Machado-Joseph Disease

A number sign (#) is used with this entry because Machado-Joseph disease (MJD), also known as spinocerebellar ataxia-3 (SCA3), is caused by a heterozygous (CAG)n trinucleotide repeat expansion encoding glutamine repeats in the ataxin-3 gene (ATXN3; 607047) on chromosome 14q32.

Normal individuals have up to 44 glutamine repeats, and MJD patients have between 52 and 86 glutamine repeats. Incomplete penetrance is associated with 45 to 51 repeats (Todd and Paulson, 2010).

For a general discussion of autosomal dominant spinocerebellar ataxia, see SCA1 (164400).

Description

Machado-Joseph disease, named for affected families of Azorean extraction, is an autosomal dominant progressive neurologic disorder characterized principally by ataxia, spasticity, and ocular movement abnormalities. Although independently described as a seemingly separate disorder, spinocerebellar ataxia-3 is now known to be the same as Machado-Joseph disease.

Three classic clinical subtypes of MJD are recognized: type 1 with early onset and marked pyramidal and dystonic signs; type 2, or pure, with predominant cerebellar ataxia; and type 3 with later-onset and peripheral neuropathy (Franca et al., 2008).

Clinical Features

Early Descriptions, Diagnostic Uncertainties, and Geographic Distribution

Among Portuguese immigrants living in New England, Nakano et al. (1972) described a form of dominantly inherited ataxia occurring in descendants of William Machado, a native of an island in the Portuguese Azores. The disorder began as ataxic gait after age 40. Six patients studied in detail showed abnormally large amounts of air in the posterior fossa on pneumoencephalogram, denervation atrophy of muscle, and diabetes mellitus. Other families of Azorean origin living in Massachusetts (Romanul et al., 1977; Woods and Schaumburg, 1972) and in California (Rosenberg et al., 1976) were reported. Romanul et al. (1977) suggested that all 4 reported kindreds had the same mutant gene despite differences in expression. The progressive neurologic disorder was characterized by gait ataxia, features similar to those in Parkinson disease (PD; 168600) in some patients, limitation of eye movements, widespread fasciculations of muscles, loss of reflexes in the lower limbs, followed by nystagmus, mild cerebellar tremors, and extensor plantar responses. Postmortem examinations showed loss of neurons and gliosis in the substantia nigra, nuclei pontis (and in the putamen in one case) as well as the nuclei of the vestibular and cranial nerves, columns of Clarke and anterior horns. Rosenberg (1977) referred to the disorder he and his colleagues described as Joseph disease (Rosenberg et al., 1976) and questioned that one can be certain of its identity to the disorder in other families of Azorean origin.

In January 1976, Corino Andrade (Coutinho et al., 1977) 'went to the Azores...to investigate a degenerative disease of the central nervous system known to exist there. We saw 40 patients belonging to 15 families (in the islands of Flores and St. Michael)...It is our opinion that different families just mentioned, which have been taken as separate diseases, are only clinically diverse forms of the same disorder, of which symptomatic pleomorphism is a conspicuous feature.' In the same year, Romanul et al. (1977) arrived at the same conclusion. The full paper by Coutinho and Andrade (1978) appeared the next year. Lima and Coutinho (1980) described a mainland Portuguese family. The possibility that the Joseph family was originally Sephardic Jewish was raised by Sequeiros and Coutinho (1981). Mainland families originated in a mountainous and relatively inaccessible region of northeastern Portugal where large communities of Sephardic Jews settled at one time.

Under the designation 'spinopontine degeneration,' Boller and Segarra (1969) reported 24 persons with late-onset ataxia in 4 generations of an Anglo-Saxon family. Taniguchi and Konigsmark (1971) described 16 affected persons in 3 generations of a black family. The pathologic findings were similar in the 2 families. The cerebellum was relatively spared and the inferior olives were normal. The spinal cord showed loss of myelinated fibers in the spinocerebellar tracts and posterior funiculi. There was also marked loss of nuclei basis ponti. Pogacar et al. (1978) followed up on the Boller-Segarra family (members of which had lived in northern Rhode Island for over 300 years). In 2 clinical cases and 1 autopsy, they questioned the separation from olivopontocerebellar ataxia (SCA1; 164400), because they found abolished tendon reflexes and flexion contractures of the legs in 1 patient, and onset at 18 years of age, palatal myoclonus and optic atrophy in the second. Dementia developed in both. Pathologic findings, in contrast to earlier reports, showed involvement of the cerebellum and inferior olivary nuclei.

Coutinho and Andrade (1978) proposed a 3-way phenotypic classification for MJD: cerebellar ataxia, external ophthalmoplegia and pyramidal signs (type 2), additional predominant extrapyramidal signs (type 1), and additional distal muscular atrophy (type 3). Although not completely specific to MJD, dystonia, facial and lingual fasciculations, and peculiar, bulging eyes represent a constellation strongly suggestive of this disease. Rosenberg (1983) added a fourth phenotype: neuropathy and parkinsonism.

