Molecular Psychiatry (2005) 10, 1097–1103 & 2005 Nature Publishing Group All rights reserved 1359-4184/05 $30.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE A haplotype within the DISC1 is associated with visual memory functions in families with a high density of W Hennah1, A Tuulio-Henriksson2, T Paunio1,2,3, J Ekelund1,2, T Varilo1,4, T Partonen2, TD Cannon5, JLo¨nnqvist2,3 and L Peltonen1,4,6 1Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland; 2Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland; 3Department of Psychiatry, Helsinki University Central Hospital, Helsinki, Finland; 4Department of Medical Genetics, University of Helsinki, Helsinki, Finland; 5Departments of Psychology, Psychiatry and Biobehavioral Sciences, UCLA, CA, USA; 6The David Geffen School of Medicine at UCLA, CA, USA

We have previously reported evidence of linkage and association between markers on 1q42 and schizophrenia in a study sample of 498 multiply affected Finnish nuclear families, leading to the recent identification of four significantly associated haplotypes that specifically implicate the Translin-Associated Factor X (TRAX) and Disrupted in Schizophrenia 1 and 2 (DISC1 and DISC2) in the genetic etiology of schizophrenia. Previously, the DISC genes were found to be disrupted by a balanced translocation (1;11)(q42.1;q14.3) that cosegregated with schizophrenia and related disorders in a large Scottish pedigree. Interestingly, we also reported earlier suggestive linkage between endophenotypic quantitative traits of visual and verbal memory and microsatellite markers in close proximity to TRAX/DISC, on 1q41. Here, we tested if the identified allelic haplotypes of TRAX/DISC would be associated with visual and/or verbal memory function impairments that are known to aggregate with schizophrenia in families. One haplotype of DISC1, HEP3, displayed association with poorer performance on tests assessing short-term visual memory and attention. Analysis of affected and unaffected offspring separately revealed that both samples contribute to the observed association to visual . These results provide genetic support to the view that the DISC1 gene contributes to sensitivity to schizophrenia and associated disturbances and affects short-term visual memory functions. This finding should stimulate studies aiming at the molecular characterization of how the specific alleles of DISC1 affect the visual memory functions and eventually participates in the development of schizophrenia. Molecular Psychiatry (2005) 10, 1097–1103. doi:10.1038/sj.mp.4001731; published online 16 August 2005 Keywords: QTL; Finland; DISC1; schizophrenia

Schizophrenia is a severe mental disorder with a consistently observed to be significantly undertrans- strong genetic component,1–3 inherited in a complex mitted to affected individuals. HEP3 also displayed manner. Recently, specific alleles of numerous genes a sex difference in transmission distortion.6 This have been introduced as potential etiological factors association between DISC1 and schizophrenia and of schizophrenia using haplotype association meth- related disorders has recently been replicated.7 How- ods.4,5 On 1q42, haplotypes of a containing ever, another study evaluating promoter variants of three such schizophrenia candidate genes, Translin- DISC1 failed to detect association with schizophre- Associated Factor X (TRAX), and Disrupted in nia8 (Figure 1). Schizophrenia 1 and 2 (DISC1 and DISC2), were Initially, linkage analysis had identified 1q42 as a observed to be associated with schizophrenia. Speci- good candidate region for the disorder, when a fically, four different haplotypes (assigned as HEP1– balanced translocation (1;11)(q42.1;q14.3), directly HEP4) were identified in 498 multiply affected disrupting the DISC1 and 2 genes, was observed to Finnish nuclear families with HEP2 and 3 being be linked to schizophrenia and its related disorders in a large Scottish pedigree.9,10 Later, it was shown that a DISC1 intragenic microsatellite marker was linked to Correspondence: Dr L Peltonen, Department of Medical Genetics, schizophrenia in a large Finnish family sample,11 and University of Helsinki, Biomedicum/KTL, MLO, PL 104, 00251, the linkage was replicated with another DISC1 Helsinki, Finland. intragenic marker in an independent Finnish sam- E-mail: [email protected] 12 Received 25 March 2005; revised 6 July 2005; accepted 19 July ple. Observations of linkage to 1q42 to schizophre- 2005; published online 16 August 2005 nia have since been observed in the Taiwanese,13 and DISC1 and quantitative neurocognitive traits W Hennah et al 1098

