Faculty of Science, Department of Biological Sciences, Faculty of Medicine, University of Manitoba, Winnipeg, Canada
Department of Human Anatomy and Cell Science, Faculty of Medicine, University of Manitoba, Winnipeg, Canada
Correspondence to: Dr. Judy E. Anderson, Room 206, Biological Sciences Building, Faculty of Science, University of Manitoba, 50 Sifton Rd, Winnipeg, MB, Canada R3T 2N2. E-mail: Judy_Anderson@umanitoba.ca
Dystrophin deficiency from skeletal muscle causes the progressive, lethal weakness of Duchenne muscular dystrophy (DMD); it also affects the brain, resulting in variable cognitive dysfunction. Dystrophin linkage to proteins of the dystrophin-associated glycoprotein complex (DGC) in both muscle and brain tissues is established through interactions of the transmembrane protein β-dystroglycan (bDG) and the extracellular α-dystroglycan (aDG). In muscle, aDG binds to laminin in the extracellular matrix (ECM) (Campbell and Stull, 2003). In this way in skeletal muscle, the intracellular molecule, nitric oxide synthase 1 (NOS-1), links via syntrophin to the DGC (Blake et al., 2002; Perronnet and Vaillend, 2010; Ervasti, 2007) and fills a mechano-sensitive signaling function (Tatsumi, 2010; Wozniak et al., 2003;Wozniak and Anderson, 2007, 2009). The origin of the mental retardation that accompanies DMD in about a third of patients (Cotton et al., 2001) is not completely understood, although loss of dystrophin is particularly pronounced in cerebellar Purkinje neurons, where it is normally highly expressed (Lidov et al., 1990). Laminin is not present in the brain ECM and, by contrast, bDG binds to neurexin in neurons (Sugita et al., 2001).
Transmembrane bDG has a critical role in mediating the severity of various disease phenotypes in muscle and brain tissues. Proteomic studies show that aDG and bDG, translated from the same gene transcript (Ibraghimov-Beskrovnaya et al., 1992), are both reduced in the absence of dystrophin in muscle. Interestingly, only aDG has been reported to be reduced in dystrophin-deficient mouse brain (Finn and Ohlendieck, 1997; Culligan et al., 2001; Lewis and Ohlendieck, 2010; Lewis et al., 2009; Ibraghimov-Beskrovnaya et al., 1992). In adult muscle, bDG also binds rapsyn at nerve-muscle junctions and acts as a scaffold in clustering acetylcholine receptors (Haenggi and Fritschy, 2006). Although the critical function of bDG in linking the ECM and cytoskeleton appears to preclude survival from mutations to bDG itself, the glycosylation pattern of bDG, mutations to its binding and signaling-peptide domain, and other post-translational modifications, regulate its binding of dystrophin, and mutations in its glycosylation genes cause disease (Haenggi and Fritschy, 2006; Satz et al., 2010; Michele and Campbell, 2003). Now-classic literature suggested that the severity of dystrophy in brain was not correlated with that in muscle in DMD (Engel et al., 1994; Dubowitz, 1965). By comparison, two muscles in mdx mice that demonstrate only mild dystrophy (toe and some extraocular muscles) have only a modest decrease in bDG compared to other more severely affected muscles (Dowling et al., 2003).
NOS-1 also has a key role in brain and muscle function. Production of NO acts in synaptic signaling in the central nervous system (Leonelli et al., 2009; Garthwaite, 2008; Pierucci et al., 2011). Cerebellar long-term depression, which is impaired in the mdx brain, is NO-dependent (Ogasawara et al., 2008). In muscle, satellite cell activation, myoblast fusion, fiber regeneration, and the severity of muscular dystrophy are mediated by genetic, pharmacological, and activity-dependent changes in the level of NO in vivo and in vitro (Anderson, 2000; Mizunoya et al., 2011; Pisconti et al., 2006). Displacement or down-regulation of NOS-1 in dystrophin deficiency (Brenman et al., 1995; Grozdanovic and Baumgarten, 1999) and NOS-1 knockout transgenic mice (Percival et al., 2008) perturbs key processes that regulate muscle stem cell activation by NO (Wozniak et al., 2003; Wozniak and Anderson, 2007, 2009).
