Co-author Virginia Keyes (K) Lasseter is deceased.
Article first published online: 28 DEC 2010
Copyright © 2010 Wiley-Liss, Inc.
American Journal of Medical Genetics Part B: Neuropsychiatric Genetics
Volume 156, Issue 2, pages 188–197, March 2011
How to Cite
Fallin, M. D., Lasseter, V. K., Liu, Y., Avramopoulos, D., McGrath, J., Wolyniec, P. S., Nestadt, G., Liang, K.-Y., Chen, P.-L., Valle, D. and Pulver, A. E. (2011), Linkage and association on 8p21.2-p21.1 in schizophrenia . Am. J. Med. Genet., 156: 188–197. doi: 10.1002/ajmg.b.31154
How to Cite this Article: Fallin MD, Lasseter VK, Liu Y, Avramopoulos D, McGrath J, Wolyniec PS, Nestadt G, Liang KY, Chen PL, Valle D, Pulver AE. 2011. Linkage and Association on 8p21.2-p21.1 in Schizophrenia. Am J Med Genet Part B 156:188–197.
The authors wish to dedicate this Manuscript to the memory of Ms. Lasseter.
- Issue published online: 8 FEB 2011
- Article first published online: 28 DEC 2010
- Manuscript Accepted: 17 NOV 2010
- Manuscript Received: 30 JUN 2010
- AMDec Foundation, Inc.
- NIMH. Grant Number: R01MH068406
- chromosome 8;
- Top of page
- MATERIALS AND METHODS
- Supporting Information
In the past decade, we and others have consistently reported linkage to a schizophrenia (SZ) susceptibility region on chromosome 8p21. Most recently, in the largest SZ linkage sample to date, a multi-site international collaboration performed a SNP-based linkage scan (∼6,000 SNPs; 831 pedigrees; 121 from Johns Hopkins (JHU)), that showed the strongest evidence for linkage in a 1 Mb region of chr 8p21 from rs1561817 to rs9797 (Zmax = 3.22, P = 0.0004) [Holmans et al. 2009. Mol Psychiatry]. We have investigated this 8p21 peak region further in two ways: first by linkage and family-based association in 106 8p-linked European-Caucasian (EUC) JHU pedigrees using 1,402 SNPs across a 4.4 Mb region surrounding the peak; second, by an independent case-control association study in the genetically more homogeneous Ashkenazim (AJ) (709 cases, 1,547 controls) using 970 SNPs in a further narrowed 2.8 Mb region. Family-based association analyses in EUC pedigrees and case-control analyses in AJ samples reveal significant associations for SNPs in and around DPYSL2 and ADRA1A, candidate genes previously associated with SZ in our work and others. Further, several independent gene expression studies have shown that DPYSL2 is differentially expressed in SZ brains [Beasley et al. 2006. Proteomics 6(11):3414–3425; Edgar et al. 2000. Mol Psychiatry 5(1):85–90; Johnston-Wilson et al. 2000. Mol Psychiatry 5(2):142–149] or in response to psychosis-inducing pharmaceuticals [Iwazaki et al. 2007. Proteomics 7(7):1131–1139; Paulson et al. 2004. Proteomics 4(3):819–825]. Taken together, this work further supports DPYSL2 and the surrounding genomic region as a susceptibility locus for SZ. © 2010 Wiley-Liss, Inc.
- Top of page
- MATERIALS AND METHODS
- Supporting Information
Schizophrenia (SZ) is a common (1% worldwide) heritable psychiatric disorder of thought and emotions characterized by psychotic manifestations including hallucinations and delusions, and by amotivation and anhedonia. SZ is associated with a high degree of disability, stigmatization, and reduced quality of life. Recent studies show patients with SZ have alarmingly high rates of risk factors for cardiovascular disease, consume less medical care than needed, and have a 20% shorter life expectancy when compared to non-schizophrenics [Hennekens, 2007]. The burden to society globally and to affected families is substantial [Rice, 1999]. Elucidating the genetic basis for SZ may lead to new biological targets for treatment or early identification of those at risk to begin early intervention.
Heritability of SZ is high (82–89%) [Cardno et al., 1999] suggesting genome-wide linkage and association efforts to define major susceptibility loci should be fruitful, even with the caveat that inheritance is likely influenced by both additive and epistatic genes, environment, and gene/environment interactions [Riley and Kendler, 2006]. There remains controversy as to what extent common variants of modest effect and rarer alleles of moderate effect contribute to SZ susceptibility [McClellan et al., 2007]. While GWAS efforts in SZ have now shown a few consistent candidate genes likely reflecting common variant risk factors [O'Donovan et al., 2008; Shifman et al., 2008; Kirov et al., 2009; Purcell et al., 2009; Shi et al., 2009; Stefansson et al., 2009; Owen et al., 2010], the uncertainty about whether rare versus common variants cause SZ [Owen et al., 2010] supports utilization of simultaneous linkage and association approaches to SZ gene mapping.
Over the past two decades, dozens of genome-wide linkage scans have been conducted, in both genetically outbred and other more isolated populations. Most studies have combined SZ with the related schizoaffective disorder (SZA) when searching for genes, assuming a common genetic etiology for these disorders. Linkage evidence for an SZ/SZA susceptibility locus (SSL) has been reported across almost every human chromosome, with notable signals on chromosomes 1q, 5q, 6p, 6q, 8p, 9q, 10q, 13q, and 22q [Owen et al., 2004]. Perhaps the most compelling evidence is for one or more SSL on chromosome 8p based on not only multiple independent strong linkage reports but also on various meta-analytical statistical applications [Tosato et al., 2005], though the region implicated in some studies was as large as the entire short arm of chromosome 8 (32 Mb) [Suarez et al., 2006].
