American Journal of Medical Genetics Part B (Neuropsychiatric Genetics) 139B:109–114 (2005)

Haplotype Diversity and Somatic Instability inNormal and Expanded SCA8 AllelesSandra Martins,1 Ana I. Seixas,1,2 Paula Magalhaes,1 Paula Coutinho,3 Jorge Sequeiros,1,2 and Isabel Silveira1*

1UnIGENe, IBMC, Universidade do Porto, Portugal2Laboratorio de Genetica Medica, ICBAS, Universidade do Porto, Portugal3Servico Neurologia, Hospital Sao Sebastiao, Santa Maria da Feira; Portugal

Spinocerebellar ataxia type 8 (SCA8) is an auto-somal dominant late-onset neurodegenerative dis-order, belonging to the group of diseases caused bytrinucleotide repeat expansions. SCA8 remainsone of the most intriguing SCAs, regarding thereduced disease penetrance, and the high instabil-ity and poorly understood functional meaning ofthe (CTA)n(CTG)n expansion. We performed haplo-type and sequencing analysis in a large region,encompassing the repeat, in four SCA8 and 20control Portuguese families. The results from thehaplotype study including the combined repeatand six SNP markers showed two different haplo-types, AG-Exp-GTTG and AG-Exp-CTTG, in theSCA8 families. Among the control population,these were also the most frequent, in a total of fivehaplotypes found unequally distributed through-out repeat sizes. From cloning fragments of con-trol, unstable normal and expanded chromosomes,eleven different base substitutions were identifiedin exon A of the SCA8 gene. In some instances,somatic variability in repeat size or base composi-tion was found for a same chromosome, regardlessof its normal or expanded nature. In conclusion,our results in Portuguese families with ataxiashow that SCA8 expansions arose in commonbackgrounds; in addition, this region seems to beunstable beyond the repeat. � 2005 Wiley-Liss, Inc.

KEY WORDS: spinocerebellar ataxia; CTGrepeat; somatic mosaicism; haplo-type

INTRODUCTION

Spinocerebellar ataxia type 8 (SCA8) is an autosomaldominant late-onset neurodegenerative disorder,mainly char-acterized by dysarthria, limb and gait ataxia, and limbspasticity. Fromtheever-growing groupof trinucleotide repeatdiseases, SCA8 has been the most controversial. This SCA,caused by a non-coding (CTA)n(CTG)n expansion on chromo-some 13q21, has first been reported in a large kindred [Koobet al., 1999]. In patients from this family, expanded allelesranged from 110 to 130 combined repeats, whereas in the

general population no alleles with more than 92 repeats werefound. Further studies, however, reported expanded SCA8alleles in some healthy controls, psychiatric affected subjects,and non-ataxic patients [Stevanin et al., 2000; Vincent et al.,2000; Worth et al., 2000]. These findings, together with thereduced disease penetrance observed in families, led to thequestioning of thepathological role of this expanded repeat andto suggest that it could be a rare polymorphism [Vincent et al.,2000; Worth et al., 2000; Stevanin et al., 2000]. However, thehigh frequency of ataxia in carriers of SCA8 expanded allelescompared to controls and the co-segregation of the expansionand ataxia in families support a direct etiological role for therepeat expansion in the disease [Ikeda et al., 2000, 2004;Juvonen et al., 2000; Silveira et al., 2000].

In the germinal line, a high instability at the SCA8 locus inspermatogenesis both in normal and expanded alleles has beenpreviously reported by us [Silveira et al., 2000]. Moreover,some dramatic repeat contractions in sperm samples ofindividuals with SCA8 expansions (ranging from 80 to 800repeats in blood) have been observed, being suggested thatthey may underlie the reduced penetrance associated topaternal transmissions [Moseley et al., 2000].

In SCA8, like in myotonic dystrophy (MD), the repeat islocated in anuntranslated region. The discovery that theSCA8transcript is an endogenous antisense RNA contributed toelucidate the pathogenic role of expansions that occur inuntranslated regions [Nemes et al., 2000]. InMD, themutatedmyotonic protein kinase transcript binds CUG-binding pro-teins causing a defect in the mRNA processing and aconsequent reduction of protein expression, leading to adysregulation of muscle excitability [Strong and Brewster,1997]. In SCA8, the fact that the antisense RNA results fromtranscription through the KLHL1 gene, suggests that it canregulate the expression of theKLHL1 sensegene.Additionally,transfection studies indicated that the KLHL1 protein mayplaya role in organizing theactin cytoskeleton of thebrain cellsin which it is expressed [Nemes et al., 2000].

