Absence of laminin α2 chain leads to a severe form of congenital muscular dystrophy (MDC1A) associated with peripheral neuropathy. Hence, future therapies should be aimed at alleviating both muscle and neurological dysfunctions. Pre-clinical studies in animal models have mainly focused on ameliorating the muscle phenotype. Here we show that transgenic expression of laminin α1 chain in muscles and the peripheral nervous system of laminin α2 chain deficient mice reduced muscular dystrophy and largely corrected the peripheral nerve defects. The presence of laminin α1 chain in the peripheral nervous system resulted in near-normal myelination, restored Schwann cell basement membranes and improved rotarod performance. In summary, we postulate that laminin α1 chain is an excellent substitute for laminin α2 chain in multiple tissues and suggest that treatment with laminin α1 chain may be beneficial for MDC1A in humans.
Laminins are heterotrimeric multidomain molecules composed of genetically distinct chains (α, β and γ) (1). Laminin α2 subunit is the major laminin α chain in basement membranes surrounding muscle fibers and Schwann cells. Mutations in the LAMA2 gene encoding laminin α2 chain cause congenital muscular dystrophy type 1A (MDC1A) (2). Lack of laminin α2 chain results in severe muscle weakness, hypotonia, joint contractures and white matter abnormalities with onset at birth or in young infancy (3,4). Moreover, laminin α2 chain deficiency affects the peripheral nervous system (PNS). MDC1A patients display delay of peripheral nerve conduction (5,6) as well as hypo- and hypermyelination and tomacula with uncompacted myelin (7,8). Also, mouse models for MDC1A, the dy/dy and dyw/dyw mice with partial laminin α2 chain deficiency, the dy3K/dy3K mouse, which is completely deficient in laminin α2 chain and the dy2J/dy2J mouse, which expresses an aberrant laminin α2 chain, display severe muscular dystrophy accompanied by prominent abnormalities in the PNS. Neuropathy is pronounced with the presence of bundles of unmyelinated axons in the proximal PNS, hypomyelination, paranodal abnormalities, reduced voltage-gated sodium channel clustering in the distal PNS and reduction in conduction velocity (9–20).
Considering that neuropathy is a significant part of MDC1A, future therapies aimed at alleviating the neurological dysfunction should be considered. The muscular dystrophy phenotype has been ameliorated by various means in mouse models for MDC1A, but so far improvements of the neurological phenotype have not been reported or only been described very preliminarily (18,21–25). Several attempts to overexpress laminin α2 chain in the PNS have failed (21) and neuropathy is not prevented by muscle-specific expression of laminin α2 chain (18). Moreover, although the agrin minigene ameliorates muscle abnormalities in the dyw/dyw mouse model (22), expression of miniagrin in peripheral nerves driven by adeno-associated vector delivery does not prevent the nerve defects (23). Further, inactivation of the proapoptosis protein Bax in dyw/dyw mice is reported to be beneficial for the condition of motor neurons, but no detailed analyses support these data (24).
We previously demonstrated that transgenically expressed laminin α1 reduced muscular dystrophy and restored fertility in laminin α2 chain deficient mice (25,26). In this transgenic line, laminin α1 chain was poorly expressed in the sciatic nerve, and the neurological phenotype was thus not improved (25). Here, we have generated laminin α2 chain deficient mice expressing laminin α1 chain in the PNS and skeletal muscles. Dy3K/dy3K mice harboring the laminin α1 transgene in these tissues (dy3KLNα1TG-8) displayed reduced muscular dystrophy. In addition, reduced motor dysfunction symptoms and reversed histopathological features as assessed by rotarod test and analyses of myelin thickness and axon diameter, respectively, were seen in this mouse model. Moreover, compensatory changes of expression of laminin α4, β2 and γ2 chains were normalized and Schwann cell basement membranes were partially restored in the PNS upon overexpression of laminin α1 subunit. In conclusion, these findings demonstrate that laminin α1 chain is an excellent replacement for laminin α2 chain in multiple tissues in mice and suggest that treatment with laminin α1 chain may be beneficial for MDC1A in humans.
