The potential of sarcospan in adhesion complex replacement therapeutics for the treatment of muscular dystrophy

Authors

  • Jamie L. Marshall,

    1. Department of Integrative Biology and Physiology, University of California, Los Angeles, CA, USA
    2. Center for Duchenne Muscular Dystrophy, University of California, Los Angeles, CA, USA
    Search for more papers by this author
  • Yukwah Kwok,

    1. Department of Integrative Biology and Physiology, University of California, Los Angeles, CA, USA
    2. Center for Duchenne Muscular Dystrophy, University of California, Los Angeles, CA, USA
    Search for more papers by this author
  • Brian J. McMorran,

    1. Center for Duchenne Muscular Dystrophy, University of California, Los Angeles, CA, USA
    2. Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
    Search for more papers by this author
  • Linda G. Baum,

    1. Center for Duchenne Muscular Dystrophy, University of California, Los Angeles, CA, USA
    2. Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA, USA
    3. Molecular Biology Institute, University of California, Los Angeles, CA, USA
    Search for more papers by this author
  • Rachelle H. Crosbie-Watson

    Corresponding author
    1. Center for Duchenne Muscular Dystrophy, University of California, Los Angeles, CA, USA
    2. Molecular Biology Institute, University of California, Los Angeles, CA, USA
    3. Department of Neurology, University of California, Los Angeles, CA, USA
    • Department of Integrative Biology and Physiology, University of California, Los Angeles, CA, USA
    Search for more papers by this author

Correspondence

R. H. Crosbie-Watson, Department of Integrative Biology and Physiology, University of California Los Angeles, 610 Charles E. Young Drive East, Terasaki Life Sciences Building Room 1121, Los Angeles, CA 90095, USA

Fax: +1 310 206 3987

Tel: +1 310 794 2103

E-mail: rcrosbie@physci.ucla.edu

Abstract

Three adhesion complexes span the sarcolemma and facilitate critical connections between the extracellular matrix and the actin cytoskeleton: the dystrophin– and utrophin–glycoprotein complexes and α7β1 integrin. Loss of individual protein components results in a loss of the entire protein complex and muscular dystrophy. Muscular dystrophy is a progressive, lethal wasting disease characterized by repetitive cycles of myofiber degeneration and regeneration. Protein-replacement therapy offers a promising approach for the treatment of muscular dystrophy. Recently, we demonstrated that sarcospan facilitates protein–protein interactions amongst the adhesion complexes and is an important potential therapeutic target. Here, we review current protein-replacement strategies, discuss the potential benefits of sarcospan expression, and identify important experiments that must be addressed for sarcospan to move to the clinic.

Abbreviations
MDC1A

merosin-deficient congenital muscular dystrophy 1A

AAV

adeno-associated virus

CMD

congenital muscular dystrophy

DG

dystroglycan

DGC

dystrophin-glycoprotein complex

DMD

Duchenne muscular dystrophy

ECM

extracellular matrix

ER

endoplasmic reticulum

FKRP

fukutin-related protein

GTDC2

glycosyltransferase-like domain containing

Hsp

heat shock protein

ILK

integrin-linked kinase

ISPD

isoprenoid synthase domain containing

LARGE

like-acetylglucosaminyltransferase

MTJ

myotendinous junction

NMJ

neuromuscular junction

nNOS

neuronal nitric oxide synthase

POMGnT1

protein O-linked mannose β-1,2-acetylglucosaminyltransferase

POMT

protein O-mannosyl transferase

SG

sarcoglycan

SSPN

sarcospan

TMEM5

transmembrane protein 5

UGC

utrophin-glycoprotein complex

WFA

Wisteria floribunda agglutinin

Introduction

Duchenne muscular dystrophy (DMD) is a progressive muscle-wasting disease caused by mutations in the dystrophin gene leading to a loss of the dystrophin–glycoprotein complex (DGC) at the sarcolemma [1]. DMD occurs in ~ 1 in 3500 male children, leading to respiratory or cardiac failure in the second decade of life. In 1986, the gene encoding dystrophin was identified, and mutations in the dystrophin gene were determined to be responsible for DMD [2, 3]. Although dystrophin was identified over 25 years ago, there is still no cure for the disease [2]. The N-terminus of dystrophin binds directly to F-actin in the cytoplasm of the myofiber and the C-terminus of dystrophin is associated with a group of proteins including neuronal nitric oxide synthase (nNOS), α-syntrophin, α-dystrobrevin and β-dystroglycan (DG) [4, 5]. Dystrobrevin binds to the intermediate filament protein syncoilin, which provides a connection between desmin and the DGC that is thought to be important for maintaining mechanical strength and structural organization necessary for muscle contractions [6]. nNOS requires binding to both dystrophin and syntrophin for sarcolemmal localization and produces nitric oxide, which stimulates the blood flow necessary to meet the metabolic demands of healthy muscle [7-11]. The DGC functions to stabilize the sarcolemma during muscle contractions by providing a critical connection between the extracellular matrix (ECM) and the intracellular actin cytoskeleton [12]. α-DG, a peripheral membrane protein, provides the connection between the transmembrane β-DG and laminin-211 in the ECM [13, 14]. The central mucin domain of α-DG is heavily glycosylated and the extent of the glycosylation determines the affinity of DGC attachment to the ECM [5, 13-18]. The sarcoglycan–sarcospan (SG–SSPN) subcomplex stabilizes α-DG's association with β-DG at the cell surface [19-21]. Loss of the DGC in DMD renders the sarcolemma susceptible to membrane ruptures, which initiates the cycles of myofiber degeneration and regeneration characteristic of DMD [1, 12, 15, 22].