We first reported linkage evidence to this region in 1995 [Pulver et al., 1995], followed by an additional report of linkage that achieved genome-wide significance (P = 0.0001) at D8S1771 in 54 outbred European-Caucasian (EUC) pedigrees and further support from a second independent set of 41 EUC pedigrees (P < 0.01) [Blouin et al., 1998]. Many other investigators subsequently reported strong linkage to this region [Kendler et al., 1996; Kaufmann et al., 1998; Brzustowicz et al., 1999; Gurling et al., 2001; DeLisi et al., 2002; Stefansson et al., 2002; Straub et al., 2002; Liu et al., 2005; Suarez et al., 2006] in family samples primarily composed of outbred European-Caucasian, Icelandic, African American, or Taiwanese subjects. Independent meta-analyses also implicate 8p in SZ/SZA susceptibility [Badner and Gershon, 2002; Lewis et al., 2003]. Recently, a highly informative multi-site collaborative SNP-based genome-wide linkage scan (∼6,000 SNPs; 7,476 affected in 831 pedigrees including 121 pedigrees from Johns Hopkins (JHU)), confirmed a significant linkage peak over a 1 Mb region (26.591–27.647 Mb) from SNP rs1561817 to SNP rs9797 on chromosome 8p21 (Zmax = 3.22, P = 0.0004) [Holmans et al., 2009]. Finally, a current meta-analysis of 31 scans (3,215 pedigrees, including all pedigrees in the recent multi-site collaborative SNP scan) identified two 30 cM bins meeting established criteria for suggestive genome-wide linkage: one on chromosome 2q and one on chr 8p. The 8p signal (P = 0.00826) improved when analyses were restricted to only European-ancestry pedigrees (P = 0.00065) [Ng et al., 2009].
The first fine mapping effort on 8p for SZ focused on a region localized in 33 extended Icelandic pedigrees and identified a risk haplotype (OR = 2.2) spanning ∼300 kb at the 5′ end of a plausible candidate gene, neuregulin 1 (NRG1) [Stefansson et al., 2002] at 32.624–32.720 Mb, which was later replicated in a Scottish case-control sample [Stefansson et al., 2003]. Findings for this gene among other groups have yielded mixed results, and some investigators suggest that the striking inconsistencies among variant alleles and haplotypes make this association difficult to interpret [Gogos and Gerber, 2006; Karayiorgou and Gogos, 2006]. In support of a role for NRG1 variants in SZ, Weinberger and his colleagues have reported that certain of the associated risk variants are in an alternative promoter that regulates the fetal expression of a brain-specific alternative NRG1 transcript (type IV NRG1) with the risk alleles conferring increased expression [Law et al., 2006; Tan et al., 2007]. In a systematic review of the association between NRG1 and SZ, Tosato et al., suggested that there may be other susceptibility genes on 8p [Tosato et al., 2005], and in support of this hypothesis, association with SZ or other neurodevelopmental disorders has been described for several other genes in this region [Kamiya et al., 2008; Tabares-Seisdedos and Rubenstein, 2009].
Several factors likely contribute to the failure of positional cloning and identification of the causal variants for an 8p SZ locus, including locus and allelic heterogeneity, gene/gene or gene/environment interactions, population admixture, phenotypic heterogeneity, or diagnostic imprecision, and limitations in technology and marker availability. To overcome these limitations, we took several simultaneous approaches. First, we performed follow-up genotyping of this region in both family and case-control samples. Further, we have focused our family-based localization efforts on those families showing the strongest 8p linkage. In an attempt to increase power via larger sample sizes and reduced allelic heterogeneity, we have also carried out an independent case-control study with participants of Ashkenazi Jewish (AJ) descent. In both samples, we selected SNPs within this critical region to ensure coverage of r2 ≥ 0.8 for all HapMap II SNPs above 1% frequency.
We began with a 4.4 Mb (24.806–29.166 Mb) region based on linkage in the 8-site consortium linkage analysis, where our 121 European-Caucasian families had the strongest signal in the recent SNP-based linkage scan. These families included the 54 families in our initial linkage reports of an SSL in this region. This gene-rich region contains peak markers common to multiple other studies [Blouin et al., 1998; Liu et al., 2005; Suarez et al., 2006] and includes 32 annotated genes and 5 hypothetical proteins. Eight of the genes have previously been studied in SZ (NEF3, NEFL, GNRH1, DPYSL2, ADRA1A, CHRNA2, PNOC, FZD3) [Tabares-Seisdedos and Rubenstein, 2009]. Based on our EUC linkage and family-based association in this region, we further narrowed our case-control efforts among AJ samples to a 2.8 Mb (25.697–28.490 Mb) region containing 28 of these genes. The strongest prior statistical evidence for association with SZ from these genes is for DPYSL2, based on transmission disequilibrium studies in our 274 Ashkenazi Jewish parent/child trios (P < 0.001) [Fallin et al., 2005] and reports of others [Nakata et al., 2003; Hong et al., 2005; Ujike et al., 2006]. Notably, the critical linkage region does not include the more centromeric NRG1 [Stefansson et al., 2002] or another recently associated more telomeric candidate, PCM1 [Gurling et al., 2006; Kamiya et al., 2008].
