentiation (Hebrok et al., 1994; Rohwedel et al., 1995) and (4) the E-boxes bound by MyoD and found in the promoter region of miR-206 are conserved between mouse and human (Fig. 4B). Primary cells were isolated from a 4-week-old mouse; these express the Twist-1 gene at a low level (Dupont et al., 2001). Inhibition of MyoD or miR-206 caused an increase in Twist-1 protein levels in primary cells, in agreement with the results from human muscle cells (Fig. 7A). Furthermore, inhibition of MyoD or miR-206 reduced the capacity of primary cells to differentiate in vitro compared to that of the control cells (Fig. 7B,C). In order to investigate whether MyoD inhibits Twist-1 expression through miR-206, MyoD was inhibited and, at the same time, miR-206 was overexpressed in primary muscle cells. As a result, Twist-1 levels were decreased compared to those in the cells that had been subjected to inhibition of MyoD or amiR-206 alone (Fig. 7A). Furthermore, Twist-1 levels were similar to those in the control primary cells, and the differentiation capacity of the cells was restored (Fig. 7B,C).

Therefore, these results support the findings of earlier experiments performed in cultured cells and suggest that, through this newly identified mechanism, MyoD induces myogenesis by inhibiting Twist-1 through miR-206 (Fig. 7D).

The results of this study so far have shown that MyoD might exert its positive effect on myogenesis through an additional pathway, and more specifically through miRNA-mediated inhibition of Twist-1. The final step of this project aimed to investigate the role of this newly identified pathway in a disease setting. DM1 is an inherited neuromuscular disease that usually manifests during adulthood and is caused by a CTG expansion in the 3ʹ-UTR of the dystrophia myotonica protein kinase (DMPK) gene (Miller et al., 2000). A rarer and more severe congenital form of DM1 that is associated with a large CTG expansion also exists (Meola, 2013), and defects in the myogenic differentiation program have been reported in numerous DM1 cases (Amack and Mahadevan, 2004).

DM1 human myoblast cells, which have a low differentiation capacity, were used to evaluate the role of the newly identified pathway in the defective differentiation program of the disease. Three different DM1 muscle cell cultures derived from congenital DM1 individuals were examined. In all three cases, MyoD protein levels were inhibited compared to those in healthy cells (Fig. 8A). Twist-1 was found to have an inversely proportional expression profile compared to that of MyoD during differentiation both in DM1 and the healthy control muscle cells (Fig. 8A, supplementary material Fig. S4). Furthermore, the miR-206 levels were increased during the differentiation of healthy cells but inhibited in DM1 cells (Fig. 8A). Protein analysis showed that the levels of Twist-1 protein were increased in the DM1 cells that had been co-transfected with MyoD and amiR-206 compared to those in the cells that overexpressed either MyoD or miR-206 alone (Fig. 8B). Furthermore, the levels of Twist-1 protein in the double-transfected cells were similar to those in the control cells (Fig. 8B). Moreover, overexpression of MyoD together with amiR-206 resulted in a decrease in the capacity of cells to differentiate in vitro compared to expression of either MyoD or miR-206 alone. Co-transfection of MyoD and amiR-206 resulted in a similar differentiation capacity to that of control cells (Fig. 8C,D).

Therefore, these results further support the notion that this new myogenic pathway involving MyoD, Twist-1 and miR-206 might be used to overcome the defective differentiation program in DM1 in order to promote the formation of mature muscle cells.

The aim of this study was to identify miRNAs that regulate Twist-1 in muscle cells. Twist-1 is a transcription factor that belongs to the bHLH family, and it has been previously found to act as an inhibitor of muscle cell differentiation during embryonic development (Hebrok et al., 1994; Rohwedel et al., 1995; Spicer et al., 1996; Hjiantoniou et al., 2008; Koutsoulidou et al., 2011a). Results from this study show that MyoD induces the expression of miR-206, which post-transcriptionally regulates Twist-1 during skeletal myogenesis.

