We have recently shown that a deletion in the Large gene, encoding a putative glycosyltransferase, is the molecular defect underlying the myodystrophy (previously myd; now Largemyd) mouse. Here we show that the muscular dystrophy phenotype is not confined to skeletal muscle, but is also present in the heart and tongue. Immunohistochemistry indicates disruption of the dystrophin-associated glycoprotein complex (DGC) in skeletal and cardiac muscle. Quantitative western blotting shows a general increase in the expression of DGC proteins and of dysferlin and caveolin-3 in mutant skeletal muscle. In contrast, the expression of DGC proteins is reduced in cardiac muscle. Overlay assays show loss of laminin binding by α-dystroglycan in Largemyd skeletal and cardiac muscle and in brain. We also show that the phenotype of Largemyd mice is not restricted to muscular dystrophy, but also includes ophthalmic and central nervous system (CNS) defects. Electroretinograms of homozygous mutant mice show gross abnormalities of b-wave characteristics, indicative of a complex defect in retinal transmission. The laminar architecture of the cortices of the cerebrum and the cerebellum is disturbed, indicating defective neuronal migration. Thus, the phenotype of the Largemyd mouse shows similarities to the heterogeneous group of human muscle eye brain diseases characterized by severe congenital muscular dystrophy, eye abnormalities and CNS neuronal migration defects. These diseases include Fukuyama-type muscular dystrophy and muscle–eye–brain disease, both of which are also due to mutations in predicted glycosylation enzymes. Therefore, the Largemyd mouse represents an important animal model for studying the function of glycosylation in muscle, brain and retina.
The sporadic, recessive mouse mutant myodystrophy (myd) is characterized by a congenital progressive muscular dystrophy associated with motor deficits such as a shuffling gait and abnormal posturing when suspended by the tail (1,2). Homozygous mice have elevated levels of serum creatine phosphokinase (3), they are smaller than littermate controls and their lifespan is shortened (1). myd mice have also been reported to exhibit sensorineural hearing impairment (2) and defective myelination in spinal root nerves (4).
The myd locus maps to mouse chromosome 8 (1,5), and, based on homology of this region to human chromosome 4q, was proposed as an animal model of the human autosomal dominant neuromuscular disorder facioscapulohumeral muscular dystrophy (FSHD) (2,5). However, refined mapping (6) and cloning of the myd locus (7,8) excluded it as a candidate for FSHD and identified the causative mutation in myd as a deletion of exons 5–7 of the Large gene (7). This deletion results in a frameshift in the corresponding mRNA, leading to a premature termination codon. The myd mutation is now designated by the symbol Largemyd (www.jax.org). Large encodes a type 2 membrane protein with homology to several glycosyltransferases, although its biochemical activity has not yet been confirmed (7,9). As the human homologue (LARGE) maps to chromosome 22q (9), Largemyd is not a model of FSHD.
Because the predicted function of Large is a glycosyltransferase (9) and muscular dystrophy is a predominant phenotype of Largemyd, we investigated the glycosylation status of components of the dystrophin-associated glycoprotein complex (DGC) by immunoblotting. An important role of the DGC is to maintain sarcolemma membrane integrity by linking cytoskeletal actin (via dystrophin) to the extracellular matrix, and mutations in almost all of the genes encoding members of the DGC are associated with muscular dystrophy phenotypes (10,11). On immunoblots of protein extracts from myd brain, skeletal and cardiac muscle, we observed a marked reduction in α-dystroglycan (α-DG) signal using the monoclonal antibody VIA41, while a polyclonal antibody raised against a peptide gave equivalent signals in mutant and control tissue (7). Dystroglycan is produced as a precursor polypeptide that is post-translationally modified and cleaved into α and β subunits. α-DG is a peripheral membrane protein linking the DGC to the extracellular matrix, while the β subunit (β-DG) is a transmembrane protein interacting with both α-DG and dystrophin (12). α-DG is extensively and variably O-glycosylated in a tissue-specific manner. We speculated that the loss of VIA41 immunoreactivity in Largemyd mice is a consequence of altered glycosylation of α-DG (7).
