Cell–matrix interactions in muscle disease

Authors


  • No conflicts of interest were declared.

Abstract

The extracellular matrix (ECM) provides a solid scaffold and signals to cells through ECM receptors. The cell–matrix interactions are crucial for normal biological processes and when disrupted they may lead to pathological processes. In particular, the biological importance of ECM–cell membrane–cytoskeleton interactions in skeletal muscle is accentuated by the number of inherited muscle diseases caused by mutations in proteins conferring these interactions. In this review we introduce laminins, collagens, dystroglycan, integrins, dystrophin and sarcoglycans. Mutations in corresponding genes cause various forms of muscular dystrophy. The muscle disorders are presented as well as advances toward the development of treatment. Copyright © 2011 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Introduction

The extracellular matrix (ECM) is a dynamic structure that provides support and anchorage for cells, segregates tissues from one another and initiates signal transduction pathways. The ECM is composed primarily of glycoproteins, collagens and proteoglycans that are secreted and assembled locally into an organized network to which cells adhere via cell surface receptors. An ECM is already present in the mammalian embryo from the two-cell stage and is a component of the environment of all cell types, although the composition of the ECM and the spatial relationship between cells and the ECM differ between tissues. The ECM includes the interstitial matrix and the basement membrane.

The interstitial matrix is present between cells, while the basement membrane is a thin sheet-like deposition of ECM that surrounds cells (eg muscle cells) or underlies cells (eg epithelial cells) 1, 2. The major components of the basement membranes are laminins, collagen type IV, nidogens and the heparan sulphate proteoglycan perlecan. Laminins and collagen type IV form independent networks that are connected by nidogens and perlecan and the basement membrane is associated with cells through interactions with cell surface receptors (eg integrins and dystroglycan). Hence, a major role of the basement membrane is to provide a solid scaffold for the cells and to separate them from the surrounding interstitial matrix. Apart from providing tissue structure, the basement membrane is also crucial for survival and differentiation of cells. Furthermore, it represents an important barrier that limits bacterial/viral infections or malignant cell infiltration into tissues 3. Consequently, altered basement membranes are responsible for various human diseases. Particularly, in skeletal muscle, mutations in genes encoding ECM proteins and their receptors are accountable for several types of muscular dystrophy and, in particular, congenital muscular dystrophy 4.

Congenital muscular dystrophy forms a heterogeneous group of progressive genetic diseases that are mostly, but not exclusively, inherited in an autosomal recessive manner. The clinical symptoms are present from birth or during the first months of life and include neonatal hypotonia, muscle weakness, delayed motor milestones and joint contractions. Some forms are also associated with central nervous system or peripheral nerve defects. The main morphological features, observed in muscle biopsies, are degeneration–regeneration processes (evidenced by the presence of fibres with centrally located nuclei) and fibrosis infiltration within the muscle tissue 5. Phenotypical variability is present in this group of diseases and ranges from severe and lethal forms to milder types compatible with a normal life span. Depending on the genetic defect, it is possible to distinguish four forms of congenital muscular dystrophy linked to ECM or matrix receptor defects; two associated with mutations in laminin α2 chain and collagen VI (congenital muscular dystrophy type 1A and Ullrich/Bethlem congenital muscular dystrophy, respectively) and two correlated with defective laminin α2 chain receptors (dystroglycanopathies and integrin α7 deficient congenital myopathy, respectively) 6–16. Moreover, putative heterozygous mutations in collagen IV α1 chain were recently described in patients affected by muscle–eye–brain disease and Walker–Warburg syndrome 17. Previously, these two disorders were mainly attributed to alterations in glycosylation of α-dystroglycan 18.

α-Dystroglycan is a member of the dystrophin–glycoprotein complex (DGC), a large complex of membrane-associated proteins that is critical for the integrity of skeletal muscle fibres. The DGC provides a link between the ECM and the cytoskeleton and is believed to protect the muscle fibre from contraction-induced damage 19, 20. The DGC is further composed of β-dystroglycan, dystrophin, the sarcoglycans (α, β, γ and δ), sarcospan, the syntrophins (α1, β1) and α-dystrobrevin 21–24. Notably, mutations in many of its components cause impaired ECM–cell membrane–cytoskeletal interactions, resulting in severe skeletal muscle disease. Particularly, human mutations in dystrophin-, sarcoglycan- and dystroglycan-encoding genes are associated with Duchenne and its allelic variant Becker muscular dystrophy; limb-girdle muscular dystrophy types 2C–F and primary dystroglycanopathy, respectively 15, 25–31. Duchenne muscular dystrophy (DMD), which is inherited in an X-linked recessive pattern, is the most common type of muscular dystrophy, affecting around 1 in 3500 boys. Unlike congenital muscular dystrophies that are manifested at birth, patients with DMD have a childhood onset phenotype, while the age of onset in patients with limb-girdle muscular dystrophy varies from childhood to adulthood.

