Lack of Association Between the hSKCa3 ChannelGene CAG Polymorphism and Schizophrenia

Ridha Joober,1–3* Chawki Benkelfat,3 Kateri Brisebois,1 Andre Toulouse,1 Ronald G.A. Lafreniere,1Gustavo Turecki,1 Samarthji Lal,1,2 David Bloom,2 Alain Labelle,4 Pierre Lalonde,5 Diane Fortin,1Martin Alda,4 Roberta Palmour,3 and Guy A. Rouleau1

1Montreal General Hospital Research Institute, Montreal, Quebec, Canada2Douglas Hospital Research Center, Verdun, Quebec, Canada3Department of Psychiatry, McGill University, Montreal, Quebec, Canada4Department of Psychiatry, University of Ottawa and Royal Ottawa Hospital, Ottawa, Ontario, Canada5Department of Psychiatry, Universite de Montreal and L.H. Lafontaine Hospital, Montreal, Quebec, Canada

Genetic anticipation, a phenomenon char-acterized by increased severity of symptomsand earlier age at onset of a disease in suc-cessive generations, is believed to be pre-sent in schizophrenia. In several neurode-generative diseases showing anticipation,the mutation causing the disease is an ex-panded trinucleotide repeat. Therefore,genes containing trinucleotide repeatsprone to expansion have become a suitablefamily of candidate genes in schizophrenia.A human calcium-activated potassiumchannel gene (hSKCa3), possibly mappingto chromosome 22q11–13, a region previ-ously linked to schizophrenia, was recentlydescribed. This gene contains two contigu-ous expressed CAG repeat stretches. Re-cently, long allelic variants of one of theseCAG repeats were found to be overrepre-sented in schizophrenic patients comparedto normal controls. In this study we at-tempted to replicate this result and to studythe relationship between the length of thisCAG repeat on the one hand and the sever-ity and age at onset of the disease on theother hand. No association with the diseaseor correlation with the severity of schizo-phrenia was identified. In addition, hSKCa3was mapped to chromosome 1. Our resultsdo not support the involvement of this par-ticular CAG repeat-containing gene inschizophrenia. Am. J. Med. Genet. (Neuro-psychiatr. Genet.) 88:154–157, 1999.© 1999 Wiley-Liss, Inc.

KEY WORDS: schizophrenia; trinucleotide

repeats; potassium channel;genetic anticipation

INTRODUCTION

Schizophrenia is a complex disorder, which resultsfrom both genetic and environmental factors [Kendlerand Diehl, 1993]. To date, despite intensive study, nogenetic markers or genes linked to or associated withschizophrenia have been firmly identified [Karayiorgouand Gogos, 1997]. It is thought that this may resultfrom several problems, including complex genotype/phenotype relationships [Tsuang and Faraone, 1995],complex mode of inheritance [Risch, 1990], and lack ofexternal validation for the diagnosis of schizophrenia[Faraone et al., 1996].

In the last few years, trinucleotide repeat expansions(TRE) have been identified as the cause of several neu-rodegenerative [Paulson and Fischbeck, 1996] and neu-rodevelopmental [Reiss et al., 1986] diseases. Geneticanticipation, i.e., increased expressivity/penetrance ofthe disease in subsequent generations, is a constantclinical manifestation in the diseases caused by TRE.Several studies have shown anticipation in schizophre-nia [Bassett and Husted, 1997; Gorwood et al., 1996],prompting research into the possible involvement oftrinucleotide repeats in this disease. Although specificmolecular evidence for the involvement of TRE inschizophrenia is still lacking, several studies using therepeat expansion detection method found an overrep-resentation of large CAG arrays in schizophrenic pa-tients compared to normal controls [O’Donovan et al.,1996]. These studies also indicate that, in contrast withneurodegenerative diseases caused by expanded CAGrepeat expansions, the ranges of CAG repeat sizes ob-served in schizophrenic patients and in normal controlsoverlap significantly. This observation may indicatethat allelic variants in the upper range of the size dis-tribution, rather than rare mutations, are implicatedin complex diseases such as schizophrenia. Consistent