Tandem changes in NOS-1 and bDG are also important. Recently, both NOS-1 and bDG were shown to be involved in the decreasing responsiveness of muscle stem cells to stretch-activation with aging; the two proteins changed in distribution and amount as fiber leakiness increased with aging or was decreased by NO-donor treatment (Leiter and Anderson, 2010; Leiter et al., 2012b; Mizunoya et al., 2011). In a study of sarcopenic female mice, an increase in the NOS-1/bDG ratio after treatment with a NO-donor drug boosted the benefit of exercise and promoted muscle fiber growth and a 25% increase in muscle mass (Leiter et al., 2012b). Over-expression of NOS-1 also attenuated mdx mouse muscular dystrophy (Wehling et al., 2001), and a novel drug treatment that raises NO concentration decreased the severity of mdx mouse muscular dystrophy and promoted muscle regeneration (Mizunoya et al., 2011). Based on these and other studies, the potential of a treatment for DMD that promotes muscle repair via effects on NOS-1 activity and increased NO levels in muscle is under investigation (Brunelli et al., 2007; De Palma and Clementi, 2012; Sciorati et al., 2011; Percival et al., 2012).
The importance of the DGC and NOS-1 perturbations is well-established in the functioning of adult skeletal muscle and brain. As well, mutations in DGC proteins in skeletal muscle lead to a wide range of myopathies (e.g., Straub et al., 1997; Sampaolesi et al., 2003; Saito et al., 2005; Roberds et al., 1993; Kanagawa et al., 2005; Segalat et al., 2005; Lapidos et al., 2004). While regulation of the DGC is intensely scrutinized as a possible therapeutic target for the muscular dystrophies, much less is known about bDG and NOS-1 in developing muscle and brain or the effects of dystrophin deficiency during development. For example, it is unclear whether the genetic loss of dystrophin disrupts NOS-1 localization and expression early during muscle development or, alternatively, if that disruption occurs only after the onset of dystrophic damage in a muscle. This timeline is important in understanding the functional impact of a mutation that affects skeletal muscle and brain, where the DGC cytoskeleton differs; unlike the isoforms in muscle, the gamma isoforms of brain syntrophin found in Purkinje neurons are not anchored to the membrane by dystrophin (Alessi et al., 2006). For that reason, study of transmembrane bDG by systematic sampling of brain and muscle together may reveal the basis of the differential impact of dystrophin deficiency in mediating disease severity and determining the age of onset of the dystrophic phenotype during the functional changes in development.
The current study of mouse brain and skeletal muscle during development was designed to provide new information on the two tissues in parallel, as a foundation for further investigation, and to complement ongoing research on the basis of cognitive impairments in DMD and the muscle-specific severity of the disease (Snow et al., 2013a, 2013b). We compared two brain regions from control and mdx mice, the cerebellum and forebrain, and two muscles, the diaphragm with more severe dystrophy and the quadriceps with more moderate disease (Anderson et al., 1998; Stedman et al., 1991; Mizunoya et al., 2011) over the time-course of development when alterations in NOS-1 and bDG may have the largest impact on, or result from, functional changes during maturation.
Muscle Histology and Protein Localization
Quadriceps and diaphragm muscles displayed the classical patterns typical of muscle development in control mice between E18 and P28 and the development of dystrophy in mdx mice (Fig. 1 illustrates quadriceps development). In control muscles, small centrally nucleated fibers were loosely grouped into fascicles at E18, and from P1 to P28 were increasingly differentiated and had a greater frequency of peripheral myonuclei. Muscles from dystrophic mice showed the same pattern of development from E18 to P10. Inflammation, fiber degeneration, and regenerating fibers (marked by central nuclei) were only present at P28 in mdx muscles, and were especially notable in diaphragm, as expected. Fiber permeability to EBD was used as a further index of active dystrophy. Staining was absent from all E18 muscles and embryos (EBD did not cross the placenta), all the control muscles from P1–P28, and all post-natal dystrophic muscles at P1 and P10. At P28, however, 9 ± 0.85% of sampled quadriceps fibers were EBD+. In diaphragm muscle, 17 ± 8.8% of sampled fibers were EBD+ (range: 1.6–33.5%).
Strain-related differences in fiber growth with age and development were studied by measuring fiber diameter in sections of quadriceps muscle (Fig. 2). While fiber diameter increased with age in both strains (P < 0.0001), the mean and distribution of fiber diameter in dystrophic quadriceps were both lower than in controls at P28 (age X strain interaction, P = 0.04; P < 0.0001 [data not shown], respectively).