More recently, a study concerning the SCA8 locus dynamicswas performed through a comparative genetic analysis be-tween humans and great apes, and among different humanpopulations [Andres et al., 2003]. No distinct allele-sizedistributions were found among wild-type SCA8 alleles fromAfrican, European, Indian, and East Asian populations,suggesting that population-specific selective pressures arenot shaping intraspecific variation within humans. Thisfinding indicates that different risks for SCA8 expansion indifferent ethnic backgrounds are not expected. This would betrue even in the presence of a close association between thefrequency of large normal alleles and the disease prevalence,like in other SCAs [Takano et al., 1998].

To gain insight into the biological and pathological roles ofthe SCA8 locus, we performed: (1) an extended haplotypeanalysis, using six single nucleotide polymorphism (SNP)markers spanning a region of 3.8 kb, encompassing the repeat,and (2) a detailed analysis of the (CTA)n(CTG)n and flankingregions.

Grant sponsor: FCT; Grant sponsor: FEDER.

*Correspondence to: Isabel Silveira, Ph.D., UnIGENe, IBMC,Rua do Campo Alegre 823, 4150-180, Porto, Portugal.E-mail: isilveir@ibmc.up.pt

Received 10 March 2005; Accepted 21 July 2005

DOI 10.1002/ajmg.b.30235

� 2005 Wiley-Liss, Inc.

SUBJECTS AND METHODS

Samples

Genomic DNA from four previously diagnosed SCA8 and 20control non-SCA8 families of Portuguese origin was obtainedas previously reported [Silveira et al., 2000], after writteninformed consent.

Genotyping

Repeat sizes at theSCA8 locus in individuals from the controlpopulation and SCA8 families was assessed by PCR amplifica-tion, using previously described primers (SCA8-F4, SCA8-R4)and conditions [Koob et al., 1999; Silveira et al., 2000].

SNP genotyping was performed by PCR-SSCP. Amplificationreactionswere carried out in a total volumeof 25ml, with1 mMofeach primer (Table I); 200 mM of dNTPs; 1 mM MgCl2; 10 mMTris, pH 9.0; 50 mM KCl; 1 U of Taq polymerase; and 2%formamide. Amplification conditions were as follows: initialdenaturationstepat948C, for5min;30cyclesconsistingof948C,for 1min, annealing condition (Table I) and 728C, for 1min; anda final extension of 7 min, at 728C. SSCP electrophoresis wascarried out at 158C, using a 15% polyacrylamide gel.

For each SNP, sequencing analysis was performed in, atleast, one heterozygous and two different homozygous indivi-duals with Big Dye Terminator vs 1.1 (Applied Biosystems,Foster city, CA), using the ABI Prism 310 automatic sequencer(Applied Biosystems, Foster city, CA).

Haplotype Study

TheSNPsused for the haplotype studywere selected fromthedbSNP database (NCBI) or developed during this work. Haplo-types were constructed with these polymorphisms, spanning a�3.8 kb region, flanking on both sides of the repeat (Table I). ToassessallelicphaseofSNPs,undisclosedbysegregationanalysis,we used the PHASE 2.0.2 software [www.stat.washington.edu/stephens/software.html; Stephens et al., 2001]. Based on homo-zygosity, and taking into account the known allelic phasesintroduced,theprobabilityofallpossiblehaplotypecombinationswas calculated for each individual.

Cloning and Sequencing the SCA8 Repeatand Flanking Regions

Fragments of �3.4 kb, containing the repeat and flankingregions, were amplified by PCR, with primers SCA8 long 1F-GGTCCTTCTAGTTTGGCAGAGC and SCA8 long 3385R-CTCAAGTAGAAACCTGGCCCAC, using the proofreadingenzyme Expand Long Template PCR System (Roche, Basel,

Switzerland), according to manufacturers’ instructions. PCRproducts were purified and cloned into a TOPO pcR4 vector,from the TOPO TA cloning kit (Invitrogen, Carlsbad, CA).After isolating the plasmid DNA from single colonies, byminiprep (Qiagen kit, Viagen, Germany), the presence ofnormal or expanded alleles was assessed by PCR amplificationof the SCA8 repeat. The three internal primers SCA8 long474R: GCTCCACCATGGTAAAATGTGC, SCA8 long 989R:CACAATGACAATAGAACC, andSCA8 long 2694F:CTGAGT-CAAACTATGCTAAGTGC were used for sequencing regionsother than those enabled by SNP primers (Table I). Although aproofreading polymerase was used, possible PCR-based arte-facts were excluded by sequencing independent clones fromeach PCR product.

