The spinocerebellar ataxias (SCAs) are a clinically and genetically heterogeneous group of dominantly inherited degenerative disorders characterised primarily by progressive ataxia. SCA numbers are applied sequentially as new forms are delineated: currently SCA 1 through to SCA 29.1 2 The numbering is imperfect: 9 and 24 are “vacant,” SCA 24 having been reassigned as SCAR 4; the Japanese dominant ataxia due to PLEKHG4 mutation, its locus close to SCA 4 on chromosome 16q22,3 has not been accorded its own SCA number; SCA 16 has recently been shown to be identical to SCA 15,4 5 and SCAs 22 and 29 have not been excluded as representing allelic variants of SCAs 19 and 15 respectively.6 7 The causative gene has been discovered for SCAs 1–3, 5–8, 10–15, 17, 27, and the Japanese dominant ataxia due to PLEKHG4 mutation. In the remainder (SCAs 4, 18–23, 25, 26, 28 and 29), assignment has been on the basis of a linkage having been established.1 2 Some SCAs, such as SCA 6, are relatively “pure” ataxias (with minor pyramidal features allowed), while others have a wider pattern of neurological involvement, ranging from neuropathy (eg, SCA 4) through retinopathy (SCA 7) to epilepsy (SCA 10) and dementia (eg, SCA 2, SCA 17). One, SCA 20, has a pathognomonic neuroradiological sign: early dentate nucleus calcification.8 While clinical distinction among the “pure” ataxias is difficult, the tempo of progression (eg, very slow in SCA 15) and the presence of certain findings (eg, early axial myoclonus in some patients with SCA 14) can provide useful clinical clues.9

Here we report a family with a new dominantly inherited pure ataxia. Linkage has been excluded to all currently known SCA loci, and established to chromosome 4q34.3–q35.1. The designation SCA 30 has been applied. A preliminary report was presented at the 2008 American Academy of Neurology Annual Meeting.10

SUBJECTS AND METHODS

All available adult family members underwent a standardised neurological examination (ES and MF). Vestibulo-ocular reflex (VOR) gain was assessed as normal if dynamic visual acuity (with head oscillation at 2 Hz) dropped less than two lines compared with static acuity on a Snellen chart,11 and if head oscillation during ophthalmoscopy did not result in retinal movement. Quantitative sensory testing of large fibre function was performed utilising the Rydel–Seiffer tuning fork.12 MRI brain scans and lower-limb nerve conduction studies were performed on two representative affected family members.

SCAs 1, 2, 3, 6 and 7 were excluded by direct genetic testing. An autosomal genome-wide linkage study was undertaken using the Affymetrix Xba 10 K gene chip platform. Error checking of the data was done with respect to detection of non-Mendelian inheritance using the program PedCheck,13 and of apparent double recombinants using Merlin.14 Parametric genome-wide logarithm (base 10) of odds (LOD) scores were generated using Allegro, on the assumption of a rare fully penetrant autosomal dominant disease, with an abnormal allele frequency of 0.0001. Marker allele frequencies were obtained from the HapMap dataset.15 The sites of the known SCA loci were obtained from EntrezGene and from the Washington University Neuromuscular website.2 Physical locations were converted into genetic locations using the deCODE map.16 The likely haplotypes of the chromosome 4 region of interest were generated with the Allegro program.17

Endeavour,18 one of several new in silico gene prioritisation programs, was used to assemble a list of candidate genes for the SCA 30 locus. To test the hypothesis that the SCA genes have commonalities, we performed a simulation study of the 17 identified SCA genes. We selected a 10 Mb (∼10 cM) linkage region spanning each locus, and used the remaining 16 genes in each case to prioritise the genes in each linkage region. If the SCA genes had commonalities, the resultant rankings should be more accurate than those resulting from 16 random non-SCA genes. The results were assessed by calculating the area under the Receiver Operating Characteristic curves.

RESULTS

The pedigree is shown in simplified form in fig 1A, together with deduced haplotypes in the linkage region. Six affected family members from two generations were available for assessment. The mean reported age of onset was 52 (range 45–76), but the onset was typically so insidious that this was difficult to establish accurately, and we consider it probable that in fact symptoms would have appeared earlier than stated. There were insufficient parent–child pairs to assess possible anticipation. The mean illness duration at assessment was 8 years (range 6–10). All had mild-moderate dysarthria, appendicular ataxia and gait ataxia. One (II:7, aged 85 when assessed) needed a walking frame, and one (III:8, aged 55) required an assistant to walk. The VOR gain was clinically normal in all. Visual suppression of the VOR was uniformly slightly impaired, but none showed obviously broken-up smooth pursuit. Saccadic hypermetria was universal (typically horizontal and into down gaze); saccadic velocity was clinically normal. Only one displayed (slight) gaze-evoked nystagmus, on extreme right gaze only. Four had mild lower limb hyper-reflexia (three with slight crossed adductor spread), but none had spastic tone or extensor plantar responses. None had impaired vibration perception at the great toes (III:9 not tested). Two (III:6 and III:9) had normal lower-limb nerve conduction studies, comprising tibial motor conduction to abductor hallucis brevis, tibial F-waves, and sural sensory nerve action potentials. Only one (III:6), who had been on olanzapine for 5 years for a reported schizo-affective disorder, displayed (oromandibular) dystonia. Only the most severely affected (III:8) had an index-finger tapping rate >1 SD below average for age, gender and education. No myoclonus or tremor was evident.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) Partial pedigree, and haplotype construction using single nucleotide polymorphism (SNP) markers from the 4q34.3–q35.1 chromosomal region. Filled symbol = affected, personal examination. N in symbol = normal to personal examination. Grey symbol = anecdotally affected (I:2, II:3, 5) and/or obligate heterozygote (II:1, 3, 5). Of the apparently unaffected living persons, only those from whom blood samples had been drawn are included here (including one unaffected spouse, II:4). Inferred haplotypes, according to the most parsimonious construction, are displayed beneath that person’s symbol. The chromosomal segment which includes the putative spinocerebellar ataxia 30 mutation is shown in grey, the normal chromosome from the affected parent in white, and the other chromosomes from the unaffected parents in various cross-hatchings. The informative recombinations, delineating the extent of the candidate region, are indicated by arrowheads on the chromosomes of II:7 and III:5. (See text for molecular detail.) (B) Parametric logarithm (base 10) of odds (LOD) scores on chromosome 4, fully penetrant model. The peak region reaches a value of 3.0, and extends from SNP marker rs1397413 to marker rs2175476.

