Twist reverses muscle cell differentiation through transcriptional down-regulation of myogenin

The course of myogenesis is a well-characterized example of terminal differentiation. Myoblasts are capable of proliferation and upon demand to form skeletal muscle, these cells exit the cell cycle and through the activation of muscle-specific transcription factors they fuse into multinucleated terminally differentiated myotubes [1,2]. MRFs (myogenic regulatory factors), myogenin, MyoD, MRF4 (Myf6) and Myf5 are bHLH (basic helix-loop-helix) transcription factors that regulate myogenesis [3–9]. Myogenin is essential during differentiation. Mice lacking the myogenin gene die at birth due to severe skeletal muscle deficiency, as myoblasts are unable to fuse into multinucleated myotubes [10]. Furthermore, MyoD and Myf5 are unable to substitute myogenin's function during differentiation [11]. Moreover, mice lacking the myogenin gene express normal levels of MyoD and Myf5 [10].

Unlike mammals, vertebrates such as zebrafish and salamanders can display unique regenerative abilities through dedifferentiation or differentiation of precursor cells [12]. Following injury, these vertebrates are able to induce reversal of the differentiation state, which leads to a series of events that aim to generate proliferating regenerative progenitor cells with the ability to restore the lost tissue in a precise way [12–14]. Some research groups have attempted to induce dedifferentiation of muscle cells by exogenous genes or chemicals. Mouse C2C12 myotubes treated with limb regeneration extracts were able to induce myotubes to reenter the cell cycle, exhibited reduced levels of muscle differentiation proteins and cleaved to produce smaller myotubes or proliferating mononucleated cells [15]. In another study, combination of growth medium and ectopic msx1 expression caused the reduction of muscle-specific proteins and the cleavage of these myotubes into proliferating mononucleated cells that were able to redifferentiate into muscle or trans-differentiate into various cell types [16]. Microinjection of Barx2 cDNA into immature myotubes derived from primary cells led to cleavage and formation of mononucleated cells that were able to proliferate [17]. Using a chemical approach, terminal differentiated myotubes were incubated with a triazine compound. Myotubes showed to cellularize into smaller myotubes or mononucleated cells, which were able to survive and divide [18]. Similarly, myoseverin, a trisubstituted purine, was shown to induce reversible fission of multinucleated myotubes into mononucleated cells, which were able to enter the cell cycle [19]. Recently, mammalian skeletal muscle cells were induced to dedifferentiate into proliferating mononuclear cells following treatment with myoseverin and temporary p21 suppression. These cells were further induced to act as multipotent stromal cells by further treatment with the small molecule, reversine (2-(4-morpholinoanilino)-6-cyclohexylaminopurine) and simple chemical modifications of the culture media [20]. When cell cycle inhibitors, p21 and p27 were depleted from terminal differentiated mouse myotubes, incomplete DNA replication and apoptosis was observed. In contrast, when p21 and p27 were depleted from quiescent, non-terminal differentiated fibroblasts and muscle cells, DNA replication was fully recovered and apoptosis was no longer observed. These cells were able to proliferate in the absence of growth factors [21]. Recently, evidence for natural dedifferentiation of muscle cells following injury was reported by using a Cre/Lox-β-galactosidase system [22,23]. Finally, we have recently reported that down-regulation of myogenin leads myotubes to a reversal of muscle cell differentiation [24].

TWIST is a bHLH transcription factor initially identified in Drosophila [25]. Twist orthologues have subsequently been identified in other species, including mouse and human [26,27]. It forms functional homodimers as well as heterodimers with various bHLH protein partners and binds to the promoter of target genes. TWIST is expressed during embryonic development and plays critical roles in diverse developmental systems such as mesoderm formation, myogenesis, cardiogenesis and neurogenesis [28].