Dy3K/dy3K mice overexpressing laminin α1 chain in the PNS and muscles
Laminin α1 chain is not present in the neuromuscular system of wild-type animals (25). To achieve expression of laminin α1 chain in various tissues, we created transgenic mice overexpressing laminin α1 chain under the control of the CMV enhancer and β-actin promoter (25). Transgenic line no. 12 with high expression of laminin α1 chain in skeletal muscles, but not in the sciatic nerve, was previously described (25). However, transgenic mice derived from line no. 8 expressed laminin α1 subunit in both skeletal muscles and the PNS. In muscle, laminin α1 chain was present at the sarcolemma, myotendinous and neuromuscular junctions, as previously described for transgenic mice from line no. 12 (Fig. 1A), although at a lower level (∼2-fold) (Fig. 1B). Laminin α1 chain immunoreactivity in the PNS was localized to endoneurium but not to perineurium of transgenic sciatic nerve (Fig. 1A).
Animals from both lines were further used to produce dy3K/dy3K mice overexpressing laminin α1 chain (dy3KLNα1TG-12 and dy3KLNα1TG-8). Dy3KLNα1TG-12 transgenic animals have been shown to be long-lived with remarkably reduced dystrophic symptoms in skeletal muscle (25). Forty-two dy3KLNα1TG-8 animals were produced, and among those, six were more affected, possibly indicating that the inadequate phenotype correction might be influenced by individual variability. However, those that were not sacrificed for experimental procedures have remained in good health. Dy3KLNα1TG-8 animals have a near-normal lifespan, the average body weight is not significantly different from wild-type mice (Fig. 1C, data not shown) and the dystrophic phenotype is largely corrected. Histological analyses revealed slightly more advanced myopathy in 2-month-old quadriceps femoris and triceps brachii muscles of dy3KLNα1TG-8 compared with dy3KLNα1TG-12 mice, with areas containing centrally nucleated fibers (Fig. 1D, data not shown). Quadriceps femoris muscle from dy3K/dy3K mice showed 24% fibers with centrally located nuclei, whereas quadriceps femoris muscles from dy3KLNα1TG-12 and dy3KLNα1TG-8 mice showed 12 and 19% fibers, respectively, with centrally located nuclei (Fig. 1E). Moreover, dy3K/dy3K muscles display pathological fibrous tissue (16,25). Notably, fibrosis occurred very rarely in dy3KLNα1TG-12 muscles, as previously described (25), and occurred rarely in dy3KLNα1TG-8 muscles, as evaluated by hematoxylin and eosin staining and tenascin-C immunostaining (Fig. 1D, data not shown).
A common feature of dy3K/dy3K mice is that they flex their hind legs to the trunk when lifted by the tail and so do often dy3KLNα1TG-12 animals. We observed that transgenic mice from line no. 8 seldom flex their hind limbs. Moreover, they displayed less severe leg paralysis than dy3K/dy3K and dy3KLNα1TG-12 mice. These first observations prompted us to analyze in detail whether laminin α1 chain reversed the development of peripheral nerve defects in dystrophic mice lacking laminin α2 chain. As expected, laminin α1 chain was expressed in endoneurium of the sciatic nerve from dy3KLNα1TG-8 animals (Fig. 3A). Surprisingly, we also detected weak laminin α1 chain expression in sciatic nerves of dy3KLNα1TG-12 animals (Fig. 3A) despite the lack of staining in sciatic nerves of dy3K/dy3K and LNα1TG-12 animals, respectively (Figs 1A and 3A). However, more laminin α1 chain was seen in sciatic nerves of dy3KLNα1TG-8 animals compared with dy3KLNα1TG-12 animals. Western blot analysis demonstrated ∼2-fold more laminin α1 chain in sciatic nerves of dy3KLNα1TG-8 animals compared with dy3KLNα1TG-12 animals. Very little, if any, laminin α1 chain was detected in wild-type sciatic nerve (Fig. 3B). Thus, the dy3KLNα1TG-8 mouse model is better for studying peripheral neuropathy and we hypothesized that peripheral nerve dysfunction would be prevented in dy3KLNα1TG-8 animals but not in dy3KLNα1TG-12 animals.
