integrated genetic and genomic approach in the singapore translational and clinical research in...

Review Article

Integrated genetic and genomic approach in theSingapore translational and clinical research in

psychosis study: an overvieweip_272 91..99

Kang Sim,1,2 Jimmy Lee,1,2 Mythily Subramaniam,2 Jian Jun Liu,3 Richard Keefe,4 Xiao Dong Zhang,4

Tih Shih Lee4 and Siow Ann Chong2

1Department of General Psychiatry,Woodbridge Hospital/Institute of MentalHealth; 2Research Division, Institute ofMental Health; 3Genome Institute ofSingapore, Agency for Science,Technology and Research; and4Duke-NUS, Singapore

Corresponding author: Dr Kang Sim,Department of General Psychiatry,Woodbridge Hospital/ Institute ofMental Health, 10, Buangkok View,Singapore 539747, Singapore.Email: [email protected]

Received 29 July 2010; accepted 13February 2011

Abstract

Aims: Schizophrenia is a severemental disorder with onset frequentlyin adolescence and followed bya chronic and disabling course.Although the exact pathophysiologyof this devastating disorder has notbeen clearly elucidated, a large part ofit has been attributed to genetic influ-ences. This article seeks to provide anoverview on what our group hasembarked on – to elucidate geneticrisk factors for schizophrenia withinthe Chinese ethnic group.

Methods: We plan to conduct anintegrated approach to interrogatecomprehensively the genome fromdifferent angles and in stages. Thefirst stage involves a genome-wideassociation study of 1000 cases ofschizophrenia-control pairs, with afollow-up replication study in another1000 cases of schizophrenia and in1000 controls, and combinationanalyses with groups from other

places including China and HongKong. Other than the genome-wideassociation study, gene sequencingfor purported candidate genes andcopy number variation analysis willbe performed. Neurocognitive inter-mediate phenotypes will be employedto deconstruct the complex schizo-phrenia phenotype in a bid toimprove association findings. Promis-ing leads from longitudinal gene andprotein expression in ultra-high-risksubjects who develop psychosis andschizophrenia (in a parallel study) willbe followed up as candidate genesand sequenced in the genetic analy-sis. Functional analysis forms the laststage of this integrated approach.

Conclusion: This integrated geneticand genomic approach will hopefullyhelp in further characterizing anddeepening our understanding ofmolecular pathophysiological me-chanisms underlying schizophrenia.

Key words: endophenotype, genetic, psychosis, schizophrenia.

INTRODUCTION

Schizophrenia (SZ) and related psychotic disordersare among the most disabling disorders to afflictmankind. Despite the low lifetime prevalence ofpsychosis and SZ, estimated to be about 3% and0.8% worldwide, they generate an enormous burdenin both economic cost and human suffering.1,2 Thereis currently no cure for SZ, and its aetiology is stillnot fully understood.3 The heritability rates for SZare 0.70–0.80,4 making genetic risk factors naturalkey targets for research.

The search for the genetic risk factors for SZ hasbeen pursued using family-based linkage5,6 andpopulation-based association analysis.7,8 Intensivelinkage analyses have been done in SZ using bothlarge pedigrees and small nuclear families. But,with few exceptions, the outcomes of such effortsare largely negative and inconsistent.5,6 The lessthan promising outcomes from the large joint-linkage analyses of multiple genome-wide datasetshave further highlighted the limitation of family-based linkage analysis. The factors limiting geneticstudies (including linkage analysis) in psychiatric

Early Intervention in Psychiatry 2011; 5: 91–99 doi:10.1111/j.1751-7893.2011.00272.x

First Impact Factor released in June 2010and now listed in MEDLINE!

© 2011 Blackwell Publishing Asia Pty Ltd 91

conditions are likely multifactorial, with the het-erogeneity of the diagnostic phenotype expected tobe an important contributor.9 The high clinical het-erogeneity of the disease phenotype of SZ is likely aresult of the underlying genetic heterogeneity thatwill greatly reduce the power of linkage analysis toprovide significant and consistent evidence, evenwhen analyzing a large number of families.