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
- Supporting Information
European-Caucasian schizophrenia families
Our European-Caucasian family sample (EUC) includes 701 genotyped individuals from 121 families recruited through multiple sources including our Maryland Epidemiologic Sample (MES) [Pulver and Bale, 1989], our nationwide advertising efforts, and national and international collaborators (US, Italy, Poland, Greece). All families are of European-Caucasian ancestry. These families include 296 cases (62.5% male, average age onset (AAO) = 22.0 years), with an average of 2.5 cases per family and 202 affected sibling pairs. All participants were examined by a psychiatrist or PhD level psychologist using a semi-structured instrument. Personality characteristics were assessed using the SID-P. All interviews were audiotaped and informants knowledgeable about the participant were interviewed in person. Diagnoses were made according to DSM-IV criteria, with final diagnoses and age at onset assigned through a consensus procedure [Blouin et al., 1998]. These 121 families were included as part of an international consortium to perform the largest linkage scan to date in 707 European-ancestry families [Holmans et al., 2009].
Based on the consortium-wide and our JHU-only linkage results, 106 linked families (88%) were chosen for follow-up linkage and association fine-mapping based on allele sharing amongst affected relatives at the SNP closest to the linkage peak in our families (rs9797). Characteristics of these families are reported in Table 1.
|Sample||Count||% male||Mean AAO/age*||% SZA|
|EUC families||Cases||296 (121 probands)||62.5||22.5||8.8|
|8p-linked EUC families||Cases||228 (106 probands)||63.5||20.9||8.8|
Ashkenazi case-control samples
The final case-control sample includes 709 cases and 1,419 controls. Of 772 AJ cases in our repository, 716 were from independent families, and 709 of these passed all quality control measures as described below. All cases met probable or definite DSM-IV schizophrenia or schizoaffective disorder criteria based on a consensus diagnosis (see diagnostic methods in [Fallin et al., 2005]). Controls were collected from two sources: the Ashkenazi Jewish Control Repository (AJCR) at Johns Hopkins and the New York Cancer Project (NYCP) Biorepository at the North Shore Long Island Jewish Research Institute. The NYCP is a collection of healthy volunteers (ages 30–69) who live in the New York metropolitan area [Foulkes et al., 2002] and has over 20,000 DNAs available for medical research purposes, with selection possible by age, sex, race, religion, and/or medical conditions. The AJCR samples were screened to exclude psychiatric disorders (See [Chen et al., 2009]). After quality control measures, 729 AJCR controls and 690 NYCP controls were used in the final analyses.
All cases and AJCR controls are of self-identified Ashkenazi Jewish ethnicity based on response to an ancestry questionnaire inquiring about ethnicity and country of origin of all four grandparents. Cases were from 38 US states (N = 686) and Canada (N = 23), with 29.8% from NY, 13.8% from CA, 8.3% from FL, 6.1% from PA, 5.4% from NJ, and <5% each from other states. The majority of AJ case grandparents were born in Europe with 65.8% from Eastern Europe 11.6% from Western Europe, 8.3% from Northern Europe, 1.4% from Southern Europe, and 12.9% from other regions or unknown. Grandparental origins of the AJCR sample include 55.2% Eastern Europe, 9.2% Western Europe, 7.0% Northern Europe, none from Southern Europe, and 29% unknown. For the NYCP, samples were initially selection based on having any of 6 available self-identified ethnicity codes that included 1007(=Ashkenazi). Of those, we excluded individuals with any ethnicity code indicating non-Caucasian or suggesting Sephardic Jewish heritage (specifically excluding Southern Europe, Eastern Mediterranean, North Africa, Middle East), or specific non-Ashkenazi Jewish. The grandparental origin is not known for the NYCP controls.
Characteristics of the final sample used in analysis, including only individuals and SNPs that passed QC filters and only unrelated cases, are presented in Table 1.
EUC genome-wide linkage
A panel of 6,008 SNPs (Illumina 4.0 linkage panel) were genotyped at the Center for Inherited Disease Research (CIDR) using the Illumina GoldenGate assay as part of an international collaborative linkage effort [Holmans et al., 2009]. SNPs were removed due to poor genotyping quality on the Illumina system, repeated Mendelian inconsistencies, and deviation from Hardy–Weinberg equilibrium at P < 0.001, leaving 5,861 autosomal or X chromosome SNPs for analyses. Further details of data cleaning for this linkage analysis are given in the collaborative linkage report [Holmans et al., 2009].
We selected 1,536 SNPs in a 4.4 Mb region of 8p (24,806,708–29,166,073) centered around the best linkage peak in preliminary analyses of the collaborative linkage scan (rs9797). SNPs were chosen based on Caucasian HapMap data (NCBI build 35) to have minor allele frequency (MAF) >0.01, record of submission to dbSNP by more than one source, and reported validation. SNPs were chosen to represent all Caucasian (CEU) HapMap SNPs in the region with r2 ≥ 0.8, with all non-synonymous SNPs also included. SNPs with Illumina Design Score <0.6 or with another SNP less than 100 bp away were excluded and replaced with comparable SNPs where possible. Genotyping was carried out on the Illumina Integrated BeadArrayTM System at the Johns Hopkins SNP Center. SNPs were filtered due to poorly defined clusters, excessive replicate or Mendelian errors, >50% missing genotypes (134 SNPs filtered).
AJ case-control fine-mapping
A target region of 2.8 Mb on 8p (25,697,329–28,490,949, build 36.1) centered around rs7817434 was chosen based on linkage and association evidence in the EUC families. 1,064 SNPs were chosen from public databases (NCBI, HapMap) in an iterative process beginning with all CEU HapMap genotypes in the region (4,977 SNPs). Tag SNPs were chosen to cover all HapMap SNPs with MAF >1% at r2 > 0.95, then supplemented with all coding SNPs not selected as a tag SNP, regardless of MAF. We further added any previously associated SNPs from our prior analyses [Fallin et al., 2005] and family-based association above and any novel polymorphisms identified through our own sequencing efforts focusing on the candidate gene, DPYSL2 [Liu et al., 2008]. Any SNPs on this list with Illumina Design Scores <0.8 were replaced. If this was not possible, design scores >0.4 were allowed. No pairs of SNPs were chosen within 60 bp of each other. SNPs were genotyped at the Johns Hopkins SNP Center using the Illumina Integrated BeadArrayTM System. Details of SNP quality for this genotyping effort are given in the results section.