Nuclear position in fibers was used as an index of fiber maturation, and in mdx mouse muscles as an indicator of the progression of dystrophy. The central nucleation index (CNI0, a marker of immature developing fibers and regenerating fibers, decreased with increasing age (P < 0.0001, Fig. 2) and was higher in dystrophic than control muscles at every age (P < 0.0001), suggesting a delay in maturation in mdx muscles, especially the diaphragm. CNI was also higher in mdx mouse diaphragm than quadriceps (strain X muscle interaction, P < 0.0001). In mdx quadriceps, CNI had decreased at P10 to nearly zero and increased again by P28 as a result of regeneration following dystrophic damage to fibers. By comparison in mdx diaphragm, CNI decreased only to approximately 0.35 at P10. Interestingly, CNI in control mice at E18 was higher in quadriceps than diaphragm (0.9 vs. 0.5, respectively) (P < 0.001), whereas CNI in E18 dystrophic mice was high in both muscles (approximately 0.9). Satellite cell proliferation, identified at 400× as BrdU+ nuclei in flattened cells directly adjacent to the outline of non-degenerating fibers, decreased with increasing developmental age in both control and dystrophic quadriceps (P < 0.001, Fig. 2). Proliferation was higher in dystrophic than control muscles at P1 and P28 (P < 0.05, P < 0.01).
The localization of bDG and NOS-1 also changed with developmental age and dystrophy (Fig. 3A–C). bDG was localized at the fiber membrane in control diaphragm and quadriceps muscles, initially at E18 in a discontinuous, fine-scale patchy distribution that by P1 (diaphragm) and P10 (quadriceps) formed a smooth continuous outline around fibers. NOS-1 was distributed within the sarcoplasm of control muscles at E18 and by P10 was mainly localized at the periphery of fibers. In dystrophic muscles, while bDG was increasingly localized at the periphery of fibers from E18 to P1, the distribution was in uneven patches at the fiber membrane from E18 to P28. NOS-1 protein was typically localized inside the membrane of dystrophic fibers at all ages rather than primarily at the fiber periphery (observed more frequently in control muscle); this was especially notable at P28. Regenerated fibers with centrally located nuclei also displayed uneven peripheral bDG and largely sarcoplasmic NOS-1.
NOS-1 and bDG Levels in Muscle
The concentration of both NOS-1 and bDG proteins changed with age and in muscular dystrophy (Fig. 4). NOS-1 showed tissue-specific changes in significant interactions between strain and developmental age (P < 0.01). In the diaphragm, NOS-1 was generally lower in dystrophic than control mice (strain X tissue interaction, P < 0.001) and decreased with age (age X strain interaction, P < 0.001). NOS-1 was significantly lower in dystrophic than control diaphragm early in development at E18 and P1 (P < 0.001).
bDG levels decreased from P1 to P28 (e.g., at P28, P < 0.0001 vs. E18–P10) and were lower in dystrophic than control diaphragm muscle (P < 0.0001). bDG was greater in diaphragm than quadriceps (P < 0.0001), and bDG was higher in control than dystrophic diaphragm (age X strain interaction, P < 0.0001). As well, bDG changes with age varied between muscles; in the diaphragm, bDG was lower at P10 and P28 than at E18, and highest at P1 (age X tissue interaction, P < 0.0001) whereas it did not change with developmental age in quadriceps.
The NOS-1/ bDG ratio, previously of interest in relation to the response to exercise in age-related muscle atrophy after NO-donor treatment (Leiter et al., 2012a), was higher in P10 and P28 mdx diaphragm than in age-matched control diaphragm (P = 0.03). The ratio was also higher at P28 than at younger ages in quadriceps (age X tissue interaction, P < 0.0001) in both mdx and control mice, higher in mdx mouse diaphragm at P10 and P28 than earlier in development at E18 and P1 (P < 0.05), and approximately 10-fold higher in quadriceps than diaphragm (P < 0.0001).
Brain Histology and Protein Localization
Sections of brain tissues (forebrain and cerebellum) stained with H&E both showed the typical histological patterning of external molecular layer, Purkinje neuron monolayer, and inner granule cell layer in developing and mature control and mdx mice; this is illustrated for P28 (Fig. 5A). bDG staining was punctate along the Purkinje neuron membrane, with intense labeling of vascular smooth muscle, similar to the pattern seen for dystrophin (Snow et al., 2013b). In contrast, NOS-1 staining was widespread within neurons throughout sections of cerebellum, and especially prominent in Purkinje neurons (Fig. 5B,C).
