White matter integrity as an intermed

White matter integrity as an intermediate phenotype:Exploratory

genome-wide association analysis in individuals at high risk of

bipolar disorder

Emma Sprooten a,n,Kathryn M.Fleming a,Pippa A.Thomson b,Mark E.Bastin c,Heather C.Whalley a, Jeremy Hall a,Jess E.Sussmann a,James McKirdy a,Douglas Blackwood a,b,Stephen M.Lawrie a, Andrew M.McIntosh a

a Division of Psychiatry,University of Edinburgh,Kennedy Tower,Royal Edinburgh Hospital,Morningside Park,Edinburgh EH105HF,UK

b Molecular Medicine Centre,University of Edinburgh,Edinburgh,UK

c Brain Research Imaging Centre,University of Edinburgh,Edinburgh,UK

a r t i c l e i n f o

Article history:

Received6April2012

Received in revised form

14August2012

Accepted1November2012

Keywords:

Endophenotype

Quantitative phenotype

Psychiatric genetics

Diffusion tensor imaging

Fractional anisotropy

Pathway analysis

a b s t r a c t

White matter integrity,as measured using diffusion tensor imaging(DTI),is reduced in individuals with

bipolar disorder(BD),their unaffected relatives and carriers of specific risk-alleles.Fractional

anisotropy(FA),an index of white matter integrity,is highly heritable but the genetic architecture of

this trait has received little investigation.In this study we performed a genome-wide association study

with FA as quantitative phenotype,in unaffected relatives of patients with BD(N¼70)and a matched

control group(N¼80).Amongst our top results were SNPs located in genes involved in cell adhesion,

white matter development and neuronal plasticity.Pathway analysis of the top associated polymorph-

isms and genes confirmed the enrichment of processes relevant to BD and white matter development,

including axon guidance,ErbB-signalling neurotrophin signalling,phosphatidylinositol signalling,and

cell adhesion.The majority of genes implicated in these pathways were differentially associated with

FA in individuals at high familial risk,suggesting interactions with genetic background or environ-

mental factors secondary to familial risk for BD.Although the presentfindings require independent

replication,the results encourage the use of global FA as a quantitative phenotype in future large-scale

studies which may help to identify the biological processes underlying reduced FA in BD and other

psychiatric disorders.

&2012Elsevier Ireland Ltd.All rights reserved.

1.Introduction

The heritability of bipolar disorder(BD)has been estimated at

approximately80%(Gershon et al.,1982;McGuffin et al.,2003),

and genetic linkage and association studies,including genome-

wide association studies(GWAS),have now identified a few well-

replicated risk-variants(Craddock and Sklar,2009;O’Donovan

et al.,2009;Sklar et al.,2011).However,the substantial clinical

and genetic heterogeneity and lack of biological markers of BD

have led to slower progress than in other complex disorders in

which the underlying disease biology is better understood

(Manolio et al.,2008;Cichon et al.,2009).Endophenotypes are

markers of genetic vulnerability to a disorder and have a genetic

architecture that is in theory less complex than the clinical

syndrome.As constructs that lie biologically intermediate

between gene function and the clinical phenotype,endopheno-

types can provide clues to disease pathology,and may aid in the

discovery of new genetic risk variants(Gottesman and Gould,

2003;Glahn et al.,2007).

The integrity of the brain’s white matter can be measured

in vivo using diffusion tensor imaging(DTI),an MRI technique

sensitive to the magnitude and directionality of water molecule

displacement in human tissue.Fractional anisotropy(FA)is a

scalar metric derived from the diffusion tensor with values

between0and1that quantifies the degree to which water

diffusion is anisotropic.In healthy white matter,water molecules

diffuse preferentially along rather than across the mainfibre

direction,resulting in anisotropic diffusion and evaluated values

of FA.Conversely,in less healthy white matter with fewer intact

cellular boundaries,water molecules diffuse more equally in all

directions,and FA falls in value.As such,FA reflects a combination

of axonal density,fibre packing and organisation,and myelin

thickness,and is therefore assumed to be an approximation of

white matter integrity in general(Budde et al.,2007;Cercignani

Contents lists available at SciVerse ScienceDirect

journal homepage:www.elsevier.com/locate/psychres

Psychiatry Research

0165-1781/$-see front matter&2012Elsevier Ireland Ltd.All rights reserved.

http://dx.doi.org/10.1016/j.psychres.2012.11.002

n Corresponding author.Tel.:þ441315376502;fax:þ441315376531.

E-mail address:e.sprooten@ed.ac.uk(E.Sprooten).

Psychiatry Research](]]]])]]]–]]]et al.,2001;Gouw et al.,2008;MacDonald et al.,2007;Schmierer et al.,2007;Underwood et al.,2011).

FA is reduced in patients with BD(Heng et al.,2010;Mahon et al.,2010;McIntosh et al.,2008b;Sexton et al.,2009;Sussmann et al.,2009)and these reductions are found independently of present mood state in euthymic,manic and depressed patients (Adler et al.,2006;Benedetti et al.,2011;Wang et al.,2008a, 2008b).Importantly,the heritability of FA is estimated between 50%and85%,depending on the brain region,with higher estimates for young people(Chiang et al.,2009,2011;Kochunov et al.,2010).Recently,we have also found subtle,but widespread, FA reductions in unaffected relatives of patients with BD.These reductions were inversely correlated with temperamental mood fluctuations previously associated with disease liability(Sprooten et al.,2011a).The abovefindings support white matter integrity, as measured by FA,as an endophenotype for BD according to the criteria of Gottesman and Gould(2003).

In line with this endophenotypic quality,several risk variants in BD candidate genes have been associated with white matter integrity in healthy individuals(Konrad et al.,2009;McIntosh et al.2008a;Sprooten et al.,2009,2011b;Winterer et al.,2008; Zuliani et al.,2011).However,such a candidate gene approach is inherently biased because gene selection is generally inferred from prior biological knowledge or linkage studies.In contrast, GWAS allow for the unbiased investigation of the whole genome in association with a clinical phenotype or a quantitative heritable trait.Previous studies have successfully used hippocampal volume(Stein et al.,2012)or dorsolateral prefrontal brain activation(Potkin et al.,2009a,2009b)as quantitative traits, thereby discovering novel candidate risk-genes for psychiatric disorders.Recently,a genome-wide analysis of a global measure of FA was performed in the Lothian Birth Cohort1936,a large sample of ageing people(mean age¼73years),and two sugges-tive single-nucleotide polymorphisms(SNPs)including one in an ADAM-family gene,were identified(Lopez et al.,2012).

In most GWAS,conservative genome-wide significance thresh-olds are applied in order to compensate for type-I error inflation resulting from simultaneously testing of hundreds of thousands of associations.However,significance levels of around10À7inevi-tably give rise to a large number of false negatives,especially if the phenotype has a polygenic origin.In BD,two recent studies have found that a multitude of risk-alleles with small effect size, at extremely liberal p-values(p o0.5,contribute substantially to the accuracy at which genetic variation can predict affection status of BD,whilst a model including only alleles of reasonably high significance(p o10À4)performed only at chance level (Evans et al.,2009;Purcell et al.,2009).Complimentary to considering markers at genome-wide significance levels,in path-way analysis less significant genetic variation is retained,and a large set of top SNPs are clustered according to their biological and molecular functions.This then allows the statistical testing of whether any relevant pathways are significantly over-represented (‘‘enriched’’)in the top-associated SNPs.This approach appears to more accurately capture the polygenic architecture of BD,while at the same time providing important information on the biological functions is likely affected by the associated genes.Furthermore, individuals with different clinical presentations can carry differ-ent risk-alleles to a disease,yet their downstream effects can converge onto the same molecular pathway and/or be measurable in the same endophenotype.Thus,the use of both pathways and endophenotypes in GWAS are likely to capture some of the genetic and clinical heterogeneity of complex disorders that can be missed when looking at individual alleles and behavioural phenotypes alone.Pathway analysis has previously been applied to case-control GWAS(Askland et al.,2009;O’Dushlaine et al., 2011;Torkamani et al.,2008),and more recently to cognitive (Luciano et al.,2011)and brain imaging phenotypes(Lopez et al., 2012;Meda et al.,2012).

Here we performed a genome-wide association study using a global measure of white matter integrity as quantitative pheno-type,followed by pathway analyses of the top SNPs and genes. The phenotype,average FA across the centres of the brain’s white matter,was one we previously found to be significantly reduced in unaffected relatives of patients with BD(Sprooten et al., 2011a),while Kochunov et al.(2010)showed it is also highly heritable.We performed this analysis in a sample of healthy individuals,including participants at high genetic risk of BD.This design allowed us to test for main(i.e.additive)effects of SNPs on FA,as well as for SNP-by-risk interaction effects on FA.Main effects would be expected to identify genetic variants and molecular pathways involved in white matter development, plasticity and maintenance,while interaction effects,akin to a gene-by-diagnosis interaction as previously modelled in Potkin et al.(2009a),can identify SNPs whose associations to FA are modulated by familial risk and/or vulnerable genetic background.

