CAG Repeats in SCA6

Molecular genetics has had a major impact on neurology. The genes for many neurodevelopmental and neurodegenerative disorders have been identified, and molecular diagnosis has replaced invasive and cumbersome diagnostic procedures. The hereditary ataxias may have benefited most from molecular studies. Genetic mapping and gene identification have classified these heterogeneous disorders, offered patients accurate diagnoses, and made genetic counseling feasible. In addition to these clinical benefits, exciting biological discoveries have been made possible by the molecular study of many neurologic diseases. For example, gene dosage of the peripheral myelin protein PMP22 is critical for peripheral nerve integrity and normal function¹; dynamic trinucleotide repeat mutations are the cause of numerous neurologic diseases²; and genes essential for normal development in Drosophila play critical roles in mammalian nervous system development.³ Among these discoveries, the expansion of trinucleotide repeats as a mutational mechanism has been one of the most intriguing breakthroughs. These unstable repeats have been found in both coding and noncoding regions of genes. The mechanisms by which expanded repeats cause disease vary depending on the position of the repeat within the respective gene. In fragile X syndromes(FMR1 and FMR2), the repeat interferes with normal transcription and translation of the gene, leading to loss of function of the protein.² In myotonic dystrophy, expansion of a CTG repeat in the 3′ untranslated region of the gene leads to disease most likely because of a dominant effect at the level of RNA.⁴ In Friedreich ataxia, the expansion of the GAA repeat in the first intron of the gene interferes with the transcription and translation of frataxin.² A number of neurodegenerative diseases, notably the dominantly inherited spinocerebellar ataxias (SCAs), share a common mutational mechanism: the modest expansion of a translated CAG repeat predicted to encode a polyglutamine tract.^5,6 The pathogenetic mechanism in this group of diseases is believed to involve a gain of function, toxic to the cell, due to the expanded polyglutamine tract. The repeat sizes typically range from 6 to 34 on normal alleles and 35 to 135 on expanded disease alleles.^5,6 Although the size of the smallest disease-causing allele may vary depending on the gene, for seven well-characterized disorders (spinobulbar muscular atrophy, Huntington's disease, SCA1, SCA2, SCA3/MJD, SCA7, and dentatorubral-pallidoluysian atrophy), alleles with 30 or less repeats are within the normal range.^2,5,6 This observation remained true for 5 years until the discovery that very small expansions (21 to 27 CAG repeats) within the α_1A-voltage-dependent calcium channel (CACNL1A4) gene lead to cerebellar ataxia and neuronal degeneration in SCA6.⁷ To validate that these small expansions are indeed pathogenic and are not polymorphisms closely linked to the disease-causing mutation, Zhuchenko and colleagues analyzed more than 900 normal alleles and determined that the expanded alleles were observed only in patients with SCA6 and that they segregated with the disease in each of the kindreds evaluated.⁷ In this issue of Neurology, three new studies confirm that small CAG expansions within CACNL1A4 are disease-causing in ataxic patients from diverse ethnic backgrounds. The new data provide a large sample for genotype-phenotype correlation analysis and confirm that the pathogenicity of an expanded CAG tract depends on the protein context as well as the repeat size.^8-10

Prevalence of the SCA6 mutation and genotype-phenotype correlations. To date, many patients with ataxia have been genotyped for the CAG repeat expansion within the CACNL1A4 gene. The data clearly demonstrate that the prevalence of this mutation varies depending on the ethnicity of the population studied. Matsumura et al. found that SCA6 accounts for 31% of the dominantly inherited ataxias in Japan, second only to SCA3, which accounts for 39%.⁸ Within Japan, the prevalence rates vary depending on the region studied. In the Chugoku area of western Japan, SCA6 accounts for 30% of dominant ataxias whereas SCA1, SCA2, SCA3, and DRPLA each account for less than 5% of this group of diseases, suggesting that a founder effect may underlie the high prevalence rate of SCA6.¹¹ Studies of ethnically heterogenous populations in the United demonstrate that SCA6 accounts for 5% to 10% of dominantly inherited ataxia.^10,12 In Europe, the prevalence rates vary considerably; it is estimated that SCA6 accounts for approximately 10% of SCAs in Germany,¹³ whereas it is rare (1 to 2%) cause of inherited in France.⁹

Based on data from six published studies on SCA6, the CAG repeat number varies from 4 to 18 repeats on normal alleles (more than 1700 alleles analyzed) and 21 to 30 on SCA6 chromosomes (more than 150 alleles analyzed).^7-11,13 The clinical data gathered on SCA6 patients in the three studies published in this issue are valuable and contribute to understanding the harmful pathogenic effects of this mutation. The first important observation is the strong correlation between the smaller size of the repeat and the later age of onset of disease. Combined data from the three studies demonstrate that 65% to 75% of the variation in the age of onset can be accounted for by the number of CAG repeat units.^8-10 This is a striking finding given the very small variability in repeat size (21 to 29) and suggests that very minimal changes in repeat number have a dramatic effect on the age of onset.