Several experiments, mainly involving overexpression of TWIST in cell lines, have demonstrated its role in inhibition of muscle cell differentiation. Nevertheless, in Drosophila, TWIST has been reported to promote myogenesis [29], whereas in mice it inhibits muscle cell differentiation [30,31]. Published results, mainly by overexpression of TWIST in cell lines and mice, show that TWIST inhibits myogenesis by inhibiting MyoD, which is activated by binding to MEF (myocyte enhancer factor)-2 and E-proteins [25,32]. Not a single mechanism has been assigned to MyoD inhibition by TWIST. There is however evidence of direct TWIST binding to MyoD [33] or TWIST binding and sequestration of MEF-2 or E-proteins [34,35].

We have previously shown that overexpression of TWIST in terminally differentiated myotubes caused their cleavage to mononucleated cells in the presence of growth factors and re-entry to the cell cycle. This was accompanied by a reduction of MRF (myogenic regulatory factor) levels [36]. Here, we wanted to investigate the direct mechanism by which TWIST reverses muscle cell differentiation. We show that TWIST causes reversal of muscle cell differentiation through binding and down-regulation of myogenin. Moreover, reversal of cellular morphology (myotube fragmentation) was possible through this pathway, in the absence of growth factors.

C2C12 mouse myoblasts (ECACC) were grown to confluency under 5% (v/v) CO₂ at 37°C in the GM (growth medium), DMEM (Dulbecco's modified Eagle's medium) (Gibco) supplemented with 10% (v/v) FBS (Gibco), 2 mM glutamine (Gibco) and penicillin-streptomycin (100 μg/ml–100 units/ml) (Gibco). For C2C12 muscle cell differentiation, cells were then switched to DM (differentiation medium), DMEM supplemented with 2% (v/v) horse serum (Gibco), 2 mM glutamine and penicillin–streptomycin (100 μg/ml–100 units/ml) for 4 days. During the first 2 days of differentiation, Ara-C (cytosine β-D-arabinofuranoside) (Sigma) (4 μg/ml) was included in order to eliminate most undifferentiated myoblasts (for time-lapse microscopy purposes). Medium was then replaced with the fresh DM medium without Ara-C in the presence or absence of adenoviral vectors, for another 2 days (day 4 of differentiation). For adenoviral transductions on differentiated myotubes, 100 MOI (multiplicity of infection) of each of the adenoviral vectors (VectorBiolabs) was used. Cells were maintained in 37°C/5%CO₂ or in a temperature and CO₂-regulated time-lapse microscope (AxioVision, Zeiss).

C2C12 cells were grown to confluency after which DM was added. On day 2 of differentiation cells were transfected with AdT (TWIST-overexpressing adenoviral vector) or AdMyoD (MyoD-overexpressing adenoviral vector). ChIP (chromatin immunoprecipitation) assay was performed on 4-day differentiated C2C12 myotubes using MAGnify™ Chromatin Immunoprecipitation System (Invitrogen) according to the manufacturer's instructions. Briefly, dynabeads were coupled to the MyoD and TWIST primary antibodies (Abcam) prior to the crosslinking of chromatin. Following cell lysis, samples were subjected to chromatin binding to the antibody dynabeads complexes in order to isolate only the DNA of interest after a series of washes. The cells were immunoprecipitated using TWIST or MyoD antibodies, respectively (Abcam). For ChIP 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) was added. For negative control, 1 μg of rabbit or mouse IgG antibody was used.

Primers for ChiP assay were E1 F: 5′- GTTTCTGTGGCGTTGGCTAT -3′, E1 R: 5′ AAGGCTTGTTCCTGCCACT 3′. E2 F: 5′- AATCAAATTACAGCCGACGG -3′, E2 R: 5′- GAAACGTCTTGATGTGCAGC -3′. E3 F: 5′- CAAGGAACTGAAGGGGTCTG -3′, E3 R: 5′- CCCTGTACTGGGGCATATAGTT -3′. Sat2 F: 5′- AGCAGATGGCTTTGGAGAGA 3′, Sat2 R: 5′- CTGGGAGCAACCCTTATTCA -3′ (Table 1).