2.Methods

2.1.Sample

Participants were recruited as part of the Bipolar Family Study, a sample of young individuals at high genetic risk(HR)for BD and demographically matched healthy controls(HC).Individuals were considered at HR if they had at least onefirst-degree,or two second degree,family members with bipolar I disorder.Partici-pants were excluded if they fulfilled SCID criteria for an axis-I mood or psychotic disorder,had a major neurological disorder, history of head injury,history of learning disability or metallic implants or other contraindications to MRI examination.For more details of participant recruitment and screening and demo-graphics,see Sprooten et al.(2011a)and Whalley et al.(2011). In this sample,a subset of150unrelated participants provided high quality DTI data and DNA for the generation of whole-genome data,including70HR(34male;mean age¼21.6years) and80HC(37male;mean age¼21.3years)subjects,all between the ages of16and26years at the time of recruitment.

2.2.Scan acquisition and DTI processing

DTI data were acquired and processed as previously described in Sprooten et al.(2011a).MRI data were collected on a GE Signa Horizon HDX1.5T clinical scanner equipped with a self-shielding gradient set(22mT/m maximum gradient strength)and manufacturer-supplied‘birdcage’quadrature head coil.Whole brain DTI data were acquired for each subject using a single-shot pulsed gradient spin-echo echo-planar imaging(EPI)sequence with diffusion gradients(b¼1000s/mm2)applied innon-collinear directions,and seven T2-weighted EPI baseline(b¼0s/mm2)scans. Fifty-three2.5mm contiguous axial slices were acquired with a field-of-view of240Â240mm2,and an acquisition matrix of 96Â96(zero-filled to128Â128),giving an acquisition isotropic voxel size of 2.5mm.In addition,a T1-weighted volume was acquired with afield-of-view of240Â240mm2,an acquisition matrix of192Â192,and1801.2mm thick slices,giving a voxel dimension of1.25Â1.25Â1.20mm3.The sequence had an inver-sion time of500ms,an echo time of4ms,and aflip angle of81.

The DTI data were converted to4D NIfTI volumes and pre-processed using standard tools available from FSL(http://www. fmrib.ox.ac.uk/fsl).This included the following processes:correc-tion for eddy current induced distortions and bulk subject motion in the scanner by registering the diffusion weighed EPI volumes to

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]] 2thefirst T2-weighted EPI volume within each subject;brain extraction;and calculation of diffusion tensor characteristics including principal eigenvectors and FA values using DTIFIT.

Tract-based spatial statistics(TBSS)was carried out according to standard FSL procedures(Smith et al.,2006;http://www.fmrib. ox.ac.uk/fsl).First,all subject’s FA volumes were linearly and non-linearly registered to a standard FA template available in the FSL package(‘‘FMRIB58_FA’’).Second,a mean of all registered FA volumes was calculated and a white matter‘‘skeleton’’was created.This was achieved by searching for the maximum FA values in directions perpendicular to the local tract direction in the mean FA map.A threshold of FA40.25was applied to the FA skeleton to exclude predominantly non-white matter voxels. Third,for each subject’s FA volume,at each point on the skeleton the maximum voxel perpendicular to the local skeleton direction was projected onto the skeleton.This results in one FA skeleton map per subject,assumed to contain anatomically corresponding centres of white matter structure.For each individual,the mean FA was extracted from the voxels within the skeleton.All the following analyses were performed using this skeleton-wide FA average as a quantitative phenotype for genome-wide regressions.

2.3.Genotyping and genome-wide association of SNPs

DNA was extracted from whole-blood samples and genotyped using the Human OmniExpress Illumina Array(http://www.illu mina.com/applications/gwas.ilmn),containing$730K tagging SNPs.Afterfiltering for call rate(495%),minor allele frequency (40.05)and deviation from Hardy–Weinberg equilibrium (40.001),565404SNPs remained to be included in the associa-tion analysis.There were no subjects who had more than5% missing genotypes.Genome-wide association analysis was per-formed using two quantitative linear models in PLINK(Purcell et al.,2007;http://pngu.mgh.harvard.edu/purcell/plink):(1)a ‘‘main model’’to test for main effect of SNPs on FA,but covarying for group effects;and(2)an‘‘interaction model’’,to test for group-by-SNP interaction effects on FA.Sex,age and thefirst four components resulting from multidimensional scaling(MDS)ana-lysis in PLINK,were also modelled in both regression analyses. Considering group-status and SNPs may be partly explaining overlapping variance in FA,a third exploratory regression(‘‘sim-ple model’’)was performed without group as a co-variate.

2.4.Genome-wide association of genes(VEGAS)

By chance alone,larger genes are more likely to contain SNPs with lower P-values.For this reason,gene-based p-values were calculated using Versatile Gene-based Association Study(VEGAS; http://gump.qimr.edu.au/VEGAS;Liu et al.,2010).First,SNPs are allocated to genes according to the UCSC hg18database,750kb (http://genome.ucsc.edu).Second,gene-wide statistics were cal-culated by summing up all w2values within a gene(w2values are derived from the raw p-values).These observed gene-wide statistics were then compared to an empirical null-distribution obtained using simulations from the multivariate distribution. Importantly,the simulated null-distribution calculated by VEGAS not only accounts for gene size but also for the underlying linkage disequilibrium(LD)structure.Of note,the gene-corrected p-values were not genome-wide corrected,and were mainly calculated in order to correct for gene-size,which can otherwise significantly bias pathway analysis towards brain-related pro-cesses(Raychaudhuri et al.,2010).

2.5.KEGG pathway analysis

The top1%SNPs(equivalent to5400SNPs,all p o0.01for both models)were tested for enrichment of the molecular pathways of the Kyoto Encyclopaedia of Genes and Genomes(KEGG)using WebGestalt(http://bioinfo.vanderbilt.edu/webgestalt).This freely available programme compares the selected SNPs against the whole genome with a Hypergeometric test to assess the like-lihood that the observed number of SNPs would be involved in a KEGG pathway by mere chance.One statistic is calculated for each pathway,and all reported p-values are adjusted for the number of pathways tested using false discovery rate correction (Benjamini and Hochberg,1995).The minimum number of genes per process was set at2.KEGG pathway enrichment analyses were performed for each top SNP set resulting from the two PLINK association analyses.

The top1%SNPs for each model mapped to approximately 1000genes.The SNP-enrichment results were thus cross-validated with gene-enrichment results,entering the top1000 genes resulting from VEGAS(all p o0.057)into WebGestalt,using the same parameters as in the SNP-based pathway analysis.

3.Results

3.1.Top SNP results

Q–Q plots of the SNP p-values for each model showed that the results followed a normal distribution,with a tendency for smaller than expected p-values in the upper tail(Fig.1).Genomic inflation factors were equal to or close to1(l main¼1.0; l interation¼1.02;prior to MDS-component correction),

indicating Fig.1.QQ-plots of results for main(SNP)and interaction(SNPÂGroup)models.

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]]3

the absence of population stratification and other systematic biases to significance inflation.No markers reached the current genome-wide significance threshold at p o 5Â10À8.The top 10SNP results for each model are listed in Tables 1and 2.

With respect to the main model,several top markers in association with FA were located in introns in ADAM 7,EPS 15L 1,AK 127888,and CRBN .Two SNPs in the main model’s top 10were intergenic but within 10kb of genes SAP 30L ,HAND 1,TMEM 239,and C 20orf 141.As shown in Table 1,most top markers in the main model were associated with FA in both HR and HC subjects separately,except for the SNPs in AK 127888(all in almost perfect LD).Indeed,these SNPs in AK 127888were also marginally sig-nificant for the group-by-SNP interaction (p o 0.02),and their effect in the main model was clearly driven by the HR participants.Results for the ‘‘simple model’’without group as a covariate did not increase statistical power,with very similar p -values (Supplemen-tary Fig.1)and eight out of the 10top SNPs being identical between the two models.

In the interaction model,all top SNPs were significantly asso-ciated with FA within each group,but in opposite directions of effect.Several of the top SNPs were located in or near genes with functions related to cell-adhesion,namely LPP,HEPACAM,HEPN 1,and ROBO 4.In addition,two SNPs mapped to an intron in a gene encoding a calcium/calmodulin protein kinase (CAMK 2D),and another was located near the highly brain-expressed transcription regulator,WDR 5.

3.2.Pathway analysis

Table 3shows a selection of relevant significant KEGG categories resulting from the top 1%SNP and the top 1000genes analyses in WebGestalt.The top enriched pathway for the main model was arrhythmogenic right ventricular cardiomyopathy (p ¼8.55Â10À9),followed by axon guidance (p ¼4.73Â10À8).For the interaction model,the MAPK signalling pathway was most significantly enriched (p ¼3.05Â10À10).Almost all relevant KEGG categories significantly enriched in the main model were more significant in the interaction model,including axon guidance,cell adhesion molecules (CAMs),focal adhesion,calcium signalling,tight junction,long-term potentiation,the phosphatidylinositol (PI)system,and Wnt signalling.ErbB signalling was significantly enriched for the interaction model only.All of these pathways,apart from calcium signalling,replicated in the gene-based analysis,indicating that their enrichment was not merely driven by gene-size or LD structure.Neurotrophin signalling was the only pathway that seemed more powerful in the gene-based than in the SNP-based analysis,indicat-ing that different independent SNPs within genes were responsible for this signal.For a complete list of significant pathways (p o 0.05)along with the genes contributing to their enrichment,see Supple-mentary Tables 2and 3.The results of pathway analysis for top SNPs in the ‘‘simple model’’were very similar to the results of the main model,and all pathways in Table 3remained similarly significant when not co-varying for group.

