23 The motor neurone disorders

The motor neurone diseases are a group of disorders in which there is selective loss of function of upper and/or lower motor neurones in the motor cortex, brainstem, and spinal cord resulting in impairment in the nervous system control of voluntary movement. The term ‘motor neurone disease’, often abbreviated to ‘MND’, is used differently in different countries. In the United Kingdom it is used as an umbrella term to cover the related group of neurodegenerative disorders including amyotrophic lateral sclerosis, the commonest variant, as well as progressive muscular atrophy, primary lateral sclerosis, and progressive bulbar palsy. However, in many other countries amyotrophi\c lateral sclerosis, referred to as ALS, has been adopted as the umbrella term for this group of clinical variants of motor system degeneration. There is a tendency now internationally to use the ALS/MND abbreviation to cover this group of conditions. Careful diagnosis within the motor neurone diseases is essential for advising about prognosis, potential genetic implications, and for identifying those with acquired lower motor neurone syndromes who may benefit for the administration of immunomodulatory therapy. Table 23.1 classifies the motor neurone disorders according to whether the upper motor neurone, the lower motor neurone, or both groups of cells are affected.

In patients with motor neurone diseases, precise differential diagnosis requires clinical and neurophysiological assessment as to whether the disorder involves the upper motor neurones, the lower motor neurones, or both. Consideration should then be given to factors such as age of onset, rapidity of progression, evidence of an inherited disorder, and the anatomical distribution of the clinical features to determine the likely diagnosis. As a general rule, the sensory system, sphincter control, and cognitive function are usually preserved in motor neurone diseases.

The clinical features of lower motor neurone involvement are muscle wasting, fasciculations, and flaccid weakness. Tendon reflexes are usually preserved until the muscle denervation is severe. In amyotrophic lateral sclerosis, the upper motor neurone component tends to preserve the reflexes and the preservation of reflexes in a limb with severe muscular atrophy may be an important clue to the diagnosis. Motor neurone diseases will usually give rise to significant denervation atrophy of weakened muscles. If muscle bulk is relatively preserved in a weak muscle it should raise the possibility of conduction block rather than denervation, as found in multifocal motor neuropathy with conduction block. Fasciculations are visible flickering movements within the muscle belly which are insufficient to produce movement at the joint and are most commonly observed in the large proximal muscle such as deltoid and quadriceps which have large motor units. Fasciculations which are not visible clinically may be detected by electromyography. Fasciculations can only be regarded as indicative of motor neurone disease if associated with evidence of denervation by clinical or electrophysiological assessment.

In patients with suspected motor neurone disease, nerve conduction studies will rule out sensorimotor polyneuropathy; pure motor demyelinating neuropathies or multifocal motor neuropathy with conduction block. Maximal motor conduction velocity is often reduced in nerves supplying denervated muscles as a consequence of degeneration of large motor axons. However, in amyotrophic lateral sclerosis the motor conduction velocity rarely falls below 80 per cent of the lower limit of normal and F waves or distal motor latencies rarely exceed 1.25 times the upper limit of normal. Results outside these limits should raise the possibility of a primary demyelinating motor neuropathy. Electromyography helps to distinguish denervation from myopathy and may also detect subclinical denervation in limbs which are clinically normal in patients with amyotrophic lateral sclerosis. Muscle biopsy may occasionally be required to rule out myopathy, particularly in patients with atypical features or slowly progressive proximal weakness.

Signs of upper motor neurone involvement in motor neuron diseases include increased muscle tone, clonus, pyramidal distribution weakness, and extensor plantar responses. In many patients with amyotrophic lateral sclerosis the presence of clonus and extensor plantar responses is often obscured by the profound denervation changes in the distal lower limb muscles, thereby obscuring clinical confirmation of upper motor neurone involvement. Changes in threshold and parameters of central motor conduction following transcranial magnetic stimulation of the motor cortex may be seen in patients with amyotrophic lateral sclerosis, but the changes observed are complex and may vary with the stage or anatomy of the disease and are insufficiently sensitive to provide a robust and useful tool for detecting subclinical involvement of the upper motor neurone.

The group of patients with amyotrophic lateral sclerosis who do appear to have a pure lower motor neurone disorder without clear clinical signs of upper motor neurone involvement, present a diagnostic challenge. In this situation it is important to exclude treatable lower motor neurone disorders such as multifocal motor neuropathy with conduction block and a trial of immunomodulatory therapy may be indicated where diagnostic uncertainty persists. The passage of time may resolve the diagnostic uncertainty as patients with the pure lower motor neurone variant of amyotrophic lateral sclerosis, known as progressive muscular atrophy, will typically show a brisk rate of disease progression.

The commonest diagnostic problem is to distinguish amyotrophic lateral sclerosis from other motor neurone disorders carrying a better prognosis, particularly if signs of upper motor neurone involvement are lacking. Multifocal motor neuropathy with conduction block usually develops over many years, tends to present with asymmetrical involvement of the upper limbs, may present with difficulty in extending one or more fingers, is associated with electrophysiological evidence of motor nerve conduction slowing or block, and the patient may have antiganglioside antibodies detectable in blood (Section 21.11.3). Patients with post-polio syndrome tend to deteriorate rather slowly by comparison with amyotrophic lateral sclerosis and present with slight further loss of neuromuscular function some decades after an earlier attack of acute poliomyelitis (Section 23.3.7). Kennedy’s disease should be suspected in patients who appear to have amyotrophic lateral sclerosis with bulbar involvement, but who do not deteriorate as rapidly as expected (Section 23.5.1). In men with pure lower motor neurone disorders the presence of gynaecomastia should always be looked for. Other diagnostic alerts for Kennedy’s disease include the presence of mentalis contractions or fasciculations and clinical or neurophysiological evidence of sensory nerve involvement.

Muscle fasciculations are common in the normal population and are frequently observed in the calf muscles after exercise. Patients with benign fasciculations who come to neurological attention are frequently individuals with medical knowledge who fear the development of motor neurone disease, or those with a family history of amyotrophic lateral sclerosis. Benign fasciculations are most commonly observed in the calf muscles, are not associated with any other abnormal clinical signs and electromyography shows no evidence of denervation.

Historical descriptions of amyotrophic lateral sclerosis, or ALS, were reported as early as 1824 by Charles Bell and others (Goldblatt 1968; Tyler and Sheffner 1991). However, Charcot was the first to describe the condition in detail, based on careful clinico-pathological correlation and he coined the term amyotrophic lateral sclerosis (Charcot and Joffroy 1869). Motor neurone disease, or MND, as used by Brain in the first edition of this textbook (Brain 1933), is the name used for the disease most commonly in the United Kingdom and represents an umbrella term encompassing amyotrophic lateral sclerosis and three other disorders that are considered its clinical variants: primary lateral sclerosis, progressive muscular atrophy, and progressive bulbar palsy. In the USA, amyotrophic lateral sclerosis is sometimes known as Lou Gehrig’s disease after the famous Yankee baseball player who developed the disease at the peak of his sporting career and died in 1941. Amyotrophic lateral sclerosis is sometimes referred to as ALS/MND.

Amyotrophic lateral sclerosis is a neurodegenerative disorder that causes progressive injury and cell death of lower motor neurones within the brainstem and spinal cord and upper motor neurones in the motor cortex. The disease has an incidence of about 2 per 100 000 and a prevalence of 6–8 /100 000 (Chancellor and Warlow 1992; Traynor et al. 1999). There are approximately 5000 individuals with amyotrophic lateral sclerosis at any one time in the UK. The incidence is fairly uniform throughout the world, with the exception of a few high incidence foci, for example, on the Kii peninsula of Japan and the Western Pacific island of Guam. The disease predominantly affects middle-aged and elderly individuals, with a mean age of onset of 55 years, though younger individuals can also be affected. For reasons that are not understood, men are affected more commonly than women, with a male/female ratio of approximately 1.6/1. Motor neurone disease is sporadic in 90 per cent of cases, but in 5–10 per cent of cases the disease is familial, usually with an autosomal dominant mode of inheritance.

Amyotrophic lateral sclerosis is characterized and defined by the presence of clinical features reflecting degeneration of upper motor neurones of the cerebral cortex, and lower motor neurones of the brainstem and spinal cord. Lower motor neurone degeneration causes weakness, atrophy, and fasciculation of the limb and bulbar musculature. Muscle cramps are a common symptom, and patients may be aware of fasciculation as twitching or flickering movements of their muscles. Clinical signs resulting from degeneration of the upper motor neurones include the incongruous presence of active or brisk tendon reflexes in a wasted limb, increased muscle tone, and sometimes the presence of Hoffman’s or Babinski’s signs. Upper motor neurone dysfunction within the bulbar territory causes pseudobulbar palsy, where the snout and jaw reflexes may be exaggerated, and increased tone within the bulbar muscles may cause slowing of repetitive movements of the tongue as well as strained effortful speech. Emotional lability, producing difficulty in controlling episodes of laughing or crying, often accompanies upper motor neurone signs in the bulbar region.

In amyotrophic lateral sclerosis the disease often starts focally and asymmetrically in the upper limb, lower limb, or bulbar territories. This is followed by progressive spread of injury to contiguous groups of motor neurones, so that the patient’s clinical features often show an anatomically logical progression. By the end of the disease course most patients will have features of upper and lower motor neurone dysfunction affecting all four limbs and the bulbar musculature.

The first clinical problem is present in the upper or lower limbs in approximately 75 per cent of patients. At first asymmetry or unilaterality of symptoms and signs is common and the first noticeable problem for the patient commonly involves the distal limb muscles. Affected individuals may notice weakness, wasting or clumsiness of one hand, or simply difficulty with everyday actions such as turning a key or lifting a heavy object. In the lower limbs, foot drop or a tendency to trip over the toes of one foot is a common presenting symptom. The gait may become slowed or clumsy due to weakness and spasticity of one or both lower limbs. The patient may be aware of limb muscle cramps which can sometimes precede other clinical features of amyotrophic lateral sclerosis by months or years. In addition, muscle fasciculation may be noticed by the patient, particularly arising from the large proximal limb muscles such as biceps, triceps, pectoralis major, and quadriceps. As the condition progresses, patients often develop a characteristic pattern of limb muscle weakness. In the upper limbs, the intrinsic hand muscles, particularly the thenar group, tend to be affected severely and early, whereas other muscle groups such as triceps and the finger flexors are relatively spared until late in the disease course. In the lower limbs, the pattern of weakness is often in a pyramidal distribution, with flexors weaker than extensors, early weakness of hip flexion and ankle dorsiflexion and often more severe weakness of the distal muscles. Examination of the patient will frequently reveal a combination of upper and lower motor neurone features with proximal fasciculation, muscle wasting, particularly distally (Fig. 23.1), the characteristic pattern of weakness described above, but in addition tone may be mildly increased, the reflexes are brisk and extensor plantar responses may be present.

Bulbar symptoms are the presenting clinical problem in approximately 25 per cent of patients with amyotrophic lateral sclerosis and this presentation is particularly common in elderly women. The onset of amyotrophic lateral sclerosis in the bulbar regions has a less favourable prognosis than limb-onset disease (Chancellor et al. 1993; del Aguila et al. 2003). The first problem is usually slurring of speech which initially may only be apparent when the individual is tired at the end of the day, after prolonged use of the voice, or after alcohol. The dysarthria may rarely be accompanied by dysphonia. The patient may notice particular difficulty with specific sounds such as ‘s’ and may notice other features of bulbar muscle weakness such as difficulty in pursing the lips or whistling. The onset of dysarthria is often initially attributed to a ‘minor stroke’, but it soon becomes apparent that the speech difficulty is a progressive problem. Patients with amyotrophic lateral sclerosis often have a mixed spastic/flaccid dysarthria. The speech develops a tight, strangled quality due to the upper motor neurone component, but superimposed on this, the flaccid lower motor neurone weakness of the palate and nasopharynx give the speech a nasal quality. Examination of the patient will often reveal a combination of upper and lower motor neurone bulbar signs with weakness of the facial

muscles, a spastic, weak, wasted, and fasciculating tongue and a brisk jaw jerk. As the disease progresses the patient may develop weakness of the palate and muscles of mastication. Occasionally forcible jaw clenching is a troublesome symptom causing trauma to the mouth. Dysphagia becomes a problem later, often several weeks or months after the onset of the speech problem. Initially dysphagia tends to be more pronounced for liquids compared to solids. Gradually certain food items such as lettuce, foods with a crumbly texture, or items with a sharp taste become difficult to swallow and may cause choking episodes. Mealtimes may become prolonged and arduous, with frequent episodes of coughing. When dysphagia reaches a significant level, the patient will have a problem with excess saliva pooling in the mouth. If accompanied by lower facial weakness, the patient will suffer from drooling which causes considerable social embarrassment. As the bulbar dysfunction becomes increasingly severe, the patient may develop complete anarthria. Severe dysphagia causes the patient to be at risk of aspiration and chest infection.

Weight loss. This is a common feature of amyotrophic lateral sclerosis. Multiple factors may contribute to this, with loss of muscle bulk due to amyotrophy, difficulty maintaining nutrition in the face of dysphagia, and loss of appetite resulting from reactive anxiety and depression or from immobility. Severe weight loss and nutritional deficiencies may themselves exacerbate muscle weakness.

Neck muscle weakness. This is common later in the course of disease, causing difficulty holding the head upright, the ‘dropped head syndrome’. This problem commonly causes pain and increases difficulties with swallowing and communication.

Respiratory muscle weakness. Occasionally, onset of the disease in motor neurones innervating the respiratory muscles causes breathlessness as a presenting symptom (Nightingale et al. 1982; De Carvalho et al. 1996). More commonly, respiratory failure develops insidiously during the course of the disease, causing dyspnoea and orthopnoea. Significant diaphragmatic weakness may be apparent from paradoxical movement of the abdominal wall during inspiration and from a marked decline in the forced vital capacity when measured in the supine compared to the upright position. Symptoms of nocturnal carbon dioxide retention may develop, including morning headaches, anorexia, and daytime somnolence.

Extramotor features. Amyotrophic lateral sclerosis is traditionally considered to be a pure motor disorder, which spares sensation, cognition, and autonomic function. However, there is increasing evidence that, despite the relative vulnerability of motor neurones to the degenerative process, other types of neurones are also affected. Thus, amyotrophic lateral sclerosis is now regarded as a multisystem neurodegenerative disorder in which the earliest and most severe degeneration tends to involve motor neurones. In most patients, the evolution of motor dysfunction is lethal before the development of overt signs of central nervous system pathology in other regions, although occasional patients may spontaneously develop a severe multisystem degeneration (Machida et al. 1999). Patients whose survival is prolonged by ventilatory support may develop widespread features of extramotor system involvement.

Cognitive impairment. Frontotemporal dementia (Section 34.6.5) encompasses a range of clinical features that result from atrophy of the frontal and anterior temporal regions of the brain. In the frontal variant of frontotemporal dementia, there is prominent personality change and conduct disorder, and affected individuals often become disinhibited or apathetic with emotional blunting, and loss of insight. Behaviour tends to become ritualized or stereotypical, and eating behaviour often becomes abnormal. Visuo-spatial function and memory are relatively preserved. Two syndromes of progressive language dysfunction are seen with frontotemporal dementia. Primary progressive aphasia, which may progress to mutism, is characterized by effortful speech production with phonological and grammatical errors, and word retrieval difficulties. Patients may also develop semantic dementia, with impairment of naming and word comprehension but preservation of fluent, grammatical speech output (Neary et al. 1998; McKhann et al. 2001; Hodges et al. 2004). There is overlap between these clinical syndromes, as patients with either syndrome may later develop deterioration in personality and behaviour (Bak et al. 2001), while language abnormalities may be found in patients with frontotemporal dementia (Neary et al. 1990; Strong et al. 1996).

Approximately 5 per cent of patients with motor neurone disease will develop overt features of frontotemporal dementia (Hudson 1981). Cognitive dysfunction may precede, follow, or coincide with the features of motor dysfunction. The most common symptoms seen in frontotemporal dementia in association with motor neurone disease are progressive deterioration in personality and behaviour (Ringholz et al. 2005), but primary progressive aphasia is also described.

