Biglycan targets dystrobrevin, syntrophin and nNOS to the muscle cell membrane

Abstract

The dystrophin associated protein complex (DAPC) provides a linkage between the cytoskeleton and the extracellular matrix and is also a scaffold for a host of signaling molecules. The constituents of the DAPC must be targeted to the sarcolemma in order to properly function. Biglycan is an extracellular matrix molecule that associates with the DAPC. Here, we show that biglycan null mice exhibit a mild dystrophic phenotype and display a selective reduction in the localization of α-dystrobrevin -1 and -2, α- and β1- syntrophin, and nNOS at the sarcolemma. Purified biglycan induces nNOS redistribution to the plasma membrane in cultured muscle cells. Biglycan protein injected into muscle becomes stably associated with the sarcolemma and extracellular matrix for at least two weeks. This injected biglycan restores the sarcolemmal expression of α-dystrobrevin-1 and -2, and β1- and β2-syntrophin in biglycan null mice. We conclude that biglycan is important for the maintenance of muscle cell integrity and plays a direct role in regulating the expression and sarcolemmal localization of the intracellular signaling proteins dystrobrevin-1 and -2, α- and β1-syntrophin and nNOS.

Introduction

The dystrophin-associated protein complex (DAPC) provides a critical link between the basal lamina and the cytoskeleton in skeletal muscle. It is comprised of three subcomplexes: dystrophin-dystroglycans (α-, β-); sarcoglycans (α-, β-, γ- δ-)-sarcospan; and the cytosolic dystrobrevin-syntrophin-nNOS (1, 2). Dystrophin binds actin and the cytoplasmic tail of β-dystroglycan. α-Dystroglycan in turn connects β-dystroglycan with the extracellular matrix (ECM) molecules laminin, agrin, perlecan and biglycan (2, 3). The maintenance of a linkage between the ECM and the cytoskeleton is important for maintaining skeletal muscle function and stability. At the cellular level, mutations in genes that encode for proteins of the DAPC cause destabilization of the DAPC, loss of membrane integrity and degeneration of muscle fibers (4, 5).

We have previously demonstrated that the extracellular matrix molecule biglycan is expressed on the skeletal muscle surface and binds to the DAPC protein α-dystroglycan (3). Biglycan is a member of the small leucine-rich repeat proteoglycan (SLRP) family. It contains within its 38 kD polypeptide core sequence 10 leucine-rich repeats flanked by two cysteine-rich domains, and two glycosaminoglycan (GAG) attachment sites at its N-terminus. Biglycan deficient mice have been shown to display decreased growth rates as well as reduced bone mass (first detected at 6 months of age) and premature osteoarthritis (6, 7). Muscle expresses both a “proteoglycan” (PG) form and a “core” polypeptide that lacks the GAG chains. Biglycan has been shown to play a role in skeletal muscle development and regeneration. It becomes upregulated during muscle fiber regeneration and mice deficient in biglycan display delayed regeneration following injury (8). In addition, mice overexpressing biglycan demonstrate aberrant eyelid muscle development (9). However, it remains unknown how extracellular biglycan affects processes that rely upon signaling inside the muscle cell and how biglycan influences the composition of the DAPC complex.

Here we show that adult biglycan null mice are mildly dystrophic, displaying increased muscle fiber degeneration and regeneration, abnormal fiber size distribution and muscle membrane weakness. Examination of DAPC elements revealed a selective reduction in the localization of α-dystrobrevin-1 and -2, α- and β1-syntrophin, and nNOS to the sarcolemma. Exogenous biglycan induced the relocalization of nNOS from the cytosol to the plasma membrane in cultured biglycan null muscle cells. Finally, intramuscular injection of recombinant biglycan restored the expression of α-dystrobrevin-1, α-dystrobrevin-2, and β1-syntrophin to the sarcolemma of biglycan null mice. Thus, biglycan is important for the maintenance of muscle cell integrity and plays a direct role in regulating the expression and sarcolemmal localization of the intracellular DAPC signaling proteins dystrobrevin-1 and -2, α-and β1-syntrophin, and nNOS.

Methods

Expression and purification of recombinant biglycan

Recombinant biglycan was produced using a stable transfected 293-EBNA cell line. The cell line was created by transferring a biglycan polyhistidine fusion construct, originally created for expression using a vaccinia virus expression system into the pCEP4 (Invitrogen Corporation, Carlsbad, CA) expression vector. The resulting plasmid was transfected into 293-EBNA cells and stable expressing cells selected by culturing in the presence of hygromycin B. Following amplification, cells were seeded into a Celligen Plus bioreactor (New Brunswick Scientific, Edison, NJ) containing 100 g Fibra-Cel disks and grown to saturation. Protein production was initiated by replacing the culture media with serum free DMEM. Conditioned media was collected every 48 h with fresh media being added back to the bioreactor. Following concentration of the conditioned media using a Pellicon 2 Tangential Flow system (Millipore Corporation, Bedford, MA), recombinant biglycan was purified using nickel chelating chromatography and elution with a gradient of 0–250 mM imidazole in 20 mM Tris-HCl, 500 mM NaCl, 0.2% CHAPS, pH 8.0. Biglycan polypeptide with “proteoglycan” and with “core” GAG side chains was subsequently separated from proteoglycan following anion exchange chromatography on Q-Sepharose and elution with a linear gradient of 0.15–2 M NaCl in PBS, 0.2% CHAPS. Note that both forms are N-glycosylated (10).

