Transforming Growth Factor β (TGF-β) Signaling Is Regulated by Electrical Activity in Skeletal Muscle Cells

Transforming growth factor (TGF-β) is involved in several cellular processes such as cell proliferation, differentiation, and apoptosis. At the cell surface, TGF-β binds to serine-threonine kinase transmembrane receptors (type II and type I) to initiate Smad-dependent intracellular signaling cascades. During the early stages of skeletal muscle differentiation, myotubes start to evoke spontaneous electrical activity in association with contractions that arise following the maturation of the excitation-contraction apparatus. In this work, we report that TGF-β-dependent signaling is regulated by electrical activity in developing rat primary myotubes, as determined by Smad2 phosphorylation, Smad4 nuclear translocation, and p3TPLux reporter activity. This electrical activity-dependent regulation is associated with changes in TGF-β type I receptor (TβRI) levels, correlated with changes in transducing receptors at the cell membrane (measured through radiolabeling binding assays). The inhibition of electrical activity with tetrodotoxin, a voltage-dependent sodium channel blocker, increases TβRI levels via a transcription-dependent mechanism. In contrast, the promotion of electrical activity in myotube cultures, induced by the up-regulation of voltage-dependent sodium channels or by direct stimulation with extracellular electrodes, causes TβRI levels to decrease. Similar results were obtained in denervated adult muscles, suggesting that electrical activity-dependent regulation of TβRI also occurs in vivo. Additional results suggest that this activity-dependent regulation is mediated by myogenin. Altogether, these findings support the possibility for a novel regulatory mechanism acting on TGF-β signaling cascade in skeletal muscle cells.

ptosis (1,2). TGF-␤ signaling cascades initiate after ligand binding with the heterodimerization of type II (T␤RII) and type I (T␤RI) TGF-␤ receptors at the cell surface. This induces the phosphorylation of R-Smad proteins, such as Smad2 and Smad3, the association with co-Smads, such as Smad4, and their translocation to the nucleus in order to promote the transcription of target genes (3). Several regulatory mechanisms for these TGF-␤ signaling pathways have been described, including modification of transducing receptor turnover (induced by ligand binding, cell membrane distribution, and I-Smads) and cytoplasm-nuclear transport and half-life of Smad proteins (4).
In skeletal muscle cells, TGF-␤ acts as a strong inhibitor of myogenesis. It is known that TGF-␤ can inhibit myoblast differentiation in vitro, affecting the expression of muscle proteins such as myosin heavy chain and creatine kinase (5). Cell sensitivity to TGF-␤-derived inhibition decreases during myoblast differentiation, and recent studies support a Smad3-mediated mechanism for this suppression (6). On the other hand, skeletal muscle cells develop excitability and contractile properties during the first stages of myogenesis. This phenotype has been associated to the onset of voltage-dependent channel protein expression (sodium and calcium channels) and to the maturation of the excitation-contraction apparatus (7)(8)(9). At more advanced developmental stages, muscle activity is controlled by motor innervation (10). Denervation has been the classical model to study the in vivo regulatory properties of electrical activity on muscle protein expression. Studies using this model have demonstrated that several myogenic factors, such as myogenin and MyoD, and ligand-activated and voltage-dependent channels, such as nicotinic acetylcholine receptor (nAChR) and sodium channels, are up-regulated after denervation (11,12). Similar results, whereby voltage-dependent sodium channels and nAChR are up-regulated, have been reported for primary myotube cultures after blocking spontaneous electrical activity with tetrodotoxin (TTX) (13,14).
The aim of the work presented here was to evaluate whether TGF-␤ signaling cascade could be modulated in skeletal muscle cells by an intrinsic property such as electrical activity. Our results demonstrate that such electrical activity modulation occurs in rat primary myotubes through the modification of TGF-␤ transducing receptor levels. T␤RI at the cell surface was up-regulated following inhibition of spontaneous electrical activity involving transcriptional activation and down-regulated when this activity was promoted, pointing to a novel mechanism for the control of TGF-␤ signaling pathways in skeletal muscle cells. In vivo experiments demonstrated that this electrical activity-dependent modulation also occurred in adult skeletal muscles after motor denervation, and additional analyses suggested a myogenin-dependent mechanism.