Many forms of muscular dystrophy result from a loss of muscle cell attachment to its surrounding ECM. In addition to the DGC, two adhesion complexes span the sarcolemma and facilitate this connection: the utrophin–glycoprotein complex (UGC) and α7β1 integrin. The UGC is homologous to the DGC, where utrophin replaces dystrophin [23, 24]. α7β1 integrin is the main heterodimeric integrin expressed in adult skeletal muscle [25-27]. Mutations in genes encoding the protein components of the DGC/UGC, α7β1 integrin and ECM cause various forms of muscular dystrophy, which are classified based on the severity of disease and the muscle groups predominantly affected. Autosomal recessive limb-girdle muscular dystrophies encompass a wide array of genetic disorders and disease onset typically occurs between 10 and 30 years of age. Mutations in α-, β-, γ- and δ-SG and mutations in glycosyltransferases, resulting in the hypoglycosylation of α-DG cause autosomal recessive limb-girdle muscular dystrophies subtypes 2C-2F, 2K and 2N–2O respectively [28-37]. Congenital muscular dystrophies (CMDs) are an autosomal recessive disease characterized by severe muscular dystrophy with onset occurring within the first year of life. Various forms of CMD are also caused by hypoglycosylation of α-DG due to mutations in glycosyltransferases as well as mutations in laminin α2, collagen 6 and α7 integrin [38-42]. Notably, patient mutations in SSPN and β1D integrin have not been identified [43-45].

There are many strategies currently being investigated for the treatment of muscular dystrophy. Several therapeutic approaches focus on adenoviral delivery of gene replacement therapy or exon skipping to produce a shortened, but functional dystrophin protein. This review focuses on therapeutic strategies to replace the DGC with compensatory adhesion complexes, UGC and α7β1 integrin. In particular, this review discusses the potential of SSPN as a therapeutic target to upregulate both the UGC and α7β1 integrin. SSPN is a tetraspanin-like protein that was discovered as a core component of the DGC [46]. Although SSPN has been shown to ameliorate the pathology of the mdx mouse model of DMD [47, 48], many questions remain to be answered before SSPN-related therapies are suitable for patients.

Restoring myofiber function with compensatory adhesion complexes

The adhesion complexes responsible for stabilizing the myofiber membrane have distinct biochemical and functional properties, including sarcolemmal localization and ECM/actin connections in normal muscle. In muscle, the DGC is the most widely expressed adhesion complex, and is found at all regions of the sarcolemma including the neuromuscular and myotendinous junctions (NMJs and MTJs) [3, 49-52]. The UGC, a homologous complex to the DGC, is restricted to NMJ and MTJ regions of the sarcolemma, where utrophin replaces dystrophin [53, 54]. Although utrophin and dystrophin are structurally homologous, they bind to F-actin through distinct sites [55, 56] and only dystrophin contains nNOS-binding sites [57]. It is thought that muscle ischemia occurs in DMD because nNOS is not anchored to the sarcolemmal membrane [57]. The UGC also differs from the DGC in the glycosylation of α-DG, which is a critical determinant of the binding specificity for ECM ligands [47, 58-60]. α7β1 integrins, which are also highly enriched within NMJ and MTJ structures and expressed at low levels at nonjunctional regions of the sarcolemma, are the predominant integrin heterodimers expressed in adult muscle [25-27]. Integrins bind F-actin through a complex of proteins including integrin-linked kinase (ILK), PINCH and α/β-parvin [61-63]. The ILK, PINCH and parvin protein complex has been implicated in facilitating cell signaling through Akt/PKB, GSK3β/β-catenin, JNK and α-Pix/Rac1 pathways [61, 62, 64]. ILK interacts specifically with β1 integrin and deletion of ILK causes muscular dystrophy that resembles the α7-integrin-deficient mouse model [65]. Laminin-211 binds α7β1 integrin through laminin-type G domains (LG) 1–3, whereas binding to α-DG is mediated through LG1–3 and LG4–5 [66-70]. Given that many differences exist between the three adhesion complexes, the extent to which UGC- and integrin-based therapeutics for DMD fully replace DGC function remains an unanswered question.

The concept of the UGC and integrins acting as compensatory adhesion complexes and their potential as therapeutic targets to replace the DGC arose from studies in the mdx mouse model of DMD. Muscle pathology in mdx muscle is less severe than that observed in boys with DMD and the mdx mouse maintains a fairly normal lifespan [71]. It was hypothesized that increased abundance of the UGC and α7β1 integrin at extra-junctional regions of the sarcolemma in mdx muscle may partially compensate for the loss of the DGC, resulting in moderate dystrophic pathology without affecting lifespan, which is in stark contrast to the severe dystrophic pathology and premature lethality observed in the DMD population. To experimentally address this hypothesis, mdx:utrophin-null and mdx:α7 integrin-null double knockout mice were created. The additional loss of utrophin or α7 integrin in mdx mice exacerbated dystrophic symptoms to more closely resemble DMD. Furthermore, lifespan was shortened to 20 and 4 weeks, respectively for utrophin-deficient and α7 integrin-deficient mdx mice [72-75]. Significantly, overexpression of either utrophin or β1D integrin in mdx muscle prevented muscular dystrophy, demonstrating that the UGC and integrins are capable of replacing the DGC in the mouse model of DMD [76, 77]. Mental retardation and brain abnormalities have been detected in ~ 30% of patients with DMD [78]. Utrophin and its smaller isoforms are expressed in the blood vessels in all brain regions, walls of the lateral ventricle, cortex, subiculum, thalamus, brainstem nuclei, superior colliculus and deep cerebellar nuclei [79]. Although utrophin stabilizes the sarcolemma in mdx mice, utrophin was not upregulated in a compensatory fashion in the brain of mdx mice [79]. Arginine butyrate alleviated mdx muscle disease through utrophin upregulation and increased all utrophin isoforms in the brain of mdx mice, but did not rescue cognitive deficits observed in behavioral assays including exploration, emotional reactivity, and spatial and fear memories [79]. These results suggest that utrophin cannot fully compensate for the loss of dystrophin in all tissues. The ability of α7β1 integrin to ameliorate cognitive deficits in mdx mice has not been tested.