Laminin chains in the PNS of dy3K/dy3K and dy3KLNα1TG-8 mice
The loss of laminin α2 chain from various tissues and transgenic introduction of laminin α1 chain have been shown to influence the expression of other laminin subunits in different mouse models for MDC1A (19,22,25–28). Thus, we analyzed the expression of all known laminin chains in wild-type, dy3K/dy3K and dy3KLNα1TG-8 sciatic nerves. The expression pattern of laminin α1 chain is described earlier and presented in Figure 3A. Noteworthy, we found that laminin α1 chain is not present in sciatic nerve of dy3K/dy3K mice. This is in contrast to dy2J/dy2J animals, which express laminin α1 chain in sciatic nerve (28). As expected, laminin α2 chain was completely absent from dy3KLNα1TG-8 sciatic nerve (Fig. 2). Laminin α3 chain was confined to blood vessels, but it was also found in perineurium of wild-type, dy3K/dy3K and dy3KLNα1TG-8 mice (Fig. 2). It has previously been reported that laminin α4 chain is upregulated in peripheral nerves of dy/dy and dy3K/dy3K mice (19,27). We also detected a moderate upregulation of laminin α4 chain in endoneurium of dy3K/dy3K animals and normalized expression in dy3KLNα1TG-8 mice (Fig. 2). Laminin α5 staining was detected in vessels and perineurium and weakly in endoneurium in nerves of all investigated genotypes (Fig. 2). Laminin β1 chain was restricted to endoneurium in wild-type, dy3K/dy3K and dy3KLNα1TG-8 PNS (Fig. 2). Laminin β2 subunit has been shown to be reduced in dy3K/dy3K and normalized in dy3KLNα1TG-12 muscles (25). Interestingly, we observed the same phenomenon in the PNS. Laminin β2 chain was downregulated in endoneurial basement membranes in dy3K/dy3K mice and restored to wild-type levels in dy3KLNα1TG-8 sciatic nerve (Fig. 2). However, blood vessel and perineurium stainings appeared unchanged in the sciatic nerve of dy3K/dy3K and dy3KLNα1TG-8 animals (Fig. 2). Laminin β3 chain was absent from the PNS (Fig. 2). Both perineurial and endoneurial basement membranes were strongly stained with the antibody against laminin γ1 chain in wild-type, dy3K/dy3K and dy3KLNα1TG-8 animals (Fig. 2). Laminin γ2 subunit was concentrated in perineurium, and weak staining in endoneurium in wild-type sciatic nerve was also detected. Interestingly, γ2 chain was reduced in endoneurium in dy3K/dy3K PNS but not in dy3KLNα1TG-8 endoneurium (Fig. 2). These studies implicate the presence of hitherto undescribed laminin trimers composed of α1 and α2 chains associated with γ2 chain in endoneurium. Laminin γ3 chain was present in endoneurium of wild-type mice but absent from laminin α2 chain deficient sciatic nerve, and expression was not restored upon transgenic overexpression of laminin α1 subunit in dy3KLNα1TG-8 PNS (Fig. 2). Similarly, we have shown before that forced expression of laminin α1 chain in dy3K/dy3K testis does not lead to normalized expression of laminin γ3 chain (26).
Other basement membrane components such as perlecan and collagen IV appeared by immunofluorescence analyses to be normally expressed in dy3K/dy3K and dy3KLNα1TG-8 sciatic nerves (data not shown). Also, expression of major laminin receptors in the PNS (28) (integrin α6, integrin β1, integrin β4 and dystroglycan) remained unchanged (data not shown).
Corrected PNS morphology in dy3KLNα1TG-8 animals
Next, we analyzed the effect of laminin α1 overexpression on the morphological features of laminin α2 chain deficient sciatic nerve in 2- and 7-month-old dy3KLNα1TG-12 and dy3KLNα1TG-8 animals. Toluidine blue staining revealed occasional bundles of unmyelinated axons in 4-week-old dy3K/dy3K sciatic nerves (Fig. 3C). However, most of the axons were myelinated properly by Schwann cells in dy3K/dy3K animals (Fig. 3C). The areas with unsorted and unsheathed axon bundles were also seen in dy3KLNα1TG-12 sciatic nerves (Fig. 3C, arrows). In contrast, they were rarely detected in 2-month-old dy3KLNα1TG-8 sciatic nerves and their general appearance did not differ from the wild-type specimen (Fig. 3C). It has been suggested that reduced conduction velocity in dy3K/dy3K PNS might be due to small axon diameters and relatively thin myelin sheaths (19). Thus, we measured axon diameters in 2-month-old wild-type, dy3KLNα1TG-12 and dy3KLNα1TG-8 mice and in 4-week-old dy3K/dy3K animals (Fig. 3D). Although dy3KLNα1TG-8 axons remained significantly smaller than wild-type axons (P<0.0001), they were also significantly bigger than axons in dy3K/dy3K and dy3KLNα1TG-12 sciatic nerves (P<0.0001 and P=0.0224, respectively) (Fig. 3D). Moreover, in dy3K/dy3K and dy3KLNα1TG-12 sciatic nerves, thinner myelin was formed compared with dy3KLNα1TG-8 nerves (P<0.0001) (Fig. 3E). Further, we compared axon diameters and myelin thickness in 7-month-old wild-type, dy3KLNα1TG-12 and dy3KLNα1TG-8 mice. With age, the axon sizes were normalized in dy3KLNα1TG-8 sciatic nerve, and the axon diameters were not significantly different from wild-type axons (P=0.3201) (Fig. 3D). The role of laminin α1 chain in normalization of the axon growth is emphasized by the fact that the axons in dy3KLNα1TG-12 animals weakly expressing laminin α1 transgene in the PNS remained significantly smaller than dy3KLNα1TG-8 axons (P<0.0001) (Fig. 3D). Similar to 2-month-old animals, myelin sheath was thicker in dy3KLNα1TG-8 axons than in dy3KLNα1TG-12 axons (P=0.0012); however, it was thinner than in wild-type mice (P=0.0189) (Fig. 3E).
