The spinal muscular atrophies (SMAs) have been historically conceptualized as hereditary disorders preferentially affecting anterior horn cells and selected motor cranial nerve nuclei.¹ As in all disorders caused or influenced by genetics, molecular biology has served to confound as much as clarify the nosology. We have become very aware that the historical boundaries of hereditary neuromuscular disease are inaccurate. Part of this confusion arises from phenotypic overlap. For example, although lower motor neuron (LMN) morbidity dominates most SMA phenotypes, upper motor neuron (UMN) features may occur in some forms of distal SMA. Conversely, hereditary spastic paraplegia is a predominantly UMN disorder but may have notable LMN features in some genotypes. Even more damaging to the historical nosology of hereditary neuromuscular disease is the discovery that mutations of a single gene may produce variable phenotypes that have been historically represented as two or more diseases (Table 8-1).

In this chapter, we will discuss the SMAs related to mutations of the survival motor neuron (SMN) gene, the non-SMN infantile forms of the disease, the rare childhood bulbar forms of motor neuron disease (MND), Hirayama disease, Kennedy disease, the distal SMAs, the scapuloperoneal forms of SMA, and the uncommon SMA phenotypes that occur in association with multisystem disorders (Tables 8-2 and 8-3).^2–6 We emphasize this predominantly phenotypic classification as this remains, for the most part, the most practical means by which these disorders are recognized if not diagnosed.

The history of SMA dates to the independent descriptions of children with progressive weakness by Werdnig and Hoffman in the last decade of the 19th century.^7,8 Ironically, their cases would be classified today as SMA II rather than the more severe infantile form (SMA I) that bears their eponym. The molecular genetic era in SMA began in earnest in 1990 when a gene locus 5q13 was linked to the majority of childhood onset SMA cases.⁹ In 1995, deficiency of the survival motor neuron protein type 1 (SMN 1) was identified in approximately 95% of cases as the primary cause of the disease.¹⁰

SMA related to mutations of the SMN 1 is currently classified into five types, SMA I–IV with SMA III subdivided into SMA IIIa and SMA IIIb.¹¹ SMA I–III represent the traditional infantile, intermediate, and juvenile forms of the disease. SMA IV is the adult form of the disease which is less frequently associated with SMN mutations than other forms.¹² The childhood SMAs have been historically distinguished from one another, by age of onset and the milestones achieved rather than by significant phenotypic differences. SMN-related SMAs are recessively inherited. They do not differ significantly by phenotype, only by severity which is in turn related to contributions of the SMN 2 gene as will be described subsequently. They can be considered as a continuum of a single disorder.¹³ Progression and life expectancy correlate with age of onset which in turn correlates with the genetic signature as described below. Age of onset and clinical course tend to be similar in siblings.¹³

CLINICAL FEATURES

Werdnig–Hoffman disease or SMA I is the most common form of MND and the prototype of these disorders. Its incidence is estimated to occur in a range of four to ten × 10⁵ live births, depending on the geographic cohort studied.¹³ Clinical manifestations are evident within the first 6 months of life. Affected infants are hypotonic with a symmetric, generalized, or proximally predominant pattern of weakness. The legs are usually affected to a greater degree than the arms (Fig. 8-1). As in most MNDs, facial weakness is mild and extraocular muscles are spared. Fasciculations are seen in the tongue but rarely in limb muscles, presumably due to the ample subcutaneous tissue of neonates. Manual tremor, characteristic of SMA II and SMA III, occurs uncommonly. Deep tendon reflexes are typically absent. Abdominal breathing, and bulbar symptoms such as a weak cry, poor suck and feeding, and impaired secretion clearance are commonplace. The characteristic appearance includes pectus excavatum with a diminished anterior–posterior diameter of the chest, a bell-shaped chest, and a protuberant abdomen. These features are due to the relative diaphragmatic sparing in comparison to external intercostals early in the disease course. Mild contractures may occur, but arthrogryposis is not part of the classic phenotype. There is no intellectual impairment. Children with SMA I never develop the capability of independent sitting. Without mechanical ventilation, the large majority die in the first two years of life usually as a direct or indirect consequence of bulbar and/or ventilatory muscle weakness. Eight percent of individuals will survive to 10 years of age. A 20-year lifespan is unexpected.¹⁴

