Abstract

Recent molecular studies have implicated common alleles of small to moderate effect and rare alleles with larger effect sizes in the genetic architecture of schizophrenia (SCZ). It is expected that the reliable detection of risk variants with very small effect sizes can only be achieved through the recruitment of very large samples of patients and controls (that is tens of thousands), or large, potentially more homogeneous samples that have been recruited from confined geographical areas using identical diagnostic criteria. Applying the latter strategy, we performed a genome-wide association study (GWAS) of 1169 clinically well characterized and ethnically homogeneous SCZ patients from a confined area of Western Europe (464 from Germany, 705 from The Netherlands) and 3714 ethnically matched controls (1272 and 2442, respectively). In a subsequent follow-up study of our top GWAS results, we included an additional 2569 SCZ patients and 4088 controls (from Germany, The Netherlands and Denmark). Genetic variation in a region on chromosome 11 that contains the candidate genes AMBRA1, DGKZ, CHRM4 and MDK was significantly associated with SCZ in the combined sample (n=11 540; P=3.89 × 10−9, odds ratio (OR)=1.25). This finding was replicated in 23 206 independent samples of European ancestry (P=0.0029, OR=1.11). In a subsequent imaging genetics study, healthy carriers of the risk allele exhibited altered activation in the cingulate cortex during a cognitive control task. The area of interest is a critical interface between emotion regulation and cognition that is structurally and functionally abnormal in SCZ and bipolar disorder.

Introduction

Schizophrenia (SCZ) is a severe psychiatric disorder characterized by fundamental and characteristic distortions of thought and perception. It has a lifetime prevalence of around 0.5–1%. Formal genetic studies have shown that genetic factors contribute substantially to the disease risk. Meta-analyses of pooled data from 12 twin studies have estimated that susceptibility to SCZ has a heritability of 80%.1 Recent molecular studies have implicated common alleles of small to moderate effect and rare alleles with larger effect sizes in the genetic architecture of SCZ.2 Genome-wide association studies (GWAS) using single nucleotide polymorphism (SNP) array technologies have been applied to detect common risk alleles, and a total of 12 GWAS of SCZ have been published to date (Supplementary Table S3). Common risk alleles in the major histocompatibility region on chromosome 6 have so far shown the most statistically significant evidence of association.3, 4, 5 Interestingly, the major histocompatibility region, which is involved in the immune response, has long been postulated to harbor variants conferring a risk for SCZ as there is evidence for linkage in this region6 and research has suggested the involvement of infection in disease development.7 Genome-wide significance has also been reported for risk alleles at TCF4,3 NGRN3 and ZNF804A.8

All GWAS of SCZ performed to date have indicated that the strongest common genetic risk factors have odds ratios (OR) that are no greater than 1.15–1.20. In fact, recent molecular genetic evidence points to a substantial polygenic component to the risk of SCZ that involves a large number of common risk alleles of very small effect.4 It is likely that the reliable detection of risk variants with very small effects can only be achieved through the study of very large samples of patients and controls (that is, tens of thousands) or large, potentially more homogeneous samples that have been recruited from a confined geographical area using the same diagnostic criteria. Using the latter strategy, we performed a GWAS of clinically well characterized and ethnically homogeneous SCZ patients recruited from a confined area of Western Europe (Germany, The Netherlands and Denmark). Our results suggest that genetic variation in a region on chromosome 11, which contains the candidate genes AMBRA1, DGKZ, CHRM4 and MDK, is implicated in the etiology of SCZ.

Patients and methods

The following section provides details of sample recruitment and quality control (QC) for the GWAS data set. Information concerning specific aspects of the GWAS data set and data from the follow-up samples are provided in the Supplementary Information Patients and Methods.

Sample ascertainment and selection for the GWAS sample

All participating individuals provided written informed consent. The study protocols were approved by the respective institutional review boards or ethics committees.

German sample (Bonn—Mannheim)

The German SCZ patients used in the GWAS step (n=487) were recruited from consecutive hospital admissions and were all of German descent. Lifetime best estimate diagnoses were assigned according to DSM-IV criteria on the basis of multiple sources of information including structured interviews with the SCID9 or SADS-L,10 the OPCRIT,11 medical records and family history. Best estimate diagnoses were assigned by at least two experienced psychiatrists/psychologists. The controls were drawn from three population-based epidemiological studies: (A) PopGen12 (n=490), (B) KORA13 (n=488) and (C) the Heinz Nixdorf Recall (n=383) study14 (risk factors, evaluation of coronary calcification and lifestyle). The recruitment areas for PopGen, KORA and HNR were located in the following: (i) Schleswig-Holstein (Northern Germany); (ii) Essen, Bochum and Mühlheim (Ruhr area); and (iii) Augsburg (Southern Germany), respectively. Ethnicity was assigned to patients and controls on the basis of self-reported ancestry.

Dutch sample (Utrecht and Rotterdam)

Inpatients and outpatients (n=804) were recruited from various psychiatric hospitals and institutions throughout the Netherlands. Detailed medical and psychiatric histories were collected, and this process included use of the Comprehensive Assessment of Symptoms and History (CASH),15 an instrument for assessing diagnosis and psychopathology. Only those patients with a DSM-IV diagnosis of SCZ were finally included as cases. The controls from Utrecht (n=704) were volunteers with no history of psychiatric disorder. The Rotterdam control individuals (n=2302) were drawn from a large population-based project on the genetics of complex traits and diseases, which is financed by the Dutch government through the Netherlands Scientific Organization—Large Investments (NWO Groot; 175.010.2005.011). This prospective population-based cohort study of chronic disabling conditions in Dutch individuals aged 55 years and above is described elsewhere.16, 17 All patients and controls had at least three grandparents of Dutch ancestry.