NOS-1 and bDG in Brain
The levels of NOS-1 in control forebrain and in both control and mdx mouse cerebellum increased significantly with developmental age (P < 0.0001) (Fig. 6). For control mice, NOS-1 was higher in brain tissues (cerebellum > forebrain) than in muscle (P < 0.0001, compare with Fig. 4). By comparison, bDG in the control forebrain had the highest level at P10 and decreased by P28. In the cerebellum, bDG was higher at P10 and P28 than at earlier time points in controls.
The time-dependent levels of both NOS-1 and bDG were significantly affected by the loss of dystrophin in mdx mouse brain tissues. Dystrophic forebrain had markedly lower NOS-1 and bDG than controls (P < 0.0001 for both comparisons). In the cerebellum, NOS-1 differed between mdx and control tissues (age X strain interaction, P < 0.01), and was higher in mdx at P1 and lower in mdx mice at P10, compared to age-matched controls (P < 0.05, P < 0.01 respectively). bDG was higher in dystrophic than control cerebellum at P10 (P < 0.01).
The NOS-1/ bDG ratio in the forebrain was highest at P28 compared to the ratio in younger mice (P < 0.0001); in the cerebellum, the ratio was higher in controls at P28 than at E18 and P1 (P < 0.01). However, the NOS-1/bDG ratio for mdx mouse cerebellum was significantly lower at P10 compared to the levels in cerebellum in mdx mice at E18, P1, and P28 (P < 0.01). Forebrain and quadriceps had a similar pattern of changes in NOS-1/bDG ratio during development, which was approximately 10-fold higher in brain than diaphragm.
The study of NOS-1 and bDG in developing skeletal muscle and brain demonstrated tissue-specific changes related to age and the presence or absence of dystrophin. Localization patterns and fiber differentiation in control muscle matured earlier in diaphragm than quadriceps. The distinct developmental patterns in dystrophic mice illustrated that DGC maturation is muscle-specific and dystrophin-dependent. Differential changes in protein level and localization between brain and muscle were expected, due to the distinctive composition of the DGC in the two tissues (reviewed in (Snow et al., 2013a), and provide insights into the maintenance of membrane stability in non-contractile versus contractile cells. Interestingly, modifications in the localization of bDG and NOS-1 preceded the hallmarks of dystrophy, indicating an epigenetic influence on DGC development that implicates the impact of physiological demands during maturation.
The earlier changes to bDG and NOS-1 in diaphragm than quadriceps could be explained due to differential fiber loading in the two muscles during development (i.e., from prenatal to perinatal and then post-natal conditions) that would affect the diaphragm earlier than limb muscle. This would thereby account for correspondingly earlier perturbations in these two DGC proteins in diaphragm versus limb muscles in the absence of dystrophin. Locomotion matures after regular breathing movements as peripheral and central nervous system pathways develop post-natally while limb muscles come under voluntary control and are coordinated. Initial rhythmic leg movements shift to crawling before transitioning from plantar to digital stance with the onset of full weight-bearing (Clarac et al., 2004). By comparison, the diaphragm supports prenatal changes in intra-thoracic pressure and responds instantaneously to increased contraction with the onset of air breathing at birth. Architecturally (Anderson et al., 1998; Laskowski and Sanes, 1987), the diaphragm is more likely to be modified for fiber stability within the ECM, as demonstrated by earlier onset of pathology in dystrophin-deficient diaphragm than limb muscle (Stedman et al., 1991). Developmental and muscle-specific differences in ECM proteins are also reported between small antagonist flexors and extensors in the rat foot (Vinay et al., 2000). Results of the present study differ from a reported fiber misalignment and reduced fiber density during development of unspecified mdx muscles from E15.5–E17.5 (Merrick et al., 2009), likely due to our focus on particular muscles and later time points. In the absence of dystrophin from conception, muscle-specific and time-dependent changes in bDG and NOS-1 are interpreted to result from non-genetic influences that occurred during development, epigenetic changes that likely contribute to differential severity of the dystrophic phenotype (Ramaswamy et al., 2011).