Family based linkage
For the initial genome-wide linkage scan, Kong and Cox non-parametric LOD scores were estimated via the Spairs statistic under an exponential model in the MERLIN package [Abecasis et al., 2002], using independent SNP clusters with r2 < 0.16 between them to guard against bias due to LD [Boyles et al., 2005]. To select a subset of 8p-linked families for follow-up linkage and association, haplotypes were estimated for each family member in GeneHunter [Kruglyak et al., 1996] to allow selection of families with excess allele sharing at our peak marker (rs9797).
Family-based tests of association in these 106 8p-linked families were conducted using the FBAT software implemented in GoldenHelix [Horvath et al., 2001, 2004] (www.goldenhelix.com), assuming an additive genetic model. HaploView [Barrett et al., 2005] was used to examine linkage disequilibrium (LD) and haplotype block definitions (solid spine D′ > 0.8).
Initial data filtering and descriptive analyses were carried out in the R statistical environment (v2.7). The “hwexact” command in the hwde package (see cran.r-project.org) was used to test for Hardy–Weinberg disequilibrium using an exact test among controls. For SNPs also in HapMap, SNP genotypes from Illumina were recoded to match the strand genotyping of HapMap in order to combine samples for ancestry analysis. SNPs requiring allele recoding were identified through a custom script written in SAS v9.0 based on analysis of allele frequencies among HapMap CEU samples and our AJ samples. 449 SNPs were recoded: 327 A/G C/T; 73 A/C G/T; 27 C/G G/C; 22 A/T T/A. These were consistent with recodings identified in PLINK [Purcell et al., 2007].
Case and control genotypes were combined with unrelated YRI, CHB, and CEU HapMap genotypes to perform population ancestry analysis in STRUCTURE [Pritchard et al., 2000; Falush et al., 2003], assuming 3 parent populations. Ancestry proportions were estimated for each individual and plotted via the STRUCTURE software. To identify individuals likely to be outliers in our Ashkenazi set, we performed five separate STRUCTURE runs with new random seeds, and considered individuals with CEU membership estimates <0.5 in at least 3 of these 5 runs to be outliers. These are reported in the results section.
Genotype frequencies between NYCP controls and JHU controls were compared for all 970 SNPs via likelihood ratio tests in the logistic regression setting. The distribution of P values for this control-control comparison did not differ from expectation assuming they are from the same population, and control groups were combined for all case-control analyses. Odds ratios and confidence intervals for the increase in odds of SZ per copy of each minor allele (log-additive model) were estimated via generalized linear models (logistic regression) in R. Tests of association were based on likelihood ratios between models with and without the SNP term. P values were not corrected for multiple tests. Plots were generated in excel and R, after modifications of the package provided by de Bakker [Saxena et al., 2007].
- Top of page
- MATERIALS AND METHODS
- Supporting Information
EUC Genome-Wide Linkage
The strongest linkage evidence in the genome-wide analyses among all EUC families of the 8-site consortium was on chromosome 8p, with a peak between rs1561817 and rs9797, 26.59–27.65 Mb (NCBI build 36) [Holmans et al., 2009]. In our 121 JHU EUC families, the maximum non-parametric LOD score was 1.79 (P = 0.002) near rs9797 at 27,647,768 (Fig. 1).
We genotyped 1,536 SNPs in a 4.4 Mb region from 24.807 to 29,166 Mb (build 36, blue lines in Fig. 1). Of these, 1,402 SNPs met quality control criteria and were used for linkage confirmation and family-based association testing. Final intermarker density was 3.1 kb on average, with 78% coverage of CEU HapMap SNPs with MAF > 1%. P values for family-based association tests are plotted by position in Figure 2. The strongest association signal was for SNP rs7817434 at 26.818 Mb (P = 0.000301), located 39 kb on the 5′ side of ADRA1A, and 331 kb from the gene STMN4. Five SNPs in this region showed P values <0.01 (between 26.818 and 26.951 Mb). In addition, 6 SNPs had P values <0.01 in a 29 kb region spanning 25.938 to 25.967 Mb that overlaps the EBF2 gene. One other notable region contained several SNPs with P values <0.05 overlapping this SCARA3 gene. Both ADRA1A and EBF2 have been previously reported to be associated with schizophrenia [Clark et al., 2005]. Based on these results, we narrowed our region of interest to a 2.8 Mb region that contained all nominally significant associations (green boundaries, Fig. 2) for further follow-up in the independent set of Ashkenazi cases and controls.
AJ Case-Control Fine-mapping
Quality control by SNP
We manually reviewed intensity clusters for all 1,064 SNPs using the Illumina Bead Studio software (v3.2) to optimize genotype clustering, then set any genotype with an Illumina genotype quality score <0.6 to missing, assuming poor quality. Based on the distribution of missing rates by SNP showing a natural break between 0.4 and 0.8 proportion missing, we filtered all SNPs with >40% missing data (62 SNPs, 6%). We then removed remaining SNPs with MAF < 1% (27, 2.5%) and removed an additional 5 (0.5%) for extreme departures from Hardy–Weinberg proportions among controls (P < 10−4). The remaining 970 SNPs were used in all subsequent analyses. These SNPs represent 90% of all HapMap (phase II) CEU SNPs with MAF ≥ 1% at r2 ≥ 0.9. SNPs were recoded to match HapMap coding as described in the methods. SNP frequencies in our AJ samples were similar to those of the HapMap CEU samples (Supplementary Fig. 1).