Sample ascertainment and selection for the imaging genetics study sample

Subjects (n=122) were drawn from an ongoing large-scale multicenter imaging genetics study18 (Esslinger et al. plus seven subsequently scanned subjects) that is being conducted at two sites in Mannheim and Bonn, Germany. All participants were healthy German volunteers with parents and grandparents of European origin. None of the participants had any self-reported lifetime or family history of SCZ or affective disorder. All subjects provided written informed consent. The study was approved by the local ethics committees of the Universities of Heidelberg and Bonn.

Genotyping and QC for the GWAS and the imaging genetics study

Ethylenediaminetetraacetic acid anti-coagulated venous blood samples were collected from all participating individuals (GWAS and imaging genetics study). Lymphocyte DNA was isolated by salting-out19 with saturated sodium chloride solution or by a Chemagic Magnetic Separation Module I (Chemagen, Baesweiler, Germany) used according to the manufacturer's recommendations. The GWAS data set was assembled from seven sub-data sets, which were genotyped by the following: (i) Illumina's customer service, San Diego, CA, USA (all PopGen controls, all German SCZ patients); (ii) the Department of Genomics, Life & Brain Center, University of Bonn (all HNR controls); (iii) the Helmholtz Zentrum München, Germany (all KORA controls); (iv) The Southern California Genotyping Consortium (SCGC) at UCLA, Los Angeles, USA (all Dutch SCZ patients and n=704 Dutch controls); and (v) the genotyping facility of the ErasmusMC Biomics core facility, Rotterdam, The Netherlands (n=2302 Dutch controls, including the controls for the follow-up sample). All genome-wide genotyping for the GWAS was performed on HumanHap550v3 BeadArrays using the Infinium II assay (Illumina). The imaging genetics sample (n=122) was genotyped by the Department of Genomics, Life & Brain Center, University of Bonn. The genome-wide genotyping was performed on Human610-quad beadchips using the Infinium HD assay (Illumina). For a discussion on the possible confounding of case/control status with plate and genotyping procedures please see the Supplementary Information Patients and Methods.

We developed a protocol of filters for the stringent QC of whole-genome and sub-whole genome data sets. This accounted for call rates, heterozygosity, cross-contamination, population stratification, relatedness, non-random-missingness, Hardy–Weinberg equilibrium, minor allele frequency and others. With the exception of the cluster plot investigations for the follow-up SNPs, each QC filter for the GWAS, the imaging genetics study and the follow-up data sets was performed using the PLINK toolset.20 The samples from the Rotterdam study (controls for the GWAS and the replication step) underwent QC together with the other GWAS samples.

The QC protocol was applied to a total of 1291 patients and 4367 controls for the GWAS step. As a result, 375 individuals including 122 patients and 253 controls (representing 6.70, 9.73 and 5.80%, of the sample) as well as 86 037 SNPs (15.32% of the HumanHap550v3 BeadArray content) were excluded before the association analysis. The final GWAS data set was comprised of 464 030 autosomal (including PAR1+2) and 11 397 X-chromosomal SNPs, genotyped in 1169 SCZ patients and 3714 controls. An additional 400 controls were added to the Dutch follow-up data set described above.

The QC protocol was also applied to the 122 healthy German individuals who participated in the imaging genetics study. As a result, one individual and 39 356 SNPs (6.64% of the Human610-quad BeadArray content of evaluable 592 532 SNPs) were excluded before the association analyses. The mean age of the remaining 121 individuals (43.4% males) was 33.1 (s.d. 10.3). For the analysis of the imaging genetics data, only information for rs11819869, which was present in the post QC data set, was extracted and analyzed.

A detailed description of the QC filters is provided in the Supplementary Information Patients and Methods.

Statistical analyses for GWAS and follow-up

All association analyses were performed using PLINK20 (v1.07). In the single-marker analysis, all autosomal (including PAR1 and PAR2) and gonosomal SNPs that passed QC checks were tested for association with SCZ using the Cochran–Mantel–Haenszel test (CMH) with 2 × 2 × K stratified tables {disease × SNP disease cluster} with K=2 (assigning German and Dutch individuals to one cluster each) for the GWAS step, K=4 (one cluster for the German, Dutch, and Danish samples, respectively) for Replication 1, and K=6 for the combined analysis. Following analysis of the initial GWAS, a set of 60 SNPs was chosen for the follow-up step using a top-down approach (Replication 1; more information on the selection of SNPs is provided in the Supplementary Information Patients and Methods). The CMH was also applied to the replication sample as described above. In the replication analysis, the test statistic was one-tailed if the OR was in the same direction as in the GWAS. SNPs that exceeded the threshold for nominal significance (P=0.05) were identified. Finally, for all SNPs included in the replication step, a combined analysis of individuals from the GWAS and the replication step was performed using the CMH (K=6; test statistic always two-tailed). The Breslow–Day test was used to investigate the homogeneity of the ORs for the replicated SNPs. The meta-analysis for the 15 independent samples of European ancestry (Replication 2) was performed using PLINK20 (v1.07) and considering a fixed effect model. There was no evidence of heterogeneity (Cochrane's Q). In cases where genome-wide data were available, the two-tailed P-value of the TREND test was corrected by the genomic inflation factor for the specific sample as calculated from the genome-wide data.

A note on possible stratification

We have assumed that our sample is characterized by a high degree of ethnic (and presumably genetic) homogeneity. Nonetheless, several steps were taken to minimize population stratification. Patients and controls were only included when a minimum of three grandparents originated from the respective country. Outliers were identified using two methods: (1) multidimensional scaling and (2) outlier detection diagnostics as implemented in PLINK.20 This was performed for the Dutch and German samples individually, and then in the combined sample (Supplementary Figure S6, further information is provided in the Supplementary Information Patients and Methods). Tests were applied to detect differences in allele frequencies and missingness patterns between the single Dutch and German control samples (Rotterdam, Utrecht, Kiel, Munich and Essen). Finally, association analyses were performed using a CMH test and a logistic regression with covariates derived from multidimensional scaling analyses (first six dimensions) to minimize potential stratification effects (Supplementary Figure S5, more information is provided in the Supplementary Information Patients and Methods).

