Centronuclear (myotubular) myopathy exists in X-linked, autosomal-recessive and autosomal-dominant forms.

The X-linked recessive form ("myotubular myopathy") has been genetically well characterised. Following initial linkage studies and assignment of a locus to chromosome Xq28 40868788899091929394 mutations in the myotubularin (MTM1) gene have now been identified in more than 90% of affected males 199596979899100; molecular genetic analysis of the MTM1 gene is now widely available as a routine diagnostic service 34101102. Disease-causing sequence changes include deletions/insertions, nonsense, missense and splice mutations (approximately 25% each) 97102. Three substitutions account for 15% of all MTM1 mutations; these are the splice mutation c.1261-10A>G (intronic, upstream of exon 12) resulting in the insertion of three amino acids FIQ at position 420 (7.3%), R241C encoded by exon 9 (4%), and c.141-144 delAGAA resulting in a frameshift at amino acid 48 in exon 4 (4%) 28. Other mutations have been reported in a few families or are unique. MTM1 mutations are distributed throughout the entire coding sequence, but localise most frequently (in descending order) to exons 12, 4, 11, 8 and 9 979899102103104105106107. Maternal carrier state of MTM1 mutations is estimated at 85% and is thus more common than statistically expected for a severe X-linked disease 97102; maternal mosaicism has been reported in a few families 97108109 with important implications for genetic counselling regarding future pregnancies.

Genotype-phenotype correlative studies have been difficult because many mutations are private to individual families and clinical severity associated with specific mutations may vary even within the same families; however, one large series demonstrated that, while most mutations are associated with the severe phenotype, some non-truncating mutations outside of the catalytic domain may carry a more favourable prognosis 192728110.

Screening of the MTM1 gene should be considered in females with suggestive clinical and histopathological features; although usually asymptomatic or only mildly affected, carriers may manifest severe symptoms in the presence of skewed X-inactivation and/or structural alterations involving the X-chromosome such as interstitial deletions 34353637383940.

Myotubularin belongs to the large family of dual-specificity phosphatases, playing a role in the epigenetic regulation of signalling pathways involved in growth and differentiation; mutations in some human myotubularin homologues have been associated with two specific forms of peripheral neuropathies of the Charcot-Marie-Tooth (CMT) type, CMT 4B1 111112 and CMT 4B2 113114. A specific function has been proposed for myotubularin in dephosphorylating phosphatidylinositol 3-phosphate [PtdIns3P] and PtdIns(3,5)P. These two phospholipids are second messengers with a crucial role in membrane trafficking; by dephosphorylation of PtdIns(3,5)P2, myotubularin also produces PtdIns5P, whose function is not fully characterised 95115116117118119120121122123124125126. In addition to the catalytic site, myotubularins form homo- and heterodimers and contain lipid and protein binding sites; these domains include a GRAM-PH, a coiled-coil region and a putative PDZ binding site. Concerning myotubularin, no protein interactors have been characterised to date in skeletal muscle. The deleterious effect of specific MTM1 mutations may be due to either destabilisation of the 3-D structure or loss of enzymatic activity, although it is possible that a few mutations affect existing but not yet identified protein-protein interactions in muscle.

Observations in an MTM1-related mouse model 127 suggest a role of myotubularin in muscle fibre maintenance but not in myogenesis. A recent gene expression profiling study in muscle harbouring MTM1 mutations revealed upregulation of transcripts for cytoskeletal and extracellular matrix proteins within or around atrophic myofibres, indicating that remodelling of cytoskeletal and extracellular architecture plays a role in the atrophy and intracellular disorganization observed in X-linked myotubular myopathy 128. Prolonged expression but eventual decrease of developmentally regulated proteins in muscle from affected infants (see also paragraph on Diagnostic methods below) suggests maturational delay rather than complete developmental arrest in this condition.

