Skip to main content
PMC home page
Human Molecular Genetics logoLink to Human Molecular Genetics
. 2009 May 4;18(14):2719–2727. doi: 10.1093/hmg/ddp204

The DISC locus and schizophrenia: evidence from an association study in a central European sample and from a meta-analysis across different European populations

Johannes Schumacher 1,2,*, Gonzalo Laje 1, Rami Abou Jamra 2, Tim Becker 3, Thomas W Mühleisen 4, Catalina Vasilescu 4, Manuel Mattheisen 2,4, Stefan Herms 4, Per Hoffmann 4, Axel M Hillmer 4, Alexander Georgi 5, Christine Herold 3, Thomas G Schulze 1,5, Peter Propping 2, Marcella Rietschel 5, Francis J McMahon 1, Markus M Nöthen 2,4, Sven Cichon 2,4
PMCID: PMC2701338  PMID: 19414483

Abstract

Association studies, as well as the initial translocation family study, identified the gene Disrupted-In-Schizophrenia-1 (DISC1) as a risk factor for schizophrenia. DISC1 encodes a multifunctional scaffold protein involved in neurodevelopmental processes implicated in the etiology of schizophrenia. The present study explores the contribution of the DISC locus to schizophrenia using three different approaches: (i) systematic association mapping aimed at detecting DISC risk variants in a schizophrenia sample from a central European population (556 SNPs, n = 1621 individuals). In this homogenous sample, a circumscribed DISC1 interval in intron 9 was significantly associated with schizophrenia in females (P = 4 × 10−5) and contributed most strongly to early-onset cases (P = 9 × 10−5). The odds ratios (ORs) were in the range of 1.46–1.88. (ii) The same sample was used to test for the locus-specific SNP–SNP interaction most recently associated with schizophrenia. Our results confirm the SNP interplay effect between rs1538979 and rs821633 that significantly conferred disease risk in male patients with schizophrenia (P = 0.016, OR 1.57). (iii) In order to detect additional schizophrenia variants, a meta-analysis was performed using nine schizophrenia samples from different European populations (50 SNPs, n = 10 064 individuals maximum, n = 3694 minimum). We found evidence for a common schizophrenia risk interval within DISC1 intron 4–6 (P = 0.002, OR 1.27). The findings point to a complex association between schizophrenia and DISC, including the presence of different risk loci and SNP interplay effects. Furthermore, our phenotype–genotype results—including the consideration of sex-specific effects—highlight the value of homogenous samples in mapping risk genes for schizophrenia in general, and at the DISC locus in particular.

INTRODUCTION

Disrupted-In-Schizophrenia-1 (DISC1, MIM 605210) is one of the most frequently implicated risk genes for schizophrenia (MIM 181500) (1). The DISC locus was first discovered through a breakpoint mapping on chromosome 1q42 in a Scottish family with a balanced translocation (1,11) (q42.1;q14.3) that was found to segregate with a wide spectrum of mental disorders (2,3). Linkage disequilibrium (LD) studies using different schizophrenia samples were then performed and provided evidence that the DISC locus is also involved in more common forms of schizophrenia (4) (Supplementary Material, Table S1). Furthermore, DISC1 is an interesting gene for neuropsychiatric disorders at the functional level. It encodes a multifunctional scaffold protein involved in neuronal migration, cortical layering and hippocampal formation, all of which are implicated in the etiology of schizophrenia (5). Despite these encouraging findings, the genetic risk variants at the DISC locus remain unidentified. In addition, different DISC intervals have shown association across studies, and this is complicated by the fact that most studies analyzed different markers and that their coverage often did not suffice to systematically detect LD (Supplementary Material, Table S1). To date, only two studies have performed a locus-specific LD mapping using a sufficient marker grid (6,7).

This study investigated the contribution of the DISC locus in a large schizophrenia case–control sample from a central European population (n = 1621 individuals). In order to map the schizophrenia risk variants in this homogenous sample as precisely as possible, we performed a systematic LD mapping using 556 variants [121 genotyped, 435 imputed single nucleotide polymorphisms (SNPs)]. In addition, we tested for locus-specific interaction effects by analyzing the marker–marker interplay at the DISC locus most recently identified in schizophrenia (7). The study also attempted to identify more common schizophrenia variants at the DISC locus across nine case–control samples obtained from different European populations. Notably, this combined analysis (50 markers, n = 10 064 individuals maximum, n = 3694 individuals minimum) is the first DISC meta-analysis in schizophrenia to date.

RESULTS

LD mapping in the central European schizophrenia sample

Sample homogeneity and haplotype structure.

Our initial step was to assess the genetic homogeneity of our cases and controls and— in accordance with previous studies using this sample (8)—the Structure analysis suggested that our sample represented a homogenous population (see Materials and Methods and Supplementary Material).

Next, we explored the genomic architecture at the DISC locus. Using the haplotype definition of Gabriel et al. (9) and all 556 analyzed markers, we identified a total of 32 different haplotype blocks across the target region (see Materials and Methods, Supplementary Material, Table S2 and Supplementary Material, Fig. S1).

Single-marker association analysis.

For the DISC LD mapping, we stratified our sample according to gender. This procedure represents an a priori approach, as previous studies have reported sex-specific DISC associations in schizophrenia (Supplementary Material, Table S1). In addition, we stratified the patient sample in a post hoc analysis for specific phenotypic characteristics that may have represented greater etiological homogeneity (see Materials and Methods). To prevent type-I association errors, we performed 10 000 permutations of each of the markers and samples (all, female, male and using all phenotypic strata tested in all three samples).