Motor neurone disease patients without overt dementia may show more subtle features of frontal lobe dysfunction. This problem may be under-recognized in the normal clinic setting, partly because of the difficulty in assessing cognitive function in patients with bulbar dysfunction and severe motor deficits. Detailed neuropsychological assessment demonstrates that up to 50 per cent of patients with sporadic motor neurone disease develop features of frontal lobe dysfunction (Ringholz et al. 2005), There is evidence indicating that cognitive impairment may be more common in patients with primary lateral sclerosis or predominant upper motor neurone signs (Caselli et al. 1995), and in those with prominent bulbar involvement (Schreiber et al. 2005).

Parkinsonism. This is described in cases of sporadic motor neurone disease, and occurs more frequently in cases with accompanying dementia (Qureshi et al. 1996). The prevalence of Parkinsonism in motor neurone disease may be underestimated, as extrapyramidal features can be masked in patients with muscle weakness, wasting, and spasticity. Subclinical defects in dopaminergic transmission have been demonstrated by positron emission tomography scanning in patients with sporadic motor neurone disease, without clinical evidence of an extrapyramidal disorder (Takahashi et al. 1993).

Disorders of eye movement. Supranuclear ophthalmoplegia, including limitation of ocular movement, slow ocular movement, and spasmodic gaze fixation, have been described in motor neurone disease. Although the majority of these patients develop these features after a disease course extended by invasive ventilation, they can occasionally be seen in ‘natural disease’ (Hayashi et al. 1989; Mizutani et al. 1990). Difficulty initiating eyelid opening, with preserved reflex eyelid movements, often referred to as ‘eyelid apraxia’, is also seen in patients with motor neurone disease, and may be associated with supranuclear vertical gaze impairment (Abe et al. 1995).

Sensory impairment. Patients with amyotrophic lateral sclerosis not infrequently complain of non-specific sensory phenomena, such as tingling in the fingers. Sensory examination is usually normal, but nerve conduction studies and somatosensory-evoked potentials reveal the presence of abnormalities in peripheral and central sensory pathways in up to 60 per cent of patients (Bosch et al. 1985; Shefner et al. 1991). There are isolated case reports of more profound sensory involvement in association with amyotrophic lateral sclerosis (Wakabayashi et al. 1998).

Selective sparing of specific motor neurone groups. The eye movements tend to be spared in motor neurone disease. Even in advanced disease, when patients would otherwise be ‘locked in’, often limited communication can be retained by movements of the eyes. Similarly, the strength of the pelvic floor muscles is relatively preserved, so that patients with motor neurone disease usually remain continent throughout the course of the disease. These selectively spared muscle groups reflect the fact that motor neurones in the oculomotor nuclei of the brainstem and Onuf’s nucleus in the sacral spinal cord are less vulnerable to the pathological process in motor neurone disease compared to those innervating limb and bulbar musculature.

There are several clinical variants that describe the predominant clinical features of the patient at the time of presentation. As the disease progresses, however, the majority of patients will develop features of upper and lower motor neurone degeneration affecting the limbs and bulbar musculature, the most common variant of amyotrophic lateral sclerosis.

Progressive muscular atrophy. Approximately 5–10 per cent of patients with motor neurone disease will present with clinical features reflecting only degeneration of lower motor neurone groups in the spinal cord, in the absence of any evidence of upper motor neurone pathology. Some of these patients will later develop brisk reflexes or extensor plantar responses. Clinical examination may not reliably detect upper motor neurone involvement late in the disease, when there is severe global weakness and amyotrophy. It is noteworthy that in those patients who develop no clinical upper motor neurone signs during life, approximately 50 per cent will have pathological evidence of corticospinal tract pathology at autopsy (Ince et al. 2003). This suggests that progressive muscular atrophy is part of the same pathological spectrum of disease as amyotrophic lateral sclerosis.

Primary lateral sclerosis. Primary lateral sclerosis, first described by Erb (1875), is a degenerative disorder of the upper motor neurone pathways causing spasticity of the limb and bulbar muscles. It is usually a sporadic disorder of insidious onset, and commonly starts in the fifth decade or later as a spastic paraparesis. Bulbar or upper limb onset has also been described. The course is gradually progressive, and although patients ultimately develop a severe spastic spinobulbar paresis, survival is usually prolonged compared to patients with classical amyotrophic lateral sclerosis (Pringle et al. 1992). Many patients with primary lateral sclerosis will survive more than 10 to 15 years after symptom onset.

Not all case series of primary lateral sclerosis have adequately excluded alternative diagnoses, particularly prior to the availability of modern imaging techniques. Pringle and colleagues described a series of 8 patients with progressive symmetrical spinobulbar spasticity in whom alternative diagnoses were carefully excluded, and proposed diagnostic criteria for primary lateral sclerosis (Pringle et al. 1992). Although the cardinal feature of primary lateral sclerosis is upper motor neurone involvement, in most case series described, there is usually evidence clinically or electrophysiologically of some lower motor neurone dysfunction, which may only develop after several years (Kuipers-Upmeijer et al. 2001; Le Forestier et al. 2001). The evolution of primary lateral sclerosis into amyotrophic lateral sclerosis has been described, with the onset of generalized amyotrophy developing after many years of a slowly progressive degeneration of the upper motor neurone system, (Bruyn et al. 1995), and primary lateral sclerosis may occur as a phenotypic manifestation of familial amyotrophic lateral sclerosis. These findings strongly indicate that primary lateral sclerosis is part of the spectrum of amyotrophic lateral sclerosis, with predominant but not exclusive degeneration of upper motor neurones.

Progressive bulbar palsy. Bulbar onset motor neurone disease occurs in approximately 25 percent of patients and is most common in women of middle to elderly age. Progressive bulbar palsy usually progresses to involve the limbs, although clinical and electrophysiological abnormalities affecting the limbs may not be found at the time of presentation.

Segmental variants. Several variants of amyotrophic lateral sclerosis have been described, in which the disease follows a more segmental pattern than is typical in classical disease. The flail arm syndrome, progressive amyotrophic diplegia, may occur in up to 10 per cent of patients with motor neurone disease (Fig. 23.2) (Gamez et al. 1999; Katz et al. 1999). It is much more common in men than

in women, and has a longer median survival than classical amyotrophic lateral sclerosis (Hu et al. 1998). It is characterized by profound symmetrical weakness and wasting of the upper limbs with hyporeflexia. Although signs of pyramidal tract involvement may be seen in the lower limbs, there is little or no functional impairment of the bulbar muscles or legs at presentation, and these regions of the motor system are usually involved only late in the disease course. A similar focal presentation in the lower limbs with progressive paraparesis is recognized. Other forms of segmental motor neurone degeneration are described in Section 23.3.4.

Madras form. A specific subtype of early onset sporadic amyotrophic lateral sclerosis was identified in Southern India in 1970 (Meenakshisundaram et al. 1970; Saha et al. 1997; Gourie-Devi and Nalini 2003b). The phenotype is that of young onset amyotrophic lateral sclerosis with bulbar and limb involvement. Lower motor neurone features tend to predominate, but upper motor neurone signs are present and include brisk tendon reflexes and extensor plantar responses. The onset is usually in the second or third decades and the disorder is more common in males. Typical early features comprise distal upper limb weakness with wasting, dysarthria, dysphagia, and progressive sensorineural hearing loss. Progressive deafness is found in more than 50 per cent of affected individuals and is an important feature distinguishing Madras amyotrophic lateral sclerosis from other forms with early sporadic onset. Spasticity and distal leg weakness frequently develop and the disease course is one of progressive disability but with long-term survival. Electromyography studies show chronic partial denervation. Pathological findings from one case have been described (Shankar et al. 2000). There was depletion of lower motor neurones from the spinal cord and brainstem; neuronal loss and gliosis in the cochlear nucleus; demyelination and axonal loss in the cochlear nerve, and degeneration of the corticospinal tracts with accompanying gliosis. There was no description of whether ubiquitinated inclusion bodies were present in surviving motor neurones.

Symptoms and signs are often focal at the time the patient with amyotrophic lateral sclerosis first presents to medical attention, though neurophysiological evidence of motor neurone injury may be more widespread. In the majority of cases there is an inexorable progression of the pathological process, often seemingly to contiguous groups of motor neurones. At the end stage of the disease, the patient will usually have significant motor dysfunction affecting all four limbs and the bulbar muscles as well as compromise of respiratory function. The patient may become entirely dependent upon nursing care, bulbar dysfunction may progress to anarthria, and enteral feeding may be required due to severe dysphagia.

Certain clinical features are associated with a worse prognosis for the clinical course of the disease including: older age at onset of symptoms; early compromise of respiratory function; bulbar onset of symptoms and more rapid presentation of the patient to medical attention (Chancellor et al. 1993; del Aguila et al. 2003; Millul et al. 2005). The mean survival from symptom onset is approximately 3 years, although the rate of progression varies between individuals. Rapid variants of disease may progress to death within a few months, but approximately 4 per cent of patients will survive for more than 10 years (Turner et al. 2003). The usual cause of death is progressive respiratory failure, which may be accompanied by bronchopneumonia.

The El Escorial Diagnostic criteria for amyotrophic lateral sclerosis were agreed as an international consensus to facilitate therapeutic trials and multinational research collaborations (Brooks 1994). These criteria were subsequently revised and the key features of the revised criteria are shown in Table 23.2 (Brooks et al 2000). In essence there must be a combination of upper and lower motor neurone signs, evidence of progression over at least 6 months, and other conditions that may mimic amyotrophic lateral sclerosis must be excluded by appropriate investigations. These criteria are stricter than the burden of proof for the diagnosis of amyotrophic lateral sclerosis usually applied in clinical practice and indeed some individuals die from the effects of amyotrophic lateral sclerosis without ever reaching the classification criteria required for definite diagnosis (Traynor et al. 2000b). Clinicians should be cautious about applying these criteria in clinical practice as patients may be confused and less able to come to terms with their illness by being told that they have probable or possible amyotrophic lateral sclerosis rather than a more definite diagnosis. Nevertheless, the criteria provide a structured approach to the evaluation of patients with amyotrophic lateral sclerosis and facilitate the inclusion of uniform populations of patients in clinical research studies.

The main conditions to be considered in the differential diagnosis of amyotrophic lateral sclerosis are listed in Table 23.3. Diagnostic errors are not uncommon in amyotrophic lateral sclerosis and an important aspect of follow-up care is to review the diagnosis in patients whose symptoms or disease course have atypical features. Of crucial importance is not to miss the diagnosis of potentially treatable disorders or those with a more benign prognosis. In a population-based survey from Scotland (Davenport et al. 1996) amyotrophic lateral sclerosis was misdiagnosed in 8 per cent of 552 patients included on the register. In an Irish population-based study which reported 437 patients’ referrals diagnosed initially as amyotrophic lateral sclerosis, an alternative diagnosis was found in 7 per cent (Traynor et al. 2000a). Most of the misdiagnosed cases had lower motor neurone syndromes: 22 per cent had multifocal motor neuropathy with conduction block (Sections 23.3.6 and 21.11.3) and 13 per cent had Kennedy’s disease (Section 23.5.1). Other forms of motor neuropathy should also be considered including acute motor axonal neuropathy (Section 21.10.2), a variant of Guillain–Barré syndrome (Griffin et al. 1996; Léger and Salachas 2001) and acute porphyric neuropathy (Section 21.8.6). Myasthenia gravis (Section 24.10.1), particularly with prominent bulbar involvement, may occasionally mimic features of amyotrophic lateral sclerosis, though upper motor neurone clinical signs will be absent.

Benign fasciculation. This is relatively common, affecting approximately 1 per cent of the population, and often causes great anxiety, particularly in individuals with medical knowledge. The syndrome comprises chronic muscle fasciculation, most commonly in the calf muscles, which may be accompanied by cramps and is often worsened by exercise. Fasciculation may be detected by electromyography, but is not accompanied by evidence of progressive denervation (Blexrud et al. 1993). One study describing 121 patients with a diagnosis of benign fasciculation, 33 per cent of them health workers, seen at the Mayo clinic, reported no cases developing amyotrophic lateral sclerosis during a prolonged period of follow-up (Blexrud et al. 1993). However, individuals presenting with fasciculations and cramps may very rarely go on to develop amyotrophic lateral sclerosis (de Carvalho and Swash 2004). Muscle cramps and twitching in the presence or absence of weakness may occur in peripheral nerve hyper-excitability with myokymia (Hart et al. 2002; Gutmann and Gutmann 2004; Lagueny 2005). Myokymia usually consists of rippling continuous muscle contractions, rather than the discrete random twitching typically seen with fasciculation.

Focal structural lesions. These lesions of the brainstem or spinal cord may sometimes be confused with amyotrophic lateral sclerosis, particularly multilevel spondylotic radiculomyelopathy which may produce upper motor neurone signs accompanied by lower motor neurone signs in the upper and lower limbs. Other structural lesions which may occasionally be mistaken for amyotrophic lateral sclerosis include mass lesions at the skull base; slowly growing tumours of the cervical spinal cord, cauda equina or conus medullaris, and syringomyelia. Many of these patients will have pain and sensory disturbance as well as motor dysfunction. Other more peripheral lesions such as infiltration or compression of the lower roots of the brachial plexus by a Pancoast tumour (Section 22.4.2) or a cervical rib (Section 22.5.3) will also usually be accompanied by pain and sensory features. Occasionally patients with infiltrative lesions of the base of tongue may mimic the features of bulbar onset amyotrophic lateral sclerosis and the accompanying pain is an important diagnostic clue. In the presence of these structural lesions electrophysiological and imaging studies will usually lead to the correct diagnosis.

Hereditary spastic paraplegia variants (Section 23.4.2). These may have lower motor neurone features as well as the typical slowly progressive upper motor neurone disorder primarily affecting the lower limbs. There is clinical overlap between hereditary spastic paraplegia and the primary lateral sclerosis variant of motor neurone disease (Brugman et al. 2005; Strong and Gordon 2005). Patients with the Troyer syndrome, SPG 20, and Silver syndrome, SPG 17, typically have both upper and lower motor neurone features, but a much more slowly progressive clinical course than typical amyotrophic lateral sclerosis. Adult onset forms of spinal muscular atrophy and hereditary motor neuronopathy can pose diagnostic difficulties. Individuals with apparently sporadic lower motor neurone syndromes who survive more than 4 years from the onset of symptoms may turn out to have relatively benign forms of late onset spinal muscular atrophy (van den Berg-Vos et al. 2003a,b) (Section 23.3). Clues to the diagnosis of hereditary motor neuronopathy or spinal muscular atrophy include slow progression, family history, pes cavus, and other features such as vocal cord involvement.

Inclusion body myositis (Section 24.7.3) may mimic the progressive muscular atrophy variant. This is one of the most common types of myopathy presenting in those over 50 years of age (Griggs et al. 1995; Dalakas 2006). In one series of 70 patients with inclusion body myositis, 13 per cent had originally received a diagnosis of amyotrophic lateral sclerosis (Dabby et al. 2001). Important clues to the diagnosis of inclusion body myositis include early and prominent weakness of finger flexors and quadriceps muscles, which tend only to become weak late in the course of amyotrophic lateral sclerosis. Dysphagia may occur. Re-examination of patients both clinically and by electrophysiology where there is diagnostic uncertainty is very important and muscle biopsy may sometimes be indicated. Polymyositis may sometimes be misdiagnosed as amyotrophic lateral sclerosis. For example one reported case of biopsy-proven polymyositis was reported to show clinical features including dysphagia, possible upper motor neurone signs, normal creatine kinase level, and fasciculations on electromyography, causing understandable diagnostic uncertainty (Ryan et al. 2003).