Dominant forms of centronuclear myopathy have been associated with mutations in two genes, the dynamin 2 (DNM2) gene on chromosome 19p13.2 76 also implicated in dominant intermediate (CMTDIB) 80 and axonal (CMT2) 129130 forms of Charcot-Marie-Tooth disease, and the skeletal muscle ryanodine receptor (RYR1) gene on chromosome 19q13.1 in one isolated case 82.

The DNM2 gene consists of 22 exons 80 and encodes a large GTPase protein involved in actin cytoskeleton assembly 131 and centrosome cohesion 132. In addition, DNM2 is implicated in membrane trafficking from the plasma membrane and Golgi, to allow the formation and fission of budding vesicles 133. It is of interest to note that myotubularin, implicated in the X-linked form of CNM, has also been implicated in membrane trafficking and endocytosis, although its precise function remains to be determined. Recent studies on cells transfected with CNM-related DNM2 mutants suggest lack of localisation to the centrosome and centrosome malfunction as a possible pathogenetic mechanism in DNM2-related CNM. However, the impacts of mutations on allosteric enzymatic activity and membrane remodelling properties of dynamin 2 remain to be investigated. Recurrent and de novo DNM2 mutations were originally identified following a positional candidate approach in 11 families with a mild form of autosomal-dominant centronuclear myopathy 76. The most common mutation is a 1393C>T change found in 6 unrelated families resulting in an arginine to tryptophane substitution at position 465. Dynamins are structurally complex proteins composed of 5 different domains; CNM-causing mutations identified to date mainly localise to the middle domain involved in protein self assembly and centrosome localisation 76, whereas those associated with CMTDIB have been identified within the pleckstrin homology (PH) domain 80. More recently, heterozygous de novo dominant DNM2 mutations affecting the PH domain have also been identified in a more severe CNM phenotype without any early peripheral nerve involvement and characterised by neonatal onset but gradual improvement over time 81. The potential overlap between myogenic and neurogenic findings in families with mutations in this region is currently being explored 7779.

Features of centronuclear myopathy associated with the skeletal muscle ryanodine receptor (RYR1) gene have to date only been reported in one single case with a de novo dominant mutation resulting in a serine to leucine substitution at position 4112 82; the RYR1 gene had been considered as a candidate in this patient because of the frequent observation of multiple central nuclei in other RYR1-related phenotypes and the suggestion of a clinical continuum and overlap of radiological features on muscle MRI. Functional studies on patient-derived, MyoD-transformed fibroblasts indicated that cells harbouring this mutation may be hypersensitive to depolarization, but it remains unclear how the change gives rise to the appearance of CNM. The frequency of the RYR1-related form of CNM is currently uncertain.

Mutations in the amphiphysin-2 (BIN1) gene on chromosome 2q14 have been recently identified in a small proportion of cases with the recessive form of centronuclear myopathy 59 but further genetic heterogeneity is expected. The BIN1 gene is organised in 20 exons, the protein exists in at least 10 different isoforms subject to alternative splicing and, in addition to muscle, the gene is expressed in a number of different tissues including central and peripheral nervous systems 134135. BIN1 was considered a candidate for genetically unresolved forms of CNM because of functional characteristics shared with other CNM-associated genes, namely a phosphoinositide-regulated role in membrane modelling 136, and the presence of a muscle phenotype in the Drosophila melanogaster mutant 137. The amphiphysin 2 muscle-specific isoform features an N-terminal amphipathic helix thought to be involved in creating membrane curvature, a BAR (Bin1, Amphiphysin, RVS167) domain that homodimerizes and maintains the curvature, a phosphoinositide-binding domain, and an SH3 domain interacting with dynamin 2 and other proteins 136138. Functional studies on the three mutations identified to date suggest that BIN1 missense mutations in the N-BAR domain affect membrane curvature, whilst a truncating mutation in the SH3 domain appears to abolish amphiphysin2-dynamin 2 interactions 59; this indicates the importance of amphiphysin 2-dynamin 2 coupling for normal muscle function and suggests a possible alteration of T-tubule organisation, as amphiphysin 2 was previously proposed to have a role in this process 137139.