RESULTS

SCA8 Haplotypes

Haplotype analysis of the combined (CTA)n(CTG)n and sixflanking SNPs showed two haplotypes (AG-Exp-GTTG andAG-Exp-CTTG) in expanded alleles from SCA8 patients(Fig. 1); actually, these two haplotypes differ only in SNP C(dbNCBI_167308), which is that with the highest observedheterozygosity (73%). By segregation analysis, theGallelewasobserved on expanded chromosomes from families 1, 2, and 3,whereas the C allele segregated with the expansion in family4. Among the 74 analyzed control chromosomes, 5 differenthaplotypes were found unequally distributed throughoutrepeat sizes (Fig. 2). The AG-Nor-CTTG haplotype was sharedonly by alleles with 27 repeats or less (18–27 repeats). On theother hand, the AG-Nor-GTTG haplotype tended to beassociated with larger alleles (23–33 repeats), being the mostfrequent among those with over 27 repeats and reaching afrequency of 22% for alleles with 23 repeats. These twohaplotypes, the only observed in expanded alleles, were alsothe most frequent among normal chromosomes, attainingtogether a frequency of 84%. Regarding the other threehaplotypes, GG-Nor-CCGA was shared by alleles with 25–28repeat units, while GG-Nor-CTGA and AA-Nor-GTTG werefound only in phase with normal alleles with 18 and 28, and 23repeat units, respectively.

SCA8 Instability

Sequencing analysis of the combined repeat from genomicDNA and from cloned DNA fragments showed a (CTA)nvariation from 8 to 13 repeats in alleles of normal size (Fig. 3).Unstable tracts were observed on four normal chromosomesfrom two SCA8 families (Fig. 3b). In family 2, individual II: 3, a

TABLE I. SNPs Selected for the Haplotype Study

SNP RefSNP IDBase

substitutionDistance from

(CTA)n(CTG)n (bp) Primer sequence (50–30) Annealing conditions

A rs302019 A>Gb �524 F-CAGTAAGGCTGGTAGGGTGG 528C, 1 minR-CAAATTCCATACACTAAACAATTA

B BV210162a G>Ac �60 F-GCCTTTTCTGACTCCCAGCTTC 528C, 30 secR-TTTTCTAACATGAAGGACCCAGG

C rs167308 C>Gb þ2343 F-TGGGGAACAAAAGCAAAAA 558C, 1 minR-CCCCACTCAAACTCCAAAAA

D BV210163a T>Cb þ2439 The same as for SNP C 558C, 1 minE rs302016 G>Tb þ3233 F-AAAAATCAAAAGTTGGCAAGAT 558C, 30 sec

R-TGGCTTCATAACTGCACCATF rs302015 A>Gb þ3322 F-GTCTCAAAATCCCGACCTCA 608C, 1 min

R-CCAGATTATCTTGCCAACTTTTG

aGenBank accession numbers of the new SNPs found in this study and submitted to the NCBI.bAccording to the AF252279 NCBI sequence orientation.cAs in AF126749 reference sequence.

110 Martins et al.

(CTA)13(CTG)14 allele was found in a cloning fragment, where-as by direct sequencing of genomic DNA a (CTA)13(CTG)15allele (shared by the other two sibs) was present. In this samefamily, somatic mosaicism was observed in individual III: 2,who showed alleles (CTA)10(CTG)12 and (CTA)11(CTG)12present in genomic DNA, with the shortest one detected alsoin a cloning fragment. In the othermaternal allele, inheritedbyindividual III:1 (according to the reconstructed haplotypes;Fig. 1), a pointmutation in addition to aCTAdeletion occurred,resulting in the (CTA)9CAA(CTG)12 allele, found in twodifferent cloning fragments from this subject. Instability wasalso observed in family 3, with a substitution in the seventhCTAof individual II: 1, though, again in this case, thiswas onlydetected by sequencing cloning fragments.

Compared to the published sequence AF126749, 11 altera-tions were found in exon A of the SCA8 gene from cloningfragments of control, normal unstable and expanded chromo-somes (Fig. 4). One base substitution was found in two expan-ded chromosomes, either by sequencing cloning fragments, orby direct sequencing of genomic DNA. The other base changesfound in expanded alleles were detected in a cloning fragment,but not by DNA direct sequencing.