In this regard, a population-based associationanalysis has been promoted as another approachfor identifying moderate-risk alleles.10 The recentcompletion of the human HapMap project as wellas the rapid development of high-throughputgenotyping technology has further opened thedoor for pursuing genetic association analysis atthe whole genome level. The first-publishedgenome-wide association study (GWAS) on SZ wasperformed by Lencz and colleagues with a smallnumber of SZ patients and controls.11 This studyhas shown that advances in the availability of novelgenomic technologies and the improvementof genetic statistical analysis have opened upsome possibility for the identification of commongenetic risk factors for SZ. Since then, severalGWASs for SZ have been published.7,8,12–14 Suscep-tibility genes that are associated at genome-widesignificance level with SZ include major histocom-patibility complex region on chromosomes 6p21.3–22.1, a marker located upstream of the neurograningene on 11q24.2 and a marker in intron fourof transcription factor 4 on 18q21.2.8 The copynumber variation (CNV) studies suggest that whileCNVs are found in psychosis, the large deletionsand duplications are more likely found in SZ ratherthan in bipolar disorder.15 Overall, GWAS and CNVstudies have shown potential in providing insightsinto the biological basis of SZ. Several points needto be considered to further advance this under-standing. First, better integration of GWAS andCNV studies with other functional findings such asdata from gene expression studies and animalmodels would allow a better appreciation of theunderlying biological basis of SZ.16 Second, futurestudies need larger cohorts with distinct and betterdefined phenotypes (including intermediate phe-notypes) that would require greater internationalcollaborations. This has the potential to unravelhitherto unknown genetic foci underlying psy-chotic conditions like SZ. Third, to date, there havebeen slightly more than a dozen GWAS on SZ thathave been predominantly conducted on westernpopulations both at the discovery and validationstages of their investigations.7,8,11–14 It is expectedthat some of the genetic causes for SZ may bepopulation-specific. Therefore, by targeting the

Chinese population, our study hopes to providethe much-needed information on the largest popu-lation in the world. Our study is very muchcomplementary to the ongoing efforts in westernpopulations, and it will offer unique opportunitiesfor comparing the results from different popula-tions and for understanding population differencesin the genetic basis of SZ.

In view of the aforementioned issues, in thecurrent study, we will adopt an integrated genomicapproach for studying SZ, and particularly focus onneurocognitive endophenotypes in the process.The development of inheritable diseases includingSZ is likely to involve a series of genetic, molecular,cellular and physiological changes.17 Traditionally,the study of common diseases has been doneby interrogating such changes at one level, suchas association analysis of genetic variations ormicroarray analysis of mRNA expression change.Although such studies have made some progress inunderstanding the genetic and molecular mecha-nism of common diseases, the more powerfulapproach is to interrogate a disease at multiplelevels by integrating evidences from multiplegenome-scale genetic, molecular, cellular andphysiological, and animal model data.16,18 The mostrecent success of such an integrated approach instudying SZ is the identification of oligodendrocytelineage transcription factor 2 and its interactinggenes as potential susceptibility genes of SZthrough crossmatching the genetic association andgene expression.19 By investigating illness-relatedchanges at DNA, RNA, lipid and protein levels inthe same group of patients, we have a great oppor-tunity to pursue an integrated genomic andmolecular study of SZ. This integrated genetic andgenomic approach can be illustrated with methodsadopted in the different inter-related componentswithin the projects of the Singapore Translationaland Clinical Research in Psychosis (STCRP). In thisoverview, we will briefly describe the four compo-nents of this approach as well as outline some ofthe genetic/genomic strategies and analyses thatwill be conducted within the study.

Component 1: comprehensive genetic studyof SZ

In this component, a comprehensive genetic studyof the diagnostic phenotype will be carried outamong the Chinese population in Singapore bycombining GWAS and CNV analyses with a directre-sequencing analysis of important genes andpathways. The GWAS of diagnostic phenotype willbe done in a multistage fashion to maximize the

Translational research in psychosis

92 © 2011 Blackwell Publishing Asia Pty Ltd

power, efficiency and speed of study. At the firststage, a genome-wide scan will be done in the first1000 Chinese case-control pairs from Singapore.The top findings from this analysis (up to 1500single-nucleotide polymorphisms (SNPs) ) of thediagnostic phenotype in the first 1000 Chinese case-control pairs will be replicated in another 1000 case-control pairs. To further improve the power of ourgenome-wide analysis of diagnostic phenotype, weplan to pursue a joint analysis of our genome-widedata combined with the ones from other countriessuch as China, Hong Kong and the United States.

Recognizing the limitation of GWAS in identifyingrare genetic risk factors, we will carry out a large-scale direct sequencing analysis. We will focus on agroup of genes and pathways that are known or sug-gested to be important in the development of SZ. Wewill also conduct an all-exon re-sequencing study toidentify novel genes and pathways with rare variantsthat influence disease risk. Furthermore, CNVanalysis using indirect (array-based) and direct(sequencing level) approaches will also be carriedout.