Quality control by DNA
We began with 716 unrelated cases, 740 controls from AJCR, and 781 from the NYCP. Two AJCR controls and 74 NYCP controls failed DNA screening, resulting in 1,445 controls with genotype data. SNP genotype frequencies did not differ across control groups. Analysis of population substructure, seeded with CEU, YRI, and CHB samples from HapMap, showed no major ancestry differences between cases and controls (Supplementary Fig. 2). However, there were 33 individuals with <0.5 probability of membership in the Caucasian cluster (7 cases, 9 AJCR, 17 NYCP controls), and these individuals were removed from subsequent analyses. The sex and age information for the final analytic set of cases and controls is shown in Table 1.
P values for the association between each SNP and SZ are shown by physical location in Figure 3. The strongest signal was for rs12155555 at 26,576,477 bp (OR = 0.66; CI: 0.53, 0.81; P = 0.00006), located between DPYSL2 (4 kb), and ADRA1A (85 kb). A 458 kb region around this SNP contains 29 SNPs with P values <0.01, and a smaller region, from 26.509 to 26.576 Mb over the 5′ end of DPYSL2, contains 3 SNPs with P values <0.001 (Fig. 4). Only 1 SNP located in ADRA1A showed nominal association. Two other regions showed clusters of association. The first, in the EBF2 gene, contained 10 SNPs with marginal significance (P < 0.05)(Fig. 3). This gene also showed suggestive evidence of association in our family-based EUC analyses, and has been associated with SZ in other samples [Duan et al., 2007]. The second region, near PNOC, showed a cluster of 6 SNPs with marginal significance (P < 0.05).
Thus, our linkage evidence and two independent association analyses using both family-based and homogenous AJ case-control approaches all point to an interval of chromosome 8 that includes DPYSL2. Table 2 shows the frequencies and effect sizes in case-control analyses for SNPs in this region. Since sex may be an important source of heterogeneity in SZ presentation and etiology [Wolyniec et al., 1992; Leung and Chue, 2000; Aleman et al., 2003; McGrath et al., 2004], we further restricted our analyses in this case-control sample to only male cases and controls. The best signal in this region improved when restricting analyses to only men (Table 2). There were too few females to provide reliable female-only estimates.
|SNP name||Pos||MAF||Gene||Case-control analyses||Males only|
|OR||P value||N||OR||P value||N|
- Top of page
- MATERIALS AND METHODS
- Supporting Information
We have shown linkage and association evidence in European-Caucasian and the genetically more homogenous Ashkenazim for a SSL on chromosome 8p21. These results are consistent with many previous linkage and association findings for SZ [Tosato et al., 2005], and with a recent review of the importance of this region in psychiatric conditions generally [Tabares-Seisdedos and Rubenstein, 2009].
In our samples, SNP rs12155555, near DPYSL2, showed the strongest association to SZ (P = 0.000063). and was stronger for males (P = 0.00003). Associations with DPYSL2, specifically SNP rs17666, have been observed in other samples, including Japanese and separate samples of US Caucasians, although evidence in African Americans and Han Chinese has been weak [Nakata et al., 2003; Zhao et al., 2006; Allen et al., 2008; Tabares-Seisdedos and Rubenstein, 2009], and few other genetic association studies have been carried out specific to this gene. In additional to the previous genetic association findings with this gene, three proteomics studies in SZ have reported changes for DPYSL2 [Edgar et al., 2000; Johnston-Wilson et al., 2000; Beasley et al., 2006] while two studies have reported changes in rat brains in response to methamphetamine [Iwazaki et al., 2007] and to MK-801 [Paulson et al., 2004], a non-competitive NMDA antagonist that produces psychosis-like states. DPYSL2 or DRP-2, encodes dihydropyrimidinase-like 2, also called collapsin response mediator protein 2 (CRMP-2), the mammalian homolog of UNC-33, a C. elegans protein required for normal axon extension [Goshima et al., 1995]. In mammals, the DPYSL2 gene product regulates axonal differentiation by binding tubulin heterodimers and promoting microtubule assembly and is itself regulated by phosphorylation by GSK-3b [Fukata et al., 2002; Charrier et al., 2003; Arimura et al., 2004; Kawano et al., 2005; Yoshimura et al., 2005; Rogemond et al., 2008]. Taken together, our genetic data, those of others, and the relevant neurobiology provide independent and diverse evidence supporting a possible involvement of DPYSL2 in schizophrenia.
Genetic studies of schizophrenia have traditionally shown limited consistency due to variable phenotyping strategies, designs, populations, and gene coverage. We have addressed these limitations by using unified, rigorous phenotyping across a range of study designs and ancestries, and we have achieved deep SNP coverage of this region. Both family-based and case-control association approaches show association to this part of chromosome 8p, with the most compelling evidence around DPYSL2. The Ashkenazim provide an opportunity to see stronger association patterns due to reduction of allelic heterogeneity, yet we show that allele frequencies and LD patterns between our AJ samples and the HapMap CEU samples are comparable and that findings in our AJ samples are relevant to general Caucasian populations.