In cancer, Twist-1 has been shown to be regulated by a series of miRNAs, such as miR-543, miR-720 and miR-181a (Liu et al., 2013; Bing et al., 2014; Li et al., 2014). There is no information, however, regarding Twist-1 regulation through miRNAs in muscle. By using bioinformatics analysis of the 3′-UTR of Twist-1, one of the four myomiRs, miR-206, was predicted to be a post-transcriptional regulator of Twist-1. miR-206 was shown to target the 3′-UTR of Twist-1 mRNA, demonstrating that it is a strong candidate for binding to endogenous human Twist-1 mRNA.

Some targets of miR-206 are well established in myogenesis. Pax3 and Pax7 are two of the experimentally verified targets of miR-206. Pax7 and Pax3 prevent the early differentiation of myoblasts during myogenesis. Overexpression of miR-206 during the early stages of skeletal muscle development reduce the expression levels of Pax3 and Pax7, thus promoting muscle cell differentiation (Chen et al., 2010). Connexin 43 (Cx43) is another experimentally verified target of miR-206. It is an important gap junction channel that is necessary for the fusion of myoblasts to form differentiated mature myotubes. More specifically, Cx43 allows the passage of signalling molecules and metabolites that are necessary for the maturation of the myotubes (Kalderon et al., 1977; Anderson et al., 2006). miR-206 inhibits myoblast proliferation and promotes myoblast fusion through the downregulation of the Cx43 protein (Anderson et al., 2006). Another experimentally verified target of miR-206 that has been identified is Pola1. Pola1 is the largest subunit of DNA polymerase α active during DNA synthesis (Pellegrini, 2012). miR-206 has been shown to directly inhibit Pola1 expression during differentiation, and is thus associated with inhibition of the cell cycle during differentiation (Kim et al., 2006). miR-206 also negatively regulates follistatin-like 1 and utrophin through binding to the 3′-UTR of these mRNAs. Both of these are implicated in muscle cell differentiation (Rosenberg et al., 2006).

Apart from scoring highly using target-predicting software, miR-206 was chosen because of its importance in promoting muscle cell differentiation and because it is known that its expression levels are inversely proportional to those of Twist-1 during differentiation (Koutsoulidou et al., 2011a,b). In order to confirm the binding of miR-206 to the Twist-1 3′-UTR, a series of experiments were performed, including (1) luciferase assays using mutant plasmids for the seed-matched region of miR-206 located on the 3′-UTR of Twist-1 with a miR-206 mimic, which confirmed that only co-transfection with the wild-type reporter and the miR-206 mimic could reduce luciferase activity; (2) overexpression studies of miR-206 in which Twist-1 translation was suppressed and the differentiation capacity of human myoblasts that had been isolated from a 14-week-old foetus (high levels of Twist-1, low differentiation capacity) was increased; and (3) inhibition studies of miR-206 in a human muscle cell line that had been isolated from a newborn (low levels of Twist-1, high differentiation capacity), which showed increased endogenous Twist-1 protein levels and a decreased capacity for cell differentiation. These above findings reveal a new target for miR-206 through which muscle cell differentiation is promoted.

Because it is known that MyoD promotes the expression of miR-206 directly by binding to its promoter, we next investigated the possibility that this is also implicated in the regulatory pathway of Twist-1 (Sassoon, 1993; Berkes and Tapscott, 2005; Rao et al., 2006). MyoD regulates many pathways during muscle cell differentiation. Specifically, it has been shown to activate p21 expression during differentiation of murine muscle cells (Halevy et al., 1995). MyoD also interacts with several proteins, such as cyclin-dependent kinase 4, HDAC1, STAT3 and many others, in order to promote muscle cell differentiation (Zhang et al., 1999a,b; Mal et al., 2001; Puri et al., 2001; Kataoka et al., 2003).

MyoD is known to bind to the E-box sequence and to regulate the expression of muscle-specific genes (Weintraub et al., 1994; Shklover et al., 2007). By scanning the promoter of human miR-206, two highly conserved E-boxes were found to be present. The binding of MyoD to both E-boxes on the promoter of miR-206 was confirmed by chromatin immunoprecipitation analyses, which confirmed the direct binding of MyoD to the promoter of miR-206 to both sites, and by using luciferase assays, which confirmed that binding of MyoD to the miR-206 promoter induces miR-206 expression.