After the permutation, none of the DISC markers showed an association in the entire schizophrenia sample or in the male sub-sample (Supplementary Material, Table S2). In contrast, the female sub-sample showed a significant association in a circumscribed DISC1 intron 9 interval. Alleles of three markers, all belonging to haplotype block 22, were significantly more frequent in female patients with schizophrenia than female controls (P-values between 0.0002–0.0006 and 0.0225–0.0475 corrected) (Table 1, Supplementary Material, Fig. S1). The corresponding odds ratios (ORs) ranged between 1.46 and 1.48. In addition, two genotyped and 17 imputed SNPs within block 22 showed a nominally significant association (Table 1).

Table 1.

Single-marker association analysis in female cases and female controls within DISC1 haplotype block 22

SNPa Positionb Inter-bp MAc Controlsd All (n = 254)
 EAO (n = 81)
Casese P-valuef OR P pg Casee P-valuef OR P pg
rs9432024 230073842 G 0.397 0.313 0.0008 0.69 0.12 0.265 0.0020 0.55 0.3
rs1015101 230074317 475 G 0.363 0.458 0.0002 1.48 0.02 0.519 0.0003 1.88 0.03
rs9432025 230074347 30 C 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs9431706 230074383 36 A 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs7553949 230074415 32 G 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs9432028 230075733 1318 A 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs9431708 230076025 292 G 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs999710 230077566 1541 T 0.363 0.458 0.0002 1.48 0.02 0.519 0.0003 1.88 0.03
rs7549204 230077793 227 A 0.228 0.186 0.0473 0.77 0.99 0.13 0.0051 0.5 0.52
rs999708 230078214 421 G 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs1009587 230078572 358 A 0.364 0.457 0.0003 1.47 0.05 0.519 0.0003 1.87 0.07
rs9431711 230080135 1563 A 0.23 0.183 0.0268 0.75 0.83 0.13 0.0051 0.5 0.31
rs9432040 230080739 604 G 0.514 0.436 0.0031 0.73 0.36 0.377 0.0020 0.58 0.3
rs7535913 230081033 294 T 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs2356606 230081663 630 G 0.23 0.186 0.0409 0.77 0.99 0.13 0.0051 0.5 0.52
rs7534369 230082857 1194 A 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs12740724 230083249 392 G 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs11122357 230083676 427 A 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs9431714 230087871 4195 A 0.409 0.333 0.0028 0.72 0.34 0.284 0.0031 0.57 0.38
rs4658890 230090704 2833 G 0.529 0.441 0.0008 0.7 0.07 0.605 0.0018 1.73 0.14
rs4333837 230091773 1069 G 0.33 0.417 0.0006 1.46 0.04 0.469 0.0009 1.79 0.08
rs9659649 230094131 2358 T 0.332 0.285 0.0532 0.8 1 0.235 0.0154 0.62 0.84
rs1407599 230094534 403 T 0.327 0.415 0.0005 1.46 0.08 0.469 0.0006 1.82 0.12
DEP (n = 127) FAM (n = 67)
Casese P-valuef OR P pg Casese P-valuef OR P pg
rs9432024 230073842 0.342 0.1372 0.79 1.00 0.331 0.1773 0.77 1.00
rs1015101 230074317 0.445 0.0238 1.40 0.78 0.418 0.2409 1.25 1.00
rs9432025 230074347 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs9431706 230074383 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs7553949 230074415 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs9432028 230075733 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs9431708 230076025 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs999710 230077566 0.445 0.0238 1.40 0.78 0.418 0.2409 1.25 1.00
rs7549204 230077793 0.173 0.0626 0.71 1.00 0.172 0.1409 0.7 1.00
rs999708 230078214 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs1009587 230078572 0.445 0.0264 1.39 0.94 0.418 0.2537 1.24 1.00
rs9431711 230080135 0.173 0.0626 0.71 0.97 0.172 0.1409 0.7 1.00
rs9432040 230080739 0.469 0.2426 0.84 1.00 0.522 0.4757 1.14 1.00
rs7535913 230081033 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs2356606 230081663 0.173 0.0626 0.71 1.00 0.172 0.1409 0.7 1.00
rs7534369 230082857 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs12740724 230083249 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs11122357 230083676 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs9431714 230087871 0.362 0.1883 0.82 1.00 0.358 0.2689 0.81 1.00
rs4658890 230090704 0.551 0.0249 1.39 0.79 0.522 0.2586 1.24 1.00
rs4333837 230091773 0.409 0.0245 1.40 0.79 0.343 0.7819 1.06 1.00
rs9659649 230094131 0.256 0.0238 0.69 0.93 0.269 0.1473 0.74 1.00
rs1407599 230094534 0.409 0.0176 1.43 0.87 0.343 0.7097 1.08 1.00
Open in a new tab

Only markers with significant association findings are presented and only P-values <0.05 are given in bold. None of the SNPs showed Hardy–Weinberg equilibrium (HWE) deviation in controls (P > 0.05).

aSNP markers in bold are genotyped, all other SNP markers are imputed.

bSNP marker positions according to NCBI Build 36.

cMinor allele in controls (MA).

dMA frequencies (MAF) in controls.

eMAF in patients.

fP-values of the single-marker association analysis.

gP-values after permutation correction.

The detailed phenotype–genotype analysis revealed that female patients with an early age of onset (EAO; less than 21 years) predominantly contributed to this association. In this phenotypically more homogeneous sample, the genetic effect size increased by an OR of 1.88 at two of three formerly associated markers (P-values of 0.0003 and 0.0312, corrected) (Table 1). The additional strata [lifetime history of depressive symptoms (DEP), and family history of psychiatric illness (FAM)] did not strengthen the association in this sample (Table 1).

Our haplotype block 22 findings are consistent with some previously reported schizophrenia associations. Hodgkinson et al. (10) found an association in a European-American schizophrenia sample at all three markers tested in this region (rs1122359: P = 0.003; rs999710: P = 0.003; rs9432024: P = 0.019). We genotyped two of these markers and both showed a significant association in females with schizophrenia in our sample (rs999710: P = 0.0002; rs9432024: P = 0.0013) (Table 1, Supplementary Material, Table S3). Notably, the same alleles were associated with schizophrenia in both studies. Furthermore, some previous studies that did not use a sex-specific analysis detected no association between schizophrenia and rs999710 (11,12) or rs9432024 (13) (Supplementary Material, Table S3); these findings are consistent with our results, which revealed no association between schizophrenia and these markers in the all-gender analysis (Supplementary Material, Tables S2 and S3).