Although the UGC and α7β1 integrin do not fully replace the DGC, they remain excellent candidates for protein-replacement therapies. The packaging limit of adeno-associated virus (AAV) capsids prevents the delivery of the most ideal protein for the treatment of DMD, the entire dystrophin gene [80]. Additionally, recent evidence that the expression of AAV-delivered mini-dystrophin is circumvented by an immune response to dystrophin in patient clinical trials [81], strengthens the importance of UGC and α7β1 integrin protein-replacement therapies. A thorough investigation of the biochemical, structural and functional differences between the UGC, α7β1 integrin and the DGC may lead to the future development of combinatorial therapies designed to replace the DGC. The first goal is to determine whether UGC- and α7β1-integrin-based therapies are safe and beneficial to patients with DMD. To date, only three therapies designed to replace the DGC with the UGC or integrins are near clinical trials. Two therapies designed to upregulate utrophin are currently in or near Phase 1 clinical trials: BMN195 (SMT-C1100, Summit plc, Abingdon, Oxfordshire, UK) and biglycan (Trivorsan Pharmaceuticals, Providence, RI, USA). BMN195, a 5-(ethylsulfonyl)-2-(naphthalene-2-yl) benzodoxazole, was identified in a screen for small molecules, which upregulated utrophin mRNA 25% in human myoblasts, increased utrophin protein levels twofold in DMD patient cells, and demonstrated efficacy in treatment of mdx mice [82]. Specifically, BMN195 reduced regeneration, inflammation, serum CK levels and fibrosis, and prevented membrane damage due to eccentric contractions in treated mdx mice [82]. Although BMN195 was shown to be safe at all doses in a Phase 1 clinical trial by BioMarin Pharmaceuticals (Novato, CA, USA), it did not achieve plasma concentrations, even at the highest doses, required to increase utrophin expression (http://investors.bmrn.com/releasedetail.cfm?ReleaseID=703221). Summit plc has since reformulated BMN195 to allow for better absorption and has initiated Phase 1 clinical trials. The second utrophin upregulation therapy, biglycan, is an extracellular protein that is highly expressed, similar to utrophin, in regenerating and developing muscle [83, 84]. Injection of recombinant human biglycan protein improves muscle pathology in the mdx mouse by increasing cell-surface expression of utrophin and γ-SG [85]. Furthermore, the inability of biglycan to ameliorate dystrophic pathology of the utrophin-deficient mdx mouse demonstrates the requirement of utrophin for biglycan therapies [85]. Biglycan is currently being developed for Phase 1 clinical trials by Trivorsan Pharmaceuticals. The only therapeutic designed to upregulate α7 integrin near Phase 1 clinical trials is laminin-111 protein. Laminin-111 is an ECM protein present in cardiac and skeletal muscles during embryonic development that is replaced by laminin-211 in adult muscle [69, 70]. Injection of Engelbreth–Holt–Swarm-derived purified natural mouse laminin-111 protein improves dystrophic pathology in the mdx and laminin α2-deficient dyW mouse models [86-89] through the upregulation of both utrophin and integrins at the sarcolemma. Prothelia (Milford, MA, USA) is currently developing laminin-111 for clinical trials and a Phase 1 trial will occur upon the completion of preclinical trials. Laminin-111 has the potential for therapeutic use in merosin-deficient congenital muscular dystrophy, DMD and LGMD2I [mutations in fukutin-related protein (FKRP)]. The outcome of these clinical trials may illuminate whether utrophin and integrin therapeutics will alleviate the symptoms of DMD (Fig. 1).

Figure 1.

UGC- and α7β1 integrin-mediated replacement therapy for the DGC in DMD. The DGC, UGC and α7β1 integrins function to prevent contraction-induced damage of the sarcolemma by maintaining connections between the actin cytoskeleton and ECM. The DGC is composed of dystrophin, the dystroglycans (α- and β-DG), the sarcoglycans (α-, β-, γ- and δ-SG), sarcospan (SSPN) and the syntrophins (α- and β-subunits). Neuronal nitric oxide synthase (nNOS) requires dystrophin and syntrophin to be anchored to the sarcolemmal membrane, where it is thought to function in preventing functional muscle ischemia. The UGC is homologous to the DGC, where utrophin replaces dystrophin. However, many differences exist between the UGC and DGC, including the glycosylation of α-DG, the domains in which actin binding occurs, and the lack of nNOS-binding sites on utrophin. α7β1 integrin differs from the UGC/DGC in the globular domains by which laminin binds and the presence of adaptor proteins that facilitate actin binding, including ILK and β-parvin. BMN195 and biglycan are two utrophin upregulation therapeutics that are near/in clinical trials and laminin-111 is the only dual (utrophin and integrin) target therapy that is near clinical trials. This review proposes that AAV delivery of SSPN should be considered as an additional dual target therapy.

Although introduction of recombinant laminin-111 protein was initially discovered to increase α7β1 integrin protein levels, it has also been shown to increase the levels of utrophin in mdx muscle [89]. Given that the overexpression of neither utrophin nor α7β1 integrin fully rescues dystrophic symptoms, it would be interesting to determine whether the upregulation of both adhesion complexes would be more beneficial than targeting a single complex alone. Experiments to directly test the UGC and integrin association have not been performed and numerous murine genetic experiments are often complicated by the compensatory actions of other proteins. Migration of the three adhesion complexes in the same sucrose gradient fractions after biochemical isolation of DG-bound proteins using succinylated wheatgerm agglutinin lectin raises the interesting question of whether integrins and the UGC/DGC physically interact at the sarcolemma [90, 91]. If laminin-111 is successful in Phase 1 trials, the potential benefit of targeting both the UGC and α7β1 integrin may be addressed in DMD patients. Similar to laminin-111, Adam12 and SSPN also secondarily upregulate both the UGC and integrins at extrajunctional regions of the sarcolemma when overexpressed in mdx mice [47, 48, 92] (Fig. 1). Adam12 is an active metalloproteinase that is expressed during muscle development and regeneration. The mechanism by which it promotes cell adhesion through the stabilization of the UGC and α7β1 integrin is currently unknown [93-95].