Coutinho et al. (1982) described the presumedly homozygotic son of 2 affected parents; the son had onset at age 8 and died of the disease at age 15. Another son of these parents had onset at age 7. As with other late-onset dominant spinocerebellar degenerations (notably the olivopontocerebellar degenerations), there is considerable phenotypic variation even within the same family. Barbeau et al. (1984) gave an extensive review.

Sequeiros (1985) pointed out that the diagnosis of Machado-Joseph disease had been made (Healton et al., 1980) in an American black family originating from North Carolina; that on further check this proved to be the family reported by Taniguchi and Konigsmark (1971); that Coutinho et al. (1982), in commenting on the neuropathology of Machado-Joseph disease, noted the similarity to the spinopontine atrophy reported by Boller and Segarra (1969), Taniguchi and Konigsmark (1971), and Ishino et al. (1971); and, finally, that the disorder reported in the last family, Japanese, had been proved to be Machado-Joseph disease. See Sequeiros and Suite (1986). Lazzarini et al. (1992) expanded on the pedigree of the family first reported by Boller and Segarra (1969) and concluded that the disorder represented a spinocerebellar ataxia phenotypically similar to that of spinocerebellar ataxia type 1, which shows linkage to HLA. However, linkage to HLA was excluded in this kindred, leading to the designation SCA2 (183090) for this and other HLA-unlinked SCA kindreds. Silveira et al. (1993) demonstrated that the disorder designated Holguin ataxia, or SCA2, that is frequent in Cubans, is genetically distinct from MJD; MJD was excluded from a location on 12q where linkage studies showed the SCA2 locus to be situated.

Eto et al. (1990) described a family of German extraction with progressive ataxia, eye movement abnormalities, peripheral sensory loss, and spinal muscular atrophy of adult onset. The pedigree pattern in 4 generations was consistent with autosomal dominant inheritance. Eto et al. (1990) suggested that the form of spinopontine atrophy might be different from Machado-Joseph disease: the eyes were not protuberant, extraocular movements were abnormal to a minor degree, and neuropathologically the substantia nigra and dentate nucleus were spared. Eto et al. (1990) considered their family to resemble most that reported by Boller and Segarra (1969).

Takiyama et al. (1994) compared the clinical and pathologic features of SCA1 and SCA2 to those in a large Japanese family with Machado-Joseph disease that had previously been linked to markers on chromosome 14q. Although many of the clinical features and the age of onset were similar to those of SCA1 and SCA2, other features were more distinctive for Machado-Joseph disease. These included dystonia, difficulty in opening of the eyelids, slowness of movements, bulging eyes, and facial-lingual fasciculations. One autopsy showed few changes in either the inferior olive or the Purkinje cells, in sharp contrast to SCA1 and SCA2 where such changes are pronounced. The subthalamopallidal system of the MJD patient showed marked degeneration, which has not been described in SCA1 or SCA2.

Seto and Tsujihata (1999) studied a cluster of MJD in a small rural town near Nagasaki City, Japan. They stated that Sakai et al. (1983) described the first family with MJD in Japan, and that Japan had the largest number of reported MJD families in the world. One family studied by Seto and Tsujihata (1999) had 20 affected persons among 73 descending from an ancestor born in 1839. This ancestor had been told that he was a child of unknown non-Japanese parentage (probably Portuguese). The second family had 12 affected persons among 43 with a common ancestor born in 1897. Unsteady gait was the most frequent initial symptom. Age at onset varied from 11 to 51 years with a mean in males of 36.5 and in females of 39.7 years. Anticipation was observed in both families. Three patients had shown only ocular signs: nystagmus, external ophthalmoplegia, and/or blepharoptosis. Bulging eyes were found in only 4 patients. The authors stated that Nagasaki was the only open Japanese port during the Edo period (1635 to 1868).

Livingstone and Sequeiros (1984) noted that 28 families with Machado-Joseph disease had been described in the Azorean Islands, mainly Flores and Sao Miguel, and 3 non-Azorean families in northeast Portugal. Burt et al. (1993) described a dominantly inherited form of ataxia resembling Machado-Joseph disease in members of 4 families of the Arnhem Land Aboriginal people of northern Australia. Portuguese ancestry was possible, although not proven. Goldberg-Stern et al. (1994) reported a family of Machado-Joseph disease in a Yemenite Jewish kindred that originated from a remote village named Ta'izz. This family, incidentally named Yoseph, had no documentation of Portuguese ancestry. Portuguese trade connections with the Yemenites most likely did not reach Ta'izz which is far from the coast and is almost inaccessible because of a wall of high mountains.

Oculomotor Abnormalities

Among 65 patients with SCA1, SCA2, or SCA3, Burk et al. (1996) found reduced saccade velocity in 56%, 100%, and 30% of patients, respectively. MRI showed severe olivopontocerebellar atrophy in SCA2, similar but milder changes in SCA1, and very mild atrophy with sparing of the olives in SCA3. Careful examination of 3 major criteria of eye movements, saccade amplitude, saccade velocity, and presence of gaze-evoked nystagmus, permitted Rivaud-Pechoux et al. (1998) to assign over 90% of patients with SCA1, SCA2, or SCA3 to their genetically confirmed patient group. In SCA1, saccade amplitude was significantly increased, resulting in hypermetria. In SCA2, saccade velocity was markedly decreased. In SCA3, the most characteristic finding was the presence of gaze-evoked nystagmus.