In summary, it is evident that the ECM–cell membrane–cytoskeleton linkage is crucial for maintenance of skeletal muscle function. In this review, we describe the structure and functions of laminins, collagens, laminin receptors and dystroglycan-associated proteins. We also discuss the molecular pathogenesis of the muscular dystrophies that are caused by primary genetic or secondary functional defects in these molecules as well as how these disorders may be therapeutically targeted.

Laminins

Structure and function

Laminins comprise a group of high molecular weight glycoproteins (400–900 kDa), abundant in the basement membrane 3, 32. Mature laminin proteins are cross- or T-shaped, composed by one α, one β and one γ chain that associate to form a heterotrimeric molecule (Figure 1). Up to now, five α, three β and three γ chains have been characterized, and they combine to form at least 15 different laminin isoforms 33, 34. The laminin heterotrimers are assembled inside the cell, but extracellular maturation by proteolytic cleavage may occur in various chains 33. All laminin chains share a common domain structure, with a number of globular and rod-like domains. In most laminins, but not all, the short arm is formed by the N-terminal extremity (LN domain) of α, β and γ chains, followed by a variable number of globular domains separated by rod-like spacers. The long arm consists of the three chains joined in a coiled-coil domain and the C-terminal extremity of the α chain, which is composed of five homologous globular domains (LG domains 1–5). Laminin polymerization occurs through connections between the LN domains in each chain and is dependent on the presence of the laminin receptors dystroglycan and integrin 35, 36. The laminin LG domains are engaged in binding dystroglycan and integrin and these interactions are crucial for basement membrane formation, cell differentiation and cell survival 37. Finally, laminins contain binding domains for other ECM macromolecules. For instance, a single domain within the laminin γ1 and γ3 chains, respectively, binds nidogens 38–41 and perlecan also binds laminins 42.

Figure 1.

Representative model of a cruciform laminin molecule. The laminins are composed of three polypeptide chains; α (in pink), β (in green) and γ (in blue), which interact forming a coiled-coil domain. The globular domain at the N-terminal of each chain is involved in network formation. The C-terminal end of the α chain is composed of five laminin globular domains that bind dystroglycan and integrins (as well as heparin, sulphatides, perlecan and fibulins)

Congenital muscular dystrophy with laminin α2 chain deficiency

In skeletal muscle, the major laminin isoform is composed by α2, β1 and γ1 chains 43, making up laminin-211 (formerly called merosin). Laminin α2 chain is encoded by the LAMA2 gene and homozygous mutations in this gene are responsible for congenital muscular dystrophy type 1A (MDC1A) (MIM ID #607 855) 14. It represents approximately 50% of all congenital muscular dystrophies 44. This disorder is characterized by severe muscle weakness, hypotonia, joint contractures, dysmyelinating peripheral neuropathy and brain defects 45. Missense, nonsense, splice-site mutations and deletions in the whole LAMA2 gene have been reported 46 and are summarized in the LAMA2 gene database (www.dmd.nl/LAMA2_seqvar.html). Depending on the type of mutation, complete or partial absence of laminin α2 chain can be observed in skeletal muscle. Except for rare cases, patients displaying complete laminin-211 deficiency generally appear to develop a homogenous severe phenotype with an onset of muscle weakness within the 6 first months of life, incapacity to acquire independent ambulation, joint contractures and white matter abnormalities 47–49. In this group, the mutations are localized throughout the whole gene, but four hot-spots (in exons 14, 25, 26 and 27) represent 55% of the mutations 49. In the group of patients with partial deficiency, the majority of mutations are splice-site and frameshift mutations, but missense and nonsense mutations can also occur 50, 51. They are associated with a more heterogeneous course of the disease with either a severe or a milder phenotype 49, 52.

How does the absence/reduction of laminin α2 chain lead to muscle degeneration? This is not fully understood. However, typical features of MDC1A include disrupted basement membranes 53 and increased apoptosis 54, 55. Hence, the laminin α2 chain might mediate survival signals through both integrin α7β1 and dystroglycan 56, 57, and it has been proposed that laminin α2 chain binding to α-dystroglycan strengthens the cell membrane integrity by anchoring the basement membrane to the sarcolemma 58. It was also recently demonstrated that enhanced activity of the proteasome and the autophagic process, respectively, is pathogenic in laminin α2 chain-deficient muscle 59, 60. Still very little is known regarding the laminin α2 chain-induced signal transduction pathways leading to increased apoptosis and enhanced protein degradation. Deciphering such pathways remains essential in order to further clarify the details of laminin α2 chain function.

