Molecular Psychiatry (2002) 7, 1006–1011  2002 Nature Publishing Group All rights reserved 1359-4184/02 $25.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE The −2 bp deletion in exon 6 of the ‘alpha 7-like’ nicotinic receptor subunit gene is a risk factor for the P50 sensory gating deficit G Raux1, F Bonnet-Brilhault1,2, S Louchart2, E Houy1,2, R Gantier1, D Levillain3, G Allio2, S Haouzir2, M Petit1,2, M Martinez4, T Frebourg1, F Thibaut1,2 and D Campion1

1INSERM EMI 9906, IRFMP, Faculte´ de Medecine et de Pharmacie, Rouen, France; 2Service Hospitalo-Universitaire de Psychiatrie, CHU Charles Nicolle and CHSR, Rouen, France; 3CHSR, Rouen, France; 4INSERM EPI 0006, Evry, France

Keywords: ; ; P50 sensory potential mechanism for the failure to reduce the P50 gating deficit; nicotinic receptor; association; multiplex PCR response.4–6 Furthermore, a genome-wide linkage Abnormality in the P50 auditory-evoked potential gating analysis, using P50 sensory gating deficit as phenotype, is an endophenotype associated with schizophrenia. has shown that this defect was linked to a dinucleotide Biochemical and genetic studies have suggested that polymorphism on chromosome 15q13–14 where the the alpha 7 nicotinic acetylcholine receptor (nAChR) is alpha 7 nicotinic receptor subunit gene (CHRNA7) involved in this sensory gating deficit. Two related alpha maps.7 7 genes (CHRNA7 and CHRNA7-like gene) resulting Two related alpha 7 genes resulting from a partial from a partial duplication (from exon 5 to exon 10) are present in the human genome. Two types of genetic duplication and located approximately 500 kb apart are variation, a large deletion and a −2 base-pair deletion in present on chromosome 15. In both genes, exons 5–10 exon 6 resulting in a truncation of the open reading are nearly identical. The first locus (ie CHRNA7 gene) frame, affect specifically the CHRNA7-like gene. We contains specific sequences corresponding to exons 1– developed a simple multiplex PCR assay on genomic 4 whereas, at the second locus (ie CHRNA7-like gene), DNA, allowing the quantification of the number of exons these sequences are replaced by exons D, C, B and A.8 6 and the distinction of all possible exon 6 genotypes. Recently, it has been reported that in a subset of sub- Genotyping of 70 schizophrenic patients and 77 con- jects the CHRNA7-like gene is missing.9,10 In addition, trols showed that carrying at least one −2 bp deletion of a −2 bp deletion located in exon 6 and presumably spe- exon 6 did not constitute a risk factor for schizophrenia. 8 In contrast, the distribution of genotypes differed sig- cific of the CHRNA7-like gene has also been reported. nificantly between subjects with normal and abnormal To investigate the involvement of these genetic vari- P50 ratios, with an over-representation of genotypes ations in an abnormal P50 inhibition, we developed a carrying at least one −2 bp deletion of exon 6 among quantitative analysis of exon 6 on genomic DNA11 subjects exhibiting an abnormal P50 ratio. We thus con- which allowed us: (i) to determine in each individual clude that the −2 bp deletion within the CHRNA7-like the copy number of exon 6; and (ii) to discriminate gene is a risk factor for P50 sensory gating deficit. Inter- exons 6 carrying the −2 bp deletion. Distributions of estingly, most of the effect came from the non schizo- phrenic group, which may suggest that in schizophrenic genotypes were then compared between schizophrenic patients other risk factors account for the large pro- patients and controls, and between subjects with nor- portion of subjects exhibiting an abnormal P50 ratio. mal and abnormal P50 phenotype. This analysis was Molecular