Antibodies

The following primary antibodies were used: anti-biglycan (clones 2A5 and 4C6) (Creely et al., submitted), pan-specific anti-dystrobrevin (BD Transduction Laboratories) (San Jose, CA), isoform-specific syntrophin antibodies SYN259 (α-syntrophin), SYN29 (β2-syntrophin) and SYN37 (β1-syntrophin) (11), dystrobrevin antibodies 638 (α-dystrobrevin–1) and DB2 (α-dystrobrevin–2) (12), anti-nNOS (Immunostar, Hudson, WI), anti-6-10 (dystrophin) (a generous gift from Timothy J. Byers and Louis M. Kunkel (Harvard Medical School and Children's Hospital, Boston, MA)), anti-merosin (Alexis Biochemicals, San Diego, CA) and NCL-dys2 (dystrophin), NCL-g-sarc (γ-sarcoglycan), NCL-b-sarc (β-sarcoglycan), and NCL-a-sarc (α-sarcoglycan) (NovoCastra, Newcastle upon Tyne, UK). The following secondary antibodies were used: Alexa 488-goat anti-rabbit IgG, Alexa 488-goat anti-mouse IgG (Molecular Probes, Eugene, OR), Cy3-goat anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA) horseradish peroxidase-goat anti-rabbit IgG, and horseradish peroxidase-goat anti-mouse IgG (Amersham).

Biglycan antibody production

A monoclonal antibody capable of detecting biglycan immunohistochemically was produced. Adult biglycan null mice were injected with vaccinia virus/T7 bacteriophage system-expressed core and proteoglycan forms of biglycan (13) in Titermax Gold adjuvant (Sigma, St. Louis, MO). Six booster injections of biglycan core and proteoglycan were performed. Splenocytes were fused by Green Mountain Antibodies (Burlington, VT) and specific antibody production was tested via immunohistochemistry on quadriceps sections from wild type and biglycan null mice. Hybridomas secreting antibodies that immunostained wild type but not biglycan-deficient muscle sections were considered positive and were subcloned. Further characterization showed that these antibodies recognized purified biglycan in solid phase binding assays. Monoclonal antibodies 2A5 and 4C6 were used in the present studies.

Skeletal muscle membrane preparations

Quadriceps from 5-week-old mice were homogenized in dissection buffer containing 0.3 M sucrose, 35 mM Tris (pH 7.4), 10 mM EDTA, 10mM EGTA, a protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN), and sodium azide. Samples were sonicated on ice for 3 × 10 s and centrifuged at 7000 × g at 4°C for 20 min. Solid KCl was added to a final concentration of 0.6 M and the samples were centrifuged 7000 × g for 20 min. The membranes were then collected by centrifugation for 140,000 × g for 60 min at 4°C. Protein concentrations were determined by the BCA (Pierce, Rockford, IL) protein concentration assay.

Western blot analysis

Proteins were transferred from SDS-PAGE gels to nitrocellulose membranes in transfer buffer at 100 V for 1 h at 4°C. Membranes were blocked in TBS, 0.1% Tween-20, 4.0 % normal goat serum, 5.0 % milk for 1 h at room temperature and incubated with primary antibody in blocking buffer overnight at 4°C. Membranes were washed and incubated for 1 h at room temperature with goat anti-rabbit or anti-mouse IgG conjugated to horseradish peroxidase (Amersham Biosciences, Piscataway, NJ) diluted 1:1000 in blocking buffer. After washing, bound antibody was detected using enhanced chemiluminescence according to the manufacturer's protocol (Amersham).

Biglycan null mice

The biglycan null mice used were generated by inserting the PGK-neo cassette from the pPNT vector into exon 2 at the Tth111I site as described (6). The original C57BL/6 background mice were rederived and then backcrossed to C3H for several generations to create congenic animals (14) and were obtained from Jackson Labs. Five or eight week-old wild type and biglycan-deficient mice on C3H background were utilized in this set of experiments. Similar results were observed at both ages. (In preliminary experiments we observed that wild type and biglycan null littermates on the C57BL/6 background gave results indistinguishable from those seen with the C3H background animals.) The animals were housed at room temperature with a 24 h night-day cycle and fed with pellets and water ad libitum. All protocols were conducted under strict accordance and with the formal approval of Brown University's Institutional Animal Care and Use Committee.

Quantitative PCR analysis

RNA extraction from p35 quadriceps muscle was performed using the Trizol method (Invitrogen, Carlsbad, CA, Carlsbad, CA). Purified RNA was converted to cDNA using the Superscript III First-Strand Synthesis System Kit (Invitrogen, Carlsbad, CA). qPCR reactions were performed using the SYBR-Green method (Invitrogen, Carlsbad, CA) on the ABI PRISM 7300 real-time thermocycler. Primers were designed using DS Gene primer design software (Accelrys, San Diego, CA). ATP synthase was used for normalization. Data analysis was performed using the standard curve method (15). All experiments were performed in triplicate using three pairs of wild type and biglycan null mice.