Supplementary Material, Table S3 lists three additional SNPs—in haplotype blocks 1, 12 and 27—where we and other independent studies observed a disease association in the same direction. However, none of the association signals in our sample were significant after the permutation correction.

Haplotype association analysis.

For the multi-marker analysis, we used the haplotype block boundaries defined by Gabriel et al. (9) and only individually genotyped markers. All haplotypes that showed significant differences in their case–control distributions are presented in Supplementary Material, Table S4. Consistent with the single-marker analysis, a nine-marker haplotype in block 22—GTCGACTGG—was significantly more frequent in females with schizophrenia than in female controls (41.1% versus 31.1%, P = 0.00009) (Table 2).

Table 2.

Haplotype association analysis in female cases and female controls within DISC1 haplotype blocks 22 and 23

Haplotype block Sample/sub-sample Haplotype % Patientsa % Controlsa P-value Global P-value
Block 22
9-SNP-haplotypeb Female GTCGACTGG 0.411 0.311 0.00009 0.0184
ACCAACGAG 0.162 0.210 0.01860
Female EAO GTCGACTGG 0.469 0.307 0.00008 0.0090
ACCAACGAG 0.130 0.213 0.01586
Block 23
4-SNP-haplotypec Female GTAT 0.070 0.036 0.00372 0.0448
Female EAO GTAT 0.098 0.035 0.00074 0.0155
Block 22–23
13-SNP-haplotyped Female GTCGACTGGGTGT 0.325 0.252 0.00279 0.0265
GTCGACTGGGTAT 0.066 0.033 0.00483
ACCAACGAGGTGT 0.151 0.201 0.01428
Female EAO GTCGACTGGGTGT 0.365 0.254 0.00464 0.0072
GTCGACTGGGTAT 0.089 0.033 0.00207
ACCAACGAGGTGT 0.132 0.204 0.03770
Block 22–23
12-SNP-haplotypee Female GTCGACTGGGTT 0.388 0.284 0.00004 0.0367
ACCAACGAGGTT 0.149 0.200 0.01146
Female EAO GTCGACTGGGTT 0.444 0.285 0.00009 0.0239
ACCAACGAGGTT 0.129 0.204 0.02841
Open in a new tab

Only haplotypes with significant differences between cases and controls are presented and only P-values <0.05 are given in bold.

aEstimated haplotype frequencies in patients and controls. Estimations are based on each analyzed sample (all female patients and all female controls, EAO female patients and all female controls).

bSNP-marker combination rs1015101–rs999710–rs11122355–rs9431711–rs17820909–rs11122362–rs4658890–rs4333837–rs11122366.

cSNP-marker combination rs11585831–rs967433–rs17770256–rs821723.

dSNP-marker combination rs1015101–rs999710–rs11122355–rs9431711–rs17820909–rs11122362–rs4658890–rs4333837–rs11122366–rs11585831–rs967433–rs17770256–rs821723.

eSNP-marker combination rs1015101–rs999710–rs11122355–rs9431711–rs17820909–rs11122362–rs4658890–rs4333837–rs11122366–rs11585831–rs967433–rs821723.

In addition, the neighboring haplotype block 23 showed an association in females with schizophrenia. The four-marker haplotype GTAT was significantly more frequent in individuals with schizophrenia than controls (7.0% versus 3.6%, P = 0.0037) (Table 2). Because the results suggested an extended haplotype architecture at this locus, we conducted a combined block 22–23 analysis. This revealed two 13-marker risk haplotypes—GTCGACTGGGTGT and GTCGACTGGGTAT—that were both significant as well as more frequent in females with schizophrenia than female controls (32.5% versus 25.2%, P = 0.0027, and 6.6% versus 3.3%, P = 0.0048) (Table 2). The less frequent 13-marker risk haplotype differed at SNP rs17770256 only compared with the common risk haplotype (alleles G versus A, underlined above and below). This haplotype most likely arose from the common risk haplotype after the mutation event had occurred. The combined 12-marker risk haplotype GTCGACTGGGT-G/A-T was more common in individuals with schizophrenia (38.8% versus 28.4% in controls) and was significantly associated with the disease (P = 0.00004) (Table 2).

Consistent with the single-marker results, the risk haplotypes were even more frequent in EAO females with schizophrenia (Table 2). In this homogenous sample, the most discriminative 12-marker risk haplotype had a frequency of 44.4% (versus 28.4% in controls) and was significantly associated with schizophrenia (P = 0.00009) (Table 2).

Interplay and functional DISC variants in the central European schizophrenia sample

Next, we tested for the locus-specific interaction observed in the schizophrenia study by Hennah et al. (7) and adopted their marker–marker–interplay approach (see Materials and Methods). At both of the previously implicated markers (7)—rs1538979 and rs821633—we stratified our cases and controls by grouping homo- and heterozygote carriers of the same alleles together (e.g. grouping of all individuals carrying the genotypes AA and AG as well as all individuals carrying the genotypes GG and AG at rs1538979). Within the allele-stratified samples, we assessed the association at the other reported interaction-marker. Because the procedure represents a confirmative test approach, no correction was applied in this analysis.

We observed the same rs1538979–rs821633 interaction that had previously been described in schizophrenia (7). The SNP rs1538979 (allele A) was significantly associated in males with schizophrenia (15.3% cases versus 10.3% in controls, P = 0.016, OR 1.57) by using rs821633 (allele G) as a conditional marker (Supplementary Material, Table S5). The association at rs1538979 was thus detectable only under the interaction model (Supplementary Material, Table S5). We also found that, on its own, allele G at rs821633 was a significant protective factor against schizophrenia (31.3% versus 35.9% in controls, P = 0.038, OR 0.83) (Supplementary Material, Table S2); notably, the same effect had been observed by Hennah et al. (7).