The phenotype of patients with SCA6 is characterized predominantly by a gait and limb ataxia and dysarthia. Many patients have horizontal gaze-evoked nystagmus, and some have limitation of eye studies reveal cerebellar atrophy with relative sparing of the brain stem.^8-10 Although these findings tend to distinguish SCA6 from SCA1, and 3 because SCA6 is a predominantly pure cerebellar syndrome, there is sufficient inter- and intrafamilial variability that makes it difficult to rely on clinical findings for an accurate diagnosis. Furthermore, there are patients with molecularly proven SCA6 who have overlapping clinical features with the other ataxias,^7,10 and there are neuropathologic data demonstrating degeneration of inferior olive neurons in this disorder.¹⁴

Three interesting clinical observations are noted in the published data. First, some patients experience episodic ataxia early in the course of the disease, raising the possibility that episodic ataxia and SCA6 may represent a phenotypic continuum of the same disease.^7,10 Second, patients homozygous or compound heterozygous for the expanded alleles have either a similar phenotype to patients heterozygous for the mutation¹¹ or a more severe phenotype characterized by earlier age of onset and more severe cerebellar atrophy.^8,10 These findings are important when trying to understand the pathogenesis of SCA6 in relation to the normal function of CANCL1A4 (see following section). Lastly, the SCA6 mutation has been identified in patients who do not have a positive family history of ataxia.^10,13 Although these patients may represent true sporadic cases, the most likely explanation is that the parents of these patients have not yet manifested the symptoms of SCA6, given that symptoms may be mild, episodic, or have their onset in the sixth or seventh decade. Nevertheless, this suggests that SCA6 testing should be considered in patients with a negative family history for SCA and apparently sporadic disease.

The expanded CAG repeat within CACNL1A4 occurring in SCA6 differs from other disease-causing CAG repeats because the expansion is small and the repeat is relatively stable. This repeat must undergo expansions, hence the finding of alleles with 20 or more repeats in patients with ataxia. However, unlike other unstable repeats, intergenerational instability rarely is observed. To date, more than 50 parent-child transmission have been evaluated^7-11,13 and only a single expansion event (24 expanded to 26) has been observed in a father-son pair.¹¹ Despite the relative stability of the repeat on intergenerational transmission, clinical anticipation is observed in SCA6.^9,11 These data suggest that there are factors, in addition to the CAG repeat size, that contribute to the clinical anticipation.

CACNL1A4 mutations in neurologic disease: Insight into pathogenesis of neurodegeneration? Voltage-sensitive calcium channels are multimeric complexes composed of an α₁ subunit, which is sufficient to form the structural channel and confer voltage dependence, and α2,σ, and β subunits that play regulatory roles.¹⁵ Topologically, the α₁ subunit is characterized by four repeated domains (I to IV), spanning segments (1 to 6) and one pore-forming (P) segment between segments 5 and 6 (figure). There are at least six genes (A,B,C,D,E, and S) in the α₁ subunit family, each with unique pharmacologic and voltage-dependent activation and inactivation properties. The α_1A subunit (CACNL1A4) encodes P- and Q-type calcium channels that were identified in cerebellar Purkinje cells¹⁶ and granule neurons,¹⁷ respectively. The P- and Q-type channels can be distinguished by their pharmacologic properties and inactivation kinetics, which in turn may be determined by the α_1A subunit splice variants.^7,18 The α_1A subunit gene is expressed extensively in the central nervous system, with the highest expression in the cerebellum.¹⁹

• Download figure
• Open in new tab
• Download powerpoint

Figure. Topology of CACNL1A4 with approximate positions of the mutations that cause familial hemiplegic migraine (FHM), episodic ataxia type 2 (EA-2), the tottering and tottering leaner mutations in mice, and spinocerebellar ataxia type 6 (SCA6). P is pore-forming segment and is present between segments 5 and 6 in each of the four transmembrane domains.