EXPERIMENTAL PROCEDURES
Animals and Muscle Denervation-2-month-old male Sprague-Dawley rats were anesthetized with xylazine/ketamine, and right legs were denervated by a sciatic section of 0.5 cm in the hip region of the hind limb as previously described (15). After 72 h, tibialis anterior muscle was excised for analysis. Muscles from contralateral legs were used as controls. For protein extracts, muscle tissue was frozen with liquid nitrogen, triturated, and homogenized in buffer containing 1% SDS, 1% Triton X-100, and protease inhibitors. Extracts were then centrifuged at 20,000 ϫ g for 10 min and supernatants collected and analyzed for protein quantification.
Cell Cultures-Rat skeletal myoblasts were isolated from the hind legs of day 19 Sprague-Dawley rat fetuses. Briefly, muscle mass was dissected from bones, subjected to mechanic dissociation, and filtered through a 70-m cell strainer. Cell suspensions were preplated to eliminate fibroblasts, and non-adherent cells were plated onto collagencoated culture dishes at a density of 100,000 cells/ml. Myoblasts were grown at 37°C with 8% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% bovine fetal serum and 10% horse serum. After 48 h, the medium was changed to Dulbecco's modified Eagle's medium 10% horse serum to induce myotube formation with 10 M cytosine-␤-D-arabinofuranoside to inhibit fibroblast proliferation. Under these conditions, primary myotubes showed spontaneous contractions as of day 2 in differentiation medium. To inhibit electrical activity, 1 g/ml of TTX (Alomone, Jerusalem, Israel) was added to cell cultures at day 1 of differentiation.
Electrical Stimulation-Primary cultures were stimulated as described by Chahine et al. (16). Stimulation electrodes were immersed in culture medium, and myotubes were stimulated using a Grass stimulator with 0.4-ms 4-V pulses (threshold required to induce visible contractions) in 100-Hz trains lasting 1 s, applied once every 100 s for 12 h.
DNA Transfections-Primary myoblast cultures were transfected with 2 g of p3TP-Lux or 2 g of ␦-47MEKLuc and 0.01 g of pRLSV40 using the calcium phosphate DNA precipitation method (17,18). After 6 h, cell cultures were subjected to a glycerol shock for 45 s and then induced to differentiate. Three days later, p3TPLux activity was determined by incubating cultures with 1 ng/ml of TGF-␤1 for 12 h and then harvesting for luciferase assay.
Cross-linking Assays-TGF-␤1 was radiolabeled with 125 I using chloramine T, and affinity labeling was carried out for 4 h at 4°C using disuccinimidyl suberate as cross-linking reagent (Pierce). Samples were analyzed by 8 -12% gradient SDS-PAGE as previously described (19).
Immunocytochemistry Analysis-Cells were fixed in 3% paraformaldehyde, permeabilized with 0.05% Triton X-100, and incubated with 1:100 monoclonal anti-Smad4 antibody (Santa Cruz Biotechnology). Bound antibodies were detected by incubating the cells with 1:100 affinity-purified fluorescein isothiocyanate-conjugated anti-mouse IgG (Pierce). After rinsing, the slides were viewed through a Nikon upright microscope equipped with epifluorescence.
Semi-quantitative RT-PCR Analysis-Total RNA was isolated from primary myotubes using TRIzol (Invitrogen). cDNA was synthesized from total RNA previously treated with DNase I, primed with random primers, and reversed transcribed with Moloney murine leukemia virus reverse transcriptase. PCR cycling conditions for rat T␤RI were 31 cycles consisting of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s, followed by a 7-min final extension at 72°C. For ␤-actin 25 cycles were used for the same cycling protocol. The PCR products were analyzed by agarose gel electrophoresis. The primers for T␤RI were 5-TGGTCTTGCCCATCT-TCACA-3 (sense) and 5-ATTGCATAGATGTCAGCACG-3 (antisense), which yielded a PCR product of 279 bp (20). The primers for rat ␤-actin were 5-TCTACAATGAGCTGCGTGTG-3 (sense) and 5-TACATG-GCTGGGGTGTTGAA-3 (antisense), which yielded a PCR product of 131 bp.