Subsequently, four inherited human muscular dystrophies varying in accompanying pathologies affecting the heart and the tongue, the visual system and the brain, have been found to arise from mutations in genes encoding known or putative glycosylation enzymes (13–16). Fukuyama-type congenital muscular dystrophy (FCMD; MIM 253800) is associated with cerebral and cerebellar cortical dysplasia, profound mental retardation and ophthalmological abnormalities in addition to the muscle defect (17). FCMD is due to mutations in the fukutin gene on human chromosome 9q (18). A bioinformatics study proposed that fukutin is an enzyme involved in carbohydrate modification (19). Using VIA41, the same monoclonal antibody as in our mouse study, Hayashi et al. (16) demonstrated that FCMD patients show loss of α-DG signal on immunoblots of skeletal muscle, whereas α-DG was detected in brain tissue of the same FCMD patients at only slightly reduced amounts but normal molecular weights. The fukutin-related protein gene (FKRP) was identified on the basis of homology to fukutin, and is mutated in a severe form of congenital muscular dystrophy without brain involvement (MDC1C; MIM 606612) (14). Immunoblotting, again with monoclonal antibody VIA41, showed a reduction in the size of α-DG in skeletal muscle from two MDC1C individuals, which may reflect reduced glycosylation. FKRP mutations are also associated with limb girdle muscular dystrophy type 2I (LGMD2I), an allelic and milder variant of MDC1C (13). Finally, muscle–eye–brain disease (MEB; MIM 253280), characterized by congenital muscular dystrophy, ocular abnormalities and type II lissencephaly, is due to mutations in POMGnT1, which encodes O-mannose β-1,2-N-acetylglucosaminyltransferase (15). POMGnT1 is an enzyme involved in the formation of the GlcNAcβ1–2Man linkage found in O-mannosylglycans (15,20). This unusual glycan (21) is known to be present on α-DG (22–24), implicating a role for POMGnT1 in post-translational modification of this DGC protein.
Here, we show that in the Largemyd mouse, progressive dystrophy encompasses not only the limb muscles and diaphragm, but also the cardiac muscle and most severely the intrinsic muscles of the tongue. We also demonstrate ocular defects (homozygous mutants have abnormal electroretinograms) and a central nervous system (CNS) phenotype characterized by neuronal migration defects in the cerebrum and cerebellum.
The pronounced muscular dystrophy in Largemyd mice is also present in the diaphragm, heart and tongue
In order to obtain an overview of the distribution and severity of the muscular dystrophy in Largemyd mice, we investigated a series of muscle types from animals of different ages (ranging from 3 to 35 weeks) using conventional histology and electron microscopy. In order to identify disruptions of the muscle fibre membranes, some animals were injected with Evans blue dye (EBD) prior to the histological assessment. EBD does not cross into skeletal muscle fibres in normal mice, and accumulation within fibres is indicative of loss of sarcolemmal integrity (25).
Largemyd mice show an age-dependent progression of muscle pathology in all muscles of the limbs and of the trunk (Fig. 1). These changes are characterized by muscle fibre necrosis and regeneration, internalization of myonuclei, and proliferation of endomysial connective and fatty tissue. In young animals (3 weeks), the diaphragm appears more severely affected by the dystrophic process than the limbs, with multiple groups of EBD-positive fibres (Fig. 1A: inset). Although dystrophic processes were clearly visible in proximal muscle groups of the limbs, the most severely affected muscles were the psoas (Fig. 1B), the iliacus and the paraxial muscles of the back (not shown). The severity of histological changes in different muscles displayed good correlation with the number of EBD-positive fibres, suggestive of a defect in sarcolemmal integrity.