Mutations in the α_1A subunit gene have been identified in familial hemiplegic migraine (FHM) and episodic ataxia type 2 (EA-2) in humans,²⁰ and the tottering and leaner mice.¹⁹ Familial hemiplegic migraine is an autosomal dominant disorder characterized by migraine, ictal hemisparesis, and in some families, progressive cerebellar atrophy.²⁰ Episodic ataxia type 2 is an autosomal dominant paroxysmal disorder characterized by acetazolamide-responsive ataxia that often is precipitated by stress. Affected individuals have interictal nystagmus, and some have progressive cerebellar degeneration.²⁰ The tottering mutation (tg) is an autosomal recessive mutation that leads to spike and wave discharges, intermittent convulsions, ataxia, and absence-like seizures in mice.¹⁹ The tottering leaner (tg^la) allele is also a recessive neurologic disorder in mice, characterized by chronic ataxia and Purkinje and granule neuron degeneration.

Four different missense mutations have been identified in FHM, and two of these are predicted to affect the intracellular mouth of the ion pore, given their location within segment six transmembrane alpha helices (seefigure). Similarly, the mutation in tg is a missense mutation very close to the conserved pore-lining segment in the second transmembrane domain. The mutations in EA-2 and tg^la are in splice sites in the 3′ end of the gene, and both are likely to produce truncated proteins that are not functional. The CAG repeat in CACNL14A is translated in splice variants the extend the coding region by 239 amino acids and putatively place the repeat within the intracellular carboxyl terminal portion of the protein (seefigure).

Given the molecular data on FHM, EA-2, SCA6, the tottering, and tottering leaner mice, as well as the clinical data on all these disorders, what can we conclude about the pathogenesis of these various neurologic syndromes? For EA-2 and tottering leaner mice, the loss of functional voltage-sensitive calcium channel may explain the phenotype. Haploinsufficiency causes the episodic features and mild cerebellar atrophy in EA-2, whereas loss of both alleles in the mice causes the chronic ataxia and neuronal degeneration in the cerebellum. Mice heterozygous for the tg^la allele might have unrecognized episodic ataxia. The exact mechanisms by which the missense mutations in FHM and tg mice lead to disease are unknown but given the promixity of the mutations to the pore, ion sensitivity or permeability may be compromised, leading to abnormal neurotransmitter release. Geschwind et al. provided clinical data that demonstrate that EA-2 and SCA6 overlap clinically. This finding, together with the observation that a patient who is a compound heterozygous for the expansion has earlier onset and more severe form of disease, led the authors to propose that CAG repeat expansions most likely lead to loss of normal function of the protein.¹⁰ Similarly, Matsumura et al. noted that a patient homozygous for an expanded allele has a more severe course than her sister who is heterozygous for the same expanded allele.⁸ Although these findings generally support a loss of function rather than a gain of function-pathogenetic mechanism, the data available to date do not totally support this model. Patients homozygous for an expanded allele within the SCA3/MJD gene have a worse phenotype than heterozygous patients.²¹ In SCA1 transgenic mice that express a mutant form of ataxin-1 with 82 repeats in Purkinje cells, ataxia and Purkinje cell degeneration develop at an earlier age when the mice are bred to homozygosity, suggesting that gene dosage does affect the course of the disease.²² An important aspect of SCA6 pathogenesis is the contribution of the repeat size to the age of onset. Although similar to other CAG repeat disorders, it is much more pronounced. This suggests similar pathogenetic mechanisms and confirms earlier observations that the context of the repeat is crucial in determining the pathogenetic effect of particular repeat tract.⁵ If there were no knowledge about the function of CACNL1A4, the SCA6 gene product would have been named ataxin-6, and SCA6 would have been put in the same category as the other polyglutamine disorders. However, the knowledge of the function of the SCA6 gene product and its expression pattern is tantalizing and raises interesting questions about alternative mechanism(s) of pathogenesis. CACNL1A4 may lose some functional activity because of the expansion, just as the androgen receptor partially loses normal activity in the presence of longer repeats. However, as for the androgen receptor, the neurodegeneration may be mediated by a novel or toxic function secondary to the polyglutamine expansion.

The availability of the tg and tg^la mice, as well as the feasibility of generating transgenic mice expressing mutant FHM, EA-2, or SCA6 alleles, will permit physiologic and pharmacologic studies to address the functional consequences of each mutation. Such studies may give new insight about the role of calcium homeostasis in neurodegeneration and eventually could lead to rational therapeutic options.

Acknowledgments

The author thanks Dr. Cheng Chi Lee for an exciting collaboration on SCA6, and Dr. Xi Lin for preparing the figure. Huda Y. Zoghbi is an investigator with Howard Hughes Medical Institute.