Figure 1 Schematic diagram of the approximate 600 kb region of 1q42 where the TRAX and DISC genes are located. The diagram shows the exonic structure of the TRAX (gray boxes) and DISC (black boxes) genes, as well as the location of the intragenic exons (white boxes). Above the schematic are marked the linkage findings, and below are marked the association studies, with both positive (black line) and no association (dashed line) findings. The haplotypes from Hodgkinson et al7 referred to here as Hap1 to Hap4 are the haplotypes they refer in their article to as haploblocks.

to in a UK and Icelandic sample14 most telomeric of these markers (ie those immediately (Figure 1). The suggested functions of the TRAX and proximal to TRAX/DISC) displayed the strongest DISC genes also make them highly relevant for the linkage and association to quantitative measures of etiology of psychotic disorders. The DISC1 is visuospatial working memory and visual attention.24 implied to play a role in neurite outgrowth, neuronal Also, a family-based, genome-wide scan using a migration, synaptogenesis, and glutamatergic neuro- number of quantitative traits of cognitive functions transmission.15–20 The TRAX protein forms brain- observed some evidence of linkage (lod41.5) to enriched complexes with a protein called Translin, verbal memory functions at the 1q41 locus25 (Figure which is vaguely thought to be involved in the 1). Here, we have tested linked quantitative traits of development and function of the central nervous cognitive functions for association with the consis- system.21 The DISC2 gene has been hypothesized to tently associated haplotypes of 1q42 in Finnish act through its RNA as a regulator of DISC1.9 schizophrenia families. Although multiple findings imply that TRAX/DISC may be involved in schizophrenia, there is currently little data to support any specific etiological model. Materials and methods One approach to address this issue is to use intermediate phenotypes, endophenotypes, of the Study sample clinical diagnosis with a possibly more direct rela- This study used a subsample of families from the tionship to specific molecular pathways. Any mea- original association analysis on 1q42.6 The sample surable quantitative endophenotype of schizophrenia consists of 215 families containing 1437 individuals should represent a heritable state-independent trait, of which 400 are classified as affected according to revealing variations associated with a vulnerability to increasingly inclusive liability classes (LC) with the disorder. Their molecular background is likely to criteria based on the Diagnostic and Statistical be more straightforward, with less genetic hetero- Manual of Mental Disorders, fourth edition (DSM- geneity contributing to the variation in the trait, when IV).26 LC1 constitutes schizophrenia only, LC2 added compared to the diagnostic phenotype of schizophre- those individuals affected with schizoaffective dis- nia. For schizophrenia, the most prominent candidate order, LC3 added individuals with schizophrenia endophenotypes are measures of the structure and spectrum disorder,27 and LC4 added individuals with function of a number of brain systems.22 bipolar disorder or major depressive disorder. From Linkage studies of Finnish schizophrenia samples these 1437 individuals, 746 have undergone extensive have highlighted the 1q41–q42 region as a potential neuropsychological assessment, including 356 off- candidate locus for cognitive endophenotypes. A study spring currently unaffected under LC4 (Table 1). The of Finnish twins used microsatellite markers that 746 assessed individuals have an average age of 50.2 had previously been shown to form a schizophrenia years (SD, 11.6), with 356 female and 390 male linked allelic haplotype spanning 1q32.2–q41.23 The subjects.