By electron microscopy, we observed multiple structural defects of the muscle fibre sarcolemma. The plasma membrane of the muscle fibres appears discontinuous (Fig. 1C), while the sarcolemma contains severe lesions where little or no basal lamina is visible and the underlying plasma membrane is absent (Fig. 1D). In general, irrespective of the presence or absence of severe structural abnormalities, the basal lamina of the muscle fibres is thin, disorganized and frequently detached from the underlying plasma membranes (Fig. 1C). The basal lamina surrounding small endomysial blood vessels and capillaries always appears normal (Fig. 1C and E).
Our previous data suggested that glycosylation of α-DG is abnormal in heart and brain, as well as skeletal muscle, in Largemyd mice (7). Previous detailed phenotypic studies did not include the heart (1,2), although an early report mentioned the presence of focal lesions in the myocardium (3). We found evidence of cardiomyopathy in Largemyd mice (Fig. 2), which was more pronounced in older animals (20–25 weeks). Histological analysis showed multiple areas of focal fibrosis, mainly in the left ventricular walls (Fig. 2A), and scattered groups of EBD-positive muscle fibres throughout (Fig. 2B). Electron micrographs showed disorganized myofibrils within multiple cardiac muscle cells, as well as degenerative vacuolar changes (Fig. 2C). However, no signs of necrotic cardiomyocytes were visible, either by light or by electron microscopy.
Because hypertrophy of the tongue muscles has been reported as a frequent finding in LGMD2I (13), we were also interested in determining whether the pathology in Largemyd mice also involves the tongue. Comparison of the weights of whole tongues, the maximal anterior–posterior distances and the cross-sectional areas of corresponding sections showed that none of these parameters differed significantly between the two groups (data not shown), indicating that tongue hypertrophy is not a prominent feature. Histological examination of the outermost subepithelial muscle layers of Largemyd tongues at different anterior/posterior locations revealed only mild pathology, such as scattered centrally nucleated muscle fibres and minor endomysial fibrosis (Fig. 3A and B). Anatomically, the deeper portions of the tongue muscles can be identified as separate layers of muscles in normal mice. However, in Largemyd animals, these muscle layers were completely missing and were replaced by atypical fibrous tissue (Fig. 3C and D). At higher magnification, we identified small myotubes embedded in the connective tissue (Fig. 3E and F).
Alterations in sarcolemmal localization of DGC proteins in Largemyd mice
The monoclonal antibody VIA41 showed an even sarcolemmal localization of α-DG in normal muscle fibres. In agreement with our previously published immunoblot data (7), this antibody showed almost no reactivity on sections from Largemyd muscles (Fig. 4A and B). In contrast, sarcolemmal α-DG immunoreactivity was seen in the mutant muscle with the polyclonal antibody, although compared to control tissue this was patchy with additional diffuse sarcoplasmic staining (Fig. 4C and D). Sarcolemmal expression of β-DG in Largemyd skeletal muscle is also reduced and irregular, with positive staining in the sarcoplasm of numerous smaller muscle fibres (Fig. 4E and D′). Michele et al. (26) recently reported that another monoclonal antibody to a carbohydrate epitope of α-DG (IIH6) also shows loss of immunoreactivity in Largemyd muscle; however, normal sarcolemmal expression was observed using an α-DG polyclonal antibody (GT20ADG). On western blots, α-DG shows a significant reduction in molecular mass, demonstrating hypoglycosylation of the protein in Largemyd muscle and brain.
The sarcolemmal localization of other DGC proteins appears to be altered in Largemyd muscle (Fig. 5). As well as unaffected control mice, we compared our findings with those from dystrophin-deficient mdx mice, where there is secondary loss of other DGC components (27). A uniform immunoreactivity of the sarcolemma was obtained with all of the sarcoglycan antibodies in wild-type muscles, while, as expected, staining in mdx skeletal muscles was almost completely absent. In Largemyd muscle, only α-sarcoglycan (α-SG) showed uniform labelling of the sarcolemma, the localization of other SGs was disrupted, with extensive sarcoplasmic staining of numerous dystrophic muscle fibres. Dystrophin also showed weak and patchy sarcolemmal staining. Localization of caveolin-3 (cav-3), the main component of membrane-bound caveolae in skeletal muscle, was also disrupted, with increased sarcoplasmic expression in many muscle fibres (Fig. 5).
