Muscular Dystrophies and Allied Disorders III: Congenital Muscular Dystrophies and Associated Disorders

History and Background

The term ‘congenital muscular dystrophy’ (CMD) has been widely used to describe a group of infants with weakness and hypotonia from birth or within the first few months of life. Severe, early contractures are common, and most cases have delayed motor milestones. Detailed clinical studies combined with pathological and molecular studies have revolutionized this field and led to the characterization of several conditions and identified a novel pathogenic mechanism in muscle that involves the modification of proteins, in particular by glycosylation but also by phosphorylation. With advances in molecular analysis and whole genome/exome sequencing, novel genes are being identified, and the phenotype associated with already known genes has widened. Although several of these early-onset disorders with hypotonia are described as congenital muscular dystrophies, it can be questioned whether they are true muscular dystrophies with muscle necrosis. However, in this chapter we discuss disorders currently classified as a CMD, and novel disorders, in which pathological studies have been performed. We also include Bethlem myopathy, as some cases are within the clinical spectrum of a CMD and the underlying pathogenesis of both Ullrich CMD and Bethlem myopathy is defects in collagen VI. Some disorders often described as a CMD are discussed elsewhere (e.g. the disorder associated with the PTRF gene is discussed in Ch. 11 , the CHKB gene in Ch. 18 and INPPK5 and EPG5 in Ch. 16 ).

International consortia and a series of workshops sponsored by the European Neuromuscular Centre (ENMC) have made a major contribution to the advances in the field ( , ). The majority of identified forms have an autosomal recessive mode of inheritance, with the exception of some de novo dominant cases with collagen VI-related CMD, and cases with de novo lamin A/C mutations (see below). The clinical features, severity and progression are variable in the different entities ( Tables 12.1 and 12.2 ), in particular with regard to the degree of involvement of the central nervous system.

Despite being designated as muscular dystrophies and the sometimes striking pathological picture, some cases remain relatively static or progress only very slightly or slowly ( ) and, indeed, some patients may actually improve with time, passing various motor milestones and even achieving the ability to walk. The muscle biopsy may look considerably worse than the clinical picture, and one may be surprised that a patient with such extensive pathological change in a limb muscle may actually be ambulant. The degree of pathology cannot therefore be used as an index of severity of the disease or as a basis for prognosis.

To some extent the name ‘dystrophy’ is a misnomer, as unequivocal necrosis is not always a feature, but fibrosis and the presence of adipose tissue are often marked, and the presence of fibres with immature myosin isoforms suggests loss of muscle fibres and regeneration. It is difficult, however, to suggest an alternative name since a non-specific term such as ‘congenital myopathy’ is used for conditions defined by particular structural features (see Ch. 15 ). The amount of actual muscle tissue is low in CMDs so there is some kind of ‘dystrophic’ process with a reduced number of muscle fibres, for whatever reason, even though the process may not be one of necrosis. Some of the reduced bulk of muscle, however, may relate to feeding difficulties and failure to thrive, and gastric tube feeding often results in weight gain. In addition, serum creatine kinase (CK) is not elevated in some forms in contrast to other muscular dystrophies.

The incidence of CMD is difficult to estimate since the identification of cases, the clinical spectra and number of disease entities is increasing. There is also significant geographical variation in the incidence of the various forms (see ). For example, the CMD described by is one of the most common forms of muscular dystrophy in Japan after Duchenne dystrophy, because of a founder effect, but is rare elsewhere in the world. In contrast, mutations in the gene for the fukutin-related protein responsible for congenital muscular dystrophy type 1C (MDC1C) and limb-girdle muscular dystrophy type 2I (LGMD2I) (see Ch. 11 ) are common in the northern European population but rare in Asia. An epidemiological study in north-east Italy estimated the incidence of CMD to be 4.65 × 10 ⁻⁵ , with a prevalence of 6.8 × 10 ⁻⁶ ( ), indicating that CMD is one of the most common neuromuscular disorders. In our CMD referral centre in London, an audit of the frequency of various forms of CMD has shown that LAMA2 -related CMD, Ullrich/collagen 6-related CMD and dystroglycan-related CMDs are the most common forms in our population ( ). A number of other studies have also indicated the prevalence of congenital muscular dystrophies in different populations ( ). Figures are undoubtedly an underestimate in the light of recent developments.