The primers used were: ATPSase forward: 5′-TGG GAA AAT CGG ACT CTT TG-3′; ATPSase reverse: 5′-AGT AAC CAC CAT GGG CTT TG-3′; α-syntrophin forward, 5′-CTG AAG AGG ATC GTT CAT C-3′; α-syntrophin reverse: 5′-TCA GGC TGG TCT CTG AG -3′; β1-syntrophin forward, 5′-GAA CAG AGA GGC GAC TTG CC-3′; β1-syntrophin reverse: 5′-CAT GTG ACT CCT TAA ACC TG -3′; β2-syntrophin forward: 5′-GCA ACA ACA AAG AAG C-3′; β2-syntrophin reverse: 5′CCT GTT GTG GTC CAG CAG TG -3′; α-dystrobrevin-1 forward, 5′-TGA AGA ACA CAG GCT GAT CG-3′; α-dystrobrevin-1 reverse: 5′-GCA TCG ATG GTG AAG GAG AT-3′; α-dystrobrevin-2 forward, 5′-CCT CTT GTC TTG TTC CCT GTG-3′; α-dystrobrevin-2 reverse: 5′-CAG CGC CCT AAA AAC AGA AA-3′; nNOS forward, 5′-GGG CAA ACA GTC TCC TAC CA -3′; nNOS reverse: 5′-AGG GTG TCA GTG AGG ACC AC -3′.

Evans blue dye staining and serum creatine kinase

A 10 mg/ml solution of Evans Blue dye in PBS was injected into the tail vein of wild type, biglycan-null and mdx mice. Quadriceps muscles were collected 3-6 h later and cryosectioned - Evans Blue dye-positive myofibers were observed under the fluorescent microscope using a rhodamine filter set. One hundred myofibers per section for each of three mice per genotype were counted. For serum creatine kinase measurements, blood was collected from anesthetized animals by retro-orbital bleed and clotted and enzyme levels were measured using a commercial kit (Sigma).

Histology and Immunohistochemistry

Quadriceps femoris muscles were isolated and flash-frozen in liquid nitrogen-cooled isopentane. In all experiments 10 μm sections from age-matched, congenic bgn null and wild type muscle were mounted on the same slide. Sections were post-fixed with 4% freshly prepared paraformaldehyde for 5 minutes at room temperature, blocked in Vector blocking reagent for 1 h at room temperature and incubated with the indicated primary antibody overnight at 4° C followed by secondary antibody for 1 h at room temperature. The Vector M.O.M. Basic kit was used according to the manufacturer's protocol. Sections were mounted in Permafluor (Thermo Electron Corporation, Pittsburgh, PA) and analyzed using confocal laser scanning microscopy (Leica TCS SP2 Acousto-Optical Beam Splitter (AOBS)). No immunostaining was observed if non-immune IgG or an irrelevant antibody were substituted for the first layer; controls also showed that the secondary antibodies were species-specific. All comparison images presented here were captured from sections of bgn null and wild type muscle that had been mounted on the same slide. Images were acquired using Leica LCS acquisition software and imported into Adobe Photoshop. Sections were also observed using a Nikon (Melville, NY) Eclipse E800 microscope and images acquired with Scanalytics (Fairfax, VA) IP Lab Spectrum software.

Cell Culture

The biglycan null immortalized muscle cell line was generated using established protocols (16). Biglycan null cells were grown to ∼75% confluence on gelatin-coated Permanox chamberslides (Nalge Nunc, Naperville, IL) at 33°C and 10% carbon dioxide in growth medium containing Dulbecco's modified Eagle's medium (DMEM) high glucose, 20% fetal bovine serum, 2% chick embryo extract, 1% L-glutamine, 1% penicillin/streptomycin and 1U INF-γ. Cells were differentiated for 4-9 days in medium containing DMEM high glucose, 5% horse serum, 1% L-glutamine, and 1% penicillin/streptomycin.

Immunocytochemistry

Differentiated biglycan null cells were incubated for 4 h at 33°C with 0.7 nM biglycan core. Acetylcholine receptors were labeled with rhodamine-α-bungarotoxin for 30 minutes at 33°C. Further incubations were completed at room temperature and cells were rinsed in MEM-H after each step. Cells were fixed in 1% paraformaldehyde and permeabilized with 0.05% saponin. To detect nNOS, cells were incubated in primary antibody for 1 h then incubated in goat-anti-mouse IgG Alexa 488 for 1 h. Cells were fixed with methanol for 5 min at −20°C and mounted in Vectashield Hard Set with DAPI (Vector laboratories, Burlinghame, CA). Images were acquired on a Nikon Eclipse E800 microscope using a 60× objective. Based on the captured images, myotube segments were scored blind to condition on a 0-4 scale that was based upon the extent of nNOS on the myotube surface as delineated by AChR localization. A segment with no surface-localized nNOS received a score of 0 and a segment with the greatest extent received a 4.

Intramuscular biglycan injections

Fifty μg of purified recombinant biglycan core in 50 μL 20 mM Tris, 0.5M NaCl, 0.2% CHAPS, pH 8.0/1.0% India ink was injected intramuscularly into the right quadriceps femoris of 2 week old biglycan null mice using a 29 1/2G insulin syringe (BD Biosciences, San Jose, CA). Fifty μL of 20 mM Tris, 0.5M NaCl, 0.2% CHAPS, pH 8.0/1.0% India ink was injected into the left quadriceps femoris of each animal to serve as an internal control. The injected quadriceps were isolated 4, 7, 11, and 14 days following biglycan injection. The injections on two-week-old animals were performed using three bgn null litters. The increase in α-dystrobrevin-1, α-dystrobrevin-2, β1-syntrophin, and β2-syntrophin at 11 d following injection was seen in 4 out of 4 animals and the increase at 14 d following injection was observed in 4 out of 4 animals. No increase was observed in any of the vehicle-injected muscles.