At P28, the dystrophic phenotype was notably more severe in the diaphragm, as demonstrated by higher numbers of EBD+ fibers and higher CNI than in quadriceps. Shown here for the first time, developmental changes in fiber CNI related to maturation, the localization of NOS-1 and bDG, and the amounts and ratio of the two proteins were also earlier in diaphragm than quadriceps. These findings illustrate that loss of dystrophin affects development more broadly than mediating the capacity to sustain membrane stability in the face of active contractile forces and shear between the internal cytoskeleton and the ECM. Dystrophin deficiency is known to perturb signaling between fibers and adjacent satellite cells that overlie the NOS-1 source of NO inside the membrane (Anderson, 2000). Higher satellite cell proliferation at P1 in dystrophic versus control quadriceps may be “normalized” (i.e., decreased) by treatment to raise NO concentration in dystrophic muscle, since L-arginine treatment of mdx mice upregulated expression and activity of NOS-1μ and expression of CAPON, a NOS-1 anchor protein in muscle (Anderson and Vargas 2003; Segalat et al., 2005) with significant long-term functional benefit (Archer et al., 2006). NO-donor drugs, including isosorbide dinitrate and a new NO-donor, guaifenesin dinitrate (Wang et al., 2009), may also up-regulate NOS-1 and bDG specifically in diaphragm, as both improved the outcome of prednisone treatment in mdx mice sufficient to decrease disease severity (measured by EBD infiltration into fibers) and accelerated diaphragm fiber regeneration (Mizunoya et al., 2011).
In the present study, particular changes with development were interesting, whereas others were puzzling. The level of bDG was higher in control than dystrophic quadriceps at P10, and consistently higher in control than mdx diaphragm from E18 to P28; this indicates that muscle-specific regulation of DGC development is modified by the absence of dystrophin. bDG was also higher at E18 and P1 in control diaphragm than quadriceps, suggesting that fiber membranes need to be more resilient to mechanical stress, particularly in perinatal stages in diaphragm than quadriceps. This possibility is consistent with the observation that bDG decreased with age in control quadriceps but not diaphragm. Unexpectedly, there were no significant age-related differences in NOS-1 levels between control and dystrophic quadriceps; NOS-1 in P28 quadriceps only tended to be lower in mdx than control mice (P = 0.06). The level of NOS-1 in DMD muscle has been reported to increase in comparison to normal human fibers, in conjunction with its displacement from the DGC at the sarcolemma, particularly in early regenerating segments of muscle containing developmental myosins (Punkt et al., 2006, 2007). The majority of reports, however, indicate NOS-1 is down-regulated at the transcriptional level, in correlation with the absence of dystrophin (Brenman et al., 1995; Grozdanovic and Baumgarten 1999; Percival et al., 2008) and in other conditions, possibly by unknown neurogenic features (Grozdanovic et al., 1997). NOS-1 localization at the sarcolemma is suggested as an index of DGC restoration after treatment to restore dystrophin in mdx mouse muscle (Wells et al., 2003) and dystrophin recruits NOS-1 to the sarcolemma (Li et al., 2010). Therefore, decreased NOS-1 in dystrophic diaphragm and its displacement to the cytoplasm during regeneration in mdx muscle are highly consistent with the reported down-regulation of NOS-1 in human DMD.
The developmental profiles of NOS-1 and bDG in brain were quite distinct from muscles. The NOS/bDG ratio was highest at P28 in forebrain versus cerebellum and quadriceps versus diaphragm. This may result from an age-dependent decrease in bDG, noted here and previously (Henion et al., 2003), rather than changes in NOS-1 alone. Whereas NOS-1 levels increased in the cerebellum and forebrain with developmental age, and were highest at P28, NOS-1 was lower in mdx forebrain and diaphragm compared to controls, which varies from other reports (Sillitoe et al., 2003; Deng et al., 2009), likely due to different sampling of forebrain tissues. However, in cerebellum, maintenance of NOS-1 in the brain despite the absence of dystrophin may occur on multiple levels. Notably, Purkinje neurons are one of many sources of NOS-1 in cerebellum. Secondly, cytosolic NOS-1 in Purkinje neurons relates to localization of its binding partner, the gamma isoform of syntrophin, in the endoplasmic reticulum; this is quite distinct from the cytoskeletal localization in muscle where alpha- and beta-syntrophin interact strongly with dystrophin (Alessi et al., 2006). For these reasons, cognitive dysfunction in dystrophin-deficient dystrophy is likely independent of NO signaling or NOS-1 interactions with syntrophin, even in the absence of long and all shorter forms of dystrophin, as seen in the mdx3cv dystrophic mouse (Moukhles and Carbonetto, 2001).