The clinical phenotype has provided the basis for the classification of the variants of CMD and in 1994 the first ENMC workshop on CMD was convened to define and designate individual syndromes for genetic study ( ). This workshop concentrated on CMDs with brain and eye involvement and did not include Ullrich and Bethlem forms, which were discussed by a separate consortium. At that time the genetic defects and pathogenesis underlying the heterogeneity of CMDs with brain involvement were not known. Three clinical syndromes, however, were recognized—Fukuyama congenital muscular dystrophy (FCMD), muscle-eye-brain disease (MEB; Santavuori et al 198) and Walker–Warburg syndrome (WWS; )—which were distinguished by the degree of mental retardation, eye involvement and structural brain changes such as cobblestone lissencephaly. A major step forward the same year was the finding of abnormal expression of laminin α2 (also referred to as merosin) in nearly half of the cases of ‘classical’ CMD ( ). This correlated closely with the more severe cases with inability to walk unaided, joint contractures, and white matter changes on brain imaging. The cause was later found to be a primary defect in the LAMA2 gene encoding laminin α2 on chromosome 6q ( ). Since then, the understanding of the pathogenesis of the CMDs has increased dramatically: not only have many causative genes been identified but also the involvement of several extracellular matrix proteins and important post-translational modification pathways (e.g. glycosylation).

Molecular analysis soon led to the identification of genes responsible for the various forms, and defects in a growing list of genes have been shown to cause a CMD (see Tables 12.1 and 12.5 ). These can be grouped according to the type of protein the gene encodes ( ); proteins of the extracellular matrix such as laminin α2, collagen VI and integrin α7 have an important role at the sarcolemma; several genes have an important role in the glycosylation pathway of α-dystroglycan; selenoprotein N is an enzyme in the endoplasmic reticulum of unknown function; and lamin A/C and nesprins are nuclear proteins ( ). There is still further genetic heterogeneity, as a genetic defect has not yet been identified in several cases of CMD. As is apparent in Tables 12.1–12.6 , there is considerable clinical overlap between the different forms, in particular disorders that involve hypoglycosylation of α-dystroglycan and a similar phenotype may arise from a defect in more than one gene (see Table 12.5 ). In addition, defects in the same genes can cause a limb-girdle phenotype (see Ch. 11 ; ). For example, severe cases with defects in the gene for FKRP can result in a phenotype that resembles MEB or WWS. In addition, it is not always possible to distinguish a primary gene defect that may only have a mild effect on protein expression from a secondary one. This adds to problems for the pathologist and emphasizes the importance of combining pathological and clinical data in trying to arrive at an accurate diagnosis ( ).

The nomenclature of the CMDs (MDC1A–1D) was assigned by the Online Mendelian Inheritance in Man database (OMIM). As the acronym CMD had already been assigned to cardiomyopathies, MDC is used. As yet, there is no neuromuscular disorder designated MDC2. Recently, OMIM adopted a new nomenclature for the CMDs associated with hypoglycosylation of α-dystroglycan based on the clinical phenotype and the presence or absence of brain involvement but it is not widely used. The groups are subdivided into MDDGA, congenital forms with brain and eye involvement, MDDGB for congenital forms with or without mental retardation and MDDGC forms with a ‘limb-girdle’ phenotype with or without mental retardation. The original clinical terms for the forms with severe brain involvement are still in use and remain clinically appropriate (Fukuyama; MEB; WWS) and are caused by defects in several genes.

We summarize here the main pathological features of muscle biopsies in the various forms of CMD but this is an ever-expanding field. For further clinical details and citation of original articles readers are referred to various articles ( ).

General Pathological Features of Congenital Muscular Dystrophies

As the various forms of CMD share similar pathological features, we summarize these collectively here and then cover the individual disorders together with the application of immunohistochemistry. It is not possible to identify a particular form of CMD from the histological and histochemical features alone. Representative views of the typical histological and histochemical features seen in CMDs are shown in Fig. 12.1 .

All CMDs show an abnormal variation in fibre size, and the shape of the fibres is often rounded. Fibre atrophy and hypertrophy is common, and the hypertrophic fibres may be split. Internal nuclei occur but are not numerous. The amount of fibrosis and adipose tissue is variable, but both may be very extensive. The presence of endomysial connective tissue around individual atrophic fibres can help distinguish these fibres from the group atrophy seen in spinal muscular atrophy (SMA). One biopsy in our series of laminin alpha 2-deficient MDC1A cases had a surprisingly similar histological appearance to a case of SMA, but the grouped atrophic fibres were surrounded by endomysial connective tissue ( Fig. 12.2 ).

Necrotic fibres and basophilic regenerating fibres may be seen, particularly in early stages of the disease, but not in all cases. Inflammation is rare but may be prominent in some cases such as those with a mutation in the LMNA (see below; ) and may partly account for a positive response to corticosteroids in some cases ( ). Fibre typing with histochemical techniques may be indistinct or show a predominance of type 1 fibres. Oxidative enzyme staining may show disruption of myofibrils and abnormalities in mitochondrial distribution, including aggregation resembling lobulation or minicores (see Fig. 12.1g and h ). Both these have been noted in particular in Ullrich CMD and the form associated with rigid spine (RSMD1), caused by mutations in the genes for collagen VI and SEPN, respectively.