For immunohistochemical labeling of injected muscle tissue, quadriceps femoris muscles were isolated and sectioned as described above. In all experiments, sections from vehicle- and biglycan- injected muscle from the same animal were mounted on the same slide. Sections were double-immunolabeled with anti-biglycan monoclonal antibodies 2A5 or 4C6 and a rabbit polyclonal antibody against the specified dystrobrevin or syntrophin isoform or nNOS (see Antibodies). The secondary antibodies used for detection were Alexa Fluor 488-goat anti-mouse IgG and Cy3-goat anti-rabbit IgG both diluted 1:2000.

Results

Biglycan null mice display a dystrophic phenotype

Mutations in any one of several DAPC components cause a wide range of muscular dystrophies (5, 17). Since biglycan binds to three DAPC components, α-dystroglycan, α-sarcoglycan, and γ-sarcoglycan (3) (M.S.R., H.H., and J.R.F., submitted), we tested whether biglycan null mice display a dystrophic phenotype (Fig. 1). First, we performed the Evans blue dye uptake assay to determine whether muscle membrane integrity is compromised in the biglycan null mice. Bgn null, wild type littermates and dystrophin-null mdx mice were intravenously injected with Evans blue dye and the extent of uptake in skeletal muscle fibers was assessed after 6 h. Wild type muscle cells exhibited little detectable dye permeability (<0.5% of cells), while >90% of mdx myofibers showed dye uptake (18) (Fig. 1 A). Bgn null fibers displayed an intermediate pattern. One population of fibers was completely permeable and showed a uniform dye distribution (7.4% ± 1.3; n=3), while others showed a perimembranous distribution. This localization at the cell periphery was not observed in control muscles from wild type littermates. Measurement of serum creatine kinase levels also indicated modest sarcolemmal permeability in the bgn null mice. We observed levels of 240 ± 43; 870 ± 104 and 6435 ± 385 for wild type, bgn null and mdx mice, respectively; U/L, ± SE, n=3). Thus, a subpopulation of myofibers in bgn null mice shows evidence of leaky membranes and loss of sarcolemmal integrity.

Figure 1. Dystrophic phenotype in bgn null mice.

(A) Evans Blue Dye (EBD) uptake. Wild type and bgn null (bgn -/o) mice (8 wk old) were intravenously injected with EBD and dye uptake into quadriceps femoris muscles was assessed 6 h later by fluorescence microscopy. A subset of bgn null muscle fibers exhibited complete dye permeation (7.4% ± 1.3; n=3), while other fibers displayed a perimembranous uptake. Essentially no uptake (<0.5% of fibers) was observed in muscle from wild type animals, while >90% of dystrophin-null mdx fibers showed complete permeation. (B) Hematoxylin and eosin stained quadriceps femoris sections from wild type and bgn -/o mice. Bgn null mice exhibit groups of muscle fibers with centrally nucleated nuclei (arrows) characteristic of cells that have regenerated in the adult animal. Central nuclei are rarely detected in the muscle fibers of wild type animals. Scale bar = 10 μm. (C) Distributions of cell diameters of wild type and bgn null muscle fibers. Muscle fiber diameters from wild type and bgn null quadriceps femoris, gastrocnemius, and diaphragm were measured and their distributions plotted. White bars represent wild type (WT) cell diameters and grey bars represent bgn -/o cell diameters. The number of fibers counted per muscle type is indicated. Bgn null quadriceps and gastrocnemius fibers are smaller than wild type, while bgn null diaphragm fibers display a wider size distribution.

We next examined the histology of the muscles from bgn null mice. As shown in Fig. 1 B, the majority of fibers in bgn null muscle were similar to those in wild type as judged by hematoxylin and eosin staining. However, a fraction of the myofibers in bgn null mice displayed centrally localized nuclei, a hallmark of regenerated fibers (5.08% ± 0.58 and 9.65% ± 1.34 in quadriceps and diaphragm, respectively). In contrast <0.5% of fibers in wild type animals showed central nuclei. Similar percentages of centrally nucleated fibers were observed in quadriceps muscles from 4-, 12-, and 24-week-old bgn knockouts (data not shown). We did not observe extensive mononuclear cell infiltration or fibrosis in the bgn null muscle.

Abnormal muscle is often characterized by an increased variability in fiber size. We therefore measured fiber diameters from quadriceps femoris, gastrocnemius, and diaphragm (Fig. 1 C). In the quadriceps and gastrocnemius, bgn null fibers were smaller than wild type fibers. The mean quadriceps fiber diameter in wild type was 36.89 ± 3.88 μm vs. 30.56 ± 2.73 μm in the biglycan knockout. The mean gastrocnemius fiber diameter in wild type was 38.14 ± 4.69 μm vs. 33.67 ± 3.52 μm in the biglycan knockout. On the other hand, the diaphragm of bgn null mice displayed a wider fiber size distribution (Fig. 1 C). The mean fiber size diameter in wild type diaphragm was 23.20 ± 3.08 μm vs. 27.89 μm ± 5.21μm in bgn null diaphragm. Taken together, these results indicate that bgn null mice display a mild dystrophic phenotype.

Selective reduction of α-dystrobrevin, syntrophin, and nNOS expression at the sarcolemma of bgn null mice

We next compared the expression of individual DAPC proteins in adult wild type and bgn null mice. In all cases we compared bgn null mice to either normal littermates or to congenic, age-matched wild type animals. The results were consistent between both sets of animals. We first examined the expression levels of dystrophin, the major transmembrane complexes, and the most prominent basal lamina ligand of the DAPC, laminin α2. As shown in Fig. 2, immunofluorescent staining of muscle sections revealed that the sarcolemmal expression of dystrophin and α-, β-, and γ- sarcoglycan is unchanged in bgn null animals. The expression of α- and β- dystroglycan and the α2 chain of laminin were also indistinguishable between wild type and bgn null muscle (Fig. 2).