We then focused on those DISC1 variants that might be of functional relevance; these SNPs cause an amino acid substitution and were associated with schizophrenia in some previous studies (4). We found no association for rs6675281 (Leu607Phe) or rs821616 (Ser704Cys) in our overall or gender-specific analysis of individuals with schizophrenia (Supplementary Material, Table S2), as well as in our phenotypic strata analysis (data not shown).

DISC meta-analysis across European schizophrenia samples and studies

The third part of our study was a DISC meta-analysis using nine schizophrenia samples from different European populations (n = 10 064 maximum sample size/SNP, n = 3694 minimum sample size/SNP). All studies that used a gene-based association approach (DISC marker set of more than five SNPs/study) and that were performed on independent case–control samples with a broad definition of European ancestry (Supplementary Material, Table S6) were included. In total, we tested 50 locus-specific SNPs for association using the weighted Z-score method and Comprehensive Meta Analysis v. 2 (see Materials and Methods).

In this mixed European sample, evidence for common schizophrenia risk variants was found at 10 DISC SNPs (P-values between 0.002 and 0.048) (Table 3). All studies contributed equally to the findings (heterogeneity test: P-values between 0.988 and 0.100) (Table 3) and the funnel-plot analysis did not indicate the presence of publication bias effects for the associated variants (data not shown).

Table 3.

DISC meta-analysis in nine schizophrenia case–control samples from European populations

SNP Positiona Inter-bp Study and Sample Cases Controls OR z-value P-value P-value Heterogeneity
rs4658879 229809198 Present Study 782 839 1.21 2.087 0.037
Wood et al. (11) 311 291 1.00 0.000 1.000
Hennah et al. (7), Edinburgh 328 315 1.06 0.428 0.668
Hennah et al. (7), Helsinki 343 342 1.05 0.353 0.724
Hennah et al. (7), Aberdeen 256 255 1.22 1.256 0.209
Hennah et al. (7), London 348 351 1.08 0.538 0.591
Meta-analysis 2368 2393 1.12 2.126 0.033 0.848
rs1572899 229892113 82915 Present study 782 839 1.08 0.858 0.391
Hennah et al. (7), Edinburgh 328 315 1.17 1.117 0.264
Hennah et al. (7), Helsinki 343 342 1.00 0.022 0.982
Hennah et al. (7), Aberdeen 256 255 1.35 1.856 0.063
Hennah et al. (7), London 348 351 1.06 0.402 0.688
Hodgkinson et al. (10) 196 217 1.46 2.127 0.033
Meta-analysis 2253 2319 1.13 2.353 0.019 0.470
rs2812385 229930625 38512 Present study 782 839 1.04 0.447 0.655
Wood et al. (11) 311 291 1.04 0.256 0.798
Hennah et al. (7), Edinburgh 328 315 1.14 0.889 0.374
Hennah et al. (7), Helsinki 343 342 1.17 1.141 0.254
Hennah et al. (7), Aberdeen 256 255 1.46 2.343 0.019
Hennah et al. (7), London 348 351 1.08 0.571 0.568
Meta-analysis 2368 2393 1.12 2.097 0.036 0.582
rs1538979 229963491 32866 Present study 782 839 1.14 1.421 0.155
Sullivan et al. (13) 738 733 1.12 1.195 0.232
Hennah et al. (7), Edinburgh 328 315 1.03 0.229 0.819
Hennah et al. (7), Helsinki 343 342 1.37 2.269 0.023
Hennah et al. (7), Aberdeen 256 255 1.29 1.588 0.112
Hennah et al. (7), London 348 351 1.06 0.398 0.691
Meta-analysis 2795 2835 1.04 2.860 0.004 0.677
rs16854957 229979516 16025 Present study 782 839 0.85 0.171 0.864
Wood et al. (11) 311 291 0.79 0.330 0.741
Sullivan et al. (13) 738 733 0.84 0.066 0.947
Hennah et al. (7), Edinburgh 328 315 0.98 2.075 0.038
Hennah et al. (7), Helsinki 343 342 0.84 0.707 0.480
Hennah et al. (7), Aberdeen 256 255 0.93 2.394 0.017
Hennah et al. (7), London 348 351 0.96 1.677 0.093
Meta-analysis 3106 3126 1.22 2.368 0.017 0.247
rs17817356 229986844 7328 Present study 782 839 0.86 0.254 0.800
Sullivan et al. (13) 738 733 0.93 1.137 0.256
Hennah et al. (7), Edinburgh 328 315 0.77 0.166 0.868
Hennah et al. (7), Helsinki 343 342 0.98 2.137 0.033
Hennah et al. (7), Aberdeen 256 255 0.82 3.157 0.002
Hennah et al. (7), London 348 351 0.94 1.534 0.125
Meta-analysis 2795 2835 1.27 3.007 0.002 0.100
rs1322784 229995558 8714 Present study 782 839 0.92 1.012 0.311
Sanders et al. (12) 1870 2002 0.93 0.714 0.475
Hennah et al. (7). - Edinburgh 328 315 0.81 0.492 0.623
Hennah et al. (7), Helsinki 343 342 0.96 1.638 0.101
Hennah et al. (7), Aberdeen 256 255 0.80 0.576 0.565
Hennah et al. (7), London 348 351 0.91 1.260 0.208
Hodgkinson et al. (10) 196 217 0.97 1.800 0.072
Meta-analysis 4123 4321 1.19 2.430 0.015 0.716
rs860275 230102768 107210 Present study 782 839 1.00 2.004 0.045
Wood et al. (11) 311 291 0.79 0.331 0.741
Hennah et al. (7), Edinburgh 328 315 0.85 0.840 0.401
Hennah et al. (7), Helsinki 343 342 0.93 1.438 0.151
Hennah et al. (7), Aberdeen 256 255 0.93 1.519 0.129
Hennah et al. (7), London 348 351 0.84 0.674 0.501
Meta-analysis 2368 2393 1.29 2.896 0.004 0.938
rs7541019 230117808 15040 Present study 782 839 0.93 1.208 0.227
Sullivan et al. (13) 738 733 0.90 0.837 0.403
Hennah et al. (7), Edinburgh 328 315 0.94 1.523 0.128
Hennah et al. (7), Helsinki 343 342 0.82 0.572 0.568
Hennah et al. (7), Aberdeen 256 255 0.86 1.004 0.315
Hennah et al. (7), London 348 351 0.81 0.383 0.702
Meta-analysis 2795 2835 1.22 2.227 0.025 0.963
rs2038636 230121872 4064 Present study 782 839 0.95 1.361 0.174
Wood et al. (11) 311 291 0.87 0.997 0.319
Sanders et al. (12) 1870 2002 0.96 1.264 0.206
Meta-analysis 2963 3132 1.20 2.023 0.043 0.838
rs821577 230133680 11808 Present study 782 839 0.92 1.066 0.286
Sullivan et al. (13) 738 733 0.94 1.254 0.210
Hennah et al. (7), Edinburgh 328 315 0.80 0.373 0.709
Hennah et al. (7), Helsinki 343 342 0.91 1.303 0.193
Hennah et al. (7), Aberdeen 256 255 0.85 0.982 0.326
Hennah et al. (7), London 348 351 0.82 0.556 0.578
Meta-analysis 2795 2835 1.23 2.285 0.022 0.988
Open in a new tab

