Abstract

Loss of muscle mass occurs in a variety of diseases, including cancer, chronic heart failure, aquired immunodeficiency syndrome, diabetes, and renal failure, often aggravating pathological progression. Preventing muscle wasting by promoting muscle growth has been proposed as a possible therapeutic approach. Myostatin is an important negative modulator of muscle growth during myogenesis, and myostatin inhibitors are attractive drug targets. However, the role of the myostatin pathway in adulthood and the transcription factors involved in the signaling are unclear. Moreover, recent results confirm that other transforming growth factor-β (TGF-β) members control muscle mass. Using genetic tools, we perturbed this pathway in adult myofibers, in vivo, to characterize the downstream targets and their ability to control muscle mass. Smad2 and Smad3 are the transcription factors downstream of myostatin/TGF-β and induce an atrophy program that is muscle RING-finger protein 1 (MuRF1) independent. Furthermore, Smad2/3 inhibition promotes muscle hypertrophy independent of satellite cells but partially dependent of mammalian target of rapamycin (mTOR) signaling. Thus myostatin and Akt pathways cross-talk at different levels. These findings point to myostatin inhibitors as good drugs to promote muscle growth during rehabilitation, especially when they are combined with IGF-1-Akt activators.

  • muscle atrophy
  • hypertrophy
  • Akt
  • myostatin

the growth of skeletal muscle mass, like the mass of any other tissue, depends on protein and cellular turnover (38). Cellular turnover plays a major role during muscle development in both embryonic and postnatal growth. On the other hand, in adult muscle the physiological conditions promoting muscle growth do so mainly by increasing protein synthesis and decreasing protein degradation (35). Several genetic and biochemical studies have shown that the IGF-1/phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) pathway controls protein synthesis and muscle hypertrophy in adults. Muscle-specific overexpression of either IGF-1 or Akt in transgenic mice results in muscle hypertrophy and, importantly, in a physiological increase of muscle strength (11, 13, 27, 28). The downstream target of Akt is mTOR, because its inhibition completely blunts the Akt effects on muscle growth (35). Finally, S6 kinase 1 (S6K1) appears to be a further mediator of the Akt/mTOR pathway, because muscle fibers are smaller in S6K1-null mice and their hypertrophic response to IGF-1 and to activated Akt is blocked (29). Importantly, while promoting muscle growth, this pathway also inhibits the proteolytic systems and prevents muscle loss during diseases (3, 30, 34, 37, 41).

Muscle wasting is a prolonged event that occurs in many conditions, including disuse, fasting, cancer, diabetes, aquired immunodeficiency syndrome, and cardiac and renal failure, and involves a common program of changes in gene expression that affect around 100 genes termed atrophy-related genes or atrogenes (15, 33). Furthermore, two genes showing the highest level of induction are two novel muscle-specific ubiquitin ligases, atrogin-1/muscle atrophy F box (MAFbx) and muscle RING-finger protein 1 (MuRF1), that are responsible for the increased protein degradation through the ubiquitin-proteasome system (2, 8). Up to now, these two genes are actually the best markers for muscle atrophy and could be considered as master genes for muscle wasting. We and others have recently shown that these ubiquitin ligases, and protein breakdown in general, are blocked by the growth-promoting IGF-1/PI3K/Akt pathway (20, 37, 41). Members of the FoxO family, downstream targets of Akt, were identified as the main transcription factors regulating not only atrogin-1 expression but also the atrophy program that leads to loss of muscle mass (37).

Myostatin, a member of the transforming growth family-β (TGF-β) family, is expressed and secreted predominantly by skeletal muscle and mediates pleiotropic effects on different cell systems including adipose and connective tissues (17, 42). In skeletal muscle it modulates satellite cell activation and differentiation and it functions as a negative regulator of muscle growth (17). Mutations of the myostatin gene lead to a hypertrophic phenotype in mice, sheep, dogs, and cattle, and a loss-of-function mutation in the human myostatin gene was also found to induce increased muscle mass (5, 18, 24, 26, 39). The increase in muscle mass is a consequence of hyperplasia, which is an increase in cell number, and hypertrophy, which is an increase in cell size. Very few studies explore the effect of myostatin inhibition in adults and, consistent with the genetic studies, they confirm the growth-promoting effect when the pathway is blocked (43, 45). However, whether muscle hypertrophy is caused by activation of satellite cells or by increased protein synthesis in adult fiber has not been addressed. Thus, it is not clear whether myostatin can signal into adult fibers. Myostatin binds to the activin receptor IIB (ActRIIB) (19), a type II TGF-β receptor, which, in turn, activates activin receptor-like kinase 4 (ALK4) or ALK5, both type I serine/threonine kinase receptors (31). Muscle hypertrophy is induced in transgenic mice expressing a truncated and inactive ActRIIB (18). However, the downstream targets of the myostatin pathway and their role in protein synthesis as well as protein degradation are still to be determined. Gain of function studies suggest a contribution of this pathway to muscle loss, but results are conflicting. When Chinese hamster ovary cells engineered to express myostatin were injected into skeletal muscles, they caused severe muscle loss that culminated with the death of some animals (47). However, transgenic mice, which express myostatin specifically in skeletal muscles, show only mild atrophy in males and no phenotype in females (32). Electroporation experiments show that myostatin expression in adult muscle induces a degree of atrophy comparable to that observed in transgenic mice (7). However, whether this atrophy requires activation of atrophy-related genes and involves increased protein degradation is uncertain. Furthermore, it is still unclear whether Smads, the transcription factors downstream of TGF-β signaling, are mediating some of the myostatin effects in adult muscles.