*Correspondence to: Dr. Ridha Joober, Douglas Hospital Re-search Center, 6875 Boulevard LaSalle, Verdun, H4H 1R3 Que-bec, Canada. E-mail: rjoobe@po-box.mcgill.ca

Received 13 March 1998; Accepted 14 June 1998

American Journal of Medical Genetics (Neuropsychiatric Genetics) 88:154–157 (1999)

© 1999 Wiley-Liss, Inc.

with this hypothesis, long CAG repeat allelic variantsof a human calcium-activated potassium channel(hSKCa3) were recently found to be significantly morefrequent in schizophrenic patients compared to normalcontrols [Chandy et al., 1998]. This finding is particu-larly interesting because hSKCa3 plays a critical rolein determining the firing pattern of neurons and be-cause the corresponding gene was reported to map inchromosome 22q11–13 (between markers DS22301 andTOP1–P2) [Chandy et al., 1997], a region previouslylinked to schizophrenia. Its possible role in schizophre-nia needs further investigation.

As part of an ongoing association study betweengenes harboring polymorphic and expressed CAG re-peats and schizophrenia and/or its phenotypic variabil-ity (long-term outcome, neuroleptic responsiveness),we studied the hSKCa3 gene and attempted to confirmits chromosomal location. Here, we report on the re-sults of an association study between schizophreniaand the hSKCa3 CAG repeat alleles and further inves-tigated the relationship between the phenotypic vari-ability of this disease and the CAG repeat length atthis locus. In particular, we predicted that, if this geneconfers some susceptibility to schizophrenia, longerCAG repeats will be more frequent in severely affectedand neuroleptic-resistant compared to mildly affectedand neuroleptic-responsive schizophrenic patients.

MATERIALS AND METHODSClinical Methods

Schizophrenic patients were recruited according to apriori defined criteria of responsiveness to neurolepticmedication and long-term outcome. We recruited 63neuroleptic-nonresponsive (NRs) patients and 43 neu-roleptic-responsive (Rs) patients. All NRs (n 4 63) pa-tients were diagnosed according to DSM-IV criteriaand were characterized by 1) continuous psychoticsymptoms with no significant remission within thepast 2 years, 2) at least three periods of treatment withtypical neuroleptics at optimal clinical requirementswith no significant relief of symptoms in the preceding5 years, and 3) inability to function without supervisionin all or nearly all domains of social and vocationalactivities within the last 12 months.

Neuroleptic-responders were DSM-IV-diagnosed andhad at least one admission to a psychiatric care facilitybecause of an acute psychotic episode. They always ex-perienced full or partial remission in response to treat-ment with typical neuroleptics and were able to func-tion with only occasional supervision in all or nearly alldomains of social and vocational activities. They were

required to be in remission or quasiremission if andwhen compliant with their medication.

All schizophrenic patients were directly interviewedusing the Diagnostic Interview for Genetic Studies(DIGS) [Nurnberger et al., 1994], and their medicalrecords were comprehensively reviewed by a researchpsychiatrist trained in the use of structured interviews.Complementary information from the treating physi-cian and nurses was also obtained. Diagnosis was es-tablished on the basis of all available information. Theseverity of disease at the time of evaluation was evalu-ated using the Brief Psychiatric Rating Scale (BPRS)global score, the overall pattern of severity score[Nurnberger et al., 1994], and the score of responsive-ness to neuroleptic medication [Brenner et al., 1990].The control group (C) consisted of healthy volunteers (n4 93) who underwent a full psychiatric interview. Halfof the controls were recruited through advertisementin local papers and were matched to a schizophrenicsubject according to the ethnic background of both par-ents. The other controls were married-in individualsfrom a linkage study. These controls as well as the restof the schizophrenic patients were Caucasians. All con-trols were screened for DSM-IV axis I disorders.