TGF-␤-dependent Signaling Is Heightened in Inactive Primary
Myotubes-To determine whether electrical activity can modulate TGF-␤ cascade components in skeletal muscle cells undergoing differentiation, we analyzed the effect of inhibiting electrical activity in rat primary myotubes using TTX. These cells develop spontaneous electrical activity, together with visible contractions, early during in vitro differentiation. To begin with, the TGF-␤-dependent intracellular cascade that triggers Smad-dependent pathways was examined by measuring the phosphorylation of Smad2 induced by TGF-␤ in control and TTXtreated primary myotubes. Fig. 1A shows that in TTX-treated cells the TGF-␤-dependent phosphorylation of Smad2 was detectable at lower ligand concentrations than in control myotubes. In the same figure, the dose-response curves for this TGF-␤-dependent phosphorylation are shown following band quantification.
Next, we determined whether the Smad-dependent downstream steps of the TGF-␤ signaling cascade were also affected by TTX treatment. For this, the nuclear translocation of Smad4 was analyzed by indirect immunolocalization. Fig. 1B shows that after TGF-␤ incubation the distribution of Smad4 in active primary myotubes remained cytosolic and excluded the nuclei, whereas in TTX-treated myotubes it concentrated in the nuclei, supporting the idea that in inactive myotubes the responsiveness to TGF-␤ is increased.
To evaluate whether the TGF-␤-induced transcription of target genes was also affected in TTX-treated myotubes, we transfected primary cultures with the TGF-␤-responsive reporter p3TP-Lux (21). Reporter activity increased ϳ4-fold in TTX-treated myotubes with respect to control cells (Fig. 1C), suggesting a more activated Smad-dependent TGF-␤ pathway in inactive myotubes than in control cells. Moreover, Fig. 1D shows that total Smad2 and Smad4 protein levels were not altered by TTX treatment, implying that the enhanced Smad2 phosphorylation and Smad4 nuclear translocation observed in inactive myotubes in response to TGF-␤ must be associated to an upstream variation in the signaling cascade.
Total and Cell Surface TGF-␤ Type I Receptor Levels Are Increased in TTX-treated Myotubes-To better understand the increased TGF-␤ responsiveness of inactive myotubes, we decided to measure protein levels of T␤RI using Western blot analysis. A rise in T␤RI protein levels was seen during primary myoblast differentiation ( Fig. 2A), with a more notable increase observed upon inhibition of electrical activity at 48 or 72 h (cultures at day 4 or 5 of differentiation, respectively). To determine whether this augmentation in total T␤RI levels corresponded to an increase in the number of cell surface receptors, we analyzed the receptor binding of radiolabeled 125 I-TGF-␤1 through affinity labeling experiments. Fig. 2B shows that labeling associated to TGF-␤/T␤RI complex was higher in TTX-treated myotubes than in control cells (153% Ϯ 4.1 S.E., n ϭ 3), suggesting that the rise in T␤RI protein levels obtained after blocking electrical activity also corresponds to an increase in membrane receptors at the cell surface. Altogether, these results indicate that the heightened responsiveness to TGF-␤ of inactive myotubes as evaluated by determining the dose-response curves of Smad2 phosphorylation, Smad4 nuclear translocation, and p3TPLux reporter activity is consist-ent with the up-regulation of the transducing receptor T␤RI as a result of TTX exposure.