The recent failure of myostatin inhibitor during PhaseI/II Trial for muscular dystrophy (42) suggests that other TGF-β members can regulate muscle mass. Overexpression of follistatin, an inhibitor of myostatin and other TGF-β members including growth differentiation factor 11 (GDF-11) (18), promotes a great increase in muscle size that is more robust than that obtained by the deletion of the myostatin gene (18). Indeed, mice resulting from follistatin transgenic and myostatin-knockout mice show a tremendous increase in muscle mass (16). Altogether, these results confirm that, to obtain an efficient inhibition of this pathway, we should either identify other myostatin-like molecules or move downstream of myostatin/TGF-β ligands. In this study we investigated the contribution of TGF-β downstream targets to the control of adult muscle mass. To dissect the pathway, we transfected adult fibers with different mutants of type I and type II receptors to perturb the TGF-β signaling exclusively in terminally differentiated muscle cells. TGF-β inhibition has a positive effect on muscle growth independent from mTOR and satellite cell activation. Conversely, TGF-β activation induces muscle atrophy that is independent from MuRF1 upregulation. Then, we defined that Smad2 and Smad3 are the transcription factors downstream of myostatin signaling and that Akt is dominant over TGF-β pathway.

MATERIALS AND METHODS

Animals and in vivo transfection experiments.

Animals were handled by specialized personnel under the control of inspectors of the Veterinary Service of the Local Sanitary Service (AUSL 16 - Padova), the local officers of the Ministry of Health. All procedures are specified in the projects approved by the Italian Ministero Salute, Ufficio VI (authorization numbers C36 and C42). All experiments were performed on adult 2-mo-old male CD1 mice (28–32 g). The inducible Akt transgenic mice are described elsewhere (21). In vivo transfection experiments were performed by intramuscular injection of plasmid DNA in tibialis anterior muscle followed by electroporation as described previously (36, 37). Cross-sectional area of transfected fibers was measured as described previously (37) and compared with the area of surrounding untransfected myofibers (control). Denervation was performed by cutting the sciatic nerve of left limb while right limb was used as control. Muscles were collected 14 days after denervation for morphological analysis. Bromodeoxyuridine (BrdU) was injected every day starting from day 5 after transfection until the animal was killed. Every injection contained 50 mg/kg BrdU dissolved in 100 μl of 0.9% NaCl. Positive controls for BrdU incorporation are described in the supplemental data (data supplement is available online at the American Journal of Physiology-Cell Physiology website).

Plasmids and antibodies.

See supplemental data.

Cell culture and transient transfections.

Murine embryonic fibroblast (MEF) cells and C2C12 myogenic cell line were cultured in DMEM (GIBCO-Invitrogen) supplemented with 10% fetal calf serum and 1% penicillin-streptomycin mixture at 37°C and 5% CO2 until cells reached confluence. MEF and myoblasts were transfected using Lipofectamine2000 (Invitrogen) according to the manufacturer's instructions.

Immunohistochemistry.

Mouse muscle fibers expressing hemagglutinin (HA)-tagged, Flag-tagged, or myc-tagged proteins were stained in cryosections fixed with 4% paraformaldehyde. Immunohistochemistry with anti-HA polyclonal antibody (Santa Cruz), anti-Flag polyclonal antibody (Sigma), and anti-myc polyclonal antibody (Santa Cruz) was as previously described (37). Cryosections of transfected muscles were examined using a fluorescence microscope as previously described (37). To identify BrdU-positive nuclei, 4–6 μm cryostat sections were incubated with 4 M HCl and trypsin to expose the DNA to the antibody.

Immunoblotting.

Cells were lysed and immunoblotted as previously described (21). Blots were stripped using Restore Western blotting stripping buffer (Pierce) and reprobed if necessary. The list of antibodies is provided in supplemental data.

DNA microarrays.