Component 2: elucidating the genetic structureof neurocognitive intermediate phenotypesin SZ

SZ, as a clinical phenotype, is likely an accumula-tion of several intermediate endophenotypes thatmay also be a variable within the general popula-tion. Carlson et al.20 state that ‘studies using asingle clinical endpoint are akin to a shot at themoon with only one chance for success, whereasstudies that collect multiple phenotypes are morelikely to help us understand the genetic contribu-tion of genetic factors to components of disease’.Thus, to complement our analysis of the clinicaldiagnosis of SZ, we have chosen to focus on neu-rocognitive endophenotypes. Genetic factors con-tributing to these neurocognitive intermediatephenotypes such as working memory, attentionand executive functioning domains may be easierto dissect because of the improved signal-to-noiseratio in the fraction of variance explained by anygiven factor.19 The neurocognitive battery consistsof tests including the Brief Assessment of Cogni-tion in SZ (to assess verbal memory, workingmemory, motor speed, reasoning and problemsolving, attention and processing speed), the Con-tinuous Performance Test (to assess vigilance), theWisconsin Card Sorting Test (to assess executivefunctioning) and the Benton Line Orientation test(to assess visual processing). The genome-wideassociation analysis of a non-neurocognitive inter-

mediate phenotype, namely the schizotypal per-sonality trait (using the Schizotypal PersonalityQuestionnaire), will also be done in 1000 Chinesecontrol samples whereby this phenotype is freefrom the variety of confounding influences inpatients with SZ, such as antipsychotic treatment,chronic disability and psychosis. In addition, theState–Trait Personality Inventory is used to assessboth depressive and anxious state and trait fea-tures in patients and healthy controls. The assess-ment of both subject state and trait features allowsfor the control of possible confounding by thesefactors that may influence neurocognitive perfor-mance. The statistical analysis will be done usingboth standard and novel quantitative trait locus-mapping techniques. The significant findings(SNPs) will then be a subject to a follow-up analysisin the bigger sample of Chinese case-control pairsto evaluate their roles in the development of SZ.

Component 3: genomic component of theLongitudinal Youth at Risk Study

For a better understanding of their clinical impact,genetic risk factors identified by our comprehensivegenetic analysis will be further evaluated by a pro-spective analysis in ultra-high-risk (UHR) individu-als for psychosis, about 20% of whom are expectedto convert to psychosis within 2 years. Throughthis prospective analysis, we will evaluate the pre-dictive power of the identified genetic risk factors,together with other neurocognitive and neuroimag-ing factors, to predict the conversion to psychosis.Furthermore, because multiple neuropsychiatricconditions are expected to develop in this prospec-tive cohort of UHR individuals, such an analysis willalso allow us to understand the spectrum of involve-ment of these genetic risk factors in other neurop-sychiatric disorders.

In addition to the genetic analysis, a longitudinalexpression study will be performed in this compo-nent, aiming to identify gene expression factors thatare up- or downregulated during the developmentof psychosis in the UHR individuals. Access to mul-tiple lines of biological data provides us a uniqueopportunity to pursue integrated genomic analysisof the progress of SZ, which is more powerful than ifthe study were to rely on a single source of informa-tion.21 Along the lines of integrated analysis ofgenetic and expression data, one compelling way isto look for the convergence of the genetic andexpression data at individual genes and gene net-works. This approach may include investigatingthe convergence of multiple lines of evidence at a

K. Sim et al.


particular pathway or mechanism, through aBayesian-like integration of multiple independentlines of evidence.22

Component 4: animal models

Significant findings from the genetic and genomicanalyses of components 1–3 would suggest newgenetic animal models (e.g. knockout or knock-down mice) to create and study functional aspectsof these genes. Other leads can be inferred from thepharmacological actions of drugs. For instance, twoprominent theories of SZ pertain to hyperdopamin-ergic and/or hypoglutamatergic states.23–25 Patientswith SZ show enhanced sensitivity to dopaminereceptor agonists such as amphetamine (AMPH),26

and these drugs can exacerbate psychotic symp-toms.27,28 With respect to hypoglutaminergic state,antagonists of the N-methyl-D-aspartate (NMDA)receptor such as phencyclidine (PCP) can elicit awide range of psychotomimetic effects in humans.29