In an investigation of oculomotor function, Buttner et al. (1998) found that all 3 patients with SCA1, all 7 patients with SCA3, and all 5 patients with SCA6 (183086) had gaze-evoked nystagmus. Three of 5 patients with SCA2 did not have gaze-evoked nystagmus, perhaps because they could not generate corrective fast components. Rebound nystagmus occurred in all SCA3 patients, 33% of SCA1 patients, 40% of SCA6 patients, and none of SCA2. Spontaneous downbeat nystagmus only occurred in SCA6. Peak saccade velocity was decreased in 100% of patients with SCA2, 1 patient with SCA1, and no patients with SCA3 or SCA6. Saccade hypermetria was found in all types, but was most common in SCA3. Burk et al. (1999) found that gaze-evoked nystagmus was not associated with SCA2. However, severe saccade slowing was highly characteristic of SCA2. Saccade velocity in SCA3 was normal to mildly reduced. The gain in vestibuloocular reflex was significantly impaired in SCA3 and SCA1. Eye movement disorders of SCA1 overlapped with both SCA2 and SCA3.

The reticulotegmental nucleus of the pons (RTTG), also known as the nucleus of Bechterew, is a precerebellar nucleus important in the premotor oculomotor circuits crucial for the accuracy of horizontal saccades and the generation of horizontal smooth pursuit. By postmortem examination, Rub et al. (2004) identified neuronal loss and astrogliosis in the RTTG in 1 of 2 SCA1 patients, 2 of 4 SCA2 patients, and 4 of 4 SCA3 patients that correlated with clinical findings of hypometric saccades and slowed and saccadic smooth pursuits. The 3 patients without these specific oculomotor findings had intact RTTG regions. The authors concluded that the neurodegeneration associated with SCA1, SCA2, and SCA3 affects premotor networks in addition to motor nuclei in a subset of patients.

Other Features

In 19 of 27 (70%) patients with confirmed SCA types 1, 2, 3, 6, or 7 (164500), van de Warrenburg et al. (2004) found electrophysiologic evidence of peripheral nerve involvement. Eight patients (30%) had findings compatible with a dying-back axonopathy, whereas 11 patients (40%) had findings consistent with a primary neuronopathy involving dorsal root ganglion and/or anterior horn cells; the 2 types were clinically almost indistinguishable. Of 8 patients with SCA3, 5 had a neuronopathy and 4 had a sensorimotor axonopathy.

In a detailed neuropsychologic study, Kawai et al. (2004) found that 16 Japanese MJD patients had verbal and visual memory deficits, impaired verbal fluency, and impaired visuospatial and constructional function compared to controls. In addition, the patients were more depressed and anxious than controls. There was no correlation between cognitive impairment and CAG repeat length. The findings were consistent with widespread dysfunction of the cerebral cortex and/or impairment of the cerebellar cortical circuits.

Yeh et al. (2005) reported autonomic dysfunction among patients with MJD confirmed by genetic analysis. Ten (66%) of 15 patients reported at least 3 diverse autonomic symptoms, most commonly nocturia, cold intolerance, orthostatic dizziness, dry eyes, dry mouth, and impaired near vision. Electrophysiologic studies showed parasympathetic cardiovagal dysfunction in 71% of patients and sympathetic sudomotor dysfunction in 73% of patients.

Franca et al. (2007) found that 33 (47%) of 70 patients with MJD reported chronic pain, most often in the lumbar back and lower limbs.

Franca et al. (2008) observed muscle excitability abnormalities in 41 (82%) of 50 men with MJD, 10 (20%) of whom reported muscle cramps as the presenting complaint. Fifteen patients had fasciculations on clinical exam, and 25 had fasciculations identified on EMG testing. Those with fasciculations had a higher frequency of peripheral neuropathy. Franca et al. (2008) noted that damage to motor axons in classic motor neuron disease leads to collateral nerve sprouting with overexpression of ionic channels that results in spontaneous ectopic activity and muscle cramping. While this mechanism may be at work in some MJD patients, others may have cramps and/or fasciculations due to altered excitatory inputs from damaged corticospinal fibers. Kanai and Kuwabara (2009) commented that they considered muscle cramps in MJD to be primarily a symptom of peripheral motor nerve sprouting and hyperexcitability, particularly in the early stages of the disease.

Clinical Variability

Munchau et al. (1999) described a German woman who presented with severe generalized dystonia beginning at the age of 18 years when she noticed involuntary twisting and cramping of her right hand and twisting of both feet shortly thereafter. Symptoms worsened when she was stressed. At the age of 19 years, she began to grimace when talking and laughing, and her speech became difficult to understand. Over a period of 2 years her symptoms deteriorated, and she became unable to walk without support. She was found to be heterozygous for the ATXN3 gene, with a CAG repeat length of 81 +/- 2 and 14 +/- 1 in the mutated expanded allele and in the normal allele, respectively. Remarkably, cerebellar function was normal apart from mild oculomotor abnormalities. Severe dystonia as a presenting feature had never been described in patients from Germany, where MJD represented 50% of autosomal dominant cerebellar ataxia (ADCA) cases.