RNA was prepared from mouse gastrocnemius muscles using the TRIzol method followed by Qiagen RNeasy cleanup. Transgenic mice were induced to express Akt in muscle by tamoxifen treatment for 24 or 48 h. As a control, transgenic mice were treated with vehicle for 24 h. RNA from three control mice or two transgenic mice was pooled, labeled, and hybridized to Affymetrix Mouse Genome 430 2.0 Arrays according to standard Affymetrix protocols. Expression values were summarized using the Mas 5.0 algorithm. Genes that were up- or downregulated upon Akt expression in transgenic mice compared with nontransgenic mice were determined using Excel software. Data are available by using GEO accession number GSE15397 (available at http://www.ncbi.nlm.nih.gov/geo).

Gene expression analyses.

Total RNA was prepared from skeletal muscles using Promega SV Total RNA Isolation kit. Complementary DNA generated with Invitrogen SuperScript III Reverse Transcriptase was analyzed by quantitative real-time RT-PCR using Qiagen QuantiTect SYBR Green PCR Kit. All data were normalized to GAPDH expression. The oligonucleotide primers used are described in supplemental Table S1.

RNA interference in vivo.

In vivo RNA interference (RNAi) experiments were performed as previously described (37) using at least four different sequences for each gene. The sequences are shown in supplemental Table S2 and supplemental material. For the validation of small hairpin RNA (shRNA) constructs, MEF cells were maintained in DMEM-10% FBS and transfected with shRNA constructs using Lipofectamine 2000 (Invitrogen). Cells were lysed 48 h later, and immunoblotting was performed as described above.

Statistical analysis.

Data were analyzed by two-tailed Student's t-test. For all graphs, data are represented as means ± SE.

RESULTS

Modulation of TGF-β pathway in adult myofibers.

Myostatin is a hormone that is secreted into the blood and mediates pleiotropic effects on different cell systems. In muscle it modulates satellite cell activation and differentiation. It is not clear whether myostatin can signal into adult fibers. Moreover, other myostatin-like molecules, belonging to TGF-β family, affect muscle mass (16, 42). We used our established transfection technique (6, 21, 36, 37) to express mutants of type I and type II myostatin/TGF-β receptors only in terminally differentiated muscle fibers and not in satellite or interstitial cells (supplemental Fig. S1). This approach allows a high transfection efficiency without interfering with physiological homeostasis of adult fibers (6, 36, 37). Moreover, transfection of mock control plasmids does not affect myofiber size (6, 21, 36, 37). In vivo and in vitro studies have shown that myostatin binds to ActRIIB, which in turn signals through ALK4 and ALK5 to Smads. First, we tested the capability of our receptor mutants in modulating the pathway. A dominant-negative (d.n.) ActRIIB and constitutively active (c.a.) ALK4 and ALK5 were transfected in muscle cells, and the levels of phosphorylation of the downstream target, Smad3, were tested (Fig. 1A). Furthermore, transfection of c.a.ALK5, in vivo, induced myc-Smad3 or Flag-Smad2 nuclear relocalization (Fig. 1B). Next, we tested whether endogenous Smad2/3 can be regulated by the type I and II mutant receptors. The best way to monitor Smad transactivation activity is by in vivo transfection of Smad-sensors. Both c.a.ALK4 and c.a.ALK5 greatly increased luciferase activity of Smad-responsive reporter while d.n.ActRIIB reduced reporter activity (Fig. 1C). To further confirm the inhibitory effect of type II mutant receptor, we cotransfected Smad reporter with myc-Smad3 in the presence or absence of d.n.ActRIIB. Inactive type II receptor completely antagonized myc-Smad3 activity on Smad sensor (Fig. 1D). This first set of experiments confirms that we can efficiently modulate TGF-β signaling in adult myofibers.

Fig. 1.

Modulation of transforming growth factor-β (TGF-β) pathway by overexpressing constitutively active or dominant negative mutants of type I and type II myostatin receptors in adult muscles. A: Smad3 phosphorylation revealed by immunoblotting. Myogenic C2C12 cells were transfected with vectors expressing dominant negative (d.n.) activin receptor IIB (ActRIIB) and constitutively active (c.a.) activin receptor-like kinase 4 (ALK4) and ALK5. HA, hemagglutinin. B: activation of myostatin pathway induces Smad nuclear translocation. Adult muscle fibers were transfected with myc-tagged Smad3 and Flag-tagged Smad2 in the presence or absence of c.a.ALK5. Seven days later, muscles were removed and analyzed for Smad localization by immunostaining. While in control muscles Smad immunoreaction shows a cytosol staining, the fibers that coexpress activated type I receptors depict a nuclear localization of Smad. Arrows show Smad2/3 nuclear relocalization in myofibers overexpressing c.a. ALK5. Ab, antibody. C: transcriptional activity of endogenous Smads after activation or inhibition of the myostatin/TGF-β pathway in adult muscles. A Smad-dependent reporter, in which the Smad-binding site (CAGA) is repeated 12 times, was transfected into adult tibialis anterior muscle in the presence or absence of mutants for type I and II receptors. A Renilla luciferase vector was cotransfected to normalize for transfection efficiency. Seven days later, firefly/Renilla luciferase activity was determined. d.n.ActRIIB caused a 20% decrease of Smad-sensor activity. Shown in the bottom graph is the inhibitory effect of d.n.ActRIIB. *P < 0.001, #P < 0.01; n = 5. D: transcriptional activity of myc-Smad3 after inhibition of myostatin/TGF-β pathway. A Smad-dependent reporter was transfected into adult tibialis anterior muscle in the presence or absence of myc-Smad3 and d.n.ActRIIB. Mutant of type II myostatin receptor completely abolished Smad3 activity on CAGA reporter; n = 4.