Molecular Psychiatry DISC1 and quantitative neurocognitive traits W Hennah et al 1099 Phenotypic assessment respectively. Semantic clustering and intrusions were The neuropsychological test battery, from which the logarithmically transformed to normality for all cognitive traits included in this present study were analyses. obtained, has been described elsewhere in detail.28 All tests were well validated, internationally used Laboratory and statistical methods neuropsychological instruments, and include the All single nucleotide polymorphisms (SNPs) had 29 been previously genotyped as part of the original Wechsler Memory Scale-Revised (WMS-R), Califor- 6 nia Verbal Learning Test (CVLT),30 and the Wechsler analysis, in which association was observed for four Adult Intelligence Scale-Revised (WAIS-R).31 The separate haplotypes to LC4, a broad phenotype that tests were administered to the subjects in a fixed predominantly consists of schizophrenia, but also order by experienced psychologists or psychiatric includes individuals affected with other schizophre- nurses who had received extensive training with the nia spectrum diagnoses as well as affective disorders. test battery. All scoring was performed by experi- Initial analysis of this smaller sample tested the level of association between LC4 and the haplotypes, and enced psychologists. In the present study, five 34 neuropsychological test variables were chosen for was performed using the TRANSMIT program. This analysis: visual attention, visual working memory, is a program that tests for transmission distortion of a verbal learning, using semantic clusters as a learning haplotype even when phase is unknown and in the strategy, and performing intrusive recall errors (Table presence of missing genotype information. 2). These variables have previously shown to be In our further analysis, we wanted to directly test strong endophenotypic traits of schizophrenia:32,33 for transmission distortion between a haplotype and a the visual span forward and backward subtests quantitative trait. This meant that it was necessary to of the WMS-R were used to assess visual attention construct the haplotypes for each individual in the and visual working memory, respectively. Verbal analysis. There is currently no method available for learning and memory, using semantic clusters as a reliably performing this on data that consist of multiple tightly linked SNPs within family-based learning strategy, and the intrusive recall errors score, 35 all assessed by the CVLT, have been abbreviated to samples. We used the Simwalk2 program, a Markov verbal learning, semantic clusters, and intrusions, chain Monte Carlo (MCMC) and simulated annealing program for haplotype analysis, and in order to increase the reliability of the predicted haplotypes, we used all possible markers that had been genotyped Table 1 Breakdown of sample showing numbers of indivi- in the 1q42 region, which included 28 SNPs and four duals for which neuropsychological test variables and 1q42 microsatellite markers, and used our entire sample of genotype information were available 498 nuclear families. The frequencies of the resulting haplotypes were identical to those estimated from the Total Neuropsycho- Genotyped Assessed genotypes by TRANSMIT in the original analysis. The logically and haplotypes were recoded to form ‘biallelic markers’, assessed genotyped so as to test the hypothesized haplotype allele against all others combined, for use within the QTDT Total 1437 746 890 612 program.36 This is a variance component method for Parents 444 104 181 90 Affected 368 286 321 249 testing for transmission distortion of an allele with a offspring quantitative trait. In this study, age, sex, and affection Unaffected 625 356 388 273 according to LC4 were used as covariates, and 100 000 offspring permutations were performed to derive the empirical P-values.

Table 2 Descriptive statistics for the five cognitive traits within the sample used in this analysis

Mean Standard deviation Kurtosis Skewness

Variable A B C A B C A B C A B C

Visual attention 7.9 7.5 8.4 1.8 1.9 1.7 0.4 0.7 0.2 0.2 0.2 0.3 Visual working memory 7.1 6.6 7.8 2.2 2.3 1.9 0.3 0.1 À0.2 À0.5 À0.5 À0.2 Verbal learning 42.3 35.9 48.5 13.7 13.6 11.3 À0.7 À0.6 À0.6 À0.2 0.1 À0.3 Semantic clusteringa 2.3 1.9 2.6 0.8 0.8 0.6 0.5 0.2 0.4 À0.6 À0.5 À0.4 Intrusionsa 1.3 1.5 1.1 1.0 1.0 0.9 À1.0 À0.9 À1.0 0.1 À0.0 0.2

A: descriptive statistics of whole sample; B: descriptive statistics of offspring affected under LC4; C: descriptive statistics of offspring currently unaffected under LC4. aTrait logarithmically transformed.