While the unique binding properties of gamma-syntrophin differentiate it from the muscle isoforms of syntrophin that use PDZ domains to link with NOS-1, the DGC still may be involved in the impact of dystrophin deficiency in the brain. Alpha-DG, localized in glial end-feet at the ECM, is involved in glial function, neuronal migration, and forebrain development (Satz et al., 2010). Alpha-DG mutation was recently reported to cause a limb-girdle type of muscular dystrophy (Hara et al., 2011). DG at neuronal post-synaptic membranes and perivascular end-feet of forebrain astrocytes (Culligan et al., 2001; Tian et al., 1996) is involved in synaptic plasticity as it stabilizes synapse formation and axodendritic contacts during development (Ferreira and Paganoni, 2002). In contrast to our findings, others report no reduction in brain bDG levels in mdx mice, whereas bDG is reduced in the forebrain in human DMD (Finn et al., 1998). Unlike the case in muscle, the neuronal DGC has not been isolated as a complete complex (Waite et al., 2012). Thus, knowledge of precise interactions among DGC proteins in neurons is limited, and the impact of decreased bDG in the developing mdx forebrain has yet to be determined.
Given their noted involvement in neuronal function (i.e., synaptic plasticity), alterations in DGC components in the absence of dystrophin, including bDG, are assumed to play a role in the cognitive deficits reported in DMD. However, there is considerable heterogeneity in the cognitive phenotype of DMD (as reviewed in Snow et al., 2013a); intelligence quotient (IQ) scores are normally distributed, fall one standard deviation below mean scores, and about one-third of those with DMD exhibits intellectual disability (Cotton et al., 2001) . The mutation site of the dystrophin gene, which encodes multiple proteins, has been correlated with the extent of cognitive impairment in DMD. Those in the more distal region of sequence encoding shorter isoforms for Dp140 (Felisari et al., 2000; Moizard et al., 1998; Taylor et al., 2010; Wingeier et al., 2011) and Dp71 (Moizard et al., 1998, 2000), both present in glial cells (Pilgram et al., 2010), are associated with more severe cognitive dysfunction in DMD. Mutations in the promoter region encoding brain-type full-length dystrophin, in contrast, are compatible with a normal IQ range (den Dunnen et al., 1991; Rapaport et al., 1992). Careful studies using neuropsychological tests revealed specific cognitive deficits in DMD, irrespective of general intelligence, including deficits in limited verbal memory (Hinton et al., 2000). In the mdx model of DMD, in which only full-length neuronal dystrophin is absent, mice are impaired in passive avoidance learning (Muntoni et al., 1991) and spatial memory retention (Vaillend et al., 2004). On this background, results of the present study are interpreted to suggest that alterations in DGC components such as bDG subsequent to gene mutations affecting neuronal dystrophin may contribute to development of cognitive deficits.
We recently found that the distribution of dystrophin puncta on membranes of control Purkinje neurons in the lateral cerebellum was greater than in the vermal region (Snow et al., 2013b); this was associated with abolition of regional differences in membrane electrophysiology in mdx mice: regional hyperpolarization and a lower frequency of spontaneous action potentials in vermal versus lateral cerebellar Purkinje neurons in control mice (Snow et al., unpublished data). Dystrophin-dependent changes in intrinsic electrical properties of the lateral cerebellum extended the idea that cerebellar dysfunction contributes to impaired cognitive function in DMD (Cyrulnik and Hinton 2008; Cyrulnik et al., 2007; Hinton et al., 2009; Snow et al., 2013a) and implicated DGC proteins in modulating membrane properties that regulate neuronal excitability.
In summary, the results of this developmental study to examine protein localization and levels indicate that the severity of dystrophy in muscle is related, at least in chronology, to the displacement of NOS-1 from linkage with the DGC. Dystrophy was accompanied by a lower level of NOS-1, a lower level of bDG and at P10 and P28, and a higher ratio of NOS/bDG in the most severely affected diaphragm muscle. Findings suggest that a higher bDG concentration contributes to membrane resilience to contraction during the developmental progression of increasing functional demands on skeletal muscles.