In a family of African descent in which 3 members presented with phenotypic features reminiscent of typical Parkinson disease (PD; 168600), Gwinn-Hardy et al. (2001) identified pathogenic expansions in the ATXN3 gene (607047). Features suggestive of PD included bradykinesis, facial masking, rigidity, postural instability, shuffling, asymmetric onset, dopamine responsiveness, and lack of atypical features often associated with SCA3. A fourth, mildly symptomatic patient also carried the repeat expansion. The authors suggested that the low numbers of repeats in this family (67-75; normal, 16-34) presenting with parkinsonism may be associated with ethnic background and that evaluation for SCA3 should be considered in similar cases.

In a study of 412 individuals with MJD, Kieling et al. (2007) found that the estimated mean survival time was 63.96 years, compared to 78.61 years in unaffected relatives. For a subset of 366 patients, mean age at onset was 36.37 years with a survival of 21.18 years. Early onset and increased CAG length predicted shorter overall survival times.

Zeng et al. (2015) reported a Chinese man, born of consanguineous parents, who was homozygous for a pathogenic ATXN3 repeat expansion (71/71) and showed onset of symptoms at age 18 years. He initially developed gait disturbances and slurred speech. Several years later, he had spastic gait, dysphagia, nystagmus, saccade hypermetria, and mild hearing loss. Brain imaging did not show cerebellar or brainstem atrophy. His parents, who were in their mid forties, showed only mild symptoms of the disorder.

Inheritance

Machado-Joseph disease is an autosomal dominant disorder. Sequeiros and Coutinho (1981) identified 9 cases of 'skipped generations' (penetrance = 94.5%).

Some individuals, usually born of consanguineous unions, may be homozygous for a pathogenic ATXN3 allele. These individuals usually show an earlier age at disease onset and more severe symptoms (summary by Zeng et al., 2015).

Diagnosis

Dawson et al. (1982) suggested that the electrooculogram may be useful in early detection.

The finding of 'intermediate alleles' presented a problem in the Portuguese MJD Predictive Testing Program. A second problem was the issue of homoallelism, i.e., homozygosity for 2 normal alleles with exactly the same (CAG)n length, which was found in about 10% of all test results. Maciel et al. (2001) reported a study in which an affected patient carried a 71 and a 51 CAG repeat and 2 asymptomatic relatives carried the 51 CAG repeat and normal-size alleles. The results suggested that the 51 CAG repeat is not associated with disease. The intermediate alleles were not present in a large sample of the healthy population from the same region. Intragenic polymorphisms allowed distinction of the 2 different normal alleles in all cases of homoallelism. An improved protocol for molecular testing for MJD was proposed.

Mapping

In 7 French autosomal dominant SCA families, previously excluded from linkage to the region of chromosome 6 carrying SCA1, Gispert et al. (1993) also excluded linkage to the region of chromosome 12 carrying the SCA2 locus (183090), thus providing evidence for the existence of a third SCA locus, SCA3.

Stevanin et al. (1994) reported linkage studies in 3 of these French families, in 2 of which location of the gene at 14q24.3-qter was possible. Combined analysis of the families placed the SCA3 locus in a 15-cM interval between markers D14S67 and D14S81. Stevanin et al. (1995) narrowed the mapping of SCA3 to a 3-cM interval on 14q. In the third family, Stevanin et al. (1994) excluded linkage to the sites of SCA1, SCA2, and SCA3, thus indicating the existence of a fourth ADCA type I locus.

In Japanese kindreds with MJD, Takiyama et al. (1993) assigned the disease locus to 14q24.3-q32 by genetic linkage to microsatellite loci D14S55 and D14S48; multipoint maximum lod score = 9.719. Using 4 microsatellite DNA polymorphisms (STRPs), Sequeiros et al. (1994) likewise mapped the MJD gene to 14q. Using HOMOG, Sequeiros et al. (1994) could find no evidence for heterogeneity with the 5 Japanese families in whom linkage had been reported. St. George-Hyslop et al. (1994) provided evidence that MJD in 5 pedigrees of Azorean descent was also linked to 14q in an 18-cM region between the markers D14S67 and AACT (107280); multipoint lod score = 7.00 near D14S81. They also reported molecular evidence for homozygosity at the MJD locus in an MJD-affected subject with severe, early-onset symptoms.

Twist et al. (1995) studied 6 MJD families of Portuguese/Azorean origin and 1 of Brazilian origin, using 9 microsatellite markers mapped to 14q24.3-q32.

A fourth SCA locus was suggested by the report of Twells et al. (1994) in which linkage to the regions of chromosomes 6, 12, and 14, where forms of SCA had previously been mapped, was excluded in a large Thai kindred in which dominant cerebellar ataxia was often combined with frontal lobe signs and dementia. Similarly, Lopes-Cendes et al. (1994) excluded linkage with these 3 loci in a large French-Canadian kindred with 4 generations of living affected individuals in 4 generations.