Other genes in this region with suggestive evidence of association include ADRA1A, EBF2, and PNOC1. ADRA1A (OMIM # 104221) codes for the A subtype of the Alpha 1 Adrenergic receptor, a G protein-coupled transmembrane receptor with affinity for epinephrine and norepinephrine. Although most directly involved in the autonomic/sympathetic nervous system, it is of interest that antipsychotic drugs have affinity for ADRA1A, a property suggested to underlie autonomic side effects [Nourian et al., 2008]. Genetic variation in ADRA1A has been previously associated to schizophrenia [Clark et al., 2005], ADHD [Lasky-Su et al., 2008; Elia et al., 2009] and heroin addiction. EBF2 (OMIM # 609934) is a helix-loop-helix transcription factor, a member of the Olf-1/EBF-like HLH transcription factor family. It is important for neuronal differentiation and migration [Garcia-Dominguez et al., 2003] and EBF2 knockout mice have a defect in the formation of the neuroendocrine axis and peripheral neuropathy. Finally, PNOC1 (OMIM # 601459) codes for prepronociceptin which is processed to nociceptin, a 17 aminoacid neuropeptide that is the natural agonist of opioid receptor like-1 and is thought to be important for multiple brain and other functions including pain, anxiety, stress-induced anorexia, drug dependence, depression, dementia, and Parkinsonism [Meis, 2003; Chiou et al., 2007].
It is worth noting that our region does not include NRG1 [Stefansson et al., 2002, 2003]. Our linkage and association work in this region provides stronger evidence for a schizophrenia susceptibility locus at least 3.1 Mb telomeric from NRG1. This result is consistent with those of several other groups who have identified linkage or association on chromosome 8 that is distinct from NRG1 [Tosato et al., 2005; Tabares-Seisdedos and Rubenstein, 2009]. Further, our previous candidate gene work in a subset of the samples in this current report did not show evidence for association to NRG1 [Fallin et al., 2005].
Recent genome-wide association studies (GWAS) in schizophrenia have not highlighted this region as one of the top hits [O'Donovan et al., 2008; Shifman et al., 2008; Kirov et al., 2009; Purcell et al., 2009; Shi et al., 2009; Stefansson et al., 2009], focusing rather on the MHC locus of chromosome 6, among others. However, this region was among the top associations in the GWAS study in MGS among African Americans (at rs4732838, 28.10 Mb, P = 1.68 × 10−5) [Shi et al., 2009], and in analyses combining their European and African ancestry samples (rs12547975, 25.69 Mb, P = 9.59 × 10−6, very near EBF2) [Shi et al.].
In summary, this region of chromosome 8 is a promising location for a schizophrenia locus based on numerous linkage studies and meta-analyses. The chromosome has also been implicated as important for psychiatric genetics more broadly [Tabares-Seisdedos and Rubenstein, 2009]. This was the strongest linkage signal in our recent high-density linkage consortium effort, with our JHU families showing the strongest evidence of linkage in this region [Holmans et al., 2009]. The follow-up linkage and association fine-mapping presented here implicate DPYSL2 as a susceptibility locus for SZ, consistent with other reports, although other genes in this region, such as EBF2 and ADRDA1 are also worth additional attention.
- Top of page
- MATERIALS AND METHODS
- Supporting Information
We would like to thank the participants in all studies included in these analyses. The NYCP is funded by the AMDec Foundation, Inc. This work was also supported by NIMH (R01MH068406; Pulver, PI).
- Top of page
- MATERIALS AND METHODS
- Supporting Information
- 2002. Merlin-rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30(1): 97–101. , , , .
- 2003. Sex differences in the risk of schizophrenia: Evidence from meta-analysis. Arch Gen Psychiatry 60(6): 565–571. , , .
- 2008. Systematic meta-analyses and field synopsis of genetic association studies in schizophrenia: The SzGene database. Nat Genet 40(7): 827–834. , , , , , , , .
- 2004. Role of CRMP-2 in neuronal polarity. J Neurobiol 58(1): 34–47. , , , .
- 2002. Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia. Mol Psychiatry 7(4): 405–411. , .
- 2005. Haploview: Analysis and visualization of LD and haplotype maps. Bioinformatics 21(2): 263–265. , , , .
- 2006. Proteomic analysis of the anterior cingulate cortex in the major psychiatric disorders: Evidence for disease-associated changes. Proteomics 6(11): 3414–3425. , , , , , .
- 1998. Schizophrenia susceptibility loci on chromosomes 13q32 and 8p21. Nat Genet 20(1): 70–73. , , , , , , , , , , et al.
- 2005. Linkage disequilibrium inflates type I error rates in multipoint linkage analysis when parental genotypes are missing. Hum Hered 59(4): 220–227. , , , , , , , , , , et al.
- 1999. Linkage of familial schizophrenia to chromosome 13q32. Am J Hum Genet 65(4): 1096–1103. , , , , , , .
- 1999. Heritability estimates for psychotic disorders: The Maudsley twin psychosis series. Arch Gen Psychiatry 56(2): 162–168. , , , , , , , , , , et al.
- 2003. Collapsin response mediator proteins (CRMPs): Involvement in nervous system development and adult neurodegenerative disorders. Mol Neurobiol 28(1): 51–64. , , , , , .
- 2009. Fine mapping on chromosome 10q22-q23 implicates neuregulin 3 (NRG3) in schizophrenia. Am J Hum Genet 84(1): 21–34. , , , , , , , , , , et al.
- 2007. Nociceptin/orphanin FQ peptide receptors: Pharmacology and clinical implications. Curr Drug Targets 8(1): 117–135. , , , , , , .
- 2005. Polymorphisms in the promoter region of the alpha1A-adrenoceptor gene are associated with schizophrenia/schizoaffective disorder in a Spanish isolate population. Biol Psychiatry 58(6): 435–439. , , , , , .
- 2002. A genome-wide scan for linkage to chromosomal regions in 382 sibling pairs with schizophrenia or schizoaffective disorder. Am J Psychiatry 159(5): 803–812. , , , , , , , , , , et al.
- 2007. Linkage Disequilibrium Mapping for Schizophrenia Susceptibility Genes on 8p23.3-p12 in a large European Ancestry Sample. San Diego, CA: American Society of Human genetics meeting. , , , , , , , , , , et al.