DISCUSSION

Through an extended haplotype study of the SCA8 locus, adetailed intergenerational analysis of the repeat and sequen-cing of the flanking SCA8 region, we found haplotype diversity

Fig. 1. Pedigrees and haplotypes from four families with SCA8. Symbols with black dots indicate asymptomatic carriers of expanded alleles. Haplotypesof six SNPs, spanning a �3.8 kb region flanking the repeat. Haplotypes that segregate with the expansion are boxed. * denotes somatic mosaicism, withrepeat sizes differing in one unit for the given chromosome.

Haplotype Diversity in SCA8 111

and somatic instability, both in normal and expanded SCA8chromosomes.

Two haplotypes were associated with expanded chromo-somes in four SCA8 families. As these were also the mostfrequent haplotypes in the normal population, it suggests theymayhave independently expandedup to themutation state; onthe otherhand, as theydiffer only in oneSNP, thehypothesis ofa recurrent mutation in an original background cannot beruled out. Taking into account the high frequencies of

haplotypes AG-Nor-CTTG and AG-Nor-GTTG in the controlpopulation, and thatG alleles are the ancestrals of SNPsC andF (as published in NCBI), we may assume the AG-Nor-GTTGas the ancestral haplotype. The existence of two SNP-basedhaplotypes, in only four affected Portuguese families, seems agreatdiversity compared toa recent studywithmicrosatellites,spanning�1Mb region, in which three haplotypes were foundto segregate with ataxia in 37 families from the United States,Canada, Japan, and Mexico [Ikeda et al., 2004]. Moreover, a

Fig. 2. Haplotype distribution of SCA8 alleles from the normal population (n¼ 74), according to repeat number.

Fig. 3. SCA8 allele configurations assessed by sequencing analysis. The sequence configurations are shown by symbols in the box. The analysedsequences were from: (a) normal alleles, from SCA8 families, without instability; (b) normal unstable alleles (‘‘g’’ denoting genomic DNA and ‘‘c’’ clonedfragments); (c) expanded alleles from SCA8 families; and (d) normal alleles found in the control population.

112 Martins et al.

partially shared core haplotype was found among expandedchromosomes from these multiethnic ataxia families, psychia-tric affected subjects and controls. Whether these and otherworldwide SCA8 expanded chromosomes share any of the twoSNP-based haplotypes described here needs to be known tounravel the events leading to expansion.

Expanded repeat alleles show somatic mosaicism in tissuesfrom patients with SCA8 [Moseley et al., 2000; Silveira et al.,2000]. We found, for a given subject, cloning fragments fromthe samenormal chromosomediffering in repeat sizes fromoneunit; this indicates that somatic mosaicism also occurs innormal repeat alleles in this gene. In addition, base substitu-tions were found in single cloning fragments, but not by directsequencing of genomic DNA. This further supports thepresence of mitotic instability in this region, which may playa role in clinical variability as well as in reduced diseasepenetrance.

Haplotype analysis did not help to explain reduced pene-trance inSCA8, in this or previous studies [Silveira et al., 2000;Ikeda et al., 2004]. Another triplet repeat disease, fragile-X-associated tremor/ataxia syndrome (FXTAS) affects only one-third of male premutation carriers of the fragile X mentalretardation gene FMR1 [Jacquemont et al., 2003]. In thisdisorder, neuropathological findings together with the pre-sence of elevated levels of FMR1 mRNA in premutationcarriers support a role for FMR1 transcripts in neurodegen-eration. Whether this is also the case in SCA8 needs to be fullyinvestigated.

In conclusion, our results show that SCA8 expansions mayhave arisen in two different backgrounds in our Portuguesefamilies; these are also the most common in normal alleles.Additional studies are needed to understand how this repeatexpands to cause neurodegeneration, and if the model of RNA-mediated neurodegeneration proposed is also applicable toSCA8.

ACKNOWLEDGMENTS

This work was supported by research grant POCTI/34517/MGI/2000, and Financiamento Plurianual de Unidades deInvestigacao, both from FCT (Fundacao para a Ciencia e aTecnologia) andFEDER.S.M. andA.I.S. are also the recipientsof scholarships from FCT, Portugal.

REFERENCES

Andres AM, Soldevila M, Saitou N, Volpini V, Calafell F, Bertranpetit J.2003. Understanding the dynamics of Spinocerebellar Ataxia 8 (SCA8)locus through a comparative genetic approach in humans and apes.Neurosci Lett 336:143–146.