Collagen IV

Structure and function

Collagen IV is also a ubiquitous protein of the basement membrane. Its primary structure is composed of 90% of Gly–Xaa–Yaa repeats, forming a collagenous domain that allows the assembly of a triple-helix structure. In particular, the glycine residues are crucial for the triple-helix formation and many human mutations affect the glycine residues. Collagen IV also has a short non-helical N-terminal domain (7S) and a highly conserved C-terminal globular domain (NC1) (Figure 2). After synthesis, NC1 domain interactions initiate the formation of three different trimers, encoded from six distinct genes (COL4A1–6); the α1α1α2 trimer (which is ubiquitously expressed) and the α3α4α5 and α5α5α6 trimers (with a more restricted expression pattern). Unlike fibril-forming collagens, collagen IV forms a network. In the extracellular space, the type IV collagen network is formed by complex interactions involving the formation of NC1 hexamers, 7S domain heterotrimers and lateral interactions between the helical domains 61, 62. Similarly to laminins, the collagen IV network interacts with the cell surface, mainly through β1 chain-containing integrins and several non-integrin receptors 63–68. In addition, there are several binding sites for laminin in the collagen IV triple helix, as well as binding sites for heparan sulphate proteoglycans and nidogens 69, 70.

Figure 2.

Schematic diagram of a collagen IV protomer. Collagen IV is composed of three polypeptide chains (different combinations of α1–α6 chains). The central collagen domain is flanked by the N-terminal 7S domain and the C-terminal non-collagenous domain 1 (NC1). Through complex interactions, the heterotrimeric collagen IV forms a network. Integrin binding sites are located in the NC1 domain and in a domain located 100 nm away from the N-terminal

Congenital muscular dystrophy due to collagen IVα1 mutations

Until recently, no mutation in the six genes encoding the collagen IV chains had been associated with a muscle disorder. Instead, mutations found in collagen IV-encoding genes have been demonstrated to be linked to several other diseases, and in particular kidney disorders 60. Also, mutations in COL4A1 are coupled to rare genetic diseases. In humans and mice with mutated COL4A1, a phenotypic variability is often observed but in general the abnormalities include cerebrovascular disease, ocular and renal defects 71–88. Interestingly, it has now been described that the congenital muscular dystrophies muscle–eye–brain disease and Walker–Warburg syndrome could be linked to heterozygous mutations in the collagenous domain of COL4A117. Also, heterozygous Col4a1 mutant mice display ocular dysgenesis, neuronal localization defects and myopathy, all characteristics of muscle–eye–brain disease and Walker–Warburg syndrome. The molecular mechanism triggering the disease is apparently independent of α-dystroglycan glycosylation and instead collagen IV may not be properly secreted, leading to less deposition of collagen IV in the basement membrane and hence ruptured basement membranes 17.

Collagen VI

Structure and function

Collagen VI, a beaded filament-forming collagen, is an abundant interstitial matrix protein that is closely associated with basement membranes in many organs. For example, in skeletal muscle, collagen VI is mainly produced by the fibroblasts 89 but is involved in anchoring the basement membrane to the adjacent connective tissue 90, 91. The complex collagen VI network interacts with numerous proteins and has a crucial role in cell proliferation and survival 92, 93. Collagen VI is composed by three major polypeptidic chains (α1, α2 and α3) encoded by three different genes, COL6A1, COL6A2 and COL6A3, respectively. Also, three additional genes have been described, encoding α4, α5 and α6 chains with a more restricted expression pattern 94. Each chain has an N-terminal and a C-terminal large globular domain, which are connected by a central short triple-helical domain of Gly–Xaa–Yaa amino acid repeat sequences 95. Before secretion into the extracellular space, the three chains associate to form a trimer, which further assembles into disulphide-bonded anti-parallel dimers and then tetramers 96. After secretion into the extracellular space, the tetrameric molecules form the microfibrillar network by an end-to-end association with overlapping N-terminal domains, giving rise to a typical beaded appearance with a periodicity of 105 nm (Figure 3) 97. Collagen VI microfibrils interact with several components of the basement membrane (eg collagen IV and perlecan) 98, 99 and with the cell surface through β1 chain containing integrins 100. Interactions with other constituents of ECM, such as fibronectin 101, biglycan and decorin 102, 103, have also been described.

Figure 3.

Schematic description of collagen VI assembly. Inside the cell, heterotrimeric monomers first assemble into dimers and then tetramers that are secreted. Extracellularly, the tetramers compile into long molecular chains (microfibrils)

Ullrich congenital muscular dystrophy and Bethlem myopathy

Mutations in COL6A1, COL6A2 and COL6A3 genes are associated with Ullrich congenital muscular dystrophy (UCMD, MIM ID No. 254 090) and Bethlem myopathy (MIM ID No. 158 810) 16. UCMD represents the second most frequent form of congenital muscular dystrophy after MDC1A. Initially, UCMD was believed to be recessively inherited while dominant mutations were considered to cause Bethlem myopathy. However, an increasing number of heterozygous and severely affected carriers have been reported. Now these two entities are considered as two extremes of the same clinical spectrum. Classical UCMD is characterized by a severe generalized early-onset muscle weakness, slowly progressive proximal limb contractures, distal hyperlaxity, rigid spine and severe respiratory deficiency 9, 11. UCMD patients have normal IQ and MRI images show a normal development of the brain. The patients affected by classical Bethlem myopathy develop a milder, slowly progressive or static generalized muscle weakness and joints contractures 104. A high level of heterogeneity regarding the phenotype and expression levels of collagen VI is observed in this disease and even within the same family. A growing number of de novo mutations have been reported, suggesting that peculiar regions of the genes are susceptible to neomutations 10, 105. Numerous polymorphisms in the COL6A genes could also contribute to the phenotypic variability. Because of this heterogeneity, several studies aimed at delineating genotype–phenotype correlations have been carried out. By analysing collagen VI expression and secretion in patient fibroblasts and muscle biopsies, a connection between expression and clinical severity was recently demonstrated 10–12. Yet, this correlation was not demonstrated in other reports 106, 107. Nevertheless, the use of quantitative RT–PCR was lately proposed to accelerate the identification of the gene defect, since reduced transcript levels are associated with all types of mutations in COL6A genes 10.