In bold results with significant disease association across samples are presented (P-values association < 0.05, P-values heterogeneity > 0.05) using the weighted Z-score method (Comprehensive Meta Analysis v. 2)

aSNP marker positions according to NCBI Build 36.

Nevertheless, after correction by a factor of 23 was applied, only one marker, rs17817356, still remained significant (P = 0.002, P = 0.046 corrected, OR 1.27) (Table 3); this correction factor represents the number of haplotype blocks in which the 50 meta-analysis markers were located. Although the use of the haplotype structure observed in our sample might not be representative for all samples tested in the meta-analysis, a Bonferroni correction by a factor of 50 (number of tested markers) seemed to be overly conservative, given the strong inter-marker LD within each haplotype block. The significant SNP (rs17817356) is located within a DISC1 intron 4–6 interval, where four additional markers showed a nominally significant association in the meta-analysis (rs1538979 to rs1322784) (Table 3). Interestingly, rs1538979—the marker that showed an interplay effect in our study as well as the one by Hennah et al. (7)—is one of the associated markers in the identified intron 4–6 interval.

DISCUSSION

Notably, this analysis is one of the most comprehensive DISC studies in schizophrenia to date and comprises the application of three different, but complementary, approaches. Specifically, (i) in the first schizophrenia sample from a central European population, we analyzed 556 DISC markers (n = 1621). The LD mapping included a detailed phenotype–genotype examination that further increased sample homogeneity. In this population, a circumscribed DISC1 intron 9 interval was associated with schizophrenia in females (P = 0.00004), and this risk effect was strongest in EAO female patients (P = 0.00009, OR 1.88); (ii) we then assessed the previously reported locus-specific DISC interaction (7) using the same central European sample. Our results confirmed the SNP interplay effect at rs1538979–rs821633 that significantly conferred risk of schizophrenia in males (P = 0.016, OR 1.57); (iii) finally, we performed a meta-analysis using 50 DISC markers and nine schizophrenia samples (n = 10 064 individuals maximum, n = 3694 minimum) with the goal of detecting additional schizophrenia variants across European populations. We found evidence for a common risk interval within DISC1 intron 4–6 (P = 0.002, OR 1.27).

The DISC findings described here point to a complex association profile in schizophrenia, including the presence of different risk loci and SNP interplay effects. In addition, our phenotype–genotype results highlight the value of homogenous samples in mapping risk genes for complex disorders. Below, we review some of the relevant findings in this area, and discuss the implications of the current study.

Different schizophrenia risk intervals at the DISC locus

The DISC1 intron 9 risk region implicated in schizophrenia risk for females in the central European sample has, to date, attracted only limited attention. Hodgkinson et al. (10) found an association at three SNPs within this locus in a European-American schizophrenia sample. Two of those three SNPs were analyzed in our study and showed a significant association (rs999710: P = 0.0002; rs9432024: P = 0.0013). In addition, Hennah et al. (7) found an association within haplotype block 22 in two schizophrenia samples (from London and Edinburgh). Consistent with the results of the present study, this association was most prominent in females. However, the same alleles were implicated by the present study and the one by Hodgkinson et al. (10), whereas the findings of Hennah et al. (7) differ on the allelic level. This may reflect the presence of a different haplotype structure or risk haplotypes in the UK/Scottish and central European populations.

In addition, the meta-analysis points to a common schizophrenia risk interval in DISC1 intron 4–6. Within this region, one marker was significantly associated with disease risk, and four neighboring markers showed a nominally significant association. The clustering may reflect the robustness of our findings. In addition, one of the associated variants in this interval was rs1538979, which conferred risk of schizophrenia in the present study, as well as in a previous one (7), through a SNP–SNP interaction.

On the basis of our findings and those of independent association studies, it appears that both DISC1 regions (within intron 9 and intron 4–6) are interesting candidates for schizophrenia follow-up studies. Furthermore, the extent of the identified risk intervals are circumscribed; the intron 9 interval spans between 27 and 47 kb (defined by markers of haplotype blocks 22 and 23 as well as 21 and 24), and the size of the intron 4–6 interval spans between 32 and 62 kb (inter-marker distances of the associated SNPs and first flanking markers without association). The limited extent of both regions should facilitate re-sequencing efforts, an important next step in determining the true underlying DISC risk variant(s) at both loci.