TTX-dependent Increase in TGF-␤ Type I Receptor Levels Involves Transcriptional Activation-To determine whether the electrical activity-dependent up-regulation of T␤RI occurs at a transductional and/or transcriptional level, at first we performed semi-quantitative RT-PCR analysis. Fig. 3A shows that PCR product corresponding to T␤RI transcripts is increased in TTX-treated myotubes compared with control cultures (lanes 1 and 4). To estimate half-life values for T␤RI transcripts in control and inactive cultures we analyzed the time-dependent decrease in PCR product after incubation with the RNA polymerase II inhibitor DRB (5,6-dichloro-1-␤-D-ribofuranosylben-FIGURE 1. TGF-␤ signaling increases in TTX-treated primary myotubes. A, Western blot of total cell extracts showing immunodetection of phospho-Smad2 (P-Smad2) in untreated (control) and TTX-treated myotubes at day 3 of differentiation after incubation with TGF-␤1 for 15 min at the indicated concentrations (ng/ml). The plot shows the dose-response curves obtained after band quantification using GADPH. B, indirect immunofluorescence images showing the subcellular localization of Smad4 in control and TTX-treated cultures at day 3 of differentiation after incubation with TGF-␤1 for 1 h. Scale bar, 50 m. C, p3TP-Lux reporter activity in control and TTX-treated myotubes. Primary myoblasts were transiently co-transfected with p3TP-Lux and pRLSV40 to normalize transfection. At day 3 of differentiation, myotubes were incubated with 1 ng/ml of TGF-␤1 for 12 h and harvested at day 4 of differentiation (d4), after which dual luciferase activity was measured. Data are expressed as the means Ϯ S.E. of three measurements from a representative experiment repeated twice. D, Western blot for total Smad2 and Smad4 obtained from total cell extracts of control and TTX-treated cultures. JULY 7, 2006 • VOLUME 281 • NUMBER 27 zimidazole). Fig. 3A, lower plots, show that the half-life values extracted from decay of PCR product are almost identical in control and TTXtreated cultures (9.5 and 9.4 h, respectively). These results suggest that increased T␤RI RNA levels in TTX-treated cultures are due to an increased transcription rate and not to an increased stability (half-life) of transcripts.

Electrical Activity-dependent Regulation of TGF-␤ Signaling
To confirm that TTX-dependent increase in T␤RI receptor levels requires transcription we studied the effect of DRB on receptor protein levels. Fig. 3B shows that TTX-induced T␤RI up-regulation could be blocked by DRB, indicating that this activity-dependent up-regulation required transcription as a prior step. Furthermore, the induction of T␤RI by chronic exposure to TTX (for 72 h) could also be progressively inhibited by DRB as seen in Fig. 3C. These results demonstrate that the onset and maintenance of T␤RI up-regulation in paralyzed cultures is dependent on transcription.
On the other hand, to evaluate whether the increase in T␤RI observed at the plasma membrane in TTX-treated myotubes could result from a lower receptor degradation rate, we analyzed the recovery of receptor levels using chloroquine (a lysosomal degradation inhibitor) in cycloheximide-treated cultures. Fig. 4 shows that the inhibition of the lysosomal pathway had a lesser effect on receptor recovery in TTX-treated cultures, pointing to the attenuation of this pathway in inactive myotubes. Similarly, a reduced recovery of T␤RI was also noted in TTXtreated cells after inhibition of the proteasome protein degradation pathway, using MG132 (lanes 4 and 8). Altogether, these results support the notion that the up-regulation of T␤RI in inactive myotubes involves both a transcriptional activation and a reduction in the lysosome-and proteasome-dependent degradation of the receptor.
Myotube Hyperexcitability and Direct Stimulation Decrease TGF-␤ Type I Receptor Protein Levels-To test whether enhanced electrical activity in primary myotubes could negatively regulate T␤RI protein levels, we analyzed the influence of cell excitability under two experimental conditions (Figs. 5 and 6). The action of TTX over voltage-dependent sodium channels is known to be reversible, and its inhibitory effect can be restored by rinsing the toxin (22). When rat primary myotubes chronically treated with TTX were washed to restore their elec-trical activity, this activity together with visible contractions was higher than in control cultures. This hyperexcitable phenotype is consistent with a TTX-dependent up-regulation of voltage-sensitive sodium channels (see below). Therefore, T␤RI protein levels were determined under these conditions in which primary myotubes were allowed to recuperate their electrical activity. Fig. 5A shows that T␤RI protein levels were significantly lower in myotubes treated with TTX, washed, and lysed 12 h later (w12) than in untreated (C, control) or unwashed (TTX) myotubes. The lower panel shows that myogenin protein levels followed the same pattern of electrical activity-dependent regulation. It is remarkable that after such a short period of time as 12 h the up-regulation of T␤RI could be reverted and levels seen to decrease below those of controls. These findings indicate that receptor expression can be modulated in both directions (up-and down-regulated) by modifying the  electrical activity of the myotubes. As Fig. 5B shows, the change in receptor protein levels in w12 myotubes also correlated with a reduced binding of radiolabeled TGF-␤1, suggesting that receptors at the cell surface were also diminished. This activity-dependent decrease in TGF-␤1 receptor binding was greatest after 24 h. Considering that in Western blot experiments electrical activity recovery was evident after 12 h, a temporal delay between transductional responses and receptor translocation to the cell membrane can be inferred. Fig. 5B also shows that the re-inhibition of electrical activity 12 h after rinsing (lane reTTX) resulted in higher T␤RI ligand binding compared with cultures washed withoutthereadditionofTTX(w24),demonstratingthatelectricalactivitydependent down-regulation of T␤RI is reversible. Furthermore, T␤RI down-regulation caused an effect on the Smad-dependent intracellular signaling cascade, as seen in Fig. 5C. TGF-␤-dependent p3TPLux reporter activity in blocked and washed cultures (w12) was below that measured in untreated myotubes. All these results indicate that recovered electrical activity in w12 myotubes down-regulates T␤RI protein levels, decreasing the number of functional transducer receptors at the cell surface and lowering the responsiveness of primary myotubes to TGF-␤.