RNAi-mediated approach in adult fibers efficiently blocks Smad2 and 3.

To determine whether Smad2 and Smad3 are downstream and mediate the effects of TGF-β in vivo, we used an RNAi approach to selectively knock down Smad2 and 3. Eight different specific shRNAs have been tested to specifically reduce Smad2 or 3 protein levels (supplemental Table S2). Two of them efficiently knocked down Smad3, while only one reduced Smad2 protein levels (Fig. 2A). To further prove the functional inhibition of Smad3, we cotransfected plasmids expressing c.a.ALK5, a Smad reporter, and shRNAs specific for Smad3 or for LacZ into adult muscles. Knockdown of Smad3 with two different shRNAs efficiently reduced Smad3 reporter activity in fibers expressing c.a.ALK5 (Fig. 2, B and C). However, between the two shRNAs, oligo2 completely prevented Smad reporter activation, suggesting a more efficient Smad3 knockdown. Identical results were obtained by using c.a.ALK4 (not shown). The result of the RNAi experiments was validated by a rescue experiment in which we used a human Smad3 cDNA. There are three mismatches between the mouse Smad3(2) small interfering RNA (siRNA) sequence and the corresponding human sequence; therefore human Smad3 should not be silenced by the mouse siRNA. In fact, cotransfection of human Smad3 cDNA restored reporter activity in muscle fibers in which mouse Smad3 was knocked down (Fig. 2D). Altogether, these functional experiments validate, in vivo, the RNAi approach against Smads.

Fig. 2.

RNA interference (RNAi)-mediated approach in adult fibers efficiently blocked Smad2 and 3. A: RNAi-mediated knockdown of Smad3 and Smad2 revealed by immunoblotting. Murine embryonic fibroblasts (MEFs) were transfected with vectors expressing two different small hairpin RNAs (shRNAs) against Smad3 and one against Smad2. B and C: Smad-sensor activity is inhibited by Smad 3 knockdown. Smad-dependent reporter was transfected into adult tibialis anterior muscle with vectors expressing shRNAs against Smad3 or LacZ with or without c.a.ALK5. A Renilla luciferase vector was cotransfected to normalize for transfection efficiency. Seven days later, firefly/Renilla luciferase activity was determined. *P < 0.01, n = 4. D: inhibition of Smad-reporter activity by Smad3 RNAi is rescued by overexpression of human Smad3 (hSmad3). Adult skeletal muscles were transfected with Smad-reporter, c.a.ALK5, shRNA specific for mouse Smad3, and human Smad3. Experimental conditions were as in B. *P < 0.01; n = 4.

Activation of TGF-β pathway induces muscle atrophy via Smad2/3 without inducing MuRF1 expression.

We then monitored whether the activation of TGF-β signaling was sufficient to trigger muscle atrophy in adult fibers. Independent of which activated type I receptor was expressed, cross-sectional area was reduced by 20% in myofibers 2 wk after transfection (Fig. 3, A and B). To determine whether Smads are required for the atrophic action of the myostatin pathway, we knocked down Smad2 and 3 in fibers expressing c.a.ALK5. Smad3 and 2 inhibition completely blunted the atrophic effects of ALK5, suggesting that they are necessary for the myostatin pathway (Fig. 3, C and D). Furthermore, knocking down Smads is sufficient to promote muscle growth independent of the expression of activated type I receptor. Next, we tested whether Smad2/3-dependent atrophy requires the upregulation of the critical atrophy-related ubiquitin ligases atrogin-1/MAFbx and MuRF1. We cotransfected atrogin-1/MAFbx and MuRF1 promoters with activated type I receptors. These promoters have been already shown to contain the major regulatory elements for the correct upregulation of these ubiquitin ligases in atrophying muscles (4, 10, 37). Expression of c.a.ALK5 weakly activated atrogin-1 promoter and did not affect MuRF1 promoter (Fig. 3, E and F). Thus myostatin signaling can trigger an atrophy program in adult myofiber independent of MuRF1 and partially dependent on atrogin-1.

Fig. 3.