Psychiatry (2002) 7, 1006–1011. performed in a population of 147 subjects (70 schizo- doi:10.1038/sj.mp.4001140 phrenic patients and 77 control subjects). In accord- ance with previous results, 91% of patients and 18% There is convincing evidence that diminished inhibi- of controls had abnormal P50 T/C ratio, respectively. − tory gating of the P50 wave of the auditory-evoked Exons 6 without the 2 bp deletion were counted as − response to repeated auditory stimuli reflects an ‘wild type’ (wt) and exons 6 with the 2 bp deletion ⌬ inherited trait (endophenotype) predisposing to were counted as ‘TG deleted’ ( TG). Six genotypes, ⌬ ⌬ ⌬ schizophrenia. Approximately 90% of schizophrenic corresponding to 2 wt/2 TG, 3 wt/1 TG, 2 wt/1 patients and about 50% of their first-degree relatives TG, 4 wt/0 ⌬ TG, 3 wt/0 ⌬ TG and 2 wt/0 ⌬ TG (Figure fail to gate the auditory-evoked response to paired 1 and Table 1), were identified among the 147 subjects. stimuli, compared to about 10% of normal subjects.1,2 Thus, in a subset of subjects, only three or, more rarely, A strong genetic component is likely to influence this two exons 6 were present instead of four, in accordance trait since monozygotic twins have significantly higher with the recent finding that a large deletion involving intra-pair similarity for the P50 test/conditioning (T/C) the CHRNA7-like locus is prevalent in the general ratio than dizygotic twins.3 Several studies have sug- population, either at the heterozygous or at the homo- gested that nicotinic cholinergic dysfunction is a zygous state.9,10 We found that, in these subjects, the Nicotinic receptor and sensory gating deficit G Raux et al 1007

Figure 1 Multiplex PCR of CHRNA7 and CHRNA7-like genes. Electropherograms of six genotypes, corresponding to 2 wt/2 ⌬TG, 3 wt/1 ⌬TG, 2 wt/1 ⌬TG, 4 wt/0 ⌬TG, 3 wt/0 ⌬TG and 2 wt/0 ⌬TG are displayed. The y axis displays fluorescence in arbitrary units, and the x axis indicates the size of the amplicons in bp. Blue line corresponds to the reference subject. Red line corresponds to the tested subject. All samples were normalized upon an external marker, ie exon 21 of ARVCF gene. The peak identified as exon 6 corresponds to the amplification of wild-type exons 6 of CHRNA7 and CHRNA7-like genes (171 bp); the arrow indicates the amplification of exon 6 carrying the ⌬TG deletion specificofCHRNA7-like gene (169 bp); the peak identified as exon 2 corresponds to the amplification of exon 2 specificofCHRNA7 gene (226 bp); the peak identified as control corresponds to the amplification of exon 21 of ARVCF gene (235 bp).

number of exons 2, which are specific of the CHRNA7 gene, was indeed unaffected. Table 1 Distribution of exon 6 genotypes among schizo- Genotypes were then determined in six subjects from phrenics and controls whom samples were available. Specific PCR- amplification and sequencing of the cDNAs derived Genotypes Control subjects Schizophrenic subjects from CHRNA7 and CHRNA7-like genes, respectively, (n = 77) (n = 70) showed that the −2 bp deletion was specific of the 8 ⌬ CHRNA7-like gene, as previously suggested (data not 4wt/0 TG 18 17 − 3wt/0⌬ TG 13 6 shown). Since the functional significance of the 2bp 2wt/0⌬ TG 1 1 deletion is unknown (ie gain or loss of function at the 3wt/1⌬ TG 22 26 protein level), we used two different classifications to 2wt/2⌬ TG 18 14 define the potential ‘at risk’ genotype. First, we con- 2wt/1⌬ TG 5 6 sidered 2 wt/2 ⌬ TG, 3 wt/1 ⌬ TG, 2 wt/1 ⌬ TG, 3 wt/0 ⌬ TG and 2 wt/0 ⌬ TG genotypes as ‘at risk’ genotype