Overexpression of MyoD promotes muscle cell differentiation and the expression of miR-206 (Crescenzi et al., 1990; Rao et al., 2006). From our results, it was found that overexpression of MyoD also reduced the expression level of Twist-1 in 14-week-old foetus myoblasts. Competition experiments with MyoD and Twist-1 overexpression proved this specific association. Overexpression of MyoD induced muscle cell differentiation, whereas the overexpression of both Twist-1 and MyoD brought the levels of cell differentiation almost back to normal, suggesting that MyoD regulates differentiation through downregulation of Twist-1. This was further investigated, and our results showed that MyoD inhibited Twist-1 through miR-206 induction and subsequently promoted muscle cell differentiation. The implication of this new pathway in myogenesis was also demonstrated in primary mouse muscle cells.

Finally, we wanted to investigate whether this newly identified pathway is involved in a disease causing defective muscle cell differentiation. DM1 is the most common form of muscular dystrophy in adults. Symptoms of DM1 include myotonia, muscle weakness and progressive muscle atrophy (Meola, 2000). DM1 is caused by an unstable expansion of CTG trinucleotide repeats found in the 3′-UTR of the DMPK gene (Brook et al., 1992; Fu et al., 1992; Mahadevan et al., 1992). C2C12 cells containing mutant DMPK 3′-UTR transcripts assemble to nuclear foci and do not undergo differentiation (Amack et al., 1999). This differentiation deficiency might represent the muscle development abnormalities found in congenital DM1 individuals (Sarnat and Silbert, 1976; Farkas-Bargeton et al., 1988). Differentiation inhibition has been confirmed in cultured myogenic satellite cells taken from DM1 individuals carrying a large CTG expansion (Furling et al., 2001; Timchenko et al., 2001). Mutant DMPK 3′-UTR transcripts have also been shown to disrupt myoblast differentiation by reducing MyoD levels (Amack et al., 2002), and an increase in MyoD levels is sufficient to rescue the differentiation defect in DM1 myoblasts (Amack and Mahadevan, 2004). Our results showed that congenital DM1 cells, which have a defective differentiation program, have low levels of MyoD and miR-206 but high Twist-1 levels. This seems rational based on the properties that characterise these three molecules during muscle cell differentiation. In an attempt to prove that this pathway plays a role in the defective differentiation in those cells, we overexpressed MyoD and determined the downstream effects. As expected, MyoD overexpression did correct muscle cell differentiation and, moreover, miR-206 and Twist-1 levels were upregulated and downregulated, respectively. Moreover, competition with an antagomiR for miR-206 demonstrated the specificity and effectiveness of the pathway with regards to the promotion of cell differentiation. Therefore, based on these results, the MyoD–miR-206–Twist-1 pathways is compromised in DM1 cells that exhibit a defective differentiation program.

In summary, our results newly identify a mechanism by which MyoD induces muscle cell differentiation through the induction of miR-206 expression and the subsequent inhibition of Twist-1. This mechanism might play a valuable role in myogenesis and also in diseases where differentiation is defective.

