Myostatin or GDF-8 (growth and differentiation factor-8) is a member of TGF-β (transforming growth factor-beta) superfamily that is highly conserved among species (reviewed in Lee, 2004
). Solid evidence indicates that myostatin is a potent, negative regulator of muscle growth during development and adult life. The physiological function of myostatin likely consists in limiting an excessive growth of skeletal muscles. Spontaneous mutations of myostatin have been originally detected in cattles (McPherron and Lee, 1997
) and other animals displaying an excessive skeletal muscle development and myofibers stronger and with larger size than normal (Mosher et al., 2007
). A mutation at the myostatin locus that leads to the absence of myostatin expression and an abnormal muscle growth has also been reported in a child (Schuelke et al., 2004
). This “hypermuscular” phenotype has been replicated in mice either by genetic ablation of the myostatin gene (McPherron et al., 1997
) or by pharmacological blockade of myostatin protein (reviewed in Lee, 2004
). Importantly, inactivation of myostatin in dystrophic mice exerted beneficial effects on disease progression (Wagner et al., 2002
; Bogdanovich et al., 2002
), suggesting that myostatin is a primary target of pharmacological interventions in MDs. Since myostatin activity results from complex interactions with other members of the TGF-β superfamily, it is reasonable to extend this concept to the entire TGF-β signaling to muscle regeneration.
In the next paragraphs, we will illustrate the different levels of regulation of myostatin activity and will describe the successful applications of interventions targeting key interactions between myostatin and other TGF-β family members, which control muscle regeneration and other processes (i.e. fibrosis) that participate to the progression of MD.
3.1. Regulation of myostatin activity by extracellular factors
Members of the TGF-β family undergo a complex regulation at the level of biosynthesis. Myostatin is synthesized as a precursor, which is processed by furin proteases to generate a dimer composed by an N-terminal propeptide that is non-covalently bound to a 110 amino acid, biologically active C-terminal fragment (McPherron et al., 1997
). This complex is secreted and circulates as an inactive, latent myostatin form (Lee and McPherron, 2001
). One mechanism for activating myostatin consists of a proteolytic cleavage of the propeptide at the residue Asp76 by members of the bone morphogenetic protein-1/tolloid (BMP-1/TLD) family of metalloproteinase (Wolfman et al., 2003
). The importance of propeptide proteolysis in the activation of latent myostatin is highlighted by experiments showing an increased muscle mass in transgenic mice over-expressing the propeptide (Lee and McPherron, 2001
). Alternatively, the carboxyl-terminal is regulated in the extracellular environment through interactions with binding proteins, such as follistatin, follistatin related gene (FLRG), and Gasp-1 (Lee, 2004
). Although each of these proteins can inhibit myostatin activity in vitro
, only the inhibitory activity of follistatin has been demonstrated in vivo
(Lee and McPherron, 2001
; Zimmers et al., 2002
). Consistently, follistatin knockout mice have reduced muscle mass at birth (Matzuk et al., 1995
), and transgenic expression of follistatin produced a hypermuscular phenotype that resembles that of myostatin knock out (Nakatani et al., 2007
). Notably, transgenic expression of follistatin in myostatin knock out mice produces an additive effect (a quadruplication of muscle mass, in comparison to the duplication observed in myostatin knock-out mice) (Lee, 2007
). This suggests that follistatin can regulate muscle mass by mechanism(s) that are independent on myostatin inactivation.
3.2. Myostatin intracellular signaling
The unbound C-terminal myostatin dimer binds to the activin type II receptors, ACVR2A and ACVR2B and elicits an intracellular signaling pathway that is typical of members of the TGFβ family (Lee, 2004
). Myostatin/activin type II receptor complex engages type I receptors, such as ALK4 and/or ALK5, leading to phosphorylation of downstream signaling components—the R-Smads, Smad2 and Smad3. Phosphorylated Smad2 and 3 associate with co-Smad, Smad4, and enter the nucleus to regulate transcription of downstream genes (Lee et al., 2005
). Importantly, the identity of the genes that mediate myostatin effects on muscle regeneration is still unknown.