In addition to focal myofibrillar disruption (minicores), electron microscopy reveals abnormalities of the basal lamina and nuclei ( ) and fewer satellite cells are seen in CMD patients compared with controls ( ).

Congenital Muscular Dystrophies Associated with Sarcolemmal Proteins

Laminin α2

A primary deficiency of laminin α 2 (‘merosin deficiency’; MDC1A) is caused by mutations in the LAMA2 gene on chromosome 6q. It has been quoted as representing 30–40% of all cases and being one of the most common forms. This figure, however, was based on the original screening by various international groups, before the identification of other causative genes. From the experience of our own patient pool, LAMA2 defects account for about 37% of CMDs, but an increasing number of patients with other gene defects, especially collagen VI, responsible for the Ullrich form of CMD and dystroglycanopathies are being identified ( ).

Cases of MDC1A invariably present at birth, or soon after. The hypotonia and muscle weakness may be associated with failure to thrive and respiratory and feeding problems. Contractures are usually present but severe arthrogryposis is rare. Limitation of eye movement, in particular upward gaze, is also common. Patients with a complete absence of laminin α2 protein rarely achieve the ability to walk independently but are usually able to sit unsupported. Those with only a partial reduction of protein usually show a milder, more variable phenotype ( ). Thus, as with many disorders, some residual localization of protein produces a milder phenotype than a complete absence. Serum CK levels are always elevated and abnormal cerebral white matter changes are invariably seen with T2-weighted magnetic resonance imaging (MRI) of the brain by the age of 6 months but can be seen earlier. To date, all patients with a total absence of laminin α2 and a mutation in the LAMA2 gene have shown increased signal intensity in the white matter. Although the white matter changes resemble a leucodystrophy with demyelination, loss of myelin has not been shown, and the abnormalities may relate to abnormal myelination rather than loss of myelin and/or to the presence of water. Some cases also show structural brain changes but not of the type in other forms of CMD ( ; see Table 12.2 ). Milder cases with a partial deficiency of laminin α2 may not always show white matter changes and CK levels may not be as high ( ). Phenotype may not always be suggestive of a laminin α2-related disorder, and it is therefore worth considering routinely screening biopsies with antibodies (see Ch. 6 ).

Laminins are components of the basal lamina, and all 12 variants that have been identified are composed of a heterotrimer of an α, β and γ chain. The increasing diversity of laminin variants led to a numbering system for the nomenclature of each heterotrimer, replacing the previous names of merosin (M), laminin A, B1 and B2, and more recently a numerical one for each trimer has been adopted in which the merosin trimer is called 211 ( ). The most abundant trimers in muscle are laminin-2 (211, merosin; composed of α2-β1-γ1 chains) and laminin-4 (221, S-merosin; composed of α2-β2-γ1 chains). Thus, mutations in the gene for laminin α2 affect both variants. The presence of laminin α2 on the sarcolemma of cardiomyocytes probably accounts for the cardiac problems in some cases. Laminin α2 is also present on Schwann cells, in association with laminin β2 and γ1 (S laminin; laminin-4; laminin 221), and mutations in the LAMA2 gene affect motor nerve myelination, reducing motor nerve conduction velocities ( ). Sensory nerve function, however, is unaffected. In the brain, laminin α2 is present on blood vessels, but it is not detected on blood vessels in muscle (see Ch. 6 ). Laminin β2 is also present on the whole of the sarcolemma, not only at neuromuscular junctions, as often stated, but the influence of mutations of laminin α2 on the sarcolemmal trimer of α2β2γ1 is not known. In cardiac muscle, laminin is also localized to the T-tubule system, but this is not seen in skeletal muscle.

Immunohistochemical studies are essential for the analysis of all cases of CMD, to visualize and localize various proteins. Immunolabelling of laminin α2 in cases with a mutation in the LAMA2 gene may show a complete absence, slight traces on a few fibres, or a reduction on several or only a few fibres ( Fig. 12.3 ). The latter is often more apparent with an antibody to the N-terminal fragment of the protein ( Fig. 12.4 ). Laminin α2 is processed into two fragments when denatured for immunoblots, one of 80 kDa and one of 300 kDa. The commercial antibody from Millipore (previously Chemicon), MAB1922, recognizes the 80 kDa fragment and is the only commercial antibody that we are aware of that works adequately on immunoblots. The antibody from Alexis and other companies (4H8) has been shown by immunoprecipitation to recognize the N-terminal 300 kDa fragment, substantiated by the studies of , and often shows the partial reduction of laminin α2 better than the Millipore antibody ( ). Like the Alexis antibody to the 300 kDa fragment, the antibody from Leica/Novocastra (NCL-MER3) also shows the partial reduction better than the one from Millipore. This antibody does not recognize denatured laminin α2 on immunoblots but its epitope may be on the C-terminal LG3 or LG4 domain ( ). This antibody should not therefore be cited as recognizing the N-terminus, as frequently occurs in the literature.