Figure 2. Expression of dystrophin, sarcoglycans, dystroglycans and laminin α2 in wild type and bgn null skeletal muscle.

A comparison of the expression of DAPC proteins in skeletal muscle of wild type and bgn null (bgn -/o) animals. Quadriceps femoris sections from age-matched wild type and bgn null animals were mounted on the same slide and stained using antibodies against dystrophin (DYS), α-, β-, and γ- sarcoglycan (α-, β-, or γ- SG), α- and β- dystroglycan (α- or β- DG), and the laminin α2 chain (LN-2). The sarcolemmal expression levels of each of this subset of DAPC proteins are unchanged in bgn null muscle. Results are representative of 6 pairs of 5-week-old wild type and bgn null muscles. Scale bar = 20 μm.

We next used immunohistochemistry to examine the expression of the intracellular DAPC proteins involved in signaling and scaffolding: the dystrobrevins, syntrophins and nNOS. In each experiment we also immunostained for dystrophin, which is unchanged and thus serves as an internal, positive control (Fig. 2 and 3 A, B). As shown in Fig. 3 A, α-dystrobrevin-1 and -2 are expressed strongly at the muscle sarcolemma of wild type muscle. Moreover, these proteins are localized in a continuous, uninterrupted distribution at the sarcolemma. In contrast, in bgn null muscle, α-dystrobrevin-1 and -2 are expressed at a lower level at the sarcolemma and are distributed in an irregular, punctuate pattern.

Figure 3. Expression of dystrobrevins, nNOS, and syntrophins in wild type and bgn null skeletal muscle sections and microsomal membrane fractions.

Quadriceps femoris sections from 5-week-old wild type and bgn null (bgn -/o) animals were stained using the indicated antibodies. Results are representative of 6 pairs of 5-week-old wild type and bgn null muscles. (A) As shown in Fig. 2, the levels of dystrophin are indistinguishable in the wild type and bgn null muscle. In contrast, there is a selective reduction of α-dystrobrevin-1 (α-DB-1), α-dystrobrevin-2 (α-DB-2), α-syntrophin (α-Syn), β1-syntrophin (β1-Syn) and nNOS at the sarcolemma of bgn null muscle. (B) Higher magnification images of wild type and bgn null quadriceps femoris sections stained using the indicated antibodies. The intracellular levels of α-syntrophin are increased in the biglycan null as compared to the wild type muscle. The intracellular levels of dystrophin (DYS) remain unchanged. Scale bar = 20 μm. (C) A comparison of expression of the dystrobrevins, syntrophins and nNOS in KCl-washed microsomal membranes from wild type and bgn null muscle. Ten μg of membrane proteins from 5-week-old quadriceps femoris muscles from wild type and bgn null animals were separated by SDS-PAGE and immunoblotted with antibodies against α-dystrobrevin-1 (α-DB-1), α-dystrobrevin-2 (α-DB-2), α-syntrophin (α-Syn), β1-syntrophin (β1-Syn), and β2-syntrophin (β2-Syn). The expression levels of α-DB-1 and α-DB-2 are decreased in bgn null microsomes, while the levels of α-Syn, β1-Syn, and β2-Syn are increased. (D) Proteins (10 μg) from non-KCl-washed microsomal membranes isolated from 5 week old quadriceps femoris muscles from wild type and bgn null animals were separated via SDS-PAGE and immunoblotted using antibodies against nNOS and β-actin (loading control). Note the selective decrease in nNOS expression in bgn null muscle membranes. Similar results were obtained in membrane fractions obtained from 3 pairs of wild type and bgn null muscles.

We also observed that α-syntrophin and β1-syntrophin are reduced at the sarcolemma in bgn deficient animals (Fig. 3 A). The reduction in β1-syntrophin at the sarcolemma was pronounced, while that of α-syntrophin was subtler. On the other hand, the expression of β2-syntrophin at the sarcolemma of bgn null muscle did not differ from wild type. Interestingly, the intracellular level of all three syntrophins was increased in bgn null as compared to wild type muscle (Fig. 3 A, B).

Finally, nNOS has been shown to co-immunoprecipitate with α-syntrophin and to be lost at the sarcolemma of α-syntrophin-null muscle (19, 20). Therefore, we asked whether the skeletal muscle expression of nNOS is altered in bgn knockouts. As shown in Fig. 3 A, sarcolemmal nNOS levels were lower in bgn null muscle as compared to wild type. Taken together, these results demonstrate that there is a selective reduction in the expression of the dystrobrevin-syntrophin-nNOS complex in the sarcolemma of bgn null mice.

Virtually all components of the DAPC are expressed at the neuromuscular junction, with many of them being selectively enriched at this site. In addition, individuals with denervation disorders show specific changes in the expression patterns of the syntrophins and dystrobrevins, but not the rest of the DAPC (21). We thus asked whether the expression of the α-dystrobrevins, syntrophins and nNOS was altered at the synapse in bgn null muscle. The expression of each of these proteins was compared to that of acetylcholine receptors, whose levels are unaffected by the loss of biglycan. We find that the expression levels and distribution of α-dystrobrevin-1, α-dystrobrevin-2, α-syntrophin, β1-syntrophin, β2-syntrophin and nNOS remain unchanged in bgn null neuromuscular junctions (data not shown). Thus, this DAPC subcomplex is reduced at the sarcolemma but maintained at the neuromuscular junction.