Molecular Genetics

Kawaguchi et al. (1994) identified a common mutation in the MJD gene as the cause of Machado-Joseph disease. In normal individuals, the gene was found to contain between 13 and 36 CAG repeats, whereas most of the patients with clinically diagnosed MJD and all of the affected members of a family with the clinical and pathologic diagnosis of MJD showed expansion of the repeat number to the range of 68 to 79 (607047.0001). Schols et al. (1995) provided definitive proof that mutation in the ATXN3 gene cause SCA3.

Giunti et al. (1995) surveyed members of 63 families with a variety of autosomal dominant late-onset cerebellar ataxias for the CAG repeat expansion described in association with Machado-Joseph disease. The MJD mutation was identified in 9 families segregating progressive adult-onset cerebellar degeneration with variable supranuclear ophthalmoplegia, optic atrophy, mild dementia, peripheral neuropathy, or extrapyramidal dysfunction, corresponding to Harding's classification of ADCA type I (Harding, 1982). Most of the patients with ADCA type I have olivopontocerebellar atrophy at autopsy. Giunti et al. (1995) noted that this mutation was also identified in a further family affected with parkinsonism, peripheral neuropathy and dystonia but little cerebellar disease. The origins of these 10 families were the United Kingdom, India, Pakistan, the West Indies, France, Brazil, and Ghana. The authors could find no clinical feature that distinguished ADCA type I patients with the SCA3 mutation from those who did not have it. Giunti et al. (1995) found that the CAG repeat length ranged from 13 to 41 copies on normal chromosomes and 62 to 80 copies on affected chromosomes. The families in which Giunti et al. (1995) detected the Machado-Joseph disease trinucleotide repeat expansion included the historic 'Drew family of Walworth' (Harding, 1982).

Since some clinical features of MJD overlap with those of SCA, Schols et al. (1995) sought MJD mutations in 38 German families with autosomal dominant SCA. The MJD (CAG)n trinucleotide expansion was identified in 19 families. In contrast, the trinucleotide expansion was not observed in 21 ataxia patients without a family history of the disease. Analysis of the (CAG)n repeat length in 30 patients revealed an inverse correlation with the age of onset. The (CAG)n stretch of the affected allele varied between 67 and 78 trinucleotide units; the normal alleles carried between 12 and 28 simple repeats. These results demonstrated that the MJD mutation causes the disease phenotype of most SCA patients in Germany. Schols et al. (1995) pointed out that in SCA3 as observed in Germany, features characteristic of Machado-Joseph disease, such as dystonia, bulging eyes, and faciolingual fasciculations, are rare.

Durr et al. (1996) screened 173 index patients with adult-onset cerebellar ataxia of whom 125 were classified as ADCA type I (cerebellar signs with supranuclear ophthalmoplegia, extrapyramidal signs, dementia, and amyotrophy); 9 of whom were ADCA type II (cerebellar ataxia with retinal degeneration in all family members); and 4 were ADCA type III (pure cerebellar signs after a disease duration of more than 10 years). The SCA3-MJD mutation represented 28% of all their ADCA type I families, whereas SCA1 only accounted for 13% in their population. The number of CAG repeats in the expanded allele ranged from 64 to 82 with a median of 73. In contrast, normal alleles contained between 14 and 40 CAG repeats. The mean expansion between generations was +0.86 CAG repeat units without a statistically significant difference between paternally and maternally transmitted alleles. Durr et al. (1996) found no correlation between the CAG repeat length and the tendency to expansion. All SCA3 patients had cerebellar ataxia; 46% had extensor plantar responses; 55% had decreased vibratory sensation; and supranuclear ophthalmoplegia was present in 47% of the patients. Dystonia and parkinsonian signs were only found in 18% of the patients. Two of 49 patients had retinal degeneration; 60% of patients had axonal neuropathy. Bulging eyes were noticed in 23% of SCA3 patients, which was similar to the frequency observed in SCA1 patients.

Lopes-Cendes et al. (1997) reported 25 unrelated Brazilian families with MJD. Molecular analysis showed that normal alleles ranged from 12 to 33 CAG repeats, whereas expanded pathogenic alleles ranged from 66 to 78 CAG repeats. There was a significant negative correlation between age at onset and length of CAG tract. However, repeat contractions were also detected, and Lopes-Cendes et al. (1997) estimated that only 40% of the variation in age at disease onset could be attributed to length of the expanded repeat.

Ramesar et al. (1997) investigated 14 South African kindreds and 22 sporadic individuals with SCA for expanded SCA1 (601556.0001) and MJD repeats. The authors stated that SCA1 mutations accounted for 43% of known ataxia families in the Western Cape region of South Africa. They found that expanded SCA1 and CAG repeats cosegregated with the disorder in 6 of the families, 5 of mixed ancestry and 1 Caucasian, and were also observed in a sporadic case from the indigenous Black African population. The use of the microsatellite markers D6S260, D6S89, and D6S274 provided evidence that the expanded SCA1 repeats segregated with 3 distinct haplotypes in the 6 families. None of the families nor the sporadic individuals showed expansion of the MJD repeat.