Animals and Tissues
Animals were housed and treated according to the Canadian Council on Animal Care in the University of Manitoba (protocol F07-008). Tissues from control wild-type and the dystrophic mdx mutant strains of C57BL/10ScSn (n≥4 per group) were collected at embryonic day 18 (E18) after timed matings, and post-natal (P) days P1, P10, and P28. The E18 time was chosen to sample tissues in utero after formation of muscle anlages, when the diaphragm would be physically loaded by pre-natal breathing movements (Anderson et al., 1993; Scott et al., 1994), whereas quadriceps would be unloaded. P1 was selected as a time when diaphragm loading would have just increased by the demand to move air immediately post-natally, and the quadriceps would be only modestly loaded in neonates. At P10, both muscles would be loaded, while at P28, functional locomotion and respiration would have fully matured capacity. Further, the sampling of ages was chosen based on findings that mdx mice do not display evidence of significant muscle pathology or degradation in limb muscles until P21 (Pastoret and Sebille, 1995); thus, the post-natal days chosen ensured data were gathered before and after the presence of typical muscle pathology to evaluate the changes in bDG and NOS-1 with disease progression. Mice (or dams) were injected with Evans blue dye (EBD, 100 mg/kg subcut.) 24 hr before euthanasia to identify the loss of sarcolemmal integrity as an indicator of active dystrophy (Archer et al., 2006). For the study of muscle cell proliferation in quadriceps development, mice were injected with bromodeoxyuridine (BrdU; 100 mg/kg i.p.) to label DNA synthesis, 2 hr before euthanasia by cervical dislocation under isoflurane anaesthesia (Baxter Corporation, Mississauga, ON).
The brain was dissected rapidly from the calvarium and dissected into forebrain (olfactory bulbs removed) and cerebellum, and bisected down the mid-sagittal sulcus and vermis, respectively. Half of each region was collected into an Eppendorf tube, snap frozen in liquid nitrogen, and stored at −80°C until Western blot analysis. The other half was fixed for histology in fresh 4% paraformaldehyde overnight at 4°C and transferred to 30% sucrose (in 0.1 M phosphate buffer with 0.01% sodium azide) for 3–7 days at 4° (until it sank), prior to embedding in Cryomatrix (Thermo Scientific, Waltham, MA) and freezing in isopentane cooled below −50°C. Cryosections (8 μm thick) were collected on glass slides and stored at −20°C until staining in hematoxylin and eosin (H&E) or processing for immune detection of NOS-1 and bDG proteins.
Quadriceps and diaphragm muscles were rapidly dissected from the same mice from which brain tissues were collected; each muscle was bisected. The proximal quadriceps and right diaphragm were placed in Eppendorf tubes and frozen for Western blot analysis. The distal quadriceps and left diaphragm were prepared for cryosectioning and histological studies.
All assays of brain and muscle tissues were conducted on coded samples, without knowledge of source group; full sets of brain and muscle samples were each processed and analyzed together, to minimize variation.
The location of myonuclei was quantified to track the developmental shift from internally located myonuclei (referred to as central nuclei) in early or primary fibers to a peripheral location in differentiated fibers. Observations at P28 in sections of dystrophic muscle also tracked the appearance of regenerated fibers with central nuclei following earlier segmental damage due to disease progression. The central nucleation index (CNI) was calculated as the proportion of sampled fibers (100–300 per section, scanned across its longest chord) that displayed central nucleation. Fields with muscle-tendon junctions were avoided, as they have noticeably more central nuclei than other areas of the muscle, even in normal, non-dystrophic mature muscle (Wang et al., 2010).
Satellite cell proliferation in the quadriceps muscle was assessed by counting the number of BrdU-positive nuclei in cells in the satellite position, juxtaposed to fibers. BrdU was detected as previously reported (Mizunoya et al., 2011) using anti-BrdU (mouse monoclonal, Roche Diagnostics, Indianapolis, IN; diluted 1:50) and biotin-conjugated AffiniPure fab Fragment goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA; 1:200) and the diaminobenzidiene method in conjunction with streptavidin-conjugated horseradish peroxidase (Sigma, St. Louis, MO). Darkly stained nuclei in a satellite position on non-degenerating fibers (100–300 per section, scanned across its longest chord) and the total number of such fibers within that field (to exclude the possibility of labeled non-myogenic cells surrounding or invading degenerating fibers) were counted at 400× without knowledge of source group.
Infiltration of Evans Blue Dye (EBD) into fibers was used to assess the extent of membrane damage from active dystrophy in the 24 hr prior to euthanasia (Hamer et al., 2002). Sections were immunostained according to previous reports (Mizunoya et al., 2011) for laminin by incubating in primary rabbit anti-laminin antibody (diluted 1:100, Sigma L9393) and Alexa488-conjugated AffiniPure Fab fragment of anti-rabbit IgG (diluted 1:200, Jackson ImmunoResearch Laboratories, Inc.). Five non-overlapping images of each section were captured through a 20× objective using a Zeiss Axio Imager Z1 (Carl Zeiss Canada Ltd.). Bright red EBD fluorescence within a green-laminin outline indicated a fiber positive for EBD. EBD+ fibers in cross-section were counted and expressed as a proportion (%) of total fibers counted (100–300 per section).