Locus-specific DISC interaction

The rs1538979–rs821633 interaction confirms the recently reported SNP interplay effects at the DISC locus (7). In that study as well as in the present one, the combination of the same alleles at both markers increased schizophrenia risk. Furthermore, we confirmed that rs821633-allele G by itself was a protective factor (7).

Nevertheless, the results obtained here are not a complete replication. Hennah et al. (7) obtained their results in female individuals with schizophrenia, whereas we found ours in males with schizophrenia. The findings may point to differences in the underlying genetic structure of schizophrenia between both populations. Alternatively, additional variants (at other loci) may be involved in this two-marker interaction and they show sex differences between the populations. Finally, we cannot exclude the possibility that our interplay findings represent type-I errors, although they (ad hoc) confirm both of the previously reported DISC effects (rs1538979–rs821633 interplay and protective effect of rs821633). We therefore recommend that the potential epistasis between both variants be an a priori consideration in all future schizophrenia DISC1 association studies.

Phenotype–genotype results

Our analysis points to a gender-specific DISC contribution in schizophrenia. This is supported by most other DISC studies that stratified their samples according to gender (7,14–16). The results are also in accordance with clinical findings, where age of onset, premorbid functioning, symptomatic characteristics and course of illness are well-established gender differences (17). Furthermore, the present phenotype–genotype analysis suggests that the identified DISC intron 9 interval contributes most strongly to EAO in females with schizophrenia. The finding is particularly consistent with the expectation that genetic effects should be more pronounced in patients with an earlier phenotypic expression. Unfortunately, we could not use a phenotypically more homogenous sample within our meta-analysis, because the raw data necessary for a detailed stratification (e. g. for gender) were not available for most of the included studies (see Materials and Methods). Despite this limitation, the comparatively large size of this sample provided sufficient power to detect more common risk loci at the DISC locus that contribute to schizophrenia across European populations.

Conclusion

Our results concur with previous DISC findings and point to a complex genetic architecture underlying schizophrenia at this locus, including the presence of private mutations as observed in the initially reported ‘breakpoint family’. To address this complexity, we applied two different, but complementary approaches. We identified a circumscribed risk interval in DISC1 intron 9 by using a sample that might better represent greater etiological homogeneity (central European population, gender-stratification, EAO). In addition, we identified a risk interval in the DISC1 intron 4–6 across European populations by performing a meta-analysis on a more heterogeneous but maximally sized sample. Both approaches are recommended when studying complex genetic disorders in general, and both led to the successful identification of two DISC risk intervals in the present study. On the basis of our results, future research to identify the underlying DISC risk variants at both loci can be conducted, with the ultimate goal of increasing our understanding of the etiology of schizophrenia.

MATERIALS AND METHODS

SNP marker selection

We genotyped 121 SNP markers, tagging all haplotypes at a frequency > 1% in the CEPH Utah (CEU) population using HapMap phase I+II data (HapMap release #20). Haplotype blocks were defined by the algorithm developed by Gabriel et al. (9) implemented in HaploView (v. 3.32) using the default settings. The selected SNPs covered 507 660 bp of the DISC locus (Supplementary Material, Table S2 and Fig. S1). The boundaries of the analyzed region ranged from 229 736 450 to 230 244 110 bp and the average inter-marker distance was 4195 bp.

Sample characteristics and diagnostic assessment

All 121 SNP markers were genotyped in a sample of 782 patients with schizophrenia (354 females, 428 males) and 839 controls (358 females, 481 males). All individuals were of German descent and were recruited in the local areas around the cities of Bonn and Mannheim (central western part of Germany). The genetic homogeneity of the sample was also determined using Structure v. 2.2 (18,19). (see Supplementary Material). The diagnostic assessment of patients was done using a best estimate approach and IPGS, a comprehensive inventory for phenotype characterization (20). This included a SCID-I Interview for Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) and the OPCRIT system, as well as a review of medical records (21,22). All control subjects were screened for psychiatric disorders (see Supplementary Material).

SNP marker genotyping and association statistics

Genotyping was performed using Golden Gate Assay (Illumina, San Diego, CA, USA). Genotype call rate was > 98% for all markers, and there were no replicate errors among the 2% of the samples genotyped in duplicate. Departure from Hardy–Weinberg equilibrium (HWE), single-marker and multi-marker haplotype association analyses were performed using PLINK software (http://pngu.mgh.harvard.edu/~purcell/plink/index.shtml) (23). SNPs showing significant deviation from HWE are listed in Supplementary Material, Table S7. To prevent type-I association errors, we also used PLINK to perform 10 000 permutations of each of the markers and samples studied [all, female, male, and using specific phenotypic characteristics (see below)]. However, PLINK only allows permutation results from single-marker association analyses. The results achieved by the multi-marker analyses were not permutated and are, thus, uncorrected. We also performed a power analysis of our samples using Genetic Power Calculator (http://pngu.mgh.harvard.edu/~purcell/gpc) (24). Under two different assumptions on the genotypic relative risk (GRR 1.25 and 1.5), the statistical power of the entire sample was between 0.53 and 0.96 [between 0.26 and 0.67 for the female sub-sample and between 0.32 and 0.78 for the male sub-sample (80% power, alpha 0.05, allelic test (df 1), risk allele frequency 0.10, disease prevalence 0.015)].

Haplotype definition and marker imputation

To obtain the most precise estimate of locus-specific genomic architecture, we used the genotypes of all 121 analyzed markers in our sample of 1621 individuals (cases and controls) and imputed genotypes of 435 additional SNPs using the CEU data released by the most updated phase of HapMap (HapMap release #22), as well as a Markov-chain approach as implemented by MACH (25). The haplotype block structure was defined according to Gabriel et al. (9) using all 556 markers and HaploView (v. 4.0 beta 14).

