Spinal Muscular Atrophies

Type I SMA

Following the initial description of infantile SMA by Werdnig and Hoffman in the early 1890s and further descriptions made by Sylvestre in 1899 and Beevor in 1903, infantile or type I SMA was described again in detail, both clinically and pathologically, by Randolph Byers and Betty Banker at Boston Children’s Hospital in 1961. Patients with type I SMA, also known as Werdnig-Hoffman disease, present between birth and 6 months of age. SMA type I has been further subdivided into three groups: type IA (or type 0 in certain reports ) with onset in utero and presentation at birth, type IB with onset of symptoms prior to 3 months of age, and type IC with onset between 3 and 6 months of age. Infants with SMA type I have progressive proximal weakness that affects the legs more than the arms. They have poor head control, hypotonia that causes them to assume a “frog-leg” posture when lying and to “slip through” on vertical suspension, and areflexia. They are never able to sit (“non-sitters”) ( Table 8.1 ). There is also weakness of the intercostal muscles with relative sparing of the diaphragm, producing a bell-shaped chest and a pattern of paradoxical or “belly”-breathing. Infants with type I SMA also classically exhibit tongue fasciculations and eventually have difficulty swallowing, with risk for aspiration and failure to thrive. Other cranial nerves are not as affected, although facial weakness does occur at later stages of the disease. Cognition is normal, and they are often noted at diagnosis to have a bright, alert expression that contrasts with their generalized weakness. Infants with type I SMA usually develop respiratory failure by age 2 years, or much earlier, and in the past most did not survive past 2 years; however, there has been some increase in survival in recent years with the use of assisted ventilation (to be described in more detail in later sections) and other interventions. Prior natural history studies have reported shortened life span, with 68% mortality within 2 years and 82% by 4 years of age. The application of nutritional and respiratory interventions has reduced the mortality to approximately 30% at 2 years, but about half of the survivors are fully dependent on noninvasive ventilation.

It was previously believed that SMA is a purely motor neuron disorder. However, recent studies have shown that severe type I SMA can result in various organ manifestations, apart from spinal cord motor neurons such as brain, cardiac, vascular, and even sensory nerve involvement. Recent autopsy studies have shown increasing evidence of congenital heart disorders in severe SMA, with the most common association being hypoplastic left heart syndrome. In a study, 3 out of 4 (75%) patients with 1 SMN2 copy had congenital SMA (SMA type 0) associated with hemodynamically significant atrial or ventricular septal defects, which underscores the relevance of the SMN protein for normal cardiac development ; despite the reported cases, however, a chance association has not been firmly excluded. Severe SMA mouse models have shown autonomic disturbances including bradyarrhythmia and progressive heart block, as well as dilated cardiomyopathy and decreased contractility ; in humans with molecularly proven SMA there has been no evidence of dilative or congestive cardiomyopathy, nor have there been reports of AV conduction anomalies or arrhythmias besides bradyarrhythmias. Studies done on various mouse models have also found that severe SMN protein deficiency might present as vasculopathy. This was also noted in the case reports of two unrelated patients with severe SMA type I. Both of these infants developed ulcerations and necroses of the fingers and toes. Sural nerve biopsy studies on them showed thrombotic occlusion of small vessels without inflammation resulting in perfusion abnormality and subsequent tissue necrosis in the absence of a significant peripheral neuropathy. Autonomic dysfunction is thought to be the primary cause of this vasculopathy. While some of the reported patients had only one SMN2 copy, two cases of infants with SMA type I in whom a distal necrosis developed had two SMN2 copies. In a different biopsy study done by Rudnik-Schöneborn et al. on 19 patients with infantile SMA, significant sensory nerve pathology was found in severe SMA type I patients, whereas no sensory involvement was found in SMA type II and type III patients. Out of these, axonal degeneration was noted in seven and abnormal sensory conduction in five SMA type I patients. The abnormal nerve conduction study patients were followed with biopsy, which showed marked fiber density reduction in four patients, thus indicating that severe SMA type I patients can show both clinical and morphological sensory nerve involvement.