Two (III:6, III:9) underwent MRI scanning. This showed atrophy of the superior and dorsal cerebellar vermis, less prominent cerebellar hemisphere atrophy and no involvement of brainstem or supratentorial structures (fig 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

MRI scans of III:6 (left) and III:9 (right), showing atrophy of superior and dorsal vermis and cerebellar hemispheres, with sparing of inferior vermis, brainstem and supratentorial structures.

Four at-risk pedigree members (III:1, 2, 4 and 5), being the offspring of presumed obligate heterozygotes, showed no evidence of ataxia on careful examination at ages 71, 67, 60 and 67, respectively, and were included in the linkage analysis as unaffected persons with an age-dependent liability estimated from the ages of onset of the affected individuals. These conservative penetrances are Pr(dis|AA) = Pr(dis|Aa) = 0.8 and Pr(dis|aa) = 0. This approach allows that these individuals might yet develop the disease, albeit with a very low probability, reflecting their older ages at examination.

We had anecdotal information on five deceased family members. Individual I:2 died at age 74. She was said to have had a condition “similar to Parkinson’s, with slurred speech, an unsteady gait, and all movements restricted.” Her brother (not shown in the pedigree) is said to have had “a similar condition.” Individual II:1 was regarded by the family as unaffected at death at age 75, but given the insidious development of the disorder (eg, a clearly affected individual, III:3, was also regarded by herself and the family as unaffected), this suggestion of non-penetrance may not be accurate. The two other sons of I:2, II:3 and II:5, were described possibly to have been affected: II-3’s speech was “difficult to understand,” while II:5 was diagnosed by a neurologist as having progressive supranuclear palsy, although perusal of the clinical notes indicated a number of unusual features suggesting an alternative diagnosis of a progressive ataxia: notably, the record referred to normal downward saccades and an “unusual,” broad-based gait.

SCAs 1, 2, 3, 6 and 7 were excluded in individual III:9 on direct genetic testing. Of the 27 reported SCA loci, all were excluded with LOD scores of less than zero. A single region of interest was identified at chromosome 4q34.3–q35.1, with a LOD score of 3.0. The haplotype region that is defined by using both affecteds and unaffecteds is 5 Mb (∼12 cM) long, bounded by single nucleotide polymorphism markers rs1397413 at physical map position 179450350, and rs2175476 at physical map position 184452826 on chromosome 4, according to the UCSC Human Genome Browser.19 The centromeric bound is given by recombination in an affected individual, II:7, while the telomeric bound is due to III:5, who was unaffected on targeted examination at age 67. The linkage results are shown in fig 1B. A total of 19 known and predicted genes are contained within the delineated region, none of which were compelling candidates.

The in silico simulation study indicated that, although the SCA genes are heterogeneous, subsets do share commonalities, and this can validly be utilised to improve likelihood ranking among genes in each linkage region. A ranking of the 19 genes in the SCA 30 region of interest produced ODZ3 (odd Oz/ten-m homologue 3), a known gene with brain expression,20 as the top candidate.

DISCUSSION

This family displays a late-onset, relatively “pure” cerebellar ataxia, mapping to a region of chromosome 4 (4q34.3–q35.1) that contains none of the currently known SCA loci. This therefore presumably represents a previously undescribed condition. The rubric SCA 30 has been applied, with the approval of the Human Genome Nomenclature Committee. The differential diagnosis of SCA 30 would include the other dominantly inherited “pure” SCAs: 5, 6, 8, 11, 14, 15, 16, 22 and 26. The lack of prominent nystagmus in several affected members of a pedigree would weigh against SCAs 5, 6, 8, 11, 15/16 and 22, in which nystagmus is present in the majority of affected individuals,2 but there are no other distinctive features. The fact that II:7 and III:3 regarded themselves as unaffected, and II:1 was so regarded by his family, suggests that the onset of the disorder may pass unrecognised for some years, and that the progression is slower than according to the affected individuals’ histories.