To confirm that lower T␤RI levels in w12 myotubes correlated with a rise in electrical activity in these cultures, we transfected primary myotubes with the ␦-47MEKLuc reporter plasmid in which a minimal promoter sequence of encephalin (MEK) is under the control of an enhancer sequence (47 bp) from the nAChR ␦ subunit. This enhancer sequence confers electrical activity-dependent regulatory properties to a heterologous promoter that result in the down-regulation of gene expression (23). As shown in Fig. 5D, w12 myotubes presented a lower reporter activity than control cells, denoting greater electrical activity in recovered myotubes than in untreated cultures. The increased reporter activity in TTX-treated cells is in concordance with the inhibition of spontaneous electrical activity as previously described for denervated muscles and inactive primary cultures (23,24). Voltage-dependent sodium channels are required to generate action potentials, giving rise in cultured skeletal muscle cells to a linear relationship between ion channel levels and the frequency of action potentials (7). In addition, the up-regulation of sodium channels can be induced by the denervation of adult muscle as well as after blockade of electrical activity in muscle cell cultures by TTX (12,13). Fig. 5E shows a progressive increase in sodium channel levels during the differentiation of primary myotubes, with a clear up-regulation in TTX-treated myotubes. Moreover, Fig. 5F shows that 12 h after rinsing off TTX (w12) sodium channel protein levels remained higher than in control myotubes. This result is consistent with a half-life Ͼ24 h for muscle sodium channels, estimated from experiments using DRB (data not shown). Together, these results show that w12 myotubes washed after chronic exposure to TTX exhibit enhanced electrical activity with regard to untreated cultures as measured through a reduction in ␦-47MEKLuc reporter activity. This hyperexcitable phenotype, compatible with a higher sodium channel content, leads to a reduction in T␤RI protein levels, suggesting that electrical activity negatively regulates this receptor in skeletal muscle cells.
The second approach used to study the electrical activity-dependent regulation of T␤RI was to investigate whether direct electrical stimulation of primary myotubes could also negatively modulate receptor levels. Electrical activity was induced in primary myotubes by stimulating with extracellular electrodes. This method has been successfully employed in studying the electrical activity-dependent expression of several muscle proteins, such nAChR and myogenin (25,26). Fig. 6A shows that electrical stimulation for 12 h (STIM12) significantly decreased T␤RI protein levels in primary myotubes, as expected. Fig. 6B shows that the electric stimulation protocol was also effective in downregulating myogenin expression as described before (25,26).

Electrical Activity-dependent Regulation of TGF-␤ Type I Receptor Occurs in Vivo in Adult Skeletal
Muscle-The next question was to determine whether the activity-dependent regulation of T␤RI observed in primary cultures also occurred in adult skeletal muscle in vivo. For this, we measured receptor protein levels in rat hind limb muscles after short-term motor denervation. Fig. 7 shows that 72 h after denervation, receptor levels increased in tibialis anterior compared with innervated contralateral muscles. Additionally, the middle panel shows that denervated muscles exhibited increased myogenin levels with respect to controls, as has been extensively documented (11). This result demonstrates that in adult skeletal muscles in vivo, T␤RI protein levels are also modulated by electrical activity, supporting the notion that this control mechanism is maintained after differentiation of skeletal muscle cells.