Table 1

Top markers resulting from the main model.Location SNP Gene a

SNP function MAF P ALL

P HC

P HR

8p21rs11781622ADAM7Intron 0.14 3.48E-0060.00480.003119p13rs875622EPS15L1

Intron 0.19 4.06E-006 6.60E-0040.00997p22rs6977940CARD11,GNA12Intergenic 0.11 4.54E-0060.0277 2.88E-00418q12b rs4800279AK127888Intron 0.06 5.42E-0060.3154 6.00E-0063p26rs1669338CRBN Intron 0.17 5.98E-006 1.43E-0050.00382p25rs2598592450kb

Intergenic 0.39 6.85E-0069.18E-0040.0015

20p13rs7360412TMEM239,C20orf141Intergenic 0.067.49E-0060.0404c

5q33rs17116175SAP30L,HAND1Intergenic 0.10 1.08E-0050.0286 4.21E-0045p15rs11133604Fam173B Intron 0.49 1.09E-0050.00230.003716q23

rs117294

VAT1L

Intron

0.33

1.10E-005

0.0095

0.0012

Top markers resulting from the main model,and annotated genes (within 10kb).MAF =minor allele frequency;P HC and P HR :P -values in controls and high-risk samples separately.

a Genes are annotated if SNP lies within 50k

b of the gene.

b Multiple loci in perfect LD:includes rs19409and rs8085344at same p -values.c

SNP failed frequency criterion in this sample.

Table 2

Top markers resulting from the interaction model.Location SNP Gene a

SNP function MAF P ALL

P HC

P HR

b HC

3q13rs488628450kb

Intergenic 0.21 6.33E-0070.0249 1.28E-005þ3p14rs1036797MIR548A2,ADAMTS9-AS2Intron 0.26 2.09E-0060.0091 4.32E-004þ4q26rs1880529CAMK2D Intron 0.43 3.86E-0060.00120.0020À3q28rs1975991LPP

Intron 0.32 6.40E-0060.0127 2.45E-004À2p25rs17491951AK095310Intron 0.38 6.78E-006 3.40E-0040.0022þ1p21rs66305AGL,FRRS1Intergenic 0.198.56E-0060.00160.0011

þ9q34rs4292757WDR5

Intergenic 0.06 1.18E-0050.0133b

þ11q24rs12222810CCDC15,HEPACAM,HEPN1,ROBO4Intergenic 0.12 1.55E-005 2.38E-0040.0327

À4q21rs76591LOC339966Intergenic 0.08 1.67E-0050.0887b

þ4p15

rs12503381

450kb

Intergenic

0.07

1.70E-005

0.0049

0.0066

þ

Top markers resulting from the main model,and annotated genes (within 10kb).MAF =minor allele frequency;P HC and P HR :P -values in controls and high-risk samples separately.The sign of the standardised b in HC is listed in the last column,which was invariably opposite to the b in HR.

a Genes are annotated if SNP lies within 50k

b of the gene.b

SNP failed frequency criterion in this sample.

E.Sprooten et al./Psychiatry Research ](]]]])]]]–]]]

4

4.Discussion

White matter integrity is highly heritable(Chiang et al.,2009, 2011;Kochunov et al.,2010),reduced in BD patients and their unaffected relatives(Chaddock et al.,2009;McIntosh et al.,2008b; Mun˜oz Maniega et al.,2008;Sussmann et al.,2009;Sprooten et al., 2011a),and associated with several risk-genes for BD(Konrad et al., 2009;McIntosh et al.,2008a;Sprooten et al.,2009,2011b;Winterer et al.,2008;Zuliani et al.,2011).Together,thesefindings indicate that white matter integrity,as measured by DTI,could be a potent endophenotype of BD.Here,we implemented this knowledge, by using average FA across the TBSS skeleton as a quantitative phenotype in a genome-wide association analysis in unaffected relatives of BD and a matched control sample.

4.1.Single SNPs

Cell adhesion,cell growth and calcium signalling were com-mon functions of genes containing the top SNPs from both main and interaction models.One top SNP in the main model was located in EPS15L1,an EPS15homologue which is involved in epidermal growth factor receptor signalling.Although EPS15L1 function in the brain specifically is poorly characterised,it is highly brain expressed,and its calcium-binding property suggests it contributes to activity dependent neuronal growth and survival. ADAM7,also supported by the main model,encodes an ADAM-family member,a group of proteins that interact with epidermal growth factors(Seals and Courtneidge,2003),although they are better known as cell adhesion proteins(Reiss and Saftig,2009). Of note,the one other GWAS conducted on a global measure of FA in the elderly also reported a suggestive association with a SNP in an ADAM-family gene,namely ADAMTS18(Lopez et al.,2012). Thefinding that ADAM-family genes appeared in the top‘‘hits’’in association with FA in two demographically different(but Scottish) samples is quite remarkable and supports cell adhesion as a path-way of interest in the genetics of white matter integrity.Also related to cell-adhesion is LPP,which encodes a shuttle protein,and plays a role in cell proliferation,differentiation and migration,presumably through its interactions with the Wnt signalling pathway (Grunewald et al.,2009).Better known for its role in white matter, however,is HEPACAM,a gene encoding the adhesion protein GlialCAM,which is predominantly located in axons and myelin, and mutations in this gene have been associated with macroence-phaly,autism and mental retardation(Lo´pez-Herna´ndez et al., 2011).Equally relevant to white matter is ROBO4,which,as a ROBO-family member,is critical for axon guidance and angiogenesis during embryonic development(Couch and Condron,2002).

Two genes in the top-SNP results are closely involved with calcium signalling in the brain:CRBN and CAMK2D.CRBN,or cereblon,controls the density and number of calcium-dependent potassium channels on the neuronal surface,thereby influencing the excitability of neurons(Chang and Stewart,2011).CRBN mutations in humans are linked to mental retardation,and CRBN knock-out mice exhibit selective memory impairment (Rajadhyaksha et al.,2012).CAMK2D encodes a multifunctional calcium/calmodulin-dependent protein kinase,which are essen-tial mediators of long-term potentiation and are thus important in neuronal plasiticity and learning.

Although all genes in our top results are brain-expressed to some extent,some of their functions are poorly characterised or too general to warrant more in depth interpretation in this context.In general,these top SNP results are presented as an illustration of what was obtained with the current genome-wide association method,but of course in the absence of independent replication and genome-wide significance,they should be inter-preted with extreme caution.Nevertheless,as explained above,a few of these genes are consistent with prominent roles in brain development and plasticity,including at least two(HEPACAM and ROBO4)with particular relevance to white matter.

4.2.Pathways

Simple SNP associations resulting from GWAS require very con-servative significance thresholds in order to accommodate the type-I error resulting from testing hundreds of thousands of SNPs at the same time.Pathway analysis provides a complimentary approach to address the contribution of many,modestly associated SNPs across

Table3

Relevant pathways significantly enriched for top SNP and top gene results.

KEGG pathway P-value(BH adj)

for SNP enrichment P-value(BH adj)

for gene enrichment

Axon guidance

Main 4.73E-080.0216

Interaction7.60E-080.0405

MAPK signalling pathway

Main 3.87E-060.006

Interaction 3.05E-100.0003

Calcium signalling pathway

Main 4.00E-06À

Interaction 2.09E-08À

Regulation of actin cytoskeleton

Main 1.36E-050.004

Interaction 1.92E-060.0014

Cell adhesion molecules(CAMs)

Main 2.71E-05À

Interaction 1.32E-050.0459

Tight junction

Main9.76E-05À

Interaction 1.27E-070.0009

Focal adhesion

Main 2.00E-040.0307

Interaction8.78E-090.005

Long-term potentiation

Main 2.58E-02À

Interaction 6.86E-050.0348

Phosphatidylinositol signalling system

Main 3.39E-020.0466

Interaction 4.19E-060.0047

Wnt signalling pathway

Main 3.87E-020.0386

Interaction 1.63E-02À

ErbB signalling pathway

MainÀÀ

Interaction7.98E-050.0026

Neurotrophin signalling pathway

MainÀ0.0391

Interaction 1.90E-030.0017

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]]5the genome.This approach is therefore more in line with the polygenic architecture of BD(Purcell et al.,2009;Evans et al.,2009).

Amongst our top-enriched pathways were pathways previously implicated in BD,as well as pathways known to be involved in brain development and plasticity,some of which are known to play a significant role in white matter development particularly.Axon guidance is a complex process that is essential for the establish-ment of long-range structural connections in the developing brain. Axon guidance has previously been reported in the top pathways in association to BD(Askland et al.,2009;Torkamani et al.,2008).The present results confirm these previousfindings,and are in line with known white matter pathology in BD patients and their relatives (Mahon et al.,2010;Sprooten et al.,2011a).