Mouse myogenin promoter, as shown in figure 1, was cloned into a luciferase-pcDNA3 plasmid upstream of the luciferase gene. The myogenin promoter/luciferase plasmid was mutated using the GeneArt Site-Directed Mutagenesis System (Invitrogen) at myogenin promoter E-box site (CATATG) (E3) by using two different sets of primers: Mutant 1 F: 5′- AGAGCTCATGTCTCTAGCTGCGGATGTAGCAGAA -3′, R: 5′- GCAGCTAGAGACATGAGCTCTGGGGGTACTGG -3′ and Mutant2 F: 5′- AGAGCTCATGTCTCTAGCTGCTGGTATAGCAGAAGAT -3′, R: 5′- GCAGCTAGAGACATGAGCTCTGGGGGTACTGG -3′. C2C12 cells were stably transfected with the wild-type and mutant myogenin/luciferase plasmids. Cells expressing the wild-type or mutant myogenin promoters were differentiated for 2 days before being transfected with AdT or AdC (control adenoviral vector) after which cells were differentiated for a further 2 days. Luciferase gene expression was then measured using Dual-Luciferase Reporter Assay (Promega). Cells were lysed and mixed with Luciferase Assay Substrate (LAR II) before measuring activity in a luminometer (Berthold). In order to ensure that luciferase readings reflect results originating from viable cells, the viability of cells in experiment was measured using a cell counter (Countess Automated cell counter, Invitrogen).

Cell cycle studies were performed using Click-iT EdU Imaging kit (Invitrogen). Myotubes were transfected with adenoviral vectors (day 2 of differentiation) and incubated in GM or DM for a further 2 days. Cells were then assayed with EdU (5-ethynyl-2′-deoxyuridine) for 3 h (Invitrogen). Following incubation with EdU cells were then fixed with 4% (v/v) paraformaldehyde in PBS for 20 min, washed twice with 3% (w/v) BSA in PBS, before and after permeabilization with 0.5% (v/v) Triton-X-100 in PBS for 20 min at room temperature (25°C) prior to 30 min incubation in a mixture containing Alexa fluor 647. Nuclear staining was performed with DAPI II (1.5 ng/μl) (Vysis) and observed under a fluorescence microscope (AxioVision, Zeiss).

Total RNA was extracted from transfected or untransfected myotubes (Perfect RNAEukaryotic Mini kit, Eppendorf) and 500 ng of RNA was subjected to reverse transcription using MuLV reverse transcriptase (NEB). For PCR, mouse-specific primers were used for the analysis of expression for the following molecules: MyoD F 5′- AGTGAATGA GGCCTTCGAG A -3′, R 5′- GCATCTGAGTCGCCACTGTA -3′. Myogenin F 5′-CATCCAGTACATTG AGCGCCTA-3′, R 5′- GAGCAAATGATCTCCTGGGTTG -3′. Myf6 F 5′- ATG GTACCCTATCCCGTTGC -3′, R 5′- TAGCTGCTTTCCGACGATCT -3′. Myf5 F 5′- TGAAGGATGGACATGACGGACG -3′,R 5′-TTGTGTGCTCCGAAGGCTGCTA -3′. Cyclin D1 F 5′- GGCACCTGGATTGTTCTGCT -3′, R 5′- CAGCTTGC TAGGGAACTTGG -3′. Cyclin E2 F 5′ -GGAACCACAGATGAGGTC -3′, R 5′- CG TAAGCAAACTCTTGGAG -3′. Pax7 F: 5′- GAGTTCGATTAGCCGAGTGC -3′, Pax7 R: 5′- CGGGTTCTGATTCCACATCT -3′. GAPDH F 5′-TCATCATCTCCGCCC-CTTCT-3′, R 5′- GAGGGGCCATCCACAGTCTT -3′ (Table 2).

Cells were lysed using a protein lysis buffer including protease inhibitor. 40–60 μg of protein extracts were incubated with myogenin (1/200; BD), MyoD, Myf5, Myf6 (1/300; Santa-Cruz), cyclin D1 (1/400; Abcam), cyclin E2 (1/200, Abcam) and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (1/2000; Santa-Cruz) primary antibodies followed by incubation with goat anti-mouse IgG or donkey anti-rabbit IgG secondary antibodies conjugated to horseradish peroxidase (Santa-Cruz).