Another NMDA receptor antagonist, dizocilpine,may induce features akin to those seen in patientswith SZ.23,24 Nevertheless, because SZ is a heteroge-neous syndrome, and there are numerous interac-tions between the dopaminergic and glutamatergicsystems,25 abnormalities in one, both or additionalinteracting systems may contribute to the variouspositive, negative and cognitive symptoms of SZ.Over the years, a number of different animal modelsof SZ-like behaviours have been developed andevaluated at Duke University Medical Center. Theseinclude genetic (dopamine transporter knockout

(DAT-KO) and NR1 subunit NMDA receptor knock-down) and pharmacologic-induced (AMPH, PCP ordizocilpine) mouse models of SZ. The DAT-KO micerecapitulate many of the positive-like symptoms,30,31

while the NR1 knock-down animals display manynegative-like symptoms of SZ.32,33 The AMPHand PCP pharmacologic models complement theDAT-KO and NMDA receptor knock-down geneticmodels, respectively. AMPH- or apomorphine-induced hyperactivity and stereotypy are postulatedto represent psychosis-like behaviours, and thisanimal model has been used extensively as a screenfor drugs with antipsychotic efficacy.33,34 The utilityof the animal model component includes the facili-tation of specific animal behavioural studies (con-ditioned avoidance response, open field, prepulseinhibition, etc.) and the development of transgenicmouse models to aid in the interpretation of inte-grated genomic data from the other three compo-nents. Figure 1 illustrates the interactions betweenthe different components within the study.

Genotyping and statistical analysis

All the genotyping analyses will be performed at theGenome Institute of Singapore, which harbours anexperienced research team and strong technicalplatform for a large-scale genotyping analysis. OnlySNPs with (i) call rates > 98%; (ii) minor allele fre-quencies > 0.01 and (iii) no significant deviationsfrom Hardy–Weinberg equilibrium in the controlswill be retained for further statistical analysis.Samples with call rates < 98% will be dropped from

FIGURE 1. Illustration of the interactions between the different components within the study.

Animal models(Component 4)

Follow-up analysis of top findings insecond 1000 case-control pairs from

Singapore

Joint genome-wide analysis of all the 2000 case-control pairs from Singapore and additional case-

control pairs from China, Hong Kong and USA

Genome-wide analysis in the first 1000Chinese case-control pairs

(Component 1)

Genome-wide QTL analysis of neurocognitivephenotypes in Chinese controls

(Component 2)

Genes implicated by longitudinalexpression analysis (Component 3)

Functional studies includinggene expression, lipidomics

CNV analysis by genome-wide SNP data in 2000 case-controlpairs



further analysis. Relationship between pairs of indi-viduals will be inferred using the programmePLINK. Upon the identification of a first-degreerelative pair, only one of the pairs (with higher callrate) will be used for association analysis. The pres-ence of the population stratification in case andcontrol samples will be evaluated, and any geneticoutliers will be identified and removed from furtherassociation analysis. Genome-wide associationanalysis of diagnostic phenotype will be performedusing the Armitage trend test and corrected forpopulation stratification. In addition to the SNP-based trend test, we will also perform a genome-wide haplotype-based association analysis becausehaplotype analysis may have better power undercertain conditions.35 We will use two approaches forhaplotype associations: (i) sliding windows of fixedsizes implemented in the programme WHAP; and(ii) sliding windows of varying sizes developed byour group.36 Furthermore, we will also explore themulti-marker association analysis that was sug-gested to be more powerful than the single SNP testwhen several SNPs with marginal effects jointly con-tribute to the disease.37

Fine-mapping analysis

Upon the identification of significant associationevidence, fine-mapping analysis will be performedto narrow down the observed association evidenceto a small set of SNPs. The identified set of SNPs willthen be subjected to the following functional analy-sis for identifying the causal risk allele(s) respon-sible for the observed association evidence. Thelinkage-disequilibrium block harbouring the identi-fied association evidence will be sequenced toreveal all the genetic variants evaluated for theirassociation with phenotype.

Candidate gene selection for sequence analysis

A large-scale candidate gene-based re-sequencinganalysis will be performed. Candidate genes will beselected from the following pathways or sources ofinformation: (i) pharmacological pathways such asdopamine and serotonin; (ii) genes whose mRNAand protein expression patterns are shown to beassociated with the development of SZ in our longi-tudinal expression profiling study; (iii) candidategenes implicated by previous genetic associationanalysis; and (iv) genes within the linkage loci sug-gested by previous genome-wide linkage studies. Inaddition, candidate genes or regions implicated byour GWAS will be subjected to direct sequenceanalysis. Our selection of candidate genes will be a

dynamic process, meaning that additional candi-date genes will be added into the study when newbiological and genetic findings become available.