The ErbB signalling pathway is also a well-studied pathway in relation to both BD and white matter.Several SNPs and haplo-types within ErbB4and NRG1,its activating molecule,have been associated with BD in candidate-gene studies(Stefansson et al., 2002,2003;Hall et al.,2006;Ke´ri et al.,2009;Li et al.,2006; Norton et al.,2006;Silberberg et al.,2006;Thomson et al.,2007). On a molecular level,ErbB4and NRG1interact to serve many functions key to brain development and plasticity,including glial proliferation and differentiation,axon guidance,and axon–glia signalling to promote myelination(Chen et al.,2006;Mei and Xiong,2008;Corfas et al.,2004;Lo´pez-Bendito et al.,2006; Taveggia et al.,2008).Correspondingly,several structural MRI and DTI studies have shown association of SNPs in ErbB4and NRG1with white matter density or integrity(Konrad et al.,2009; McIntosh et al.,2008a;Sprooten et al.,2009;Winterer et al., 2008;Zuliani et al.,2011).

Other catergories of pathways listed in Table3have so far demonstrated no relevance to white matter per se,but are known to be involved in the pathology of BD for various other reasons.Of special interest may be calcium signalling,a pathway essential for synaptic plasticity and regulation of neurotransmitter release,and which has been significantly associated to BD in previous path-way analyses(Askland et al.,2009;Torkamani et al.,2008;Wang et al.,2010).Furthermore,the involvement in BD is supported by several candidate genes including genome-wide supported CAC-NA1C(Ferreira et al.,2008;Sklar et al.2011).Another pathway with well-known links to BD is the PI pathway,which is a therapeutic target of lithium action.Furthermore,the candidate gene DGKH is part of the PI pathway(Baum et al.,2008),and our group recently found an interaction of this genome-wide sup-ported DGKH SNP with genetic risk on brain activation in the Bipolar Family Study(Whalley et al.,2012).The PI pathway was also the top-enriched pathway in BD in a previous genome-wide pathway analysis(Peng et al.,2010).

A few other significantly enriched pathways,including focal adhesion,CAMs,and regulation of actin cytoskeleton,are important for the development and maintenance of cell and tissue structure, including white matter tissue.CAMs have also previously been listed in the top-enriched pathways in BD GWAS(O’Dushlaine et al.,2011;Askland et al.,2009),and in relation to FA(Lopez et al., 2012).

Some of these selected pathways(Table3)were enriched in the top-SNPs and top-genes of both the main and the interaction models,but the vast majority were more significant for the interaction model than for the main model.This reflects the fact that genes involved in these processes were differently associated with white matter integrity in HR than in HC subjects,rather than equally associated with white matter across both groups.This finding tentatively suggests that the‘‘normal’’processes in which these molecular pathways affect white matter integrity may be disrupted in BD.Alternatively it could also mean that in HR subjects the gene-FA associations are modulated or obscured by environmental factors related to familial risk to BD.4.3.Limitations

There are a few limitations to this study which require further elaboration.Firstly,our sample size is small compared to most other GWAS.Samples that have combined genome-wide and DTI data are rare to date,and difficult and expensive to obtain.The current sample size is similar to previous single-site GWAS using brain imaging phenotypes(Lopez et al.,2012;Potkin et al.,2009a, 2009b),and it is reassuring that many of our results are in line with known biological mechanisms of white matter development and structural integrity,as well as with recent advances in the genetics of BD.Nevertheless,the small samples size is the main limitation of the present study,and our analysis should be interpreted as exploratory until the results have been replicated independently.

None of our top single markers reached a p-value of o5Â10À8,the current genome-wide significance threshold. However,significance values were equivalent to those reported in BD case-control GWAS of samples10times larger than our sample(Sklar et al.,2008;Smith et al.,2011).The results are therefore promising for future genome-wide imaging studies and support the notion that endophenotypes require smaller sample sizes than case-control studies.

Our participants were relatively young and included individuals at high risk for BD,enabling the investigation of interactions between general familial risk and specific SNPs.Of note,given the young age of the unaffected relatives,a small proportion of them can be expected to develop BD later,which may have influenced our results.On the other hand,an older sample of healthy unaffected relatives could have biased the genetic risk background towards protective genes,which could obscure the effects of risk variants.In the future,clinical follow-up of this young at-risk sample should enable the separation of premorbid effects from familial-risk effects.

Finally,although the application of pathway analysis is both informative for the identification of molecular mechanisms under-lying disease,and increases power by taking a more polygenic approach,there remain several drawbacks to its application.Firstly, variability in gene size biases the SNP-based results towards path-ways containing larger genes.Although the gene-wide p-values calculated by VEGAS improve this bias,they may partly be over-correct for it,as mentioned in Section3.Another bias is related to pathway size,in that pathways containing more genes have more power,although in using a hypergeometric test comparing a gene-set to the whole genome,as in our case,this is not accompanied by type-I error inflation.A bias truly inherent to pathway analysis is its dependence on current knowledge of molecular functions and interactomes,and completeness of the databases.Due to different availability of resources and variable scientific progress across disciplines,some pathways are more likely to have annotated functions related to some diseases rather than others,i.e.heart disease and cancer.Thus,certain pathways may seem irrelevant to BD or white matter structure given our current knowledge,but future progress in biomedical sciences could reveal otherwise.

4.4.Conclusions

In our sample,a global measure of white matter integrity was associated with genetic variation in genes with functions broadly related to cell adhesion,axon guidance,and neuronal plasticity. Many of these variants were differently associated to FA in individuals at high genetic risk for BD,suggesting that their functions may be disrupted by the presence of other risk-variants or environmental factors associated with familial risk. Although the present top SNPsfindings require independent replication,the results are consistent with our current genetic

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]] 6

knowledge of BD,and therefore encourage the use of global FA as an intermediate phenotype of BD in future GWAS. Acknowledgements

All authors confirm they have nofinancial or other conflicts of interest with regard to this study.ES and AMM are supported by a NARSAD independent investigator award.JES is supported by a Clinical Research Fellowship from the Wellcome Trust.JH is supported by a Senior Clinical Fellowship from the Chief Scientist Office in Scotland.HCW is supported by a Dorothy Hodgkin Fellowship from the Royal Society(DH080018).HCW,JH,SML and AMM have receivedfinancial support from Pfizer(formerly Wyeth)in relation to other imaging studies of people with schizophrenia and bipolar disorder.We would like to thank all of the participants who took part in the study and the radio-graphers who acquired the MRI scans.This study was conducted at the Brain Research Imaging Centre(http://www.bric.ed.ac.uk) which is supported by SINAPSE(Scottish Imaging Network,a Platform for Scientific Excellence,http://www.sinapse.ac.uk).The investigators also acknowledge thefinancial support of National Health Service(NHS)Research Scotland,through the Scottish Mental Health Research Network(http://www.smhrn.org.uk) who provided assistance with subject recruitment and cognitive assessments.All imaging aspects also receivedfinancial support from the Dr.Mortimer and Theresa Sackler Foundation. Appendix A.Supporting information

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.psychres. 2012.11.002.

References

Adler,C.M.,Adams,J.,DelBello,M.P.,Holland,S.K.,Schmithorst,V.,Levine,A., Jarvis,K.,Strakowski,S.M.,2006.Evidence of white matter pathology in bipolar disorder adolescents experiencing theirfirst episode of mania:a diffusion tensor imaging study.American Journal of Psychiatry163(2), 322–324.

Askland,K.,Read,C.,Moore,J.,2009.Pathways-based analyses of whole-genome association study data in bipolar disorder reveal genes mediating ion channel activity and synaptic neurotransmission.Human Genetics125(1),63–79. Baum,A.E.,Akula,N.,Cabanero,M.,Cardona,I.,Corona,W.,Klemens,B.,Schulze, T.G.,Cichon,S.,Rietschel,M.,N¨othen,M.M.,Georgi, A.,Schumacher,J., Schwarz,M.,Abou Jamra,R.,H¨ofels,S.,Propping,P.,Satagopan,J.,Detera-Wadleigh,S.D.,Hardy,J.,McMahon,F.J.,2008.A genome-wide association study implicates diacylglycerol kinase eta(DGKH)and several other genes in the etiology of bipolar disorder.Molecular Psychiatry13(2),197–207. Benedetti,F.,Yeh,P.-H.,Bellani,M.,Radaelli,D.,Nicoletti,M.A.,Poletti,S.,Falini,A., Dallaspezia,S.,Colombo,C.,Scotti,G.,Smeraldi,E.,Soares,J.C.,Brambilla,P., 2011.Disruption of white matter integrity in bipolar depression as a possible structural marker of illness.Biological Psychiatry69(4),309–317. Benjamini,Y.,Hochberg,Y.,1995.Controlling the false discovery rate:a practical and powerful approach to multiple testing.Journal of the Royal Statistical Society Series B57,2–300.