Multisystem degenerative disorders. Disorders in which features of upper and lower motor neurone damage may occur include the frontotemporal dementia syndromes; subtypes of spinocerebellar atrophy including SCA6 and SCA3 (Section 39.8) (Ohara et al. 2002; Seilhean et al. 2004), multiple system atrophy (Section 40.3.8), progressive supranuclear palsy (Section 40.3.9), and corticobasal degeneration (Section 40.3.10) (Neary et al. 1990; Strong et al. 2003; Kertesz et al. 2005; Mott et al. 2005). These multisystem disorders all have prominent features of involvement of the central nervous system outside the upper and lower motor neurones including involvement of basal ganglia, the cerebellum, and cortical structures, so that diagnostic confusion with amyotrophic lateral sclerosisis is uncommon.

Infective. Infective and post-infective disorders occasionally need to be considered in the differential diagnosis of amyotrophic lateral sclerosis. HTLV1 associated myelopathy (Sections 28.5.8; 42.4.1) typically produces a slowly progressive spastic paraparesis with early involvement of the bladder. It has been reported to occasionally be associated with amyotrophic lateral sclerosis-like features, with wasting and fasciculation of the tongue and limbs and electrophysiological evidence of widespread denervation (Matsuzaki et al. 2000; Silva et al. 2005). Patients have been reported with an amyotrophic lateral sclerosis-like disorder in the presence of HIV infection and some patients have apparently improved neurologically with highly active anti-retroviral treatment, ‘HARRT’ (Moulignier et al. 2001; Calza et al. 2004). Often the motor disorder occurring in the context of HIV is not entirely typical of amyotrophic lateral sclerosis, with a younger age of onset and a more rapidly progressive disease. The post-polio progressive muscular atrophy syndrome (Section 23.3.7) may sometimes be mistaken for amyotrophic lateral sclerosis, but its clinical course is much more slowly progressive. It may affect up to one-third of individuals who suffered acute paralytic poliomyelitis and is usually regarded as a syndrome of motor decompensation related to ageing (Ramlow et al. 1992; Ragonese et al. 2005). Lyme disease (Section 42.5.2) has been described as causing a progressive motor neurone disorder which may improve following the administration of antibiotic therapy (Hemmer et al. 1997). Lyme disease is not established as a cause of amyotrophic lateral sclerosis, but it is reasonable to consider the condition as a potential mimicking syndrome in individuals with exposure to tick bites, or when the CSF contains oligoclonal immunoglobulin bands. It is noteworthy that antibodies to Borrelia burgdorferi are commonly encountered in people living in endemic areas (Halperin et al. 1990). Shoulder girdle amyotrophy is a feature of infection by the tick borne encephalitis virus (Logina et al. 2006).

Malignancy. Most association studies have failed to demonstrate a convincing link between amyotrophic lateral sclerosis and malignant disease (Section 38.4.4) (Jokelainen 1976; Evans et al. 1990; Rosenfeld and Posner 1991; Freedman et al. 2005). However, some reports have suggested a possible association with several types of haematological malignancy and melanoma (Younger et al. 1990; Rowland et al. 1995; Gordon et al. 1997; Freedman et al. 2005). An amyotrophic lateral sclerosis-like syndrome has been reported in association with anti-Hu antibodies, and breast cancer may be linked to a primary lateral sclerosis like disorder without anti-Hu antibodies (Forsyth et al. 1997). Predominantly motor neuropathies or neuronopathies can be associated with anti-Hu or anti-Yo antibodies (Khwaja et al. 1998; Graus et al. 2001). Overall, a causative association between amyotrophic lateral sclerosis and cancer is not generally regarded as proven, although there are occasional reports of remission of motor dysfunction following therapy for malignant disease.

Metabolic disorders. These disorders occasionally mimic the features of amyotrophic lateral sclerosis. Hyperthyroidism may present with muscle weakness, wasting, and fasciculation (Rosati et al. 1980; Chotmongkol 1999). Some patients have superimposed upper motor neurone signs, a combination resulting in a clinical picture similar to that of amyotrophic lateral sclerosis (Fisher et al. 1985; Shaw et al. 1988). Treatment of the hyperthyroid state usually results in complete or near complete recovery of the upper motor neurone signs. Hyperparathyroidism may present with weakness and brisk reflexes and a causal relationship between hyperparathyroidism has been suggested (Patten and Pages 1984), but not established with certainty (Jackson et al. 1998).

The diagnosis of amyotrophic lateral sclerosis is essentially clinical and there is no specific diagnostic test. Neurophysiological evaluation and MRI of the brain and spine are the most useful investigations. Sensory nerve conduction is usually normal and motor nerve conduction velocity is also normal unless there has been severe depletion of large diameter motor axons. It is important that conduction block is carefully excluded and this may require evaluation of proximal nerve segments. Electromyography is valuable in demonstrating neurogenic changes that cannot be explained by a single nerve, root, or plexus lesion (Mills 2003). Assessment of the thoracic paraspinal muscles can be particularly valuable in differentiating amyotrophic lateral sclerosis from multilevel spondylotic radiculomyelopathy (Kuncl et al. 1988). Imaging is useful to exlude the presence of structural disorders. A variety of blood tests may be helpful in distinguishing the amyotrophic lateral sclerosis mimic syndromes outlined above. Muscle biopsy is only indicated in rare or atypical cases where diagnostic uncertainty persists in the light of the initial investigation results.

Motor neurone disease has been considered as traditionally a pure motor disorder, but the selectivity of the disease process for the motor system is now recognized to be relative rather than absolute. Careful clinical and pathological studies have revealed involvement in extra-motor parts of the central nervous system such as changes in other long tracts, including sensory and spinocerebellar pathways, and in neuronal groups such as substantia nigra neurones and dentate granule cells in the hippocampus. The description of ubiquitinated intraneuronal inclusions in 1988 highlighted a common molecular pathology in motor neurone disease (Leigh et al. 1988; Lowe et al. 1989) and has provided evidence of widespread involvement of extramotor regions of the central nervous system. Furthermore, the ubiquitinated inclusion is described in a number of related disorders, including primary lateral sclerosis, progressive muscular atrophy, and frontotemporal dementia, supporting the hypothesis that these conditions represent a clinico-pathological spectrum of the same disease process. Thus, motor neurone disease is now regarded as a multi-system disease in which the motor neurones tend to be affected earliest and most severely (Ince et al. 1998a).

Gross pathological changes. The gross pathological changes of amyotrophic lateral sclerosis consist of atrophy of the cerebral precentral gyrus, and shrinkage, sclerosis, and pallor of the lateral and anterior corticospinal tracts tracts of the spinal cord. Thinning of the hypoglossal nerves and anterior spinal roots may be observed, and there is atrophy of the somatic musculature.

Lower motor neurone pathology. Motor neurone disease patients will typically have lost 50 per cent of the lower motor neurones in the limb enlargement areas of the spinal cord at autopsy (Ince 2000). Many of the remaining lower motor neurones show atrophic and basophilic changes that are likely to represent part of the spectrum of an apoptosis, programmed cell death pathway (Martin 1999). The depletion of lower motor neurones is accompanied by diffuse astrocytic gliosis in the spinal grey matter. There is relative preservation of motor neurones in the nucleus of Onufrowitz, Onuf’s nucleus, in the sacral spinal cord (Mannen et al. 1977), which innervates skeletal muscles of the pelvic floor, and in the cranial motor nuclei of the oculomotor, trochlear, and abducens nerves which control eye movements. The selective resistance of these two groups of motor neurones in motor neurone disease is unexplained.

A cardinal feature of lower motor neurone pathology in motor neurone disease is the presence of inclusion bodies within the soma and proximal dendrites. Ubiquitinated inclusions are the most frequent neuronal lesion, and are found in virtually 100 per cent of cases (Ince et al. 2003; Piao et al. 2003). These inclusions show a range of appearances including thread-like profiles; skeins (Fig. 23.3A) of varying compactness and more compact spherical bodies (Fig. 23.3C) (Ince et al. 1998a). Ubiquitin is a small highly conserved protein which becomes covalently bound to intracellular proteins targeted for disposal. The resulting tagged protein is degraded in the proteolytic channel of the 26S proteasome. Ubiquitin immunoreactivity is a prominent feature of the neuropathology of many neurodegenerative diseases, where it is found as a component of various inclusion bodies. Whereas in several other neurodegenerative diseases, the protein to which ubiquitin binds is known, for instance tau or α-synuclein ubiquitin-positive inclusions, in motor neurone disease the protein substrate for ubiquitination is not currently known with certainty, though TD43 represents an important candidate (Neumann et al. 2006).

Bunina bodies, first described in 1962 are eosinophilic inclusions present in the soma of lower motor neurones (Bunina 1962). They are found in 86 per cent of amyotrophic lateral sclerosis cases (Piao et al. 2003), and are specific to the disease. They have been shown to be immunoreactive to the proteinase inhibitor cystatin C, but their nature, origin, and significance remain unclear (Okamoto et al. 1993).

Neurofilament conglomerate inclusions are large inclusion bodies detectable with silver stains and with antibodies directed at both the phosphorylated and non-phosphorylated forms of heavy- and medium-chain neurofilament proteins (Ince et al. 1998b). Motor neurones containing these prominent inclusion bodies lose the normal neurofilament cytoskeleton from the remainder of the cell body. Neurofilament conglomerate inclusions are seen only infrequently in sporadic cases of motor neurone disease and are most commonly described in familial motor neurone disease caused by specific mutations in the SOD1 gene (Fig. 23.3B) (Rouleau et al. 1996; Ince et al. 1998). In normal motor neurones, neurofilament proteins within the cell body are predominantly non-phosphorylated, and there is progressive phosphorylation of neurofilaments in the axonal compartment. Several reports have documented a diffuse increase in neurofilament phosphorylation within the soma of spinal motor neurones in motor neurone disease (Munoz et al. 1988; Sobue et al. 1990). Swellings termed spheroids have been described in the axons of lower motor neurones in motor neurone disease. These are composed of abnormally orientated accumulations of neurofilaments, and are presumed to represent focal abnormalities of axonal cytoskeletal regulation, with failure of axonal transport (Ince 2000). Axonal spheroids are not however, disease specific, and are present in normal subjects, although the numbers of spheroids may be increased in motor neurone disease. Spheroids have been described in the cell body of spinal motor neurones (Hirano et al. 1967), and by immunocytochmistry appear similar to neurofilament conglomerate inclusions suggesting that the same dysregulation of cytoskeletal function may underlie both types of cellular lesion.

Upper motor neurone pathology. A major feature of the pathology of motor neurone disease is axonal loss within the descending pyramidal motor pathway, associated with secondary myelin pallor and gliosis of the corticospinal tracts. This pathology gives rise to the ‘lateral sclerosis’ within the spinal cord originally described by Charcot. Myelin pallor of the corticospinal tract is most prominent in the cervical cord and medullary pyramids, and in many cases is not demonstrable above the level of the medulla. This finding suggests that upper motor neurone changes in motor neurone disease arise from a dying back axonopathy, with distal denervation due to axonal loss preceding degeneration of the cell bodies of the upper motor neurones and loss of pyramidal cells from the cortex (Ince 2000).

In the motor cortex, pathological changes are highly variable, even in patients with well-established signs of upper motor neurone involvement clinically. Severely affected cases will usually show a reduction in the population of giant pyramidal neurones, Betz cells, in the motor cortex, either due to loss of these neurones, or a reduction in their size so that they are indistinguishable from smaller neighbouring pyramidal cells. Intracellular inclusions, including ubiquitinated inclusions, have not been convincingly described in Betz cells. The motor cortex may show astrocytic gliosis of varying severity in the grey matter and underlying subcortical white matter. Evidence of microglial activation, can be detected with immunocyotchemical markers in the corticospinal tract within the brain and spinal cord (Troost et al. 1990; Ince et al. 2003).

Sensory and cerebellar pathways. Despite the paucity of sensory signs clinically, the ascending sensory pathways of the dorsal column are commonly affected in amyotrophic lateral sclerosis. In the Japanese literature, the prevalence of dorsal column pallor in familial disease is emphasized, with the concept of ‘familial amyotrophic lateral sclerosis with posterior column involvement’. However, more recent work suggests that this change is detectable at autopsy in up to 50 per cent of all sporadic amyotrophic lateral sclerosis cases (Ince et al. 2003). There is also evidence of degenerative changes in peripheral sensory nerves, and loss of large afferent nerve fibres (Kawamura et al. 1981). Cerebellar involvement in amyotrophic lateral sclerosis is not usually recognized clinically, although late involvement of these pathways could go undetected in the presence of severe weakness. Pathologically, degeneration of the spinocerebellar pathway, reflected by cell loss from the thoracic nucleus of Clarke, and by pallor of ascending spinocerebellar pathways, is a relatively frequent finding in sporadic motor neurone disease (Brownell et al. 1970).

Muscle pathology. Loss of lower motor neurones result in denervation and atrophy of skeletal muscle. Histologically denervated muscle shows clusters of angular atrophic fibres and fibre-type grouping which results from serial denervation and reinnervation arising from collateral sprouting of axons of surviving motor neurones within muscle (Ince 2000).

Primary lateral sclerosis and progressive muscular atrophy. The pathological features of several patients with primary lateral sclerosis have been described (Pringle et al. 1992; Watanabe et al. 1997). The most prominent features described are atrophy of the prefrontal gyrus, with loss of Betz cell somata, and pallor and atrophy of the lateral corticospinal tracts. In cases where ubiquitin immunochemistry was undertaken, classical ubiquitinated inclusions were seen in a small number of lower motor neurones, and gliosis of the spinal ventral horns has also been described (Pringle et al. 1992; Watanabe et al. 1997). Ubiquitinated inclusion bodies in the cerebral cortex in a distribution typically found in amyotrophic lateral sclerosis-dementia cases has also been described (Kawashima et al. 1998), suggesting that primary lateral sclerosis may merge into the spectrum of amyotrophic lateral sclerosis dementia.

A pathological study of 14 patients with the clinical phenotype of progressive muscular atrophy demonstrated typical ubiquitinated inclusions in the spinal cord or bulbar motor neurone groups in most cases (Ince et al. 2003). Interestingly, 50 per cent of this series were also shown to have evidence of corticospinal tract involvement demonstrated by immunostaining for active microglia and macrophages.

The pathological findings of occasional abnormalities of the lower motor neurones in primary lateral sclerosis, and of corticospinal tract involvement in progressive muscular atrophy support the clinical evidence of overlap between these conditions and amyotrophic lateral sclerosis (Rowland 1999). Ubiquitinated inclusions are the characteristic pathological feature of amyotrophic lateral sclerosis, primary lateral sclerosis, and progressive muscular atrophy suggesting that these disorders share common pathophysiological mechanisms (Mackenzie and Feldman 2005). Primary lateral sclerosis and progressive muscular atrophy should be considered part of the spectrum of amyotrophic lateral sclerosis.

Pathological features of prolonged disease. Patients whose survival is prolonged by invasive ventilation may develop clinical and pathological features of involvement of less vulnerable motor neurone groups and eventually widespread involvement of the central nervous system (Hayashi and Kato 1989; Mizutami et al. 1990; Sasaki et al. 1992). These patients may lose voluntary control of the ocular muscles, and become totally ‘locked-in’. They also lose control of external sphincters, and develop decubitus ulcers, features not typically seen in motor neurone disease. Electroencephalograms performed at later stages of disease show diffuse slowing (Hayashi et al. 1989) and MR imaging demonstrates progressive cerebral atrophy, including the frontal and temporal lobes, precentral gyrus, postcentral gyrus, anterior cingulate gyrus, and corpus callosum (Kato et al. 1993).

Pathology of cognitive impairment. Patients with motor neurone disease and dementia have both the characteristic motor system ubiquitinated inclusions, together with cerebral pathology, which consists of small globular ubiquitinated inclusions within the dentate granule cells (Fig. 23.4), and a variable component of neocortical ubiquitinated neurites and small neuronal ubiquitinated inclusions (Munoz et al. 2003; Mackenzie and Feldman 2003). Other limbic structures, including the amygdala and parahippocampal gyrus may also show ubiquitinated inclusion pathology. Autopsy series indicate that 20–50 per cent of non-demented motor neurone disease patients have similar cerebral pathology, although the severity and distribution tends to be less extensive (Wilson et al. 2001; Mackenzie et al. 2003; Mackenzie and Feldman 2005). The degree of overlap in pathological findings, neuropsychological deficits, and imaging studies in classical motor neurone disease without dementia, motor neurone disease inclusion dementia, and amyotrophic lateral sclerosis-fronto-temporal dementia suggest that they represent a spectrum of clinical disease with a common pathological substrate.