Dysregulation of α-dystrobrevin-1 and -2, α-, β1-, and β2- syntrophin, and nNOS content in membrane fractions from bgn null muscle

The immunohistochemical data described above indicate that the levels of α-dystrobrevin- 1 and -2 are reduced at the plasma membrane of biglycan-deficient muscle fibers. To determine whether the biochemical profile of these proteins is also changed, we assessed their levels in KCl-washed heavy microsomal membrane fractions, which are a mixture of plasma and intracellular membranes (22). To compare the levels of membrane-associated dystrobrevin, we isolated fractions from wild type and bgn null quadriceps muscle and probed them using a pan-specific anti-dystrobrevin antibody. Fig. 3 C shows that the levels of both α-dystrobrevin -1 and -2 are reduced in these membrane fractions.

We next examined the expression of α-, β1-, and β2- syntrophins in KCl-washed membrane fractions. Immunoblotting with isoform-specific antibodies revealed that the levels of all three syntrophins were increased in membrane fractions from bgn knockouts compared to wild type controls (Fig. 3 C). This increased expression is likely to reflect the contribution of syntrophins associated with intracellular membranes; indeed, it is in agreement with our immunohistochemical results showing elevated intracellular levels of all syntrophins in the bgn null muscle (Fig. 3 A, B). These findings suggest that biglycan serves to appropriately target syntrophins to the muscle cell surface and/or to regulate the intracellular trafficking of these molecules.

We also compared the expression of nNOS in wild type and bgn null quadriceps. Since KCl treatment strips nNOS from muscle membrane fractions (20), we isolated heavy microsomal membrane fractions with no KCl wash. As shown in Fig. 3 D, the expression levels of nNOS at the membranes of bgn null animals are decreased compared to wild type, in agreement with the immunohistochemical results.

Selective changes in β1-syntrophin and nNOS mRNA expression levels in skeletal muscle of bgn null mice

We next asked whether the changes in the expression of syntrophin-dystrobrevin-nNOS complex members observed in bgn null was reflected in changes transcript levels (Fig. 4). Quantitative real time PCR (qRT-PCR) analysis showed that β1-syntrophin transcripts are significantly upregulated in the bgn knockout as compared to the wild type (1.88 ± 0.27 vs. 1.00, p < .03; Student's t-test), while nNOS message levels are downregulated (0.18 ± 0.27 vs. 1.00, p < .03; Student's t-test). On the other hand, the levels of both α-dystrobrevins and of α- and β2-syntrophin messages are equivalent in wild type and bgn null muscle.

Figure 4. Relative mRNA expression of dystrobrevins, syntrophins, and nNOS in wild type and bgn null skeletal muscle.

Expression levels of the mRNAs encoding components of the dystrobrevin-syntrophin-nNOS complex in 5-week-old bgn knockout quadriceps femoris muscle are shown relative to wild type. Levels were determined by quantitative real-time PCR and are normalized to the expression levels of ATP synthase. * = The changes in β1-syntrophin and nNOS transcript levels are statistically significant (P<0.03; Student's t-test). n = 3.

Biglycan induces the redistribution of nNOS to the plasma membrane

We next developed a cell culture system to further examine the role of biglycan in targeting DAPC components to the cell surface. We cultured biglycan-deficient myotubes and incubated them with either purified recombinant biglycan core polypeptide or with vehicle alone for 4 h. Living myotubes were labeled with rhodamine-α-bungarotoxin to visualize AChRs and to demarcate the plasma membrane. We then fixed and permeabilized the cells and immunostained for the dystrobrevins, syntrophins and nNOS. As shown in Fig. 4 A, nNOS is distributed throughout the cytoplasm in untreated cells, with little labeling observed in the region of the plasma membrane. However, biglycan treatment increases the levels of nNOS localized subadjacent to the myotube surface. We quantified this redistribution of nNOS by assigning each cell a score (0-4, 0 being no nNOS at the myotube surface, 4 being the highest change in nNOS at the surface; scored blind to treatment conditions) representing the extent to which surface nNOS was present. In the absence of biglycan, the mean score was 2.0 ± 0.11, whereas in the presence of biglycan, the mean score was 2.94 ± 0.12. The histogram in Fig. 4 B displays the number of untreated or biglycan-treated cells receiving each score. These data demonstrate that the treatment of bgn null myotubes with biglycan core increases the amount of nNOS on the myotube plasma membrane (p < .01; Kolmogorov-Smirnov test).

In parallel experiments, we observed that α-dystrobrevin -1 and -2 and β2-syntrophin staining were widely expressed within the cell, but little cell surface-proximal expression was detected. On the other hand, α-syntrophin and β1-syntrophin were expressed throughout the cell including in a surface-proximal disposition. However, biglycan treatment did not alter the localization of these proteins.

The injection of purified biglycan protein restores the expression of α-dystrobrevins and β-syntrophins to the sarcolemma of bgn null muscle fibers in vivo

The results described above show that a subcomplex of the DAPC is reduced at the sarcolemma of adult bgn null mice. However, since biglycan is absent throughout development, it is not possible to determine when biglycan is playing a role in the localization of these components. Moreover, since biglycan is expressed in other locales, including bone and tendon (23, 24), it is possible that the muscle phenotype is secondary to a defect in another tissue. Finally, the results with the bgn null mice do not allow a determination of which biglycan form(s) (core or proteoglycan) is required for directing the correct localization of the DAPC subcomplex.