Budde,M.D.,Kim,J.H.,Liang,H.F.,Schmidt,R.E.,Russell,J.H.,Cross,A.H.,Song,S.K., 2007.Toward accurate diagnosis of white matter pathology using diffusion tensor imaging.Magnetic Resonance in Medicine57,688–695.

Cercignani,M.,Inglese,M.,Pagani,E.,Comi,G.,Filippi,M.,2001.Mean diffusivity and fractional anisotropy histograms of patients with multiple sclerosis.

American Journal of Neuroradiology22,952–958.

Chaddock,C.A.,Barker,G.J.,Marshall,N.,Schulze,K.,Hall,M.H.,Fern,A.,Walshe, M.,Bramon,E.,Chitnis,X.A.,Murray,R.,McDonald,C.,2009.White matter microstructural impairments and genetic liability to familial bipolar I disorder.

British Journal of Psychiatry194(6),527–534.

Chang,X.B.,Stewart,A.K.,2011.What is the functional role of the thalidomide binding protein cereblon?International Journal of Biochemistry and Molecular Biology2(3),287–294.

Chen,S.,Velardez,M.O.,Warot,X.,Yu,Z.-X.,Miller,S.J.,Cros,D.,Corfas,G.,2006.

Neuregulin1-erbB signalling is necessary for normal myelination and sensory

function.Journal of Neuroscience:The Official Journal of the Society for Neuroscience26(12),3079–3086.

Chiang,M.-C.,Barysheva,M.,Shattuck,D.W.,Lee,A.D.,Madsen,S.K.,Avedissian,C., Klunder, A.D.,Toga, A.W.,McMahon,K.L.,de Zubicaray,G.I.,Wright,M.J., Srivastava, A.,Balov,N.,Thompson,P.M.,2009.Genetics of brainfiber architecture and intellectual performance.Journal of Neuroscience29(7), 2212–2224.

Chiang,M.-C.,McMahon,K.L.,de Zubicaray,G.I.,Martin,N.G.,Hickie,I.,Toga,A.W., Wright,M.J.,Thompson,P.M.,2011.Genetics of white matter development:a DTI study of705twins and their siblings aged12to29.NeuroImage54(3), 2308–2317.

Cichon,S.,Craddock,N.,Daly,M.,Faraone,S.V.,Gejman,P.V.,Kelsoe,J.,Lehner,T., Levinson,D.F.,Moran,A.,Sklar,P.,Sullivan,P.F.,Psychiatric GWAS Consortium Coordinating Committee,2009.Genomewide association studies:history, rationale,and prospects for psychiatric disorders.American Journal of Psy-chiatry166(5),540–556.

Corfas,G.,Roy,K.,Buxbaum,J.D.,2004.Neuregulin1-erbB signalling and the molecular/cellular basis of schizophrenia.Nature Neuroscience7(6),575–580. Couch,J.,Condron,B.,2002.Axon guidance:Comm hither,Robo.Current Biology 12(21),R741–R742.

Craddock,N.,Sklar,P.,2009.Genetics of bipolar disorder:successful start to a long journey.Trends in Genetics25(2),99–105.

Evans,D.M.,Visscher,P.M.,Wray,N.R.,2009.Harnessing the information con-tained within genome-wide association studies to improve individual predic-tion of complex disease risk.Human Molecular Genetics18(18),3525–3531. Ferreira,M.A.,O’Donovan,M.C.,Meng,Y.A.,Jones,I.R.,Ruderfer,D.M.,Jones,L.,Fan, J.,Kirov,G.,Perlis,R.H.,Green,E.K.,Smoller,J.W.,Grozeva,D.,Stone,J.,Nikolov,

I.,Chambert,K.,Hamshere,M.L.,Nimgaonkar,V.L.,Moskvina,V.,Thase,M.E.,

Caesar,S.,Sachs,G.S.,Franklin,J.,Gordon-Smith,K.,Ardlie,K.G.,Gabriel,S.B., Fraser,C.,Blumenstiel,B.,Defelice,M.,Breen,G.,Gill,M.,Morris,D.W.,Elkin,

A.,Muir,W.J.,McGhee,K.A.,Williamson,R.,MacIntyre,D.J.,MacLean,A.W.,St,

C.D.,Robinson,M.,Van Beck,M.,Pereira,A.C.,Kandaswamy,R.,McQuillin,A.,

Collier, D.A.,Bass,N.J.,Young, A.H.,Lawrence,J.,Ferrier,I.N.,Anjorin, A., Farmer, A.,Curtis, D.,Scolnick, E.M.,McGuffin,P.,Daly,M.J.,Corvin, A.P., Holmans,P.A.,Blackwood,D.H.,Gurling,H.M.,Owen,M.J.,Purcell,S.M.,Sklar, P.,Craddock,N.,Wellcome Trust Case Control Consortium,2008.Collaborative genome-wide association analysis supports a role for ANK3and CACNA1C in bipolar disorder.Nature Genetics40(9),1056–1058.

Gershon,E.S.,Hamovit,J.,Guroff,J.J.,Dibble,E.,Leckman,J.F.,Sceery,W.,Targum, S.D.,Nurnberger,J.I.,Goldin,L.R.,Bunney,W.E.,1982.A family study of schizoaffective,bipolar I,bipolar II,unipolar,and normal control probands.

Archives of General Psychiatry39(10),1157–1167.

Glahn,D.C.,Thompson,P.M.,Blangero,J.,2007.Neuroimaging endophenotypes: strategies forfinding genes influencing brain structure and function.Human Brain Mapping28(6),488–501.

Gottesman,I.I.,Gould,T.D.,2003.The endophenotype concept in psychiatry: etymology and strategic intentions.American Journal of Psychiatry160(4), 636–5.

Gouw, A.A.,Seewann, A.,Vrenken,H.,van der Flier,W.M.,Rozemuller,J.M., Barkhof,F.,Scheltens,P.,Geurts,J.J.G.,2008.Heterogeneity of white matter hyperintensities in Alzheimer’s disease:post-mortem quantitative MRI and neuropathology.Brain131,3286–3298.

Grunewald,T.G.,Pasedag,S.M.,Butt,E.,2009.Cell adhesion and transcriptional activity—defining the role of the novel protooncogene LPP.Translational Oncology2(3),107–116.

Hall,J.,Whalley,H.C.,Job,D.E.,Baig,B.J.,McIntosh,A.M.,Evans,K.L.,Thomson,P.A., Porteous,D.J.,Cunningham-Owens,D.G.,Johnstone,E.C.,Lawrie,S.M.,2006.A neuregulin1variant associated with abnormal cortical function and psychotic symptoms.Nature Neuroscience9(12),1477–1478.

Heng,S.,Song,A.W.,Sim,K.,2010.White matter abnormalities in bipolar disorder: insights from diffusion tensor imaging studies.Journal of Neural Transmission (Vienna,Austria:1996)117(5),639–654.

Ke´ri,S.,Kiss,I.,Kelemen,O.,2009.Effects of a neuregulin1variant on conversion to schizophrenia and schizophreniform disorder in people at high risk for psychosis.Molecular Psychiatry14(2),118–119.

Kochunov,P.,Glahn,D.C.,Lancaster,J.L.,Winkler,A.M.,Smith,S.,Thompson,P.M., Almasy,L.,2010.Genetics of microstructure of cerebral white matter using diffusion tensor imaging.NeuroImage53(3),1109–1116.

Konrad,A.,Vucurevic,G.,Musso,F.,Stoeter,P.,Dahmen,N.,Winterer,G.,2009.

ErbB4genotype predicts left frontotemporal structural connectivity in human brain.Neuropsychopharmacology:Official Publication of the American College of Neuropsychopharmacology34(3),1–650.

Li,D.,Collier,D.A.,He,L.,2006.Meta-analysis shows strong positive association of the neuregulin1(NRG1)gene with schizophrenia.Human Molecular Genetics 15(12),1995–2002.

Liu,J.Z.,McRae,A.F.,Nyholt,D.R.,Medland,S.E.,Wray,N.R.,Brown,K.M.,AMFS Investigators,Hayward,N.K.,Montgomery,G.W.,Visscher,P.M.,Martin,N.G., Macgregor,S.,2010.A versatile gene-based test for genome-wide association studies.American Journal of Human Genetics87(1),139–145.

Lopez,L.M.,Bastin,M.E.,Mun˜oz Maniega,S.M.,Penke,L.,Davies,G.,Christoforou,

A.,Valde´s Herna´ndez,M.C.,Royle,N.A.,Tenesa,A.,Starr,J.M.,Porteous,D.J.,

Wardlaw,J.M.,Deary,I.J.,2012.A genome-wide search for genetic influences and biological pathways related to the brain’s white matter integrity.Neuro-biology of Aging33(8)1847.e1-1847.e14.

Lo´pez-Bendito,G.,Cautinat,A.,Sa´nchez,J.A.,Bielle,F.,Flames,N.,Garratt,A.N., Talmage,D.A.,Role,L.W.,Charnay,P.,Marı´n,O.,Garel,S.,2006.Tangential

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]]7neuronal migration controls axon guidance:a role for neuregulin-1in thalamocortical axon navigation.Cell125(1),127–142.