By immunocytochemistry, the expression of several DGC components was reduced in Largemyd cardiac muscle (Fig. 6). Dystrophin showed a weak and patchy distribution compared with controls. Control cardiac muscle showed strong sarcolemmal staining for β-DG and γ-sarcoglycan (γ-SG), in addition to sarcoplasmic immunoreactivity for both antibodies. In contrast, we observed no sarcolemmal localization of either β-DG or γ-SG in Largemyd hearts, in addition to reduced or absent sarcoplasmic signals.
Levels of DGC proteins are increased in Largemyd skeletal muscle but decreased in cardiac muscle
We then used quantitative western blotting to compare the relative amounts of these proteins in control and Largemyd skeletal and cardiac muscle (Fig. 7). In agreement with previous data (7), the monoclonal antibody VIA41 produced a diffuse α-DG band (Molecular weight ∼150 kDa) in control skeletal muscle, but no signal in mutant tissue (Fig. 7). In contrast, the anti-peptide polyclonal antibody produces a very sharp discrete band on mouse western blots in both wild-type and mutant tissue (7). It is likely that this polyclonal antibody recognizes only a subset of glycosylated isoforms of α-DG, and therefore is not suitable for quantitative analysis.
The intensities of the β-DG, γ-SG and dystrophin bands were consistently increased in all Largemyd muscles, irrespective of the age of the mice. This increase was reproducibly observed in different skeletal muscles, including quadriceps femoris, tibialis anterior and gastrocnemius. Figure 7 shows data from quadriceps femoris and cardiac muscles from 8-week-old Largemyd and control mice. For densitometric analysis, all values were normalized to the intensities of the α-actinin bands, and the protein levels detected in control extracts were assumed to be 100%. We found that dystrophin levels were only moderately increased in Largemyd skeletal muscle (∼120%), whereas both β-DG and γ-SG were markedly increased (∼700% and ∼200%, respectively; Fig. 7A′). Dysferlin and cav-3 were also expressed at higher levels in the mutant (∼600% and ∼150%, respectively), indicating that the alteration in sarcolemmal proteins is not restricted to components of the DGC (Fig. 7B′).
In agreement with the immunohistochemical data, expression of all the DGC proteins tested was diminished in cardiac muscle from Largemyd mice (Fig. 7C and C′). The level of dystrophin was ∼75% that of wild type, while the expression of β-DG and γ-SG was about half that of controls.
Disruption of the α-DG–laminin interaction in Largemyd mice
These findings suggested loss of muscle sarcolemmal integrity in Largemyd mice. While the numerous EBD-positive muscle fibres in skeletal and cardiac muscle were suggestive of structural defects of the plasma membrane, the electron-microscopic observations showed gross structural abnormalities of the muscle basement membranes (Fig. 1). Laminin, which is an important component of basement membranes, is assembled as a heterotrimer of α, β and γ chains; variations in chain combination producing at least 12 isoforms (28). The α2 chain is highly expressed in skeletal muscle, and mutations in the gene (LAMA2) produce severe congenital muscular dystrophies (CMD) that are often associated with peripheral and CNS defects (29–31). Secondary loss of laminin, without mutations in LAMA2, is also seen in some other forms of CMD (32). A pan-laminin antibody (33) showed bright labelling at the muscle fibre peripheries and around small endomysial blood vessels in normal skeletal muscle (Fig. 8A). There is a moderate reduction in staining of sections from dystrophic dy/dy mice, which are deficient for the laminin α2 chain (Fig. 8B), but a greater reduction in laminin reactivity in Largemyd skeletal muscle (Fig. 8C).