After complete description of the study to the sub-jects, written informed consent was obtained.

Genetic Methods

Using the basic local alignment search tool (BLAST)[Altschul et al., 1990], we identified several CAG-containing expressed sequence tags (EST), includingone EST (GenBank accession number Y08263, locushSAAD14) containing two CAG repeat arrays. ThisEST aligns with a human ortholog of a rat calcium-activated potassium channel gene (U69884). PCRprimers able to amplify the two CAG separately weredesigned using DNASTAR, Inc. (Madison, WI) software(see Table I for primer sequences and PCR conditions).

Genomic DNA was isolated from peripheral lympho-cytes using standard methods. The two CAG repeatstretches were PCR-amplified using two specific pairsof primers (Table I), and the PCR products were elec-trophoresed on denaturing 6% polyacrylamide gels andvisualized by autoradiography.

Absolute allele sizes were estimated according to anM13 (Life Science, Cleveland, OH) sequence ladder.Since differences in absolute allele sizes were in allcases multiples of three base pairs, we assumed thatvariations in allele sizes were due to differences in thenumber of trinucleotide repeat units in the amplifiedsequences. By convention, we designated the most com-mon allele as 0, with less common alleles as positive or

TABLE I. Characteristics of the Two CAG Repeats of the hSKCa3 Gene and PCR Conditions Used to Amplify Each ofThese Repeats*

T°C PNQ Primer sequence No. of alleles H (%)

38 CAG repeat 59 14 58-GGGGTGGGGGACTTGGATGAA-38 13 7758-CAGGGGCTGCTGGGGGACTACT-38

58 CAG repeat 61 12 58-CAGCCTCAGCCTCCGCAGCTTCAG-38 3 158-CAGGCCAGGGTGGACGGACTG-38

*PNQ, number of encoded polyglutamines (as identified in the EST database); T°C, annealing temperature; H, percent of heterozygosity.

hSKCa Gene CAG Repeats and Schizophrenia 155

negative integers according to their number of tri-nucleotide repeats (e.g., if allele 0 had 20 repeats, al-leles +2 and −2 would have 22 and 18 repeats, respec-tively). All gels were read blind to the phenotype andthe response status.

We used the NIGMS mapping panel number 2, ver-sion 3 (Corriel Institute for Medical Research, Camden,NJ) to map the hSKCa3 gene.

Statistical Analysis

Data were analyzed by performing an overall x2 sta-tistic with the appropriate degrees of freedom. We firstcompared allelic frequencies between affected subjectsand normal controls, and then we compared the threegroups of subjects (schizophrenic patients stratified ac-cording to their response to neuroleptic medication andcontrols). These analyses were carried out first by in-cluding all identified alleles, and second by groupingalleles into two categories as reported by Chandy et al.[1998] (equal to or longer than, and shorter than themost common allele, named allele 0). Since both pa-tient and control groups included an important numberof subjects of French Canadian ethnic origin, we reana-lyzed our data after stratifying subjects according tothe ethnic origin of their parents (both parents ofFrench Canadian origin vs. at least one parent of non-French Canadian origin). Because the French Cana-dian population was reported to have a strong foundereffect [Brais et al., 1998], this stratification may in-crease the power of detecting differences between pa-tients and controls in addition to controlling, at leastpartially, for population stratification bias. Finally,correlation between age at onset, responsiveness toneuroleptic medication scores, and overall pattern ofseverity of the disease on the one hand and the CAGrepeat length on the other were investigated using thePearson correlation coefficient. Despite the fact that wetested several CAG repeat-containing genes, a differ-ence at a 5% significance level in the allelic frequenciesbetween patients and controls could be suggestive of atrue association in this study. Indeed, given the re-ported map position of hSKCa3 and its association withschizophrenia in the report of Chandy et al. [1998], aswell as its putative involvement in neural cell function,hSKCa3 is a good candidate gene for schizophrenia.