Lo´pez-Herna´ndez,T.,Ridder,M.C.,Montolio,M.,Capdevila-Nortes,X.,Polder,E., Sirisi,S.,Duarri,A.,Schulte,U.,Fakler,B.,Nunes,V.,Scheper,G.C.,Martı´nez,A., Este´vez,R.,van der Knaap,M.S.,2011.Mutant GlialCAM causes megalencephalic leukoencephalopathy with subcortical cysts,benign familial macrocephaly,and macrocephaly with retardation and autism.American Journal of Human Genet-ics88(4),422–432.

Luciano,M.,Hansell,N.K.,Lahti,J.,Davies,G.,Medland,S.E.,R¨aikk¨onen,K.,Tenesa,

A.,Widen,E.,McGhee,K.A.,Palotie,A.,Liewald,D.,Porteous,D.J.,Starr,J.M.,

Montgomery,G.W.,Martin,N.G.,Eriksson,J.G.,Wright,M.J.,Deary,I.J.,2011.

Whole genome association scan for genetic polymorphisms influencing information processing speed.Biological Psychology86,193–202. MacDonald,C.L.,Dikranian,K.,Bayly,P.,Holtzman,D.,Brody,D.,2007.Diffusion tensor imaging reliably detects experimental traumatic axonal injury and indicates approximate time of injury.Journal of Neuroscience27, 11869–11876.

Mahon,K.,Burdick,K.E.,Szeszko,P.R.,2010.A role for white matter abnormalities in the pathophysiology of bipolar disorder.Neuroscience and Biobehavioral Reviews34(4),533–554.

Manolio,T.A.,Brooks,L.D.,Collins,F.S.,2008.A HapMap harvest of insights into the genetics of common disease.Journal of Clinical Investigation118(5), 1590–1605.

McGuffin,P.,Rijsdijk,F.,Andrew,M.,Sham,P.,Katz,R.,Cardno,A.,2003.The heritability of bipolar affective disorder and the genetic relationship to unipolar depression.Archives of General Psychiatry60(5),497–502. McIntosh, A.M.,Moorhead,T.W.J.,Job, D.,Lymer,G.K.S.,Mun˜oz Maniega,S., McKirdy,J.,Sussmann,J.E.D.,Baig,B.J.,Bastin,M.E.,Porteous,D.,Evans,K.L., Johnstone,E.C.,Lawrie,S.M.,Hall,J.,2008a.The effects of a neuregulin1variant on white matter density and integrity.Molecular Psychiatry13(11), 1054–1059.

McIntosh,A.M.,Mun˜oz Maniega,S.,Lymer,G.K.S.,McKirdy,J.,Hall,J.,Sussmann, J.E.D.,Bastin,M.E.,Clayden,J.D.,Johnstone,E.C.,Lawrie,S.M.,2008b.White matter tractography in bipolar disorder and schizophrenia.Biological Psychia-try(12),1088–1092.

Meda,S.A.,Narayanan,B.,Liu,J.,Perrone-Bizzozero,N.I.,Stevens,M.C.,Calhoun,V.D., Glahn,D.C.,Shen,L.,Risacher,S.L.,Saykin,A.J.,Pearlson,G.D.,2012.A large scale multivariate parallel ICA method reveals novel imaging-genetic relationships for Alzheimer’s disease in the ADNI cohort.Neuroimage60,1608–1621.

Mei,L.,Xiong,W.C.,2008.Neuregulin1in neural development,synaptic plasticity and schizophrenia.Nature Reviews.Neuroscience9(6),437–452.

Mun˜oz Maniega,S.,Lymer,G.K.S.,Bastin,M.E.,Marjoram,D.,Job,D.E.,Moorhead, T.W.J.,Owens, D.G.,Clayden,J.D.,Johnstone, E.C.,Lawrie,S.M.,2008.A diffusion tensor MRI study of white matter integrity in subjects at high genetic risk of schizophrenia.Schizophrenia Research106(2–3),132–139. Norton,N.,Moskvina,V.,Morris, D.W.,Bray,N.J.,Zammit,S.,Williams,N.M., Williams,H.J.,Preece,A.C.,Dwyer,S.,Wilkinson,J.C.,Spurlock,G.,Kirov,G., Buckland,P.,Waddington,J.L.,Gill,M.,Corvin,A.P.,Owen,M.J.,O’Donovan, M.C.,2006.Evidence that interaction between neuregulin1and its receptor erbB4increases susceptibility to schizophrenia.American Journal of Medical Genetics.Part B,Neuropsychiatric Genetics141B(1),96–101.

O’Donovan,M.C.,Craddock,N.J.,Owen,M.J.,2009.Genetics of psychosis;insights from views across the genome.Human Genetics126(1),3–12.

O’Dushlaine,C.,Kenny,E.,Heron,E.,Donohoe,G.,Gill,M.,Morris,D.,Corvin,A.,2011.

Molecular pathways involved in neuronal cell adhesion and membrane scaffold-ing contribute to schizophrenia and bipolar disorder susceptibility.Molecular Psychiatry16(3),286–292.

Peng,G.,Luo,L.,Siu,H.,Zhu,Y.,Hu,P.,Hong,S.,Zhao,J.,Zhou,X.,Reveille,J.D.,Jin, L.,Amos,C.I.,Xiong,M.,2010.Gene and pathway-based second-wave analysis of genome-wide association studies.European Journal of Human Genetics18

(1),111–117.

Potkin,S.G.,Turner,J.A.,Fallon,J.A.,Lakatos,A.,Keator,D.B.,Guffanti,G.,Macciardi,F., 2009a.Gene discovery through imaging genetics:identification of two novel genes associated with schizophrenia.Molecular Psychiatry14(4),416–428. Potkin,S.G.,Turner,J.A.,Guffanti,G.,Lakatos,A.,Fallon,J.H.,Nguyen,D.D.,Mathalon,

D.,Ford,J.,Lauriello,J.,Macciardi,F.,2009b.A genome-wide association study of

schizophrenia using brain activation as a quantitative phenotype.Schizophrenia Bulletin35(1),96–108.

Purcell,S.,Neale, B.,Todd-Brown,K.,Thomas,L.,Ferreira,M.A.R.,Bender, D., Maller,J.,Sklar,P.,de Bakker,P.I.,Daly,M.J.,Sham,P.C.,2007.PLINK:a tool set for whole-genome association and population-based linkage analyses.Amer-ican Journal of Human Genetics81(3),559–575.

Purcell,S.M.,Wray,N.R.,Stone,J.L.,Visscher,P.M.,O’Donovan,M.C.,Sullivan,P.F., Sklar,P.,2009.Common polygenic variation contributes to risk of schizo-phrenia and bipolar disorder.Nature460(7256),748–752.

Rajadhyaksha,A.M.,Ra,S.,Kishinevsky,S.,Lee,A.S.,Romanienko,P.,DuBoff,M., Yang,C.,Zupan,B.,Byrne,M.,Daruwalla,Z.R.,Mark,W.,Kosofsky,B.E.,Toth, M.,Higgins,J.J.,2012.Behavioral characterization of cereblon forebrain-specific conditional null mice:a model for human non-syndromic intellectual disability.Behavioural Brain Research226(2),428–434.

Raychaudhuri,S.,Korn,J.M.,McCarroll,S.A.,Altshuler,D.,Sklar,P.,Purcell,S.,Daly, M.J.,2010.Accurately assessing the risk of schizophrenia conferred by rare copy-number variation affecting genes with brain function.PLoS Genetics6,9. Reiss,K.,Saftig,P.,2009.The‘a disintegrin and metalloprotease’(ADAM)family of sheddases:physiological and cellular functions.Seminars in Cell Develop-mental Biology20(2),126–137.Schmierer,K.,Wheeler-Kingshott,C.A.M.,Boulby,P.A.,Scaravilli,F.,Altmann,D.R., Barker,G.J.,Tofts,P.S.,Miller,D.H.,2007.Diffusion tensor imaging of post mortem multiple sclerosis brain.Neuroimage35,467–477.

Seals, D.F.,Courtneidge,S.A.,2003.The ADAMs family of metalloproteases: multidomain proteins with multiple functions.Genes and Development17

(1),7–30.

Sexton,C.E.,Mackay,C.E.,Ebmeier,K.P.,2009.A systematic review of diffusion tensor imaging studies in affective disorders.Biological Psychiatry66(9), 814–823.

Silberberg,G.,Darvasi,A.,Pinkas-Kramarski,R.,Navon,R.,2006.The involvement of ErbB4with schizophrenia:association and expression studies.American Journal of Medical Genetics.Part B,Neuropsychiatric Genetics141B(2), 142–148.