In addition to the BIN1-related form, heterozygous missense variants in hJUMPY, a novel phosphoinositide phosphatase with functional similarities to myotubularin, were recently identified in two sporadic cases with features of centronuclear myopathy and an additional DNM2 mutation in one case 140. These variants were shown to decrease the enzymatic activity of hJUMPY in in vitro and in cellulo experiments. It is unclear whether the phenotype in those cases is due to digeny or recessive inheritance with an undetected second mutation, and clarification of the implication of hJUMPY awaits the characterisation of additional patients with mutations in this gene.

Genes implicated in various forms of centronuclear myopathy have been summarised in Table 1.

The diagnosis of CNM depends on the presence of typical histopathological findings on muscle biopsy in combination with suggestive clinical features; muscle MR imaging may complement clinical assessment and inform genetic testing in cases with equivocal features.

On muscle biopsy, centronuclear (myotubular) myopathy is characterised by centrally placed nuclei surrounded by a perinuclear halo devoid of myofilaments 2 and occupied by mitochondrial and glycogen aggregates (Figure 2) The characteristic central nuclei are seen in all muscles, including extra-ocular muscles 141, and may affect up to 90% of fibres 52. Some autosomal cases of centronuclear myopathy may also feature a radial arrangement of sarcoplasmic strands on NADH staining 50; it currently appears that this is a feature in most cases of centronuclear myopathy caused by mutations in the DNM2 gene 76: Whilst radial arrangement of sarcoplasmic strands appears to be common in mild forms of DNM2-related CNM due to mutations affecting the middle domain 77, this finding appears not to be as prominent in more severe and early presentations due to mutations affecting the pleckstrin homology (PH) domain, or in other genetically distinct forms of CNM. Type 1 predominance and hypotrophy 14142 are commonly associated features, and may precede the appearance of internal nuclei 143; there may be compensatory type 2 hypertrophy in a small number of fibres 25, and a deficiency of type 2B fibres with relative increase in undifferentiated type 2C fibres 142.

More recently, Pierson et al. 110 were able to correlate MTM1 mutation type and pathologic findings and could demonstrate that missense mutations are associated with increased myofibre diameter compared to nonsense mutations.

Histopathological changes may progress over time and marked increases in fat 144 and connective tissue 3778145 can at times be a striking feature. Associated core-like structures have occasionally been reported 47144 and may be associated with mutations in both the skeletal muscle ryanodine receptor (RYR1) gene 82 and the DNM2 gene 78.

Reported histopathological findings in carriers of the X-linked form range from normal appearance in clinically asymptomatic mothers 11 to findings similar to those in affected males, as reported in a female with the full clinical picture of myotubular myopathy due to skewed X-inactivation 37.

On electron microscopy (EM), immaturity of neuromuscular junctions and junctional changes comprising reduction of acetylcholine receptors on immunoperoxidase stains 141 and simplification of the postsynaptic membrane with paucity of secondary synaptic clefts 146 have been reported but the molecular basis for this observation remains uncertain. The great majority of CNM-related electron microscopy studies either predate the molecular resolution of these conditions or concern X-linked myotubular myopathy, and there are currently not sufficient data for more detailed EM genotype-phenotype correlative studies with a view to the more recently identified genes.

Immunohistochemical studies in CNM are mainly available for the X-linked form and have demonstrated consistent but non-specific abnormalities: persistent foetal expression pattern of various proteins including the cell surface protein N-CAM 146, myosin 32147, vimentin and desmin 145148149 have been reported in male infants with the X-linked forms, but more recent immunohistochemical studies on sequential biopsies in long-term survivors 150 suggest that the expression of developmentally regulated proteins eventually decreases as in healthy individuals. Other proteins abnormally expressed in myotubular myopathy include laminin and collagen components 145.