RESULTS

Table II show the demographic and clinical charac-teristics of the three groups of subjects (Rs, NRs, andC). As expected, the two patient groups were signifi-cantly different according to the severity of psychosis(BPRS scores, F 4 .280, df 4 1, P < 0.001), and thepercent of time spent as an inpatient since their firstcontact with the psychiatric institution (F 4 81, df 4 1,P < 0.001). More pertinent to the context of this studyis the difference in age at first contact with psychiatriccare facilities, with NRs patients having on average, 5years’ earlier contact than Rs patients (F 4 47, df 4 1,P < 0.001).

The 58 CAG repeat was not very polymorphic, withone major (allele 0, 99.4%) and two rare alleles (allele−3, 0.2%, and allele 1, 0.4%). The 38 CAG repeat washighly polymorphic (heterozygosity, 77.0%). Table IIIshows the allelic distribution in affected and normalcontrols. No excess of long alleles was identified inschizophrenic patients (x2 4 16.4, df 4 12, P 4 0.17).No differences were detected when Rs and NRs pa-tients were analyzed as two separate groups (x2 423.9, df 4 24, P 4 0.46). After dichotomizing allelesinto two classes (longer than or equal to and shorterthan allele 0), no differences were observed whetherthe schizophrenic patients were grouped together (x2

4 0.00, df 4 1, P 4 0.97) or separated into Rs and NRsgroups (x2 4 0.00, df 4 2, P 4 0.99). No significantdifferences were found when only subjects having bothparents of French Canadian ethnic background wereanalyzed separately (x2 4 0.058, df 4 1, P 4 0.80).

No correlation between age of onset, and pattern ofseverity of the disease as measured by the percent oftime spent as inpatient or by a scale of overall patternof severity, were detected. In contrast to the map posi-tion reported by Chandy et al. [1997], the hSKCa3 genemapped to chromosome 1, confirming previous resultsfrom Philibert et al. [1998], who mapped the hSKCa3gene to chromosome 1q21–22, using both somatic-cellhybrid panels and the FISH technique.

DISCUSSION

We tested two CAG repeats of the hSKCa3 gene forassociation with two groups of schizophrenic patients

TABLE II. Demographic and Clinical Characteristics of Patients and Controls†

Nonresponsive (63) Responsive (43) Controls (93)

Mean age (years) 38 ± 7 40 ± 10 44 ± 12**Sex, percent M 74 67 47**FC/other 27/36 26/17 43/50Mean age at 1C (years) 18 ± 3.9 (55)* 24 ± 4.8 NAIllness duration (years) 20 ± 7.0 (55) 16 ± 8.8 NAPercent of time as inpatient 62% (61)* 8.2% NABPRS total score 49 ± 8.9 (53)* 24 ± 3.9 NANLP response score 6.3 ± 0.67* 1.83 ± 0.74 (58) NAPattern of severity 4.0 ± 0.5 (55)* 1.9 ± 0.8 NA

†Numbers are mean ± SD. Number of subjects is given in parentheses unless it is equal to the whole sample size.1C, first contact with a care facility; BPRS, Brief Psychiatric Rating Scale; NLP, neuroleptics; FC/others, FrenchCanadian ethnic background/other ethnic backgrounds; M, males. We used one-way analysis of variance forquantitative variables and the chi-square test for categorical variables.*P < 0.05 for R vs. NR comparisons.

**P < 0.05 for R vs. NR vs. C comparisons.

156 Joober et al.

selected on the basis of their long-term response to neu-roleptic medication and outcome. Our results do notsupport the findings of Chandy et al. [1998]; nor did wedetect a correlation between the hSKCa3 CAG repeatlength and severity of disease, whether measured byoverall long-term outcome, responsiveness to neurolep-tic medication, or age at onset. In addition, our resultsare not consistent with the map position of hSKCa3 onchromosome 22q11–13, indicating instead that thisgene maps to chromosome 1.