Imaging genetics study

The participants of the imaging genetics study completed a battery of six cognitive tasks that have been used and validated previously in imaging genetics. Specifically, the following tasks were performed: (1) an associative learning task,21 in which subjects learn and are then tested on face-job associations, (2) an n-back working memory task,22 in which subjects are presented with a sequence of numbers and press a button corresponding either to the number currently seen (control condition, ‘0-back’) or the number seen two presentations previously (‘2-back’), (3) a theory of mind task,23 in which subjects have to make inferences about the state of mind of a human based on a series of cartoons, (4) a flanker task24 in which subjects have to perform or withhold a button press depending on a set of either congruent or incongruent stimuli, requiring cognitive control, (5) an implicit emotion recognition task,25 in which subjects match pictures of angry and fearful faces, (6) a monetary reward task,26 in which subjects receive or do not receive monetary rewards according to their performance in a reaction time task and a 5 min of rest period. Further information, including details of the imaging parameters, is provided in the Supplementary Information Patients and Methods.

For analysis of the influence of rs11819869, contrast images of incongruent > congruent trials were subjected to a two-sample t-test with age, gender and scanning site as covariates of no interest. To avoid small cell sizes, CT heterozygotes and TT homozygotes were combined and compared with the CC homozygous group.

In view of our exploratory approach, we used a significant threshold of P<0.0083 (equivalent to P<0.05 with Bonferroni correction for six tasks), correcting for multiple comparisons across all voxels of the brain using family-wise error, based on Gaussian random fields theory. Further information is provided in the Supplementary Information Patients and Methods.

Results

We performed a GWAS of 1169 clinically well characterized and ethnically homogeneous SCZ patients who had been recruited from a confined area of Western Europe (n=464 from Germany, n=705 from The Netherlands) and 3714 ethnically matched controls (n=1272 and n=2442, respectively; Table 1). A total of 475 427 SNPs (Illumina's HumanHap550v3 BeadArray) were tested using the CMH test (CMH, for 2 × 2 × K stratified tables, Patients and methods section). No marker exceeded the widely acknowledged genome-wide significance threshold of 5 × 10−8, or a Bonferroni-corrected significance threshold of 1.1 × 10−7. The best result (PGWAS=4.50 × 10−7) was for the SNP rs11154491, which is located in an intron of the Rho GTPase activating protein 18 gene on chromosome 6 (ARGHAP18, Supplementary Figure S2). An overview of the GWAS results is provided in Figure 1a, a quantile–quantile plot of the P-values is shown in Supplementary Figure S4.

Table 1: Descriptive data for schizophrenia patients and controls following quality control
Figure 1
Figure 1

Association results for the genome-wide association study (GWAS) and the two best-supported genes from the follow-up study (Replication 1). (a) Manhattan plot. Regional plots depicting (b) AMBRA1/CHRM4/DGKZ/MDK region and (c) CCDC68/TCF4 region. The best-associated marker from the GWAS (enlarged red diamond) is centered in a genomic window of 1 Mb (hg18, RefSeq genes); its P-value from the combined analysis (PCOMB) is shown (enlarged blue diamond). The linkage disequilibrium (LD) strength (r2) between the sentinel single nucleotide polymorphism from the GWAS and its flanking markers is demonstrated by the red (high) to white (low) colored bar. The recombination rate (cM/Mb; second y axis) is plotted in blue, according to HapMap27 CEU.

Under the assumptions that the most significantly associated SNPs included true SCZ susceptibility factors and that the failure to reach genome-wide significance in the GWAS step was a result of insufficient power, we performed a subsequent follow-up analysis (Replication 1). A total of 43 SNPs were selected using a top-down P-value approach in 2569 additional patients and 4088 controls from Western Europe (Germany/Munich: 913 patients/1668 controls; Bonn/Mannheim: 600/1146; The Netherlands: 178/400; Denmark: 878/874; Table 1). In all 9 of the 43 SNPs (21%) showed nominal significance in the combined replication samples, and all with the same alleles as in the GWAS (Table 2 and Supplementary Figure 1 and Supplementary Table S1a and b). The strongest evidence for association was found for four highly correlated SNPs (rs7112229, rs11819869, rs7130141 and rs12575668). These four SNPs are located in intronic regions of the activating molecule in beclin-1-regulated gene (AMBRA1) on chromosome 11. The best SNP (rs11819869) showed PREPLI=5.04 × 10−5 (PCORR=0.0022). One further SNP (rs4309482) withstood Bonferroni correction (PCORR=0.013) for the number of SNPs tested in the replication step. This SNP is located in an intergenic region of chromosome 18 between the coiled-coil domain-containing 68 gene (CCDC68) and the transcription factor 4 gene (TCF4). Our top GWAS SNP (rs11154491), which is located in ARGHAP18 on chromosome 6q22.33, was not replicated (PREPLI=0.38).

Table 2: Nine SNPs replicated in the follow-up study (Replication 1) with the same alleles as in the GWAS

In an analysis of the combined samples (GWAS+follow-up; n=3738 patients/7802 controls; CMH, K=6), variation in the chromosome 11 region surpassed the threshold for genome-wide significance (rs11819869, PCOMB=3.89 × 10−9, OR=1.25, Figure 1b). A Breslow–Day test was performed across all of the analyzed samples for all SNPs that had withstood Bonferroni correction in the replication step. This revealed no significant differences in ORs (P>0.05) between the investigated samples. In support of this, subtraction of any one individual replication sample failed to alter the effect sizes to any substantial degree (Supplementary Table S1b).