In summary, our data indicate that laminin α1 chain promotes myelination in vivo almost as efficiently as laminin α2 chain. Moreover, it also influences axon growth.
Aberrant myelination is also detected in the spinal root of dy3K/dy3K mice (19). Hence, we analyzed the expression of laminin α1 chain in spinal roots of wild-type, dy3K/dy3K and dy3KLNα1TG-8 animals. We detected very low expression of laminin α1 subunit in wild-type and dy3K/dy3K spinal roots (Fig. 4A). In contrast, laminin α1 chain was expressed at high levels in roots from dy3KLNα1TG-8 mice (Fig. 4A). Morphology analyses of dy3KLNα1TG-8 spinal roots confirmed that laminin α1 chain corrects the myelination defects also in laminin α2 chain deficient spinal roots. In five out of six analyzed transgenic animals, unmyelinated axon bundles were not found in the dorsal roots (Fig. 4B). Unmyelinated axons were not found in ventral roots of dy3KLNα1TG-8 animals either (data not shown). Thus, the phenotype of dy3KLNα1TG-8 spinal roots was remarkably corrected compared with roots from dy3K/dy3K mice, where extensive areas with sorting defects were found (Fig. 4B).
The improvement of myelination in dy3KLNα1TG-8 animals might be caused by the restoration of Schwann cell basement membranes, which have been shown to be perturbed, both in dy/dy and dy3K/dy3K PNSs (9,19). Consistent with Nakagawa et al. (19), we found by electron microscopy analyses that Schwann cell basement membranes are defective in dy3K/dy3K mice (Fig. 5A). Moreover, dy3KLNα1TG-12 animals, which only express low amounts of laminin α1 chain in PNS, also have discontinuous basement membranes (Fig. 5A). However, they were clearly restored around axons in dy3KLNα1TG-8 sciatic nerve (Fig. 5A), and patches of basement membranes were only occasionally found.
In addition, we studied the ultrastructure of the nodes of Ranvier, which are responsible for more rapid propagation of action potentials. It has been shown that dy/dy and dy2J/dy2J mice have aberrant nodes of Ranvier (10,20). In dy2J/dy2J nerves, the width of the node gap was found to be abnormally wide and Schwann cell microvilli were poorly formed (20). However, in dy3K/dy3K animals, the Ranvier node gaps were found to be narrower than in the wild-type littermates (19). Moreover, it has been reported that basement membranes were disrupted at the nodes of Ranvier in laminin α2 chain deficient mice, possibly causing the reduction in conduction velocity (19). Consistent with Nakagawa et al. (19), we found that the nodal gaps in dy3K/dy3K sciatic nerve were narrower than gaps of wild-type nerves (P=0.0158) (Fig. 5C). Also, the microvilli were perturbed and the nodal zone was deprived from basement membranes (Fig. 5B). In contrast, we found that microvilli were formed in four out of five analyzed nodes in dy3KLNα1TG-8 sciatic nerve (Fig. 5B, arrows) and the node gaps were neither wider nor narrower than in wild-type nodes (P=0.9505). Also, basement membranes were restored at all analyzed nodes in dy3KLNα1TG-8 sciatic nerve (Fig. 5B, arrow heads).