α-DG binds laminins 1–4 directly, via the C-terminal globular (G) domain repeats present in all known laminin α chains (34). Glycosylation of α-DG is necessary for this interaction (35), and there is evidence from blocking studies that O-mannosyl-type or sialic acid oligosaccharide structures are required (22,36). Therefore, we used a blot overlay assay (36) to test the ability of tissue extracts from normal and mutant mice to bind laminin. This result shows that laminin binding is impaired in Largemyd skeletal and cardiac muscle and brain (Fig. 8D and E), in agreement with recent data demonstrating reduction of laminin binding in Largemyd skeletal muscle and brain (26).
Retinopathy in Largemyd mice implies functional impairment of synaptic transmission
Two of the human muscle disorders associated with defective glycosylation (MEB and FCMD) have ophthalmic abnormalities as part of the phenotypic spectrum. To determine whether loss of the Large glycosyltransferase also results in an ocular phenotype, we investigated the eyes, and in particular the retinas, of the animals. No gross abnormalities were found in either anatomical compartments of the eyeball or the retina (data not shown). Abnormal electroretinogram (ERG) recordings are a feature of Duchenne and Becker muscular dystrophies as well as the dystrophin-deficient mdxCv3 mouse (37,38). Therefore, we investigated the visual transmission pathway in Largemyd mice of different ages by recording ERGs.
Single-flash ERGs were found to be abnormal in all Largemyd mice, and differences between the wild type and the mutant were discernable at all levels of stimulus intensity (Fig. 9A). Figure 9B shows an overlay of averaged responses at 5 log units above threshold stimulation. ERG recordings from Largemyd mice showed substantial pathological changes, with pronounced deviations of the b-wave amplitude, implicit time and b/a-wave amplitude ratio compared with wild-type animals. In contrast, the a-wave amplitude and implicit time (peak latency) did not differ significantly between groups (Table 1).
Neuronal migration defects in Largemyd cerebrum and cerebellum
Both FCMD and MEB are characterized by neuronal migration defects. Therefore, we examined coronal and sagittal serial sections of the brains of Largemyd mice of different ages (3–25 weeks). In all mutant mice, the architecture of the cortical lamination of both the cerebrum and the cerebellum was disturbed. In the cerebral cortex, neuronal heterotopia was seen throughout the outermost layer, the molecular layer (ML) (Fig. 10A and B). While these mis-localized neurons were mostly located diffusely within the ML, in some areas they formed small groups (Fig. 10C). These aberrant neurons were observed throughout the cerebrum. Michele et al. (26) recently reported similar neuronal migration defects in Largemyd mice. In contrast to their study, we observed no structural abnormalities of the neuronal basal lamina (not shown); moreover, we found no visible pathology in the hippocampus in the majority of mutant animals (Fig. 10E). In only two out of six Largemyd mice, we found mildly abnormal layering of the granular cells of the dentate gyrus (Fig. 10F). Furthermore, these abnormalities were only seen unilaterally, since the contralateral hippocampi of these two mice were structurally normal (not shown).
In the cerebellum of normal mice, the cortical structures were clearly discernible, as shown in Figure 10G. The molecular layer (ML) lies above the Purkinje cell layer and the tightly packed granular layer (GL) containing the majority of cortical neuronal cell bodies. The white matter, consisting mainly of myelinated axons, is beneath the GL. The subpial layers of the cerebellar cortex are essentially devoid of neurons in normal adult animals (arrows in Fig. 10G). However, the cerebella of Largemyd mice contain multiple clusters of remnants of external granular layer neurons, presumably due to a migration defect (arrows in Fig. 10H and I). Clusters of heterotopic neurons were also observed within the white matter (open arrow in Fig. 10H) or had formed small clusters beneath the pial surface (Fig. 10I). Purkinje cells normally form a layer between the ML and the GL (open arrows in Fig. 10J). In Largemyd mice, we often observed Purkinje cells that had failed to reach their correct neuroanatomical location and remained within the ML (arrows in Fig. 10J).