Studying 77 German families with autosomal dominant cerebellar ataxia of SCA types 1, 2, 3, and 6 (183086), Schols et al. (1997) found that the SCA1 mutation accounted for 9%, SCA2 for 10%, SCA3 for 42%, and SCA6 for 22%. There was no family history of ataxia in 7 of 27 SCA6 patients. Age at onset correlated inversely with repeat length in all subtypes. Yet the average effect of 1 CAG unit on age of onset was different for each SCA subtype. Schols et al. (1997) compared clinical, electrophysiologic, and magnetic resonance imaging (MRI) findings to identify phenotypic characteristics of genetically defined SCA subtypes. Slow saccades, hyporeflexia, myoclonus, and action tremor suggested SCA2. SCA3 patients frequently developed diplopia, severe spasticity or pronounced peripheral neuropathy, and impaired temperature discrimination, apart from ataxia. SCA6 presented with a predominantly cerebellar syndrome, and patients often had onset after 55 years of age. SCA1 was characterized by markedly prolonged peripheral and central motor conduction times in motor evoked potentials. MRI scans showed pontine and cerebellar atrophy in SCA1 and SCA2. In SCA3, enlargement of the fourth ventricle was the main sequel of atrophy. SCA6 presented with pure cerebellar atrophy on MRI. Overlap between the 4 SCA subtypes was broad, however.

Genotype/Phenotype Correlations

Kawaguchi et al. (1994) found a negative correlation between age of onset and CAG repeat numbers in MJD. Southern blot analyses and genomic cloning demonstrated the existence of related genes and raised the possibility that similar abnormalities in related genes may give rise to diseases similar to MJD.

Maruyama et al. (1995) examined the molecular features of the CAG repeats and the clinical manifestations in 90 MJD individuals from 62 independent Japanese MJD families and found that the MJD repeat length was inversely correlated with the age of onset (r = -0.87). The MJD chromosomes contained 61-84 repeat units, whereas normal chromosomes displayed 14-34 repeats. In the normal chromosomes, 14 repeat units were the most common and the shortest.

Takiyama et al. (1995) examined the size of the (CAG)n repeat array in the 3-prime end of the ATXN3 gene and the haplotype at a series of microsatellite markers surrounding the ATXN3 gene in a large cohort of Japanese and Caucasian subjects with MJD. Expansion of the array from the normal range of 14-37 repeats to 68-84 repeats was found, with no instances of expansions intermediate in size between those of the normal and MJD affected groups. The expanded allele associated with MJD displayed intergenerational instability, particularly in male meiosis, and this instability was associated with the clinical phenomenon of anticipation. The size of the expanded allele was not only inversely correlated with the age-of-onset of MJD, but was also correlated with the frequency of other clinical features, such as pseudoexophthalmos and pyramidal signs were more frequent in subjects with larger repeats. The disease phenotype was significantly more severe and had an early age of onset (16 years) in a subject homozygous for the expanded allele, which contrasts with Huntington disease (HD; 143100), in which the homozygous subject has a disorder indistinguishable from that in the heterozygous subject. The observation in MJD suggests that the expanded allele may exert its effect either by a dominant-negative effect (putatively excluded in HD) or by a gain-of-function effect as proposed for HD. Japanese and Caucasian subjects affected with MJD shared haplotypes at several markers surrounding the ATXN3 gene, these markers being uncommon in the normal Japanese and Caucasian populations, thus suggesting the existence either of common founders in these populations or of chromosomes susceptible to pathologic expansion of the CAG repeat in the ATXN3 gene.

Ranum et al. (1995) made use of the fact that the genes involved in 2 forms of autosomal dominant ataxia, that for MJD and that for SCA1, have been isolated to assess the frequency of trinucleotide repeat expansions among individuals diagnosed with ataxia. They collected and analyzed DNA from individuals with both disorders. In both cases, the genes responsible for the disorder were found to have an expansion of an unstable CAG trinucleotide repeat. These individuals represented 311 families with adult-onset ataxia of unknown etiology, of which 149 families had dominantly inherited ataxia. Ranum et al. (1995) found that of these, 3% had SCA1 trinucleotide repeat expansions, whereas 21% were positive for the MJD trinucleotide expansion. For the 57 patients with MJD trinucleotide repeat expansions, strong inverse correlation between CAG repeat size and age at onset was observed (r = -0.838). Among the MJD patients, the normal and affected ranges of CAG repeat size were 14 to 40 and 68 to 82 repeats, respectively. For SCA1, the normal and affected ranges were much closer, namely 19 to 38 and 40 to 81 CAG repeats, respectively.

Cancel et al. (1995) documented the marked phenotypic heterogeneity associated with expansion of the CAG repeat sequence at the SCA3/MJD locus. They studied 3 French families with type I autosomal dominant cerebellar ataxia and a French family with neuropathologic findings suggesting the ataxochoreic form of dentatorubropallidoluysian atrophy (DRPLA; 125370). A strong correlation was found between size of the expanded CAG repeat and age at onset of clinical disease. Instability of the expanded triplet repeat was not found to be affected by sex of the parent transmitting the mutation. Both somatic and gonadal mosaicism for alleles carrying expanded trinucleotide repeats was found. The 4 French families had no known Portuguese ancestry. Faciolingual myokymia, said to be a hallmark of MJD, increased tendon reflexes, ophthalmoplegia, and dystonia occur significantly more frequently among Azorean MJD patients, while decreased vibratory sense and dementia were found more often among the French cerebellar ataxia type I patients. Myoclonus, present in 1 of the 5 patients in the French family with the DRPLA-like disorder, had never been reported in SCA3 or MJD kindreds.