We then attempted to replicate our genome-wide significant result for rs11819869 in 15 independent samples of European ancestry (Replication 2; 4734 patients and 18 472 controls). We observed significant association for the T risk allele of rs11819869, as had been observed in the GWAS study (PMETA=0.0029, OR=1.11; Table 3). We then applied a functional magnetic resonance imaging test battery for imaging genetics in a total of 121 healthy subjects (53 males, 68 females) of German descent with no family history of affective disorder or SCZ. This demonstrated a significant effect of the rs11819869 genotype on medial prefrontal activation during a flanker task, which contrasted incongruent and congruent stimuli configurations (P<0.05, family wise error corrected for multiple testing, cluster size: K=9, maximum activation found at (−3, 36, −18), T(116)=5.522, PCORR<0.008 (see Patients and methods for details of the task and the statistics). Carriers of the risk allele showed increased activation during incongruent conditions, whereas homozygote carriers of the C-allele showed decreased activation (Figure 2b).

Table 3: Association results for rs11819869 at all stages of the analysis: GWAS, Replication 1 and Replication 2
Figure 2
Figure 2

Effect of the rs11819869 genotype on brain activation during a flanker task (contrast incongruent trials >congruent trials). (a) The only cluster showing a significant group difference between carriers of at least one T allele and homozygote C allele carriers after correction for multiple testing (P<0.05, family wise error corrected for the entire brain and the number of different tasks tested). (b) Mean contrast estimates (+/− standard errors) for the significant cluster, reflecting an increased activation during incongruent trials for the T-allele carriers and during congruent trials for participants homozygous for the C allele.

Discussion

The SNP rs11819869 is located in the gene AMBRA1, in a region of strong linkage disequilibrium (LD) that spans 360 kb (r2>0.8 based on HapMap Phase 2 CEU data,27 Supplementary Figure S3). Other RefSeq genes in the LD region are KIAA0652/ATG13, CHRM4, DGKZ, MDK, HARBI1, ARGAPH1 and ZNF408. In our initial GWAS, 11 of the 25 most significant SNPs were located in this region. Although this region also showed evidence for association in the GWAS study of the SGENE consortium,3 it was not among the top findings selected for follow-up.3

AMBRA1 has a major role in the development of the nervous system and has been reported to be a member of the Autophagy Interaction Network.28 In mouse models, AMBRA1 functional deficiency results in severe neural tube defects that are associated with impaired autophagy, accumulation of ubiquitinated proteins, unbalanced cell proliferation and excessive apoptotic cell death.29, 30 The gene is highly expressed in the most ventral part of the undifferentiated neural tube during embryogenesis.30 Another gene in the chromosome 11 region of strong LD (KIAA0652/ATG13) is also a member of the Autophagy Interaction Network28 and acts as a regulator of autophagy.28 Together with the association findings for genetic variation in TCF4 and NRGN,3 these reports and our findings may be regarded as supportive evidence for the role of brain-development genes in the etiology of SCZ.

Although AMBRA1 is an interesting functional candidate gene and includes the top associated SNP, (Figure 1b, Supplementary Figure S3) it is difficult to pinpoint the potential susceptibility gene on the basis of the genetic data alone due to the presence of strong LD in the associated region. Another very interesting gene that is covered by LD in this region is the muscarinic acetylcholine receptor M4 gene (CHRM4). Neuropsychopharmacological and neuroimaging studies have produced strong evidence that muscarinic cholinergic receptors are involved in SCZ.31 Studies of muscarinic receptor knockout mice have suggested that CHRM4 has an impact on the homeostatic control of cholinergic activity and dopaminergic neurotransmission in mesolimbic and hippocampal brain regions.32 Interestingly, the muscarinic agonist Xanomeline, which is selective for the CHRM1 and CHRM4 subtypes, exhibited functional dopamine antagonism and resulted in antipsychotic-like effects in rodent models predictive of antipsychotic behaviors.32 Furthermore, a recent study proposed allosteric modulation of the muscarinic M4 receptor as a potential approach to the treatment of SCZ.33 Further information concerning the impact of CHRM4 on the pharmaceutical treatment of SCZ is provided in the Supplementary Information.

Another gene of interest in the LD region of chromosome 11 (Figure 1b, Supplementary Figure S3) is the diacylglycerol kinase zeta gene (DGKZ). There are a total of 10 diacylglycerol kinase enzymes (DGKs) (DGKalpha, DGKbeta, DGKgamma, DGKdelta, DGKepsilon, DGKzeta, DGKeta, DGKtheta, DGKiota and DGKkappa) and these metabolize 1,2,diacylglycerol to phosphatidic acid. Diacylglycerol kinases are central to a wide range of signal transduction pathways of potential relevance to neuropsychiatric disorders.34 The diacylglycerol kinase eta (DGKH), a key protein in the lithium-sensitive phosphatidyl inositol pathway, has recently been implicated in the etiology of bipolar disorder.35 A further member of this protein family, DGKbeta, has been reported to promote dendritic outgrowth and spine maturation in developing hippocampal neurons.36 In addition, gene-wide evidence for association with SCZ has been reported for the gene that encodes for DGKiota (DGKI).33 One independent study3 found evidence for association between SCZ and a DGK1 variant. In view of these findings, DGKZ may represent a further promising candidate gene for SCZ in the LD region of chromosome 11.

Last but not least, Midkine (MDK) is an interesting candidate gene in the region of our top finding. It has been shown that Mdk(−/−) mice exhibited a delayed hippocampal development with impaired working memory and increased anxiety.37 In addition, midkine was found to accumulate in senile plaques in the hippocampus of patients with Alzheimer's disease.38 Ohgake et al.39 reported that Mdk(−/−) mice showed a significantly disrupted Prepulse inhibition (PPI) test when compared with Mdk(+/+) mice. This is interesting, because PPI has previously been shown to be a valuable tool for evaluating models or model organisms relevant to SCZ.40 The PPI reduction was compensated when Mdk(−/−) mice were pre-treated with either haloperidol or clozapine (both known to reverse PPI deficits).39 In view of their findings, Ohgake et al.39 hypnotized that Mdk(−/−) mice might act as a putative animal model of SCZ.