Defects in glycosylation enzymes have emerged as an important pathomechanism for muscular dystrophy; the Largemyd mouse currently represents the only animal model of such a defect. Detailed analysis of the phenotype is therefore likely to be useful in furthering our understanding of the importance of glycosylation with respect to muscle function. The data presented here expand our knowledge of the diversity of affected tissues and organs in this animal model to include the heart, retina and CNS.
The expanded muscular dystrophy phenotype in Largemyd mice: degeneration in the diaphragm, the myocardium and the tongue
Our findings reported in this paper are in overall agreement with previously reported observations concerning the pathological alterations of the limb muscles in Largemyd mice (1,2). The diaphragm is severely affected, being one of the first organs to show signs of pathological changes (3,39; this study). The myocardial lesions that we report here are very similar to those described in mdx mice (25) and the δ-SG-deficient cardiomyopathic hamster (40). Our findings are suggestive of a chronic cardiomyopathy in Largemyd mice. In this context, it is noteworthy that in patients mutated for FKRP, left ventricular cardiomyopathy of differing severity has been reported (14).
We also show that Largemyd mice have severe dystrophic changes of the intrinsic muscles of the tongue, completely disrupting the anatomical structure of the different muscle layers. These severe changes must result in a profound functional impairment of the tongue, such as with regard to chewing and swallowing, and could well account for the progressive weight loss seen in these animals after weaning (1). Hypertrophy of the tongue has been reported in patients with FKRP mutations (13,14), but we could not detect any signs of this in Largemyd mice.
Structural disruption of the muscle sarcolemma in Largemyd mice is accompanied by changes in DGC components and loss of dystroglycan–laminin binding
Based on the EBD-positive muscle fibres and cardiomyocytes, and on our electron-microscopic observations, the plasma membrane and the basal lamina are structurally disrupted in Largemyd mice. Largemyd skeletal and cardiac muscle also displays quantitative and spatial changes in expression of sarcolemmal proteins in addition to α-DG. In numerous skeletal muscle fibres, we saw a reduction in sarcolemmal localization but increased sarcoplasmic staining signals for DGC proteins and for cav-3. However, in cardiac muscle, we saw an overall reduction immunoreactivity for dystrophin, β-DG and γ-SG in both the sarcolemmal membrane and the sarcoplasm. These differences were corroborated by the quantitative western blot data. The implications of these findings with respect to the underlying molecular pathways and to the development of skeletal muscle and cardiac phenotypic manifestations are unclear, but suggest a general disruption of DGC localization and/or processing. Our data differ from those of Michele et al. (26), who did not observe any qualitative difference in skeletal muscle sarcolemmal expression of β-DG or dystrophin between Largemyd mice and controls, although no quantitative analysis was included in this study. The muscular dystrophy may be more severe in the animals that we studied owing to age and/or background strain differences (41).
The laminin overlay assay shows that, in addition to skeletal muscle and brain, cardiac α-DG in Largemyd has reduced laminin-binding activity, consistent with the altered glycosylation of α-DG and disruption of the DGC that we observe in the heart. Skeletal muscle biopsies from MEB and FCMD patients also show reduced laminin-binding activity and hypoglycosylation (26), suggesting that the muscular dystrophy phenotypes in these human diseases and in Largemyd mice are due to disruption of the DGC–laminin axis. If this is the case, we would expect the muscle phenotype in Largemyd to be rescued by overexpression of alternative laminin receptors such as α7β1 integrin, since this significantly alleviates the muscular dystrophy seen in mdx/utr−/− mice (42).
