Patients were diagnosed according to clinical observations, family history and serum creatine kinase (CK) levels. Muscle biopsies were obtained with informed consent.

Five unfixed serial cryosections, 10µm thick, were obtained from each muscle biopsy. After blocking of unspecific sites with blocking buffer (0.5% Triton X-100, 1% BSA in PBS) for 30 minutes at room temperature, the sections were incubated with the primary antibodies overnight at 4°C in a humid chamber. After washing in PBS, the sections were incubated with the appropriated fluorescent secondary antibody for 30 minutes at room temperature. All tissue sections were mounted in aqueous mountant and viewed with a fluorescence microscope (Nikon), images collected exploiting a digital Nikon camera (Coolpix 5400) using the software X-Pro.

Muscle biopsies from dystrophic patients and control sample were homogenized in lysis buffer (125mM Tris-HCl pH 7.4, 2mM EDTA, 1% SDS, 1% Triton X-100) containing protease inhibitors; supernatants (total protein extract) were collected upon centrifugation. Proteins were separated by SDS-PAGE and transferred to nitrocellulose (NT) membranes. After blocking with 5% dry milk Tris-buffered saline (TTBS; 150 mM NaCl, 10 mM Tris-HCl buffer pH 7.4 and 0.05% Tween 20), the NT membranes were incubated o/n at 4°C with primary antibody. After washing in TTBS and in TTBS containing 3% milk powder for 30 min, the blots were incubated for 1 h with anti-mouse or anti-rabbit antibodies horseradish peroxidase-conjugated (Pierce).

After washing in 3% milk powder TTBS and several washes in TTBS, blots were developed using enhanced chemiluminescence (ECL) (Pierce). The immuno-reacting bands were visualized by exposing the membranes to Kodak X-OMAT films. Glycoprotein enrichment was performed on total protein extracts. Briefly, supernatants were incubated with Wheat Germ Lectin (WGL) Sepharose 6MB (Amersham) overnight at 4°C. After extensive washing with washing buffer (WB) (50mM Tris-HCl pH 7.8, 500mM NaCl, 0.1% Triton X-100), bound glycoprotein were eluted in WB with 250mM N-acetylglucosamine and used in Western blot.

In Western blot and in immunofluorescence experiments, β-DG was detected with mouse monoclonal antibodies anti-β-DG 43DAG/8D5 (Novocastra) diluted 1:25 and anti-β-DG BD (Becton Dickinson) diluted 1:500. α-DG was detected with mouse monoclonal antibodies anti-α-DG clone VIA4-1 (lot 32685, Upstate Biotechnology) and clone IIH6 (lot 22991 Upstate Biotechnology) diluted 1:500 and 1:200, respectively. Polyclonal antibodies directed against the mouse α-DG C-terminal region previously characterized [21] were used at dilution 1:100.

Three male and four female patients affected by non-congenital muscle dystrophies were studied. Age at onset was on average 25 years (range 10-35 years), and the average age at biopsy 32 years (range 24-48 years). Three patients, P2, P4 and P7, reveal a positive family history; on detail, P4 and P7 have one brother affected respectively, whereas P2 has one brother and one sister affected. The creatin kinase level was analyzed and the values were within a range of 1 to 40 times the normal value. In particular, P1 underwent biopsy before showing increased levels of creatin kinase.

All patients had a normal respiratory and cardiac function, with the exception of P2 who was affected by a coronary artery disease.

Some patients were also genetically characterized. In particular, the calpain (CAPN3), dysferlin (DYSF), telethonin (TCAP), lamin (LMNA), γ-sarcoglyan (γ-SGC), fukutin, POMT1 and fukutin related protein (FKRP) genes were analyzed. Among the genes analyzed, only P3, P5 and P7 displayed mutations in the dysferlin gene. The clinical details and genetic analysis outcome are reported in Table 2.

Immunofluorescence analysis of the DG subunits was performed on skeletal muscle cryosections of all patients, with the exception of P4, using a panel of antibodies hitting different epitopes within the α- and β-DG subunits, as depicted in Fig. (1). The β-DG subunit was stained with two different antibodies hitting the extracellular region (BD) and the cytoplasmic domain of the protein (43-DAG). The β-DG signal in all the samples analyzed was comparable with the control section, and showed normal sarcolemma expression (Fig. 2). Also the membrane localization of the α-subunit was normal when detected with a polyclonal antibody raised against the core protein of the C-terminal domain of α-DG: furthermore, the positive staining with an antibody raised against the carbohydrate molecules revealed the correct glycosylation of the α-subunit in all patients (Fig. 2).