Type II SMA

This intermediate form of SMA was first reported in 1893 at the University of Edinburgh, United Kingdom, and described again by Byers and Banker in 1961 and in detail by Dubowitz in 1964. Patients with type II SMA, also known as intermediate SMA or Dubowitz disease, are able to sit unsupported at some point (“sitters”), but are never able to stand alone or walk ( Table 8.1 ). The onset of symptoms is between 6 and 18 months of age. They have progressive proximal weakness affecting legs more than arms, hypotonia, and areflexia. They also develop progressive scoliosis which, in combination with intercostal muscle weakness, results in significant restrictive lung disease as they grow older. They develop joint contractures and can have ankylosis of the mandible. They exhibit tremor, or polyminimyoclonus, of the hands. Although their body mass index may be low (at the 3 ^rd percentile or less when compared with normal children), the high-functioning, nonambulatory patients have a higher relative fat mass index and are at risk of becoming overweight. Cognition is normal and verbal intelligence may be above average. In a study done on 240 type II patients, survival rates were found to be 98.5% at 5 years and 68.5% at 25 years. Patients may live into the third decade, but life expectancy is shortened due to the risk of respiratory compromise.

Type III SMA

In 1956, Kugelberg and Welander described a much milder form of SMA characterized by prolonged ambulation. Patients with type III SMA, also known as Kugelberg-Welander disease, are able to stand alone and walk at some point (“walkers”). The onset of symptoms occurs after the age of 18 months; it has further been subdivided into type IIIA (onset between 18 months and 3 years) and type IIIB (onset after 3 years). They have progressive proximal weakness affecting legs more than arms and may ultimately need to use a wheelchair, but they generally develop little to no respiratory muscle weakness or severe scoliosis. Loss of ambulation increases the risk of these complications. They may have tremor or polyminimyoclonus of the hands. Sometimes, the calves of these patients can be prominent and hence type III SMA can be confused with Becker muscular atrophy. Life expectancy is not significantly different compared with the normal population ( Table 8.1 ). Case Example 8.1 describes the disease progression of a particular type III patient.

There has been much debate about the appropriate classification of patients into these three types of SMA because, as mentioned, there are patients within these categories who exhibit phenotypes of differing severities. A classification system based on a continuous rather than discrete variable (e.g. SMA “type 1.8” in the case of a less severely affected type I patient) has been proposed to better capture the clinical spectrum of these patients.

Outliers

There are also patients who are outliers on either end of the phenotypic spectrum. As discussed above, an SMA “type IA” (formerly known as type 0) has been used to describe neonates who present with severe weakness and profound hypotonia, probably of prenatal onset, as well as with a history of decreased fetal movements. The majority do not attain any motor milestones. Other findings include areflexia, facial diplegia, atrial septal defects, and joint contractures. In SMA type IA, respiratory failure forms an important cause of morbidity and mortality, requiring noninvasive ventilation and endotracheal intubation at birth. Life expectancy is reduced and most of them are unable to survive beyond 6 months of age ( Table 8.1 ). Furthermore, arthrogryposis multiplex congenita (congenital joint contractures involving at least two regions of the body) has been noted in SMA patients with SMN1 gene deletions, and congenital axonal neuropathy involving motor and sensory nerves in conjunction with facial weakness, joint contractures, ophthalmoplegia, and respiratory failure at birth has been reported in three newborn siblings with deletions in the SMA chromosomal region.

A milder adult-onset SMA, or type IV SMA, has also been described with onset of symptoms after age 21 years and essentially normal life span. Most patients with the SMA type IA and IV phenotypes have homozygous deletions of exon 7 in SMN1 , but as it will be discussed in the Genetics section, the SMN2 copy number is usually only 1 in SMA type IA and 4 to 5 in SMA type IV.

Other “Spinal Muscular Atrophies”

The non-5q13 associated spinal muscular atrophies are a heterogeneous group of motor neuron diseases associated with mutations in a variety of different genes (e.g. X-linked and autosomal dominant or recessive spinal muscular atrophies), distal or segmental spinal muscular atrophies, or distal hereditary motor neuropathies. Patients with these disorders generally have some clinical characteristics that can help differentiate them from those with 5q13-associated or classic SMA. The differential diagnosis of SMA is detailed in the respective section.

The SMN Gene

Linkage analysis studies showed that all three forms of SMA map to chromosome 5q11.1-13.3. In 1995, Lefebvre et al. identified the SMN gene within this region, which was absent or interrupted in 98.6% percent of the patients in their group. The structure of this region is complex, with a large inverted duplication of a 500 kb element. This contains the SMN1 gene, which is deleted or interrupted in patients with SMA and is evolutionarily older, in the telomeric portion of the region; and the SMN2 gene, a duplication of SMN1 that differs from it by only 5 nucleotides, in the centromeric portion ( Figure 8.1 ). The critical difference between SMN1 and SMN2 is a C to T transition that creates an exon splicing suppressor in exon 7 of SMN2 . While this splice modulator change is translationally silent (i.e. it does not change the amino acid sequence), it affects the alternative splicing of the gene, so that exon 7 is spliced out of or excluded from most SMN2 mRNA transcripts. This altered mRNA results in production of a truncated version of the SMN protein, which does not oligomerize efficiently and is degraded. Since exon 7 is not always spliced out of all SMN2 mRNA, a small amount of full-length transcript and hence functional protein is produced by SMN2 , but it yields only on average about 10% as much as that produced by SMN1 . In patients with SMA both copies of the SMN1 gene are deleted or disrupted, so the individual is left with only the small amount of SMN protein produced by the remaining copies of SMN2 . The amount of SMN protein is inversely correlated with the severity of disease.