SMA II or the intermediate form of childhood SMA typically manifests between 6 and 18 months of age.¹⁵ The disorder is clinically defined by milestone acquisition, that is, a child who sits independently but never walks. Postural hand tremor is the only significant phenotypic variance from Werdnig–Hoffman disease. Tongue fasciculations, areflexia, manual tremor, and a symmetric, proximally predominant pattern of weakness characterize the SMA II phenotype. Symptoms related to impaired bulbar function are less of an issue than in SMA I. Approximately 98% of these individuals survive to the age of 5 years and two-thirds to the age of 25 years. In view of the more protracted course and the ability to sit, patients with SMA II and SMA III patients commonly acquire kyphoscoliosis and joint contractures.

SMA III is also referred to as the Kugelberg–Welander disease or the juvenile-onset SMA.¹⁵ It differs clinically from the intermediate form by age of onset, life expectancy, and milestones achieved. SMA IIIa is distinguished from SMA IIIb predominantly by age of onset, the former defined by symptom onset between 18 months and 3 years and the latter with symptom between 3 and 21 years. Afflicted individuals develop the ability to stand and walk which are subsequently lost in childhood, adolescence, or adulthood. Initial symptoms are referable to weakness of proximal leg muscles in the vast majority of cases. For example, the patient depicted in Figure 8-2, is now 30 and still capable of standing, became aware of his problem at age 14 when the crouched position of a hockey goalie became difficult to maintain. In SMA IIIa, 70% of patients are capable of walking 10 years after symptom onset and 20% at 40 years. In SMA IIIb, almost all patients walk at 10 years and 60% of patients remain ambulatory 40 years after symptom onset.¹⁵ Life expectancy in SMA IIIb extends into the sixth decade and may be normal in many individuals. Like SMA II, hand tremor, areflexia, and tongue fasciculations, are commonplace in SMA III. Presumably related to the older age of these patients, and the diminished proportion of subcutaneous tissue, limb fasciculations are more evident in SMA III than in SMA I and SMA II.¹⁴

Recessively or dominantly inherited adult-onset SMA or SMA IV is uncommon.^16,17 Even though X-linked spinobulbar muscular atrophy is an adult-onset disorder manifesting with the same proximally predominant, symmetric pattern of weakness, it has both distinctive clinical and genetic features and will not be considered as SMA IV in this text.

SMA IV patients do not typically become aware of weakness until age 21 years or older.¹⁸ As with other SMAs, initial symptoms are typically referable to proximal lower extremity muscles. Hip flexors and extensors and knee extensors are usually the most severely affected muscles. The shoulder abductors and elbow extensors are the most affected muscles of the arms. Tongue and limb fasciculations, hand tremor, and, in some cases, calf hypertrophy occur.¹⁸ The latter can be confounding, particularly in males, as myopathies are a more common cause of proximal weakness in this age group. Life expectancy is normal.¹⁸

DIFFERENTIAL DIAGNOSIS

The differential diagnosis of SMA I is the differential diagnosis of the floppy infant (Table 8-3).¹⁹ The majority of hypotonic neonates have a central nervous system disorder. Clinical clues implicating a potential but less common neuromuscular cause of a floppy infant include preservation of alertness, depressed or absent deep tendon reflexes, the pattern of weakness, and fasciculations if present. At least two other forms of non-SMN infantile SMA are known to exist. Recessively inherited spinal muscular atrophy with respiratory distress type 1 (SMARD1) and X-linked infantile SMA with arthrogryposis will be described subsequently.^20–22 Neonatal or congenital myasthenia, congenital muscular dystrophy, neonatal myotonic dystrophy, infantile Pompe disease, severe nemaline, myotubular or other congenital myopathies, infantile botulism, and rare hypomyelinating neuropathies are the major neuromuscular considerations in a hypotonic infant.