Molecular Psychiatry Nicotinic receptor and sensory gating deficit G Raux et al 1008 as opposed to 4 wt/0 ⌬ TG genotype. In the second classification, 2 wt/2 ⌬ TG, 3 wt/1 ⌬ TG and 2 wt/1 Table 3 (a) Distribution of exon 6 ⌬ TG genotype among con- ⌬ TG genotypes were considered as ‘at risk’ as opposed trol subjects with normal or abnormal P50 ratio. (b) Distri- to 4 wt/0 ⌬ TG, 3 wt/0 ⌬ TG and 2 wt/0 ⌬ TG geno- bution of exon 6 ⌬ TG genotype among schizophrenic subjects types. Whatever the classification used no association with normal and abnormal P50 ratio was detected with the schizophrenic phenotype ⌬ (␹2 = 0.01, df = 1, NS, and ␹2 = 0.82, df = 1, NS, a TG genotype Wild-type genotype respectively). Ͼ We then searched for an association between the Ratio 0.50 12 2 = abnormal P50 ratio phenotype and the two genotypic (n 14) classifications indicated above. No association was Ratio Ͻ0.45 33 30 = detected using the first genotypic classification (n 63) (␹2 = 1.95, df = 1, NS). In contrast, when the second ␹2 = Ͻ classification was used (ie presence of at least one −2 5.24, 1 df, P 0.03. bp deletion, referred as ‘⌬ TG genotype’ hereafter, as ⌬ opposed to 4 wt/0 ⌬ TG, 3 wt/0 ⌬ TG and 2 wt/0 ⌬ TG b TG genotype Wild-type genotype genotypes, defined as ‘wild-type’ genotype hereafter), Ͼ distributions of the two groups of genotypes were sig- Ratio 0.50 44 20 = nificantly different (␹2 = 6.01, df = 1, P Ͻ 0.02) between (n 64) subjects with normal and abnormal P50 ratio (Table 2). Ratio Ͻ0.45 3 3 = The corresponding odds ratio (OR) (⌬ TG genotype vs (n 6) wild-type genotype) was 2.33 (95% CI = 1.18–4.62). ␹2 = The ⌬ TG genotype was present in 72% of subjects 0.23, 1 df, NS. with abnormal P50 ratio compared to 52% of subjects with normal P50 ratio. We next dichotomized each in a premature stop codon, affect specifically the group of subjects (ie subjects with abnormal and nor- CHRNA7-like gene. We showed that the presence of at mal P50 ratio) according to their clinical status least one –2 bp deletion (⌬ TG genotype) in the (schizophrenic or control subject) and compared the CHRNA7-like gene does not constitute a risk factor for two classes of genotypes within each subgroup (Table schizophrenia by itself, but is significantly associated 3a and b). While the two subgroups were of similar with P50 sensory gating deficit. size, the association between the ⌬ TG genotype and Since failure to gate the P50 auditory evoked poten- P50 abnormality was statistically significant (␹2 = 5.24, tials is strongly associated with schizophrenia and df = 1, P Ͻ 0.03) in the controls subgroup, whereas in since the del genotype is associated with this endo- the schizophrenics subgroup only a slight trend was phenotype, we should have detected in our sample an detected (␹2 = 0.23, df = 1, NS). The corresponding ORs association between the ⌬ TG genotype and schizo- (⌬ TG genotype vs wild-type genotype) in each sub- phrenia. To analyze this apparent discrepancy, the group were OR = 5.45 (95% CI = 1.13–26.39) and respective contribution of schizophrenic subjects and OR = 2.20, (95% CI = 0.41–11.87), respectively. How- non schizophrenic controls was examined for each ever, although most of the effect detected in the entire class of genotype and for each class of phenotype. The group was contributed by the non schizophrenic sub- proportion of individuals bearing the ⌬ TG genotype group, the ␹2 test for heterogeneity between these ORs and exhibiting the abnormal P50 ratio was increased was non significant (␹2 = 0.59, df = 1, NS). in the schizophrenic subgroup as compared to that Our data support previous reports that two types of found in the control subgroup. But, simultaneously, genetic variation, a large genomic deletion (as charac- the proportion of schizophrenic subjects without the terized in the present study by the entire exon 6 ⌬ TG