Molecular Psychiatry DISC1 and quantitative neurocognitive traits W Hennah et al 1100 Results association to any of the five traits tested (data not shown). After applying the conservative Bonferroni We tested association of two specific allelic haplo- correction to account for the multiple testing by types of TRAX/DISC with the following quantitative analyzing two haplotypes with five neurocognitive traits: visual attention, visual working memory, verbal traits (10 separate tests), only the association between learning, semantic clustering, and intrusive recall HEP3 and visual working memory remained signifi- errors. These were the most meaningful quantitative cant to the 0.05 level. traits for analysis as they each had previously In order to determine whether these associations displayed linkage to multiallelic markers on a wide are primarily to the trait itself or rather reflect the genomic region, 1q41,24,25 proximal to the TRAX/ effect of schizophrenia on the trait distribution, we DISC genes. The allelic haplotypes were selected as analyzed the association between HEP3 and the two they were observed to associate with schizophrenia in traits of visual memory functions separately for Finnish families. Analysis was carried out in a sample affected and unaffected offspring. The association to of 215 Finnish families consisting of 1437 indivi- the visual attention trait was only observed with duals, of which 746 were assessed with a neuropsy- HEP3 in the affected sample (P ¼ 0.0025) compared chological test battery (Tables 1 and 2). These families to the unaffected (PX0.10). However, for visual represent a subsample of the families originally used working memory, both affected and unaffected in the detection of association between TRAX/DISC offspring contributed approximately equally to the and schizophrenia.6 In order to examine the effect of observed association in the combined study sample, the reduced sample size on our ability to detect the P-values 0.034 and 0.048, respectively. Additionally, original association with these haplotypes, we per- the HEP3 haplotype displayed sex differences in its formed an analysis using the same TRANSMIT34 transmission distortion, only being significantly program and specifications as in the original finding,6 undertransmitted to affected female subjects, in within this extensively phenotyped sample. Despite the original study. In order to investigate if similar the reduction in sample size, the association signal, in sex-dependent effects were present in the association relation to schizophrenia diagnosis, for both HEP2 to these quantitative traits, we repeated the permu- and HEP3 remained at the suggestive level (P-values: tated QTDT analysis with HEP3, once only including HEP2, 0.0082; HEP3, 0.072) (Table 3). male offspring and then only including female off- We next performed association analysis of the HEP2 spring, mimicking the process by which the original and HEP3 haplotypes using the quantitative neurop- sex differences were observed with the TRANSMIT sychological test variables and the QTDT program.36 program. The association to poorer performance This is a variance component analysis method that in the visual working memory trait was only sig- tests for significant transmission distortion within nificant in male subjects (male subjects, P ¼ 0.0060; families for a quantitative trait. Age, sex, and affection female subjects, P40.10), yet for visual attention status were used as covariates and 100 000 permuta- the association was equally significant in both tions were performed in order to derive the empirical sexes (male subjects, P ¼ 0.025; female subjects, P-values. The haplotype HEP3 displayed a more P ¼ 0.019). significant association to poorer performance in the traits of visual attention and working memory, with P- values 0.0079 and 0.0013, respectively (Table 3), than Discussion to schizophrenia, but did not show association to any of the three traits of verbal learning and memory Here, we have tested whether specific allelic forms of (data not shown). The haplotype HEP2 showed no the TRAX/DISC genes, reflected by regional haplotypes

Table 3 Empirical P-value results from analysis of the two haplotypes and LC4 or the five quantitative traits used in this study

HEP2 HEP3 HEP3a HEP3b HEP3c HEP3d

Original finding6 Clinical diagnosise o 0.00001 0.0031 0.38 0.00024 NA NA Current study sample Clinical diagnosise 0.0082 0.072 0.34 0.096 NA NA Visual attentionf 40.10 0.0079 0.025 0.019 0.0025 40.10 Visual working memoryf 40.10 0.0013 0.0060 40.1 0.034 0.047

aAnalysis using only male offspring. bAnalysis using only female offspring. cAnalysis using only offspring affected under LC4. dAnalysis using only offspring currently unaffected under LC4. ePerformed using TRANSMIT34 with 100 000 permutations. fPerformed using QTDT36 with 100 000 permutations.