As none of the 11 top SNPs, which are all located in introns of AMBRA1, have any known disease causing function, the SNAP tool41 was used to test whether our top SNP (rs11819869) was in strong LD (r2>0.8, based on HapMap27 CEU Phase 2 and 3 data) with SNPs that are functional variants. We found that a DGKZ splice-site variant (rs2046768) was in strong LD with our top SNP (r2=0.935 for HapMap10 CEU data of Phase 2 release 22, Supplementary Figure S3). As this SNP was not present in our post QC data set, it was necessary to impute the information for the GWAS data set using the strategy described in the Supplementary Information Patients and Methods (rs2046768, PIMPUT=8.79 × 10−5). Although the result for the functional variant itself lags behind the GWAS result for the top SNP (rs11819869, PGWAS=4.71 × 10−6), we cannot excluded the possibility that this is the functionally relevant effect that was picked up by our top SNP in AMBRA1.

Irrespective of which specific gene in the region is tagged by the identified SNPs, the impact of the risk allele on brain function and SCZ can be tested using imaging genetics. This approach has high specificity when applied to risk genes for SCZ.42 The stringent statistical approach, which combines Bonferroni correction over tasks with family-wise error correction over brain regions tested, provides strong protection against false positive findings.42 The present results therefore indicate a regionally specific impact on the function of the subgenual cingulate during a cognitive control task. This represents a critical interface between emotion regulation and cognition that is structurally43 and functionally44 abnormal in SCZ and bipolar disorder. The present findings therefore provide evidence that the identified risk allele is functional in a neural system of relevance to the disorder.

Our second best result was for rs4309482 (PCOMB=9.68 × 10−7, OR=0.87, Figure 1c), which is located near CCDC68 and TCF4 on chromosome 18. This SNP, unlike the SNPs in the chromosome 11 region, was included in the follow-up of the SGENE study and achieved a combined P-value of 7.1 × 10−5 (Supplementary Table S2 in Stefansson et al.3). However, as a substantial proportion of our combined case sample (and a smaller proportion of our control sample) was included in that study, our result for this SNP cannot be regarded as an independent replication.

In summary, the present study has identified a susceptibility region for SCZ on chromosome 11, which contains four excellent functional candidate genes: AMBRA1, DGKZ, CHRM4 and MDK. In addition, we found evidence that the identified risk allele is functional in a neural system that is of relevance to the disorder. The aim of future studies will be to determine which of the genes in the region contribute to the risk for SCZ. CHRM4 is of particular interest as it may lead to novel therapeutic interventions for SCZ.

References

  1. 1.

    , , . Schizophrenia as a complex trait: evidence from a meta-analysis of twin studies. Arch Gen Psychiatry 2003; 60: 1187–1192.

  2. 2.

    , , . New findings from genetic association studies of schizophrenia. J Hum Genet 2009; 54: 9–14.

  3. 3.

    , , , , , et al. Common variants conferring risk of schizophrenia. Nature 2009; 460: 744–747.

  4. 4.

    International Schizophrenia Consortium, , , , , et al. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature 2009; 460: 748–752.

  5. 5.

    , , , , , et al. Common variants on chromosome 6p22.1 are associated with schizophrenia. Nature 2009; 460: 753–757.

  6. 6.

    , , , , , et al. Genome scan meta-analysis of schizophrenia and bipolar disorder, part II: Schizophrenia. Am J Hum Genet 2003; 73: 34–48.

  7. 7.

    , . Prenatal infection and schizophrenia: a review of epidemiologic and translational studies. Am J Psychiatry 2010; 167: 261–280.

  8. 8.

    , , , , , et al. Identification of loci associated with schizophrenia by genome-wide association and follow-up. Nat Genet 2008; 40: 1053–1055.

  9. 9.

    , , , . Structured Clinical Interview for DSM-IV Axis I Disorders. New York State Psychiatric Institute, Biometrics Research: New York, 1994.

  10. 10.

    , , . Research diagnostic criteria: rationale and reliability. Arch Gen Psychiatry 1978; 35: 773–782.

  11. 11.

    , , . A polydiagnostic application of operational criteria in studies of psychotic illness. Development and reliability of the OPCRIT system. Arch Gen Psychiatry 1991; 48: 764–770.

  12. 12.

    , , , , , . PopGen: population-based recruitment of patients and controls for the analysis of complex genotype-phenotype relationships. Community Genet 2006; 9: 55–61.

  13. 13.

    , , . KORA-gen – Resource for population genetics, controls and a broad spectrum of disease phenotypes. KORA- gen – Ressource für Bevölkerungsgenetik, Kontrolle und ein breites Spektrum an Krankheitsphänotypen. Gesundheitswesen 2005; 67: S26–S30.

  14. 14.

    , , , , , et al. Assessment of clinically silent atherosclerotic disease and established and novel risk factors for predicting myocardial infarction and cardiac death in healthy middle-aged subjects: rationale and design of the Heinz Nixdorf Recall Study. Am Heart J 2002; 144: 212–218.

  15. 15.

    , , . The comprehensive assessment of symptoms and history (CASH). An instrument for assessing diagnosis and psychopathology. Arch Gen Psychiatry 1992; 49: 615–623.

  16. 16.

    , , , , , et al. The Rotterdam Study: objectives and design update. Eur J Epidemiol 2007; 22: 819–829.

  17. 17.

    , , , . Determinants of disease and disability in the elderly: the Rotterdam Elderly Study. Eur J Epidemiol 1991; 7: 403–422.