Myogenin Is Required for Electrical Activity-dependent Regulation of TGF-␤ Type I Receptor-Considering that all changes in T␤RI induced upon modification of the electrical activity of skeletal muscle cells were accompanied by equivalent changes in myogenin levels, as reproduced in Figs. 5, 6, and 7, and that the promoter sequences of mammalian TGF-␤ receptors contain E-boxes (binding sites for basic helix-loophelix (bHLH) transcription factors such as MyoD and myogenin, which are regulated by electrical activity) (11,14), we proceeded to evaluate the participation of myogenin in the electrical activity-dependent regulation of T␤RI. For that purpose, we first analyzed the time courses and coincidence of myogenin and T␤RI up-regulation after applying TTX to inhibit spontaneous electrical activity. Fig. 8A shows that TTX-dependent up-regulation of myogenin preceded the elevation observed in T␤RI protein levels, suggesting that this bHLH myogenic factor possibly acts upstream in the intracellular cascade for the electrical activity-dependent regulation of T␤RI. To examine the effects of myogenin on T␤RI levels in primary myotubes, we next performed transient transfection experiments using an expression plasmid containing full-length myoge-  JULY 7, 2006 • VOLUME 281 • NUMBER 27 FIGURE 5. Heightened electrical activity decreases T␤RI protein levels and suppresses TGF-␤-dependent signaling. A, Western blot for T␤RI in protein extracts taken from control untreated myotubes (C ) at day 5 of differentiation, as well as TTX-treated cells (TTX) and cells treated with TTX, washed and lysed 12 h later (w12). GADPH immunostaining is shown as a loading control. The lower panel shows Western blotting for myogenin, using total extracts from myotube cells submitted to the same treatments as above. B, affinity labeling of myotubes at day 4 of differentiation using 100 pM 125 I-TGF-␤1. Phosphorimage shows total myotube cell extracts separated by SDS-PAGE. The migration position of the T␤RI/TGF-␤ complex is indicated. Treatments applied to the cells are the following: Lane c, control myotubes (d4); lane TTX, TTX-treated myotubes; lane w12, TTX-treated myotubes, washed and lysed 12 h later; lane w24, TTX-treated myotubes, washed and lysed 24 h later; lane reTTX, TTX-treated myotubes, washed, reincubated with TTX 12 h later, and lysed after 12 h. C, p3TP-Lux reporter activity in control, TTX-treated, and w12 myotubes. Primary myoblasts were transiently co-transfected with p3TP-Lux and pRLSV40 to normalize transfection. At day 3 of differentiation, myotubes were incubated with 1 ng/ml of TGF-␤1 for 12 h and then harvested at day 4 of differentiation (d4), after which dual luciferase activity was measured. Data are expressed as the means Ϯ S.E. of three measurements from a representative experiment repeated twice. D, ␦-47MEKLuc reporter activity in control, TTX-treated, and w12 myotubes at day 4 of differentiation. Primary myoblasts were transiently co-transfected with ␦-47MEKLuc and pRLSV40 to normalize transfection. Data are expressed as the means Ϯ S.E. of three measurements from a representative experiment repeated twice. E, Western blot for the ␣ subunit of voltage-dependent sodium channels (anti-Pan-Na v ) using total cell extracts of primary myotubes undergoing differentiation (day 2 to day 5; d2-d5). The last two lanes correspond to extracts from TTX-treated cultures. GADPH immunostaining is shown as a loading control. F, Western blot for the ␣ subunit of voltage-dependent sodium channels using total cell extracts from control, TTX-treated, and w12 myotubes at day 4 of differentiation. GADPH immunostaining is shown as a loading control. nin cDNA (pEMSV-myogenin). Fig. 8B shows that the overexpression of myogenin (in both control and inactive cells) correlated with an increase in T␤RI protein levels. This result implies that myogenin is a positive modulator of T␤RI expression in skeletal muscle cells. In contrast, inhibiting the transcriptional activity of myogenin through incubation of primary cultures with sodium butyrate, an inhibitor of bHLH factors such as MyoD and myogenin (27), prevented the rise in T␤RI levels induced by TTX treatment (Fig. 8C). This in turn suggests that transcriptional modulation by myogenin is required for the TTX-induced up-regulation of T␤RI. Altogether, these results (temporal correlation, overexpression, and inhibitory effects) strongly suggest that bHLH factors such as myogenin are implicated in the electrical activitydependent regulation of T␤RI in skeletal muscle cells.