Sklar,P.,Ripke,S.,Scott,L.J.,Andreassen,O.A.,Cichon,S.,Craddock,N.,Edenberg,

H.J.,Nurnberger,J.I.,Rietschel,M.,Blackwood,D.,Corvin,A.,Flickinger,M.,

Guan,W.,Mattingsdal,M.,McQuillin,A.,Kwan,P.,Wienker,T.F.,Daly,M., Dudbridge,F.,Holmans,P.A.,Lin,D.,Burmeister,M.,Greenwood,T.A.,Ham-shere,M.L.,Muglia,P.,Smith,E.N.,Zandi,P.P.,Nievergelt,C.M.,McKinney,R., Shilling,P.D.,Schork,N.J.,Bloss,C.S.,Foroud,T.,Koller,D.L.,Gershon,E.S.,Liu,

C.,Badner,J.A.,Scheftner,W.A.,Lawson,W.B.,Nwulia, E.A.,Hipolito,M.,

Coryell,W.,Rice,J.,Byerley,W.,McMahon,F.J.,Schulze,T.G.,Berrettini,W., Lohoff,F.W.,Potash,J.B.,Mahon,P.B.,McInnis,M.G.,Z¨ollner,S.,Zhang,P.,Craig,

D.W.,Szelinger,S.,Barrett,T.B.,Breuer,R.,Meier,S.,Strohmaier,J.,Witt,S.H.,

Tozzi,F.,Farmer,A.,McGuffin,P.,Strauss,J.,Xu,W.,Kennedy,J.L.,Vincent,J.B., Matthews,K.,Day,R.,Ferreira,M.A.,O’Dushlaine,C.,Perlis,R.,Raychaudhuri, S.,Ruderfer,D.,Hyoun,P.L.,Smoller,J.W.,Li,J.,Absher,D.,Thompson,R.C., Meng,F.G.,Schatzberg,A.F.,Bunney,W.E.,Barchas,J.D.,Jones,E.G.,Watson, S.J.,Myers,R.M.,Akil,H.,Boehnke,M.,Chambert,K.,Moran,J.,Scolnick,E., Djurovic,S.,Melle,I.,Morken,G.,Gill,M.,Morris,D.,Quinn,E.,M¨uhleisen,T.W., Degenhardt, F.A.,Mattheisen,M.,Schumacher,J.,Maier,W.,Steffens,M., Propping,P.,N¨othen,M.M.,Anjorin,A.,Bass,N.,Gurling,H.,Kandaswamy,R., Lawrence,J.,McGhee,K.,McIntosh,A.,McLean,A.W.,Muir,W.J.,Pickard,B.S., Breen,G.,St Clair,D.,Caesar,S.,Gordon-Smith,K.,Jones,L.,Fraser,C.,Green,

E.K.,Grozeva,D.,Jones,I.R.,Kirov,G.,Moskvina,V.,Nikolov,I.,O’Donovan,

M.C.,Owen,M.J.,Collier,D.A.,Elkin,A.,Williamson,R.,Young,A.H.,Ferrier,I.N., Stefansson,K.,Stefansson,H.,Thornorgeirsson,T.,Steinberg,S.,Gustafsson,O., Bergen,S.E.,Nimgaonkar,V.,Hultman,C.,Lande´n,M.,Lichtenstein,P.,Sullivan, P.,Schalling,M.,Osby,U.,Backlund,L.,Frise´n,L.,Langstrom,N.,Jamain,S., Leboyer,M.,Etain,B.,Bellivier,F.,Petursson,H.,Sigursson,E.,M¨uller-Mysok,

B.,Lucae,S.,Schwarz,M.,Schofield,P.R.,Martin,N.,Montgomery,G.W.,

Lathrop,M.,Oskarsson,H.,Bauer,M.,Wright,A.,Mitchell,P.B.,Hautzinger, M.,Reif,A.,Kelsoe,J.R.,Purcell,S.M.,Psychiatric GWAS Consortium Bipolar Disorder Working Group,2011.Large-scale genome-wide association analysis of bipolar disorder identifies a new susceptibility locus near ODZ4.Nature Genetics43(10),977–983.

Sklar,P.,Smoller,J.W.,Fan,J.,Ferreira,M.A.R.,Perlis,R.H.,Chambert,K.,Nimgaon-kar,V.L.,McQueen,M.B.,Faraone,S.V.,Kirby,A.,de Bakker,P.I.,Ogdie,M.N., Thase,M.E.,Sachs,G.S.,Todd-Brown,K.,Gabriel,S.B.,Sougnez,C.,Gates,C., Blumenstiel,B.,Defelice,M.,Ardlie,K.G.,Franklin,J.,Muir,W.J.,McGhee,K.A., MacIntyre,D.J.,McLean,A.,VanBeck,M.,McQuillin,A.,Bass,N.J.,Robinson,M., Lawrence,J.,Anjorin,A.,Curtis,D.,Scolnick,E.M.,Daly,M.J.,Blackwood,D.H., Gurling,H.M.,Purcell,S.M.,2008.Whole-genome association study of bipolar disorder.Molecular Psychiatry13(6),558–569.

Smith,E.N.,Koller,D.L.,Panganiban,C.,Szelinger,S.,Zhang,P.,Badner,J.A.,Barrett, T.B.,Berrettini,W.H.,Bloss,C.S.,Byerley,W.,Coryell,W.,Edenberg,H.J.,Foroud, T.,Gershon,E.S.,Greenwood,T.A.,Guo,Y.,Hipolito,M.,Keating,B.J.,Lawson, W.B.,Liu,C.,Mahon,P.B.,McInnis,M.G.,McMahon,F.J.,McKinney,R.,Murray, S.S.,Nievergelt,C.M.,Nurnberger,J.I.,Nwulia,E.A.,Potash,J.B.,Rice,J.,Schulze, T.G.,Scheftner,W.A.,Shilling,P.D.,Zandi,P.P.,Z¨ollner,S.,Craig,D.W.,Schork, N.J.,Kelsoe,J.R.,2011.Genome-wide association of bipolar disorder suggests an enrichment of replicable associations in regions near genes.PLoS Genetics7

(6),e1002134.

Smith,S.M.,Jenkinson,M.,Johansen-Berg,H.,Rueckert,D.,Nichols,T.E.,Mackay,C.E., Watkins,K.E.,Ciccarelli,O.,Cader,M.Z.,Matthews,P.M.,Behrens,T.E.,2006.

Tract-based spatial statistics:voxelwise analysis of multi-subject diffusion data.

NeuroImage31(4),1487–1505.

Sprooten,E.,Lymer,G.K.S.,Mun˜oz Maniega,S.,McKirdy,J.,Clayden,J.D.,Bastin, M.E.,Porteous,D.,Johnstone,E.C.,Lawrie,S.M.,Hall,J.,McIntosh,A.M.,2009.

The relationship of anterior thalamic radiation integrity to psychosis risk associated neuregulin-1variants.Molecular Psychiatry,143(237–238),233. Sprooten, E.,Sussmann,J.E.,Clugston, A.,Peel, A.,McKirdy,J.,William,T., Moorhead,J.,Shand,A.J.,Giles,S.,Bastin,M.E.,Hall,J.,Johnstone,E.C.,Lawrie, S.M.,McIntosh,A.M.,2011a.White matter integrity in individuals at high genetic risk of bipolar disorder.Biological Psychiatry.

Sprooten,E.,Sussmann,J.E.,Moorhead,T.W.,Whalley,H.C.,Ffrench-Constant,C., Blumberg,H.P.,Bastin,M.E.,Hall,J.,Lawrie,S.M.,McIntosh, A.M.,2011b.

Association of white matter integrity with genetic variation in an exonic DISC1 SNP.Molecular Psychiatry16(7),688–6685.

Stefansson,H.,Sarginson,J.,Kong,A.,Yates,P.,Steinthorsdottir,V.,Gudfinnsson,E., Gunnarsdottir,S.,Walker,N.,Petursson,H.,Crombie,C.,Ingason,A.,Gulcher,J.R., Stefansson St,K.,Clair,D.,2003.Association of neuregulin1with schizophrenia confirmed in a Scottish population.American Journal of Human Genetics72(1), 83–87.