The retinal pathology in Largemyd implies importance of Large for proper visual signal processing in the retina
The ERG is a sensitive tool to investigate functional abnormalities of the retina. The normal a-wave in Largemyd mice suggests that photoreceptor function is unaffected by the glycosyltransferase mutation. However, the increased implicit times and decreased amplitudes of the b-wave suggest altered signal processing in the downstream retinal circuitry (43). There could also be a role for the DGC in this aspect of the phenotype. In both humans and mice, loss of retinal isoforms of dystrophin results in a similar b-wave abnormality to that seen in Largemyd mice (Fig. 9) (44). Components of the DGC, including dystroglycan, are expressed in several regions of the retina, including the outer plexiform layer and Müller cell processes (45). Müller cells, a radial retinal glia cell type, are thought to be involved in the generation and shaping of the b-wave (43). Antibody studies have shown several glycoforms of α-DG in the retina, indicating that correct glycosylation of the protein may be functionally important (45). However, there are differences between the phenotype that we report here and MEB, where patients typically have a severe progressive myopia, retinal degeneration and cataracts (46). Furthermore, a different ERG abnormality, giant flash visual evoked potentials, is characteristic of this disease (46).
Functional Large protein is required for neuronal migration during CNS development
Our data showing features suggestive of abnormal neuronal cell migration in the CNS of Largemyd mice are generally in agreement with recently reported findings (26). We found a unilateral abnormality in neuronal layers of the hippocampus in two out of six mutants investigated. In contrast to previous findings (26), we did not observe structural abnormalities of the neuronal basal lamina, which is in good agreement with the lack of externally migrated cortical neurons in Largemyd brains (26; this study). Although the reasons for the discrepancies in neuronal phenotype between our findings and those of Michele et al. (26) are unclear, they might be due to effects of background strain on neuronal development (47).
Human type II lissencephaly, as seen in MEB and FCMD, is characterized by a ‘cobblestone’ cortex where breaches of the basement membrane allow external migration of the neurons, resulting in polymicrogyria (48). Cobblestone cortex is also seen in a third type of congenital muscular dystrophy associated with brain and eye defects, Walker–Warburg syndrome (WWS); the molecular basis of this disease is unknown (49). Selective deletion of dystroglycan in the brain produced a CNS phenotype in mice strikingly similar to the lissencephaly seen in these congenital muscular dystrophies. In combination with data showing that α-DG appears to be abnormally glycosylated or reduced in MEB and FCMD brain (16,50), this implicates dystroglycan as a key molecule in development of this CNS pathology (51). As α-DG is also hypoglycosylated in Largemyd brain (7,26), and DGC proteins, syntrophin and aquaporin fail to target to astrocyte foot processes (26), similar mechanisms to those in human congenital muscular dystrophies are likely to underlie the mouse brain abnormalities.
The phenotypic trilogy of muscle, eye and brain pathologies is common to the glycosylation-defective conditions MEB, FCMD and Largemyd, but is not a feature of MDC1C/LGMD2I patients, indicating tissue-specific differences in the functions of the enzymes involved in the respective disorders. Although abnormal glycosylation of α-DG is strongly implicated as having a causative role in the development of CNS abnormalities in MEB, FCMD and Largemyd, it might not account for the whole story—in FCMD, cobblestone lissencephaly occurs even in the presence of significant amounts of apparently normally glycosylated α-DG (16). It is interesting that Largemyd homozygotes (which apparently lack functional α-DG) are viable, yet null mutants are lethal because of failure of early embryonic development (52). No human disorder corresponding to Largemyd has yet been identified. However, mutations in the human homologue of Large may not produce a muscular dystrophy, but instead primarily present with defects of the cardiac, ocular or CNS organs.
MATERIALS AND METHODS
Animals used in this study and routine histology
Colonies of B6C3Fe-Largemyd (former myd) mice were originally obtained from Jackson Laboratories (Bar Harbour, ME) in 1993. To get a more homogeneous background, they were backcrossed twice to C57BL/6JHim, with subsequent intercrosses not to lose the mutation, and then inbred. C57BL/6-dy/+ mice, also purchased from Jackson Laboratories in 1993, and C57BL/10ScSn-mdx mice obtained from the University of Dundee in 1988 are maintained at the Institute for Laboratory Animal Sciences and Genetics, University of Vienna, Himberg. Animals of both sexes were used at ages between 3 and 30 weeks. Animals were sacrificed either by cervical dislocation or by deep anaesthesia followed by transcardial perfusion with 4% paraformaldehyde/1% cacodylate buffer. Tissues were snap-frozen in isopentane (−80°C), or processed for paraffin embedding, or embedded in epoxy resin for electron microscopy. Cryosections and paraffin sections were stained with haematoxylin and eosin or van Giesson respectively.