SMN-related SMA II–IV need to be distinguished from dominantly inherited SMAs including the Finkel type associated with mutations of the VAPB (fALS8), a disorder linked to the 14q32 locus, mutations of the lamin A/C gene (LMNA) and a disorder described in two Finnish families in which the gene has yet to be identified.^23–26 The differential also includes a wide variety of myopathic disorders, including certain muscular dystrophies (dystrophinopathies, limb-girdle, myotonic, congenital and Emery-Dreifuss), congenital myopathies; mitochondrial disorders; and lipid and glycogen storage disorders. Chronic inflammatory demyelinating polyradiculoneuropathy would be the primary neuropathic consideration. Congenital myasthenic syndromes should also be considered.

LABORATORY FEATURES

In Werdnig–Hoffman disease, creatine kinase (CK) is elevated, typically less than five times the upper limits of normal. In a patient with an SMA I–IV phenotype the most expeditious means to confirm the diagnosis is through SMN 1-targeted mutation analysis which identifies the exon 7 deletion, this is the genetic defect in the majority of SMA I–III, and some SMA IV cases.^10,27,28 This test will identify a mutation in approximately 95% of childhood and adolescence patients with an SMA phenotype and is felt to be nearly 100% specific.^16,29,30 In the remaining patients, sequence analysis may be performed to identify other mutations.

Patients with recessively inherited SMA IV are associated with homozygous SMN 1 deletions infrequently.^16,29,30 Analysis of the vesicle-associated membrane protein-associated protein B (VAPB) gene on chromosome 20, allelic to familial ALS type 8, may provide diagnostic confirmation in some patients with dominantly inherited SMA IV. Asymptomatic adults with SMN 1 genotypes have been described.^14,30–32 Although SMN mutations are highly specific for SMA, a phenotype suggesting congenital axonal neuropathy with sensory involvement has been described.^33–35

Carrier detection and prenatal testing for SMA I–III are available through gene dosage analysis although the results need to be interpreted cautiously. A mutation detected in only one parent is reassuring but does not guarantee healthy children. Although 98% of SMA children have parents who are each heterozygotes for the SMN 1 mutation, an SMA child born of only one identified heterozygote parent can occur. This can result from either a spontaneous mutation of the child’s second allele, through germline mosaicism in the seemingly normal parent or from false paternity. Interpretation of carrier testing is also complicated by consideration that both SMN 1 copies may exist on a single chromosome in 4% of individuals.³²

ELECTRODIAGNOSIS

Historically, electrodiagnosis (EDX) was the major diagnostic tool used to support the clinical diagnosis of childhood SMA. This has been supplanted by genetic testing in the majority of cases. EMG is primarily used in individuals with a SMA phenotype without a detectable SMN mutation or in individuals who have neuromuscular disorders originating from muscle, neuromuscular junction or nerve that may resemble the SMA phenotype. With SMA or other anterior horn cell diseases, the electromyographer would anticipate a characteristic pattern of abnormal parameters. These would include low-amplitude compound muscle action potentials (CMAPs), normal sensory nerve action potential, and widespread evidence of both ongoing denervation (spontaneous discharge of fibrillation potentials and positive waves) and chronic partial denervation, and reinnervation (reduced numbers of motor unit potentials of increased amplitude and duration with muscle activation). Fasciculation potentials may or may not be identified in part because of the necessary brevity of the needle examination in many children.