Activation of TGF-β pathway induces muscle atrophy via Smad2/3 without inducing muscle RING-finger protein 1 (MuRF1) expression. A and B: activation of the myostatin pathway induces mild atrophy. Adult muscle fibers were transfected with c.a.ALK5 (A) or c.a.ALK4 (B) and examined 2 wk later. Cross-sectional area of transfected fibers, identified by anti-HA immunofluorescence, was measured. *P < 0.001; n = 400. C and D: inhibition of Smads blocks ALK5-mediated muscle atrophy. Adult muscle fibers were transfected with vectors expressing shRNAs against Smads or LacZ with or without c.a.ALK5. Cross-sectional area of transfected fibers, identified by anti-HA immunofluorescence, was measured. *P < 0.001; n = 200. E: the activity of the atrogin-1 promoter is weakly induced by c.a.ALK5. Atrogin-1 promoter was transfected into adult tibialis anterior muscle in the presence or absence of c.a.ALK5. A Renilla luciferase vector was cotransfected to normalize for transfection efficiency. Seven days later, firefly/Renilla luciferase activity was determined. #P < 0.05; n = 4. F: the activity of MuRF1 promoter was not affected by c.a.ALK5. Experimental conditions were as in E; n = 4.

Inhibition of TGF-β pathway induces muscle growth independent of satellite cells and partially dependent of mTOR pathway.

Since activation of TGF-β downstream targets triggers muscle atrophy, we then ask whether blocking TGF-β signaling stimulates muscle hypertrophy. Inhibition of the myostatin pathway by expressing d.n.ActRIIB promoted a 29% increase in myofiber size after 2 wk (Fig. 4A). It has been suggested that myofiber hypertrophy may be sustained by activation and fusion of muscle stem cells. To address the possibility that fibers expressing mutant receptors can signal to adjacent satellite cells, we treated mice with BrdU after transfection. Two weeks later, we monitored for incorporation of BrdU-positive nuclei into fibers expressing d.n.ActRIIB. Despite the presence of scattered BrdU-positive nuclei in interstitium, we never detected BrdU in myonuclei of transfected fibers (Fig. 4B). Positive controls confirmed the efficiency of BrdU incorporation. In fact, BrdU-positive myonuclei were revealed in regenerated muscles and in myofibers during postnatal growth (supplemental Fig. S2, A and B). Therefore, the increase in myofiber size during myostatin inhibition is not due to the recruitment of satellite cells and addition of new nuclei into adult fibers. Since the Akt-mTOR axis is another important regulator of muscle hypertrophy, we asked whether mTOR may mediate some of the effects of myostatin inhibition. We used a pharmacological approach to inhibit mTOR by using rapamycin. The treatment efficiently blocked mTOR activity (Fig. 4C) according to our recent results (21). However, mTOR inhibition per se did not affect myofiber size and did not induce atrophy but partially prevented hypertrophy induced by d.n.ActRIIB (Fig. 4D). To confirm these results, we used a genetic approach. We recently set up an RNAi technique that efficiently blocked mTOR signaling in adult myofibers (21). Knocking down mTOR partially reduced muscle growth induced by d.n.ActRIIB (Fig. 4E). Inhibition of muscle growth, obtained by rapamycin and by RNAi experiments, in d.n.ActRIIB-transfected fiber was of 33% and 40%, respectively. The present results represent the first example of stimulated muscle growth independent of mTOR and satellite cells. These findings are in line with a recent study that showed that stimulation of protein synthesis in vivo by anti-myostatin treatment is mTOR independent (44). We next tested whether Smad inhibition could recapitulate the growth-promoting effect of d.n.ActRIIB. RNAi-mediated knockdown of Smad3 alone or of both Smad2 and 3 induced a 10% and 22% increase in fiber size after 2 wk, respectively (Fig. 4F). It is worth noting that Smads have an additive effect. Importantly, the increase of myofiber size obtained by Smad knockdown is very close to that obtained by d.n.ActRIIB (29%) (Fig. 4A).

Fig. 4.