To address these questions, we asked whether intramuscular injection of purified biglycan core or proteoglycan could restore the expression of the syntrophin-dystrobrevin-nNOS DAPC subcomplex to the sarcolemma of bgn null mice. Purified recombinant biglycan core polypeptide or proteoglycan (50 μg) was injected into the right quadriceps femoris muscles of two-week-old bgn null animals. Buffer alone was injected into the left quadriceps to provide intra-animal comparisons. The injection site was visualized by the inclusion of India ink in each solution. Quadriceps were dissected 4, 7, 11, and 14 days post-injection, sectioned, and immunostained. As shown in Fig. 5, A-D (upper left), intramuscularly injected biglycan core polypeptide becomes stably associated with the perimysium and sarcolemma, persisting for up to 14 days following a single injection. Intramuscularly injected biglycan proteoglycan also becomes stably associated with the perimysium (data not shown). No biglycan immunoreactivity was observed on the vehicle-injected side at any time examined (Fig. 5, A-D, lower left).

Figure 5. Biglycan treatment regulates the localization of nNOS in cultured myotubes.

(A) Differentiated myotubes deficient in biglycan expression were treated with medium alone or with 0.7nM recombinant biglycan core polypeptide for four hours. Living myotubes were then labeled with rh-α-bungarotoxin to label surface AChRs (red). The myotubes were then fixed, permeabilized and stained for nNOS (green). Arrows delineate the myotube plasma membrane. In the absence of biglycan, nNOS expression is largely cytosolic. Upon biglycan treatment, nNOS becomes more highly localized to the plasma membrane. (B) Quantification of biglycan-induced nNOS relocalization. Myotubes were scored blind to treatment conditions on a scale of 0 to 4 (0 being no nNOS cell surface expression, 1 being low nNOS cell surface expression and 4 being high nNOS cell surface expression). The number of untreated and bgn null cells receiving each score was graphed. The number of biglycan-treated myotubes displaying cell surface localization of nNOS was significantly higher than the number of untreated myotubes with nNOS on their surface (p < .01, Kolmogorov-Smirnov test).

Intramuscularly injected biglycan core had a striking effect on the expression of the syntrophins and dystrobrevins in the bgn null muscle. By 11 days post-injection, we observed increased α-dystrobrevin -1 and -2 and β1- and β2- syntrophin expression at the sarcolemma. Moreover, the increased expression of these intracellular DAPC proteins showed a tight spatial correlation with the exogenous biglycan (Fig. 5, A-D; compare upper left and upper right). No upregulation was observed in the vehicle-injected muscle (Fig. 5, A-D; compare upper and lower right) or in the biglycan proteoglycan-injected muscle (data not shown). The increase in α-dystrobrevin-1 and -2 and β1- and β2- syntrophin persisted at 14 days post-injection. We did not observe a change in α-syntrophin following biglycan injection. Additionally, nNOS expression was increased at the injection sites of both biglycan-treated and vehicle-treated muscle, suggesting that nNOS expression increases following mechanical injury. Indeed, there is evidence that mechanical injury induces nNOS expression in the brain (25). Taken together, these results show that biglycan can be delivered to muscle in vivo and that it can direct the localization of the α-dystrobrevins and β-syntrophins to the sarcolemma. Moreover, this activity is a property of the biglycan core polypeptide and does not require the GAG side chains. Thus, biglycan can regulate the expression of the dystrobrevin-syntrophin complex at the sarcolemma in vivo.

Discussion

In this study we demonstrate that bgn null mice have a mild muscular dystrophy and provide several lines of evidence that this extracellular protein regulates the localization and expression of a subset of intracellular DAPC components in skeletal muscle. Moreover, our data indicate that the mechanisms underlying the biglycan-mediated sarcolemmal expression of the syntrophins, dystrobrevins and nNOS are likely to be distinct. Here we discuss the evidence supporting these conclusions, the potential underlying mechanisms, and the implications of these findings for potential therapies for muscular dystrophy.

Bgn null muscle displays a mild muscular dystrophy characterized by elevated serum CK and Evans blue dye uptake, increased numbers of centrally nucleated fibers and abnormal myofiber size distribution. The diaphragm was most affected, with ∼10% of the fibers showing central nuclei. These features were observed at five weeks of age, were not progressive and were not accompanied by mononuclear cell infiltration. It is noteworthy that this phenotype is qualitatively similar to that observed in α-dystrobrevin null mice (which also show reduced nNOS at the sarcolemma (26)). In that case ∼50% central nuclei are observed, but little mononuclear infiltration was seen and the overall dystrophy is considerably less severe than in mdx. α-Syntrophin knockout mice show decreased sarcolemmal expression of α-dystrobrevin-2 and nNOS, but no myopathic symptoms (27);(28). Finally, α-dystrobrevin-1 has been shown to bind to the intermediate filament proteins desmuslin and syncoilin [Newey, 2001; Blake, 2002]. It is therefore possible that the dystrobrevins help protect muscle cell membranes from contraction-induced damage by structurally linking the DAPC to the cytoskeleton. Together these findings suggest that the reduced levels of α-dystrobrevin, at least in part, underlie the dystrophic phenotype observed in the bgn null mice.