Igarashi et al. (1996) investigated the association of intergenerational instability of the expanded CAG repeat in MJD with a CAG/CAA polymorphism in the CAG repeat and a CGG/GGG polymorphism at the 3-prime end of the CAG array. Their results strongly suggested that an interallelic interaction is involved in the intergenerational instability of the expanded CAG repeat. Igarashi et al. (1996) reported that normal chromosomes with the CGG allele are more frequently associated with larger CAG repeats than normal chromosomes with the GGG allele. They also reported that 80 of 88 independent MJD chromosomes had the CGG allele, which is in striking contrast to the CGG allele frequency in the normal chromosome. Igarashi et al. (1996) investigated the effect of gender on the intergenerational instability of the expanded CAG repeat. They obtained significant evidence that the expanded CAG repeats were less stable in paternal transmission than in maternal transmission.

Size of the expanded repeat and gene dosage are factors in the severity and early onset of MJD. Another factor pointed out by Kawakami et al. (1995) is gender. In a total of 14 sib pairs, the mean of the differences in age of onset between the sibs of different sexes was 12.7 +/-1.7 (n = 7) and between the sibs of the same sex was 3.9 +/-1.7 (n = 7). The difference was statistically significant, whereas the variance in length of CAG repeats between these 2 groups was not significant.

Van Alfen et al. (2001) reported a Dutch family in which 4 members in 2 generations had intermediate repeat lengths (53 and 54) in the ATXN3 gene. All but the youngest had a restless legs syndrome with fasciculations and a sensorimotor axonal polyneuropathy. The authors concluded that intermediate repeat lengths can be pathogenic and may predispose for restless legs and peripheral nerve disorder.

Van de Warrenburg et al. (2005) applied statistical analysis to examine the relationship between age at onset and number of expanded triplet repeats from a Dutch-French cohort of 802 patients with SCA1 (138 patients), SCA2 (166 patients), SCA3 (342 patients), SCA6 (53 patients), and SCA7 (103 patients). The size of the expanded repeat explained 66 to 75% of the variance in age at onset for SCA1, SCA2, and SCA7, but less than 50% for SCA3 and SCA6. The relation between age at onset and CAG repeat was similar for all groups except for SCA2, suggesting that the polyglutamine repeat in the ataxin-2 protein exerts its pathologic effect in a different way. A contribution of the nonexpanded allele to age at onset was observed for only SCA1 and SCA6. Van de Warrenburg et al. (2005) acknowledged that their results were purely mathematical, but suggested that they reflected biologic variations among the diseases.

Padiath et al. (2005) reported a 3-generation Indian pedigree in which the proband had 45 CAG repeats in the ATXN3 gene. The proband had clinical features of spinocerebellar ataxia as well as signs of cerebellar and brainstem atrophy. The 45-repeat allele was unstable on intergenerational transmission and was associated with a haplotype found in the majority of MJD/SCA3 patients worldwide. Padiath et al. (2005) noted that this was the smallest unstable allele in the ATXN3 gene reported to that time.

Allelic Transmission

Maruyama et al. (1995) analyzed parent-child transmission in association with the clinical anticipation of the disease and showed the unidirectional expansion of CAG repeats with no case of diminution in the affected family. The differences in CAG repeat length between parent and child and between sibs were greater in paternal transmission than in maternal transmission. Detailed analysis showed that a large degree of expansion was associated with a shorter length of the ATXN3 gene in paternal transmission. On the other hand, the increments of increase were similar for shorter and longer expansions in maternal transmission. Among the 3 clinical subtypes, type 1 MJD with dystonia showed a larger degree of expansion in CAG repeats of the gene and younger ages of onset than the other types.

Ikeuchi et al. (1996) analyzed segregation patterns in 80 transmissions in 7 MJD pedigrees and in 211 transmissions in 24 DRPLA pedigrees with the diagnoses confirmed by molecular testing. The significant distortions in favor of transmission of the mutant alleles were found in male meiosis, where the mutant alleles were transmitted to 73% of all offspring in MJD (P less than 0.01) and to 62% of all offspring in DRPLA (P less than 0.01). The results were consistent with meiotic drive in these 2 disorders. The authors commented that, since more prominent meiotic instability of the length of the CAG trinucleotide repeats is observed in male meiosis than in female meiosis and meiotic drive is observed only in male meiosis, these results raised the possibility that a common molecular mechanism underlies the meiotic drive and the meiotic instability in male meiosis.

Rubinsztein and Leggo (1997) investigated the transmission of alleles with larger versus smaller CAG repeat numbers in the ATXN3 gene in normal heterozygotes from the 40 CEPH families. Their data suggested that there was no segregation distortion in male meioses, while the smaller CAG allele was inherited in 57% of female meioses (p less than 0.016). The pattern of inheritance of smaller versus larger CAG alleles at this locus was significantly different when male and female meioses were compared. While previous data suggested that meiotic drive may be a feature of certain human diseases, including the trinucleotide disease MJD, myotonic dystrophy, and DRPLA, the data of Rubinsztein and Leggo (1997) were compatible with meiotic drive also occurring among non-disease-associated CAG sizes.