Inhibition of TGF-β pathway induces muscle growth independent of satellite cells and partially dependent of mammalian target of rapamycin (mTOR) pathway. A: inhibition of the myostatin pathway promotes muscle growth. Adult muscle fibers were transfected with d.n.ActRIIB and examined 2 wk later. Experimental conditions were as in Fig. 3A. Cross-sectional area of transfected fibers, identified by anti-myc immunofluorescence, was measured. *P < 0.001; n = 400. B: adult muscles from wild-type mice were transfected by electroporation with a plasmid expressing d.n.ActRIIB. Five days later, bromodeoxyuridine (BrdU) was injected in mice for the next 8 days. Muscles were collected and stained for BrdU (46) and laminin (red). Transfected fibers were identified on serial sections by anti-myc immunofluorescence (red), and nuclei were stained by Hoechst (blue). Myonuclei are below laminin staining while interstitial nuclei are into laminin staining. Two different muscles are depicted. C: rapamycin treatment causes dephosphorylation of S6, a downstream target of mTOR, in muscles transfected with d.n.ActRIIB. D: inhibition of the TGF-β pathway promotes muscle growth in the presence of rapamycin. Adult muscle fibers were transfected with d.n.ActRIIB and treated with rapamycin for 12 days. Values represent means of six muscles. *P < 0.001; n = 900. E: inhibition of mTOR partially prevents muscle growth induced by d.n.ActRIIB. Adult muscle fibers were cotransfected with shRNAs specific either for mTOR or for scramble and green fluorescent protein (GFP). Two weeks later, cross-sectional area of transfected fibers, identified by GFP fluorescence, was measured. *P < 0.001; n = 300. F: inhibition of Smads promotes muscle growth. Adult muscle fibers were cotransfected with shRNAs specific either for Smad3 or for Smad3 and Smad2 and GFP. Two weeks later, cross-sectional area of transfected fibers, identified by GFP fluorescence, was measured. *P < 0.001; n = 200. The increase of myofiber size obtained by Smad knockdown (22%) was closely related to that obtained by d.n.ActRIIB (29%).

Inhibition of TGF-β pathway partially protects from denervation- and fasting-induced muscle atrophy.

These findings suggest that signals from TGF-β type I and type II receptors can modulate the size of myofibers, but leave open the question of whether TGF-β signaling is necessary for the activation of an atrophy program. To test this hypothesis, we expressed d.n.ActRIIB in denervated muscle and we quantified myofiber size after 2 wk. Inhibition of the pathway did not prevent muscle atrophy (Fig. 5A). However, denervated fibers expressing d.n.ActRIIB were 20% bigger than untransfected fibers (Fig. 5A), suggesting that a partial protection against muscle loss was achieved. To confirm this result, we knocked down Smad2 and Smad3 in denervated muscles and checked fiber size after 2 wk. Knockdown of Smads increased myofiber size and partially prevented muscle loss during denervation (Fig. 5B). Similarly, expression of d.n.ActRIIB in fasted muscle results in 10% protection from muscle loss (Fig. 5C). Therefore denervation- and starvation-induced muscle atrophy does not completely require Smad2/3 transcriptional activity.

Fig. 5.

Inhibition of TGF-β pathway partially protects from denervation-induced muscle atrophy. A: inhibition of the myostatin pathway partially protects from muscle atrophy. Adult muscle fibers were transfected with myc-tagged d.n.ActRIIB, denervated, and examined 2 wk later. Cross-sectional area of transfected fibers, identified by anti-myc immunofluorescence, was measured. *P < 0.001; n = 400. B: inhibition of Smads partially prevents muscle loss during denervation. Adult muscle fibers were transfected with vectors expressing shRNAs against Smads or LacZ together with GFP and denervated. Two weeks later, cross-sectional area of transfected fibers, identified by GFP fluorescence, was measured. *P < 0.001; n = 700. C: inhibition of TGF-β pathway partially prevents muscle loss during fasting. Adult muscle fibers were transfected with myc-tagged d.n.ActRIIB; 1 wk after transfection, mice were fasted for 32 h and examined. Cross-sectional area of transfected fibers, identified by anti-myc immunofluorescence, was measured. *P < 0.001; n = 400.

Akt blocks the atrophic effect of TGF-β signaling.

The growth-promoting IGF-1/PI3K/Akt pathway has been shown to prevent muscle wasting by blocking proteasomal and lysosomal protein breakdown through inhibition of FoxO transcription factors (35). We then asked which of the two pathways is dominant and whether, besides the partial mTOR requirement for muscle growth, other cross-talk between them exists. We have recently generated a transgenic mouse line in which an Akt-estrogen receptor (Akt-ER) fusion protein can be activated in an inducible manner by tamoxifen specifically in skeletal muscle (21). In denervation experiments, Akt was activated in adult animals immediately after cutting the sciatic nerve. Overexpression of activated Akt in muscles completely prevented muscle loss in denervated soleus and TA muscles (Fig. 6A and supplemental Fig. S3). These findings confirm that Akt signaling prevents muscle wasting by blocking catabolic pathways. We then performed gene expression profiling to determine which genes of catabolic pathways are suppressed by Akt activation. Activation of Akt for 24 and 48 h led to the downregulation of atrogin-1, autophagy genes (i.e., LC3, GABARAPl1, Atg7), and, interestingly, also of ActRIIB transcripts (supplemental Fig. S4). Quantitative real-time PCR confirmed that after 4 days of Akt activation, ActRIIB transcript level is still twofold downregulated (Fig. 6B). We have shown that Smad2/3 are downstream of ActRIIB and are crucial for TGF-β effects while FoxOs are downstream of Akt and regulate expression of atrophy-related genes. To understand which of the two pathways is dominant at transcriptional level, we expressed activated type I receptors in muscles of Akt transgenic mice. To check that Akt activation does not affect Smad recruitment on target promoters, we transfected c.a.ALK5 together with Smad reporter in Akt transgenic mice. Smad activity was strongly increased by expression of activated type I receptor and was not significantly affected by Akt (Fig. 6C). Since luciferase transcription occurs in the nucleus via RNA polymerase II, activated Akt does not block Smad phosphorylation, nuclear translocation, interaction, and transactivation of CAGA promoter. Moreover, the transfection efficiency was not different between muscles with activated and nonactivated Akt (not shown). However, induction of Akt completely suppressed the atrophic action of c.a.ALK4 and c.a.ALK5 (Fig. 6D and supplemental Fig. S5). This finding suggests that besides the Akt effect on ActRIIB mRNA level there is a further cross-talk between the two pathways downstream of ALK4/5. Then we asked whether the two pathways can synergize to promote maximal growth. We expressed d.n.ActRIIB in muscles of Akt transgenic mice and we analyzed myofiber size after 2 wk from Akt activation. Activation of Akt induced muscle hypertrophy but the presence of d.n.ActRIIB greatly enhanced muscle growth (Fig. 6E). In conclusion, the two pathways cross-talk at different levels but they have different routes to promote muscle hypertrophy.