We have recently shown that down-regulation of myogenin can reverse muscle cell differentiation by cleavage into mononuclear cells which enter the cell cycle. Since TWIST is known as a transcription factor that negatively regulates muscle cell differentiation, we first looked at the putative E-boxes on the mouse myogenin promoter, which is located on chromosome 1 and is 1.5 kb in length (chr1:134288446–134290053) (Figure 1A). TWIST is known to bind to E-boxes. With this in mind, we tested the binding affinity of TWIST with E-boxes found in myogenin promoter. In order to determine whether Twist can bind to the myogenin promoter and therefore contribute to the mechanism of reversal of differentiation, C2C12 muscle cells were first differentiated in vitro to multinucleated myotubes and then transduced with an AdT. Following incubation with the viral vector, the ability of TWIST to bind to the myogenin promoter was tested by ChIP. As it can be seen in Figure 1B, TWIST was able to bind the myogenin promoter. This was possible only for an E-box (E3) that was found 1083 bp upstream of the transcription start site. The two well-characterized E-boxes (E1 and E2) of the myogenin promoter showed not to have binding affinity with TWIST (Figure 1B). Furthermore, a AdMyoD was in vitro transduced in differentiated myotubes using the same conditions to the previous AdT transfections. This was done in order to investigate whether MyoD interacts with the same E box (E3) and acts in competition to TWIST protein. Following adenoviral transductions of MyoD, ChIP assay showed that MyoD did not interact with E3 (Figure 1B). In order to determine the effect of the binding of TWIST to myogenin, myogenin promoter (sequence shown in Figure 1A) was cloned upstream of the luciferase gene in a plasmid and a stable cell line was created. Following transduction of AdT, luciferase activity was considerably reduced, whereas no reduction was observed in the transduction with a control AdC (adenoviral vector) (Figure 1C). This result demonstrates that TWIST is able to bind and down-regulate myogenin promoter activity. As a next step, two specific mutations (mutants 1 and 2) were introduced into the E3 site in order to determine the specificity of binding. In both mutants, the activity of luciferase remained unchanged, similar to AdC denoting that TWIST binding was specific.

Based on our previous reports, myogenin down-regulation caused reversal of muscle cell differentiation with cellular and molecular characteristics which resemble those of TWIST-mediated reversal of differentiation [24,36]. Binding of TWIST to myogenin promoter indicates that TWIST mediates its effect via myogenin down-regulation. All previous reports investigating induced reversal of differentiation included the addition of growth factors in the procedures. Growth factors are essential for the proliferation of muscle cells since they induce pathways involved in cell division. In vitro differentiation, on the other hand proceeds in the absence of growth factors since cells must exit the cell cycle. Previous experiments have shown that newt myotubes are able to dedifferentiate in the absence of growth factors [37]. The possibility that TWIST induces reversal of muscle cell differentiation in the absence of growth factors was next investigated.

Differentiated muscle cells were transduced with AdT and then induced to reverse their cellular fate and morphology by the addition or the absence of growth factors. Following cell incubation under a time-lapse microscope, around 80% of myotubes were shown to be cleaved in individual mononuclear cells with and without growth factors, compared with control-transduced cells (Figures 2A, 2B and 2D). To determine whether product cells initiated DNA synthesis EdU incorporation assays were performed on product cells that originated from the cleavage of myotubes in the absence of growth factors, did not show any EdU incorporation. In contrast, DNA synthesis was seen in approximately 90% of product cells in the presence of growth factors (Figures 2C and 2D). In conclusion, although cells were morphologically changed, indicating reversal of differentiation, the absence of growth factors prevented them from entering the cell cycle. This result indicates that although TWIST cannot sustain a full reversal of differentiation (cell cycle re-entry) in the absence of growth factors, it is capable of reversing the morphology of differentiated muscle cells.

