Myostatin and muscle fiber size. Focus on “Smad2 and 3 transcription factors control muscle mass in adulthood” and “Myostatin reduces Akt/TORC1/p70S6K signaling, inhibiting myoblast differentiation and myotube size”

Stephen L. Welle

the plasticity of skeletal muscle continues to fascinate physiologists and cell biologists. Many of the key molecules in the pathways that signal growth or atrophy of muscle have been identified, along with a number of hormones and growth factors that regulate these pathways. This knowledge may enable us to more effectively prevent or reverse muscle atrophy associated with a number of conditions: cancer, muscular dystrophies, old age, inability to maintain normal activity, glucocorticoid therapy, arthritis, or settings associated with high levels of catabolic hormones and cytokines. Of course, there is much more to learn about muscle growth and atrophy pathways, and two articles (9, 12) contribute to our understanding of one of the important regulators of muscle growth: myostatin [also known as growth differentiation factor-8 (GDF-8)].

The discovery of myostatin, a member of the transforming growth factor-β (TGF-β) superfamily, was reported 12 years ago (7). When myostatin was knocked out in mice, their skeletal muscles were two or more times the normal size. Remarkably, no other major effects of myostatin deficiency were noted, although subsequent research has revealed a few. Expression of the myostatin gene was high in muscle, but absent or very low in other tissues. It was soon realized that naturally occurring mutations in this gene, which is highly conserved across species, are responsible for hypermuscularity in certain breeds of cattle, sheep, and dogs. The active domain of the peptide is identical in mice and humans, and a homozygous loss of function mutation in a boy was associated with marked hypermuscularity in infancy (10). Anti-myostatin antibodies and other proteins that bind myostatin were reported to increase muscle mass and strength in normal and dystrophic mice (2, 18). All of this information suggested that myostatin could be an ideal target for drugs or biological compounds to combat muscle atrophy. The first human trial of an anti-myostatin antibody did not demonstrate functional improvement or significant muscle hypertrophy (at low doses) in adults with muscular dystrophies (14). This negative result should not discourage more research on myostatin; one of the main problems with that first trial was that no molecular markers were assessed to confirm that the administered doses of antibody had an impact on the relevant signaling pathways in muscle. A deeper understanding of the molecular and phenotypic effects of myostatin is needed before advocating or rejecting myostatin inhibition as an anti-atrophy strategy. From the standpoint of translating basic knowledge into biomedical advances, it is especially important to study postdevelopmental effects of myostatin because anti-myostatin agents generally would be used in adults.

It has been assumed that signaling by myostatin is the same as that of TGF-β, i.e., binding of the ligand to a type II receptor (activin receptor 2B appears to be the critical one for myostatin) induces phosphorylation of a type I receptor, which phosphorylates Smad2 and Smad3, thereby facilitating formation of a complex of these phospho-Smads with Smad4. This complex enters the nucleus and alters gene transcription. It had not been proven that activation or inhibition of this pathway is sufficient to alter myofiber size.

In their article, Sartori et al. (9) show that activation of the pathway in vivo in adult mice, which the authors accomplish by electroporating genes encoding a constitutively active form of either of the TGF-β/activin type I receptors [activin receptor-like kinase 4 (ALK4) or ALK5], induces myofiber atrophy. This effect was blocked by small hairpin RNAs (shRNAs) targeting Smad2 and Smad3. Administration of Smad2 and Smad3 shRNAs together induced fiber hypertrophy to an extent similar to previously reported effects of anti-myostatin treatments in adult mice (16, 18). In their article, Trendelenburg et al. (12) also demonstrated that Smad2 and Smad3 small interfering RNAs (siRNAs) block myostatin-induced and TGF- β1-induced atrophy of human myotubes. There is strong evidence that other ligands besides myostatin contribute significantly to the basal level of myostatin-like activity in muscle, although it remains to be determined which ones are involved (4). The study by Trendelenburg et al. shows that both TGF-β1 and GDF-11 are more potent activators of this pathway (as reflected by inhibition of myoblast differentiation) in human myotubes than is myostatin, although it is unknown whether endogenous levels of these specific ligands in muscle are sufficient to influence muscle mass.

Many conditions that induce myofiber atrophy do so by increasing the expression of “atrogenes,” which promote degradation of muscle proteins primarily through the ubiquitin-proteasome proteolytic pathway. Two key atrogenes are those encoding the E3 ubiquitin ligases muscle atrophy F-box (MAFbx) (atrogin-1) and muscle RING-finger 1 (MuRF1). Trendelenburg et al. report that myostatin causes atrophy of human myotubes while inhibiting the expression of these genes. This effect occurred at myostatin concentrations of 1–100 ng/ml and differs from the response to a high concentration of myostatin (3,000 ng/ml) in C2C12 myotubes, which increases atrogin-1 (but not MuRF1) expression and the ubiquitination of muscle proteins (6). A modest increase in myostatin levels in rat muscle in vivo also caused atrophy without affecting expression of atrogin-1, MuRF1, or other genes involved in ubiquitination or proteasomal degradation of proteins (1). Sartori et al. report that activation of the myostatin signaling pathway in murine muscle induces atrogin-1, but not MuRF1, promoter activity. I would put a question mark next to the lines showing inhibition of MAFbx and MuRF1 by Smad2/3 in Figure 8 of Trendelenburg et al. because in vivo studies do not support the notion that myostatin inhibits atrogin-1 or MuRF1 expression. None of these studies included a direct measure of proteolysis, which is a critical point because increased expression of atrogin-1 or MuRF1 does not necessarily increase the rate of proteolysis (13). Thus it is important to recall an early study of C2C12 myotubes in which high concentrations of myostatin (6,000 ng/ml) suppressed protein synthesis but did not affect the rate of proteolysis (11). The weight of evidence suggests that myostatin affects myofiber size by affecting protein synthesis rather than degradation (11, 15, 17).