The primary pathogenetic processes underlying amyotrophic lateral sclerosis are multifactorial and the precise mechanisms underlying selective cell death in the disease are at present incompletely understood. Current understanding of the neurodegenerative process in amyotrophic lateral sclerosis suggests that there may be a complex interplay between multiple mechanisms including genetic factors, oxidative stress, excitotoxicity, and protein aggregation as well as damage to critical cellular processes, including axonal transport and organelles such as mitochondria (Bruijn and Cleveland 1996; Bruijn et al. 2004; Shaw 2005). Recently there has been growing interest in the role played in motor neurone injury by neighbouring non-neuronal glial cells and in dysfunction of particular molecular signalling pathways. The relative importance of these different pathways may well vary in different subgroups of patients, and a very important task for the future is to further define the subgroups of amyotrophic lateral sclerosis. Evidence has also accumulated that the final process of motor neurone death is likely to occur via a caspase-dependent programmed cell death pathway resembling apoptosis.

Patients with amyotrophic lateral sclerosis report a family history of the disease in about 5–10 per cent of cases. The most common pattern of inheritance is autosomal dominant, with complete penetrance, although recessive or X-linked inheritance occurs in some pedigrees. Linkage studies in amyotrophic lateral sclerosis are rendered difficult by age-dependent onset in adults, short disease duration, heterogeneity of presentation, and misdiagnosis. Linkage to 12 different chromosomal loci has been established in familial motor neurone disease, and for 6 of these, the underlying genetic defect has been identified (Table 23.4).

ALS 1: Copper–zinc superoxide dismutase, SOD1. Familial motor neurone disease with adult onset is clinically indistinguishable from sporadic amyotrophic lateral sclerosis in individual cases. Twenty per cent of families with autosomal dominant motor neurone disease show mutations in the gene on chromosome 21q22.1 which encodes the free radical scavenging enzyme superoxide dismutase, referred to as SOD1 (Rosen et al. 1993). More than 100 mutations have now been identified, the majority of which are missense mutations. SOD1 is a ubiquitously expressed metalloenzyme whose major function is to convert intracellular superoxide free radicals to hydrogen peroxide. SOD1 is an abundant protein in the central nervous system, accounting for about 1 per cent of brain protein. SOD1 was initially thought to be confined to the cytosolic compartment of cells but it is now recognized that a small proportion of the protein is located in the intermembrane space of mitochondria (Okado-Matsumoto and Fridovich 2001). The reasons why motor neurones are especially vulnerable to injury in the presence of SOD1 mutations, are not yet clear. Despite 13 years of intensive research effort, the pathways leading to the cell death of motor neurones in the presence of SOD1 mutations have not been fully identified, although there is a convincing body of evidence that the mutant SOD1 protein exerts its detrimental effects through a toxic gain of function rather than a loss of function. Most of our current level of understanding of disease mechanisms in amyotrophic lateral sclerosis has come from the study of the effects of SOD1 mutations but, even in this defined genetic subgroups of disease, the pathways to neurodegeneration appear to be complex and multifactorial.

There is considerable variation in disease phenotype in terms of age of onset and rate of disease progression in human SOD1 related motor neurone disease. It is apparent that the clinical phenotype must be modified by other genetic and/or environmental factors. There has been much interest in the D90A SOD1 mutation, which has a dominant inheritance in some genetic backgrounds, but is recessively inherited with two mutated copies of the gene required to cause disease, in Scandinavian populations, implying a co-inherited protective factor (Andersen et al. 1996).

Mutant SOD1 transgenic mice have been genetically engineered and develop a disease which clinically and pathologically resembles human motor neurone disease. The most extensively studied are SOD1 G93A, SOD1 G37R, and SOD1 G85R (Bruijn et al. 1996). Transgenic rats, carrying G93A or H46R SOD1 also develop a motor neurone disease phenotype (Nagai et al. 2001). In addition cellular models of SOD1 related motor neurone disease have been generated which have helped to elucidate cellular mechanisms of disease (Pasinelli et al. 1998; Cookson et al. 2002). The toxic gain of function of mutant SOD1 has not yet been fully defined, but there are several pathophysiological processes which may be involved, including oxidative stress, mitochondrial dysfunction, excitotoxicity, protein aggregation, and inflammation. These mechanisms are not mutually exclusive and it is possible that all of these factors play a role in the development of motor neurone injury.

ALS 2: Alsin. In 2001 alsin was identified as the causative gene for an autosomal recessive form of juvenile amyotrophic lateral sclerosis linked to chromosome 2q33 (Hadano et al. 2001; Yang et al. 2001). Mutations in alsin can also cause a motor neurone degenerative disorder with a predominant upper motor neurone phenotype: juvenile recessive primary lateral sclerosis; infantile onset ascending hereditary spastic paralysis (Eymard-Pierre et al. 2002); or autosomal recessive complicated hereditary spastic paraplegia (Gros-Louis et al. 2003).

ALS 2 encodes a 184 KDa protein which contains three putative guanine nucleotide exchange factor domains and is alternatively spliced to generate a short and a long transcript. These factors are known to activate small GTPase proteins by stimulating the release of GDP in exchange for GTP. Given the conserved guanine nucleotide exchange factor domains of ALS2, it is predicted to function as an activator of particular small GTPases. The small GTPases control a range of important cellular processes and function as binary switches—alternating between inactive GDP-bound and active GTP-bound states. The alsin protein is widely expressed, but enriched within the central nervous system, where it is localized to the cytoplasmic face of endosomal membranes (Yamanaka et al. 2003). The functions of alsin are still being investigated, but to date it has been shown to function as an activator of the small GTPase protein Rab5 (Otomo et al. 2003; Kunita et al. 2004). This implies that alsin is important in endosomal dynamics and the working hypothesis is that alsin normally regulates trafficking of signalling molecules important for proper development and/or maintenance of health of motor neurones. ALS2 knockout mice have been generated but the motor system phenotype so far arising appears to be very mild (Kris et al. 2003).

ALS 4: Senataxin The ALS4 locus linked to chromosome 9q34 was originally identified in a single large pedigree with juvenile onset, autosomal dominant amyotrophic lateral sclerosis. The disease course in this family was indolent and did not reduce life expectancy. Three different missense mutations, L3095, R2136H, and T3I, in three families with this subtype of motor neurone disease were identified (Chen et al. 2004). The SETX gene encodes senataxin, a large 302.8KD protein of unknown function. Much of the protein has no homology with other known proteins but there is one DNA/RNA helicase domain. DNA/RNA helicase proteins are known to have roles in processes such as repair, replication, recombination or transcription of DNA and RNA processing, RNA transcript stability, and the initiation of translation. Recessive loss of function mutations in SETX are associated with ataxia-oculomotor apraxia type 2 (Moreira et al. 2004). It is predicted that the different phenotype of dominantly inherited ALS4 is likely to be caused by a toxic gain of function of the mutated senataxin protein.

ALS 8: VAPB. Nishimura and co-workers described a novel missense mutation, P565, in the VAPB gene for vesicle-associated membrane protein/synaptobrevin associated membrane protein, located on chromosome 20q13.3, in a Brazilian family with ALS8, an autosomal dominant slowly progressive disorder characterized by fasciculation, cramps, and postural tremor (Nishimura et al. 2004). They subsequently found the same mutation in six further families with different clinical phenotypes, including late onset spinal muscular atrophy and classical rapidly progressive amyotrophic lateral sclerosis. Vesicle associated proteins are intracellular membrane proteins that can associate with microtubules and have been shown to function in membrane transport. The VAPB protein has three identifiable structural domains. The first 150 residues form an MSP domain conserved between all members of this protein family; the central region contains an amphipathic helical structure predicted to form a coiled/coil protein–protein interaction motif and at the carboxy terminus is a hydrophobic region that acts as a membrane anchor. Preliminary cell biological studies have indicated that the wild-type VAPB protein localizes predominantly to the endoplasmic reticulum. The P56S mutation disrupts the subcellular distribution and induces the formation of intracellular protein aggregates (Nishimura et al. 2004).

Dynactin mutation. A mutation, G595, substitution in the gene encoding the P150 subunit of dynactin, DCTN1, has been identified in a single family with a slowly progressive lower motor neurone degenerative disorder (Puls et al. 2003). The described family had an unusual phenotype, presenting in early adulthood with respiratory difficulties due to vocal cord paralysis, progressive facial weakness, and weakness and atrophy of the hands and later development of neurogenic changes distally in the lower limbs. Mutations of the p150 subunit of DCTN1 were subsequently identified in patients with familial amyotrophic lateral sclerosis and frontotemporal dementia, and one case of apparently sporadic amyotrophic lateral sclerosis (Munch et al. 2004; Munch et al. 2005).

The dynactin–protein complex is required for dynein mediated retrograde axonal transport of vesicles and organelles along the microtubule system. The amino acid change caused by the mutation is predicted to distort the folding of the microtubule binding domain of dynactin. Overexpression of the P50 subunit of dynactin has been shown to disrupt the function of this protein complex and causes late onset progressive motor neurone degeneration in genetically engineered mice (La Monte et al. 2002).

Other amyotrophic lateral sclerosis loci. The genes for several other subtypes of amyotrophic lateral sclerosis remain to be identified. Three separate families have shown linkage to chromosome 16. Amyotrophic lateral sclerosis with fronto-temporal dementia has been mapped to a 17-cM interval chromosome 9q21 (Hosler et al. 2000; Ostojic et al. 2003), and one Swedish family with a similar phenotype without linkage to the chromosome 9 locus has recently been identified, suggesting genetic heterogeneity for this subtype of disease. Motor neurone degeneration may occur in patients with fronto-temporal dementia and Parkinson’s disease associated with mutations in the microtubule associated protein tau MAPT (Clark et al. 1998). The mutant tau protein forms filamentous inclusions and insoluble aggregates that are associated with neurodegeneration. Some patients with familial fronto- temporal dementia, Parkinsonism, and amyotrophic lateral sclerosis do not have identified mutations in tau suggesting that further genes causing this triad of features remain to be identified (Kowalska et al. 2003).

Possible genetic risk factors in sporadic amyotrophic lateral sclerosis. There have been reports of genetic variants found in individuals with apparently sporadic disease (summarized in Table 23.4) (Kunst 2004; Shaw 2005).

Cellular injury by free radical species is a major potential cause of the age-related deterioration in neuronal function which occurs in neurodegenerative diseases. There has been particular interest in the role of oxidative stress in amyotrophic lateral sclerosis given that SOD1 mutations, which encode a key cellular anti-oxidant defence protein, underlie approximately 20 per cent of familial cases. Studies of CSF and human post-mortem central nervous system tissue have demonstrated the presence of biochemical changes which represent the effects of oxidative stress and these changes are more pronounced in amyotrophic lateral sclerosis cases compared to controls (Shaw et al. 1995b; Ferrante et al. 1997; Smith et al. 1998; Tohgi et al. 1999). Fibroblasts cultured from the skin of patients with both familial and sporadic amyotrophic lateral sclerosis show increased sensitivity to oxidative insults compared to those from control cases (Aguirre et al. 1998).

In relation to the toxic gain of function of the mutant superoxide dismutase 1 protein, oxidative damage and/or metal mishandling have been strongly implicated. The main hypotheses have been that mutations alter the structure of this protein, allowing greater access of abnormal substrates to the active copper site of the dimeric enzyme, resulting in the production of damaging free radical species including peroxynitrite and hydroxyl radicals. Nitration of tyrosine residues on cellular proteins by peroxynitrite can have damaging consequences (Beckman et al. 1993). Some SOD1 mutations render the protein more likely to form a zinc deficient variant (Crow et al. 1997), which in turn makes the copper site more accessible to abnormal substrates. In vitro studies have demonstrated that zinc deficient superoxide dismutase 1 causes peroxynitrite dependent cell death (Estevez et al. 1999). However, a body of experimental work has raised questions as to whether the toxicity of mutant superoxide dismutase 1 can be explained by copper- dependent oxidative mechanisms. For example, superoxide dismutase 1 that has been manipulated not to bind copper by mutating the four histidine residues for copper binding still causes motor neurone disease in transgenic mice. Also, knock out of the gene encoding the copper chaperone protein normally required for insertion of copper into the enzyme, has no effect on the disease phenotype in SOD1 transgenic mice (Subramaniam et al. 2002). It seems plausible that mutant superoxide dismutase 1 may induce oxidative stress by a mechanism beyond its own catalytic activity and transcriptional repression of anti-oxidant response genes under the control of the transcription factor NRF2 is one potential pathway of interest (Kirby et al. 2005).

Glutamate is the major excitatory transmitter in the human central nervous system and there is great complexity in the molecular structure of the repertoire of receptors for this neurotransmitter system. Excitotoxicity is the term coined for neuronal injury induced by excessive stimulation of glutamate receptors, by mechanisms which include derangement of intracellular calcium homeostasis, and excessive free radical production. Motor neurones appear particularly susceptible to toxicity via activation of cell surface AMPA receptors (Carriedo et al. 1996). Glutamatergic toxicity has been implicated as a contributory factor to motor neurone injury in amyotrophic lateral sclerosis (Heath and Shaw 2005). The key findings are that the expression and function of the major glial glutamate re-uptake transporter protein EAAT₂ may be impaired in the central nervous system of motor neurone disease patients and that CSF, and therefore central nervous system extracellular fluid, levels of glutamate appear to be abnormally elevated at least in a proportion of motor neurone disease patients (Rothstein et al. 1995; Shaw et al. 1995a; Fray et al. 1998; Spreux-Varoquaux et al. 2002). Excitotoxicity has provided a potential mechanistic link between SOD1 mutant mediated motor neurone disease and the sporadic form of the disease. The presence of mutant superoxide dismutase 1 increases the sensitivity of motor neurones to glutamate toxicity (Kruman et al. 1999); causes alteration in AMPA receptor subunit expression (Spalloni et al. 2004); as well as reduced expression of the major glutamate re-uptake transporter EAAT₂ (Bendotti et al. 2001). Whether as a primary or a propagating process, it appears that glutamate toxicity plays a contributory role to the injury of motor neurones in amyotrophic lateral sclerosis. This is supported by the finding that anti-glutamate therapy with riluzole has some effect, albeit modest, in prolonging survival in human amyotrophic lateral sclerosis patients, and in SOD1 mutant mouse models (Lacomblez et al. 1996; Gurney et al. 1996).

Mitochondria serve multiple important intracellular functions including generation of intracellular ATP, buffering of intracellular calcium, generation of intracellular free radicals, and involvement in the initiation of apoptotic cell death. Age-related deterioration in mitochondrial function may contribute to the development of late-onset neurodegenerative diseases. There is a body of evidence emerging from investigation of human material and cellular and animal models indicating that mitochondrial dysfunction may contribute to motor neurone injury in amyotrophic lateral sclerosis and this has been reviewed (Beal 2000; Menzies et al. 2002b).