Recent studies have implicated a role for SSPN in facilitating interactions between the UGC and α7β1 integrin. SSPN, a tetraspanin-like protein, functions with the SGs to stabilize the association of α-DG with β-DG in the DGC [19, 20, 46]. Overexpression of threefold levels of SSPN ameliorates mdx muscle by increasing the levels of the UGC, α7β1 integrin and laminin-binding to α-DG [47, 48]. Interestingly, the increase in laminin-binding to α-DG was not observed with 1.5-fold levels of SSPN overexpression in mdx mice, demonstrating that there is a minimum threshold of SSPN expression required for restoring normal ECM binding to DG [47]. To determine whether laminin-binding to α-DG is required for SSPN-mediated amelioration of dystrophic symptoms, threefold SSPN transgenic mice were crossed with LARGEmyd mice. LARGE, like-acetylglucosaminyltransferase, is a glycosyltransferase responsible for elongating O-mannose glycans in the mucin domain of α-DG and mutations in LARGE abolish laminin-binding to α-DG [96]. Overexpression of SSPN does not ameliorate muscular dystrophy in the LARGEmyd mouse model of hypoglycosylated α-DG (merosin-deficient congenital muscular dystrophy 1A) despite increasing levels of the UGC at the sarcolemma, demonstrating that functionally glycosylated α-DG capable of binding to laminin is a requisite for SSPN-mediated amelioration of dystrophic pathology [47]. Recent genetic studies demonstrated that the combined loss of SSPN and α7 integrin results in extensive muscle pathology and decreased specific force production possibly due to decreased protein levels of the DGC and UGC in 4.5-month-old diaphragm muscle [90]. These studies suggest that SSPN and integrins genetically interact and affect protein abundance of the DGC/UGC, but do not establish a direct interaction between these adhesion complexes. Investigation of the dependence of SSPN-mediated amelioration of mdx mice on α7 integrin and utrophin would clarify whether both adhesion complexes are essential for SSPN-based therapeutics. Interestingly, α7 integrin overexpression in mdx:utrophin-null mice causes a 10% reduction in regeneration, a threefold increase in mean survival age and reduced severity of joint contractures [97]. These results suggest that α7 integrin reduces the severity of muscular dystrophy in mdx mice partially independent of utrophin, although the muscle pathology was not significantly reduced. It is possible that utrophin is required for the full rescue effect of α7 integrin in mdx muscle or that higher than threefold levels of α7 integrin are needed to rescue mdx muscle independently of utrophin. The ability of utrophin alone to rescue mdx:α7 integrin-null muscle has not been tested. Thus, additional evidence is needed to establish a direct SSPN-mediated interaction between the UGC and α7β1 integrin at the sarcolemma while the benefit of upregulating both adhesion complexes for the treatment of DMD remains undetermined.

Glycosylation of α-DG and amelioration of muscular dystrophy

Appropriate glycosylation of sarcolemmal proteins is integral for proper muscle function. Over 50% of all known CMDs result from hypoglycosylation of α-DG and therefore are known as dystroglycanopathies [98]. Mutations in 13 genes are known to cause aberrant glycosylation of α-DG. These gene products include: protein O-mannosyl transferase-1 and -2 (POMT1 and POMT2), protein O-linked mannose β-1,2-N-acetylglucosaminyltransferase (POMGnT1), LARGE, fukutin, FKRP, β-1,3-N-acetylgalactosaminyltransferase 2, isoprenoid synthase domain containing (ISPD), dolichyl-phosphate-mannosyltransferase subunits 2 and 3, dolichol kinase, glycosyltransferase-like domain containing (GTDC2) and transmembrane protein 5 (TMEM5) [36, 37, 39, 99-113]. POMT1/2 act collaboratively to add the initiating mannose to serine or threonine residues of the mucin domain of α-DG [114, 115]. Classical mannose glycans are then elongated by the addition of a β-1,2-N-acetylglucosamine (β1,2GlcNAc) residue, whereas laminin-binding mannose glycans are elongated by the addition of a β-1,4-N-acetylglucosamine (β1,4GlcNAC) residue (Fig. 2) [37, 99, 116]. It was recently proposed that β-1,3-N-acetylgalactosaminyltransferase 2 may be the glycosyltransferase responsible for the addition of the GalNAc residue following β1,4GlcNAc addition to laminin-binding mannose glycans, however, this precise activity remains to be proven [108]. On laminin-binding glycan structures, LARGE creates the unique moiety required for binding of α-DG to laminin as well as other ECM components (Fig. 2) [39, 117, 118]. Binding of α-DG to laminin has been shown to be dependent upon the xylosyl and glucuronyl transferase activities of LARGE [119]. Furthermore, knockout of LARGE activity causes a decrease in the molecular weight of α-DG [116]. Precise enzymatic functions of fukutin, FKRP, ISPD, GTDC2 and TMEM5 remain elusive; however, mutations in these genes lead to hypoglycosylation of α-DG and subsequent CMD phenotypes [100, 102-104, 107, 120]. Dolichol kinase and dolichyl-phosphate-mannosyltransferase subunits-2 and -3 are required for synthesis of the glycosylation precursors dolichol-phosphate and dolichol-phosphate-mannose respectively [121-123]; mutations in all three give rise to dystroglycanopathies as both glycosylation precursors are required for synthesis of N-glycans and O-mannosyl glycans [124, 125].

Figure 2.

Glycosylation of the DGC. Known and putative sites of glycosylation of DGC component proteins are depicted. Inset of laminin-binding glycan is provided along with potential site of Galgt2 modification. Colored symbols used to represent glycan structures are in accordance with the guidelines outlined by the Consortium of Functional Glycomics.

The dystroglycanopathies represent a diverse group of muscular dystrophies with broad phenotypic severities and wide ranging causative gene mutations. Walker–Warburg syndrome (WWS) cases result from the largest group of genetic mutations and can be attributed to mutations in POMT1, POMT2, LARGE, fukutin, FKRP, ISPD, β-1,3-N-acetylglucosaminyltransferase, β-1,3-N-acetylgalactosaminyltransferase 2, GTDC2 and TMEM5 [103, 107, 126, 127]. Patients presenting with muscle–eye–brain disease represent a similarly heterogeneous population of phenotypic severities and genetic mutations. Muscle–eye–brain disease has been reported to result from mutations in POMGnT1, fukutin and FKRP [37, 127, 128]. LGMD types 2K, N, M, I and O result from mutations in POMT1, POMT2, fukutin, FKRP and POMGnT1, respectively [28, 127-129]. CMD types 1C and 1D result from mutations in FKRP and LARGE, respectively [39, 130]. Furthermore, an insertion of a retrotransposon in the fukutin gene is known to cause Fukuyama congenital muscular dystrophy [102]. The genetic basis for the various dystroglycanopathies is varied and the resulting phenotypes are overlapping, which reinforces the need for a thorough understanding of the connection between specific genetic mutations and resulting biochemical dysfunctions that underlie or cause muscular dystrophy.