EDX has a limited role in the determination of SMA prognosis. The density and geographic distribution of fibrillation potentials in comparison to changes of chronic partial denervation and reinnervation is related to the rapidity with which these disorders progress. Although pragmatically difficult to apply to the pediatric patient, motor unit instability and the rate of decline of motor unit number estimation may also provide prognostic insight.^36–38

HISTOPATHOLOGY

The SMN-related SMAs are attributed to anterior horn cell pathology. This observation dates back to the original writings of Werdnig and Hoffman. In SMA, swelling of motor neurons laden with phosphorylated neurofilaments and glial bundles within ventral roots are common. The ubiquitinated inclusions of ALS are not seen.³⁹

As with EDX, the role of muscle biopsy in SMA has greatly diminished. For all intents and purposes, EDX will arrive at the same conclusion provided by the arguably more invasive muscle biopsy. EDX has the additional advantage of more readily demonstrating the geographic distribution of these findings. In SMA I, the biopsy will demonstrate sheets of rounded, atrophic fibers of both types. Hypertrophic fibers are intermixed and are exclusively type I (Fig. 8-3). Type grouping is uncommon. In SMA II, the biopsy may be similar to SMA I or may differ because of the presence of hypertrophic type II fibers and/or the presence of type grouping. In SMA III, type grouping and group atrophy of both fiber types are common. In addition, as with many chronic neurogenic disorders, “pseudomyopathic” features such as fiber splitting, increased endomysial connective tissue, and an increased number of internal nuclei may be seen.

PATHOGENESIS

SMN-related SMA is caused by a loss of function effect due to deficiency of the SMN 1 protein caused in most cases by large deletions of exon 7 or 8 or truncation of the SMN 1 gene.¹⁷ The SMN proteins are found in both the nucleus and cytoplasm of all cells where they have RNA processing functions. The SMN 1 protein appears to interact with a number of cytoplasmic proteins to facilitate the formation, nuclear importation, and regeneration of nuclear spliceosomal RNA.⁴⁰ The SMN 1 protein is also found to traffic in motor axons and may play a role in disordered axonal transport through its influence on β-actin mRNA.¹⁷

The severity of the SMA phenotype is related in part to the number of copies of the similar, but unstable and significantly less effective than the allelic protein, SMN 2.^40–42 The SMN 2 gene is identical to SMN 1 with the exception that exon 7 is excluded. SMA 0 is typically associated with one copy of the SMN 2 gene, SMA I with two copies, SMA IIIa with three or four copies and SMA IIIb invariably with four, and recessively inherited SMA IV with anywhere between four and eight gene copies.¹¹ Individuals homozygous for the SMN 1 mutation with five copies of the SMN 2 gene have been reported to be asymptomatic.^14,30–32 SMN 2 gene copy number is not the sole determinant of phenotypic severity. There are other complex genetic influences that are not as yet fully understood. Although 95% of affected individuals have homozygous mutations, 5% have more complex compound heterozygotic mutations with a typical deletion in one allele with a subtle intragenic defect on the other.⁴³ Prognostication based on SMN 2 gene copy number in SMN homozygotes should proceed cautiously. There are no known clinical consequences from mutations of the SMN 2 gene alone.

Unlike SMA 0–III, the SMA IV phenotype may be inherited in either dominant or recessive fashion. Autosomal-dominant inheritance, referred to as the Finkel type, is estimated to occur in approximately 30% of these patients.⁴⁴ This phenotype is associated with a mutation of the VAPB gene. Autosomal-dominant SMA IV is allelic to ALS 8 as some families with VAPB mutations will have UMN in addition to the more characteristic LMN findings.

MANAGEMENT

In the past decade, a number of pharmacological agents have been utilized in an attempt to increase SMN protein levels. Aminoglycosides, quinazoline derivatives and drugs that can inhibit the enzyme histone deacetylase have been used in animal models and in some cases in humans. Although valproic acid, sodium butyrate, phenylbutyrate, and trichstatin A can activate the SMN 2 promoter in vivo, no agent has demonstrated clinical benefit to date. Current strategies include the use of antisense oligonucleotides in an attempt to incorporate exon 7 into the SMN 2 gene or viral vectors to transfect the entire SMN 1 gene into the DNA of afflicted individuals. Mouse models utilizing gene therapy as well as intrathecal embryonic stem cell transplants have shown considerable promise.¹⁷