About 95% to 98% of patients with SMA harbor deletions of the telomeric SMN1 gene ( Table 8.2 ). The remainder have small intragenic mutations or have undergone gene conversions from SMN1 to SMN2 . De novo mutations occur at a rate of about 2%, which is relatively high, and explained by the fact that this region of chromosome 5 is unstable, containing not only the inverted repeat of SMN1 and SMN2 , but other surrounding low copy number repeats. This leads to a high rate of unequal crossover and de novo mutations (especially paternally derived), which could explain the relatively high carrier frequency despite the mortality rate for the most severe forms of the disease. More mildly affected patients seem more likely to undergo gene conversions of SMN1 to SMN2 rather than deletion of SMN1 .

The number of copies of SMN2 per chromosome 5 varies among normal individuals, and 10–15% of the population possess no copies of SMN2 . Among patients with SMA, a clear correlation has been established between SMN2 copy number and phenotypic severity. Feldkotter et al. in 2002 found that in their series 80% of patients with type I SMA had one or two copies of SMN2 , 82% of patients with type II had three copies of SMN2 , and 96% of patients with type III had three or four SMN2 copies ( Figure 8.2 ). Studies by Mailman et al. in 2002 and Arkblad et al. in 2009 found similar results (95–100% of type I patients had one or two copies of SMN2 , and all type III patients had at least three copies of SMN2 ). However, this correlation is not so perfect as to permit the absolute prediction of clinical severity based on SMN2 copy number, especially in the intermediate forms of the disease where there is some overlap (patients with three copies of SMN2 have been described with all three phenotypes). One reason for the overlap is that not all copies of the SMN2 gene are equal; in terms of full-length SMN protein production, some copies probably produce less and some more than 10% functional protein. In general, though, it can be said that a patient with one or two copies of SMN2 is highly likely to present with SMA type IA, IB, or IC. Interestingly, unaffected family members with homozygous SMN1 deletions and five copies of SMN2 have been described, which suggests that SMN2 copy number alone cannot be the sole modifying factor in disease severity, since there are patients with type III SMA who also have five copies of SMN2 . In Smn (−/−) knockout mice, the introduction of eight copies of human SMN2 can rescue the phenotype and result in essentially normal mice. The presence of only two copies of SMN2 in mice results in a severe SMA-like phenotype and death at 5 days. The addition of an SMN transgene lacking the exon 7 sequence ( SMNΔ7 ) along with two copies of SMN2 prolongs the lifespan of the transgenic mouse to about 14 days.

A positive modifier has been identified in the SMN2 gene. A single base substitution in exon 7 in three unrelated patients’ DNA (c.859G>C) creates a new exonic splicing enhancer that increases exon 7 inclusion and thus the amount of full-length protein by about 20%. The phenotype in these patients was less severe and did not correlate with their SMN2 copy numbers, supporting the positive modifying effect of this sequence change and the notion that not all SMN2 copies are equal. Of note, this variant has been identified in two copies in type IIIB patients with a mild phenotype, in one copy in type II patients, and not at all in severe type I patients. It seems that a 20% increase in full-length protein in patients with two copies of SMN2 will probably result in mild type IIIB SMA; therefore, it is fair to speculate that a 25% increase in full-length SMN protein in the same patients may result in total rescue of the phenotype. There are also modifiers that do not involve the SMN1 or SMN2 genes. It is well known that haploidentical siblings with the same SMN2 copy number can have SMA of different severities. Higher expression of plastin 3 ( PLS3 ) was also identified as a gender-specific protective modifier of SMA in asymptomatic SMN1 -deleted females, carrying the same number of SMN2 copies as their affected sibling. A subsequent study, however, showed that PLS3 may be an age- and/or puberty-specific and sex-specific modifier as the expression of the gene was greatest in postpubertal female patients with SMA type III, intermediate in SMA type II, and lowest in SMA type I patients, underscoring an association with disease severity as well. Further, high plastin levels have also been detected in female siblings with more severe SMA presentation. It is possible that the plastin 3 modifier is not an important modifier of SMA phenotype or it may be female-specific and incompletely penetrant. No changes in the plastin 3 gene itself have been reported; thus, the modifier role of plastin 3 in SMA remains uncertain. Sequence variants within splicing regulators of the SMN1 and SMN2 genes, either polymorphisms or mutations, may explain divergent phenotypes in haploidentical siblings.