Although DMD results from mutations in dystrophin, it is noteworthy that significant changes in glycosylation have been reported in the mouse model for DMD. The lectin Wisteria floribunda agglutinin (WFA) recognizes terminal β1,4-GalNAc residues of α-DG in skeletal muscle [60, 131] and binding is normally restricted to the NMJ and MTJ in wild-type myofibers [59]. However, WFA binds to the extrasynaptic sarcolemma in addition to the NMJ on mdx muscle cryosections [47, 58]. This redistribution of WFA staining to the extrasynaptic sarcolemma is concomitant with the redistribution of utrophin and the associated UGC in mdx mouse muscle.

Overexpression of both full-length and truncated utrophin rescues the dystrophic phenotype in mdx mice [77, 132-135]. Because WFA preferentially binds to α-DG in the UGC, increasing GalNAc modification of α-DG has been one approach to increase extrasynaptic utilization of utrophin as a potential therapy for DMD. The overexpression of terminal β-GalNAc glycosyltransferase galgt2 in wild-type mice has been shown to increase levels of UGC proteins, laminin-binding and reactivity of α-DG to WFA [59, 60]. Similar results were obtained from galgt2 overexpression in mdx mice, where rescue of laminin-binding and the dystrophic phenotype were additionally noted [59, 60]. Interestingly, transgenic overexpression of SSPN also increases the WFA reactivity of α-DG in mdx and LARGEmyd muscle, whereas loss of SSPN reduces UGC levels and the reactivity of α-DG with WFA [47]. Importantly, no mitigation of dystrophic pathology or rescue of laminin-binding is observed in LARGEmyd mice overexpressing SSPN [47]. These studies are significant because they demonstrate that laminin-binding is required for SSPN-mediated amelioration of dystrophic pathology and that the glycans detected by increasing WFA reactivity are distinct from the structure created by LARGE activity. It has been proposed that SSPN overexpression mediates changes in glycosylation via increased galgt2 activity [47]; however, this mechanism requires further validation through the creation of galgt2-deficient mdx SSPN transgenic mice. Interestingly, overexpression of galgt2 has also been demonstrated to be effective in ameliorating the dystrophic phenotype in mouse models of CMD 1A (dyW) and α-SG-deficient LGMD 2D [136, 137]. It is noteworthy, however, that the overexpression of a neuronal homolog galgt1, which adds terminal β-GalNAc residues distinctly to ganglioside glycolipids, caused muscle pathology including decreased myofiber diameter and increased central nucleation in wild-type mice [138]. These stark differences resulting from the overexpression of either galgt2 or galgt1 demonstrate the tissue and acceptor substrate specificity of glycosylation and reinforce the importance of thoroughly evaluating the effects of manipulating glycosylation.

Although genetic overexpression of a glycosyltransferase provides one potential therapeutic approach for manipulating glycosylation of sarcolemmal glycoproteins, pharmacologic approaches also provide potential therapeutics. Using high-throughput screening, the small molecule lobeline was identified as a pharmacological treatment that altered C2C12 glycosylation in vitro [139]. Compounds from the Prestwick library of ~ 1200 FDA-approved small molecules were added to myoblasts in differentiation media and changes in glycosylation were measured after two days of treatment [139]. Specifically, binding of terminal β-GalNAc modifications as detected by WFA binding increased in C2C12 cells as well as isolated wild-type and mdx myoblasts following treatment with lobeline during differentiation [139]. Lobeline treatment increased the abundance of UGC proteins and laminin-binding in wild-type and mdx primary cell cultures [139]; these results are similar to changes resulting from galgt2 overexpression. It is noteworthy that the increase in WFA binding following lobeline treatment was dependent upon complex N-glycans and not O-mannose glycans such as those required for laminin binding [139]. Furthermore, deoxymannojirimycin inhibition of complex N-glycans necessary for WFA binding caused a decrease in laminin binding as detected by laminin overlay [139]. Although lobeline was first FDA approved for smoking cessation, in vitro studies have shown that it may potentially act as a protein folding chaperone or nicotinic antagonist through the dopamine or vesicular monoamine transporter [140-143]. In the C2C12 studies, the molecular mechanism by which lobeline increased WFA binding in vitro was not determined. Although the ability of lobeline to alter muscle glycosylation in vivo remains to be demonstrated, these studies represent strong proof of the principle of pharmacologic manipulation of sarcolemmal protein glycosylation as a potential therapeutic.

Protein aggregation and amelioration in muscular dystrophies

The mdx mouse contains a point mutation in exon 23 [144], resulting in a premature termination codon that is predicted to result in a 115-kDa protein [145]. It has previously been shown that dystrophin mRNA levels are decreased in the skeletal muscle, cardiac muscle and brains of three mdx mouse strains [146]. However, dystrophin mRNA levels vary widely, depending on the causative mutation in human DMD and Becker muscular dystrophy patients [147]. It was previously thought that the 115 kDa dystrophin protein is rapidly degraded after synthesis in mdx muscle, but studies suggest that truncated dystrophin may be expressed at the sarcolemma [148-150] or retained in the ER/Golgi [47].

A 70–80-kDa dystrophin reactive to a C-terminal antibody was found in a group of Japanese Spitz dogs displaying progressive Duchenne-like muscular dystrophy [149]. These dogs displayed exercise intolerance, an abnormal gait and pain upon handling of muscles beginning at 10–12 weeks of age. The symptoms worsened over time, suggesting that the 70–80-kDa dystrophin was unable to restore sarcolemmal stability. In a separate study, immunohistochemical assays revealed expression of truncated dystrophin in myotubes isolated from three 12-week-old aborted fetuses at risk for DMD, suggesting that mutant dystrophin is synthesized in human DMD patients [148]. Recently, the truncated form of dystrophin was detected within the ER/Golgi compartments of mdx mice [47], suggesting that the 115-kDa dystrophin may be retained in the ER/Golgi compartments.