  18. 18.

    , , , , , et al. Neural mechanisms of a genome-wide supported psychosis variant. Science 2009; 324: 605.

  19. 19.

    , , . A simple salting out precedure for extracting DNA from Human nucleated cells. Nucleic Acids Res 1988; 16: 1215.

  20. 20.

    , , , , , et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575.

  21. 21.

    , , , , , et al. Altered neural function in carriers of a bipolar disorder variant with genome-wide support. Arch Gen Psychiatry 2010; 67: 803–811.

  22. 22.

    , , , , , et al. Functional magnetic resonance imaging brain mapping in psychiatry: methodological issues illustrated in a study of working memory in schizophrenia. Neuropsychopharmacology 1998; 18: 186–196.

  23. 23.

    , , , , , et al. Effects of a genome-wide supported psychosis risk variant on neural activation during a theory-of-mind task. Mol Psychiatry 2011; 16: 462–470.

  24. 24.

    , , , , , et al. Neural mechanisms of genetic risk for impulsivity and violence in humans. Proc Natl Acad Sci USA 2006; 103: 6269–6274.

  25. 25.

    , , , , . The amygdala response to emotional stimuli: a comparison of faces and scenes. Neuroimage 2002; 17: 317–323.

  26. 26.

    , , , , , et al. Anticipation of reward in a nonaversive differential conditioning paradigm and the brain reward system: an eventrelated fMRI study. Neuroimage 2003; 20: 1086–1095.

  27. 27.

    The International HapMap Consortium. A haplotype map of the human genome. Nature 2005; 437: 1299–1320.

  28. 28.

    , , , . Network organization of the human autophagy system. Nature 2010; 466: 68–76.

  29. 29.

    , , , , , et al. Ambra1 regulates autophagy and development of the nervous system. Nature 2007; 447: 1121–1125.

  30. 30.

    , , , , , et al. A novel role for autophagy in neurodevelopment. Autophagy 2007; 3: 506–508.

  31. 31.

    . Muscarinic receptors in psychiatric disorders - can we mimic ‘Health’? Neurosignals 2009; 17: 298–310.

  32. 32.

    , . Cholinergic mechanisms in schizophrenia: current concepts. Curr Psychosis Therapeutics Rep 2006; 4: 20–26.

  33. 33.

    , , , , , et al. Allosteric modulation of the muscarinic M4 receptor as an approach to treating schizophrenia. Proc Natl Acad Sci USA 2008; 105: 10978–10983.

  34. 34.

    , , , , , et al. Gene-wide analyses of genome-wide association data sets: evidence for multiple common risk alleles for schizophrenia and bipolar disorder and for overlap in genetic risk. Mol Psychiatry 2009; 14: 252–260.

  35. 35.

    , , , , , et al. A genome-wide association study implicates diacylglycerol kinase eta (DGKH) and several other genes in the etiology of bipolar disorder. Mol Psychiatry 2008; 13: 197–207.

  36. 36.

    , , , . Diacylglycerol kinase beta promotes dendritic outgrowth and spine maturation in developing hippocompal neurons. BMC Neurosci 2009; 10: 99.

  37. 37.

    , , , , , et al. Disruption of the midkine gene (Mdk) resulted in altered expression of a calcium binding protein in the hippocampus of infant mice and their abnormal behaviour. Genes Cells 1998; 3: 811–822.

  38. 38.

    , , , , , . Midkine, a novel neurpotropic factor, is present in senile plaques of Alzheimer disease. Biochem Biophys Res Commun 1993; 192: 246–251.

  39. 39.

    , , , , , et al. Dopaminergic hypofunctions and preulse inhibition deficits in mice lacking midkine. Prog Neuropsychopharmacol Biol Psychiatry 2009; 33: 541–546.

  40. 40.

    , , , , . Realistic expectations of prepulse inhibition in translational models for schizophrenia research. Psychopharmacology (Berl) 2008; 199: 331–388.

  41. 41.

    , , , , , . SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics 2008; 24: 2938–2939.

  42. 42.

    , , , , , . False positives in imaging genetics. NeuroImage 2008; 40: 655–661.

  43. 43.

    , , , , , . A cross-sectional and longitudinal magnetic resonance imaging study of cingulate gyrus gray matter volume abnormalities in first-episode schizophrenia and first-episode affective psychosis. Arch Gen Psychiatry 2008; 65: 746–760.

  44. 44.

    , , , , , et al. Evidence that altered amygdala activity in schizophrenia is related to clinical state and not genetic risk. Am J Psychiatry 2009; 166: 216–225.

Download references

Acknowledgements

We are grateful to all of the patients who contributed to this study. We also thank the probands from the community-based cohorts of PopGen, KORA, the Heinz Nixdorf Recall (HNR) study, the Rotterdam Elderly Study, as well as the Munich controls. This study was supported by the German Federal Ministry of Education and Research (BMBF), within the context of the National Genome Research Network plus (NGFNplus) and the MooDS-Net (grant 01GS08144 to SC and MMN, grant 01GS08147 to MR). MMN also received support from the Alfried Krupp von Bohlen und Halbach-Stiftung. Further funding came from the European Union Seventh Framework Programme (FP7/2007–2011) under grant agreement no. 242257 (ADAMS). The Heinz Nixdorf Recall cohort was established with the support of the Heinz Nixdorf Foundation (Dr G Schmidt, Chairman). This study was also supported by The Polish Ministry of Science and Higher Education (grant N N402 244035 to PMC), The Danish Council for Strategic Research (grant no. 2101–07-0059), H Lundbeck A/S, the Faculty of Health Sciences, Aarhus University and the Stanley Medical Research Institute. Genotyping of the Dutch samples from Utrecht was sponsored by NIMH funding, R01 MH078075.