The studies by Sartori et al. and Trendelenburg et al. both addressed the issue of cross talk between myostatin signaling and the Akt/mammalian target of rapamycin (mTOR) pathway. This pathway is responsible for the stimulation of protein synthesis and muscle hypertrophy induced by insulin-like growth factor 1 (IGF-1). IGF-1 induces phosphorylation of the serine/threonine kinase Akt. Phospho-Akt indirectly activates mTOR, which is an integrator of nutrient and growth factor signals that influence protein synthesis and cell size. Trendelenburg et al. report that myostatin inhibits Akt phosphorylation in human myotubes, a result that is consistent with findings indicating that myostatin reduces Akt phosphorylation in mouse C2C12 myotubes (6). Trendelenburg et al. demonstrate that this effect depends on Smads 2 and 3. The path between Smads and Akt is probably not a simple one, because myostatin has the opposite effect on Akt phosphorylation in muscle fibroblasts (5). IGF-1 does not block Smad transcriptional activity but nevertheless can promote myotube hypertrophy when myostatin levels are high. Sartori et al. show that the muscle fiber atrophy induced by Smad2/3 activation (transfection of a gene encoding constitutively active ALK4) is prevented by the presence of constitutively active Akt. They also report that muscle fibers transfected with a gene encoding dominant-negative activin receptor 2B (which inhibits myostatin signaling) are ∼30% larger than normal fibers; this hypertrophy occurs without recruitment of new myonuclei, is diminished by blocking mTOR activity with either rapamycin or mTOR siRNAs, and is enhanced in transgenic mice with constitutively active Akt. However, about 60–65% of the hypertrophic effect of blocking myostatin activity cannot be reversed with anti-mTOR agents. This observation is consistent with our recent observation that rapamycin does not prevent the stimulation of muscle protein synthesis induced by an anti-myostatin antibody (17). Sartori et al. also show that high Akt activity suppresses expression of the myostatin receptor (activin receptor 2B). This does not appear to explain why IGF-1 suppresses the effect of myostatin, because IGF-1 does not blunt myostatin signaling downstream of the receptor (Smad2 phosphorylation).

While the current studies answer some questions regarding myostatin signaling, several questions remain unanswered (Fig. 1). It is safe to conclude that myostatin and IGF-1 have antagonistic effects on Akt phosphorylation and that muscle is resistant to myostatin-induced atrophy when Akt activity is high. However, reduced mTOR activity is unlikely to be the primary cause of myostatin-induced atrophy because blocking mTOR (with either rapamycin or siRNAs) does not reduce muscle fiber diameter or protein synthesis below the normal level. Perhaps reduced mTOR activity is necessary but not sufficient for atrophy to occur and myostatin exerts additional effects that trigger atrophy. Moreover, mTOR signaling accounts for only a fraction of the muscle hypertrophy caused by loss of myostatin activity. Perhaps mitogen-activated protein kinase pathways (p38, ERK1/2, and JNK) are also involved (3, 8, 19). To fully understand the actions of myostatin in mature muscle, we need to know more about regulation of myonuclear domain volume. To understand effects on protein synthesis, we should not focus exclusively on regulators of translational efficiency. The fact that the ratio of RNA to DNA increases with myostatin deficiency (15, 16) suggests that myostatin might be influencing overall RNA production or stability.

Fig. 1.

Established mechanisms and unresolved questions about mechanisms of myostatin effects on muscle cell size, based on several studies cited in text. Solid green arrows signify positive effects and red dashed lines signify negative effects. Intermediary molecules that mediate many of the connections are not shown. Loss of Akt/mammalian target of rapamycin (mTOR)-mediated stimulation of protein synthesis explains less than half of the inhibitory effect of myostatin on muscle cell size, so other unidentified mechanisms must be involved in suppression of protein synthesis by myostatin. Although it is accepted that phospho-Akt blocks FoxO-mediated stimulation of atrogene expression (via FoxO phosphorylation) when IGF-1 levels are high, the current studies do not resolve whether the opposite occurs when Akt phosphorylation is inhibited by myostatin. IGF-1, insulin-like growth factor 1; p70S6K, p70 S6 kinase; TGF-β, transforming growth factor-β.