It is important to examine laminin α2 on nerve axons if it is present in a sample. Normal nerves show laminin α2 round each axon and in cases of MDC1A with an absence of sarcolemmal laminin α2 it is also absent from the axons. In rare cases, however, laminin α2 may appear almost normal, or only mildly reduced on the sarcolemma, but be absent from nerves ( ).

Laminin α2 is not expressed on the blood vessels of skeletal muscle but is detected on the blood vessels of the brain ( ). The immunolabelling of blood vessels in the brain of MDC1A patients is not known. Some labelling of laminin α2 may be observed on the inner rim of some vessels in muscle, but not the basal lamina. In MDC1A the deficiency of laminin α2 is a primary effect caused by mutations in the LAMA2 gene. In several other forms of CMD a secondary reduction occurs as a consequence of a mutation in another gene. In milder cases it may be difficult to distinguish this secondary reduction from a primary partial reduction. An important distinguishing feature, however, is the presence of white matter changes on brain MRI, as this is a consistent feature in CMD cases with LAMA2 mutations over the age of 6 months.

Secondary immunohistochemical features are of diagnostic relevance. Laminin α5 is overexpressed on mature fibres in MDC1A. This protein, however, is developmentally regulated (see Ch. 6 ) and is high on immature fibres ( ). It is therefore important to make comparisons with an antibody to embryonic and/or fetal myosin so that immature and/or regenerating fibres can be excluded from the assessment. It can also be quite high, however, on muscle fibres in neonates and the exact decline of its sarcolemmal expression in normal infants is not clear. Moderate sarcolemmal levels in neonates may relate to a delay in maturation, and differences in antibody titre and detection methods may influence the intensity of labelling observed. The commercial antibody that is often used (clone 4C7) was originally thought to recognize the laminin α1 chain (previously known as laminin A), but showed that it is probably the α5 chain that is recognized by this antibody.

No detectable difference in the labelling of laminin β1 and γ1 is usually seen in MDC1A, and these can be used to control for good preservation of the basal lamina. Similarly, labelling of β-spectrin, dystrophin and its associated proteins is usually normal. In contrast, laminin β2 shows a reduction on the sarcolemma ( ). Interpretation of this, however, also has to take immaturity of the fibres into consideration (see Ch. 6 ). The integrin complex α7β1D and α-dystroglycan are also reduced on the sarcolemma in MDC1A, which is thought to result from the deficiency of laminin α2, as both of these interact with laminin α2. Integrins, however, are also developmentally regulated, and immaturity has to be considered when assessing biopsies from very young patients.

Labelling of myosin isoforms shows that fibres with fetal myosin are abundant, and fibres with the developmental myosin isoform are also present but usually less numerous. As necrosis and, by inference, regeneration that usually follows it, are not always marked features, the presence of so many fibres with fetal myosin may not all be a reflection of regeneration (see Chapter 10, Chapter 6 ). Biopsies from cases of CMD are often taken when patients are very young, and it is possible that some fetal myosin reflects a lack of maturation, or an up-regulation. As in many muscular dystrophies, fibres with slow myosin are often predominant and some may co-express fast myosin and immature myosin isoforms.

Laminin α2 is also expressed in skin at the epidermal/dermal junction on the base of the keratinocytes, on the sensory nerves and round hair follicles ( ); thus, skin biopsies can be useful for diagnosis ( Fig. 12.5 ). Skin biopsies are particularly useful in cases where muscle is not available and in cases with pronounced muscle wasting, in which a muscle biopsy is likely to yield very few muscle fibres for assessment ( ). In some cases the reduction of laminin α2 may appear greater in skin than muscle. Studies of other laminin chains, including α5, β1 and γ1, act as good controls for studies of laminin α2 in skin.

Prenatal diagnosis of cases of MDC1A is aided by studies of the expression of laminin α2 in chorionic villus samples ( ), and affected fetuses show a complete absence of laminin α2 ( Fig. 12.6 ). It is important, however, to establish that there is an absence, or near absence, of laminin α2 in muscle or skin in the proband, as the expression of laminin α2 in chorionic villi of fetuses with a partial deficiency is not known. Laminin β2 is also reduced in chorionic villi of affected fetuses, suggesting that, as in muscle, the laminin-4 variant (laminin 221) may have a pathogenic role as well as laminin-2 (Sewry, unpublished observation; see ).