Human myoblasts were isolated from muscle biopsies obtained during autopsies of normal foetuses (14, 31, 37 weeks old) and congenital DM1 foetuses (31, 37 weeks old and newborn) in accordance with French legislation on ethics rules. Informed consent was obtained and the study conformed to the Declaration of Helsinki. Myoblasts were isolated as previously described, and grown in Dulbecco's modified Eagle medium (DMEM; Invitrogen, Carlsbad, CA) supplemented with 20% foetal bovine serum (FBS; Invitrogen, Carlsbad, CA) (Edom et al., 1994; Furling et al., 2001). When cells reached confluence, they were differentiated with DMEM supplemented with 2% horse serum (HS; Invitrogen, Carlsbad, CA). Human myoblasts were culture under 5% CO₂ at 37°C. HeLa cells were grown to 90% confluence before being subjected to transfections in growth medium using DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% (v/v) FBS (Invitrogen, Carlsbad, CA), 2 mM glutamine (Invitrogen, Carlsbad, CA) and penicillin-streptomycin (100 mg/ml-100 units/ml) (Invitrogen, Carlsbad, CA). For primary mouse cell extraction, 4-week-old mice were euthanized through cervical dislocation, and the muscles were carefully removed from the hind limb in accordance with Cyprus legislation on ethics rules. Myofibres were isolated from the extensor digitorum longus (EDL) muscle as described previously and digested in 0.2% collagenase type 1 for 1 h at 37°C (Rosenblatt et al., 1995). Muscle fibres were next purified from fibroblasts and cell debris. The washed muscle fibres were next placed in Corning 6-well plates coated with 1 mg/ml Matrigel (Corning, Tewksburym, MA) and incubated in activation medium [DMEM with 20% (v/v) foetal bovine serum (FBS) (Invitrogen, Carlsbad, CA), 10% (v/v) horse serum (Invitrogen, Carlsbad, CA), 1% (v/v) chick embryo extract (MP Biomedicals, France), 0.01% murine FGF-basic (Peprotech, Rocky Hill, NJ), and 1% (v/v) penicillin-streptomycin solution (Sigma-Aldrich, St Louis, MO)] at 37°C for 48 h. The muscle fibres were removed from the plate by gentle blowing, leaving the satellite cells attached to the plate. The satellite cells were further purified by re-plating the remaining population onto non-treated plates, where the satellite cells remained in suspension. The suspended satellite cells were then plated into Corning 6-well plates coated with 1 mg/ml Matrigel and maintained in activation medium at 37°C in 5% CO₂. After 4 days of proliferation, the medium of the primary cells was changes to differentiation medium (used for culture of human myoblasts).

For HeLa and human myoblasts, 100 pmol of each miRNA mimic (Qiagen, Limburg, The Netherlands), antimiR (Ambion, Carlsbad, CA), negative miRNA (Qiagen, Limburg, Netherlands), mutated miRNA (Eurofins Genomics, Ebersberg, Germany), or anti-miR control (Ambion, Carlsbad, CA) was transfected using X-tremeGENE 9 transfection reagent (Roche, Penzberg, Upper Bavaria, Germany) in Opti-MEM reduced serum medium (Invitrogen, Carlsbad, CA). After 24 h, cells were then transfected with the wild-type or mutated Twist-1 3′-UTR luciferase plasmid or miR-206 promoter luciferase plasmid (400 ng), and CMV-pRLRenilla (5 ng) (Promega, Madison, WI) plasmid in complex with Lipofectamine 2000 (Invitrogen, Carlsbad, CA). For transfections into primary mouse muscle cells, the same transfection procedure was repeated. For small interfering (si)RNA transfections, 100 pmol of MyoD siRNA (Invitrogen, Carlsbad, CA) and siRNA negative control (Invitrogen, Carlsbad, CA) were combined with Lipofectamine RNAi or RNAiMAX (Invitrogen, Carlsbad, CA). MyoD (AdM), (Vector Biolabs, Malvern, PA) and Twist-1 (AdT) (Vector Biolabs, Malvern, PA) adenoviruses were used to overexpress MyoD and Twist-1. AdC (Vector Biolabs, Malvern, PA) was used as a control virus, containing an empty vector with a CMV promoter. Cells were transduced with the adenoviruses a day before confluence for 48 h.

Total RNA enriched for small RNAs, including miRNAs, was isolated using mirVana™ miRNA Isolation Kit (Ambion, Carlsbad, CA) according to the manufacturer's protocol. 10 ng of the extracted RNA was subjected to reverse transcriptase PCR using the TaqMan^® MicroRNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA). miRNA levels were measured by using real-time PCR amplification using TaqMan^® MicroRNA Assays specific for miR-206 (Applied Biosystems, Carlsbad, CA), according to the manufacturer's instructions. miRNA expression was normalized to that of the RNA U6B small nuclear (RNU6B) (Applied Biosystems, Carlsbad, CA). Real-time PCR was performed using the Applied Biosystems 7900 HT Fast Real-Time PCR System (Applied Biosystems, Carlsbad, CA).

Human Twist-1 3′-UTR was amplified by using PCR from genomic DNA extracted from human myoblasts, and this was then cloned into the multiple cloning site of the pMIR-Report luciferase miRNA expression reporter (Applied Biosystems, Carlsbad, CA). The PGL3 plasmid with the promoter of miR-206 was kindly provided by the laboratory of Dr Stephen Tapscott Fred Hutchinson Cancer Research Center, Seattle, WA.