- 2000. Comparative proteome analysis of the hippocampus implicates chromosome 6q in schizophrenia. Mol Psychiatry 5(1): 85–90. , , , , , .
- 2009. Candidate gene analysis in an on-going genome-wide association study of attention-deficit hyperactivity disorder: Suggestive association signals in ADRA1A. Psychiatr Genet 19(3): 134–141. , , , , , , , .
- 2005. Bipolar I disorder and schizophrenia: A 440-single-nucleotide polymorphism screen of 64 candidate genes among Ashkenazi Jewish case-parent trios. Am J Hum Genet 77(6): 918–936. , , , , , , , , , , et al.
- 2003. Inference of population structure using multilocus genotype data: Linked loci and correlated allele frequencies. Genetics 164(4): 1567–1587. , , .
- 2002. The founder mutation MSH2*1906G–>C is an important cause of hereditary nonpolyposis colorectal cancer in the Ashkenazi Jewish population. Am J Hum Genet 71(6): 1395–1412. , , , , , , , , , , et al.
- 2002. CRMP-2 binds to tubulin heterodimers to promote microtubule assembly. Nat Cell Biol 4(8): 583–591. , , , , , , , , , , et al.
- 2003. Ebf gene function is required for coupling neuronal differentiation and cell cycle exit. Development 130(24): 6013–6025. , , , .
- 2006. Schizophrenia susceptibility genes: Emergence of positional candidates and future directions. Trends Pharmacol Sci 27(4): 226–233. , .
- 1995. Collapsin-induced growth cone collapse mediated by an intracellular protein related to UNC-33. Nature 376(6540): 509–514. , , , .
- 2006. Genetic association and brain morphology studies and the chromosome 8p22 pericentriolar material 1 (PCM1) gene in susceptibility to schizophrenia. Arch Gen Psychiatry 63(8): 844–854. , , , , , , , , , , et al.
- 2001. Genomewide genetic linkage analysis confirms the presence of susceptibility loci for schizophrenia, on chromosomes 1q32.2, 5q33.2, and 8p21-22 and provides support for linkage to schizophrenia, on chromosomes 11q23. 3-24 and 20q12.1-11.23. Am J Hum Genet 68(3): 661–673. , , , , , , , , , , et al.
- 2007. Increasing global burden of cardiovascular disease in general populations and patients with schizophrenia. J Clin Psychiatry 68 (Suppl 4): 4–7. .
- 2009. Genomewide linkage scan of schizophrenia in a large multicenter pedigree sample using single nucleotide polymorphisms. Mol Psychiatry 14(8): 786–795. , , , , , , , , , , et al.
- 2005. Dihydropyrimidinase-related protein 2 (DRP-2) gene and association to deficit and nondeficit schizophrenia. Am J Med Genet Part B Neuropsychiatr Genet 136B(1): 8–11. , , , , , , , , .
- 2001. The family based association test method: Strategies for studying general genotype–phenotype associations. Eur J Hum Genet 9(4): 301–306. , , .
- 2004. Family-based tests for associating haplotypes with general phenotype data: Application to asthma genetics. Genet Epidemiol 26(1): 61–69. , , , , , .
- 2007. Protein expression profile in the striatum of rats with methamphetamine-induced behavioral sensitization. Proteomics 7(7): 1131–1139. , , .
- 2000. Disease-specific alterations in frontal cortex brain proteins in schizophrenia, bipolar disorder, and major depressive disorder. The Stanley Neuropathology Consortium. Mol Psychiatry 5(2): 142–149. , , , , , , .
- 2008. Recruitment of PCM1 to the centrosome by the cooperative action of DISC1 and BB S4: A candidate for psychiatric illnesses. Arch Gen Psychiatry 65(9): 996–1006. , , , , , , , , , , et al.
- 2006. Schizophrenia genetics: Uncovering positional candidate genes. Eur J Hum Genet 14(5): 512–519. , .
- 1998. NIMH genetics initiative millenium schizophrenia consortium: Linkage analysis of African-American pedigrees. Am J Med Genet 81(4): 282–289. , , , , , , , , , , et al.
- 2005. CRMP-2 is involved in kinesin-1-dependent transport of the Sra-1/WAVE1 complex and axon formation. Mol Cell Biol 25(22): 9920–9935. , , , , , , , .
- 1996. Evidence for a schizophrenia vulnerability locus on chromosome 8p in the Irish Study of High-Density Schizophrenia Families. Am J Psychiatry 153(12): 1534–1540. , , , , , , , , , , et al.
- 2009. A genome-wide association study in 574 schizophrenia trios using DNA pooling. Mol Psychiatry 14(8): 796–803. , , , , , , , , , .
- 1996. Parametric and nonparametric linkage analysis: A unified multipoint approach. Am J Hum Genet 58(6): 1347–1363. , , , .
- 2008. Genome-wide association scan of quantitative traits for attention deficit hyperactivity disorder identifies novel associations and confirms candidate gene associations. Am J Med Genet Part B Neuropsychiatr Genet 147B(8): 1345–1354. , , , , , , , , , , et al.
- 2006. Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5' SNPs associated with the disease. Proc Natl Acad Sci USA 103(17): 6747–6752. , , , , , , , , .
- 2000. Sex differences in schizophrenia, a review of the literature. Acta Psychiatr Scand Suppl 401:3–38. , .
- 2003. Genome scan meta-analysis of schizophrenia and bipolar disorder, part II: Schizophrenia. Am J Hum Genet 73(1): 34–48. , , , , , , , , , , et al.
- 2005. Linkage evidence of schizophrenia to loci near neuregulin 1 gene on chromosome 8p21 in Taiwanese families. Am J Med Genet Part B Neuropsychiatr Genet 134B(1): 79–83. , , , , , , .