Molecular Psychiatry DISC1 and quantitative neurocognitive traits W Hennah et al 1101 previously observed to be significantly associated some degree of convergence in implicating the TRAX/ with schizophrenia, are also associated with quanti- DISC loci as associated with visuospatial working tative traits of cognitive function known to be memory and schizophrenia. Owing to different deficient in individuals with schizophrenia. The combinations of SNP markers in this region showing neuropsychological test data were available for 215 different patterns of associations with end state families. Thus, the study sample used for this diagnosis and quantitative traits, further work is analysis was less than half the size of the original needed to isolate the specific disease-related var- sample (498 nuclear families) resulting in a reduced iant(s) of these genes. power for the detection of association, given no In the Cannon et al (in press) study of 118 Finnish difference in phenotypic information content. We twin pairs, another haplotype of markers in the tested if the initially observed association of HEP2 translocation breakpoint region of DISC1 (HEP1) and 36 could be detected within the reduced sample seemed to be associated with schizophrenia and for the dichotomized trait of schizophrenia. Both impaired long-term verbal memory functions (Figure haplotypes still displayed some marginal signifi- 1). Despite being unable to support these findings, it cance. It can be hypothesized that the use of more should be emphasized that rather than having informative quantitative traits, if truly associated with identified a true ‘predisposing’ variant, we provide clinical vulnerability and the studied gene variations, evidence for an altered distribution of DISC1 alleles has the potential to display stronger association. We in individuals showing impaired visual working observed such a significant increase in statistical memory, when we used a hypothesis-driven approach significance for the associations involving one haplo- to our analysis. This reduced the risk of type 2 errors type, HEP3, and measures of visual working memory through over conservative corrections for excessive and attention (Table 3). The haplotype HEP3 spans multiple testing, yet does not allow for the full allelic 62 kb of the DISC1 gene from intron 1 to exon 2. The variation to be described at this locus for these traits. analysis of affected and unaffected offspring sepa- Consequently, it is probable that different study rately showed that for visual working memory the samples will show different alleles associated with association is to the endophenotype rather than to this trait, as well as identifying alleles from this locus deficiencies in the trait caused by the disorder. This is that associate with other intermediate phenotypes. consistent with our prior hypothesis that variants on This shows that the findings so far on TRAX/DISC are 1q42 play a role in the etiology of schizophrenia not unlike that for Dysbindin and Neuregulin 1, and through an effect on the brain systems governing other putative schizophrenia susceptibility loci, working memory. where different haplotypes of markers in each locus Our current and previous observations24,25 show associate with the illness and its component traits, in that visuospatial memory traits display linkage and different study samples.4,5 association to 1q42 in Finnish schizophrenia families. The results of this study support our hypothe- Additionally, the P300 event-related potential (ERP) sis that DISC1 affects visual working memory and endophenotype has been linked to 1q42, with reduced P300 in all individuals from a large Scottish family carrying the balanced translocation Table 4 Mean average trait values between individuals with (1;11)(q42.1;q14.3), not just in those individuals at least one copy of the HEP3 haplotype (HEP3 þ ) and those suffering from schizophrenia.37 ERPs represent neu- with no copies (HEP3À) ronal electrophysiological activity in response to a sensory stimulus. It has been shown that the P300 Visual Visual 38 attention working ERP is reduced in patients with schizophrenia, in memory part when performing visual recognition tasks.39 It is interesting, in this light, that our findings show visual HEP3 þ individuals 7.7 6.9 working memory measures to associate with allelic HEP3À individuals 7.9 7.1 variants of DISC1. Working memory tasks invoke the dorsolateral prefrontal cortex (DLPFC), a brain region known to be disrupted both structurally and func- tionally in patients with schizophrenia.40,41 A recent analysis of all allelic variants of the HEP1 to HEP3 Table 5 Haplotype and observed transmission frequencies haplotypes in relation to neurocognitive functions of the HEP3 haplotype in a control sample and our original and gray matter volume in a sample of Finnish twins study sample6 concordant and discordant for schizophrenia ob- served association of a rare allele of a haplotype Sample Haplotype Transmission frequency frequency consisting of the SNPs combined from HEP2 and HEP3 (HEP2/HEP3) with increased risk for schizo- Control trios 0.070 0.072 phrenia, decreased gray matter in prefrontal cortex, Original study sample 0.088 0.075 and impaired visuospatial working memory function- Male offspring 0.097 0.080 ing (Cannon, et al, in press) (Figure 1). The findings Female offspring 0.077 0.066 across these two independent samples thus provide

Molecular Psychiatry DISC1 and quantitative neurocognitive traits W Hennah et al 1102 Table 6 Re-examination of the difference between the observed and expected number of transmissions to affected offspring in our original sample