By studies of mice with defective collagen VI secretion (and humans with UCMD or Bethlem myopathy) it has been proposed that the molecular pathogenesis of collagen VI-deficient muscular dystrophy involves latent mitochondrial dysfunction accompanied by ultrastructural alterations of mitochondria and the sarcoplasmic reticulum and spontaneous apoptosis of muscle fibres 10, 91. Moreover, it was recently demonstrated that the accumulation of abnormal mitochondria and sarcoplasmic reticulum is caused by defective autophagy 108.However, it remains to be established how absence of an interstitial matrix protein can lead to muscle cell apoptosis, and therefore the identification of collagen VI muscle cell surface receptors may be an interesting task.

Dystroglycan

Structure and function

Dystroglycan is a widely expressed member of the DGC, involved in connecting the ECM with the intracellular cytoskeleton. Dystroglycan is encoded by the DAG1 gene and undergoes major post-translational modifications, producing the highly glycosylated extracellular peripheral membrane protein α-dystroglycan and the transmembrane protein β-dystroglycan (Figure 4) 8, 109. Intracellularly, the cytoplasmic domain of β-dystroglycan links the C-terminal domain of dystrophin that interacts with the actin cytoskeleton. Extracellularly, β-dystroglycan binds α-dystroglycan, which in turn binds to laminin-211 19, agrin 110 and perlecan 111 in the muscle basement membrane. At the synapse, α-dystroglycan also binds neurexin and pikachurin 112, 113. α-Dystroglycan contains a globular N-terminal and a C-terminal domain separated by a central mucin domain 114 rich in serine and threonine amino acids with attached complex O-linked glycans (eg O-mannosyl glycans). α-Dystroglycan is glycosylated by several putative and known glycosyltransferases and the presence of the oligosaccharides is crucial for the normal function of α-dystroglycan as a membrane receptor 7, 8. In particular, the ability to bind laminins and other ligands requires phosphorylation of O-mannosyl glycans, which is mediated by the glycosyltransferase LARGE 115 that interacts with the N-terminal domain of α-dystroglycan 116. Furthermore, following functional glycosylation of α-dystroglycan, the N-terminal domain is proteolytically processed by the proprotein convertase furin 116, 117.

Figure 4.

Schematic diagram of dystroglycan. The dystroglycan complex is composed of α-dystroglycan, which contains numerous O-linked oligosaccharides and β-dystroglycan. α-Dystroglycan binds laminin in the basement membrane and β-dystroglycan binds to the intracellularly located protein dystrophin

Dystroglycanopathies

Congenital muscular dystrophies linked to abnormalities in the glycosylation of α-dystroglycan are commonly denominated dystroglycanopathies. This group of diseases could either be due to a primary mutation in the DAG1 gene (primary dystroglycanopathy) or mainly due to mutations in genes encoding glycosyltransferases involved in the O-mannosyl-linked glycosylation of α-dystroglycan (secondary dystroglycanopathies). Recently, the first primary dystroglycanopathy was described. A missense mutation was discovered in a patient originally diagnosed with a mild form of limb-girdle muscular dystrophy with severe cognitive impairment. Interestingly, a mouse model harbouring the same mutation recapitulated the phenotype and it was further demonstrated that the missense mutation interferes with LARGE-mediated glycosylation of α-dystroglycan, resulting in a lower binding to laminin 15. Secondary dystroglycanopathies represent the prominent cases of dystroglycanopathies and are autosomal recessive congenital muscular dystrophies caused by mutations in protein O-mannosyltransferase 1 and 2 (POMT1, POMT2), protein O-mannose β1,2-N-acetylglucosaminyltrans- ferase 1 (POMGnT), fukutin-related protein (FKRP), fukutin and LARGE genes 8, 118–125. The resulting diseases are Walker–Warburg syndrome, muscle–eye–brain disease, congenital muscular dystrophy type 1C, Fukuyama-type congenital muscular dystrophy, congenital muscular dystrophy type 1D and limb-girdle muscular dystrophy type 2I, respectively. The clinical hallmarks of these disorders at the severe end of the spectrum include a characteristic combination of brain malformations, as well as eye defects or high myopia, but the spectrum extends from patients with mental retardation with normal brain structure to presentations of pure muscular dystrophy with normal cognition.

