Complexity of the Genetics and Clinical Presentation of Spinocerebellar Ataxia 17

Abstract

Spinocerebellar ataxia type 17 (SCA17) is a rare autosomal dominant neurodegenerative disease caused by a CAG repeat expansion in the TATA-box binding protein gene (TBP). The disease has a varied age at onset and clinical presentation. It is distinct from other SCAs for its association with dementia, psychiatric symptoms, and some patients presenting with chorea. For this reason, it is also called Huntington's disease-like 4 (HDL-4). Here we examine the distribution of SCA17 allele repeat sizes in a United Kingdom-based cohort with ataxia and find that fully penetrant pathogenic alleles are very rare (5 in 1,316 chromosomes; 0.38%). Phenotype-genotype correlation was performed on 30 individuals and the repeat structure of their TBP genes was examined. We found a negative linear correlation between total CAG repeat length and age at disease onset and, unlike SCA1, there was no correlation between the longest contiguous CAG tract and age at disease onset. We were unable to identify any particular phenotypic trait that segregated with particular CAG/CAA repeat tract structures or repeat lengths. One individual within the cohort was homozygous for variable penetrance range SCA17 alleles. This patient had a similar age at onset to heterozygotes with the same repeat sizes, but also presented with a rapidly progressive dementia. A pair of monozygotic twins within the cohort presented 3 years apart with the sibling with the earlier onset having a more severe phenotype with dementia and chorea in addition to the ataxia observed in their twin. This appears to be a case of variable expressivity, possibly influenced by other environmental or epigenetic factors. Finally, there was an asymptomatic father with a severely affected child with an age at onset in their twenties. Despite this, they share the same expanded allele repeat sizes and sequences, which would suggest that there is marked difference in the penetrance of this 51-repeat allele. We therefore propose that the variable penetrance range extend from 48 repeats to incorporate this allele. This study shows that there is variability in the presentation and penetrance of the SCA17 phenotype and highlights the complexity of this disorder.

Keywords: CAG repeat expansions; PolyQ; SCA17; ataxia; genetic counseling; neurodegeneration.

FIGURE 1

CAG/CAA repeat configuration in the TBP gene. SCA17 is caused by CAG repeat expansion in the TATA-box binding protein gene (TBP). The repeat tract is characterized by CAA interruptions and has a typical configuration as shown, and can be segmented into Domain I (red), Domain II (brown), Domain III (green), Domain IV (blue), and Domain V (purple) (Gostout et al., 1993; Gao et al., 2008). Domains II and IV are polymorphic CAG repeats.

FIGURE 2

Frequency distribution of SCA17 allele sizes from diagnostic fragment analysis performed on a United Kingdom cohort at the Neurogenetics Unit, National Hospital for Neurology and Neurosurgery, London. SCA17 alleles from 1,316 discrete chromosomes were analyzed by fragment analysis. The main graph shows the frequency of all alleles tested whilst the inset depicts an enlarged view of the alleles in the variable and fully penetrant range (41–57 repeats, n = 39). Normal alleles (24–40 repeats, n = 1277) are plotted in blue, whilst variable and fully penetrant alleles are plotted in brown. The SCA17 alleles ranged from 24 to 57 repeats, with a mean allele size of 37 ± 2 repeats (mean ± standard deviation). The most frequent allele had 38 repeats (n = 354), whilst the most frequent variable penetrance allele had 41 repeats (n = 14). Only five alleles were within the fully penetrant pathogenic range (≥49 repeats).

FIGURE 3

Correlation between SCA17 pathogenic allele size and age at disease onset. Age at disease onset data was available for 21 SCA17 patients that had been both analyzed by fragment sizing and clone sequencing. Pathogenic allele size as determined by fragment sizing shows a negative correlation with respect to age at onset and a good fit to the linear model (R² = 0.595) (A). Cloning and sequencing the SCA17 alleles permits the calculation of the mean pathogenic allele size, rounded to the nearest whole repeat, based on the total length of the CAG/CAA repeat tract sequence. Mean pathogenic allele size as determined by clone sequencing also shows a negative correlation with respect to age at onset and an improved fit to the linear model (R² = 0.6005) (B). Comparing the SCA17 pathogenic allele sizes determined by these two methods there is a very significant correlation between the two and a high-quality fit to the linear model (R² = 0.9935) (C). The bold line depicts the linear model fit result and the 95% confidence interval bounds are shown by the narrow line and shaded in gray.

FIGURE 4

Venn diagram showing the distribution of phenotypes within the cloned SCA17 cohort. Phenotype information was available for 25 patients out of the 30 individuals in the cohort. The number of patients falling into a specific symptom category is shown in black, whilst the subject numbers corresponding to the clone frequency tables (Supplementary Tables S1, S2) are shown in dark blue. Patients with particular phenotypes are enclosed by the appropriate rounded rectangle (Ataxia, blue; Dementia, green; Chorea, red; and Parkinsonism, brown). Five subjects either were asymptomatic or no phenotype information was available.

FIGURE 5

Pedigrees of two case reports from the SCA17 cloned patient cohort. The first family shows two monozygotic twins (II:1, subject #13; II:2, subject #16), who have a clone sequenced mean pathogenic allele size of 47 repeats and mean normal alleles of 38 and 37 repeats, respectively (A). Subject #13 (II:1) presented with a more severe phenotype compared to their sibling, with an age of onset 3 years prior and dementia/cognitive impairment and chorea in addition to the ataxia observed in their sibling (Subject #16, II:2). The second family shows a father (I:1, subject #5) and child (II:3; subject #4) where the father was asymptomatic at the age of 50 years old, whilst the child developed ataxic symptoms at the age of 21 years old (B). Despite the apparent anticipation, they both had a clone sequenced mean pathogenic allele size of 51 repeats, but their normal alleles differed with the child having a slightly larger mean allele size of 39 repeats compared to 37 repeats in their father. DNA was only available for individuals who were cloned, with the given subject number in italics.