The key evidence for mitochondrial dysfunction in human amyotrophic lateral sclerosis includes:

In cellular models of superoxide dismutase 1 related motor neurone disease, expression of a G93A SOD1 mutant results in the development of abnormally swollen mitochondria which display impaired activity of complexes II and IV of the mitochondrial respiratory chain, impaired cellular bioenergetic status, and alteration in the mitochondrial proteome (Menzies et al. 2002a; Wood-Allum et al. 2006). Molecular targeting of mutant superoxide dismutase 1 to the mitochondria but not to the nucleus or endoplasmic reticulum leads to activation of the apoptosis cascade and cell death (Takeuchi et al. 2002). In some strains of mutant SOD1 transgenic mice mitochondrial vacuolation within motor neurones is an early feature of the pathology (Wong et al. 1995). Whereas superoxide dismutase 1 was previously considered to be an exclusively cytosolic protein, it is now recognized also to reside in the intermembrane space of mitochondria. Superoxide dismutase 1 has been shown to accumulate in vacuolated mitochondria in mutant SOD1 mice. It has been demonstrated that the activities of several complexes of the mitochondrial respiratory chain are reduced prior to disease onset and that these changes increase with age. Oxidative damage to mitochondrial protein and lipids and decreased ATP synthesis have been reported at the onset of the murine disease. Translocation of cytochrome C, an initiator of apoptosis, from the mitochondria to the cytosol has been demonstrated during disease progression in the mice. Recent studies have reported that mutant superoxide dismutase 1 is selectively and aberrantly recruited to the cytoplasmic face of mitochondria in spinal cord tissue from mutant SOD1 transgenic mice and that the anti-apoptotic protein Bcl2 may be entrapped within large protein aggregates of superoxide dismutase 1 within spinal cord tissue, which may result in reduced availability of this protein to regulate apoptosis (Liu et al. 2004; Pasinelli et al. 2004).

Therapeutic effects of compounds which modulate mitochondrial function have begun to be investigated in SOD1 transgenic mouse models. Creatine buffers energy levels within the cell, maintains ATP levels, and stabilizes mitochondrial creatine kinase which inhibits opening of the mitochondrial permeability transition pore. Administration of creatine to G93A transgenic mice improved motor function and extended survival in a dose-dependent manner, as well as causing a decrease in biochemical indices of oxidative damage in the spinal cord (Klivenyi et al. 1999). Minocycline, a tetracycline derivative which inhibits microglial activation and blocks release of cytochrome c from mitochondria, also slows disease in mutant SOD1 mice (Zhu et al. 2002).

Motor neurones, which in the human nervous system may have axons up to 1 m in length, are highly reliant on an efficient intracellular transport system with anterograde and retrograde components. It is interesting that in SOD1 mutant mice, axonal transport is demonstrably impaired several months before clinical disease onset (Wiliamson et al. 1999). The kinesin complex of proteins are important molecular motors for anterograde axonal transport on the microtubule system. Mutations of genes encoding several kinesin proteins have been shown to cause several types of motor neurone degeneration including a hereditary spastic paraplegia, SPG10, and type 2A Charcot–Marie–Tooth disease, though have not yet been associated with amyotrophic lateral sclerosis. The dynein–dynactin complex is the important motor for retrograde transport on the microtubule system, returning components such as multivesicular bodies and neurotrophic factors back to the perikaryon. Mutations in dynein and the dynactin complex which is an activator of cytoplasmic dynein, cause progressive motor neurone disease in mice (Lamonte et al. 2002; Hafezparast et al. 2003). Mutations in the P150 subunit of dynactin may cause a form of motor neurone disease in human subjects (Puls et al. 2003).

Neurofilament proteins form a major component of the cytoskeleton of neurones and important functions include maintenance of cell shape and axonal calibre, as well as axonal transport. Neurofilament subunits are assembled in the motor neurone cell body, and transported down the axon by slow axonal transport. Accumulation and abnormal assembly of neurofilaments are common pathological hallmarks of amyotrophic lateral sclerosis. Ubiquitinated inclusions with compact or Lewy body like morphology within surviving motor neurones in amyotrophic lateral sclerosis may show immunoreactivity for neurofilament epitopes. In some cases of SOD1 mutation related amyotrophic lateral sclerosis, large argyrophilic neurofilament conglomerate inclusions have been observed in the cell bodies and axons of motor neurones (Ince et al. 1998). Approximately 1 per cent of sporadic amyotrophic lateral sclerosis cases have deletions of insertions in the KSP repeat region of the neurofilament heavy chain gene, NFH (Figlewicz et al. 1994; Tomkins et al. 1998). Pathological changes within motor neurones develop in mice overexpressing NF-light or -heavy subunits, or in mice expressing mutations in the NFL gene. Transgenic mice which carry SOD1 mutations, also show alterations in neurofilament organization, with the development of neurofilament spheroids, as well as reduced neurofilament protein and decreased transport rate in the ventral root axons. Genetic manipulations to alter the expression of neurofilament proteins have been shown to alter the disease course in SOD1 transgenic mice. Increased expression of NFH, resulting in trapping most neurofilaments within the cell body, robustly improves the disease course, by as much as 6 months in mutant SOD1 mice (Couillard-Despres et al. 1998).

Misfolding of proteins with the formation of intracellular aggregates is a key feature of multiple neurodegenerative diseases. There is continuing debate as to whether such aggregated proteins play an important role in disease pathogenesis, whether they represent harmless by-standers, or whether they could be beneficial to the cell by sequestration of toxic abnormal proteins. In the SOD1 transgenic mouse model of familial amyotrophic lateral sclerosis, the mutant superoxide dismutase 1 protein forms conspicuous cytoplasmic inclusions in motor neurones and sometimes in astrocytes, which develop before the onset of motor dysfunction. Several hypotheses have been put forward to explain how mutant superoxide dismutase 1 aggregates could generate cellular toxicity:

Overexpression of chaperone proteins has been shown to reduce mutant superoxide dismutase 1 aggregation and enhances the survival and function of motor neurones in culture (Takeuchi et al. 2002). In addition, arimoclomol, a drug which up-regulates the expression of heat shock proteins increases the life span of G93A SOD1 mice by 22 per cent (Kieran et al. 2004). Clearly protein aggregates, which can be identified by ubiquitin immunostaining, are a feature of sporadic as well as familial motor neurone disease. Superoxide dismutase 1 containing aggregates are not a characteristic feature of sporadic motor neurone disease and determining the nature of the protein inclusions in the sporadic disease is a very important research goal. A recent report has identified TDP-43, the TAR-DNA-binding protein of 43kDa, as a component of ubiquitinated inclusions in ALS (Neumann et al. 2006). TDP-43 is known to be a nuclear factor which plays a role in the regulation of transcription and alternative splicing. Further investigation is required to determine the role played by this protein in the pathogenesis of amyotrophic lateral sclerosis.

There has been much recent interest in the possibility that non-neuronal cells, including activated microglia and astrocytes, may contribute to the pathogenesis and/or propagation of the disease process in amyotrophic lateral sclerosis. Several studies in genetically engineered mouse models have suggested that expression of mutant superoxide dismutase 1 in neurones alone is insufficient to cause motor neurone degeneration and that involvement of non-neuronal cells in the vicinity of motor neurones may be required. Chimeric mice have been produced which have both normal and mutant SOD1 expressing cells (Clement et al. 2003). Motor neurones expressing mutant superoxide dismutase 1 can escape disease if surrounded by a sufficient number of normal non-neuronal cells. Conversely normal motor neurones surrounded by mutant superoxide dismutase 1 containing non-neuronal cells, developed signs of cellular injury, with the development of ubiquitinated protein deposits. Thus, the mutant enzyme may cause neurotoxicity indirectly by perturbing the function of non-neuronal cells such as microglia (Boillee et al. 2006). Microglia play a critical role as resident immunocompetent and phagocytic cells within the central nervous system. Activation is associated with transformation to phagocytic cells capable of releasing potentially cytotoxic molecules including reactive oxygen species, nitric oxide, proteases, and pro-inflammatory cytokines (Gonzales-Scarano and Baltuch 1999). Given this, there is little doubt that activated microglia can inflict significant damage on neurones but their role is complex and they are capable of stimulating neuroprotective as well as neurotoxic effects. Proliferation of activated microglia is a prominent histological feature in the spinal ventral horn both in mutant SOD1 trangenic mice and in human amyotrophic lateral sclerosis. In the mice, microglial activation is present before the onset of significant motor neurone loss or clinical signs of disease. Various inflammatory cytokines or enzymes are up- regulated in the spinal cord or CSF of amyotrophic lateral sclerosis patients: 1L-6, 1L-1β, cyclo-oxygenase 2, and prostaglandin E2 or, in the spinal cord of mutant SOD1 mice: 1L-1β, TNF-α, cyclo-oxygenase 2, and prostaglandin E2 (Almer et al. 2002; Hensley et al. 2002; Tikka et al. 2002). Microglia appear to mediate the toxicity to neurones in culture of CSF from patients with motor neurone disease by releasing factors which enhance glutamate toxicity. Minocycline, which inhibits microglial activation ameliorates disease progression in mutant SOD1 mice (Zhu et al. 2002).

There is a tendency in amyotrophic lateral sclerosis for the disease to start focally and to spread to contiguous groups of motor neurones (Brooks et al. 1995). It would be very relevant to identify molecules that contribute to this propagation and clearly those released from activated microglia would be plausible candidates.

Apoptosis describes the controlled removal of cells by an energy-dependent cell death programme. Key molecules which regulate apoptosis include: the caspase family of proteolytic enzymes which, when activated by cleavage, orchestrate cell destruction by digesting several intracellular targets including structural and regulatory proteins; the Bcl2 family of oncoproteins where the balance and subcellular distribution between pro- and anti-apoptotic members is crucial in regulating cell survival or destruction; and the apoptosis inhibitor family of proteins which suppress apoptosis by preventing proteolytic activation of specific caspases. Several pathways triggering caspase activation have been identified including: release of pro-apoptotic factors such as cytochrome c from mitochondria; activation of cell surface ligand receptor systems of the tumour necrosis factor family including Fas-Fas ligand; and stress to the endoplasmic reticulum with activation of caspase 12.

There is evidence that motor neurones may die in amyotrophic lateral sclerosis according to a programmed cell death pathway resembling apoptosis (Guegan and Przedborski 2003; Sathasivam and Shaw 2005). Key evidence from human post-mortem studies includes:

Investigation of cellular models of SOD1 mutation related amyotrophic lateral sclerosis has shown that motor neuronal cells expressing mutant SOD1 are more likely to die by apoptosis when oxidatively stressed (Cookson et al. 2002). In addition, under unstressed basal culture conditions, these mutant superoxide dismutase 1 containing cells appear to be ‘primed’ for cell death by expressing early molecular markers of apoptosis (Sathasivam et al. 2005). In the mutant SOD1 transgenic mouse model, there is evidence of DNA laddering, increased expression, and activation of caspase 1 and caspase 3 in the spinal cord of symptomatic mice, and alterations in the balance of key members of the Bcl2 protein family in a direction favouring apoptosis (Li et al. 2000). Cross breeding experiments between G93A SOD1 transgenic mice and mice genetically engineered to over-express anti-apopototic molecules results in amelioration of the murine disease. The administration of caspase inhibitors has a partial neuroprotective effect in cellular models (Sathasivam et al. 2005) and intraventricular administration of a broad spectrum caspase inhibitor to mutant SOD1 mice prolongs life span by approximately 20 per cent (Li et al. 2000).

One of the unsolved mysteries in neurodegenerative diseases, including amyotrophic lateral sclerosis, is the selective vulnerability of particular neuronal groups to the neurodegenerative process. Superoxide dismutase 1 is a ubiquitously distributed anti-oxidant defence protein, yet when the protein is mutated, it is the motor neurones which are most susceptible to injury. Certain cell specific features of motor neurones may predispose to age-related degeneration (Shaw and Eggett 2000; Durham et al. 2003). Key features are likely to include the cell size of motor neurones which has downstream consequences for intracellular transport, energy metabolism, and neurofilament content. The neurones vulnerable to injury in motor neurone disease have particular sensitivity to glutamatergic toxicity via AMPA receptor activation and differ from most other neuronal groups in expressing a high preponderance of calcium permeable AMPA receptors, lacking the GluR2 subunit (Williams et al. 1997). Motor neurones also have a relative lack of expression of calcium buffering proteins (Ince et al. 1993) and appear to have a high threshold for mounting a protective heat shock response. (Durham 2003) Recent studies suggest that the properties of mitochondria from the spinal cord may differ from those of mitochondria from other tissues (Sullivan et al. 2004).

The incidence of motor neurone disease reported from recent epidemiological studies ranges from 1 and 3 per 100 000, with point prevalence rates of 6–8 per 100 000 (Chancellor and Warlow 1992; Traynor et al. 1999). Most of the studies have been conducted from developed countries and relatively little is known of the incidence and prevalence in developing countries, or in specific racial or ethnic groups. Pockets of high incidence are described amongst the Chamorro indigenous population of the Western Pacific island of Guam, on the Kii peninsula of Japan and amongst the Auyu and Jakai people of Irian Jaya (Plato et al. 2003; Kuzuhara and Kokubo 2005). The explanation for the strikingly increased incidence and prevalence of motor neurone disease in these geographical foci remains uncertain. Although the prevalence remains high in Guam compared to typical populations in western countries, there has been a substantial decrease over the last half century.

Some studies have indicated that the overall age-related incidence of motor neurone disease has increased over several decades (Lilienfeld et al. 1989; Maasilta et al. 2001). However, it is unclear whether this is due to demographic factors, better ascertainment, or to changed exposure to unknown environmental risk factors. A gradual increase in the prevalence of motor neurone disease is expected given the changing age structure of the population and with the introduction of therapies which extend survival of patients.

The incidence of motor neurone disease increases with age, being very low under the age of 40 and peaking at approximately 75 years of age. The reported mean age of onset in sporadic motor neurone disease varies between 55 and 65 years in most studies, with a range varying between the third decade and the ninth decade (Jokelainen 1976; Kurtzke 1991). The mean age of onset in patients with familial motor neurone disease is about a decade earlier. Occasionally patients with classical motor neurone disease present in the second or third decade of life. Sporadic motor neurone disease is commoner in men than women, with a male/female ratio of around 1.6:1. This ratio approaches 1:1 in familial motor neurone disease. Women are relatively over-presented in older age groups, although the standardized age-related incidence is greater in elderly men. Bulbar onset is also more common in older patients, especially in older women (Haverkamp et al. 1995; Forbes et al. 2004). In most large clinic-based or population-based series, 5–10 per cent of cases are classified as familial.

The only risk factors proven to have an association with the development of motor neurone disease are gender, a positive family history, and increasing age. At the present time no environmental risk factors are regarded as of proven causative significance. Environmental factors which have been reported to increase the risk of developing motor neurone disease include trauma, physical activity, participation in athletic pursuits, dietary habits, alcohol consumption, cigarette smoking, residence in rural rather than urban areas, and working in certain occupations, for example the leather industry or electrical work. Using an evidence-based medicine approach, Armon (2003) concluded that smoking is probably associated with amyotrophic lateral sclerosis, but that the evidence in favour of other reported environmental factors was not strong. Recent studies have reported an increased risk of developing motor neurone disease in military personnel, in airline pilots, and in Italian professional football players (Horner et al. 2003; Chio et al. 2005; Weisskopf et al. 2005). Further investigation is required to substantiate these findings and to address the underlying causes of the reported occupational associations.

There are two main goals in the management of motor neurone disease. The first is the alleviation of symptoms which occur during the course of the disease to maintain quality of life. The second is the administration of neuroprotective therapy to slow the progression of motor neurone injury and neurodegeneration.

The management of motor neurone disease poses considerable ethical, logistical, and educational problems. Ethical issues are involved in aspects of management including the use of artificial methods for maintaining nutrition, ventilatory support, the use of neuroprotective drugs to slow disease progression, and the use of opiate medication in the terminal phase of the disease. The logistical and educational problems arise from the relative rarity of motor neurone disease and the fact that many health care professionals have little experience in dealing with the rapidly progressive weakness and bulbar and respiratory failure which may occur during the course of the disease. The coordinated action of multiple health care professionals within a multidisciplinary team can lessen the difficulties experienced by patients and by their families.

Symptomatic therapy aimed at alleviating the distressing symptoms which often arise during the course of motor neurone disease can do much to improve the quality of life for the patient. Detailed discussion of all of these therapies is beyond the scope of this chapter. Some of the common symptoms which may develop, and their symptomatic therapies are highlighted in Table 23.5. Many of the symptomatic therapies currently recommended by clinicians have not been assessed in rigorous controlled trials. The evidence base for some of these therapies was reviewed several years ago by an American Academy of Neurology task-force (Miller et al. 1999). A few areas of progress in the symptomatic management of motor neurone disease will be highlighted.