IkedaY, ShizukaM,WatanabeM,OkamotoK, ShojiM. 2000.Molecular andclinical analyses of spinocerebellar ataxia type 8 in Japan. Neurology54:950–955.

Ikeda Y, Dalton JC, Moseley ML, Gardner KL, Bird TD, Ashizawa T, SeltzerWK, Pandolfo M, Milunsky A, Potter NT, Shoji M, Vincent JB, Day JW,Ranum LP. 2004. Spinocerebellar ataxia type 8: Molecular geneticcomparisons and haplotype analysis of 37 families with ataxia. Am JHumGenet 75:3–16.

Jacquemont S, Hagerman RJ, Leehey M, Grigsby J, Zhang L, BrunbergJA, Greco C, Des Portes V, Jardini T, Levine R, Berry-Kravis E, BrownWT, Schaeffer S, Kissel J, Tassone F, Hagerman PJ. 2003. Fragile Xpremutation tremor/ataxia syndrome: Molecular, clinical, and neuroi-maging correlates. Am J Hum Genet 72:869–878.

Juvonen V, HietalaM, PaivarintaM, RantamakiM, Hakamies L, KaakkolaS, Vierimaa O, Penttinen M, Savontaus ML. 2000. Clinical and geneticfindings in Finnish ataxia patients with the spinocerebellar ataxia 8repeat expansion. Ann Neurol 48:354–361.

KoobMD,MoseleyML, Schut LJ, BenzowKA,Bird TD,Day JW,RanumLP.1999. An untranslated CTG expansion causes a novel form ofspinocerebellar ataxia (SCA8). Nat Genet 21:379–384.

Moseley ML, Schut LJ, Bird TD, Koob MD, Day JW, Ranum LP. 2000. SCA8CTG repeat: En masse contractions in sperm and intergenerationalsequence changesmay play a role in reduced penetrance.HumMol Genet9:2125–2130.

Nemes JP, BenzowKA,MoseleyML, RanumLP, KoobMD. 2000. The SCA8transcript is an antisense RNA to a brain-specific transcript encoding anovel actin-binding protein (KLHL1). Hum Mol Genet 9:1543–1551.

Silveira I, Alonso I, Guimaraes L, Mendonca P, Santos C, Maciel P, FidalgoDeMatos JM, CostaM, Barbot C, Tuna A, Barros J, Jardim L, CoutinhoP, Sequeiros J. 2000. High germinal instability of the (CTG)n at theSCA8 locus of both expanded and normal alleles. Am J Hum Genet66:830–840.

Stephens M, Smith NJ, Donnelly P, PHASE version 2.0.2. 2001. A newstatistical method for haplotype reconstruction from population data.Am J Hum Genet 68:978–989.

Stevanin G, Herman A, Durr A, Jodice C, Frontali M, Agid Y, Brice A. 2000.Are (CTG)n expansions at the SCA8 locus rare polymorphisms? NatGenet 24:213.

Strong PN, Brewster BS. 1997. Myotonic dystrophy: Molecular and cellularconsequences of expanded DNA repeats are elusive. J InheritMetab Dis20:159–170.

Fig. 4. Observed base substitutions in exonA in control, unstable normal, and expanded chromosomes. The circle sizes are proportional to the number ofindividuals with the alteration.

Haplotype Diversity in SCA8 113

Takano H, Cancel G, Ikeuchi T, Lorenzetti D, Mawad R, StevaninG, Didierjean O, Durr A, Oyake M, Shimohata T, Sasaki R,Koide R, Igarashi S, Hayashi S, Takiyama Y, Nishizawa M, Tanaka H,Zoghbi H, Brice A, Tsuji S. 1998. Close associations between pre-valences of dominantly inherited spinocerebellar ataxias with CAG-repeat expansions and frequencies of large normal CAG alleles inJapanese and Caucasian populations. Am J Hum Genet 63:1060–1066.

Vincent JB, Neves-Pereira ML, Paterson AD, Yamamoto E, Parikh SV,MacciardiF,GurlingHM,PotkinSG,PatoCN,MacedoA,KovacsM,DaviesM, Lieberman JA, Meltzer HY, Petronis A, Kennedy JL. 2000. An unstabletrinucleotide-repeat region on chromosome 13 implicated in spinocerebellarataxia: A common expansion locus. Am J HumGenet 66:819–829.

Worth PF, Houlden H, Giunti P, Davis MB, Wood NW. 2000. Large,expanded repeats in SCA8 are not confined to patients with cerebellarataxia. Nat Genet 24:214–215.

114 Martins et al.