In order to investigate the molecular changes of the absence of growth factors in the reversal of muscle cell differentiation, RNA and protein analysis was performed in TWIST-transduced and control myotubes. Our previous work demonstrated that the presence of growth factors in the AdT transductions induced the expression of cell cycle molecules such as cyclin D1 and suppressed the MRFs. Here, as it can be seen in Figures 3(A) and 3(B), the absence of growth factors had no influence on any of the cell cycle molecules or MRFs. The only molecule that is altered is myogenin which was found reduced compared to control-transduced cells. This result indicates that in the absence of growth factors, TWIST specifically binds the myogenin promoter, resulting in the down-regulation of myogenin, which drives the reversal of myotube morphology causing the cleavage of myotubes into individual mononuclear cells. Finally, the expression levels of Pax7 were assessed, since this molecule is known to be reduced when myogenin is increased [38]. Following experimentation, PAX7 was found unchanged following TWIST overexpression, both in the presence and absence of growth factors (Figure 3B).

This study reports a novel pathway which leads to the reversal of muscle cell differentiation. Moreover, results from this study indicate that TWIST can achieve reversal of muscle cell differentiation (cleavage of myotubes into mononuclear cells) in the absence of growth factors.

As of the date of submission, there is no strong evidence indicating that dedifferentiation of muscle cells occurs in mammals. There is however published work which shows that following induction with genetic or chemical molecules, reversal of differentiation is plausible. Since this pathway occurs naturally in some higher vertebrates as an additional way for regenerating muscle, it may well be beneficial for humans, if appropriate inducers can mimic this phenomenon. Induced reversal of muscle differentiation may also reveal unknown functions of molecules and novel pathways in the regulation of myogenesis. For example, reversal of muscle differentiation by down-regulating myogenin shows the important role of this MRF to maintain differentiation [24]. Interestingly, from previous studies Pax7 was found to be down-regulated as myogenin was induced. In our case however, Pax7 was not found altered following down-regulation of myogenin by TWIST either in presence or absence of growth factors [38]. This could imply that the molecular mechanism of dedifferentiation following overexpression of TWIST may involve such pathways which do not lower the levels of Pax7.

Results from this study reveal a mechanism by which overexpression of TWIST reverses muscle cell differentiation. Binding of TWIST to the E-box of the promoter of myogenin suppressed its activity and down-regulated its expression. As a result, differentiation is reversed by the subsequent reduction in the other MRFs and the entry of the product cells to the cell cycle. Interestingly, MyoD was not found to bind to the same E box that TWIST binds, implying that there is no competition in the particular E box, although MyoD has been previously shown to bind to the myogenin promoter [39,40]. Findings from this work show also that TWIST-mediated myogenin reduction leads, independently of cell cycle entry, to the cell morphological reversal of differentiation, i.e. the myotube cleavage into mononucleated cells. This, not only demonstrates that TWIST is a potent inducer of morphological reversal of muscle cells but it also indicates that myogenin down-regulation is responsible for the cleavage of terminally differentiated myotubes into mononucleated cells independently from the cell cycle re-entry.

This work provides mechanistic insights to TWIST-mediated reversal of muscle cell differentiation, which might be beneficial for muscle regeneration strategies and also for the investigation of novel pathways in myogenesis.

AdC

control adenoviral vector

AdMyoD

MyoD-overexpressing adenoviral vector

AdT

TWIST-overexpressing adenoviral vector

Ara-C

cytosine β-D-arabinofuranoside

bHLH

basic helix-loop-helix

ChIP

chromatin immunoprecipitation

DM

differentiation medium

EdU

5-ethynyl-2′-deoxyuridine

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GM

growth medium

MEF

myocyte enhancer factor

MRF

myogenic regulatory factor

Nikolaos Mastroyiannopoulos carried out the cell culture and time-lapse experiments, adenovirus transfections, RNA and protein analysis, immunocytochemistry, luciferase assay experiments and drafted the paper. Antonis Antoniou participated in the cloning of myogenin promoter. Andrie Koutsoulidou participated in the ChiP assay experiments. James Uney participated in project design. Leonidas Phylactou conceived the study and participated in its design and co-ordination. All authors read and approved the final paper.

This work was supported by the A.G. Leventis Foundation (L.A.P.)