Specialist multidisciplinary clinics. There is an increasing tendency in the United Kingdom and worldwide for patients with motor neurone disease to be managed in specialist clinics. This allows the coordination of an experienced multi-disciplinary team, that includes the neurologist, specialist nursing staff, physiotherapist, occupational therapist, speech therapist, dietician, social worker, and orthotist. Input may also be required from other specialist teams including respiratory medicine, gastroenterology, and palliative care. Considerable support is also provided by patient associations such as the Motor Neurone Disease Association.

Nutritional management. Weight loss is universal in motor neurone disease patients and may be due to dysphagia, loss of muscle mass, or anorexia. Weight loss, malnutrition, and dehydration can aggravate muscle weakness and shorten lifespan, whilst frequent choking spells can make mealtimes intolerable. Malnutrition is an independent prognostic factor for survival in amyotrophic lateral sclerosis with an almost 8-fold increased risk of death in patients who are malnourished (Desport et al. 1999). The management of the nutritional status of motor neurone disease patients has improved in recent years. When the patient first begins to develop swallowing problems a few simple measures can be helpful. Dysphagia may be increased by anxiety and the social embarrassment resulting from slowness in eating, dribbling, and choking. Patients should be encouraged to eat in as relaxed and comfortable an environment as possible. Sucking ice before meals may decrease choking spells. Attention to food consistency is important, and the family should receive advice from a dietician. If the patient is continuing to lose weight, then nutritional supplements of liquid or semi-solid consistency may be helpful and a range of these should be tried to ascertain which preparations are most palatable for the individual patient. More active therapeutic intervention for the dysphagia should be considered when the following problems are apparent:

For long-term enteral feeding percutaneous endoscopic gastrostomy, or PEG, is the procedure of choice. The placement of a percutaneous endoscopic gastrostomy is a relatively straight-forward procedure which can be performed under a local anaesthetic. After placement many patients report great relief and increased well-being, though as yet no large scale quality of life studies have been conducted. Percutaneous endoscopic gastrostomy feeding in motor neurone disease has been the subject of a recent Cochrane review (Langmore et al. 2006), and represents one of the major advances in symptomatic care for patients, leading to weight stabilization and adequate nutrional and fluid intake, although a survival benefit has yet not been convincingly shown. The need for percutaneous endoscopic gastrostomy feeding should be anticipated as the risks of the procedure are higher once the patient’s forced vital capacity falls below 50 per cent (Miller et al. 1999). In some patients, technical difficulties may be experienced in the insertion of a percutaneous endoscopic gastrostomy tube and in this situation, a radiologically guided method may be used (Thornton et al. 2002). In patients with a low vital capacity, the use of non-invasive positive pressure ventilation during percutaneous endoscopic gastrostomy insertion has been shown to improve tolerance and safety of the procedure (Gregory et al. 2002). A full strength feeding regimen providing 1500–2000 calories per day can be introduced over 48 h following feeding tube insertion. Feeding can be managed by syringe boluses at normal meal-times, or by continuous infusion by pump which can be given overnight. Some patients defer consenting to undergo percutaneous endoscopic gastrostomy insertion until an advanced disease stage is reached. In these patients, as well as attention to respiratory status, vigilance is required for the possibility of the re-feeding syndrome as a post-operative complication (Fotheringham et al. 2005).

Respiratory support. Respiratory muscle weakness develops insidiously during the course of motor neurone disease, causing dyspnoea, orthopnoea, and symptoms of carbon dioxide retention, which include daytime somnolence, morning headaches, and lack of restorative sleep, with frequent waking. The management of the respiratory complications include the following general measures. Attention should be given to the detection and prevention of aspiration pneumonia. Antiobiotic therapy should be used at the first indication of a chest infection. When the patient has difficulty in clearing secretions from the chest, chest physiotherapy and postural drainage should be used if possible. The provision of a suction machine and the prescription of a mucolytic agent such as carbocisteine to reduce the viscosity of secretions may also be helpful. Patients will breathe more comfortably during sleep if placed in a semi-upright position. When patients experience bouts of severe dyspnoea, accompanied by extreme anxiety or panic, a small dose of lorazepam may be useful, 0.5–1 mg sublingually. If breathlessness causes distress during the later stages of the disease, the use of small amounts of morphine will be useful. Further depression of respiration can usually be avoided if the initial dose is small and increments are gradual.

Assisted ventilation, coupled with appropriate nutritional support, could theoretically extend the patient’s life indefinitely, and the implications of initiating such respiratory support must be clearly thought through and discussed for each patient. There are considerable international differences in the use of assisted ventilation to manage respiratory failure in amyotrophic lateral sclerosis. The progressive nature of the condition has acted as a deterrent, in some countries, for the active management of respiratory dysfunction in many patients, particularly when other motor disabilities are extensive. Full 24-h intermittent positive pressure ventilation via a tracheostomy, is an option that is chosen only rarely by fully informed patients. The costs of tracheostomy ventilation, in terms both of financial resources and the caregiver support required, are substantial. Non-invasive intermittent positive pressure ventilation via a mask (Fig. 23.5) is a practical option for respiratory support. Non-invasive positive pressure ventilation used overnight has been shown to alleviate symptoms of chronic hypoventilation and to significantly improve several measures of quality of life (Lyall et al. 2001; Bourke et al. 2003). In the United Kingdom only a small proportion of patients with amyotrophic lateral sclerosis are treated with non-invasive positive pressure ventilation, and there is marked variation in clinical practice (Bourke et al. 2002). This may partly be due to regional variation in the availability of non-invasive positive pressure ventilation. The early signs and symptoms of hypoventilation are also subtle and easily overlooked.

Recent work has demonstrated that the criteria most predictive of symptomatic benefit from non-invasive positive pressure ventilation in patients with motor neurone disease are: orthopnoea, daytime hypercapnia, nocturnal oxygen desaturation, together with relatively preserved bulbar function (Bourke et al. 2003). A recently published randomized, controlled trial of non-invasive ventilation in motor neurone disease has shown that in patients without severe bulbar dysfunction, survival is significantly extended with a median survival benefit of approximately 7 months, with maintenance of and improvement in multiple quality of life measures (Bourke et al. 2006). The survival benefit from non-invasive positive pressure ventilation in this group is much greater than from currently available neuroprotective therapy. Patients with severe bulbar dysfunction have more difficulty in tolerating non-invasive positive pressure ventilation for hours at a time and in this group there was some improvement in sleep-related symptoms but no demonstrable survival benefit.

End of life care. In United Kingdom practice the opening up of hospice places for patients with amyotrophic lateral sclerosis has greatly improved the quality of care and support for patients in the later stages of the disease. If motor neurone disease patients are not ventilated, they will almost always die in their sleep from hypercapnic coma. In the terminal phases of illness the aim of treatment is to ensure that the patient is comfortable, and opiate and anxiolytic medication should be used as required to alleviate discomfort or distress.

There is no therapy currently available which has a dramatic effect in slowing disease progression in amyotrophic lateral sclerosis. However, some small steps have been made towards this ultimate goal in recent years. An understanding of the molecular pathways that lead to motor neurone death (Section 23.2.4), is needed in order to target therapeutic strategies. These insights into the mechanisms of neuronal degeneration have led to the development of a number of compounds which protect neurones in cell culture and in animal models of motor neurone disease. Over 50 potential neuroprotective agents have been tested in clinical trials which have been extensively reviewed elsewhere (Meininger et al. 2000; Turner and Leigh 2003). The larger recent trials, and their theoretical and experimental basis are summarized in Table 23.6.

Riluzole. Has been shown to significantly slow disease progression and is the only neuroprotective agent licensed for use in motor neurone disease. It is a sodium channel blocker whose primary mechanism of action is to reduce excitotoxicity through inhibition of glutamate release. Also it has been shown to have several other potentially neuroprotective effects. Two double-blind placebo-controlled trials of riluzole have been carried out in more than 1100 patients (Bensimon et al. 1994; Lacomblez et al. 1996). A Cochrane review of riluzole therapy in motor neurone disease concluded that there is a statistically significant, although modest, effect in prolonging survival by approximately 3 months (Miller et al. 2002). No clear effect on muscle strength was demonstrated, and neither trial evaluated quality of life. In view of the high cost to benefit ratio, there has been controversy about the use of riluzole worldwide. In the United Kingdom, its use is recommended by the National Institute for Clinical Excellence, which estimated the cost of therapy to be £34 000 to £43 500 per quality-adjusted life year, or QALY. Riluzole therapy is relatively expensive and the average survival benefit to be expected is modest, but against this must be weighed the arguments that the patient population requiring the drug is relatively small, and that these patients are facing a lethal disease for which no other therapy is available. At presentit is unknown whether the modest overall therapeutic effect ofriluzole conceals individual good responders and non-responders.

Other neuroprotective agents. Several compounds that appeared to protect neurones from degeneration in cell culture and animal models have had disappointing results in human clinical trials. There are two possible explanations for this. First, the models used may not accurately reproduce human disease or the testing in the models may be insufficiently rigorous. Several drugs effective in SOD1 transgenic mice have not been beneficial in human trials, including gabapentin, creatine, topiramate, and vitamin E. Potential explanations for failure of translation into effective human neuroprotective therapies include starting therapy presymptomatically in mice, and deficiencies in the design of mouse trials including failure to take into account gender and litter effects. Also problems with the design and methodology of human clinical trials in the past could mask a modest clinical benefit. To improve the design and implementation of clinical trials in motor neurone disease, the World Federation of Neurology published consensus guidelines in 1998 (Miller et al. 1999).

Future developments. The next few years are likely to see further progress in defining the molecular mechanisms of cell death underlying the neurodegenerative process in amyotrophic lateral sclerosis. It can be anticipated that further genetic mutations associated with familial disease and the contribution of genetic factors to the sporadic form of the disease will be identified. The continuing use and refinement of cellular and animal models will allow researchers to understand the sequential molecular events leading to motor neurone cell death and to evaluate new neuroprotective strategies.

Several potential neuroprotective agents are being evaluated or will shortly be entered into clinical trials in amyotrophic lateral sclerosis. These include minocycline, arimoclomol, glatiramer acetate, co-enzymeQ10, celecoxib, and ONO-2506. Automated laboratory assays of neurodegeneration can rapidly screen thousands of chemicals to identify lead compounds for drug development. The traditional reluctance of pharmaceutical companies to invest heavily in rarer diseases has recently been addressed by the emergence of non-profit-making biotech companies and academic institutions using high-throughput drug screening to identify compounds of interest. Future neuroprotective therapy for patients with amyotrophic lateral sclerosis may well involve a ‘cocktail’ of pharmacological agents aimed at different mechanisms contributing to the biochemical cascade of cell injury. Gene therapy approaches using viral vectors pseudotyped to ensure retrograde transport within motor neurone axons following intramuscular injection have shown great promise in murine models of motor neurone disease (Azzouz et al. 2004) and phase 1 human trials are expected in the near future. Cell replacement therapy using stem cells poses particular difficulties in relation to amyotrophic lateral sclerosis, not least because of the long axonal processes required for the normal function of motor neurones and the inhibitory signals preventing the effective growth and correct synaptic alignment of newly generated motor neuronal axons. Perhaps the earliest promise of cell replacement therapy for motor neurone disease will be to try to create a supportive environment of non-neuronal cells in the vicinity of motor neurones to achieve neuroprotection of existing differentiated motor neurones. Measures to improve supportive care for patients and their families will continue to develop including the judicious use of ventilatory support. As clinical and scientific developments allow the prospect of increasing the duration of the disease and prolonging survival of patients with motor neurone disease, very careful attention needs to be paid to the quality of life of afflicted individuals.

The term spinal muscular atrophy encompasses a group of genetically determined pure lower motor neurone disorders in which degeneration of the anterior horn cells leads to progressive, symmetrical muscle weakness, and wasting, with sparing of sensation, and absence of pyramidal tract involvement. As the bulbar musculature may be affected, and motor neurone degeneration is therefore not confined to the spinal cord, an alternative term ‘hereditary motor neuronopathy’ was proposed, and both terms are currently in use. There are difficulties with the definition and classification of spinal muscular atrophy, which are gradually being resolved as the underlying genetic defects are identified. By definition spinal muscular atrophy is genetically determined, but adult patients sometimes present with what appears to be a sporadic form, and a genetic basis is therefore unproven. The neurodegenerative process in spinal muscular atrophy has a predilection for motor neurones, but in some families, pyramidal tract and sensory involvement are also seen. Thus the condition shows some overlap with other neurodegenerative diseases affecting the central and peripheral motor systems.

This is the most common type of spinal muscular atrophy. It is inherited as an autosomal recessive disorder and is one of the most common lethal childhood autosomal recessive diseases. Patients develop predominantly proximal limb weakness, with relative sparing of the facial muscles and the diaphragm. Proximal spinal muscular atrophy is divided into subtypes, according to severity and age of onset (Zerres and Rudnik-Schoneborn 1995):

Linkage analysis revealed that the three subtypes of proximal recessive spinal muscular atrophy mapped to chromosome 5q13. This region of chromosome 5 is complex, and characterized by low copy repeats, which may account for instability of this region, and trigger frequent deletions or gene conversions. Within the spinal muscular atrophy critical region, there is a 500 kB inverted duplication, with four genes present in at least two copies, telomeric and centromeric: the survival motor neuron gene, SMN, the neuronal apoptosis inhibitory protein gene, NAIP, the gene encoding BTF2p44, a subunit of RNA polymerase II involved in transcription, and a putative RNA binding protein, H4F5. Homozygous deletion of the telomeric copies of all four genes have been described in patients with spinal muscular atrophy, but the frequency of gene deletions is much higher for SMN1 than for the other 3 genes and it is now well established mutations affecting SMN1 cause 5q13-linked spinal muscular atrophy. In a series of 525 patients with classical spinal muscular atrophy, 96 per cent were linked to chromosome 5q13, and all of these showed mutations in SMN1 (Wirth et al. 2000). Furthermore, mice possess only one survival motor neuron gene, Smn, loss of which is embryonically lethal. However, Smn^−/−;SMN2 mice that carry one or two copies of human SMN2 develop a phenotype of motor neuron degeneration that resembles spinal muscular atrophy in type (Monani et al. 2000).

Duplication of the SMN gene occurred more than 5 million years ago, before the separation of human and chimpanzee lineages and subsequent sequence divergence in Homo sapiens has led to 5 base pair differences between SMN1, and its centromeric homolog SMN2. The primary gene sequences of SMN1 and SMN2 predict identical proteins, but the translationally silent change in exon 7 of SMN2 decreases the activity of an exonic splicing enhancer, leading to skipping of exon 7 and a truncated protein in 80 per cent of the transcript produced by SMN2. Homozygous absence of SMN2, found in about 5 per cent of controls, has no clinical phenotype. The majority of 5q13-linked spinal muscular atrophy patients show homozygous absence of SMN1 exon 7. This may occur through gene deletion, often a large deletion that includes the whole gene, or several genes within the critical region for spinal muscular atrophy. Alternatively, SMN1 may be replaced by a copy of SMN2 during DNA replication, a process known as gene conversion.

Multiple more subtle intragenic mutations have been described in spinal muscular atrophy, the majority of which produce a truncated protein, either through splice site mutations that disrupt exon 7, or through nonsense or frameshift mutations which introduce a stop codon. Missense mutations show an interesting pattern of clustering which provides some insight into the functional domains of the SMN protein. A tyrosine–glycine rich sequence at the C-terminal of SMN encompasses five of the described missense mutations, and is a highly conserved sequence, identical in yeasts and nematodes. This region has homology to RNA interacting proteins. A further cluster of mutations is seen in the central region of SMN, which constitutes a so-called Tudor domain, an evolutionarily conserved sequence of unknown function found in many eukaryotic proteins.