For all dystroglycanopathies, the common molecular feature involves abnormal glycosylation of α-dystroglycan, although a strong correlation between reduced glycosylated α-dystroglycan and clinical course may not be evident 126. Hypoglycosylated α-dystroglycan in skeletal muscle and brain displays reduced capacity to bind its ECM ligands and consequently basement membrane deposition is defective 115, 127. However, in order to further understand the pathogenesis of dystroglycanopathies, it will be important to determine the exact glycan composition in normal and diseased skeletal muscle (as well as in other tissues) and to further determine the role of glycans in ligand binding.

Integrins

Structure and function

Integrins are a large family of surface receptors that also bridge the ECM to the intracellular cytoskeleton. Moreover, they are signalling receptors involved in both outside-in and inside-out signalling. Integrins consist of two non-covalently-associated subunits, α and β (Figure 5). Even if not genetically related to each other, the two subunits share a common structure. They display a large modular extracellular domain followed by a hydrophobic transmembrane α-helix and a relatively short C-terminal domain that points to the cytosol. This C-terminal domain lacks actin binding sites and enzymatic activities and consequently integrin signals are transduced through associated proteins, such as talin and integrin-linked kinase. At least 24 different receptors could be formed depending on the combination of the 18α and 8β existing subunits 128, 129. Integrin α7β1 is the main integrin isoform expressed in striated muscle 130. Several alternatively spliced isoforms of α7 and β1 chains have been described and some are differentially expressed during muscle development and regeneration 131. In skeletal muscle, α7β1 integrin interacts mainly with laminin-211 but it can also bind other laminin isoforms 132, 133. Hence, the laminin-211–integrin α7β1 interaction creates an additional important link between the ECM and the cytoskeleton in skeletal muscle 134. A functional redundancy may exist between integrin α7β1 and the DGC 135, 136 and both integrin α7β1 and dystroglycan contribute to force production in the muscle fibre. However, only the latter is involved in anchoring the basement membrane to the sarcolemma 58.

Figure 5.

Representative model of an integrin. The α subunit (blue) consists of a β-propeller (top), a thigh domain, calf-1 and calf-2 domains, a transmembrane domain and a short cytoplasmic domain. The β subunit is composed of a βA domain (top), a hybrid domain, a plexin–semaphorin–integrin domain, four epidermal growth factor-like repeats, a β-tail domain, a transmembrane domain and a short cytoplasmic domain. The β-propeller and the βA domain are engaged in ligand binding

Congenital muscular dystrophy with integrin α7 deficiency

Integrin α7 chain is encoded by the ITGA7 gene. Mutations in this gene have only been described in three patients who presented a recessive form of congenital myopathy characterized by delayed motor milestones and mental retardation for two of the three patients. Laminin α2 chain expression was normal in patient muscle but no integrin α7 subunit was present 13. The consequences of the mutation at the molecular level have not been studied, but analyses of a mouse model that lacks the α7 integrin chain have shown that this receptor is required to maintain myotendinous junctions 134. The β1 integrin accessory molecule integrin-linked kinase also stabilizes the myotendinous junctions and protects skeletal muscle from stress-induced damage 138. However, apart from integrin-linked kinase, not much is known concerning the identity and function of the linker proteins that connect integrin α7β1 to the actin cytoskeleton. In addition, the signalling cascades associated with integrin α7β1 in skeletal muscle remain largely undetermined, although one study demonstrated that integrin α7 deficiency is associated with a disruption of the Ras–MAPK signalling pathway 137. Finally, α7β1 integrin expression is impaired in laminin α2 chain-deficient muscle, suggesting that integrin α7 insufficiency may also contribute to MDC1A pathogenesis 139, 140.

Dystrophin

Structure and function

Dystrophin, composed of four distinct domains, is a 427 kDa cytoskeletal protein that is involved in attaching F-actin to the ECM 141 (Figure 6). The N-terminal extremity contains an actin-binding domain and is followed by a central rod domain, which also contains an actin-binding site 142, a cystein-rich domain and the C-terminal extremity. The rod domain encompasses 24 spectrin-like repeats that provide flexibility to the molecule. Recently, it was demonstrated that neuronal NOS (nNOS, a key enzyme involved in the production of NO) binds directly to two spectrin-like repeats in dystrophin 143. A small subregion (WW domain) composed by 30 amino acids links the central rod domain to the cystein-rich domain and provides a binding site for proline residues and binding to β-dystroglycan occurs within this part of the molecule 144. Furthermore, the cystein-rich domain harbours two EF-hand motifs for intracellular calcium fixation and a calmodulin binding site. Finally, the C-terminal extremity is composed by two coiled-coil domains involved in the oligomerization of the molecule and it also contains binding sites for dystrobrevins and syntrophins 145–148. Multiple smaller isoforms of dystrophin sharing the same C-terminal extremity are transcribed from several intronic promoters within the dystrophin gene 149. Furthermore, utrophin is a 395 kDa dystrophin homologue that is widely distributed and contains the same four domains as dystrophin 21, 150, 151.