genotype who exhibited the abnormal P50 deletion) and a −2 bp deletion in exon 6 which results phenotype was also increased (even to a greater extent) compared to that found in controls. As a result, the association between the ⌬ TG genotype and the P50 Table 2 Distribution of exon 6 ⌬ TG genotype among 147 abnormal ratio was more difficult to detect in the subjects with normal and abnormal P50 ratio regardless of schizophrenic subgroup. Despite this difficulty, no sig- their clinical status nificant difference in odds ratio values (OR ⌬ TG geno- type vs wild-type genotype) was detected among the ⌬ TG genotype Wild-type genotype two subgroups by the ␹2 test for heterogeneity, suggest- ing that the ⌬ TG genotype is a risk factor for P50 Ratio Ͼ0.50 56 22 abnormality, whatever the clinical status of the subject. (n = 78) However, we have to be cautious with this conclusion, 2 Ratio Ͻ0.45 36 33 since, given our sample size, the ␹ test for heterogen- (n = 69) eity probably lacks power to detect a true heterogen- eity. ␹2 = 6.01, 1 df, P Ͻ 0.02. In control subjects, the probability to exhibit the P50 abnormality given the ⌬ TG genotype (ie the pen-

Molecular Psychiatry Nicotinic receptor and sensory gating deficit G Raux et al 1009 etrance of the ⌬ TG genotype) was mild (27%) and this sent was obtained from every individual. P50 auditory- probability given the wild-type genotype was low evoked potentials were recorded in all subjects. (6%). In contrast, in schizophrenic patients the prob- ability to exhibit the P50 abnormality given the ⌬ TG Electrophysiological recording genotype was very high (94%) as well as the prob- Subjects were in a supine position, relaxed but awake, ability associated with the wild-type genotype (87%). with head and neck well supported on pillows. They These results suggest, first, that a protective factor were asked to refrain from smoking for half an hour (either genetic or environmental) which reduces the before the recording. Auditory stimuli were delivered penetrance of the ⌬ TG genotype in controls is lacking through headphones in a conditioning-testing para- in schizophrenic subjects. Secondly, the probability to digm consisting of click pairs (S1, conditioning click; exhibit the abnormal P50 phenotype due to another S2, testing click; 500 ms interclick interval) delivered cause than the CHRNA7-like genotype abnormality is with an interpair interval of 10 s. Clicks were produced greatly increased in schizophrenic subjects. Together, by square wave pulse of 100 ␮s duration and amplified these results show that the ⌬ TG genotype is a major to reach the intensity of 75 dB sound pressure level at risk factor for P50 abnormality in the general popu- the subject’s ear. The subject was allowed to accommo- lation. In schizophrenics, the intrinsic weight of this date to the experimental situation for 5 min without factor is lower, and other factors do share a part of recording any data. Then, three sets of 30 stimulus responsibility in P50 abnormality. Other genetics vari- pairs were delivered. Electroencephalographic activity ations, affecting either the CHRNA7-like gene or other was monitored and averaged on a Nihon Kohden com- genes in biological pathways controlling this endo- puter. EEG was recorded from a disk Ag/AgCl electrode phenotype, are likely to be involved besides the affixed to the vertex (Cz, International 10-20 system), ⌬ TG genotype. referenced to linked ears. A ground electrode was The mechanisms by which the −2 bp deletion in affixed on the right wrist. Horizontal and vertical elec- exon 6 of the CHRNA7-like gene confers the risk trooculograms were also recorded. Electrode imped- remain to be established. However, our