Modelo de Conglomerados

The data presented necessitate a careful re-evaluation o f the validity o f the SCA8 expansion as a cause o f cerebellar ataxia. Five expanded SCA8 alleles with > 100 combined repeats were found in DNA from individuals with no known neurological disease, individuals with Parkinson’s disease and HMSN 1A as well as patients with cerebellar ataxia. The large overlap between allele sizes found in the control and affected individuals precludes the clear definition o f normal and pathological ranges. The frequency o f either intermediate (43-99 CR’s) or large (>100 CR’s) alleles was not clearly different among the three groups o f subjects with cerebellar ataxia (ADCA, ILOCA and ARCA), nor was it different between controls and ataxia subjects. This finding alone argues strongly against a pathological role for the SCA8 repeat.

The finding of five large expanded alleles in the control population is particularly

noteworthy. One o f these alleles (133 CR) was o f a similar size, and another (174 CR) was

larger than those found in the SCA8 family described by Koob et al (Koob et ai^ 1999).

The possibility that these two individuals would have gone on to develop cerebellar ataxia cannot be ruled out. However, this is unlikely, as the 174 allele was found in leucocyte DNA from a healthy 36-year-old Brazilian female, with no family history o f cerebellar ataxia and the 133 repeat allele was found in a female who died o f ischaemic heart disease, aged 56 years.

An expanded allele with 208 CR was found in a patient who, together with his deceased

recessive late-onset “multisystem disorder” including a moderately-severe cerebellar syndrome. The clinical picture of the disorder in this family is therefore at variance with

that o f the SCA8 families described by Koob et al. (Koob et al.., 1999). Although, the sister’s

DNA was not available, the unaffected 92-year-old mother of both patients was found to have an expanded allele of 127 CR which appears to have undergone an expansion by 81 CTG repeats during meiosis, further demonstrating the tendency for maternally-derived

SCA8 alleles to undergo expansion. Koob et al. reported a 42-year-old asymptomatic

individual with 140 CTG repeats. However, the non-penetrance o f the 127 CR allele is remarkable; this allele falls within the pathological range defined by these authors and yet remains non-penetrant in a 92 year old.

It is possible that intermptions in the CTG tract may influence penetrance o f expanded SCA8 alleles; reports o f CCG, CTA, CTC, CCA, CTT or TTG interruptions, either single

or in tandem, within the CTG tract exist (Moseley et al, 1999; Stevanin et al, 2000),

although no consistent pattern has been observed which would account for this

phenomenon. Moreover, the CTG tracts o f the expanded alleles identified in the present study were unintermpted in both ataxia patients and controls. The length o f the CTA tract or the ratio between the number of CTAs and CTGs could also have an influence on penetrance; previous studies have reported CTA tracts o f 1-21 repeats but the current study identified alleles with 3-11 CTA with no consistent differences between controls and affected subjects. Hence the length of the CTA does not appear to be the critical factor.

Since these data were published (Worth et al., 2000), a number o f other studies have cast

further doubt on the validity of the SCA8 gene as a cause o f cerebellar ataxia (Stevanin et

ai, 2000; Vincent et al., 2000). Stevanin et al. (Stevanin et al, 2000) found alleles with 107,

111 and 123 CR among control subjects and alleles with 92-111 CR in a family with Lafora body disease. In addition, disease and expansions did not co-segregate in two families with ADCA. Moreover, in both this and the present smdy, the frequency o f large SCA8 alleles with more than 200 repeats may have been underestimated among both the ataxia and the control population. It will have been noted in the methods section o f the present study that large SCA8 expansions in common with those o f SCA7 are sometimes difficult to amplify with PCR. Some large alleles with >200 repeats may have not been amplified and this may account for the absence o f expansions among controls in two smdies where it has been

claimed that SCA8 is a relatively common cause o f cerebellar ataxia (Ikeda et al, 2000;

between 100 and 1,300 repeats were present in 14 patients with psychosis. They also found an allele segregating in control CEPH reference family 1334 with a range o f sizes from 160-900 repeats and a further 4 expanded alleles in other controls. The alleles with high

repeat numbers were identified by Southern blotting o f EcoRI digests ensuring detection of

these large alleles.

In response to the controversy that these studies have generated, the authors o f the original

SCA8 paper have restated (Moseley et al., 2000a) the evidence which, in their view, is

compelling:

1. The expanded repeat is linked to ataxia in the large family with a LOD score o f 6.8 at

0=0.00.

2. A “biologic relationship” between repeat length and disease is alleged, where affected subjects in their series have significantly longer CTG tracts (mean 117) than asymptomatic carriers (mean 92, p < 10’^).