DISCUSSION
Results presented here demonstrate that the TGF-␤ signaling cascade is susceptible to myotube excitability when undergoing differentiation. The enhanced responsiveness to TGF-␤ exhibited by inactive myotubes was associated to the up-regulation of the T␤RI. In contrast, this transducing receptor was down-regulated when myotube electrical activity was promoted. This excitability-dependent regulation of T␤RI also occurred in vivo in adult skeletal muscle after short-term motor denervation. Taken together, these findings point to a novel regulatory mechanism for TGF-␤ signaling in skeletal muscle cells.
The effects of electrical activity on T␤RI levels in primary cultures (namely, TTX-dependent up-regulation and stimulation-dependent down-regulation) were significant even after time periods as short as 12 h (see Figs. 3B, 5A, and 6), suggesting a fast turnover of TGF-␤ receptors in cultured skeletal muscle cells. These results are consistent with a protein half-life of ϳ2 h, as estimated from Western blot and radiolabeling analyses in primary cultures treated with cycloheximide (data not shown), and also with short half-lives for T␤RI detected in other cell types. For example, in pulse-chase experiments using CCL-64 lung epithelial cells, the half-life of T␤RI was estimated to be ϳ12 h (28), whereas in osteoblasts, using transcription and protein synthesis inhibitors, it was estimated at 2 h for the protein and 7 h for the transcripts (29).
Semi-quantitative RT-PCR analysis showed that T␤RI transcripts are increased in inactive myotubes without modification of RNA half-life, suggesting that TTX-induced up-regulation of T␤RI involves transcriptional activation (Fig. 3A) and not increased transcript stability. Western blot analysis of the effect of DRB over TTX-induced up-regulation of T␤RI protein levels confirmed that transcription is required for this regulation (Fig. 3, B and C). These results could be the consequence of a direct effect on the T␤RI gene or on genes encoding transcription factors that act upstream in the cellular response to muscle inactivity. It is well known that bHLH factors, such as myogenin, are required for the up-regulation of nAChR subunits upon blocking of electrical activity in primary cultures or denervation (23,24). Recent studies have shown that after denervation, myogenin up-regulation is associated with enhanced MEF2 transcriptional activity induced by the down-regulation of a histone deacetylase (30). Although the promoter sequences of T␤RI and T␤RII do not contain MEF2 binding sites, reports studying fibroblast myogenic conversion have provided evidence for a cooperative and synergic activation between MEF2 and myogenic bHLH factors in which MEF2 binding sites are not required (31).
Our results also indicate that up-regulation of T␤RI in inactive myotubes involves, in addition to transcriptional activation, a reduction in the lysosome-and proteasome-dependent degradation rates of the receptor (Fig. 4). An unequal contribution of these two pathways to the degradation of TGF-␤ receptors has been observed in epithelial cells (32). On the other hand, there may also be an inactivity-induced redistribution of transducer receptors into different endocytic compartments at cell membrane, thereby decreasing T␤RI turnover (33).
The effect on T␤RI levels of increasing the electrical activity of primary myotubes was examined under two experimental conditions: after up-regulation of sodium channels induced by chronic inhibition of electrical activity with TTX and after direct electrical stimulation. In the first approach, the observed decrease in ␦47-MEKLuc reporter activity pointed to a hyperexcitable phenotype associated with the up-regulation of sodium channels. Nevertheless, the possibility that chronic blockage of electrical activity can modify the gene expression of other ion channels, resulting in an equivalent phenotype, cannot be ruled out. A hyperexcitable state in denervated muscle associated to the up-regulation of SK3, a small conductance calcium-activated potassium channel, has indeed been described (34).

Electrical Activity-dependent Regulation of TGF-␤ Signaling
Moreover, another study demonstrated that the muscle chloride channel ClC-1 was down-regulated after denervation, leading to an increase in membrane resistance and excitability (35).