Dy3KLNα1TG-8 mice show improved rotarod performance
Finally, we analyzed by the rotarod test whether the neurological properties and motor function were improved in dy3KLNα1TG-8 mice. Dy3K/dy3K animals are very small and emaciated and presumably not able to perform the test. Instead, we analyzed dy3KLNα1TG-12 mice. These animals were not able to stay on the rotating rod as long as wild-type mice, in spite of substantial improvement of muscle phenotype (Fig. 6). Dy3KLNα1TG-8 animals, on the other hand, performed as well as wild-type mice, demonstrating that laminin α1 chain overexpression in the PNS is beneficial for motor function in the absence of laminin α2 chain (Fig. 6).
Gene targeting experiments have demonstrated crucial roles for laminins and their receptors in peripheral myelination (19,29–33). Laminin-211 (α2β1γ1) and -411 (α4β1γ1) are the major laminin isoforms in Schwann cell basement and loss of laminin α2 and α4 chain, respectively, resulting in amyelination of sciatic nerves (30). We here provide the first detailed evidence that laminin α2 chain deficient peripheral neuropathy largely can be corrected. Transgenically expressed laminin α1 chain promoted sorting and myelination in the complete absence of laminin α2 chain. It was previously hypothesized that ectopic expression of laminin α1 chain in sciatic nerves, but not in roots of dy2J/dy2J animals, may account for the less severe amyelination in sciatic nerves compared with roots (28). Indeed, our results significantly substantiate the hypothesis that laminin α1 chain may compensate for lack of normal laminin α2 chain function in the PNS. Considering preceding reports which demonstrated significantly improved survival, weight gain, fertility, muscle morphology, muscle function and, as demonstrated in this article, motor function, laminin α1 subunit (if expressed in sufficient amounts) appears to be an ideal candidate for replacing laminin α2 chain in congenital muscular dystrophy with laminin α2 chain deficiency (25,26). An additional advantage of laminin α1 chain overexpression for possible gene therapy trials is that paralogous gene therapy might be beneficial in eradicating potential immune response problems.
It is noteworthy that transgenic expression of laminin α5 chain in the PNS partially promoted myelination in dy2J/dy2J/laminin α4 chain null roots. Nevertheless, it was suggested that laminin α5 chain supports radial sorting and myelination only via collaboration with the shorter dy2J-variant of laminin α2 chain (30) and thus it might be not of apparent help in laminin α2 chain deficient PNS.
Major laminin receptors on Schwann cells include integrins (α6β1, α6β4) and dystroglycan (28). Laminin-411 mainly interacts with integrin receptor α6β1 (30,31), whereas laminin α2 chain engages additional receptors (30). Both integrin β1 subunit and dystroglycan are important for normal myelination but only the former is crucial for early steps of myelination (32,33). Absence of laminin α2 chain did not appear to alter Schwann cell expression of integrin α6, β1 and β4 subunits and α- and β-dystroglycan (data not shown). However, integrin α7 subunit is also expressed on Schwann cells (28). In skeletal muscles, lack of laminin α2 chain leads to increased integrin α7Bβ1D synthesis but reduced expression of integrin α7B at the sarcolemma. Notably, integrin α7B is reconstituted at the cell surface upon transgenic expression of laminin α1 chain in muscles (34). It will also be interesting to analyze integrin α7 expression in PNS of dy3K/dy3K and dy3KLNα1TG-8 animals.
We demonstrated that basement membranes were restored in the PNS of dy3KLNα1TG-8 mice. Early studies showed that basement membranes were important for myelination in vitro (35–37). Later studies performed with cell cultures as well as in dy3K/dy3K and laminin α4 chain deficient mice indicated that continuous basement membranes are not strictly required for myelination (19,30,38). However, the presence of laminins and signaling appear to be important for this process to occur (30,38). Yet, continuous basement membranes might promote myelination more efficiently. They can potentially coordinate the distribution of internodal molecules and molecules present at the nodes of Ranvier (e.g. sodium channels) (20,39). Thus, the presence of basement membranes in our rescue model might be beneficial. Future work in our laboratory aims at demonstrating the role of laminin α1 chain in sodium channel clustering and the influence on conduction nerve velocity in laminin α2 chain deficiency.
In conclusion, we provide detailed evidence that laminin α1 chain can correct laminin α2 chain deficient peripheral neuropathy. Hence, our data strengthen our earlier hypothesis that gene therapy with laminin α1 chain might constitute promising therapy for the multiple defects seen in MDC1A.