Detailed phenotype–genotype analysis

Given that some previous studies found evidence of a gender-specific DISC association in their schizophrenia samples (4) (Supplementary Material, Table S1), we further stratified our sample according to gender. In addition, we stratified our patient sample for specific phenotypic characteristics that might represent greater etiological homogeneity. Patients were stratified based on FAM in first- and second-degree relatives [FAM, n = 131 (n = 67 females, n = 64 males)], lifetime history of depressive symptoms [DEP, n = 232 (n = 127 females, n = 105 males)] and EAO [EAO < 21 years, n = 194 (n = 81 females, n = 113 males)]; for additional details on defining characteristics, see Supplementary Material. Supplementary Material, Table S8 provides a detailed overview of the composition of the stratified sample.

SNP marker interplay

For the locus-specific interaction, we used the marker–marker interplay approach reported by Hennah et al. (7). At the two SNPs most implicated in their study (7)—rs1538979 and rs821633—we stratified our cases and controls by grouping homo- and heterozygote carriers of the same alleles together. Using the allele-stratified samples at the conditional marker, we tested for an association at the other reported interplay SNP. Phenotypic strata were not used for the interplay tests. The sample sizes after stratifying for allele carriers and phenotypic characteristics appeared to be too small for a powerful association.

Meta-analysis across samples and studies

For the meta-analysis, we included all studies that used a gene-based association approach (defined by a marker set of more than five SNPs) and that were performed on independent case–control samples with a broad definition of European ancestry. Markers were only selected, if they had been analyzed in at least three of the included studies and showed the same allele tendencies in cases and controls across all samples. In total, we tested 50 markers equally distributed across the DISC locus (average inter-marker distance of 8.502 bp). The sample sizes analyzed per SNP ranged between 3.694 and 10.064 individuals.

For the association statistics, we applied a weighted Z-score method using Comprehensive Meta Analysis v. 2 (http://www.meta-analysis.com/pages/comparisons.html). This program also allows the detection of possible publication bias effects by performing a funnel-plot analysis, as well as an estimate of between-group heterogeneity by performing a heterogeneity test. A meta-analysis based on allelic and/or genotypic data or using information on gender could not be applied, because the raw data necessary for this test statistic were not available for most of the included samples and studies. The P-values for each marker and sample association were downloaded from the Schizophrenia Research Forum database (http://www.schizophreniaforum.org) (10–12) or, if not available on the website, results were obtained by contacting the groups personally (7,13).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

[Supplementary Data]
ddp204_index.html (872B, html)

FUNDING

This study was supported by the National Genomic Network of the ‘Bundesministerium für Bildung und Forschung’ (BMBF), a NIH/DFG Research Career Transition Award to J.S., the Alfried Krupp von Bohlen und Halbach-Stiftung to M.M.N. and the NIMH Intramural Research Program to F.J.M.

ACKNOWLEDGEMENTS

The authors are most grateful to the patients and their families for their cooperation. In addition, we wish to acknowledge the excellent technical assistance of Margrieta Alblas, Susanne Kopmann, Kathrin Kemmerling and Annegret Flosdorff. Ioline Henter provided invaluable editorial assistance and critical review during the preparation of the manuscript. Finally, the authors thank Dr William Hennah and Professor Patrick F. Sullivan for providing us with their schizophrenia association results at the DISC locus, which were published in Hennah et al. (7) and Sullivan et al. (13).

Conflict of Interest statement. None declared.