Fig. (1). Scheme of the DG subunits displaying the epitopes recognized by different antibodies.

Fig. (2). Immunofluorescence localization of α- and β-DG in human skeletal muscle sections. C: control, P1-P7: patients. The β-DG subunit was detected with two monoclonal antibodies, 43DAG/8D5 and BD, that recognized the β-DG cytoplasmic tail and its extracellular domain, respectively. α-DG was detected with the monoclonal antibody IIH6 (lot 22991), raised against an undefined carbohydrate epitope (red arrow), and with a polyclonal antibody directed against its C-terminal region.

The α- and β-DG subunits were also analyzed by Western blot using total skeletal muscle protein extracts. α-DG was probed with the monoclonal antibody VIA4-1 (Upstate), that recognises a yet unidentified carbohydrate epitope. The expected broad band of approximately 156 kDa was present in the control sample and in patients P1, P2, P3 and P7 (Fig. 3A, red arrow). The patient P5 displays a strong signal for α-DG but with a slight shift toward lower molecular mass (Fig. 3A). On the contrary, in patients P4 and P6 the 156 kDa band of α-DG was totally absent (Fig. 3A, red asterisk), although in P6 a very faint band around 120 kDa was also detected that could represent a hypoglycosylated form of α-DG still recognized by the antibody. β-DG was analyzed in total protein extracts using the monoclonal antibody NCL-43DAG that recognizes the last 15 aa of its cytoplasmic tail. All the samples analyzed showed the expected 43 kDa band (Fig. 3B).

Fig. (3). Western blot analysis of the DG subunits in total protein extracts from patients P1-P7 and from control muscle tissue (C). A: Samples were resolved with a 7.5-15% acrylamide gradient SDSPAGE and α-DG was detected with the antibody VIA4-1 (lot 32685) raised against an undefined carbohydrate epitope (red arrow). In samples P4 and P6 the expected 156 KDa band was absent (red asterisk). The additional lower bands detected in all the samples might correspond to hypoglycosylated and/or proteolytic forms of α-DG; B: Western blot upon a 12% SDS-PAGE of the total protein extracts using anti 43DAG/8D5 for β-DG detection. All the samples showed the expected 43 kDa band.

The total protein extracts of patients P4 and P6, which showed an anomalous signal of α-DG in Western blot, were submitted to a lectin-enrichment protocol using wheat germ lectin (WGL) beads that specifically bind to N-acetylglucosamine residues. This protocol is very useful for the analysis of the carbohydrate decorations of a glycoprotein like α-DG. In the normal tissue sample, WGL-enriched fractions contain both α- and β-DG subunits (Fig. 4A). The WGL-enriched fractions from patients P4 and P6, analyzed with an antibody against β-DG, showed a loss of immunodetection in comparison to the corresponding total protein extract (Fig. 4B).

Fig. (4). Western blot analysis of total protein extracts of control and patients P4 and P6 that were submitted to a lectin-enrichment protocol using wheat germ lectin beads. A: In the control tissue sample, WGL-enriched fractions contain both the α-DG and β-DG subunits. B: WGL-enriched fractions from patients P4 and P6, analysed with antibodies against β-DG. β-DG, that is present in the total protein extracts of both patients, is not enriched upon WGL incubation, confirming the presence of altered glycosylation of α-DG.

In this study, we have analyzed the α- and β-DG subunits in seven patients who had been catalogued by clinical observation as subjects affected by non-congenital muscular dystrophies (Table 2). Five of those patients were classified as affected by LGMD, one by distal myopathy and another one by Miyoshi myopathy.