3.3. Strategies of myostatin blockade in the therapy of muscular dystrophies
The pharmacological strategies to block myostatin signaling in adult muscles are mostly based on the mechanism of regulation of myostatin biosynthesis and activity described above. Compounds capable of binding and inhibiting the C-terminal dimer have been the focus of the most recent and successful strategies. Targeting of the C-terminal dimer by neutralizing monoclonal antibody (JA16) resulted in the increase of muscle mass and function in wild type mice (Whittemore et al., 2003
) and could rescue the pathological phenotype in dystrophin-deficient mdx mice (Bogdanovich et al., 2002
). This latter study showed the first evidence that myostatin blockade in dystrophic mice increased the myofiber size and alleviated signs and symptoms of the disease, such as decline in strength, fiber susceptibility to degeneration and fibrosis. This original observation provided the impetus for the evaluation of the monoclonal antibody, MYO-029, in a trial with patients with muscular dystrophy that was recently published (Wagner et al., 2008
). This study reported on the safety of the molecule, but did not demonstrate any sign of clinical improvement in the patients treated with MYO-029. It is possible that the selection of dystrophic patients at late stage of the disease – that is, when the regenerative response is exhausted – eliminated the myostatin substrate (e.g. regenerating muscle progenitors) and therefore preclude any appreciable effect of myostatin blockade. Inactivation of myostatin with a propeptide fused to an Fc domain, which enhances stability in vivo
, also alleviated the signs of disease when injected into mdx
mice (Bogdanovich et al., 2005
). Similar results were observed by genetic ablation of myostatin in mdx mice, obtained by breeding myostatin knock-out and mdx mice (Wagner et al., 2002
). However, the rescue of the pathological phenotype in this experimental setting was less evident, when compared to that described by Bogdanovich and colleagues. It is unclear the reason for such discrepancy, although it might depend on the different timing of myostatin blockade (during embryogenesis vs. adult life), on the magnitude of myostatin inactivation and the possibility that anti-myostatin antibodies could affect other pathways. Indeed, the injection of the soluble form of the ACVRIIB receptor fused to an Fc domain (ACVRIIB-Fc), which was designed to block myostatin activity, led to an increased muscle growth in myostatin-deficient mice (Lee et al., 2005
), suggesting an action through a myostatin-independent pathway. Likewise, inactivation of myostatin by either transgenic expression of follistatin (Nakatani et al., 2007
)orby single injection of adeno-associated viral (AAV) vector that delivered a follistatin-splicing variant (FS-344) (Haidet et al., 2008
) increased muscle mass and strength, and reduced the histological signs of disease in dystrophic mice. Moreover, increased levels of follistatin appear to mediate the beneficial effects of two independent therapeutic strategies in mdx mice—one based on the deacetylase inhibitor delivery (Minetti et al., 2006
) and the other on nitric oxide release (Brunelli et al., 2007
) (see following sections).
Collectively, these data indicate that interactions between myostatin, follistatin and possibly other members of the TGF-beta pathway provide a valuable target for pharmacological treatments of DMD.
The inhibition of myostatin was also effective in alleviating the pathological phenotype of caveolin 3-deficient mice (a model of LGMD 1C) (Ohsawa et al., 2006
), but not in a mouse model of lamin-deficient muscular dystrophy, which caused a more severe disease phenotype (Li et al., 2005
). Likewise, antibody-mediated inactivation of myostatin failed to revert the pathological phenotype of aged δ-sarcoglycan null mice (Parsons et al., 2006
). The lack of effect of myostatin inactivation in lamin-deficient and δ-sarcoglycan null mice correlates with the reduced regeneration and the more severe fibrosis observed in these animals. Once again, this suggests that myostatin blockade could be effective only at early stages of disease progression. However, differences in experimental conditions and distinct types of muscular dystrophies could also explain these results.
3.4. Effect of myostatin blockade on muscle regeneration
How does blockade of myostatin signaling interfere with the progression of muscular dystrophy? Myostatin signaling regulates both muscle regeneration and fibrosis, which are two interdependent processes. The data reported in the sections above show a correlation between the enhanced muscle regeneration and the reduced fibrosis, upon myostatin blockade in young dystrophic mice. Instead, an impaired ability to regenerate fibrotic muscles at late stages of the disease coincides with the therapeutic failure of myostatin blockade.