Author information

Author notes

    • M Rietschel
    •  & M Mattheisen

    These authors contributed equally to this work.

    • R A Ophoff
    •  & S Cichon

    These authors contributed equally to this work.

Affiliations

  1. Department of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, University of Heidelberg, Mannheim, Germany

    • M Rietschel
    • , J Strohmaier
    • , R Breuer
    • , C Schmael
    • , S Meier
    • , J Nikitopoulos
    •  & T G Schulze
  2. Department of Genomics, Life and Brain Center, University of Bonn, Bonn, Germany

    • M Mattheisen
    • , F Degenhardt
    • , T W Mühleisen
    • , S Herms
    • , B Haenisch
    • , L Priebe
    • , F B Basmanav
    • , A J Forstner
    • , P Hoffmann
    • , M Alexander
    • , J Schumacher
    • , M M Nöthen
    •  & S Cichon
  3. Institute of Human Genetics, University of Bonn, Bonn, Germany

    • M Mattheisen
    • , F Degenhardt
    • , T W Mühleisen
    • , S Herms
    • , B Haenisch
    • , L Priebe
    • , F B Basmanav
    • , A J Forstner
    • , P Hoffmann
    • , M Alexander
    • , J Schumacher
    • , M M Nöthen
    •  & S Cichon
  4. Institute for Medical Biometry, Informatics, and Epidemiology, University of Bonn, Bonn, Germany

    • M Mattheisen
    • , M Steffens
    •  & T F Wienker
  5. For a full list of members see Appendix

  6. Department of Psychiatry and Psychotherapy, Central Institute of Mental Health, Mannheim, Germany

    • P Kirsch
    • , C Esslinger
    •  & A Meyer-Lindenberg
  7. Department of Human Genetics, Aarhus University, Aarhus, Denmark

    • D Demontis
    •  & A D Børglum
  8. Laboratory of Psychiatric Genetics, Department of Psychiatry, Poznan University of Medical Sciences, Poznan, Poland

    • P M Czerski
    • , P Kapelski
    •  & J Hauser
  9. Division of Molecular and Clinical Neurobiology, Department of Psychiatry, Ludwig-Maximilians-University, Munich, Germany

    • I Giegling
    •  & D Rujescu
  10. Department of Medical Genetics and Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands

    • E Strengman
    •  & R A Ophoff
  11. Centre for Psychiatric Research, Aarhus University Hospital, Risskov, Denmark

    • O Mors
  12. National Centre for Register-based Research, Aarhus University, Aarhus, Denmark

    • P B Mortensen
  13. Section of Neonatal Screening and Hormones, Department Clinical Chemistry and Immunology, The State Serum Institute, Copenhagen, Denmark

    • D M Hougaard
  14. AROS Aplied Biotechnology A/S and Department of Molecular Medicine, Aarhus University Hospital, Skejby, Aarhus, Denmark

    • T Ørntoft
  15. Institute of Medical Informatics, Biometry, and Epidemiology, University Duisburg-Essen, Essen, Germany

    • S Moebus
  16. Department of Psychiatry, University of Bonn, Bonn, Germany

    • R Mössner
    •  & W Maier
  17. Institute of Epidemiology, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany

    • H-E Wichmann
  18. Institute of Medical Informatics, Biometry and Epidemiology, Ludwig-Maximilians-University, Munich, Germany

    • H-E Wichmann
  19. Institute of Clinical Molecular Biology, Christian-Albrechts University Kiel, Kiel, Germany

    • S Schreiber
  20. Department of Internal Medicine, Genetics Laboratory, Eramsmus Medical Center Rotterdam, Rotterdam, The Netherlands

    • F Rivandeneira
    •  & A G Uitterlinden
  21. Department of Epidemiology, Erasmus University Medical Center, Rotterdam, The Netherlands

    • F Rivandeneira
    • , A Hofman
    •  & A G Uitterlinden
  22. UCLA Center for Neurobehavioral Genetics, Los Angeles, CA, USA

    • R M Cantor
    •  & R A Ophoff
  23. Department of Psychiatry, Division of Mind and Brain Research, Charite Medical University Berlin, Charité Campus Mitte, Berlin, Germany

    • S Erk
    •  & H Walter
  24. Division of Medical Psychology, Department of Psychiatry and Psychotherapy, University of Bonn, Bonn, Germany

    • S Erk
    •  & H Walter
  25. Department of Psychiatry and Psychotherapy, Georg-August-University Göttingen, Göttingen, Germany

    • T G Schulze
  26. Medical Research Council (MRC) Centre for Neuropsychiatric Genetics and Genomics, Department of Psychological Medicine and Neurology, School of Medicine, Cardiff University, Cardiff, UK

    • N Craddock
    • , M J Owen
    •  & M C O'Donovan
  27. Centre for Psychiatric Research, Aarhus University Hospital, Risskov, Denmark

    • A D Børglum
  28. Institute of Neuroscience and Medicine (INM-1), Research Center Juelich, Juelich, Germany

    • S Cichon

Consortia

  1. GROUP Investigators

    1. For a full list of members see Appendix
  2. Genetic Risk and Outcome in Psychosis (GROUP Investigators)

    René S Kahn1, Don H Linszen2, Jim van Os3, Durk Wiersma4, Richard Bruggeman4, Wiepke Cahn1, Lieuwe de Haan2, Lydia Krabbendam3 and Inez Myin-Germeys3 1Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands;2Academic Medical Centre University of Amsterdam, Department of Psychiatry, Amsterdam, The Netherlands;3Maastricht University Medical Centre, South Limburg Mental Health Research and Teaching Network, Maastricht, The Netherlands and4University Medical Center Groningen, Department of Psychiatry, University of Groningen, Groningen, The Netherlands.