We have used a combination of immunofluorescence and immunoblot analysis to characterize the DG complex and to determine whether an alteration of the DG expression pattern emerged in any of these patients affected by non-congenital muscular dystrophies. Firstly, we analyzed the dystroglycan subunits expressed in skeletal muscle by immunofluorescence of tissue sections, using antibodies hitting different epitopes within the two molecules (Fig. 1). In particular, the β-DG subunit was stained using two monoclonal antibodies directed against its extracellular domain and its cytosolic C-terminal domain, respectively (Fig. 2), while the α-DG subunit was detected using a polyclonal antibody raised against the core protein at the level of its C-terminal domain, [21], and by a monoclonal antibody directed against an unknown carbohydrate epitope of α-DG (Mab IIH6 purchased from Upstate) (Fig. 2).

In all patients, all the different antibodies detected the expected membrane localization of the DG subunits (Fig. 2). From these very experiments, the DG complex does not seem to be related at all to the dystrophic phenotypes observed in those patients. However, Western blot analysis of crude muscle protein extracts, performed using a different monoclonal antibody directed against a carbohydrate epitope of α-DG (Mab VIA4-1, purchased from Upstate), showed the loss of α-DG immunodetection in patients P4 and P6, while patient P5 displayed a signal with a slight shift toward a lower molecular mass (Fig. 3A).

The discrepancy between the immunofluorescence and Western blot analysis may find an explanation in the possibility that α-DG glycosylation is indeed altered in these patients, and the carbohydrate epitope recognized by the VIA4-1 antibody is missing or is reduced. It should be noted that α-DG does not lose its sarcolemmal targeting, so it could be hypothesized that the altered glycosylation may rather affect its functional properties [22]. Such a picture seems to be confirmed by the lectin enrichment protocol that failed to co-purify β-DG from total protein extracts of P4 and P6 patients (Fig. 4B). In fact, the complete loss of immunodetection of the β-DG subunit in P4 and P6 WGL-enriched samples should depend on the impaired interaction between α-DG and the lectin, which in itself depends on the presence of N-acetylglucosamine residues in the mucin-like region of α-DG (Fig. 3A) [22].

Therefore, P4 and P6 patients may well be affected by a form of “dystroglycanopathy” that could be related to the reduced amount of N-acetylglucosamine present within the glycosylation shell of α-DG. Being POMGnT1 the glycosyltransferase that typically add the N-acetylglucosamine “brick” to the growing O-mannose sugar chains, these two patients (P4 and P6) could be tentatively catalogued as LGMD2M, even if muscle weakness displayed a distal involvement as well. Moreover, the patient P6 revealed also a deficit of calpain-3 (data not shown) that is characteristic of LGMD2A. Further genetic analysis will be essential to test this hypothesis and to identify directly and unambiguously the mutated gene responsible of the altered glycosylation of α-DG.

Our analysis of a small cohort of patients revealed that the biochemical characterization of patients affected by muscular dystrophy should include a more extensive immunological analysis of muscle samples using different antibodies.

In fact, muscle sections of all the patients showed in immunofluorescence the same expression and targeting pattern of the two DG subunits. On the other hand, when applying Western blot analysis, especially upon lectin-enrichment, two patients were unambiguously identified as patients displaying an altered glycosylation pattern of α-DG. Although the possible role of POMGnT1 in a form of LGMD also with distal involvement could be only speculated, these patients should now undergo an extensive gene analysis aiming at a definitive characterization of the genes responsible for their pathologies.

After the identification of DG as a common final target for a whole series of glycosyltransferases responsible for its O-glycosylation pattern, the DG complex has taken centre stage in neuromuscular medicine [9]. Nonetheless, many important structural and functional details of the DG complex, in the light of its involvement in human neuromuscular disorders, await further elucidation. In particular, the novel identification of a pathogenic missense mutation in the DG gene that results in neuromuscular abnormalities, points to the opportunity of including DAG1 in the list of genes whose mutations cause muscular dystrophy and cognitive impairment in humans [19, 23].

Moreover, it has not been clarified yet whether DG would represent the substrate of other still unidentified enzymes. In fact, there are several reported cases in which patients, while displaying a series of phenotypic signs that would suggest the presence of a secondary dystroglycanopathy, do not carry mutations in any of the glycosyltransferases identified so far [9]. It is also possible that additional unknown substrates of these enzymes exist, enlarging the pathological scenario of sarcolemmal instability and necrosis [23].

During the next few years, further efforts will certainly be needed in order to identify alternative biochemical and/or genetic pathways underlying this group of congenital and limb-girdle muscular dystrophies.