MATERIALS AND METHODS
Laminin α2 chain null dy3K/dy3K mice and dy3K/dy3K mice overexpressing laminin α1 chain (dy3KLNα1TG derived from line no. 12, referred as dy3KLNα1TG-12) were previously described (16,25). Transgenic line no. 8 with expression of laminin α1 transgene in both skeletal muscles and peripheral nerves was produced in the same manner as described before (25). Briefly, LNα1TG mice derived from line no. 8 were crossed with dy3K/+ mice to generate dy3KLNα1TG-8 animals lacking laminin α2 chain and expressing laminin α1 chain in muscle and peripheral nerve.
Sciatic nerves, spinal roots and skeletal muscles were collected from wild-type, dy3K/dy3K, dy3KLNα1TG-12 and dy3KLNα1TG-8 animals, immersed in Tissue Tek and frozen rapidly in liquid nitrogen. Cryosections (7 µm) were subjected to immunofluorescence analyzes. Primary antibodies were rat monoclonal Ab 200 (25), 4H8-2 (Alexis Biochemicals, Lausanne, Switzerland), 1928 (Chemicon, Hampshire, UK), MTn15 (25), 9EG7 (BD Biosciences, San Jose, USA), 346-11A (BD Biosciences) and GoH3 (Beckman Coulter, Fullerton, USA) against laminin α1, α2, β1 chains, tenascin-C, integrins β1, β4 and α6, respectively; mouse monoclonal antibody IIH6 against α-dystroglycan (Upstate Biotechnology, Lake Placid, USA); rabbit polyclonal antibodies against laminin α3 (26), α4 (25), α5 (25), β2 (25), β3 (26), γ1, γ2, γ3 (26), collagen IV (Chemicon), perlecan (25) and β-dystroglycan (25). Sections were analyzed using a Zeiss Axioplan fluorescence microscope. Images were captured using an ORCA 1394 ER digital camera with Openlab 3 software. Images were prepared for publication using Adobe Photoshop software.
Immunoblotting was performed as previously described (26) with proteins isolated from skeletal muscles and sciatic nerves from dy3KLNα1TG-12 and dy3KLNα1TG-8 animals. Extract from Engelbreth-Holm-Swarm (EHS) tumor was from Invitrogen (Carlsbad, USA). Rabbit polyclonal antibodies detecting laminin α1, β1 and γ1 chains (Sigma, St Louis, USA) and laminin α1 chain (antibody 317) (40) were used. Quantification of chemiluminescence signals was performed using a CCD camera (LAS 1000, Fujifilm, Tokyo, Japan) and the software program Image Gauge V4 (Fujifilm, Tokyo, Japan).
Skeletal muscle cryosections (7 µm) were stained with hematoxilin and eosin (H&E). Cross- or longitudinal sections of sciatic nerves and cross-sections of spinal roots were stained with toluidine blue, as previously described (25).
Light microscopy analyzes
The diameters of myelinated axons and the thickness of myelin sheaths in sciatic nerves stained with toluidine blue were measured in Openlab 3 computer imaging software. The diameter of axons was measured at their shortest axis. Sciatic nerves from 4-week-old dy3K/dy3K, 2-month-old wild-type, 2-month-old dy3KLNα1TG-8, 2-month-old dy3KLNα1TG-12, 7-month-old wild-type, 7-month-old dy3KLNα1TG-8 and 7-month-old dy3KLNα1TG-12 animals were measured. Statistical significance was examined by using Student's t-test.
Transmission electron microscopy
Transmission electron microscopy on ultrathin cross- or longitudinal sections of sciatic nerves was performed as described previously (25).
Thirteen wild-type or dy3K/+, seven dy3KLNα1TG-12 and 14 dy3KLNα1TG-8 mice were trained for 2 days before the final test. On the first day, mice were placed at the rotarod three times for 120 s at a speed of 5 rpm and once at a speed of 7 rpm. On the second day, mice were given a 120 s session at 8 rpm and two sessions at 10 rpm. Animals were tested on the following day with one 8 rpm warm-up trial and three final trials at 10 rpm. One dy3KLNα1TG-8 mouse was excluded from the test owing to its poor general condition. The mean values of time (latency to fall) mice spent on the rotarod were compared between wild-type or dy3K/+, dy3KLNα1TG-12 and dy3KLNα1TG-8 animals and analyzed by Mann-Whitney test.
This work was supported by Muscular Dystrophy Association, The Swedish Research Council and Crafoord and Thorsten and Elsa Segerfalks Foundations. We thank Dr Takako Sasaki for generously providing various laminin antibodies and we gratefully acknowledge Professor Patrik Brundin for providing access to the rotarod device.
Conflict of Interest statement. None declared.