Stefansson,H.,Sigurdsson,E.,Steinthorsdottir,V.,Bjornsdottir,S.,Sigmundsson,T., Ghosh,S.,Brynjolfsson,J.,Gunnarsdottir,S.,Ivarsson,O.,Chou,T.T.,Hjaltason,O., Birgisdottir,B.,Jonsson,H.,Gudnadottir,V.G.,Gudmundsdottir,E.,Bjornsson,A.,

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]] 8Ingvarsson,A.M.,Ingason,A.,Sigfusson,S.,Hardardottir,H.,Harvey,R.P.,Lai,D., Zhou,M.,Brunner,D.,Mutel,V.,Gonzalo,A.,Lemke,G.,Sainz,J.,Johannesson,G., Andresson,T.,Gudbjartsson,D.,Manolescu,A.,Frigge,M.L.,Gurney,M.E.,Kong,

A.,Gulcher,J.R.,Petursson,H.,Stefansson,K.,2002.Neuregulin1and suscept-

ibility to schizophrenia.American Journal of Human Genetics71(4),877–2. Stein,J.L.,Medland,S.E.,Vasquez,A.A.,Hibar,D.P.,Senstad,R.E.,Winkler,A.M., Toro,R.,Appel,K.,Bartecek,R.,Bergmann,Ø.,Bernard,M.,Brown,A.A.,Cannon,

D.M.,Chakravarty,M.M.,Christoforou,A.,Domin,M.,Grimm,O.,Hollinshead,

M.,Holmes,A.J.,Homuth,G.,Hottenga,J.J.,Langan,C.,Lopez,L.M.,Hansell, N.K.,Hwang,K.S.,Kim,S.,Laje,G.,Lee,P.H.,Liu,X.,Loth,E.,Lourdusamy,A., Mattingsdal,M.,Mohnke,S.,Maniega,S.M.,Nho,K.,Nugent,A.C.,O’Brien,C., Papmeyer,M.,P¨utz,B.,Ramasamy,A.,Rasmussen,J.,Rijpkema,M.,Risacher, S.L.,Roddey,J.C.,Rose,E.J.,Ryten,M.,Shen,L.,Sprooten,E.,Strengman,E., Teumer,A.,Trabzuni,D.,Turner,J.,van Eijk,K.,van Erp,T.G.,van,Tol,M.J., Wittfeld,K.,Wolf,C.,Woudstra,S.,Aleman,A.,Alhusaini,S.,Almasy,L.,Binder,

E.B.,Brohawn,D.G.,Cantor,R.M.,Carless,M.A.,Corvin,A.,Czisch,M.,Curran,

J.E.,Davies,G.,de Almeida,M.A.,Delanty,N.,Depondt,C.,Duggirala,R.,Dyer, T.D.,Erk,S.,Fagerness,J.,Fox,P.T.,Freimer,N.B.,Gill,M.,G¨oring,H.H.,Hagler,

D.J.,Hoehn,D.,Holsboer,F.,Hoogman,M.,Hosten,N.,Jahanshad,N.,Johnson,

M.P.,Kasperaviciute,D.,Kent,J.W.,Kochunov,P.,Lancaster,J.L.,Lawrie,S.M., Liewald,D.C.,Mandl,R.,Matarin,M.,Mattheisen,M.,Meisenzahl,E.,Melle,I., Moses,E.K.,M¨uhleisen,T.W.,Nauck,M.,N¨othen,M.M.,Olvera,R.L.,Pandolfo, M.,Pike,G.B.,Puls,R.,Reinvang,I.,Renterı´a,M.E.,Rietschel,M.,Roffman,J.L., Royle,N.A.,Rujescu,D.,Savitz,J.,Schnack,H.G.,Schnell,K.,Seiferth,N.,Smith,

C.,Steen,V.M.,Valde´s Herna´ndez,M.C.,Van den Heuvel,M.,van der Wee,N.J.,

Van Haren,N.E.,Veltman,J.A.,V¨olzke,H.,Walker,R.,Westlye,L.T.,Whelan,

C.D.,Agartz,I.,Boomsma,D.I.,Cavalleri,G.L.,Dale,A.M.,Djurovic,S.,Drevets,

W.C.,Hagoort,P.,Hall,J.,Heinz,A.,Jack,C.R.,Foroud,T.M.,Le Hellard,S., Macciardi, F.,Montgomery,G.W.,Poline,J.B.,Porteous, D.J.,Sisodiya,S.M., Starr,J.M.,Sussmann,J.,Toga,A.W.,Veltman,D.J.,Walter,H.,Weiner,M.W., Alzheimer’s Disease Neuroimaging Initiative,EPIGEN Consortium,IMAGEN Consortium,Saguenay Youth Study Group,Bis,J.C.,Ikram,M.A.,Smith,A.V., Gudnason,V.,Tzourio,C.,Vernooij,M.W.,Launer,L.J.,DeCarli,C.,Seshadri,S., Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium, Andreassen,O.A.,Apostolova,L.G.,Bastin,M.E.,Blangero,J.,Brunner,H.G., Buckner,R.L.,Cichon,S.,Coppola,G.,de Zubicaray,G.I.,Deary,I.J.,Donohoe,G., de Geus,E.J.,Espeseth,T.,Ferna´ndez,G.,Glahn,D.C.,Grabe,H.J.,Hardy,J., Hulshoff Pol,H.E.,Jenkinson,M.,Kahn,R.S.,McDonald,C.,McIntosh,A.M., McMahon,F.J.,McMahon,K.L.,Meyer-Lindenberg,A.,Morris,D.W.,M¨uller-Myhsok,B.,Nichols,T.E.,Ophoff,R.A.,Paus,T.,Pausova,Z.,Penninx,B.W., Potkin,S.G.,S¨amann,P.G.,Saykin,A.J.,Schumann,G.,Smoller,J.W.,Wardlaw, J.M.,Weale,M.E.,Martin,N.G.,Franke, B.,Wright,M.J.,Thompson,P.M., Enhancing Neuro Imaging Genetics through Meta-Analysis Consortium, 2012.Common genetic polymorphisms are associated with human hippo-campal and intracranial volumes.Nature Genetics44(5),552–561.Sussmann,J.E.,Lymer,G.K.S.,McKirdy,J.,Moorhead,T.W.J.,Mun˜oz Maniega,S.,Job,

D.,Hall,J.,Bastin,M.

E.,Johnstone,E.C.,Lawrie,S.M.,McIntosh,A.M.,2009.

White matter abnormalities in bipolar disorder and schizophrenia detected using diffusion tensor magnetic resonance imaging.Bipolar Disorders11(1), 11–18.

Taveggia,C.,Thaker,P.,Petrylak,A.,Caporaso,G.L.,Toews,A.,Falls,D.L.,Einheber, S.,Salzer,J.L.,2008.Type III neuregulin-1promotes oligodendrocyte myelina-tion.Glia56(3),284–293.

Thomson,P.A.,Christoforou,A.,Morris,S.W.,Adie,E.,Pickard,B.S.,Porteous,D.J., Muir,W.J.,Blackwood,D.H.,Evans,K.L.,2007.Association of Neuregulin1with schizophrenia and bipolar disorder in a second cohort from the Scottish population.Molecular Psychiatry12(1),94–104.

Torkamani,A.,Topol,E.J.,Schork,N.J.,2008.Pathway analysis of seven common diseases assessed by genome-wide association.Genomics92(5),265–272. Underwood,C.K.,Kurniawan,N.D.,Butler,T.J.,Cowin,G.J.,Wallace,R.H.,2011.Non-invasive diffusion tensor imaging detects white matter degeneration in the spinal cord of a mouse model of amyotrophic lateral sclerosis.Neuroimage55,455–461. Wang,F.,Jackowski,M.,Kalmar,J.H.,Chepenik,L.G.,Tie,K.,Qiu,M.,Gong,G., Pittman,B.P.,Jones,M.M.,Shah,M.P.,Spencer,L.,Papademetris,X.,Constable, R.T.,Blumberg,H.P.,2008a.Abnormal anterior cingulum integrity in bipolar disorder determined through diffusion tensor imaging.The British Journal of Psychiatry:The Journal of Mental Science193(2),126–129.

Wang,F.,Kalmar,J.H.,Edmiston,E.,Chepenik,L.G.,Bhagwagar,Z.,Spencer,L., Pittman,B.,Jackowski,M.,Papademetris,X.,Constable,R.T.,Blumberg,H.P., 2008b.Abnormal corpus callosum integrity in bipolar disorder:a diffusion tensor imaging study.Biological Psychiatry(8),730–733.

Wang,K.,Li,M.,Hakonarson,H.,2010.Analysing biological pathways in genome-wide association studies.Nature Reviews.Genetics11(12),843–854. Whalley,H.C.,Papmeyer,M.,Romaniuk,L.,Johnstone,E.C.,Hall,J.,Lawrie,S.M., Sussmann,J.E.,McIntosh, A.M.,2012.Effect of variation in diacylglycerol kinase eta(DGKH)gene on brain function in a cohort at familial risk of bipolar disorder.Neuropsychopharmacology37(4),919–928.

Whalley,H.C.,Sussmann,J.E.,Chakirova,G.,Mukerjee,P.,Peel,A.,McKirdy,J.,Hall, J.,Johnstone, E.C.,Lawrie,S.M.,McIntosh, A.M.,2011.The neural basis of familial risk and temperamental variation in individuals at high risk of bipolar disorder.Biological Psychiatry70(4),343–349.

Winterer,G.,Konrad,A.,Vucurevic,G.,Musso,F.,Stoeter,P.,Dahmen,N.,2008.

Association of50end neuregulin-1(NRG1)gene variation with subcortical medial frontal microstructure in humans.NeuroImage40(2),712–718. Zuliani,R.,Moorhead,T.W.J.,Bastin,M.E.,Johnstone,E.C.,Lawrie,S.M.,Brambilla, P.,O’Donovan,M.C.,Owen,M.J.,Hall,J.,McIntosh,A.M.,2011.Genetic variants in the ErbB4gene are associated with white matter integrity.Psychiatry Research191(2),133–137.

E.Sprooten et al./Psychiatry Research](]]]])]]]–]]]9