- 2008. Assessing the Biological Relevance of Sequence Variants in and Around DPYSL2, a Candidate Schizophrenia (SZ) Susceptibility Gene on 8p21. Philadelphia, PA: American Society of Human Genetics. , , , , , , , , , , et al.
- 2007. Schizophrenia: A common disease caused by multiple rare alleles. Br J Psychiatry 190:194–199. , , .
- 2004. A systematic review of the incidence of schizophrenia: The distribution of rates and the influence of sex, urbanicity, migrant status and methodology. BMC Med 2:13. , , , , , .
- 2003. Nociceptin/orphanin FQ: Actions within the brain. Neuroscientist 9(2): 158–168. .
- 2003. The human dihydropyrimidinase-related protein 2 gene on chromosome 8p21 is associated with paranoid-type schizophrenia. Biol Psychiatry 53(7): 571–576. , , , , , , .
- 2009. Meta-analysis of 32 genomewide linkage studies of schizophrenia. Mol Psychiatry 14(8): 774–785. , , , , , , , , , et al.
- 2008. The antagonistic effect of antipsychotic drugs on a HEK293 cell line stably expressing human alpha1A1-adrenoceptors. Eur J Pharmacol 596(1–3): 32–40. , , , , .
- 2008. Identification of loci associated with schizophrenia by genome-wide association and follow-up. Nat Genet 40(9): 1053–1055. , , , , , , , , , , et al.
- Suggestion of roles for both common and rare risk variants in genome-wide studies of schizophrenia. Arch Gen Psychiatry 67(7): 667–673. , , . 2010.
- 2004. The molecular genetics of schizophrenia: New findings promise new insights. Mol Psychiatry 9(1): 14–27. , , .
- 2004. Comparative proteome analysis of thalamus in MK-801-treated rats. Proteomics 4(3): 819–825. , , , , , , .
- 2000. Inference of population structure using multilocus genotype data. Genetics 155(2): 945–959. , , .
- 1989. Availability of schizophrenic patients and their families for genetic linkage studies: Findings from the Maryland epidemiology sample. Genet Epidemiol 6(6): 671–680. , .
- 1995. Schizophrenia: A genome scan targets chromosomes 3p and 8p as potential sites of susceptibility genes. Am J Med Genet 60(3): 252–260. , , , , , , , , , , et al.
- 2007. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81(3): 559–575. , , , , , , , , , , et al.
- 2009. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature 460(7256): 748–752. , , , , , , .
- 1999. The economic impact of schizophrenia. J Clin Psychiatry 60 (Suppl 1): 4–6 (discussion 28–30). .
- 2006. Molecular genetic studies of schizophrenia. Eur J Hum Genet 14(6): 669–680. , .
- 2008. Processing and nuclear localization of CRMP2 during brain development induce neurite outgrowth inhibition. J Biol Chem 283(21): 14751–14761. , , , , , , , .
- 2007. Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels. Science 316(5829): 1331–1336. , , , , , , , , , , et al.
- 2009. Common variants on chromosome 6p22.1 are associated with schizophrenia. Nature 460(7256): 753–757. , , , , , , , , , , et al.
- 2008. Genome-wide association identifies a common variant in the reelin gene that increases the risk of schizophrenia only in women. PLoS Genet 4(2): e28. , , , , , , , , , , et al.
- 2009. Common variants conferring risk of schizophrenia. Nature 460(7256): 744–747. , , , , , , , , , , et al.
- 2003. Association of neuregulin 1 with schizophrenia confirmed in a Scottish population. Am J Hum Genet 72(1): 83–87. , , , , , , , , , , et al.
- 2002. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet 71(4): 4. , , , , , , , , , , et al.
- 2002. Genome-wide scans of three independent sets of 90 Irish multiplex schizophrenia families and follow-up of selected regions in all families provides evidence for multiple susceptibility genes. Mol Psychiatry 7(6): 542–559. , , , , , , , , , , et al.
- 2006. Genomewide linkage scan of 409 European-ancestry and African American families with schizophrenia: Suggestive evidence of linkage at 8p23.3-p21.2 and 11p13.1-q14.1 in the combined sample. Am J Hum Genet 78(2): 315–333. , , , , , , , , , , et al.
- 2009. Chromosome 8p as a potential hub for developmental neuropsychiatric disorders: Implications for schizophrenia, autism and cancer. Mol Psychiatry 14(6): 563–589. , .
- 2007. Molecular cloning of a brain-specific, developmentally regulated neuregulin 1 (NRG1) isoform and identification of a functional promoter variant associated with schizophrenia. J Biol Chem 282(33): 24343–24351. , , , , , , , .
- 2005. Association between the neuregulin 1 gene and schizophrenia: A systematic review. Schizophr Bull 31(3): 613–617. , , .
- 2006. Association study of the dihydropyrimidinase-related protein 2 gene and methamphetamine psychosis. Ann N Y Acad Sci 1074:90–96. , , , , , , , , , , et al.
- 1992. Schizophrenia: Gender and familial risk. J Psychiatr Res 26(1): 17–27. , , , .
- 2005. GSK-3beta regulates phosphorylation of CRMP-2 and neuronal polarity. Cell 120(1): 137–149. , , , , , .
- 2006. An investigation of the dihydropyrimidinase-like 2 (DPYSL2) gene in schizophrenia: Genetic association study and expression analysis. Int J Neuropsychopharmacol 9(6): 705–712. , , , , , , , , , , et al.
- Top of page
- MATERIALS AND METHODS
- Supporting Information
Supporting information may be found in the online version of this article.
|ajmb_31147_sm_suppFig1.tif||124K||Supplementary Figure 1|
|ajmb_31147_sm_suppFig2.tif||93K||Supplementary Figure 2|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.