In 2007, guidelines for the care of the SMA patients were published based on expert consensus.¹⁸ Readers are referred to this document for further detail. In summary, physicians are urged to provide parents, and when applicable the patient, information related to the natural history of the disease, genetic implications, and the role and availability of clinical trials. Like ALS, a multidisciplinary clinic with representation from disciplines that are familiar with the management of the nutritional, psychosocial, mobility/orthopedic, and ventilatory consequences of the disease in this age group is the recommended care model. The development of kyphoscoliosis is a common problem in children who become wheelchair bound. Spine stabilization is commonly recommended in individuals whose curves exceed 50 degrees and whose vital capacities exceed 40% of the predicted normal value. The goals of this intervention are patient comfort, ease of patient management, and potential stabilization of restrictive pulmonary deficits.⁴⁵ Tracheostomy assisted long-term mechanical ventilation, and percutaneous gastrostomy feeding tube insertion are decisions with enormous emotional and financial consequence to the parents of an affected child. Noninvasive positive pressure ventilation may provide an improved quality and duration of life in child with symptoms of ventilatory insufficiency until a decision regarding tracheostomy is required.

One genetic issue of particular importance in SMA and other heritable pediatric neuromuscular disease is the role of presymptomatic testing in siblings of affected children. The current ethical perspective, at least in the United States, is that no child should undergo presymptomatic genetic testing until they are of the age to make this decision themselves, unless a treatment that makes a meaningful difference in the natural history of the child’s disease exists.

Infantile SMA with arthrogryposis, previously referred to by some as SMA 0 is an X-linked disorder associated with a mutation of the ubiquitin-activating enzyme 1 gene (UBE1) (Xp11.23). Its phenotype is very similar to SMN 1 with the exception that contractures are present at birth or early in development.^17,46

SMARD1 is also referred to as hereditary motor neuropathy (HMN) type VI or distal SMA type I as the pattern of weakness typically affects distal more than proximal muscles.^15,17,46 Its onset is in infancy and as the name implies, is associated with compromise of ventilatory muscles, particularly the diaphragm. Understandably, without mechanical ventilation, life expectancy is limited to months in most cases. It is a recessively inherited disorder resulting from a mutation of the immunoglobulin mu binding protein 2 gene (IGHMPP2) at locus 11q13.2–q13.4.

Recessively inherited SMA may occur in association with pontocerebellar hypoplasia.^15,17,46 The phenotype of this congenital or infantile onset disorder may include microcephaly, mental retardation, nystagmus, upper limb ataxia, or in some cases arthrogryposis. It occurs as a result of mutations of the vaccinia-related kinase 1 gene (VRK1) located at locus 14q32. Recessively inherited SMA may also result from abnormal mitochondrial function due to mutations of the cytochrome oxidase 2 (SCO2) gene on locus 22q13. In addition to clinical features suggestive of SMA I, an associated cardiomyopathy with lactic acidosis may occur.¹⁷ There is also a lethal arthrogryposis with anterior horn cell disorder (LAAHD) associated with severe facial deformities resulting from mutations of the GLE1 gene at locus 9q34.11.⁴⁶

The syndromes of Fazio Londe and Brown–Vialetto–Van Laere (BVVLS) historically describe two bulbar syndromes, clinically distinguished from one another by the sensorineural hearing loss that occurs in the latter. Disease onset is typically within the first two decades of life. In BVVLS, the hearing loss typically precedes the development of bulbar and limb weakness.⁴⁷ These disorders are complex with involvement of multiple neurologic systems which typically include LMN weakness of the limbs and ventilatory muscles. Less common UMN signs, as well as ataxia, optic atrophy, retinal pigmentary degeneration, seizures, and dysautonomia occur. The brainstem syndrome is dominated by involvement of cranial nerves VII, IX, and XII which are more commonly affected than III, V, or VI. The disorders are usually inherited in a recessive manner and are linked to a mutation at the C20ORF54 locus.⁴⁸ The clinical course is usually progressive and fatal within years of onset. A singular case has been reported in which treatment with a low fat diet, riboflavin, carnitine, 3-hydroxybutyrate, and glycine seemed to have provided temporary clinical stabilization.⁴⁹