The negative impact of myostatin on muscle regeneration is well documented. The widespread increase in skeletal muscle mass displayed by myostatin null mice results from the combination of muscle cell hyperplasia and hypertrophy during development (McPherron et al., 1997
). However, myostatin action is not limited to embryonic development, as myostatin regulates muscle mass also in adult animals (Lee and McPherron, 2001
; Grobet et al., 2003
). Given the strong analogies in the molecular mechanisms regulating developmental myogenesis and muscle regeneration in adult life (Snider and Tapscott, 2003
), it is tempting to speculate that myostatin signaling targets conserved effectors of these processes. Indeed, recent studies revealed that excess of myostatin downregulates the expression of Pax-3, Pax7, MyoD and Myf5 (Amthor et al., 2002
; McFarlane et al., 2008
). Receptors for myostatin are found on numerous muscle cell lines, and the effects on muscle fiber number probably result from blockade of muscle cell proliferation and differentiation (Lee, 2004
). Studies with cultured myoblasts showed that elevated concentrations of active myostatin inhibits mpc proliferation by up-regulating the cyclin-dependent kinase inhibitor p21 (Waf1/Cip1) and leading to the accumulation of unphoshorylatyed, active pRb (Thomas et al., 2000
). Studies with satellite cells showed the inverse patterns of myostatin and follistatin levels in quiescent vs. activated satellite cells, with follistatin/myostatin ratio increasing during satellite cells activation (McCroskery et al., 2003
). Taken together, these and other data support the notion that myostatin inhibits both proliferation and differentiation potential of satellite cells. Interestingly, the increased regeneration observed in myostatin null mice did not decline along the life span. And aged myostatin-deficient mdx mice, which have undergone multiple degeneration/regeneration cycles, continued to maintain an enhanced regenerative response and an increased muscle mass (Wagner et al., 2005
). This evidence indicates that the absence of myostatin counters satellite cell exhaustion to regenerate diseased or aged muscles, and indirectly suggests an effect of myostatin on symmetric division of satellite cells to maintain a reserve pool.
The interest on therapeutic blockade of myostatin extends to the possibility to increase the size and the strength of muscles in atrophic conditions. However, recent studies reported on the excessive muscle growth but impaired force generation in two independent mouse lines that harbor mutations in the myostatin gene, constitutive null (myostatin−/−) and compact (Berlin High Line, BEHc/c
) mice (Amthor et al., 2007
. These data are in conflict with the evidence that mammals with spontaneous mutation of the myostatin gene show enhanced muscle performance (Mosher et al., 2007
; Schuelke et al., 2007). However, it should be noted that the therapeutic benefit of myostatin inactivation in dystrophic muscles would derive from a decreased susceptibility to degeneration of bigger myofibers (Zammit and Partridge, 2002
), rather than from an increased force of contraction.
3.5. TGF-β-mediated regulation of muscle regeneration
It is becoming progressively clear that the TGF-β network signaling profoundly influences proliferation and differentiation of mpc. The relative expression of the TGF-β family members (BMP4, gremlin, activin) and receptors in human mpc (side population and main population) and myofibers establishes a regulatory network that reciprocally control cell proliferation in a paracrine fashion (Frank et al., 2006
). This network is implicated in the regulation of the number of mpc available for sequential waves of regeneration and is therefore a potential target for interventions aimed at ensuring long-lasting efficacy of regeneration. Quite surprisingly, the impact of TGF-β on the symmetry of mpc cell division has not been yet addressed. Furthermore, members of TGF-β family are known inhibitors of terminal differentiation of muscle cells in vitro
(Liu et al., 2001
), and increased levels of TGF-β in vivo
limit regeneration of injured muscle by inhibiting satellite cells proliferation and differentiation (Cohn et al., 2007
). In resident satellite cells of aged animals, excessive levels of TGF-β induce high pSmad3, which antagonizes Notch signaling by activating the expression of cyclin-dependent kinase inhibitors (cdks) (Carlson et al., 2008
). Moreover, TGF-β signaling regulates fibrosis in different tissues, including muscles. Consistent with the importance of the TGF-β signaling in regulating both regeneration and fibrosis, Cohn et al. showed that blockade of the angiotensin II receptor 1 (AT1) with Losartan ameliorates the pathological phenotype in dystrophic mice via the inhibition of the TGF-β signaling (Cohn et al., 2007
). In this study, both improvement in muscle regeneration and decreased fibrosis were observed in mdx mice after prolonged exposure to Losartan.
3.6. Future challenges
Future studies should optimize pharmacological strategies to maximize the benefits deriving from myostatin blockade or from manipulation of the TGF-β network, and to identify the profile of dystrophic patients suitable for such an effect. This can be achieved by the complete elucidation of the mechanism by which myostatin regulates muscle regeneration and fibrosis. It is possible that the modulation of the myostatin/follistatin pathway, and more in general the TGFβ network, could have an independent impact on different parameters, such as mpc activity, inflammation and fibrosis. While the combination of these effects might result in a global positive impact on regeneration, it would be interesting to identify individual pathways that can selectively improve specific disease features—i.e. muscle fibrosis in older patients. The discovery that increased levels of follistatin produce beneficial effects in dystrophic muscles both through myostatin blockade and via myostatin-independent pathways, suggests that strategies that upregulate follistatin might have a stronger therapeutic potential than selective myostatin inhibitors. Mouse models of dystrophies that show resistance to myostatin blockade should be exploited to address this issue.
Additionally, it will be critical to develop methods that assess the magnitude of myostatin inhibition in muscles, to monitor the treatment effectiveness. Possibly, the development of small molecules inspired by the mechanism of myostatin blockade might help to increase the selectivity of the treatment.