results argue ance was less than 5000 ohms. Electrical activity was against a loss of function of the mutant truncated pro- amplified with a bandpass filter of 1–200 Hz and digit- tein, since in this case the statistical evidence for the ized at 1000 Hz for averaging by a digital computer. All association would have been greater under the genetic trials contaminated by ocular movements and move- classification which includes the large deletion of the ment artefacts were automatically rejected (criterion = CHRNA7-like gene within the ‘at risk’ genotype. Since 40 ␮V). A single average was compiled from the three the converse is true, it could thus be hypothetized that sets of the 30 artifact-free trials. The conditioning P50 the −2 bp deletion does not act by a reduction in gene was identified as the most positive peak between 40 dosage but generates a mutant protein acting by a gain and 80 ms after the first stimulus. The test P50 was of function or by a dominant negative effect. Specifi- identified at the same latency, plus or minus 10 milli- cally, we speculate that the truncated protein might seconds.13 Latencies were measured as the time affect the assembly of the alpha 7 pentamer. Alterna- between stimulus onset and peak of the component. tively, we cannot exclude that the −2 bp deletion in Amplitudes were measured as the difference between exon 6 of the CHRNA7-like gene might be in linkage the peak of the P50 wave and the immediately preced- disequilibrium with the true causative DNA polymor- ing negative peak. The amplitude and the latency of phism. the P50 wave were measured in both the conditioning and test responses. Test/conditioning ratios (T/C) were calculated by dividing the test P50 amplitude by the Materials and methods conditioning P50 amplitude. P50 ratios lower than 0.45 Patients were considered as normal whereas P50 ratios higher Unrelated schizophrenic patients (51 males and 19 than 0.50 indicated diminished inhibitory processing. females, aged 38 ± 10.9 years) were included in this Subjects with values between 0.45 and 0.50 were not study. Diagnoses were made according to DSM-IIIR cri- included in the study. teria by two trained psychiatrists after clinical inter- views following the French version of the Schedule for Genotype analysis Affective and Schizophrenia Disorders Lifetime Ver- To determine in each individual the number of exons sion (SADS-LA). A control group was composed of 37 6 and to discriminate exons carrying the −2bp males and 40 females, (aged 39 ± 11.7 years) free of deletion, we devised a multiplex PCR assay of fluor- neurological or psychiatric diseases, and who had no escent fragments based on the approach that we family history of psychosis or neurodegenerative dis- initially developed for the detection of mismatch repair orders. Controls were free of any psychotropic treat- gene rearrangements in hereditary non polyposis col- ment. All patients received a neuroleptic treatment but orectal cancer11 and which is based on: (a) the amplifi- none had atypical neuroleptic medications, which, in cation of short fragments to reduce the difference of contrast to conventional neuroleptic treatments, may efficiency between each amplification; (b) a limited normalize P50 measures.12 All schizophrenic patients number of cycles to allow an exponential amplifi- and controls were French Caucasians originating from cation; (c) the comparison of fluorescence between the Normandy (west part of France). Written informed con- same peak generated from two samples corresponding