SMN2 produces only 20 per cent of full-length protein. This is insufficient to rescue the phenotype in homozygous deletion of SMN1, but there is strong evidence that SMN2 is a disease-modifying gene for spinal muscular atrophy. A molecular basis for the wide variation in the severity of the phenotype resides in the fact that homozygous absence of SMN1 can be due to gene deletion or to conversion to SMN2. A patient may have anything from 1 to 4 copies of SMN2, leading to a progressive increase in the amount of full-length protein. Both the copy number of SMN2 and the protein levels of SMN have been shown to correlate with severity of disease phenotype (Feldkotter et al. 2002; Lefebvre et al. 1997). However, the copy number of SMN2 in types I, II, and III spinal muscular atrophy overlaps, therefore this alone cannot expain the phenotypic variation of the disease, and other disease-modifying factors must exist.

The majority of type I spinal muscular atrophy patients display large scale 5q13 deletions, removing SMN1 and adjacent microsatellite markers, whereas type III patients tend to have small deletions or gene conversions affecting only SMN1(Rodrigues et al. 1996). This suggests that a spinal muscular-atrophy-modifying locus distinct from SMN1 lies in the 5q13 interval. NAIP is a good candidate as it functions as a negative regulator of apoptosis. NAIP deletion occurs in 45 per cent of type I spinal muscular atrophy patients, and 18 per cent of type II and III patients, therefore loss of NAIP may lead to a more severe disease phenotype (Roy et al. 1995). H4F5 lies closer to SMN1 than any other known gene. Homozgous deletions of this gene have been found in 90 per cent of type I spinal muscular atrophy cases, and it has also been proposed as a disease-modifying gene. However, deletions of both NAIP and H4F5 may simply reflect larger chromosomal deletions that involve both SMN1 and SMN2, and their role in the pathogenesis is not confirmed.

SMN produces a protein of 294 amino acids that is widely expressed. In spinal muscular atrophy patients, the level of the SMN protein is only moderately reduced in muscle and lymphoblasts, but is reduced 100-fold in the spinal cord of type I patients (Coovert et al. 1997). SMN protein shows diffuse cytoplasmic expression, but within the nucleus, is clustered in suborganelles called ‘gems’, for ‘gemini of coiled bodies’. These are similar in size and number to, and often associated with Cajal or coiled bodies, which are known to have a role in mRNA metabolism. The SMN protein oligomerizes, and associates with six proteins named Gemins, to form the SMN complex. Self-association occurs through an oligomerization domain in exon 6, and appears to be essential for its activity (Lorson et al. 1998).

The SMN complex interacts with several proteins, many of which are involved in RNA metabolism. These are ubiquitous cellular processes, which would indicate that the clinical features of spinal muscular atrophy may be caused by a particular susceptibility of lower motor neurones to defects in RNA handling. Alternatively, SMN may have as yet unidentified functions that are specific to the motor neurone. The list of proteins reported to interact with SMN also includes several which are not involved in RNA metabolism, including profilin, the FUSE binding protein, ZRP1, and p53. The functional significance of these interactions is currently unknown.

Within the cytoplasm, the SMN complex has an important role in the assembly of spliceosomal small nuclear ribonucleoproteins, snRNPs (Pellizzoni et al. 1998; Yong et al. 2004). SMN has also been shown to have a function in pre-mRNA splicing in the nucleus. SMN mutations found in patients with spinal muscular atrophy cause deficiencies in splicing regeneration activity and interactions with other key proteins, but it is uncertain whether the function of SMN in pre-mRNA splicing explains the motor neurone specific pathology in spinal muscular atrophy. The SMN complex also interacts with viral transcriptional activators, and with RNA polymerase II, pol II, which physically and functionally couples transcription, splicing, and polyadenylation, an association mediated by RNA helicase A, RHA. Expression of a dominant-negative mutant of SMN causes accumulation of pol II and RHA in the nucleus, and inhibits transcription in vivo, suggesting a role for SMN in transcriptional regulation (Pellizzoni et al. 2001).

Motor neurone-specific functions of SMN. SMN has been shown to interact with two heteronuclear ribonucleoproteins, hnRNP-R and hnRNP-Q, which are considered to play important roles in mRNA editing, transport, and splicing. HnRNP-R is predominantly located in the axons of motor neurones, where it colocalizes with SMN, a finding which led to the confirmation of a motor-neurone-specific function of SMN (Rossoll et al. 2002). In zebrafish, knock-down of SMN protein levels in the developing embryo, caused pathfinding defects specific to the motor axon (McWhorter et al. 2003). Similarly, primary motor neurones cultured from a transgenic spinal muscular atrophy mouse model show reduced axon growth, while overexpression of SMN and hnRNP-R in cultured neuronal cells promoted neurite outgrowth (Rossoll et al. 2003). These findings raise the possibility that SMN is involved in transport of mRNA molecules in the axons of motor neurones.

Other disorders may present in infancy with hypotonia and a pattern of weakness identical to Werdnig–Hoffman disease, which are distinguished by associated features. The aetiological relationship of these disorders to classical spinal muscular atrophy has been clarified by testing for SMN mutations:

A missense mutation, P56S, in the vesicle-associated membrane protein, VAPB gene on chromosome 20q13.3 was identified in seven kindreds, with three different phenotypes of autosomal dominant motor neurone disease (Nishimura et al. 2004). Some patients had an atypical slowly progressive form of amyotrophic lateral sclerosis with tremor, ALS 8, some had typical rapidly progressive familial amyotrophic lateral sclerosis, while others presented with adult onset proximal spinal muscular atrophy, Finkel Type. VAPB is discussed further in Section 23.2.5.

The distal distal spinal muscular atrophies or the distal hereditary motor neuronopathies, or dHMN, are a more diverse group of disorders. They commonly manifest as a peroneal muscular atrophy syndrome, distinct from the Charcot–Marie–Tooth syndrome (Section 21.4), which causes a similar pattern of weakness, by the lack of sensory involvement. The clinical picture is one of progressive weakness of the toes and feet which may be associated with foot deformity and which extends over time to involve the distal upper limb muscles. Some patients have unusual or additional features, including predominant involvement of the hands, vocal cord paralysis, diaphragm paralysis, and pyramidal tract signs. Harding proposed seven subtypes of distal hereditary motor neuronopathy (Harding 1993; European CMT Consortium 1998), on the basis of genetic and clinical criteria. This classification was reviewed by the European Charcot–Marie–Tooth Consortium in 1998, but this is constantly evolving with the identification of novel clinical and genetic entities, the finding that previously delineated phenotypes show genetic heterogeneity, and that single gene disorders can vary widely in phenotype (Irobi et al. 2004).

In a series of 200 patients with infantile-onset spinal muscular atrophy, Rudnik-Schoneborn et al. (1996) found that approximately 1 per cent had diaphragmatic weakness and did not have deletions of the SMN gene on chromosome 5q. This subtype of infantile spinal muscular atrophy, known by the acronym SMARD, is characterized by severe breathing difficulties and limb weakness which is predominantly distal. SMARD1 is caused by mutations in the gene encoding immunoglobulin mu-binding protein 2, IGHMBP2, which has RNA helicase activity, and is involved in pre-mRNA processing (Grohmann et al. 2001). Some affected infants have evidence of sensory and autonomic nerve involvement as well as the motor deficit. Mutations in the homologous murine gene, ighmbp2 are responsible for spinal muscular atrophy in the ‘neuromuscular degeneration’, nmd mouse, which shows close phenotypic resemblance to SMARD1 (Cox et al. 1998).

This disorder causes slowly progressive distal limb weakness and has a very variable age of onset. In the most severely affected patients, there is evidence of diaphragmatic involvement, with a decrease in vital capacity and elevation of the hemidiaphragms on chest radiographs. In a large consanguineous Lebanese pedigree, genetic linkage has been mapped to chromosome 11q though mutations in IGHMBP2 have been excluded. A distinct form of autosomal recessive distal spinal muscular atrophy, identified in families from the Jerash region of Jordan, has been mapped to chromosome 9p21.1-p12.

Pedigrees with distal spinal muscular atrophy with upper limb predominance, designated dHMN-V, often include affected individuals with evidence of mild sensory disturbance or upper motor neurone features. The disorder was linked to chromosome 7p in a large Bulgarian family (Christodoulou et al. 1995). Subsequently a Mongolian family was identified in which Charcot–Marie–Tooth disease type 2D and distal spinal muscular atrophy with upper limb predominance segregated in the same kindred. All affected members had weakness and wasting of the intrinsic muscles of the hands. Those with no sensory deficit and peroneal muscle weakness had a diagnosis of distal spinal muscular atrophy, while in other affected individuals Charcot–Marie–Tooth disease was diagnosed on the basis of a glove and stocking sensory loss. Both disorders in this family were linked to the same region of chromosome 7p. Mutations were subsequently uncovered in the gene encoding glycyl tRNA synthetase, GARS, in Charcot–Marie–Tooth disease type 2D families, in families with dHMN-V, and in the family described with both disorders (Antonellis et al. 2003). Identification of the underlying genetic disorder has therefore confirmed that Charcot–Marie–Tooth Type 2D and distal spinal muscular atrophy with upper limb predominance are phenotypic variants of a single disorder. GARS is a member of the family of aminoacyl tRNA sythetases involved in diverse cellular processes, including charging tRNAs with their appropriate amino acids. This defect would be expected to affect every glycine-containing protein, and it is at present unknown why this gene defect specifically impairs the function of neurones.

Linkage to chromosome 7p15 was excluded in a large Austrian family with dHMN-V. In this family and several others, heterozygous mutations were identified in the Berardinelli–Seip congenital lipodystrophy gene, (Windpassinger et al. 2004). The gene encodes seipin, an integral membrane protein of the endoplasmic reticulum, and null mutations cause Berardinelli–Seip congenital lipodystrophy. N88S and S90L mutations affect glycosylation of seipin, resulting in the formation of intracellular aggregates. The same mutations were also found to be associated with the Silver syndrome variant of hereditary spastic paraplegia (Section 23.4.2). This is a further example of the same genetic defect causing two quite distinctive phenotypes. The clinical features associated with BSCL2 mutations have recently been broadened to include individuals who have lower limb predominant distal amyotrophy in the absence of pyramidal tract signs, and some with a combination of spasticity and severe amyotrophy in the lower limbs (Irobi et al. 2004).

Designated dHMN-II, this is usually inherited as an autosomal dominant trait, but sporadic cases are frequently described, which may reflect late-onset recessive disease, non-genetic aetiology, or new mutations. One large Belgian pedigree was linked to chromosome 12q24 and subsequently heterozygous mutations were found in the gene encoding the small heat shock 22 kDa protein 8, HSPB8/HSP22, in four families (Irobi et al. 2004c). Missense mutations have also been identified in the interacting partner of HSP22, heat shock 27kDa protein 1, HSPB1/HSP27, both in patients with dHMN-II and in pedigrees with Charcot–Marie–Tooth Type 2F (Evgrafov et al. 2004). Most of the HSP22 and HSP27 mutations disrupt a conserved αcrystallin domain in these proteins and cell biological studies have shown that HSP22 mutants show greater binding to HSP27, resulting in the formation of intracellular aggregates (Irobi et al. 2004a). HSP27 is also involved in the organization of the neurofilament network, important for the maintenance of the axonal cytoskeleton and for axonal transport, and mutant HSP27 perturbs neurofilament assembly (Evgrafov et al. 2004).

Distal spinal muscular atrophy with vocal cord paralysis, designated dHMN-VII, characterized by distal limb weakness and dysphonia, has been described to belinked to chromosome 2q14 in two large Welsh families with common ancestry (McEntagart et al. 2001). Three disorders have been described which show similar clinical features to dHMN-VII: one family with distal spinal muscular atrophy and vocal cord paralysis also had sensorineural hearing loss, (Bolthauser et al. 1989) while Charcot–Marie–Tooth disease type IIC is characterized by progressive vocal cord paralysis, distal limb weakness, and sensory loss (Donaghy and Kennett 1999). The gene alterations underlying these two disorders are currently unknown. A third disorder, presenting with breathing difficulty due to vocal cord paralysis, progressive facial weakness, and weakness and atrophy of the hands, has been found to be caused by a missense mutation in the gene encoding dynactin, DCTN1, on chromosome 2p13. This mutation is predicted to distort the folding of the dynactin microtubule binding domain, and lead to dysfunction of dynactin-mediated retrograde axonal transport (Puls et al. 2003).

This relatively benign disorder, described in pedigrees from Holland and Canada, has been shown by linkage analysis to map to chromosome 12q23-24 (van der Vleuten et al. 1998). Again, the disorder is genetically heterogeneous, as in one family linkage to this region has been excluded.

Muscle weakness and wasting in a scapuloperoneal distribution may be myopathic or neurogenic in origin. Several large pedigrees with neurophysiological evidence of denervation have been described. In one pedigree, linkage has been mapped to chromosome 12q24.1-12q24.31 (Isozumi et al. 1996) which is close to the region of chromosome 12 to which a myopathic scapuloperoneal syndrome is linked (Wilhelmsen et al. 1996). Interestingly many of the families described with scapuloperoneal spinal muscular atrophy include individuals with both myopathic and neurogenic changes in affected muscles. Congenital non-progressive spinal muscular atrophy affecting the lower limbs maps to the same region as scapuloperoneal spinal muscular atrophy, but it has not yet been established whether these disorders are allelic.

Monomelic amyotrophy has been chiefly reported from Japan and India. Hiroyama et al. (1959) reported unilateral atrophy of the upper limb and gave it the term ‘juvenile muscular atrophy of unilateral upper extremity’. Later a report from India described cases of atrophy of the muscles of one lower limb and described it as the ‘wasted leg syndrome’ (Prabhakar et al. 1981). Gourie-Devi et al. (1984) suggested the term monomelic amyotrophy to cover both of these entities. Monomelic amyotrophy has been reported to account for 8–29 per cent of all motor neuron diseases in the series reported from India (Gourie-Devi et al. 1984; Saha et al. 1997).

Monomelic amyotrophy is a benign variant of motor neurone disease that predominantly affects young men. Lower motor neurone features of wasting and weakness are usually confined to one upper or less commonly the lower limb, without involvement of other components of the nervous system. Patients with the upper limb variant often report tremulousness of the fingers. In the upper limb, the muscles innervated by the C7 to T1 spinal segments i.e. the intrinsic hand muscles and flexors and extensors of the wrists and fingers, tend to be most severely affected. Relative sparing of the brachioradialis muscle amongst the surrounding atrophic forearm muscles is a characteristic feature of this condition (Hirayama et al. 1963). The disorder is usually sporadic, although rarely it may be familial (Nalini et al. 2004). In the lower limb variant, the muscle atrophy most commonly involves both proximal and distal muscles, but some patients have predominant involvement of either the thigh or lower leg musculature (Prabhakar et al. 1981; Gourie-Devi et al. 1984). Muscle cramps and fasciculations are observed in 20–30 per cent of patients and unilateral pes cavus may be a presenting feature in the lower limb variant.

Some patients develop similar symptoms in the contralateral limb, but the disorder usually remains strikingly asymmetrical. In most cases, the onset is insidious, with a slow progression over 2–4 years, followed by a plateau phase, although further weakness may evolve for up to 8 years (Peiris et al. 1989; Gourie-Devi and Nalini 2003a). The tendon reflexes are depressed. Seldom, if ever, does weakness spread to involve other parts of the body, and upper motor neurone signs are absent. It does not evolve to amyotrophic lateral sclerosis.

A neurogenic pattern on electromyography, and histological evidence of neurogenic atrophy of muscle indicate anterior horn cell pathology. Pathological studies in two patients, who died of coincidental causes, showed atrophy of the affected region of the spinal cord with severe, asymmetrical, bilateral loss of anterior horn motor neurones (Hirayama et al. 1987; Araki et al. 1989). Imaging studies have shown evidence of forward displacement of the dural sac during neck flexion in patients with Hirayama syndrome, and this has led to the proposal that the condition may be caused by intermittent compression and ischaemia of the cervical spinal cord. However, other reports have failed to confirm this (Schroder et al. 1999; Willeit et al. 2001) and thus the pathogenic mechanisms underlying Hirayama syndrome and other forms of monomelic amyotrophy remain uncertain. Mutations in the SMN and SOD1 genes have been excluded. A mitochondrial DNA mutation has been reported from Italy in one patient with monomelic amyotrophy and sensorineural hearing loss (Fetoni et al. 2004).