3. The absence o f alleles in the pathogenic range in 1200 control chromosomes in their series.

4. A high frequency of expansions among apparently unrelated ataxia patients (seven distinct haplotypes segregate in their eight families)

5. The expression of SCA8 transcripts mainly in central nervous tissue.

It is clear that none o f the above evidence proves that the SCA8 repeat causes cerebellar ataxia, although the points raised are variably convincing. The absence o f expanded alleles in the 1200 CEPH chromosomes analysed by Koob et al. is surprising given the three studies (including the present study) which have identified expanded alleles in normal

controls or subjects with other neurological diseases (Stevanin et at., 2000; Vincent et al.,

2000; Worth et al., 2000). Alleles with up to 250 repeats would most likely have been

detected among these CEPH controls in the study by Koob et al., as alleles up to this size were detected in the ataxia patients. Clearly, the frequency of the SCA8 expansion may vary significantly from one population to another, regardless o f its pathogenicity.

In order to explain the apparently reduced penetrance o f the SCA8 repeat, Moseley et al. hypothesise that there may be a pathogenic ‘window^; hence CTG repeat lengths either less

than or exceeding the 107-127 range identified may be non-pathological (Moseley etal.^

2000b). However, the following findings cannot be explained by this hypothesis: the non­ penetrance o f the allele with 127 CR identified in the 92-year-old mother o f a patient with cerebellar ataxia in the present study, the 117 CR allele identified in a 33-year-old control

by Vincent et al. (Vincent et al.^ 2000) and the two alleles o f 111 and 123 CR identified by

Stevanin et al. (Stevanin et al., 2000) in normal controls aged 62 and 64, respectively.

The presence o f seven different SCA8 haplotypes segregating in the eight ataxia families

described by Koob et al. (Koob et al., 1999) is an important observation. However, it will

be remembered that in seven o f these families, linkage o f the disease phenotype to the repeat is not established owing to the small number of affected individuals in these families. There are, in addition, two interpretations o f the observation o f multiple haplotypes:

1. the repeat is the causative mutation, is not in linkage disequilibrium with another gene for cerebellar ataxia on 13q21, and the haplotype data indicate multiple founders;

2. alternatively, the SCA8 repeat is non-pathological, and arises frequently de novo in the

normal population, and is therefore found at a similar frequency in both ataxia and control populations.

If it could be shown that uncommon or even unique haplotypes are present in the subjects with cerebellar ataxia who have expanded SCA8 alleles, this would provide stronger evidence o f pathogenicity than multiple haplotypes alone. This evidence would carry an important proviso: if the SCA8 mutation arose on a common haplotype, then this haplotype may not be observed any less frequently among ataxia patients with expanded SCA8 alleles. Further studies may help to clarify this point.

6.5.2 Analysis of the SCA12 CAG repeat

In this study, no SCAl 2 alleles with greater than 30 CAG repeats were found among 392 individuals with cerebellar ataxia . However, a 30-repeat allele was identified in a patient with clinically Hkely autosomal recessive cerebellar ataxia with an early age at onset. This allele is unlikely to be pathological, as it is much shorter (by 36 repeats) than the smallest

allele identified in an affected family member. Therefore, the range of normal for SCAl 2 can probably be revised to 7-30 repeats.

The SCAl 2 gene encodes a brain-specific regulatory subunit (PPP2R2B) o f a serine / threonine protein phosphatase 2A, PP2A. The cellular effect o f the SCAl2 CAG expansion remains to be demonstrated. It has previously been shown that none o f 394 normal controls or 351 individuals with unrelated neurological diseases had SCAl 2

expansions (Holmes et al.^ 1999). In addition, this study found only one asymptomatic

individual, aged 49, who had an expanded repeat. These observations lessen the possibility that the expansion is in linkage disequilibrium with another gene for cerebellar ataxia on 5q31-33.

Age at onset in SCAl 2 is 8-55 years and most individuals in the one family described present in the fourth decade with upper extremity tremor, progressing slowly to include head tremor, gait ataxia, dysmetria, dysdiadochokinesia, hyperreflexia, hypokinesia,

abnormal eye movements and dementia (Holmes et al.^ 1999). The phenotype is somewhat

unusual for the SCAs; the presenting symptom is usually tremor, and the ensuing

syndrome is suggestive o f a global neurodegenerative condition o f which ataxia is certainly one important feature. Therefore, in addition to screening documented cases of ADCA and ILOCA with no genetic diagnosis, subjects without a firm clinical diagnosis or family history, or who had features atypical for SCA were included in this screen. However, none o f these patients had SCAl2 expansions. Similarly, Holmes et al. found no expansions among approximately 300 additional familial cases o f cerebellar ataxia. Taken with our findings, it is evident that SCAl 2 is likely to be a rare cause o f cerebellar ataxia, particularly in the United Kingdom.