Muscle MRI findings have been reported in autosomal forms of CNM due to mutations in the DNM2 7778 and the RYR1 82 genes. Muscle MRI in cases of centronuclear myopathy secondary to mutations in the DNM2 gene show a characteristic progressive sequence (Figure 3) with early involvement of the ankle plantarflexors and subsequent signal changes within the hamstring muscles and, finally, the anterior thigh. This sequence and also the prominent adductor longus and rectus femoris involvement reported in one family 78 is distinct from cases with mutations in the RYR1 gene 82151 and may guide genetic testing in autosomal cases, particularly as cores on oxidative stains may be an additional finding in both DNM2- and RYR1-related forms 7882. Muscle imaging findings in MTM1-related CNM have only been reported in one manifesting female carrier 152 and are currently not documented for the recessive form due to mutations in the amphiphysin2 (BIN1) gene.

DNA sequencing of the exons and exon-intron boundaries of the implicated genes is used to confirm the diagnosis at the molecular level. For MTM1 sequencing, it is recommendable to start by investigating the more frequently implicated exons although mutations have been found distributed throughout the gene. With regard to DNM2 mutation screening, mutations identified to data are clearly concentrated in the middle and PH domains and these hotspots should be checked first. Due to lack of clear genotype-phenotype correlations and in cases where the genetic segregation of the disease cannot be assessed unambiguously, it is advisable to examine all three genes (MTM1, BIN1 and DNM2) in the search for molecular confirmation of the diagnosis.

RNA sequencing may be used if tissue or cultured cells are available from the patient, as the implicated genes appear to be ubiquitously expressed. For the X-linked form, the vast majority of the known MTM1 mutations lead to a decrease in protein levels in cultured myoblasts, fibroblasts or lymphoblastoid cell lines 153; based on only a few cases studied, mutations in the BIN1 and DNM2 gene do not seem to affect protein levels in such cell types. In addition, investigation of the RNA integrity or protein levels, although not used routinely, might reveal mutations in introns or regulatory sequences that remain undetected by DNA sequencing 100.

Central nuclei on muscle biopsy are not pathognomonic, and other neuromuscular disorders with a secondary increase in internal nuclei have to be considered in the differential diagnosis of CNM. Congenital myotonic dystrophy is a histopathological phenocopy 154 of the X-linked form, myotubular myopathy, and ought to be considered and excluded in the first instance by obtaining a detailed family history, clinical examination of the mother and specific genetic testing as indicated, before embarking on a muscle biopsy. In addition, although often already unlikely on clinical grounds, other causes of severe neonatal hypotonia ought occasionally to be excluded by more specific testing, including the other congenital myopathies, the congenital muscular dystrophies, spinal muscular atrophy, myasthenic disorders and motor neuropathies.

The autosomal-dominant form of myotubular myopathy 5061626376 also has to be differentiated from myotonic dystrophy and other autosomal-dominant disorders with numerous central nuclei on muscle biopsy, particularly in cases where mutations in the currently known CNM genes have been excluded, as clinical findings such as cataracts or electrical myotonia (i.e. myotonic bursts on needle EMG) 155156 suggest that some of the families reported in the premolecular era were affected by myotonic dystrophy rather than autosomal-dominant centronuclear myopathy. A facioscapulohumeral distribution of weakness in other families 4575 should lead to consideration of facioscapulohumeral muscular dystrophy in the differential diagnosis of autosomal-dominant centronuclear myopathy.

No curative treatment is currently available for any form of CNM and management is essentially supportive, based on a multidisciplinary approach.

X-linked myotubular myopathy is the most severe form of CNM and usually, but not invariably, follows a fatal course over days and weeks. Occasionally, long-term survival has been reported but often depends on the degree of respiratory intervention; a few male infants may be more mildly affected from the outset with better long-term prognosis 6181922272829. The decision regarding the duration of respiratory support is not an easy one, but as no firm prognostic criteria have yet been established 19, an at least initially proactive stance should be taken and any respiratory management decision should be made on an individual basis rather than on diagnosis alone. The latter approach is particularly advisable in cases with neonatal presentation where the X-linked form has been excluded, as recently described patients with recessive mutations in the amphiphysin 2 (BIN1) 59 gene and dominant mutations in the dynamin 2 (DNM2) gene 76 follow a milder course and may even improve over time. Patients who survive beyond the neonatal period without immediate ventilatory requirement will need close monitoring of their respiratory function, including polysomnography studies where needed, and are likely to benefit from initiation of non-invasive nocturnal ventilation as indicated 157. Follow-up care of those on nighttime ventilation should include regular cardiac assessments considering the risk of associated cor pulmonale 158159. Respiratory infections should be treated actively.