Whole sample Male offspring only Female offspring only

Observed Expected Observed Expected Observed Expected

(a) The original observation6 HEP3 125 145 81 85 44 60 P-value 0.086 0.61 0.039

(b) Using an expected transmission frequency based on that of a control sample HEP3 125 109 81 66 44 43 P-value 0.12 0.050 0.92

In the whole sample, there were 1678 transmissions to affected offspring, of which 1006 were to male subjects and 672 to female subjects.

attention, based on the main finding that HEP3 Acknowledgements associates with poorer performance in these traits (Table 4). However, this appears counterintuitive as We acknowledge Ms OM Palo, Ms H Kilpinen, Ms J the HEP3 haplotype was earlier observed to be Meyer, and Mr A Parker for their contributing efforts undertransmitted to affected individuals, suggesting in genotyping these samples, and doctors R Araja¨rvi, at the time that the haplotype could represent H Juvonen, M Muhonen, J Suokas, K Suominen, J a potential protective variant.6 The issue was Suvisaari, and J Hintikka for their diagnostic work. further complicated by the fact that the observed This work was partly funded by Wyeth Pharmaceu- associations each have sex-dependent effects, being ticals Inc., Millennium Pharmaceuticals Inc., Acad- detectable only in female subjects when analyzing emy of Finland, Center of Excellence in Disease the end-state diagnosis, and only in male subjects Genetics and research grant for LP, and Finnish for visual working memory. To address this issue, Cultural Foundation, Paavo Koskinen and Aili and we genotyped the two SNPs that constitute the Paul Pennanen grants for WH. Mr Hennah is a HEP3 haplotype (rs751229, rs3738401) in a group member of the Helsinki Biomedical Graduate School of 60 control trios, representing anonymous and Dr Peltonen is the Gordon and Virginia Mac samples from the Finnish population. From this Donald Distinguished Chair in Human Genetics at sample set, we were able to determine the HEP3 UCLA. haplotype and transmission frequencies, which were lower than in the schizophrenia family sample (Table 5). Using a w2 test, we re-examined the References relationship between the observed number of trans- 1 Cannon TD, Kaprio J, Lo¨nnqvist J, Huttunen M, Koskenvuo M. The missions to affected individuals in our whole sample genetic epidemiology of schizophrenia in a Finnish twin cohort. and the expected number based on the transmission Arch Gen Psychiatry 1998; 55: 67–74. frequencies from the control trio sample (Table 6). 2 Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Although this analysis has less power for the Walsh D. The Roscommon family study. I. Methods, diagnosis of detection of association, it would be indicative probands, and risk of schizophrenia in relatives. Arch Gen Psychiatry 1993; 50: 527–540. of the type of results we could expect to observe 3 Tienari P, Wynne LC, Moring J, Lahti I, Naarala M, Sorri A et al. by doing an analysis similar to the one performed by The Finnish adoptive family study of schizophrenia. Implications the higher power TRANSMIT program, but using for family research. Br J Psychiatry Suppl 1994; 23: 20–26. allele frequencies determined in a control population. 4 Harrison PJ, Weinberger DR. Schizophrenia genes, gene expres- sion, and neuropathology: on the matter of their convergence. Mol These results suggest that due to HEP3 having Psychiatry 2005; 10: 40–68. a higher frequency in the schizophrenia sample 5 Hennah W, Varilo T, Paunio T, Peltonen L. Haplotype analysis and than controls, the original number of expected identification of genes for a complex trait: examples from transmissions was upwardly biased, which in turn schizophrenia. Ann Med 2004; 36: 322–331. led to the epiphenomenon of the observed under- 6 Hennah W, Varilo T, Kestila M, Paunio T, Arajarvi R, Haukka J et al. Haplotype transmission analysis provides evidence of transmission to affected female subjects. In the light association for DISC1 to schizophrenia and suggests sex-depen- of our present results, it is a more parsimonious dent effects. Hum Mol Genet 2003; 12: 3151–3159. interpretation of our data that the HEP3 haplotype 7 Hodgkinson CA, Goldman D, Jaeger J, Persaud S, Kane JM, Lipsky actually confers risk to male subjects in our study RH et al. Disrupted in schizophrenia 1 (DISC1): association with schizophrenia, schizoaffective disorder, and bipolar disorder. Am J population, and that this risk affects schizophrenia Hum Genet 2004; 75: 862–872. and its related disorders through a role in visual 8 Kockelkorn TT, Arai M, Matsumoto H, Fukuda N, Yamada K, working memory. Minabe Y et al. Association study of polymorphisms in the 5’