6.6 CONCLUSIONS AND DIRECTIONS FOR FUTURE STUDY

With respect to SCAB, the findings o f this study and those o f other investigators cast some doubt on the validity of SCAB as a cause o f cerebellar ataxia. SCAB is the first SCA where an expanded untranslated CTG repeat was alleged to be the mechanism underlying the disease process. However, a CTG repeat is established as the cause o f Dystrophia

Myotonica (DM). It is not fully understood how this CTG expansion, which is within the

3’ untranslated region of the DMPK gene, causes an affect at the cellular level. Mouse

models (Jansen et a/.^ 1996; Reddy eta/., 1996) and the absence o f pathogenic point

simply a lack or excess o f DMPK. There is evidence that the CTG repeat may affect the

expression o f at least three genes: DMPK, DMWD and SDC5. Alternatively, DM may be

caused by abnormal interactions between the RNA containing the GUG expansion and

other cellular factors (Davis et al., 1997; Hofmann-Radvanyi et al., 1993; Krahe et al, 1995;

Taneja et al., 1995). Recent experiments suggest that RNA containing the GUG repeat

expansion results in a gain o f function through its increased affinity for GUG-binding

proteins (Philips et al, 1998; Roberts et al, 1997; Timchenko et al, 1996).

If the GTG expansion at the SCAB locus is subsequently proven to be a cause o f cerebellar ataxia, it is likely to provide important new insights into the process o f neurodegeneration. Recently, it has been shown that the most 5' exon o f the SCAB gene is transcribed through the first exon o f another gene that is transcribed in the opposite orientation

(Nemes et al, 2000). Nemes et al. suggest that the SCAB transcript is an endogenous

antisense RNA that overlaps the transcription and translation start sites as well as the first splice donor sequence o f the sense gene. This gene has been denoted Kelch-like 1 (KLHLl); the sense transcript encodes a 74B amino acid protein with a predicted domain structure typical o f a family o f actin-organizing proteins related to the Drosophila Kelch gene. The same authors have identified the full-length cDNA sequence for both the human and mouse KLHLl genes, and have elucidated the general genomic organization o f the human gene. The predicted open reading frame and promoter region are highly conserved, and both genes are primarily expressed in specific brain tissues, including the cerebellum, the tissue most affected by SGAB. Experiments have shown that the protein localizes to the cytoplasm, suggesting that it may play a role in organizing the actin cytoskeleton o f the brain cells in which it is expressed.

Despite these findings, the precise function of the KLH Ll gene and proof that the SGAB transcript acts in the way suggested is not yet established. The effect o f the expanded repeat also remains to be demonstrated. Until this proof is forthcoming, and a mechanism for non-penetrance is proposed, it has been argued that diagnostic or predictive testing for

expansions in SGAB should not be carried out (Worth et al, 2000).

The untranslated SGA12 GAG expansion appears to present an additional new mechanism o f mutation associated with cerebellar ataxia. Although no expansions in normal controls or subjects with other neurological diseases have yet been identified, the causative nature of the GAG expansion remains unproven. As discussed above, the GAG repeat is presumed

to lie in the 5’-untranslated region o f the PPP2R2B gene, but no PPP2R2B transcript containing the CAG repeat nor altered expression o f this gene in patients with an

expansion has been identified to date. More studies are needed to demonstrate the validity of this hypothesis

The rarity o f expansions in SCAl 2 is o f great interest. This has important implications for the appropriateness o f random testing o f individuals with cerebellar ataxia. However, until other families and the range o f phenotypes associated with this disorder are identified, no clear recommendations for targeted screening can be made, and therefore currendy, all patients with dominantly inherited ataxia, especially those with unusual, complicating features can be potential candidates for screening for SCAl 2. In view o f the prominence o f tremor in the early stages o f the disease, it might be prudent to consider screening selected patients with a clinical diagnosis o f essential tremor for the SCAl 2 expansion.

Together, the identification o f SCA8 and SCAl 2 have potential to define further the genetic heterogeneity of the spinocerebellar ataxias. Any potential unifying theory of neurodegeneration must now take account o f these genes. The degree to which polyglutamine expansions can account for the pathological mechanism underlying the SCAs appears to be limited. Do the SCA8 and SCAl2 transcripts play a role in the same or an unrelated injury mechanism? It wiU be o f great interest to discover if neuronal inclusions are identified in either o f these disorders. The nature o f the mutations in the many SCAs which probably remain undiscovered can, it seems, only add to the growing number of unanswered questions.

CHAPTER 7. MOLECULAR A N D CLINICAL STUDY OF 18