  3. SGENE-plus Consortium

    1. For a full list of members see Appendix
  4. (Only those persons responsible for the samples of Replication 2 are listed)

    Hreinn Stefansson1, Stacy Steinberg1, Omar Gustafsson1, Engilbert Sigurdsson2, Hannes Petursson2, Augustine Kong1, Kari Stefansson1, Olli PH Pietiläinen3,4, Annamari Tuulio-Henriksson5, Tiina Paunio6, Jouko Lonnqvist5, Jaana Suvisaari5, Leena Peltonen3,4, Mirella Ruggeri7, Sarah Tosato7, Muriel Walshe8, Robin Murray8, David A Collier8, David St Clair9, Thomas Hansen10, Andres Ingason10, Klaus D Jakobsen10, Linh Duong10, Thomas Werge10, Ingrid Melle11, Ole A Andreassen12, Srdjan Djurovic12, István Bitter13, János M Réthelyi13, Lilia Abramova14, Vasily Kaleda14, Vera Golimbet14, Erik G Jönsson15, Lars Terenius15, Ingrid Agartz15, Ruud van Winkel16,17, Gunter Kenis16, Marc De Hert17, Jan Veldink18, Carsten Wiuf19 and Michael Didriksen20 1deCODE genetics, Sturlugata 8, IS-101 Reykjavik, Iceland;2Department of Psychiatry, National University Hospital, Hringbraut, IS-101 Reykjavik, Iceland;3Institute for Molecular Medicine Finland FIMM and Department of Medical Genetics, University of Helsinki, FI-00014 Helsinki, Finland;4Wellcome Trust Sanger Institute, Cambridge CB10 1SA, UK;5Department of Mental Health and Alcohol Research, National Public Health Institute, Mannerheimintie 166, FI-00300 Helsinki, Finland;6Health Genomics Unit, National Institute for Health and Welfare THL, FI-00270 Helsinki, Finland;7Section of Psychiatry and Clinical Psychology, University of Verona, 37134 Verona, Italy;8Institute of Psychiatry, King's College, London SE5 8AF, UK;9Department of Mental Health, University of Aberdeen, Royal Cornhill Hospital, Aberdeen AB25 2ZD, UK;10Institute of Biological Psychiatry, Mental Health Centre Sct. Hans Copenhagen University Hospital, DK-4000 Roskilde, Denmark;11Department of Medical Genetics, Ulleval University Hospital, Institute of Psychiatry, University of Oslo, Kirkeveien 166, N-0407 Oslo, Norway;12Department of Psychiatry, Ullevål University Hospital and Institute of Psychiatry, University of Oslo, Kirkeveien 166, N-0407 Oslo, Norway;13Semmelweis University, Department of Psychiatry and Psychotherapy, Budapest, Hungary;14Mental Health Research Center, Russian Academy of Medical Sciences, Zagorodnoe sh. 2/2, 117152 Moscow, Russia;15Department of Clinical Neuroscience, HUBIN project, Karolinska Institutet and Hospital, R5:00, SE-171 76 Stockholm, Sweden;16Department of Psychiatry and Psychology, School of Mental Health and Neuroscience, European Graduate School of Neuroscience (EURON), South Limburg Mental Health Research and Teaching Network (SEARCH), Maastricht University Medical Centre, 6200 MD Maastricht, The Netherlands;17University Psychiatric Centre Catholic University Leuven, Leuvensesteenweg 517, B-3070 Kortenberg, Belgium;18Rudolf Magnus Institute of Neuroscience, Department of Neurology, UMC Utrecht, 3584 CG Utrecht, The Netherlands;19Bioinformatics Research Centre, Aarhus University, DK-8000 Aarhus C, Denmark and20Discovery Biology Research, H Lundbeck A/S, Ottiliavej 9, DK-2500 Valby, Denmark.

Authors

  1. Search for M Rietschel in:

  2. Search for M Mattheisen in:

  3. Search for F Degenhardt in:

  4. Search for T W Mühleisen in:

  5. Search for P Kirsch in:

  6. Search for C Esslinger in:

  7. Search for S Herms in:

  8. Search for D Demontis in:

  9. Search for M Steffens in:

  10. Search for J Strohmaier in:

  11. Search for B Haenisch in:

  12. Search for R Breuer in:

  13. Search for P M Czerski in:

  14. Search for I Giegling in:

  15. Search for E Strengman in:

  16. Search for C Schmael in:

  17. Search for O Mors in:

  18. Search for P B Mortensen in:

  19. Search for D M Hougaard in:

  20. Search for T Ørntoft in:

  21. Search for P Kapelski in:

  22. Search for L Priebe in:

  23. Search for F B Basmanav in:

  24. Search for A J Forstner in:

  25. Search for P Hoffmann in:

  26. Search for S Meier in:

  27. Search for J Nikitopoulos in:

  28. Search for S Moebus in:

  29. Search for M Alexander in:

  30. Search for R Mössner in:

  31. Search for H-E Wichmann in:

  32. Search for S Schreiber in:

  33. Search for F Rivandeneira in:

  34. Search for A Hofman in:

  35. Search for A G Uitterlinden in:

  36. Search for T F Wienker in:

  37. Search for J Schumacher in:

  38. Search for J Hauser in:

  39. Search for W Maier in:

  40. Search for R M Cantor in:

  41. Search for S Erk in:

  42. Search for T G Schulze in:

  43. Search for N Craddock in:

  44. Search for M J Owen in:

  45. Search for M C O'Donovan in:

  46. Search for A D Børglum in:

  47. Search for D Rujescu in:

  48. Search for H Walter in:

  49. Search for A Meyer-Lindenberg in:

  50. Search for M M Nöthen in:

  51. Search for R A Ophoff in:

  52. Search for S Cichon in:

Competing interests

The authors declare no conflict of interest.

Corresponding authors

Correspondence to M Rietschel or S Cichon.

Supplementary information

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/mp.2011.80

Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp)

Further reading