Molecular Psychiatry Nicotinic receptor and sensory gating deficit G Raux et al 1010 to the tested subject and to a reference subject (and not (Pharmacia) and was resuspended into 200 ␮l of water. between different peaks generated from the same Random hexamer primed cDNAs were synthesized sample). Each sample was compared with the reference from 20 ␮l mRNA in a final volume of 33 ␮l using the sample using an external marker to normalize the dif- First-Strand cDNA Synthesis Kit (Pharmacia). ferent electropherograms. CHRNA7 and CHRNA7-like cDNAs were then ampli- Short fragments of exon 2 of the CHRNA7 gene fied from 5 ␮l of the cDNA reaction using as sense pri- (sense primer: 5Ј-GGACAAGGAGCTGGTCAAGAACT- mers, 5Ј-AAGGCGAGTTCCAGAGGAAG-3Ј (exon 2) 3Ј; Antisense primer: 5Ј-CTTGGGGCCAACTA- and 5Ј-AAGTGGACACCTGAGTCAGCA-3Ј (exon C), GAGTGC-3Ј; size of the amplicon: 226 bp), exons 6 of respectively, and as antisense primer, 5Ј-CCAGGGAAA the CHRNA7 and CHRNA7-like genes (sense primer: TCTTCTCCCCGGAA-3Ј (exon 7). PCR was performed 5Ј-GGGCATATTCAAGAGTTCCTGCTAC-3Ј; Antisense in a final volume of 20 ␮l containing 0.5 ␮M of primer, Ј Ј primer: 5 -CCACTAGGTCCCATTCTCCATTG-3 ; size 0.2 mM dNTP, 1.5 mM MgCl2 and 1.5 units of Taq of the amplicon: 171 bp) and exon 21 of the ARVCF polymerase (Eurobio, Les Ulis, France). The PCR con- gene (sense primer: 5Ј-GACATGGTGCTGTGTGTG sisted of 32 cycles of 20 s at 94°C, 20 s at 55°C, and 1 AGC-3Ј; Antisense primer: 5Ј-GGTCCGCCTTTA- min at 72°C, preceded by 3 min at 94°C and followed GAAGTCCAAGT-3Ј; size of the amplicon: 235 bp) were by 5 min at 72°C. One ␮l of the initial PCR was then simultaneously PCR-amplified using sense primers lab- diluted in 100 ␮l of water, and 1 ␮l of the diluted tem- elled with the fluorescein dye 6-FAM from 100 ng of plate was submitted to a nested PCR using as sense genomic DNA extracted from peripheral blood lym- primers, for the CHRNA7 and CHRNA7-like cDNAs, 5Ј- phocytes using a standard proteinase K protocol. PCR TCATTGTGTGTGTCAGTCTTGG-3Ј (exon 4) and 5Ј- was performed in a final volume of 25 ␮l containing AAATATTGCATCTACCAGCA-3Ј (exon B) respect- ␮ Ј 0.3 M of each primer, 0.2 mM dNTP, 1.5 mM MgCl2, ively, and as antisense primer 5 -CCACTAGGTCCC and 1 unit of Thermoprime plus Taq polymerase ATTCTCCATTG-3Ј (exon 6). Dilution of the CHRNA7- (Abgene, UK). After a 3 min denaturation at 95°C, the like amplification product was amplified using a sense PCR consisted of 23 cycles of 15 s at 94°C, 15 s at 58°C primer in exon B and an antisense primer in exon 6 and 15 s at 72°C followed by a final extension of 7 min (5Ј-CCACTAGGTCCCATTCTCCATTG-3Ј). The PCR at 72°C. The PCR product (1 ␮l) was resuspended in a was performed as indicated above except that the mix containing 2.5 ␮l of deionised formamide, 0.5 ␮l annealing temperature was 50°C. PCR products were of GeneScan-500 Rox (PE Applied Biosystems, War- then purified by electrophoresis on low-melt agarose rington, UK) and 1 ␮l of loading buffer. After denatur- gel and directly sequenced on both strands using the ation for 2 min at 90°C, 2 ␮l was loaded on a 6% denat- Big Dye Terminator Kit (Applied Biosystems) and a uring polyacrylamide gel. Electrophoresis was model 377 automated sequencer (Applied Biosystems). performed for 2 h on an Applied Biosystems model 377XL automated sequencer (Applied Biosystems) with a 12 cm well to read distance, and data were ana- Acknowledgements lyzed using the Genescan Analysis v3.1 fluorescent The authors thank Dr A Delacourte for brain samples. fragment analyser (Applied Biosystems). The fluor- This study was supported by grants from MNERT pro- escence of each peak was compared to that observed jet cognitique, Pfizer, and Fondation de la Recherche in a reference subject which has been shown to have a Me´dicale (to EH). 