Figure 6.

Representation of dystrophin. Dystrophin is composed of an N-terminal domain, a long central rod domain containing spectrin repeats, a cystein-rich domain and finally a C-terminal domain. F-actin binding sites are located in the N-terminal and in the rod domain. β-Dystroglycan binds to the cystein-rich domain, whereas binding sites for syntrophins and α-dystrobrevin are located in the C-terminal extremity

Duchenne and Becker muscular dystrophies

DMD (MIM ID No. 310 200) and Becker muscular dystrophy (BMD, MIM ID No. 300 376) are allelic disorders caused by mutations in the dystrophin gene 152. The phenotypic difference between DMD and BMD patients can mostly be explained by the type of mutations and the resulting consequences at the protein level. Generally, DMD patients exhibit an absence of dystrophin while BMD patients often harbour internally deleted but partially functional dystrophin proteins 153–155.

DMD is a severe form of muscular dystrophy with an onset around 3 years of age. It is characterized by progressive wasting of skeletal muscles, eventually resulting in cardiac and respiratory failure during the twenties age range 155. Patients usually lose independent ambulation in the second decade of life. BMD is similar to DMD in the distribution of muscle wasting and weakness, which is mainly proximal, but the course is more heterogeneous than in DMD. The onset of the disease is generally later during childhood compared to DMD. Some patients have no symptoms until later in life but others can be more affected 155–157. In this group, important genotype–phenotype correlations have been described regarding the onset of muscular, cardiac and cognitive impairment 154.

In DMD patients and in the mdx mouse model of dystrophin deficiency, loss of dystrophin results in a major reduction of the entire DGC complex, and thereby the link between the muscle fibre cytoskeleton and the ECM is destabilized and the sarcolemma becomes fragile. It is hypothesized that the brittle sarcolemma renders the muscle fibre less resistant to mechanical stress, which in turn leads to escalating fibre damage with membrane leakage and altered calcium homeostasis (through membrane tears, due to loss of sarcolemmal stability and/or disrupted calcium channels) and subsequent cell death 19, 20. Yet, dystrophin deficiency may cause pathology by other mechanisms as well. For example, in the absence of dystrophin, nNOS is lost from the sarcolemma. As nNOS produces NO that diffuses to the muscle vasculature and promotes blood vessel dilatation and improves blood flow into skeletal muscle, dystrophin deficiency has been associated with muscle ischaemia 158. Furthermore, absence of nNOS at the sarcolemma also increases exercise-associated fatigue 143, 159.

Sarcoglycans

Structure and function

Sarcoglycans are single-pass transmembrane glycoproteins that belong to the DGC and, through multiple interactions, serve as anchorage for the peripheral DGC components. They display a short intracellular tail, and a large extracellular glycosylated domain rich in cystein residues. This cystein cluster is conserved in all sarcoglycans and seems to be crucial for the assembly in a subcomplex (Figure 7) 160. Six sarcoglycans (α, β, γ, δ, ε and ζ) have been described and α-, β-, γ- and δ-sarcoglycan form a tight subcomplex within skeletal muscle, while ε-, β-, γ- and δ-sarcoglycan form a tight subcomplex within smooth muscle 161, 162. ζ-Sarcoglycan is the most recently identified sarcoglycan family member and is expressed in both striated and smooth muscle 163. The sarcoglycans are cotranslationally translocated in the endoplasmic reticulum and the assembly into the subcomplex occurs during the transport from the Golgi to the plasma membrane 160, 161. At the sarcolemma, the sarcoglycan complex interacts with other DGC members as well as with other proteins, including γ-filamin. In addition, the sarcoglycan complex has been shown to cooperate with integrins in mediating cell adhesion, at least in vitro164. Yet, the precise mechanical and/or signalling functions of the sarcoglycan complex remain to be elucidated.

Figure 7.

Simplified diagram of the sarcoglycans. β-, γ- and δ-sarcoglycan have a C-terminal extracellular domain, but for α-sarcoglycan it is the N-terminal domain. Brown ovals denote N-glycosylation moieties

Sarcoglycanopathies

Sarcoglycanopathies are a group of autosomal recessive muscular dystrophy caused by mutations in one of the four sarcoglycan genes 26–30. Mutations in the genes encoding α-, β-, γ- and δ-sarcoglycan are associated with limb-girdle muscular dystrophy type 2D (LGMD2D, MIM ID No. 608 099), 2E (LGMD2E, MIM ID No. 604 286), 2C (LGMD2C, MIM ID No. 253 700) and 2F (LGMD2F, MIM ID No. 601 287), respectively. Limb-girdle muscular dystrophy is a group of progressive muscle disorders predominantly affecting proximal muscles around the scapular and pelvic girdles. The clinical phenotype of sarcoglycanopathies is very heterogeneous regarding onset, progression and severity 165. Generally, patients present progressive weakness and muscle degeneration, leading to loss of ambulation, respiratory failure and often premature death. Due to their simultaneous synthesis, defects in each sarcoglycan result in a complete destabilization of the entire sarcoglycan complex 160, 161. Furthermore, absence of the sarcoglycan complex at the sarcolemma leads to diminished α-dystroglycan expression 166. Several putative mechanisms underlying the pathogenesis of sarcoglycan-deficient muscular dystrophy have been revealed by studies in animal models. They include loss of membrane integrity, calcium channel dysregulation, mitochondrial dysfunction and vascular abnormalities 166–171.