These results should be interpreted with some cau-tion. Indeed, case-control association studies may bebiased by population stratification (leading to bothfalse-positive and false-negative results), and only halfof our cases were tightly matched to controls with re-gard to ethnicity (the rest being broadly matched).However, when subsamples of patients and controlswho were ethnically homogenous (both parents FrenchCanadian) were analyzed separately, no association be-tween schizophrenia and the CAG repeat in thishSKCa3 gene was detected.

At least two different explanations might account forthese discordant findings. First, the association ob-served in Chandy et al. [1998] may be due to differ-ences in the ethnic composition of controls and schizo-phrenic patients. Though they used controls ‘‘matched’’for ethnic background in the European and Americansamples, it is not known whether their matching wastightly defined or took into account only broadly de-fined ethnic subgroups. Second, their findings could besimply a false-positive result. Consistent with this in-terpretation is the new map position of the hSKCa3gene from Philibert et al. [1998] and our studies, indi-cating that this gene is less likely to be a good candi-date gene on the basis of previous linkage data inschizophrenia.

In conclusion, the present study does not support theinvolvement of the hSKCa3 CAG repeat in schizophre-nia or in its phenotypic variability. Despite the factthat several other studies reported the absence of ex-pansions [Petronis et al., 1996] or allelic associations[Speight et al., 1997] between genes containing poly-morphic CAG repeats and schizophrenia, the numberof CAG repeats in the human genome is estimated to beabout 700 [Gastier et al., 1995]. Therefore, to explorethe hypothesis of the involvement of trinucleotide re-peats in schizophrenia to its fullest extent, additionalCAG and other trinucleotide repeats should be system-atically studied.

REFERENCESAltschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local

alignment search tool. J Mol Biol 215:403–410.

Bassett AS, Husted J. 1997. Anticipation or ascertainment bias in schizo-phrenia? Penrose’s familial mental illness sample. Am J Hum Genet60:630–637.

Brais B, Bouchard JP, Jomphe M, Desjardins B, Dube M-P, Gosselin F, XieY, Tome F, Fardeau M, Brunet G, Samson F, Mathieu J, Prevost C,Duranceau A, Codere F, Pericak-Vance M, Bulman D, Tapscott SFT,Munsat T, Feasby T, Munsat T, Pilarski R, Morgan K, Rouleau GA.1998. When genetics and history converge: The fine-mapping andNorth American introduction and diffusion of the French Canadianoculopharyngeal muscular dystrophy mutation. Am J Hum Genet InPress

Brenner HD, Dencker SJ, Goldstein MJ, Hubbard JW, Keegan DL, KrugerG, Kulhanek F, Liberman RP, Malm U, Midha KK. 1990. Definingtreatment refractoriness in schizophrenia. Schizophr Bull 16:551–556.

Chandy KG, Fantino E, Kalman K, Gutman GA, Gatrgus JJ. 1997. Geneencoding neuronal calcium-activated potassium channel has polymor-phic CAG repeats, a candidate role in excitotoxic neurodegenerationand maps to 22q11–q13, a candidate region for bipolar disease andschizophrenia disorder 4. Am J Hum Genet 61:305 (abstract).

Chandy KG, Fantino E, Wittekindt O, Kalman K, Tong L, Ho T, GutmanGA, Crocq M, Ganguli R, Nimgaonkar V, Morris-Rosendahl DJ, GargusJJ. 1998. Isolation of a novel potassium channel gene hSKCa3 contain-ing a polymorphic CAG repeat: A candidate for schizophrenia and bi-polar disorders? Mol Psychiatry 3:32–37.

Faraone SV, Blehar M, Pepple J, Moldin SO, Norton J, Nurnberger JIX,Malaspina D, Kaufmann CA, Reich T, Cloninger CR, DePaulo JR, BergKX, Gershon ES, Kirch DG, Tsuang MT. 1996. Diagnostic accuracy andconfusability analyses: An application to the Diagnostic Interview forGenetic Studies. Psychol Med 26:401–410.