2 alleles status for the ARVCF gene (data not shown). To standardize the intensity of fluorescence peaks from the tested subject relative to the reference subject, elec- References trophoregrams were superimposed and the peaks corre- 1 Freedman R, Adler LE, Waldo M. Gating of the auditory evoked sponding to an external marker (ie the exon 21 of the potential in children and adults. Psychophysiology 1987; 24: ARVCF gene located on chromosome 22) were set to 223–227. 2 Siegel C, Waldo M, Mizner G, Adler L, Freedman D. Deficit in the same value. sensory gating in schizophrenic patients and their relatives. Arch Gen Psych 1984: 41: 607–612. Statistical methods 3 Myles-Worsley M, Coon H, Byerley W, Waldo M, Young D, Freed- Genotypes distributions were compared using a ␹2 test man R. Developmental and genetic influences on the P50 sensory gating phenotype. Biol Psychiatry 1996; 39: 289–295. with Yates correction if necessary. Odds ratios (OR) 4 Adler LE, Rose GM, Freedman R. Neurophysiologic studies of sen- were calculated and compared using a chi-square test sory gating in rats: effects of amphetamine, phencyclidine and hal- of heterogeneity, according to the Mantel–Haenszel operidol. Biol Psychiatry 1986; 21:787–798. method. 5 Luntz-Leybman V, Bickford PC, Freedman R. Cholinergic gating of response to auditory stimuli in rat . Brain Res 1992; 587:130–136. RT-PCR of the CHRNA7 and CHRNA7-like genes 6 Freedman R, Hall M, Adler LE, Leonard S. Evidence in post- mortem brain tissue for decreased numbers of hippocampal nic- Samples of frozen frontal cortex (0.5 mg) from six nor- otinic receptors in schizophrenia. Biol Psychiatry 1995; 38:22–23. mal human post-mortem were placed in 400 ␮l 7 Freedman R, Coon H, Myles-Worsley M, Orr-Urtreger A, Olincy A, of RNA lysis buffer (Pharmacia Biotech, Uppsala, Davis A et al. Linkage of a neurophysiological deficit in schizo- − ° phrenia to a chromosome 15 locus. Proc Natl Acad Sci USA 1997; Sweden) and stored at 20 C. mRNA was extracted 94:587–592. using the QuickPrep Micro mRNA purification kit 8 Gault J, Robinson M, Berger R, Drebing C, Logel J, Hopkins J et al.

Molecular Psychiatry Nicotinic receptor and sensory gating deficit G Raux et al 1011 Genomic organization and partial duplication of the human alpha7 12 Light GA, Geyer MA, Clementz BA, Cadenhead KS, Braff DL. Nor- neuronal nicotinic acetylcholine receptor gene (CHRNA7). Genom- mal P50 suppression in schizophrenia patients treated with atypi- ics 1998; 52:173–185. cal antipsychotic medications. Am J Psychiatry 2000; 15:767–771. 9 Weiland S, Bertrand D, Leonard S. Neuronal nicotinic acetylcho- 13 Nagamoto HT, Adler LE, Waldo MC, Freedman R. Gating of audi- line receptors: from the gene to the disease. Behav Brain Res 2000; tory response in schizophrenics and normal controls: effects of rec- 113:43–56. ording site and stimulation interval on the P50 wave. Schizophr 10 Riley BP, Makoff A, Mogudi-Carter M, Jenkins T, Williamson R et Res 1991; 4:31–41. al. Haplotype transmission disequilibrium and evidence for the linkage of the CHRNA7 gene region to schizophrenia in Southern Correspondence: D Campion, INSERM EMI 9906, IRFMP, Faculte´ de African Bantu families. Am J Med Genet 2000; 96: 196–201. Medecine et de Pharmacie, 22 bd Gambetta, 76183 Rouen, France. E- 11 Charbonnier F, Raux G, Wang Q, Drouot N, Cordier F, Limacher mail: [email protected] JM et al. Detection of exon deletions and duplications of the mis- Received 3 December 2001; revised 12 February 2002; accepted 14 match repair genes in hereditary nonpolyposis colorectal cancer February 2002 families using multiplex polymerase chain reaction of short fluor- escent fragments. Cancer Res 2000; 60: 2760–2763.

Molecular Psychiatry