The predicted miR-206-binding site located in the Twist-1 3'-UTR and the predicted MyoD sides on the promoter of miR-206 were mutated by a substitution of 6 bp (Twist-1 3'-UTR: CATTCT to GCCGGT, miR-206 promoter: ACAGCT to TGTCGA and GCAGCT to CGTCGA) using the GeneArt^® Site-Directed Mutagenesis System (Invitrogen, Carlsbad, CA). All of the cloned vectors were verified by sequencing.

Following transfections, HeLa cells were harvested, and assays were performed 24 h after the last transfection using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI). Firefly luciferase activity was normalized to Renilla luciferase expression (internal control). Cells were subsequently lysed with commercial cell lysis buffer (Promega, Madison, WI), and luciferase activity was measured using a luminometer (Berthold, Bad Wildbad, Germany) according to kit protocols.

One day before inducing differentiation, cells were transfected with AdM (MyoD-overexpressing adenoviral vector). The chromatin immunoprecipitation assay was performed on myotubes that had been differentiated for 2 days using MAGnify™ Chromatin Immunoprecipitation System (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Briefly, dynabeads were coupled to the anti-MyoD primary antibodies (Abcam, Cambridge, UK) before the crosslinking of chromatin. Following cell lysis, samples were subjected to chromatin binding to the antibody–dynabead complexes in order to isolate only the DNA of interest after a series of washes. The cells were immunoprecipitated using anti-MyoD antibodies (Abcam, Cambridge, UK). For chromatin immunoprecipitation assay, two controls were used – a positive- and a negative-control antibody. For positive control, 2.5 μg of unconjugated polyclonal antibody specific to human and mouse histone H3 trimethylated at lysine 9 (K9me3) (H3-K9Me3) (Invitrogen, Carlsbad, CA) was added. For negative control, 1 μg of mouse IgG antibody was used. Primers for the chromatin immunoprecipitation assay were design for the promoter of miR-206, E1 forward primer, 5′-CAGTGAACAATGGTGCTTGG-3′; E1 reverse primer, 5′-TTCCACATTCACGCAGAGAG-3′; E2 forward primer, 5′-AACCCCATCTCCCTCCAG-3′; and E2 reverse primer, 5′-GATCCTTTTGTCGGGCTTCT-3′.

Proliferating myoblasts or differentiated muscle cells were used for protein extractions. 30–50 μg protein extracts were incubated with primary antibodies against Twist-1 (1:100, Santa Cruz Biotechnology, Dallas, TX), troponin (1:200, Santa Cruz Biotechnology, Dallas, TX), skeletal actin (1:200, Santa Cruz Biotechnology) or GAPDH (1:1500, Santa Cruz Biotechnology, Dallas, TX), followed by incubation with goat anti-mouse IgG or donkey anti-rabbit IgG secondary antibodies. Differentiated human myoblasts or primary mouse myoblasts were fixed in 4% paraformaldehyde and incubated with a monoclonal antibody against myosin heavy chain (MyHC) (Sigma-Aldrich, St Louis, MO) at a concentration of 1:400 in 1% BSA in PBS or Twist-1 (Abcam, Cambridge, UK) at a concentration of 1:50 in 1% BSA in PBS and Alexa-Flour-conjugated anti-mouse secondary antibody (Life Technologies, Carlsbad, CA). Nuclei were stained with Hoechst 33342 (Invitrogen, Carlsbad, CA). Images were taken using a Zeiss Axiovision digital camera (Zeiss, Oberkochen, Germany) and then accumulated using Adobe Photoshop Software. Cells were counted at least three times from each pool of clones in ten different cellular areas to determine the fusion index (percentage of nuclei located under myotubes divided by the total number of nuclei).

ANOVA and Student's t-test were used to determine whether specific group mean differences were significant. The level of significance was set at 0.0. Data are presented as mean±s.d.

We would like to thank Dr Stephen Tapscott for providing us with the PGL3 plasmid containing the promoter of miR-206.