Development of treatment

There is currently no treatment for muscular dystrophy. Yet, improved disease management (eg non-invasive ventilation to improve respiratory function) has significantly prolonged life for patients with muscular dystrophy. Moreover, corticosteroids have been demonstrated to improve walking and increase quality of life for patients with DMD. However, the side-effects of corticosteroids are quite severe 172. Thus, there is an urgent need to develop therapies for muscular dystrophy.

Several of the ECM and ECM-associated proteins discussed in this review were identified decades ago (eg collagen IV was identified in 1966; laminin in 1979; dystrophin in 1987; dystroglycan in 1992) 3, 25, 61, 109. The structure and function of laminins, collagens, dystroglycan, integrin, dystrophin and sarcoglycans have been extensively studied and mutation analyses have been performed by numerous research laboratories around the world. Moreover, generation and characterization of animal models of laminin α2 chain-deficient muscular dystrophy, Ullrich congenital muscular dystrophy/Bethlem myopathy, dystroglycanopathies, integrin α7-deficient muscular dystrophy, DMD/BMD and sarcoglycanopathies, respectively, have also significantly contributed to understanding the pathogenetic mechanisms of these disorders (Table 1 and references therein). Still, numerous questions as to the molecular and cellular consequences of the protein deficiencies and how they lead to muscular dystrophy persist. Nevertheless, the preceding biochemical and molecular work have laid the foundation for development of molecular therapeutic strategies targeting these incapacitating disorders. These approaches can be divided into the categories of gene replacement therapy, cell therapy and mutation-specific strategies. In Table 1, we present therapies that have had success in animal models and human explants and that have been tested in clinical trials. Considering that the linkage between the ECM and the cytoskeleton is disrupted in laminin α2 chain-deficient muscular dystrophy, Ullrich congenital muscular dystrophy/Bethlem myopathy, dystroglycanopathy, integrin α7-deficient muscular dystrophy, DMD/BMD and sarcoglycanopathy, many approaches have been aimed at restoring this connection. The idea has been to replace the mutated protein with a functional copy or smaller versions thereof or to replace the defective protein with a structurally and functionally similar protein. Indeed, there are several successful examples of gene replacement therapy, using transgenic mouse models as well as virus-mediated gene therapy described in Table 1. Also, the downstream effects of the perturbed interactions, such as increased apoptosis, have been addressed. Indeed, suppression of apoptosis has been beneficial in mouse models of several forms of muscular dystrophy (Table 1). Finally, a number of other possibilities aiming at correcting dystrophin mutations have been explored. Around 15% of DMD patients display mutations inducing a premature stop codon. In these cases, stop codon read-through has been forced by the use of gentamicin and PTC124 in mdx mice followed by clinical trials. Nevertheless, it appears that these drugs still need some improvement before clinical use, in particular since the results of the Phase IIb clinical trial with PTC124 were not encouraging (participants did not show significant increases in the 3 min walking distance test; Table 1 and references therein). However, manipulating the dystrophin gene splicing by antisense oligonucleotides, in order to exclude a mutated exon and restore the open reading frame and allowing the production of a partially deleted but functional dystrophin, has been successful and led to Phase III patient trials (Table 1).