The mutation in mdx mice leads to the loss of the β-DG-binding domain, leaving only the actin-binding domains in the dystrophin fragment [145]. The N-terminus of dystrophin binds directly to F-actin in the cytoplasm of the myofiber, whereas the C-terminus of dystrophin contains cysteine-rich domains that bind the C-terminus of β-DG [4, 151]. Recent studies utilizing truncated dystrophin have provided more insight into which segments of the protein are required for sarcolemmal stability. One study utilized rAAV6 microdystrophin to evaluate the role of the actin-binding domains in sarcolemmal stabilization [152]. Four months after injection into the tibialis anterior, muscles treated with micro-dystrophins lacking actin-binding domains displayed decreased specific force and a decreased ability to protect against contraction-induced injury [152], indicating the importance of an intact actin-binding domain in maintaining sarcolemmal stability.

Other studies have analyzed different isoforms of dystrophin. The dystrophin isoform Dp116 is expressed in Schwann cells within the peripheral nervous system [153]. This isoform lacks the actin-binding domains although it retains the complete dystroglycan-binding domain [153]. Although expression of Dp116 in mdx:utrophin-null mice increased muscle mass, life span and maximal force, the specific force and histopathology did not improve [153]. It is hypothesized that Dp116, or the dystroglycan-binding domain, attempts to stabilize the DGC at the sarcolemma through other interactions with the cytoskeleton, although specifics have yet to be elucidated [153]. The expression of a non-muscle isoform of dystrophin, Dp71, which lacks N-terminal actin-binding domains and spectrin-like repeats, restores the DGC in muscle, suggesting that the β-DG-binding domain is required for assembly of the DGC [154]. Thus, it would be interesting to examine other mice with C-terminal mutations in dystrophin. Such experiments would assess the ability of truncated dystrophin to be incorporated and expressed with the DGC at the sarcolemma and restore functionality of dystrophic muscle. This provides an exciting opportunity to examine ways to express fragmented dystrophin at the sarcolemma to improve dystrophic pathology. Other studies involving protein retention within cellular compartments in different muscular dystrophies may provide insights into the therapeutic translation of these findings.

Studies in Fukuyama congenital muscular dystrophy, which is caused by mutations in fukutin [102], have provided insight into the importance of protein processing within the ER/Golgi. The exact role of fukutin remains unclear, but mutations in fukutin lead to abnormal glycosylation of α-DG [155], a reduction in laminin-binding activity [155], and the mislocalization and retention of fukutin in the ER [156]. A recent study analyzed 13 missense fukutin constructs in C2C12 cells and discovered that four were mislocalized to the ER. To understand whether mutant fukutin still leads to the production of functionally glycosylated α-DG, site-directed mutagenesis was used to generate mutants for transfection into fukutin-null mouse embryonic stem cells [156]. Expression of the four missense fukutin mutants restored α-DG reactivity against the IIH6C4 antibody and laminin-binding to α-DG, suggesting that α-DG is functionally glycosylated with the expression of mutant fukutin [156]. Using treatments to improve protein folding, including curcumin, a molecule isolated from the spice turmeric and low temperature culturing conditions, ER retention of fukutin was corrected in the four mutants shown to be mislocalized in the ER. In order to elucidate the reasons for aberrant fukutin trafficking, brefeldin A was used to examine anterograde transport and nocodazole was used to examine retrograde transport [156]. After incubation with brefeldin A, the fukutin mutants remain accumulated in the ER [156]. The results suggest that although mutant fukutin proteins are processed in the ER, they are unable to be transported to the Golgi through the anterograde pathway, providing valuable insight into how the protein-trafficking process is affected upon the mislocalization of a mutant protein. Thus, it is possible that truncated dystrophin may be retained within the ER/Golgi compartments and is not transported to the cell surface in mdx muscle. It will be important for future studies to utilize a similar strategy to elucidate the effects of 115-kDa dystrophin on retrograde and anterograde transport in mdx cells. Additionally, treatment with pharmacological agents, such as curcumin or exposure to low temperature, may potentially alleviate accumulation of truncated dystrophin in the ER/Golgi compartments and improve the transport of DGC and UGC components to the sarcolemmal membrane.

Curcumin was found to target the NF-κB pathway [157] and has been used to treat dystrophic muscle. The activation of NF-κB, which is involved in the modulation of immune responses and regulation of myogenesis, is increased in DMD patient muscle [158]. Intraperitoneal administration of curcumin inhibits NF-κB activation and reduces dystrophic pathology in mdx mice [159]. However, a different study reports that curcumin treatment is unable to inhibit the NF-κB pathway and does not improve specific force in the diaphragm muscle of mdx mice [160]. Future studies examining the effect of curcumin on the accumulation of truncated dystrophin in the ER/Golgi will be important in determining what role, if any, it plays in pathogenesis.

Protein mislocalization has also been implicated in LGMDs 2C–F pathologies caused by mutations in the SG genes [161]. The severity of pathology in LGMDs 2C–F and cellular fate of the affected SG varies according to the mutation [161]. According to data compiled from the Leiden University mutation database, the most common mutations are α-SG (p.R77C), β-SG (p.S114F) and γ-SG (p.C283Y), which cause a mild to severe phenotype [161]. These data led to an examination of the noted point mutations and their effect on intracellular fate and expression of the SGs at the cell surface. Following treatment with kifunensine, an α-mannosidase I inhibitor that prevents ER-associated degradation [162], the expression of sarcoglycans with mild mutations thought to cause the least amount of structural modification was restored at the cell surface. Future studies manipulating ER quality control in wild-type and mdx cells might provide more information on how changes in the ER can affect DGC and UGC expression at the cell surface.

Heat shock proteins and proteasome inhibition in the treatment of muscular dystrophy

Heat shock proteins (Hsp) are induced to combat cellular stress when organisms are under environmental strains such as heat, or during disease and infection [163]. Analysis of skeletal muscle from young DMD patients revealed an induction of Hsp72 and Hsp65 in hypercontracted fibers and Hsp90 in regenerating muscle [164]. Hsp72 is also increased at the mRNA level following electrical stimulation for tetanic contractions in isolated single skeletal muscle fibers from Xenopus laevis [165] and after high-intensity exercise in normal Wistar rats [166], suggesting a role for Hsp72 in protection against muscle stress. Treatment of mdx and utrophin-deficient mdx mice with BGP-15, a pharmacologic inducer of Hsp72, showed that an increase in Hsp72 expression improved SERCA function and dystrophic pathology, decreased kyphosis and ultimately extended lifespan [167]. Targeting Hsp72 to improve protein folding and quality control during cellular stress and disease progression provides a possible avenue to improve muscle function while prolonging the lifespan of patients with muscular dystrophy.