Molecular Psychiatry DISC1 and quantitative neurocognitive traits W Hennah et al 1103 upstream region of human DISC1 gene with schizophrenia. between 1q and working memory function in Neurosci Lett 2004; 368: 41–45. schizophrenia. Am J Med Genet 2003; 116B: 8–16. 9 Millar JK, Wilson-Annan JC, Anderson S, Christie S, Taylor MS, 25 Paunio T, Tuulio-Henriksson A, Hiekkalinna T, Perola M, Varilo T, Semple CAM et al. Disruption of two novel genes by a transloca- Partonen T et al. Search for cognitive trait components of tion co-segregating with schizophrenia. Hum Mol Genet 2000; 9: schizophrenia reveals a locus for verbal learning and memory on 1415–1423. 4q and for visual working memory on 2q. Hum Mol Genet 2004; 13: 10 St Clair D, Blackwood D, Muir W, Carothers A, Walker M, 1693–1702. Spowart G et al. Association within a family of a balanced 26 American Psychiatric Association. Diagnostic and Statistical autosomal translocation with major mental illness. Lancet 1990; Manual of Mental Disorders, 4th edn (DSM-IV). American 336: 13–16. Psychiatric Press Inc.: Washington, DC, 1994. 11 Ekelund J, Hovatta I, Parker A, Paunio T, Varilo T, Martin R et al. 27 Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, loci in Finnish schizophrenia families. Hum Mol Walsh D. The Roscommon Family study. III. Schizophrenia-related Genet 2001; 10: 1611–1617. personality disorders in relatives. Arch Gen Psychiatry 1993; 50: 12 Ekelund J, Hennah W, Hiekkalinna T, Parker A, Meyer J, Lonnqvist 781–788. J et al. Replication of 1q42 linkage in Finnish schizophrenia 28 Tuulio-Henriksson A, Haukka J, Partonen T, Varilo T, Paunio T, pedigrees. Mol Psychiatry 2004; 9: 1037–1041. Ekelund J et al. Heritability and number of quantitative trait loci of 13 Hwu HG, Liu CM, Fann CS, Ou-Yang WC, Lee SF. Linkage of neurocognitive functions in families with schizophrenia. Am J schizophrenia with chromosome 1q loci in Taiwanese families. Med Genet 2002; 114: 483–490. Mol Psychiatry 2003; 8: 445–452. 29 Wechsler D. Wechsler Memory Scale-Revised (WMS-R), Manual. 14 Curtis D, Kalsi G, Brynjolfsson J, McInnis M, O’Neill J, Smyth C Psychological Corporation, Harcourt Brace Jovanovich Inc.: San et al. Genome scan of pedigrees multiply affected with bipolar Antonio, 1987. disorder provides further support for the presence of a suscept- 30 Delis DC, Kramer JH, Kaplan E, Ober BA. California Verbal ibility locus on chromosome 12q23–q24, and suggests the Learning Test: Manual, Research Edition. Psychological Corpora- presence of additional loci on 1p and 1q. Psychiatr Genet 2003; tion, Harcourt Brace & Company: San Antonio, 1987. 13: 77–84. 31 Wechsler D. Wechsler Adult Intelligence Scale-Revised (WAIS-R), 15 Miyoshi K, Honda A, Baba K, Taniguchi M, Oono K, Fujita T et al. Manual. Psychological Corporation, Harcourt Brace Jovanovich Disrupted-In-Schizophrenia 1, a candidate gene for schizophrenia, Inc.: Cleveland, 1981. participates in neurite outgrowth. Mol Psychiatry 2003; 8: 32 Tuulio-Henriksson A, Arajarvi R, Partonen T, Haukka J, Varilo T, 685–694. Schreck M. et al. Familial loading associates with impairment in 16 Ozeki Y, Tomoda T, Kleiderlein J, Kamiya A, Bord L, Fujii K et al. visual span among healthy siblings of schizophrenia patients. Biol Disrupted-in-Schizophrenia-1 (DISC-1): mutant truncation Psychiatry 2003; 