REFERENCES

  • 1.Craddock N., O'Donovan M.C., Owen M.J. The genetics of schizophrenia and bipolar disorder: dissecting psychosis. J. Med. Genet. 2005;42:193–204. doi: 10.1136/jmg.2005.030718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Millar J.K., Wilson-Annan J.C., Anderson S., Christie S., Taylor M.S., Semple C.A., Devon R.S., Clair D.M., Muir W.J., Blackwood D.H., et al. Disruption of two novel genes by a translocation co-segregating with schizophrenia. Hum. Mol. Genet. 2000;9:1415–1423. doi: 10.1093/hmg/9.9.1415. [DOI] [PubMed] [Google Scholar]
  • 3.St Clair D., Blackwood D., Muir W., Carothers A., Walker M., Spowart G., Gosden C., Evans H.J. Association within a family of a balanced autosomal translocation with major mental illness. Lancet. 1990;336:13–16. doi: 10.1016/0140-6736(90)91520-k. [DOI] [PubMed] [Google Scholar]
  • 4.Chubb J.E., Bradshaw N.J., Soares D.C., Porteous D.J., Millar J.K. The DISC locus in psychiatric illness. Mol. Psychiatry. 2008;13:36–64. doi: 10.1038/sj.mp.4002106. [DOI] [PubMed] [Google Scholar]
  • 5.Hennah W., Thomson P., Peltonen L., Porteous D. Genes and schizophrenia: beyond schizophrenia: the role of DISC1 in major mental illness. Schizophr. Bull. 2006;32:409–416. doi: 10.1093/schbul/sbj079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Thomson P.A., Wray N.R., Millar J.K., Evans K.L., Hellard S.L., Condie A., Muir W.J., Blackwood D.H., Porteous D.J. Association between the TRAX/DISC locus and both bipolar disorder and schizophrenia in the Scottish population. Mol. Psychiatry. 2005;10:657–668. doi: 10.1038/sj.mp.4001669. [DOI] [PubMed] [Google Scholar]
  • 7.Hennah W., Thomson P., McQuillin A., Bass N., Loukola A., Anjorin A., Blackwood D., Curtis D., Deary I.J., Harris S.E., et al. DISC1 association, heterogeneity and interplay in schizophrenia and bipolar disorder. Mol. Psychiatry. 2008 doi: 10.1038/mp.2008.22. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
  • 8.Stefansson H., Rujescu D., Cichon S., Pietiläinen O.P., Ingason A., Steinberg S., Fossdal R., Sigurdsson E., Sigmundsson T., Buizer-Voskamp J.E., et al. Large recurrent microdeletions associated with schizophrenia. Nature. 2008;455:232–236. doi: 10.1038/nature07229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Gabriel S.B., Schaffner S.F., Nguyen H., Moore J.M., Roy J., Blumenstiel B., Higgins J., DeFelice M., Lochner A., Faggart M., et al. The structure of haplotype blocks in the human genome. Science. 2002;296:2225–2229. doi: 10.1126/science.1069424. [DOI] [PubMed] [Google Scholar]
  • 10.Hodgkinson C.A., Goldman D., Jaeger J., Persaud S., Kane J.M., Lipsky R.H., Malhotra A.K. Disrupted in schizophrenia 1 (DISC1): association with schizophrenia, schizoaffective disorder, and bipolar disorder. Am. J. Hum. Genet. 2004;75:862–872. doi: 10.1086/425586. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Wood L.S., Pickering E.H., Dechairo B.M. Significant support for DAO as a schizophrenia susceptibility locus: examination of five genes putatively associated with schizophrenia. Biol. Psychiatry. 2007;61:1195–1199. doi: 10.1016/j.biopsych.2006.07.005. [DOI] [PubMed] [Google Scholar]
  • 12.Sanders A.R., Duan J., Levinson D.F., Shi J., He D., Hou C., Burrell G.J., Rice J.P., Nertney D.A., Olincy A., et al. No significant association of 14 candidate genes with schizophrenia in a large European ancestry sample: implications for psychiatric genetics. Am. J. Psychiatry. 2008;165:497–506. doi: 10.1176/appi.ajp.2007.07101573. [DOI] [PubMed] [Google Scholar]
  • 13.Sullivan P.F., Lin D., Tzeng J.Y., van den Oord E., Perkins D., Stroup T.S., Wagner M., Lee S., Wright F.A., Zou F., et al. Genomewide association for schizophrenia in the CATIE study: results of stage 1. Mol. Psychiatry. 2008;13:570–584. doi: 10.1038/mp.2008.25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hennah W., Varilo T., Kestilä M., Paunio T., Arajärvi R., Haukka J., Parker A., Martin R., Levitzky S., Partonen T., et al. Haplotype transmission analysis provides evidence of association for DISC1 to schizophrenia and suggests sex-dependent effects. Hum. Mol. Genet. 2003;12:3151–3159. doi: 10.1093/hmg/ddg341. [DOI] [PubMed] [Google Scholar]
  • 15.Chen Q.Y., Chen Q., Feng G.Y., Lindpaintner K., Wang L.J., Chen Z.X., Gao Z.S., Tang J.S., Huang G., He L. Case–control association study of Disrupted-in-Schizophrenia-1 (DISC1) gene and schizophrenia in the Chinese population. J. Psychiatr. Res. 2007;41:428–434. doi: 10.1016/j.jpsychires.2006.01.001. [DOI] [PubMed] [Google Scholar]
  • 16.Hashimoto R., Numakawa T., Ohnishi T., Kumamaru E., Yagasaki Y., Ishimoto T., Mori T., Nemoto K., Adachi N., Izumi A., et al. Impact of the DISC1 Ser704Cys polymorphism on risk for major depression, brain morphology and ERK signaling. Hum. Mol. Genet. 2006;15:3024–3033. doi: 10.1093/hmg/ddl244. [DOI] [PubMed] [Google Scholar]
  • 17.Aleman A., Kahn R.S., Selten J.P. Sex differences in the risk of schizophrenia: evidence from meta-analysis. Arch. Gen. Psychiatry. 2003;60:565–571. doi: 10.1001/archpsyc.60.6.565. [DOI] [PubMed] [Google Scholar]
  • 18.Falush D., Stephens M., Pritchard J.K. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164:1567–1587. doi: 10.1093/genetics/164.4.1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Pritchard J.K., Stephens M., Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155:945–959. doi: 10.1093/genetics/155.2.945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Fangerau H., Ohlraun S., Granath R.O., Nöthen M.M., Rietschel M., Schulze T.G. Computer-assisted phenotype characterization for genetic research in psychiatry. Hum. Hered. 2004;58:122–130. doi: 10.1159/000083538. [DOI] [PubMed] [Google Scholar]
  • 21.First M.B., Frances A.J., Pincus H.A., Vettorello N., Davis W.W. DSM-IV in progress. Changes in substance-related, schizophrenic, and other primarily adult disorders. Hosp. Community. Psychiatry. 1994;45:18–20. doi: 10.1176/ps.45.1.18. [DOI] [PubMed] [Google Scholar]
  • 22.McGuffin P., Farmer A., Harvey I. 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. doi: 10.1001/archpsyc.1991.01810320088015. [DOI] [PubMed] [Google Scholar]
  • 23.Purcell S., Neale B., Todd-Brown K., Thomas L., Ferreira M.A., Bender D., Maller J., Sklar P., de Bakker P.I., Daly M.J., Sham P.C. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 2007;81:559–575. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Purcell S., Cherny S.S., Sham P.C. Genetic power calculator: design of linkage and association genetic mapping studies of complex traits. Bioinformatics. 2003;19:149–150. doi: 10.1093/bioinformatics/19.1.149. [DOI] [PubMed] [Google Scholar]
  • 25.Li Y., Abecasis G.R. Mach 1.0: rapid haplotype reconstruction and missing genotype inference. Am. J. Hum. Genet. 2006;79:2290. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

[Supplementary Data]
ddp204_index.html (872B, html)
ddp204_1.pdf (468.2KB, pdf)

Articles from Human Molecular Genetics are provided here courtesy of Oxford University Press

RESOURCES