It is well known that TGF-␤ negatively regulates the early stage differentiation of skeletal muscle cells (5). Considering that T␤RI is downregulated by electrical activity in primary myotubes, it can be inferred that during muscle differentiation spontaneous electrical activity atten-uates TGF-␤-derived inhibition, allowing myogenesis to continue. In addition, the absence of this regulatory phenomenon in undifferentiated myoblasts when TGF-␤ acts as an activator of proliferation (36) can acquire relevance in long-term muscle denervation or regeneration processes after damage, when the proliferation of myoblasts in association with enhanced TGF-␤ signaling is required for the formation of new fibers. On the other hand, results obtained here using adult denervated FIGURE 8. Myogenin is required for activity-dependent regulation of T␤RI. A, Western blot for myogenin and T␤RI using the same protein extracts taken from myotubes incubated with TTX for different time periods between 0 and 9 h. Plot shows band quantification for myogenin and T␤RI expression. B, Western blot for myogenin and T␤RI using the same protein extracts obtained at day 3 of differentiation (d3) from control and TTX-treated myotubes as well as untransfected cells and myotubes transfected with the pEMSVmyogenin expression plasmid. Right panels show band quantification. C, Western blot for myogenin and T␤RI using the same protein extracts taken from control myotubes and myotubes incubated for 24 h with TTX in the absence or presence of 5 mM sodium butyrate. muscles could help to explain previous reports in which muscle inactivity is related to an increase in extracellular matrix deposition (15,37). Therefore, up-regulation of T␤RI in paralyzed fibers could increase cell sensitivity to profibrotic factors such as TGF-␤, elevating extracellular matrix synthesis and accumulation and possibly leading to muscle fibrosis.
Results concerning the timing of T␤RI and myogenin up-regulation after treatment with TTX and the overexpression and inhibition of myogenin activity using butyrate suggest the involvement of this bHLH factor in the electrical activity-dependent regulation of T␤RI in primary myotubes. Myogenin activates target genes by binding to E-boxes (bHLH myogenic factor binding sequences) and by transactivation of non-myogenic factors such as Sp1 (38,39). Three E-boxes are present in the rat T␤RII promoter sequence and two in the rat T␤RI promoter (around Ϫ1800 bp, not previously described) (40 -42). Interestingly, it has been demonstrated that during differentiation of myoblasts to myotubes and following denervation the binding of Sp1 factor to the promoter region of nAChR subunits is increased (43,23). In fact, using site-directed mutagenesis of reporter constructs carrying an enhancer of the rat nAChR ␦-subunit, both Sp1 and E-boxes were demonstrated to be necessary for conferring susceptibility to electrical activity to a heterologous promoter (23). Given that the rat T␤RI promoter sequence contains seven consensus Sp1 binding sites, which is also true for T␤RII and receptor type III (T␤RIII or betaglycan) (40,41,44), and two E-boxes, our results showing that myogenin can mediate the electrical activity-dependent regulation of T␤RI indicate that direct binding to E-boxes and/or a cooperative activation process with other transcription factors, such as Sp1, could underlie the observed modulation of T␤RI by myogenin. Furthermore, the mouse T␤RIII promoter has also been reported to contain two E-boxes, localized at Ϫ1500 bp, whose transcriptional activity is positively regulated by MyoD (19). Studies employing promoter deletions will be necessary to fully determine the functional relevance of these sites (around Ϫ1800 bp) in the gene regulation of T␤RI in skeletal muscle cells.
In summary, the present study has shown that T␤RI protein levels can be modulated by electrical activity in skeletal muscle cells undergoing differentiation and also in fully differentiated muscle fibers. Our data show that activation state-dependent changes in the receptor protein content of myotubes cause modifications in the downstream steps of the TGF-␤ transduction cascade, demonstrating that affecting total and membrane T␤RI levels has a direct impact on TGF-␤ signaling. These results provide novel evidence that cell excitability acts as a regulatory factor in the TGF-␤ signaling pathway of skeletal muscle cells. In the cellular context of skeletal muscle this regulatory mechanism can acquire physiological relevance in muscle damage events and the regeneration process. TGF-␤ is released between other inflammatory cytokines, and activitydependent regulation of a TGF-␤-transducing receptor can affect the cellular responsiveness of myoblasts undergoing differentiation.