Our results indicate that distinct mechanisms underlie the reduction of dystrobrevins, syntrophins and nNOS at the sarcolemma in bgn null mice. Both immunohistochemical staining and Western blots of total muscle membranes show reductions in α-dystrobrevin -1 and -2 (Fig. 3). Moreover, the transcript levels are unchanged (Fig. 4). At present it is not known whether this post-transcriptional reduction of dystrobrevin is due to increased degradation or decreased translation. Since the levels of dystrophin are unchanged in these mice, the reduction in dystrobrevin could be due to non-dystrophin-mediated association with the membrane (28-30) and/or to defective signaling. On the other hand, nNOS seems likely to be regulated at the transcriptional level in vivo, since mRNA encoding this protein is reduced over five-fold compared to controls. Bgn could also regulate nNOS targeting, since acute treatment of myotubes with purified biglycan induces and increase in the plasma membrane-association of this enzyme. However, it should be stressed that the in vivo and the cell culture experiments are quite different. For example, the DAPC in cultured myotubes is rudimentary at best, with a fragmentary basal lamina, and substantial intracellular expression of nNOS and sarcoglycans (29, 31).

The dysregulation of syntrophins in bgn null mice is complex. All of the syntrophins show defective intracellular trafficking as evidence by increased intracellular localization and elevated expression in KCl-washed membranes. The intracellular accumulation of the syntrophins suggests that these proteins are targeted to the muscle membrane via a different mechanism than that used by the rest of the DAPC. Indeed, previous work has indeed shown that dystrophin and syntrophin take different routes during targeting to the membrane (32). Syntrophins are transported in post-Golgi vesicles that exocytose to fuse with the membrane, while dystrophin becomes directly associated with the sarcolemma following synthesis. The syntrophins also showed distinct transcriptional regulation: α– and β2- syntrophin messages were unchanged, while those encoding β1- mRNA were elevated (Fig. 4). Finally, the changes in sarcolemmal expression vary among the syntrophins: β1- syntrophin shows the most marked decrease, while α–syntrophin was less reduced and β2– was indistinguishable from wild type. Thus, while all of the syntrophins are affected in the mutant mice, biglycan is particularly important for the regulation of β1-syntrophin.

The mechanism by which biglycan signals the sarcolemmal localization of the dystrobrevin-syntrophin subcomplex or prevents the intracellular accumulation of syntrophins is not known. Biglycan binds to α-dystroglycan via its GAG side chains (3). However, the biglycan core polypeptide is active in our cell culture and in vivo assays, suggesting that the effects observed are not likely to be mediated by α-dystroglycan. In addition, since dystrophin expression is unchanged in biglycan-deficient muscle, it is unlikely that the loss of syntrophin and dystrobrevin binding sites on dystrophin is causing their mislocalization. The sarcoglycan complex can be purified with dystrobrevin and syntrophin in the absence of dystroglycan or dystrophin (31). Moreover, co-immunoprecipitation and ligand blot overlay experiments from our laboratory have demonstrated that biglycan's polypeptide core binds to α- and γ- sarcoglycan (H.H., M.R. and J.F., unpublished observations). Thus, biglycan binding to sarcoglycans, perhaps in a complex that does not contain dystrophin, could form part of the transmembrane complex by which this matrix protein regulates the localization of the dystrobrevin-syntrophin-nNOS complex.

Finally, we show that intramuscular administration of purified biglycan core polypeptide can restore the sarcolemmal expression of α-dystrobrevin-1 and -2, and β1- and β2- syntrophin in bgn null mice. This result demonstrates that biglycan can directly induce an increase in protein expression and/or a relocalization of DAPC proteins from the cytoplasm to the plasma membrane. It also demonstrates that intramuscularly injected biglycan protein appropriately localizes to the perimysium and sarcolemma and is stably associated with these sites for up to two weeks following administration. These pharmacokinetic properties of biglycan, coupled with its ability to induce specific changes in the localization of some DAPC components in muscle, suggest that it is a potential therapeutic for muscular dystrophy. It will be of interest to determine whether the administration of biglycan to dystrophin or dystrophin/utrophin knockout mice induces the appropriate expression and localization of DAPC proteins, and ameliorates the dystrophic symptoms of these mutants.

Figure 6. Intramuscular injection of biglycan core polypeptide into bgn null skeletal muscle restores sarcolemmal expression of α-dystrobrevin-1 and -2 and β1- and β2- syntrophin.

Exogenous biglycan core polypeptide increases the sarcolemmal expression of α-dystrobrevin -1 and -2 and β1- and β2- syntrophin in vivo. Two-week-old bgn null mice were injected intramuscularly into the right quadriceps femoris muscles with 50 μg biglycan core, and into the left quadriceps with buffer alone. 1.0% India ink was added to each injected solution to allow visualization of the injection site. Muscle was harvested 11 d following injection, sectioned, and immunolabeled. Biglycan core polypeptide enhances sarcolemmal expression of α-dystrobrevin-1 (α-DB-1) (A), α-dystrobrevin-2 (α-DB-2) (B), β1-syntrophin (β1-Syn) (C), and β2-syntrophin (β2-Syn) (D). The injected biglycan core polypeptide colocalizes with α-dystrobrevin and β-syntrophin expression at the sarcolemma of injected muscle. Scale bar = 20 μM.

Abbreviations List

DAPC

dystrophin-associated protein complex

nNOS

neuronal nitric oxide synthase

AChR

acetylcholine receptor

Bgn

biglycan

GAG

glycosaminoglycan

PG

proteoglycan

DB

dystrobrevin

SYN

syntrophin