Ubiquitin is an important component in proteasome regulation and can direct proteins towards degradation [168]. Ubiquitin was also found to be elevated in hypercontracted, regenerating and necrotic myofibers in the skeletal muscle of young DMD patients [164]. The increase in ubiquitin protein may suggest misregulation of the ubiquitin–proteasome pathway as well as an increase in protein degradation in DMD patients, which may contribute to the pathogenesis of DMD. Recent studies attempting to ameliorate dystrophic pathology using proteasome inhibitors have produced mixed results. Velcade and MLN273, two FDA-approved proteasome inhibitors, were injected into the gastrocnemius muscle of mdx mice [169]. A truncated 97 kDa dystrophin product, as well as an increase in α-DG, β-DG and α-SG, were detected using immunoblotting and immunohistochemistry 24 h post treatment [169]. Similarly, another proteasome inhibitor, MG-132, rescued the membrane localization of dystrophin, α-DG, β-DG and α-SG [170]. However, physiological functionality was not examined and experiments such as grip strength and force production were not performed. This information is critical for the pursuit of such pharmacological agents as treatments for DMD (Fig. 3A).

Figure 3.

Effects of SSPN overexpression in mdx mice on the cell surface protein expression and protein processing and possible outcomes of truncated dystrophin within the cell. (A) Wild-type mice express full-length dystrophin, which is 425 kDa. The mdx mutation leads to a premature stop codon that results in a truncated dystrophin of 100 kDa. Dystrophin is depicted in purple, the proteasome is blue and orange, and ubiquitin is green. Truncated dystrophin can be retained in the ER/Golgi. Alternatively, truncated dystrophin is ubiquitinated and sent to the proteasomal degradation pathway. (B) The overexpression of SSPN in mdx muscle leads to molecular events resulting in the restoration of laminin binding and rescue of mdx pathology. SSPN activates Akt, which leads to an increase in utrophin and integrins. Galgt2, one enzyme responsible for GalNAc modification of α-DG, is also increased in isolated ER/Golgi membranes. SSPN also improves utrophin-DG transportation to the sarcolemma while simultaneously restoring laminin-binding and membrane stability. SSPN's effect on the trafficking of truncated dystrophin and proteasomal degradation is still unknown. DGs (red), SGs (yellow), SSPN (green), integrins (blue) and Akt (orange) are shown. Utrophin (Utr) is depicted in gray.

Although some drugs such as Velcade and MLN273 have shown promise for the attenuation of dystrophic pathology in mdx mice, other studies have described conflicting results. In a separate study, long-term administration of a MG-132 failed to restore dystrophin expression in mdx muscle and ultimately increased susceptibility to contraction-induced damage [171]. The effects of proteasome inhibitors on grip strength, force production and DGC protein expression need to be analyzed more rigorously to better determine their efficacy in improving muscle function and dystrophic pathology.

A possible function for sarcospan as a chaperone protein

New data suggest a role for SSPN in protein trafficking to the sarcolemma. A recent study isolated ER/Golgi membranes from mdx muscle and immunoblots revealed an increase in α-DG and utrophin compared with wild-type muscle [47]. This suggests a possible compensatory mechanism is at play, whereby the system attempts to utilize utrophin upon the loss of dystrophin (Fig. 3B). Interestingly, SSPN transgenic mdx mice revealed decreased levels of WFA-reactive α-DG and utrophin in ER/Golgi preparations, whereas immunofluorescence assays showed an increased abundance of these proteins at the sarcolemma [47]. SSPN overexpression increased WFA binding 1.8-fold per α-DG molecule as detected by WFA overlay [47]. Increased extrasynaptic binding of WFA may be the result of general chaperone trafficking of glycoprotein complexes by SSPN. However, the increase in WFA-reactive glycosylation per α-DG protein demonstrates a role for SSPN in specific changes in glycosylation of α-DG. This modification in α-DG glycosylation may result from downstream transcriptional effects of increased Akt signaling [47] or could additionally be a more direct result of chaperone activities of SSPN in trafficking. SSPN might act directly as a chaperone to stabilize α-DG during ER/Golgi trafficking and addition of glycans resulting in increased glycan modification per α-DG molecule. Importantly, overexpression of SSPN drives changes in glycosylation of α-DG, which aid in the amelioration of mdx pathology [47]. These studies highlight a promising role for SSPN as a therapeutic and it would be interesting to examine SSPN overexpression in dystroglycanopathy mouse models where direct manipulation of glycosylation ameliorates dystrophic pathology [136, 137]. SSPN may possess chaperone-like functions, and the overexpression of SSPN may improve overall protein folding and quality control as well as transport to the cell surface.

How to achieve forced overexpression of SSPN

Membrane proteins are synthesized by ribosomes on the ER and disulfide bonds are synthesized and rearranged in the ER lumen [172]. Because SSPN is an integral membrane protein with disulfide bonds requiring processing through the ER, it most likely cannot be administered systemically. The small size of SSPN makes it an excellent gene to be delivered through AAV. AAV delivery of α-SG in α-SG-deficient patients and γ-SG in γ-SG-deficient patients resulted in no adverse events, demonstrating that intramuscular AAV delivery is likely to be safe in adult patients [173, 174]. The only patients that did not express SG following AAV delivery had pre-existing immunity to the AAV serotype used, demonstrating the need for prescreening of AAV serotypes [173, 174]. However, there are many challenges to be overcome before systemic delivery of AAV is feasible. Systemic delivery of therapeutics will be required for amelioration of fatal dystrophic symptoms in the diaphragm and heart muscles. Surprisingly, a T-cell-mediated immune reaction against dystrophin prevented the expression of mini-dystrophin following AAV delivery in DMD patients [81]. The precise reason for the varied T-cell immune responses (against mini-dystrophin and against self revertant dystrophin) in the six trial patients is unknown. These results warrant caution in future AAV mini-/microdystrophin and exon skipping trials. Importantly, immune responses should not be a problem in secondary approaches to upregulate utrophin or α7β1 integrin for protein-replacement therapeutics. A drug screen for compounds that upregulate SSPN offers an additional approach to AAV-mediated SSPN therapy. A similar approach led to the discovery of BMN195 for utrophin upregulation and is currently in Phase 1 clinical trials.