Immunohistochemistry, western blotting and laminin-binding assay
Ten hours prior to histological examination, some animals were injected with Evans blue dye (EBD; Sigma-Aldrich, Steinheim) in order to detect fibres with sarcolemmal defects. For immunohistochemistry, 4 µm cryosections were probed with the following primary antibodies (Novocastra Laboratories, Newcastle upon Tyne): dystrophin (dys2), β-dystroglycan, (β-DG), γ-SG (γ-SG) and δ-SG (δ-SG). α-Dystroglycan was detected using either the monoclonal antibody VIA41 (Upstate Biotechnology, Lake Placid, NY) or a polyclonal sheep-immune serum (53), kindly provided by Stephan Kröger, University of Mainz; a rat monoclonal antibody to laminin was used for immunohistochemistry (33) (Immunotech, Marseille); caveolin-3 and dysferlin were detected using the monoclonal anti-caveolin-3 antibody (Transduction Laboratories, Lexington, KY) and the monoclonal antibody NCL-Hamlet (54) kindly provided by Louise Anderson, University of Newcastle, UK. For circumventing unspecific background staining, the Vector MOM immunodetection fluorescein kit (Vector Laboratories, Peterborough, UK) was used or mouse monoclonal antibodies were directly conjugated to Cy3 (Amersham Pharmacia Biotech, Little Chalfont, UK). Immunostained sections were analysed with a confocal laser scanning microscope (Fluoview, Olympus, Tokyo).
For western blotting, tissue extracts were separated on 4–12% Tris–glycine gels (Novex, Invitrogen, UK) and electrotransferred onto Hybond-P membranes (Amersham-Pharmacia Biotech, Uppsala) as described previously (7). Detection was performed with ECL plus (Amersham-Biosciences, Freiburg). For visualization and densitometric analysis, we used the VDSCL Bioimaging System and ImageQuant software (Amersham-Pharmacia Biotech). To assess laminin binding to α-DG, laminin overlay assays were performed as described in (36).
ERGs in responses to ganzfeld single-flash stimuli were measured in anaesthetized (100 mg/kg ketamine and 9 mg/kg xylazine intraperitoneally and proparacaine HCl 0.5% eye drops) C57BL/6J (n=6) and Largemyd (n=5) mice. The light stimuli (10 ms) were generated with an electronic shutter on an optical bench with a 100 W xenon source, and fibre optics were used to guide the flashes to the right eye while the animal was kept in a dark cage (maximum stimulus intensity 3.1 mW/cm2). The head of the mouse was kept in a head holder to impede breathing artefacts in the recording signal, and the animal was placed on a heating plate to maintain normal body temperature (38–39°C). A LabView-based system was used for shutter control and signal recordings (recording bandwidth 1 Hz–3 kHz). Unpaired t-tests were used for statistical evaluations, and P-values <0.05 were considered significant.
The authors are grateful to Gabriele Schaden and Marianne Leiszer for excellent technical assistance. This work was supported by the Austrian Muscle Research Fund (P.J.H. and R.E.B.) and by the University of Nottingham Research Support Fund and the Wellcome Trust to J.E.H.
To whom correspondence should be addressed. Tel: +43 1427761182; Fax: +43 1427761198; Email: email@example.com
|a-wave amplitude||a-wave implicit time||b-wave amplitude||b/a amplitude ratio||b-wave implicit time|
|a-wave amplitude||a-wave implicit time||b-wave amplitude||b/a amplitude ratio||b-wave implicit time|
Amplitudes and peak latencies of ERG a-waves and b-waves in response to single flash (10 ms) white light stimuli. The results are presented as the mean±standard error.