Table 1. Current emerging therapies evaluated in preclinical studies and/or human trials
Mutated proteinEmerging therapyMethodSpeciesReferences
Laminin α2 chainGene over-expressionMini-agrinMouse model178–180
  Laminin α1 chainMouse model181–183
  Laminin α2 chainMouse model184
  Insulin-like growth factor 1Mouse model185
  Integrin α7Mouse model186
  Galgt2Mouse model187
 Apoptosis suppressionOver-expression BCL2Mouse model188
  OmigapilMouse model189
  DoxycyclineMouse model190
  Cyclophilin D deletionMouse model170
 Proteasome inhibitionMG-132Mouse model59
 Autophagy inhibition3-MethyladenineMouse model60
 Stop codon read-throughLaminin α2 chainPatient myotubes191
 Cell-based therapyBone marrow transplantationMouse model192
  CD90+ cell transplantationMouse model193
Collagen VIApoptosis suppressionCyclosporine AMouse/Patient trials93, 194–196
  Debio 025Mouse model197
 Stop codon read-throughsiRNA against SMG-1 or Upf1Patient fibroblasts198
 Autophagy stimulationGene expression, diet and drugsMouse model108
α-DystroglycanGene over-expressionLARGEMouse model116,199;
Integrin α7Protein deliveryLaminin-111Mouse model200
DystrophinGene over-expressionDystrophin (full length/µ/mini)Mouse model201–204
  AAV dystrophinMouse/dog/primate/patients205–212
  Lentivirus dystrophinMouse/dog/primate model213–220
  UtrophinMouse model221–224
  SarcospanMouse model225
  Integrin α7Mouse model226
  Galgt2Mouse model227,228
  ADAM12Mouse model229,230
  nNOSMouse model231,232
  CalpastatinMouse model233
  SERCA1Mouse model234
 Stop codon read-throughGentamicinMouse/Patient trials235–240
  Ataluren (PTC124)Mouse/Patient trials241,242
 Exon-skipping2′-O-MethylphosphorothioatesMouse/Patient trials243–248
  MorpholinosMouse/Patient trials249–260
  AAV (U7snRNA or U1)Mouse/dog model261,262
 Naked DNA deliveryMuscle/vein injectionMouse/dog/primate model263–270
  ElectroporationMouse/dog model271–273
 Restoration reading frameMeganucleasesMouse model274
 Utrophin up-regulationDrug inductionMouse model275–277
 Cell therapyMyogenic/satellite cellsMouse/dog/patient trials278–287
  Muscle-derived stem cellsMouse model288,289
  Bone marrow stem cellsMouse model220,261,290–297
  Side populationMouse model298,299
  PericytesMouse model300,301
  MesoangioblastsMouse/dog model216,302–303
  Placenta-derived cellsMouse model304
  iPS cellsMouse model305
 Myostatin inhibitionAntibodiesMouse/patient trials306,307
  InhibitorsMouse model308,309
  Inactivation myostatin receptorMouse model310–312
  Myostatin pro-peptideMouse model313–315
 Protein deliveryLaminin-111Mouse model316–318
  BiglycanMouse model176
 Akt pathway stimulationValproic acidMouse model319
 Proteasome inhibitionDrug targetingMouse model/patient explants320–323
 Apoptosis suppressionDebio 025Mouse model170
 Inflammation and NOHCT 1026Mouse model324
 ROS scavengingTiron/PP2Mouse model muscle fibers325
 IGF-1Systemic deliveryMouse model326
 L-ArginineSystemic deliveryMouse model327
 Angiotensin II regulationLosartanMouse model328,329
 MMP inhibitorsBatimastatMouse model330
 TNFα reductionEicosapentanoic acidMouse model331
 NF-κB inhibitionPyrrolidine dithiocarbamateMouse model332
Gene over-expressionα-sarcoglycan rAAV/AV vectorMouse/patient trials333–339
  Galgt2Mouse model340
  ε-sarcoglycanMouse model341
 Inflammation and NOIbuprofen and NO co-treatmentMouse model324,342,343
 Cell therapyMesoangioblastsMouse model344
 Proteasome inhibitionVelcadePatient muscle explants345
 Muscle wasting preventionMyostatin inhibitionMouse model309,346
β-SarcoglycanGene over-expressionβ-sarcoglycan rAAV/AV vectorMouse model339,347
γ-SarcoglycanGene over-expressionγ-sarcoglycan rAAV vectorMouse model348
δ-SarcoglycanGene over-expressionδ-sarcoglycan rAAV/AV vectorHamster/mouse model347,349–352
  SERCA1Mouse model353
 Cell therapyTransplantationHamster/mouse model354
 Mitochondrial functionMPTP protectionMouse model173
 Apoptosis suppressionCyclophilin D/Debio 025Mouse model170
 Membrane integrityOmega-3 fatty acid dietHamster model355

Conclusion

An enormous amount of data on the structure and function of laminin α2 chain, collagens type IV and VI, dystroglycan, integrin α7β1, dystrophin and α-, β-, γ- and δ-sarcoglycans have accumulated since the discovery of these proteins. In addition, proteoglycans that are essential components of the ECM may be implicated in the development of muscular dystrophy, as the expression of several proteoglycans is enriched in fibrotic areas of dystrophic muscles 173–176. Also, syndecan cell surface proteoglycans appear to be essential for skeletal muscle regeneration, a feature that is impaired as muscular dystrophy progresses 177. Finally, a great deal is known regarding the pathogenetic mechanisms underlying laminin α2 chain-deficient muscular dystrophy, Ullrich congenital muscular dystrophy/Bethlem myopathy, dystroglycanopathies, integrin α7-deficient muscular dystrophy, DMD/BMD and sarcoglycanopathies, respectively. Yet, no cure is available for these diseases, although vast progress is being made using several different lines of attack. Therefore, continued studies of the ECM–cell membrane–cytoskeleton interactions in skeletal muscle remain fundamental to in order to develop treatment for the muscular dystrophies.

Acknowledgements

We would like to thank Illustratörer.com (http://illustratorer.com/) for help with figures and Kinga Gawlik for valuable comments on the manuscript.

Author contributions

VC and MD wrote the manuscript jointly.

Ancillary