Unanswered questions for SSPN-based therapeutics

Overexpression of SSPN under the human skeletal actin promoter results in a 60% reduction in regeneration in the mdx mouse model of DMD by replacing the DGC with the UGC and α7β1 integrin [47, 48]. As with most therapeutic targets, many questions remain to be answered for SSPN-based therapeutics. For SSPN to become a viable therapeutic for the treatment of DMD, the questions below need to be addressed.

Does SSPN delivery with AAV prevent dystrophic pathology in the mdx mouse model?

This is an important question that needs to be answered, because it will determine the delivery system used for SSPN-based therapeutics. It would also be useful to further determine the feasibility of AAV delivery of SSPN in the golden retriever model of DMD, because the dog model provides a more realistic clinical model for systemic delivery.

Does SSPN eliminate/reduce dystrophic pathology in the heart and diaphragm?

The human skeletal actin promoter is not highly expressed in the diaphragm or heart muscles. Because death occurs from respiratory and/or cardiac failure in DMD patients, an ideal therapeutic target should prevent dystrophic pathology in the heart and diaphragm. This question can be addressed with systemic delivery of AAV6-SSPN in the mdx mouse model. AAV6 has been shown to infect the heart and the diaphragm [175, 176].

Can SSPN reverse or prevent dystrophic pathology after the onset of dystrophy?

The human skeletal actin promoter is turned on early in muscle development. SSPN is likely preventing the onset of dystrophic pathology rather than reversing pathology. Boys with DMD are often diagnosed with the disease well after the onset of dystrophic pathology, so effective treatments should reverse or halt pathology already in progress. To determine whether SSPN is effective after the onset of disease, a murine inducible transgenic system could be engineered to turn SSPN expression on later in life. Intramuscular or systemic injection of AAV-SSPN after the onset of pathology in mdx mice will also address this question.

Is widespread expression of SSPN safe?

The promoter used in murine studies of SSPN overexpression in the mdx mouse model restricted SSPN expression to striated skeletal muscle [177]. An effective treatment in patients with DMD would require systemic delivery of SSPN in order to target the heart and diaphragm. It is important to determine whether long-term systemic delivery of SSPN is safe. This question can also be addressed with systemic delivery of AAV-SSPN in mdx mice. Because AAV infection of particular tissues depends on the AAV serotype, it would be important to test SSPN delivery with the serotypes approved for use in patients.

Does SSPN upregulate the UGC and integrins in human myoblasts?

Many recent treatment strategies have been tested in DMD myoblasts, including BMN195 and dantrolene [82, 178]. The ability to demonstrate that a therapy acts through similar targets in human cells offers a nice proof-of-principle for the approach being tested.

Which patient populations will benefit from SSPN-based therapeutics?

It is important when developing therapies for muscular dystrophy to determine how many different muscle diseases will benefit from the treatment. This will allow for the design and recruitment of appropriate patient populations for clinical trials. SSPN is effective in ameliorating pathology in the mdx model of DMD [47, 48]. SSPN did not reduce dystrophic pathology in the LARGEmyd model of hypoglycosylation of α-DG (merosin-deficient congenital muscular dystrophy 1A) [47]. It would be interesting to determine if SSPN can ameliorate various limb-girdle and congenital muscular dystrophies.

Conclusions

In the mdx model of DMD, the SSPN transgene ameliorates dystrophic pathology by increasing the UGC and α7β1 integrin extrasynaptic sarcolemma, the synapse-specific GalNAc glycosylation of α-DG, and the transport of utrophin and α-DG from the ER/Golgi to the cell surface [47, 48, 179]. These studies raise important questions about the mechanisms by which SSPN overexpression accomplishes these observed effects. Biochemical evidence that SSPN is a component of both the DGC and UGC complexes and genetic analysis of mice lacking both SSPN and α7 integrin have led to the proposal that SSPN stabilizes the UGC and integrins through direct interactions at the sarcolemma [46, 47, 90, 179]. More experiments are needed to address the requirement of utrophin and α7 integrin for SSPN-mediated amelioration of mdx muscle, as well as the possibility that the UGC and α7β1 integrin function synergistically at the sarcolemma. In addition, SSPN may act either directly or indirectly as a chaperone protein to facilitate the efficient assembly and export of the UGC to the cell surface [47]; however, more studies are needed to elucidate the role of SSPN in the ER/Golgi. Overexpression of constitutively active Akt signaling has been shown to be beneficial in mdx muscle by increasing the UGC and integrins at the sarcolemma, reducing membrane damage and improving the force-generating capacity of muscle [180, 181]. SSPN-mediated amelioration of mdx muscle results in a similar increase in active Akt signaling and downstream muscle growth pathways [47]. Many therapeutic approaches for DMD concentrate on a single target. Although experimental strategies for the treatment of DMD have been developed for over 25 years, steroids remain the only approved drugs to slow the progression of the disease. Thus, the best treatment strategy remains an open question. SSPN is unique because it incorporates several beneficial therapeutic targets into a single protein that is small and easily packaged in AAV delivery systems. SSPN is also ubiquitously expressed in other tissues throughout the body so systemic delivery and immune response should not be an issue [46]. We are currently experimentally addressing the unanswered questions for SSPN therapeutics.

Acknowledgements

We thank A. W. Kwok, J. Lee and J. Oh for critically reading this manuscript. This work was supported by grants from the Genetic Mechanisms Pre-doctoral Training Fellowship USPHS National Research Service Award GM07104, the Edith Hyde Fellowship, the Eureka Pre-doctoral Training Fellowship, and the Ruth L. Kirschstein National Service Award T32AR059033 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases to JLM; Muscular Dystrophy Association RG 135449 and NIH P30 AR057230 to LGB and NIH/NIAMS (R01 AR048179) to RC-W.

Ancillary