Fig. 6.

TGF-β effects on muscle mass are mediated by cross-talk between Smad and Akt-FoxO axis. A: Akt activation, induced by tamoxifen in soleus muscles from Akt transgenic mice, promotes muscle growth and blocks muscle atrophy. Adult muscles from Akt transgenic mice were treated with tamoxifen or vehicle and denervated for 2 wk. Soleus muscles were collected and weighed. Values represent means of four animals. B: Akt inhibits the expression of ActRIIB. Akt transgenic mice were treated with tamoxifen or vehicle for 4 days. Quantitative PCR analysis was performed in triplicate using specific oligonucleotides. This experiment validated microarray data from skeletal muscle after 1 day and 2 days of Akt activation (supplemental Fig. S4). *P < 0.01 C: Akt does not prevent Smad activation. Smad-dependent reporter was transfected into adult tibialis anterior muscle of Akt transgenic mice with or without c.a.ALK5. A Renilla luciferase vector was cotransfected to normalize for transfection efficiency. Mice were treated with tamoxifen or vehicle for 7 days, and firefly/Renilla luciferase activity was determined. D: Akt blocks muscle atrophy induced by c.a.ALK4. Akt transgenic mice were transfected with c.a.ALK4, treated with tamoxifen or vehicle, and examined 2 wk later. Cross-sectional area of transfected fibers, identified by anti-HA immunofluorescence, was measured. At least five muscles for each point were analyzed. *P < 0.001; n = 900. E: inhibition of TGF-β pathway synergizes Akt hypertrophic effect. Akt transgenic mice were transfected with d.n.ActRIIB, treated with tamoxifen or vehicle, and examined 2 wk later. Cross-sectional area of transfected fibers, identified by anti-myc immunofluorescence, was measured. *P < 0.001; n = 900.

DISCUSSION

The present study has provided several important new insights concerning the signaling of the TGF-β pathway and its contribution to muscle loss. Specifically, we have shown that 1) inhibition of the myostatin/TGF-β pathway is sufficient to promote muscle growth in adult myofiber independent of satellite cells and partially dependent of mTOR, 2) Smad2 and 3 are the transcription factors downstream of TGF-β and activate an atrophy program that is MuRF1 independent and 3) in agreement with the results of D. J. Glass's group (41a) investigating human muscle cell culture, i.e., that the Akt pathway cross-talks with TGF-β signaling. Importantly, activated Akt overcomes most of the TGF-β effects on muscle mass.

Myostatin is a TGF-β family member that is expressed and secreted predominantly by skeletal muscle (17). The function of myostatin appears to be conserved across species, since mutations in the myostatin gene induce bigger muscles in cattle, mice, dogs, sheep, and humans (5, 18, 25, 26, 39). However, in these studies the deletion of the gene occurs early during development; therefore the increase in muscle mass is a consequence of hyperplasia, an increase in cell number, and hypertrophy, an increase in cell size. How much cellular turnover or protein turnover in adult myofibers contributes to a hypertrophic phenotype was unclear. Furthermore, it has been shown that, besides myostatin, other molecules can signal to the pathway and affect muscle mass. Overexpression of follistatin, an inhibitor of myostatin and other TGF-β members including GDF-11 (18), promotes a great increase in muscle size that is more robust than that obtained by the deletion of the myostatin gene (18). Importantly, mice resulting from follistatin transgenic and myostatin knockout mice show a tremendous increase in muscle mass (16). Altogether, these findings support the concept that other myostatin-like molecules are present and relevant for muscle growth (16). To address these questions, we used our transfection technique to perturb the pathway in terminally differentiated myofibers of adult animals. Furthermore, to dissect the pathway independently of the ligand, and to avoid the effects of myostatin secretion on satellite cells, we overexpressed mutants of myostatin receptors in adult fibers. Thus, by using gain and loss of function approaches, we showed that Smad2 and Smad3 are downstream of myostatin type II receptors. It has been reported that Smad2/3 are phosphorylated after myostatin treatment in cell culture, but their functional involvement in the pathway has never been addressed in adult muscles (14, 31). By knocking down Smad2 and 3, we ruled out their critical role in myostatin signaling. Inhibition of Smad2 and 3 had an additive effect and could completely suppress ALK4/5-mediated muscle atrophy. Moreover, inhibition of the pathway either by expressing the inactive activin receptor, d.n.ActRIIB, or by knocking down the Smads promoted an almost identical myofiber hypertrophy.