54: 623–628. prevents binding to NudE-like (NUDEL) and inhibits neurite 33 Cannon TD, Huttunen MO, Lonnqvist J, Tuulio-Henriksson A, outgrowth. Proc Natl Acad Sci USA 2003; 100: 289–294. Pirkola T, Glahn D et al. The inheritance of neuropsychological 17 Austin CP, Ma L, Ky B, Morris JA, Shughrue PJ. DISC1 (Disrupted dysfunction in twins discordant for schizophrenia. Am J Hum in Schizophrenia-1) is expressed in limbic regions of the primate Genet 2000; 67: 369–382. brain. Neuroreport 2003; 14: 951–954. 34 Clayton D. A generalization of the transmission/disequilibrium 18 Millar JK, Christie S, Porteous DJ. Yeast two-hybrid screens test for uncertain-haplotype transmission. Am J Hum Genet 1999; implicate DISC1 in brain development and function. Biochem 65: 1170–1177. Biophys Res Commun 2003; 311: 1019–1025. 35 Sobel E, Lange K. Descent graphs in pedigree analysis: applica- 19 Morris JA, Kandpal G, Ma L, Austin CP. DISC1 (Disrupted-In- tions to haplotyping, location scores, and marker-sharing statis- Schizophrenia 1) is a -associated protein that interacts tics. Am J Hum Genet 1996; 58: 1323–1337. with MAP1A, MIPT3, ATF4/5 and NUDEL: regulation and loss of 36 Abecasis GR, Cardon LR, Cookson WO. A general test of interaction with mutation. Hum Mol Genet 2003; 12: 1591–1608. association for quantitative traits in nuclear families. Am J Hum 20 Brandon NJ, Handford EJ, Schurov I, Rain JC, Pelling M, Duran- Genet 2000; 66: 279–292. Jimeniz B et al. Disrupted in Schizophrenia 1 and Nudel form a 37 Blackwood DHR, Fordyce A, Walker MT, St Clair DM, Porteous DJ, neurodevelopmentally regulated protein complex: implications Muir WJ. Schizophrenia and affective disorders–cosegregation for schizophrenia and other major neurological disorders. Mol Cell with a translocation at chromosome 1q42 that directly disrupts Neurosci 2004; 25: 42–55. brain-expressed genes: clinical and P300 findings in a family. Am J 21 Finkenstadt PM, Kang W-S, Jeon M, Taira E, Tang W, Baraban JM. Hum Genet 2001; 69: 428–433. Somatodendritic localization of translin, a component of the 38 Jeon YW, Polich J. Meta-analysis of P300 and schizophrenia: translin/TRAX RNA binding complex. J Neurochem 2000; 75: patients, paradigms, and practical implications. Psychophysiology 1754–1762. 2003; 40: 684–701. 22 Gottesman II, Gould TD. The endophenotype concept in psychia- 39 Vianin P, Posada A, Hugues E, Franck N, Bovet P, Parnas J et al. try: etymology and strategic intentions. Am J Psychiatry 2003; 160: Reduced P300 amplitude in a visual recognition task in patients 636–645. with schizophrenia. Neuroimage 2002; 17: 911–921. 23 Hovatta I, Varilo T, Suvisaari J, Terwilliger JD, Ollikainen V, 40 Winterer G, Coppola R, Goldberg TE, Egan MF, Jones DW, Sanchez Araja¨rvi R et al. A Genomewide screen for schizophrenia genes in CE et al. Prefrontal broadband noise, working memory, and genetic an isolated Finnish subpopulation, suggesting multiple suscept- risk for schizophrenia. Am J Psychiatry 2004; 161: 490–500. ibility loci. Am J Hum Genet 1999; 65: 1114–1124. 41 Levy R, Goldman-Rakic PS. Segregation of working memory 24 Gasperoni TL, Ekelund J, Huttunen M, Palmer CG, Tuulio- functions within the dorsolateral prefrontal cortex. Exp Brain Henriksson A, Lonnqvist J et al. and association Res 2000; 133: 23–32.

Molecular Psychiatry