Double immunostaining was used to visualize the localization of NOS-1 and β-dystroglycan using the IHCWorld protocol “parallel approach” (IHC World Protocols, 2008) as reported (Leiter et al., 2012a). Briefly, muscle sections were fixed in 4% paraformaldehyde. All sections were processed by antigen retrieval in citrate buffer, cooled to room temperature (RT) and blocked with a solution containing normal goat blocking serum, 2M glycine, the Fab fragment of goat anti-rabbit unconjugated antibody, and unconjugated goat anti-mouse antiserum (at ratios of 20:1:1, 1 hr at RT). After blocking, sections were incubated in a mixture of two primary antibodies, rabbit polyclonal anti-NOS-1 (sc-1025, Santa Cruz Biotechnology, Santa Cruz, CA; 1:50) and mouse monoclonal anti-bDG (NCL-43DAG, Novocastra, Bannockburn, IL; 1:100) overnight at 4°C. Negative control slides did not receive primary antibody. After washing in Tris-buffered saline with Tween (TBST, 0.5M Tris Base and 9% NaCl in water with 0.5% Tween 20), sections were incubated for 1 hr at RT with a solution containing two secondary antibodies, Alexa488-conjugated Fab fragment of goat anti-rabbit and DyLight649-conjugated Fab fragment of goat anti-mouse (Jackson ImmunoResearch Laboratories, Inc.; 1:200). After rinsing, sections were mounted under coverslips with Vectashield (Vector Laboratories, Inc., Burlingame, CA). Sections were viewed and photographed with a Zeiss Axio Imager Z1 using the 40× oil immersion objective.
NOS-1 and bDG Levels
The levels of NOS-1and bDG were assayed in the same protein samples collected from every animal, using Western blotting as previously reported (Leiter et al., 2012a; Mizunoya et al., 2011). Membranes were divided into three parts by cutting across the 75- and 25-kDa marker bands. The portion above 75 kDa was probed for NOS-1 (155 kDa); the portion from 25–75 kDa was probed first for bDG (43 kDa), then stripped and probed for actin (42 kDa) as a presumptive loading control. For blots from muscle tissues, the portion below 25 kDa was also probed for GAPDH, as a second presumptive loading control. Gels and membranes used to assay a particular protein in either muscles or brain tissues were processed in sets to minimize variation.
Membranes were blocked with 5% skimmed milk in phosphate-buffered saline with 0.5% Tween 20 for 1 hr and incubated overnight at 4°C with one of the following primary antibodies: rabbit polyclonal anti-NOS-1 (sc-1025, Santa Cruz Biotechnology, 1:200); anti-bDG (NCL-43DAG, Novocastra; Leica Microsystems, Richmond Hill, ON, 1:200); mouse anti-actin (MAB1501, Millipore, Bedford, MA, 1:200); or rabbit anti-GAPDH (Santa Cruz Biotechnology, sc-25778, 1:200). Membranes were washed and incubated with the appropriate secondary antibody (donkey anti-rabbit IgG; NA9340V, GE Healthcare, Little Chalfoot Buckinghampshire, UK) or goat anti-mouse IgG (A2304 at 1:5000, Sigma Aldrich) for 1 hr at room temperature. Bands were visualized using chemiluminescence according to the manufacturer (Santa Cruz Biotechnology, sc-2048). The optical density (OD) of bands (minus background) was measured using NIH-Image J software. Levels of both actin and GAPDH varied significantly with age (data not shown), were often more abundant than bDG and NOS-1 in extracts of brain and muscle, and varied between control and dystrophic tissues. For these reasons, protein levels are reported relative to total protein determined by Ponceau Red staining (Romero-Calvo et al., 2010; Aldridge et al., 2008; Dittmer and Dittmer, 2006).
Data were compiled in Microsoft Excel spreadsheets and analyzed using JMP-SAS statistical software by 3-way analysis of variance (ANOVA) to determine the main effects of age, dystrophy, tissue, and interactions among those effects. Pair-wise differences between means were determined using post hoc Tukey's HSD tests. A probability of P < 0.05 indicated significance.
Additional support from NSERC for a Doctoral Scholarship award (W.S.) and NSERC Undergraduate Summer Studentships (A.J.), and the technical assistance of S. Mohoric during a Summer Research Fellowship from the Manitoba Institute of Child Health are gratefully acknowledged.