Few studies have explored the effect of myostatin inhibition in adults (43, 45). Treating 24-wk-old mice with an anti-myostatin antibody for 5 wk induced a 12% increase in muscle mass (45). Furthermore, when tamoxifen-inducible Cre recombinase removed the floxed myostatin gene in 4-mo-old mice, muscle mass increased by 25% during the next 3 mo (43). These studies never addressed the mechanisms of muscle growth and whether such growth was a consequence of satellite cell activation. Since we perturbed the pathway in adult fibers with proteins that were not secreted, it is unlikely that muscle stem cells contribute to hypertrophy. This concept is supported by BrdU experiments that showed no incorporation of new myonuclei into fibers transfected with d.n.ActRIIB. Furthermore, muscle growth during myostatin inhibition is partially mTOR independent. Thus, Smads may affect muscle growth by directly impinging on the translational machinery as was recently proposed for atrogin-1. In fact, atrogin-1-dependent degradation of the translational factor eIF3f was sufficient to affect myotube size. Conversely, overexpression of eIF3f induced hypertrophy in adult muscles (12).

Systemic overexpression of myostatin leads to a variable degree of muscle loss that depends on the experimental model used (7, 32, 47). Muscle atrophy is a complex process that requires activation of many genes and coordination of different proteolytic systems. Defining which pathway is critical for muscle atrophy is a crucial issue for the correct identification of drug targets to combat muscle wasting. Our findings show that activation of the TGF-β pathway triggers muscle atrophy but this action seems to require an inhibition of Akt (Fig. 6D).

Smad transcription factors recognize the DNA sequence CAGAC, a very simple sequence that is extremely common in the genome. In fact, if Smad affinity for this frequent sequence is sufficiently high enough to allow Smad recruitment on the promoter, Smads would decorate the entire chromosome (22). Therefore, activated Smad proteins must associate with different DNA-binding cofactors for the recognition and regulation of specific target genes (22). Importantly, FoxO members, the Akt downstream targets, play such a role (9, 40). In keratinocytes, glioblastoma and neuroepithelial cells, FoxOs are critical Smad partners for the expression of p21 and p15 (40). FoxO is also essential for 11 of the 115 immediate gene activation responses to TGF-β in keratinocytes (9). Interestingly, myostatin has been reported to upregulate the critical atrophy-related ubiquitin ligases in muscle cell culture via FoxO1 (23). Furthermore, it has been reported that myostatin expression is controlled by FoxO1, supporting the notion that the myostatin pathway is synergistic with Akt-FoxO signaling (1). Thus, modulation of Smad activity may become crucial for maintaining or exacerbating an atrophy program.

The present study connects the two most important pathways for regulation of muscle mass and dissects the role of Smads in myostatin signaling. Our findings have an important impact in identifying the molecular players and in dissecting their role during muscle atrophy. Thus, myostatin inhibitors can have a rationale for the therapeutic treatment of sarcopenia during aging or of muscle wasting during diseases especially when they are combined with activators of IGF-1-Akt axis.

GRANTS

This work was supported by grants from Agenzia Spaziale Italiana OSMA (Osteoporosis and Muscular Atrophy Project); Telethon-Italy (TCP04009); the Italian Ministry of Education, University and Research (PRIN 2007); Ministero dell'Istruzione dell'Universita e della Ricerca (MiUR); and Compagnia San Paolo (to M. Sandri).

Acknowledgments

The critical reading of Kenneth Dyar is gratefully acknowledged. The d.n.ActRIIB, c.a.ALK4, myc-Smad3, HA-human-Smad3, and Flag-Smad2 were a generous gift of M. Whitman. c.a.ALK5 and pGL3(CAGA)12-luc were a gift of S. Piccolo. 5 kb MuRF1 promoter-luc was a gift of S. Lecker.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 41a.
  43. 42.
  44. 43.
  45. 44.
  46. 45.
  47. 46.
  48. 47.
View Abstract