Regarding the level of SMN protein itself, although there is in general an inverse correlation between the level of SMN protein and severity of disease, this correlation does not seem to be quite as close as that between SMN2 copy number and phenotype.

Genetic Diagnosis

Genetic testing for SMA can be performed with PCR-based targeted mutation analysis using a restriction enzyme that digests exon 7 ( Table 8.2 ). If SMN1 exon 7 is homozygously deleted (as it is in about 95–98% of patients) then the restriction fragments will be absent, the DNA will not be amplified, and a band will not be seen on the gel. If an individual is a carrier (heterozygous deletion) then a lighter band will be present ( Figure 8.2 ). However, multiplex ligation probe amplification (MLPA) methodology is currently applied in most DNA diagnostic laboratories for deletion analysis of exon 7 of the SMN1 gene in potential probands and carriers. This type of targeted mutation testing in conjunction with sequence analysis can also detect individuals who are compound heterozygotes with a deletion of exon 7 in one SMN1 allele and an intragenic point mutation in the other allele. In such a case (approximately 2–5% of individuals with clinical diagnosis of SMA), sequence analysis of the SMN gene will detect the mutation; this sequence testing, however, will not detect exonic deletions or duplications and will not determine whether the point mutation is in the SMN1 gene or SMN2 gene (if one of these genes is not deleted). Fortunately, certain point mutations have been described in more than one SMA patient already and thus the detection of a previously reported mutation supports its pathogenicity and its location in the SMN1 gene. The use of long-range PCR or subcloning can also allow specific analysis of SMN1 and confirm that the mutation is present in SMN1 . See Figure 8.3 .

Carrier testing is feasible and accurate in the parents of patients with homozygous deletions of exon 7 or compound heterozygosity using a PCR-based dosage assay, known as “SMN gene dosage analysis” ( Figure 8.4 ). It can also be performed using other techniques such as long-range PCR, MLPA, and chromosomal microarray (CMA) that includes this gene segment. Sequencing of the SMN gene will detect point mutations in nondeletion carriers. Rarely, carriers may have two copies of SMN1 on one chromosome (a so-called “2+0 carrier”); the incidence of this genotype is about 4% of the general population. In the “2+0 carriers”, the SMA dosage carrier test will be falsely normal and thus one may need to pursue other methods such as family linkage analysis to identify the disease-associated genotype in families in which a deletion mutation has been transmitted more than once from a parent with two copies of the SMN1 gene on gene dosage testing. Due to the occurrence of de novo mutations in 2% of patients with SMA, one of the parents may not be a carrier. So, in approximately 6% of parents of a child with SMA secondary to a homozygous SMN1 deletion, SMN1 gene dosage testing will be normal.

As mentioned, a general correlation has been found between SMN2 copy number and disease severity, and it is relatively straightforward to determine SMN2 copy number in individual patients using various methodologies. However, this correlation is not so strict that the severity or type of disease can be reliably predicted based on copy number, and it is hence not advisable to offer families prognostic information based on SMN2 copy number assays.

Newborn Screening

The potential for newborn screening for SMA has been of great interest because the ideal time to initiate therapy would be prior to the initial degeneration of motor neurons, and newborn screening may help to identify presymptomatic individuals. Swoboda et al. performed a prospective study in prenatally diagnosed infants with type I SMA and found that, associated with the initial onset of symptoms or decline in function in these young infants, there was electrophysiologic evidence of precipitous denervation (assessed with serial measurements of compound motor action potential amplitude and motor unit number estimation), suggesting that motor neuron loss occurs very early. This provides further support for the potential usefulness of newborn screening to identify patients prior to the period of greatest motor neuron loss, should a therapeutic intervention become available.

Regarding the issue of routine preconception and prenatal screening for SMA, the American College of Obstetricians and Gynecologists Committee on Genetics published an opinion recommending against preconception and prenatal screening for SMA in the general population, citing insufficient evidence as to its cost-effectiveness and concerns regarding the education of the public and physicians about the complex issues raised by prenatal diagnosis of this disorder. However, various other professional medical societies have stressed the importance of screening. To address this issue and create a consensus, a meeting was organized by National Institutes of Health in late 2009. The conclusion of the meeting was that a carrier screening is technically feasible but its implementation will require addressing broader issues of screening in general.