Gastier JM, Pulido JC, Sunden S, Brody T, Buetow KH, Murray JC, WeberJL, Hudson TJ, Sheffield VC, Duyk GM. 1995. Survey of trinucleotiderepeats in the human genome: Assessment of their utility as geneticmarkers. Hum Mol Genet 4:1829–1836.

Gorwood P, Leboyer M, Falissard B, Jay M, Rouillon F, Feingold J. 1996.Anticipation in schizophrenia: New light on a controversial problem.Am J Psychiatry 153:1173–1177.

Karayiorgou M, Gogos JA. 1997. A turning point in schizophrenia genetics.Neuron 19:967–979.

Kendler KS, Diehl SR. 1993. The genetic of schizophrenia: A current, ge-netic-epidemiologic perspective. Schizophr Bull 19:261–285.

Nurnberger JI Jr, Blehar MC, Kaufmann CA, York-Cooler C, Simpson SG,Harkavy-Friedman J, Severe JB, Malaspina D, Reich T. 1994. Diag-nostic interview for genetic studies. Rationale, unique features, andtraining. NIMH Genetics Initiative. Arch Gen Psychiatry 51:849–859,863–864.

O’Donovan MC, Guy C, Craddock N, Bowen T, McKeon P, Macedo A, MaierW, Wildenauer D, Aschauer HN, Sorbi S, Feldman E, Mynett-JohnsonLX, Claffey E, Nacmias B, Valente J, Dourado A, Grassi E, LenzingerEX, Heiden AM, Moorhead S, Harrison D, Williams J, McGuffin P,Owen MJ. 1996. Confirmation of association between expanded CAG/CTG repeats and both schizophrenia and bipolar disorder. Psychol Med26:1145–1153.

Paulson HL, Fischbeck KH. 1996. Trinucleotide repeats in neurogeneticdisorders. Annu Rev Neurosci 19:79–107.

Petronis A, Bassett AS, Honer WG, Vincent JB, Tatuch Y, Sasaki TX, YingDJ, Klempan TA, Kennedy JL. 1996. Search for unstable DNA inschizophrenia families with evidence for genetic anticipation. Am JHum Genet 59:905–911.

Philibert AR, Horelli-Kuitunen N, Robb AS, Lee Y-H, Long RT, Damsch-roder-Williams P, Martin BM, Brennan M, Palotie A, Ginns EI. 1998.The characterization and sequence analysis of thirty CTG-repeat con-taining genomic cosmid clones. Eur J Hum Genet 6:89–94.

Reiss AL, Feinstein C, Toomey KE, Goldsmith B, Rosenbaum K, CarusoMA. 1986. Psychiatric disability associated with the fragile X chromo-some. Am J Med Genet 23:393–401.

Risch N. 1990. Genetic linkage and complex diseases, with special refer-ence to psychiatric disorders. Genet Epidemiol 7:3–16.

Speight G, Guy C, Bowen T, Asherson P, McGuffin P, Craddock N, OwenMJ, O’Donovan MC. 1997. Exclusion of CAG/CTG trinucleotide repeatloci which map to chromosome 4 in bipolar disorder and schizophrenia.Am J Med Genet 74:204–206.

Tsuang MT, Faraone SV. 1995. The case for heterogeneity in the etiologyof schizophrenia. Schizophr Res 17:161–175.

TABLE III. Distribution of 38 CAG Repeat Alleles of thehSKCa3 Gene*

−7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 T

C 2 8 8 0 6 18 39 74 21 9 1 0 0 186R 0 3 2 1 2 13 16 32 5 9 2 1 0 86NR 0 2 2 1 5 20 25 47 12 10 0 1 1 126

T 2 13 12 2 13 51 80 153 38 28 3 2 1 398

*Alleles are numbered from −7 to −5 according to the number of CAGrepeats above (positive numbers) or below the modal allele (0). C, controls;R, neuroleptic-responsive; NR, nonresponsive patients; T, total.

hSKCa Gene CAG Repeats and Schizophrenia 157