The GM2 gangliosidoses are a group of recessively inherited disorders in which deficiency of the lysosomal enzyme, β hexosaminidase A, leads to abnormal intracellular accumulation of lipids in neurons and glia (Section 10.4.2). Tay-Sachs disease, the classical infantile form of GM2 gangliosidosis associated with mutations in both alleles of the HEXA gene, is the most frequent form of the disorder and is particularly found amongst Ashkenazi Jews. Affected children, after a few months of normal development, start to regress, lose head control and become hypotonic and apathetic, with the development of seizures and cortical blindness. Examination will frequently reveal a large head, a macular cherry red spot, spasticity, quadriparesis, and an exaggerated startle response. The disorder is usually fatal before the age of 5–6 years. GM2 gangliosidosis can also occur as a later onset disease, arbitrarily divided into juvenile, early adult, and late adult forms. The late-onset forms can cause a disorder of the upper and lower motor neurons. However, there is usually evidence clinically of a multisystem disorder with involvement of the cerebellum and its connections, the autonomic nervous system, and extrapyramidal manifestatons such as Parkinsonism, dystonia, or choreoathetosis. Patients may also have evidence of peripheral neuropathy and develop features of dementia or psychosis. Pathologically the most characteristic change is the presence of swollen ballooned neurons with storage material in lysosomes and characteristic membranous cytoplasmic inclusion bodies. Meganeurites are formed which are associated with aberrant dendritic, neuritic, and synaptic growth.

Several reports have described the phenotype of motor neurone disease in adult hexosaminidase deficiency (Kaback et al. 1978; Johnson et al. 1982; Mitsumoto et al. 1985). Gudesblatt and colleagues described 52 patients who had atypical amyotrophic lateral sclerosis and 4 of these had partial hexosaminidase A deficiency (Gudesblatt et al. 1988). None of 50 patients with typical amyotrophic lateral sclerosis had abnormal HexA activity and similar findings were reported by Drory and colleagues (Drory et al. 2003). Most cases with a motor system disorder in the context of HexA deficiency have additional neurological manifestations and a pure amyotrophic lateral sclerosis phenotype is extremely rare (Karni et al. 1988; Parboosingh et al. 1997). Muscle weakness tends to develop insidiously during the second decade of life. Weakness is usually first apparent in the proximal muscles and may be associated with muscle fasciculation. Atrophy of the intrinsic hand muscles can also be seen as an early feature. Upper motor neurone features may be present or absent. As described above, the patients on close examination or on follow up over time will often have features indicative of a progressive multisystem neurodegenerative disorder.

Multifocal motor neuropathy (Section 21.11.3) is an acquired disorder which is important to distinguish from amyotrophic lateral sclerosis because it has a much more benign prognosis and is potentially amenable to therapy with intravenous immunoglobulin (Slee et al. 2007). It is characterized by slowly progressive, asymmetrical weakness which develops gradually or with stepwise progression over several years. Initially weakness may not be accompanied by significant muscle wasting. The age of onset is usually between 20 and 50 years and the disorder is more common in men. Weakness is more common in the upper limbs compared to the legs and is usually distal. Common initial symptoms include wrist drop and weakness of grip. Weakness is often more pronounced than would be expected for the degree of muscle wasting which is a clinical clue to the presence of conduction block. However, atrophy may become pronounced in patients with long duration disease. Muscle cramps and fasciculations are reported by two-thirds of patients. Multifocal motor neuropathy is commonly initially misdiagnosed as amyotrophic lateral sclerosis (Traynor et al. 2000). The slowly progressive disease course, the absence of upper motor neurone signs, and the presence of conduction block on neurophysiological examination, which may only be found after repeated testing, are important clinical clues pointing to the correct diagnosis. Raised titres of anti-GM1 ganglioside antibodies in serum may also be a diagnostic clue, though unfortunately the presence of these antibodies is not specific for multifocal motor neuropathy (Taylor et al. 1996).

Acute poliomyelitis is described in detail in Chapter 42.4.2. Some patients who recover partially or fully from acute poliomyelitic weakness develop a new syndrome of progressive motor deficit many years after the original illness. This post-polio syndrome may affect one-third of patients who have had acute paralytic poliomyelitis. The newly reported fatigue, pain, cramps, wasting, weakness, and functional deterioration develop in areas overtly or subclinically affected during the acute attack (Trojan and Cashman 2005). Typically, several decades elapse between the acute paralysis and the onset of post-polio syndrome, with the peak incidence 20–25 years after the original illness. Often no objective change in muscle strength can be detected on manual muscle testing over several years in patients who nevertheless complain of progressive weakness or of progressive difficulty carrying out activities of daily living. Post-polio syndrome must be distinguished from the non-specific symptoms of joint instability; nerve, root, or plexus compression; and increasing scoliosis which may be late secondary effects resulting from the original weakness. A recent systematic review of the relevant literature concluded that at present conclusions cannot be drawn regarding the functional course or prognostic factors in late-onset polio sequelae and that further work, including long-term follow-up studies of unselected patient populations are needed (Stolwijk-Swuste et al. 2005). The cause has not been completely elucidated, but it is likely to be due to distal axonal degeneration of enlarged post- poliomyelitis motor units. Care must be taken to identify coincidental neurological disorders and orthopaedic complications of longstanding weakness, limb dysfunction, and spinal deformity. Although currently there is no specific disease-modifying treatment for post-polio syndrome, an interdisciplinary management programme can be useful in controlling symptoms. There is some evidence that supervised aerobic muscle training and introduction of non-invasive ventilation for patients with respiratory impairment may be helpful measures (Farbu et al. 2006). A recent randomized controlled trial of intravenous immunoglobulin showed slight improvement in some parameters including muscle strength and SF-36 subscale vitality score, but no significant change in overall quality of life or pain (Gonzalez et al. 2006).

Primary lateral sclerosis is considered as part of the spectrum of amyotrophic lateral sclerosis (Section 23.2.1).

Hereditary spastic paraplegia, first described in the 1880s, is a group of hereditary neurodegenerative or neurodevelopmental diseases which affects approximately 1 in 10 000 individuals. The main feature of the clinical phenotype is progressive lower limb spasticity due to degeneration of the corticospinal tracts within the spinal cord. Hereditary spastic paraplegia is most commonly inherited as an autosomal dominant trait, but autosomal recessive and X-linked recessive forms also exist. The hereditary spastic paraplegia phenotype may be ‘pure’ in which the spastic paraparesis occurs in isolation or ‘complicated’ where the spastic paraparesis is one component of a much more complex neurological and/or systemic disorder. Hereditary spastic paraplegia shows extreme genetic heterogeneity. To date more than 30 genetic loci have been identified and genes have been identified at 14 of these (Table 23.7). The recent discovery of multiple genes is rapidly shaping new concepts of the cellular mechanisms of degeneration of the long axons of the corticospinal tract in hereditary spastic paraplegia. It is apparent that motor neurons provide an extreme example of the potential difficulties for a cell in trafficking, transport, and energy metabolism, and that the longest axons of the central nervous system may be specifically vulnerable to several distinct biochemical perturbations.

Pure hereditary spastic paraplegia shows progressive spastic paraparesis as the major feature, but other clinical features may be observed including bladder disturbance, mild distal muscle wasting, pes cavus, dorsal column dysfunction, and loss of ankle reflexes. The most common presenting symptom in pure forms is gait disturbance, the patient often complaining of lower limb stiffness, balance difficulties, or of a tendency to fall. In young children, delayed motor milestones or a tendency to walk on the toes, may be the first indications of a problem. The major features apparent on neurological examination consist of lower limb spasticity, hyperreflexia, and extensor plantar responses which may be accompanied by a mild pyramidal distribution weakness. A characteristic feature of hereditary spastic paraplegia is that the patient’s disability usually arises from prominent spasticity and any accompanying muscle weakness is often very mild.

There is considerable variation in age at onset and severity of the spastic paraparesis even within affected members of the same pedigree, suggesting that other genetic or environmental factors may impact on the phenotype observed. The reported age at onset of pure hereditary spastic paraplegia ranges from infancy to the eighth decade (Harding 1981). Several studies have indicated that an early age of onset, <35 years, tends to be associated with relatively slow disease progression, the majority of individuals retaining the ability to walk even in elderly life. In contrast late onset, >35 years, tends to show more rapid disease progression, and many patients become non-ambulant in the seventh and eighth decades of life (Harding 1981). Hereditary spastic paraplegia does not, in general, reduce life expectancy. Approximately 25–30 per cent of patients have a subclinical phenotype in which symptoms which are so mild that the condition may not be revealed without a neurological examination.

Complicated hereditary spastic paraplegia shows the spastic paraparesis as only one component of a much more complex disorder with additional clinical features (Table 23.8). Some of the complicated hereditary spastic paraplegia phenotypes are extremely rare, having been described only in single families. Certain complicated phenotypes show characteristic clinical features associated with particular hereditary spastic paraplegia loci. For example, hereditary spastic paraplegia associated with cognitive impairment is most commonly described in patients with spastin, SPG4, mutations (White et al. 2000; Webb et al. 1998). In patients with the SPG9 subtype of hereditary spastic paraplegia there is a very distinctive clinical phenotype with the development of cataracts, severe gastroesophageal reflux, and an axonal neuropathy superimposed on the spastic paraparesis (Seri et al. 1999). Pigmentary macular degeneration is an additional feature observed in patients with hereditary spastic paraplegia linked to the SPG15 locus (Hughes et al. 2001). Peripheral neuropathy has been described in patients with hereditary spastic paraplegia linked to the SPG11 and SPG14 loci (Vazza et al. 2000; Mostacciuolo et al. 2000). In the Silver variant of hereditary spastic paraplegia, SPG17, wasting and weakness of the hands, is a striking feature (Patel et al. 2001). Patients with Troyer syndrome, SPG20, have dysarthria, distal amyotrophy, short stature, and developmental delay in addition to spastic paraparesis (Patel et al. 2002).

Epilepsy has been described as a feature complicating hereditary spastic paraplegia, including in families with spastin mutations (Gigli et al. 1993; Yih et al. 1993). There does not appear to be an association with particular types of seizures and the epilepsy may occur before or after the onset of the spastic paraparesis.

Dementia and cognitive impairment have been reported in complicated hereditary spastic paraplegia pedigrees (White et al. 2000) and have also been described as an isolated accompaniment to spastic paraparesis in both autosomal dominant and recessive families (Cross and McKusick 1967; Webb et al. 1998; Pridmore et al. 1995). The cognitive abnormalities observed, for example impairments of attention, perceptual speed, visuomotor coordination, and forgetfulness, are in keeping with a sub-cortical type of dementia and features suggesting major cortical involvement, such as dysphasia, agnosia, and dyscalculia, are usually absent. Several families with dementia complicating HSP have been described with spastin mutations, SPG4, (Webb et al. 1998; White et al. 2000). Affected individuals in SPG4 families may have subclinical cognitive impairment detectable by neuropsychological evaluation (Byrne et al. 2000).

Distal amyotrophy is one of the commonest additional features seen in patients with hereditary spastic paraplegia. Several pathologies can give rise to this amyotrophy including lower motor neurone loss, axonal neuropathy, and central axonopathy. Several characteristic syndromes of hereditary spastic paraplegia with amyotrophy have been described:

Sensory neuropathy of variable severity with onset in childhood or adult life is also a common additional feature in complicated hereditary spastic paraplegia. Severely affected patients may develop chronic painless cutaneous ulcers and neuropathic bone resorption occurring in early life, but in some patients the sensory neuropathy may be subclinical and only detected with neurophysiological testing (Schady and Smith 1994).

Detailed neuropathological findings have been reported in relatively few cases of hereditary spastic paraplegia. Therefore the extent to which the known clinical and genetic heterogeneity is reflected in pathological heterogeneity remains to be defined. In addition, because genetic characterization has only recently emerged, there are few pathological reports on genetically characterized cases. The core neuropathological features of hereditary spastic paraplegia were first described by Strümpell and confirmed in a series of subsequent reports (Sack et al. 1978; Strumpell 1886). The spinal cord shows pallor of the lateral and frequently also the anterior corticospinal tracts, with loss of axons and myelin which most markedly affects the longest descending axons in the lumbosacral region (Fig. 23.6). There is also pallor of the dorsal columns, particularly the medial fibres within the fasciculus gracilis. Involvement of the spinocerebellar tracts is described in approximately 50 per cent of cases. Depletion of Betz cells from the motor cortex is reported in some cases, but anterior horn cells in the spinal cord have usually been reported as appearing normal. Degeneration has occasionally been described in neurones of the dorsal nucleus of Clarke. Characteristically there is more severe involvement of the distal part of the corticospinal tracts and dorsal columns so that the most severe changes in these fibre pathways are observed in the lumbar and cervical cord respectively. This has led to the hypothesis that in hereditary spastic paraplegia the neurodegeneration occurs as a ‘dying back’ axonopathy, affecting the most distal part of these long axons first.

A recent study described the distribution of spastin in the normal human central nervous system, as well as the molecular pathology of three cases of hereditary spastic paraplegia with defined mutations in spastin (Wharton et al. 2003). Spastin was shown to be a neuronal protein widely distributed within the central nervous system. Within motor neurones, spastin was predominantly expressed in the cytoplasm of the cell body, with some extension of staining into the proximal neurites and axons. Interestingly, although not prominently affected clinically in SPG4, the spinal cord lower motor neurones showed evidence of cytopathology, with the presence of hyaline inclusion bodies some of which stained for β-tubulin; and variable loss of expression of n-phosphorylated neurofilament protein, β-tubulin, spastin, and mitochondria from the cell bodies, suggesting altered partitioning of cytoskeletal components and organelles. In addition all three cases showed evidence of tau pathology, with neurofibrillary tangles, neuropil threads, and glial tau pathology.

In many neuropathological reports of pure hereditary spastic paraplegia cerebral structures, with the exception of Betz cells, are said to be uninvolved. However subclinical involvement of other parts of the central nervous system is increasingly recognized and in particular there is emerging recognition of cognitive impairment is some families with hereditary spastic paraplegia. The neuropathological substrate of such cognitive impairment currently remains poorly defined, though isolated case reports have been described (Ferrer et al. 1995; White et al. 2000). In autosomal recessive forms of hereditary spastic paraplegia associated with mutation in the paraplegin gene, SPG7, muscle biopsies have shown ragged red fibres (Casari et al. 1998). Scattered muscle fibres show negative histochemical reaction for cytochrome oxidase, but preserved or elevated succinate dehydrogenase activity and peripheral accumulation of mitochondria. These changes, which are typical of oxidative phosphorylation defects in muscle, support a role for mitochondrial dysfunction in the pathogenesis of paraplegin mutation-associated hereditary spastic paraplegia. Skeletal muscle has not been widely surveyed in other types of hereditary spastic paraplegia, though it has been reported that muscle biopsies from some hereditary spastic paraplegia patients in whom spastin and paraplegin mutations had been excluded showed biochemical evidence of impairment in mitochondrial respiratory chain function (McDermott et al. 2003b).

Molecular genetic testing is increasingly used in the diagnosis. Although 14 causative genes have now been identified, testing for mutations in these genes is not necessary routinely available to the clinicians involved in patient care. Therefore, at present hereditary spastic paraplegia remains a diagnosis of exclusion and conditions which should be considered in the differential diagnosis are included in Table 23.9. It is clearly important to exclude diagnoses in which there is a treatable cause for the spastic paraparesis including structural lesions of the spinal cord, vitamin B₁₂ deficiency, multiple sclerosis, or dopa responsive dystonia. It is also important to exclude other motor system disorders such as familial amyotrophic lateral sclerosis, in which the clinical course and prognosis are significantly different from those to be expected in patients with hereditary spastic paraplegia.