Feeding difficulties usually feature in infants with X-linked myotubular myopathy and may occur also in patients with severe recessive and dominant forms of CNM, requiring input from a speech therapist who may also promote normal speech if dysarthria is present.

As in other congenital myopathies, regular physiotherapy is aimed at the preservation of muscle power and function and the prevention of contractures; considering often prominent axial involvement, exercises promoting endurance and truncal stability such as swimming and riding 160 may be particularly useful. If orthopaedic complications evolve in the course of the disease, those may be managed surgically where conservative approaches have failed, and only at centres with experience in the management of neuromuscular disorders. As in other neuromuscular conditions, post-operative mobilisation ought to be rapid in order to avoid adverse effects of prolonged immobilisation such as muscle atrophy. In the most severe cases where walking can not be achieved without additional support, independent ambulation may be promoted by appropriate rehabilitative measures such as provision of weight-bearing calipers.

Malignant hyperthermia, an abnormal response to muscle relaxants such as succinylcholine and volatile anaesthetics 161162, has been previously only reported in one genetically unresolved case of CNM 85 and may reflect the recently documented involvement of the RYR1 gene in the condition 82. As malignant hyperthermia is not a recognised feature of other genetically determined forms of CNM, this complication ought to be mainly anticipated in genetically unresolved cases or those due to RYR1 mutations, although a cautious approach is generally advisable for patients with muscle disorders considered for general anaesthesia.

Genetic counselling should be offered to all patients and families in whom a diagnosis of CNM has been made. Only the identification of the causative mutation will determine the mode of inheritance in each individual family. In families with mutations in the myotubularin gene (MTM1), it is important to note that some women, who do not show in their lymphocyte-derived DNA the mutation identified in their affected son, may still carry a risk of recurrence because of germinal mosaicism for the mutation 97108109. Mutational analysis of MTM1 is now available as a diagnostic service, and in future the same might apply to screening of the more recently identified dynamin 2 (DNM2) and amphiphysin 2 (BIN1) genes associated with dominant and recessive forms of CNM, currently mainly available on a research basis. Considering the possible important prognostic implications depending on the underlying genetic defect, there is clearly a need for the establishment of a wider diagnostic network for CNM.

The prognosis of CNM is vaguely related to the mode of inheritance (see paragraph on Clinical description) with the X-linked form being more severe than dominant or recessive forms, respectively; whilst the mortality of X-linked myotubular myopathy is very high in infancy, some rare cases who are usually milder from the outset may achieve a reasonable quality of life 2728.

As only few genetically resolved families with dominant and recessive forms of CNM due to mutations in the DNM2 and BIN1 genes have been reported to date, genotype-phenotype correlations are still emerging (See paragraph on Clinical description).

Whilst the genetic basis of the X-linked form of CNM ("myotubular myopathy") has been known for a long time, recent years have seen the genetic resolution of a proportion of the autosomal-dominant and the autosomal-recessive forms of the condition. Despite these genetic advances, a substantial proportion of CNM cases remain currently still genetically unresolved, suggesting the existence of further gene loci. The pathological mechanisms leading to the skeletal muscle defects are still not understood and, although abnormal positioning of the nuclei is the histopathological hallmark of all genetically defined forms of CNM, the link between the mutated proteins and abnormal nuclear positioning remains unknown. Finally, considering that those concern ubiquitously expressed proteins, the tissue-specific expression of some of the implicated CNM mutations remains unaccounted for. Animal models may provide the basis for an advanced understanding of CNM and for future rational therapies of the condition.

The authors declare that they have no competing interests.

The authors equally contributed to this review article. They read and approved the final version of the manuscript.