In German patients with SCA3, Riess et al. (1997) likewise found transmission distortion of the mutant alleles, but the segregation distortion was observed during maternal transmission in German families, rather than in paternal inheritance, as observed in Japanese pedigrees.

Grewal et al. (1999) performed a sperm typing study of 5 MJD patients of French descent. Analysis of the pooled data showed a ratio of mutant to normal alleles of 379:436 (46.5%:53.5%). To confirm these results, sperm typing analysis was also performed using a polymorphic marker, D14S1050, closely linked to the ATXN3 gene. Among 910 sperm analyzed, the allele linked to the disease chromosome was detected in 50.3% of the samples, and the allele linked to the normal chromosome was found in 49.6% of the sperm. The difference in frequency of these 2 alleles was not significant.

In an analysis of 428 meioses among 102 healthy Portuguese sibships, Bettencourt et al. (2008) observed preferential transmission of the smaller ATXN3 wildtype allele. There were no mutational events. There was a positive correlation between the difference in length between the 2 ATXN3 alleles of the transmitter's genotype and the frequency of transmission of the smaller alleles. The authors concluded that the genotypic composition of the transmitters in a sample should be taken into account in studies of segregation ratio distortion.

In a large population-based study of 82 MJD families from Rio Grande do Sul, Brazil, Prestes et al. (2008) found that fitness among affected individuals was increased compared to the general population and compared to unaffected family members. Affected individuals had significantly more children than unaffected relatives, with no sign of parental gender effect. In addition, affected individuals had a lower age at first delivery and earlier onset of menopause compared to unaffected relatives; however, affected women who did not have children had larger CAG tracts than those who had children. Prestes et al. (2008) noted that since disease onset usually occurs after reproductive age, most affected individuals have children before knowing their genetic status. The findings overall suggested enhanced fitness of the mutant allele.

Pathogenesis

Ikeda et al. (1996) demonstrated the induction of apoptosis in cultured cells expressing a portion of the ATXN3 gene that included the expanded CAG repeats. Cell death occurred only when the CAG repeat was translated into polyglutamine residues, which apparently precipitated in large covalently modified forms. Sisodia (1998) reviewed the significance of nuclear inclusions in glutamine repeat disorders.

Studying the link between intranuclear expression of expanded polyglutamine and neuronal dysfunction, Perez et al. (1999) demonstrated that ataxin-3 adopts a unique conformation when expressed within the nucleus of transfected cells. They found that this novel conformation of intranuclear ataxin-3 is not due to proteolysis, suggesting instead that association with nuclear protein(s) alters the structure of full-length ataxin-3, exposing the polyglutamine domain. This conformationally altered ataxin-3 was bound to the nuclear matrix. The pathologic form of ataxin-3 with an expanded polyglutamine domain also associates with the nuclear matrix. These data suggested that an early event in the pathogenesis of SCA3/MJD may be an altered conformation of ataxin-3 within the nucleus that exposes the polyglutamine domain.

Chai et al. (1999) presented 2 lines of evidence implicating the ubiquitin-proteasome pathway in the pathogenesis of SCA3/MJD. First, studies of both human disease tissue and in vitro models showed redistribution of the 26S proteasome complex into polyglutamine aggregates. In neurons from SCA3/MJD brain, the proteasome localized to intranuclear inclusions containing the mutant protein ataxin-3. In transfected cells, the proteasome redistributed into inclusions formed by 3 expanded polyglutamine proteins: a pathologic ataxin-3 fragment, full-length mutant ataxin-3, and an unrelated GFP-polyglutamine fusion protein. Inclusion formation by the full-length mutant ataxin-3 required nuclear localization of the protein and occurred within specific subnuclear structures recently implicated in the regulation of cell death. In a second set of experiments, inhibitors of the proteasome caused a repeat length-dependent increase in aggregate formation, implying that the proteasome plays a direct role in suppressing polyglutamine aggregation in disease. These results supported a central role for protein misfolding in the pathogenesis of SCA3/MJD and suggested that modulating proteasome activity is a potential approach to altering the progression of this and other polyglutamine diseases.

Evert et al. (1999) generated ataxin-3-expressing rat mesencephalic CSM14.1 cells to study the effects of long-term expression of ataxin-3. The isolated stable cell lines provided high level expression of human full-length ataxin-3 with either the normal nonexpanded CAG repeats (SCA3-Q23) or the pathogenic expanded CAG repeats (SCA3-Q70). When cultured at a nonpermissive temperature (39 degrees C), CSM14.1 cells expressing the expanded full-length ataxin-3 developed nuclear inclusion bodies, strong indentations of the nuclear envelope, and cytoplasmic vacuolation, whereas cells expressing the nonexpanded form and control cells did not. The ultrastructural alterations resembled those found in affected neurons of SCA3 patients. Cells with such changes exhibited increased spontaneous nonapoptotic cell death.

Gaspar et al. (2000) explored the possibility that frameshift