Massively parallel re-sequencing analysis

This analysis will be built on two rapidly evolvingtechnologies. The first one is the development ofhigh-throughput systems for massively parallelsequence analysis. Accompanying this is a break-through in the isolation and/or amplification oftargeted sequences in a massive-multiplexingfashion. Four different approaches have been pro-posed.38,39 While the biochemistry underlying theseprotocols is quite different, the main strategybehind these approaches can be relatively similarin terms of using high-density oligonucleotidemicroarrays or magnetic beads in solution tocapture targeted sequences and subsequentlyamplifying the enriched sequences. This technol-ogy has recently been extended to allow for theefficient capture and enrichment of all exons of thegenome (i.e. the ‘exome’), and the sensitive discov-ery of coding variants.40 Combining the isolation/amplification of targeted sequences with massivelyparallel re-sequencing, a large number of codingsequences in selected candidate genes or regionscan be analyzed for rare variants that influencedisease risk in patients with SZ and with a familyhistory of psychiatric diseases. Moreover, anexome-sequencing study will facilitate a genome-wide screen of all coding genes for an enrichmentof rare variants in disease cases compared withcontrols and thus allow for the discovery of novelgenes and pathways in SZ.

CNV analysis

CNVs are deletions or duplications of DNA rangingfrom 1 kilobase to several megabases in size, and itcan be inherited or caused by de novo mutation.41

One of the most well-known CNVs associated withSZ is the 22q11 deletion that involves several genesincluding catechol-O-methyl transferase, and it isknown to increase the risk of SZ by 20–30 fold.42 Somestudies revealed that cases with SZ had an increasedburden of CNV. Two of the largest CNV studies todate, the International Schizophrenia Consortiumand deCODE genetics, agree that deletions on 1q21and 15q13, though rare and involve less than 1% ofcases, had large effects on the risk of SZ.43,44 Given thefact that the majority of the identified CNV loci areexpected to have a low frequency, two statisticalapproaches can be employed for analyzing the CNVdata. For common CNV loci (frequency > 1%), both

K. Sim et al.


Cochran–Armitage test and the Fisher’s exact testwill be used to compare their frequency differencesbetween the cases and the controls. For rare CNVloci, comparisons are made between the levels ofgenome-wide burden between the cases and thecontrols.The same method has been recently used toreveal the strong association between multiple rarecopy number mutations and autism.45

Functional analysis

All the genetic findings will be subjected to genomicfunctional analysis. These analyses will allow func-tional genetic variants to be distinguished fromlinked DNA polymorphisms by uncovering theirrespective pathophysiological consequences, and itcan further elucidate the molecular mechanismsunderlying the pathophysiology of SZ and cognitiveimpairment. The following strategies for functionalcharacterization are adopted:

1 Gene approach

This approach will focus on identification of func-tional genetic markers by characterizing geneticvariants that affect gene transcriptional regulationand pre-mRNA splicing. Because coding sequencesrepresent ~2% of the total genome, the majority ofgenetic variants identified to date is within non-coding regions, such as promoter region, untrans-lated regions and intronic regions. We will amplifythe promoter regions of candidate genes and willestablish promoter activity assay that will addressfunctional consequences of genetic variants (e.g.genetic variants that affect binding of transcriptionfactors). We will also assess the contributions ofgenetic variants within untranslated regions inrelation to mRNA stability. Moreover, we willapply both ‘forward’ and ‘reverse’ approaches toidentify alternative splicing mutations in candidategenes.

Emerging evidence suggests that the generationof aberrant mRNA transcripts play an importantrole in many human diseases.46,47 Furthermore,alternative splicing transcripts arise not only fromintronic mutations but also from coding synony-mous and non-synonymous mutations.48 Therefore,studying candidate genes at the genomic DNA,mRNA and cDNA levels could expand our knowl-edge of the pathophysiology of human diseasesincluding SZ. The ‘forward’ approach will utilize thegenetic information obtained from Epstein–Barrvirus-transformed lymphoblastoid cell lines. Thesecell lines will allow us to assess gene function and toreveal potential alternative splicing transcriptsthrough conventional and real-time quantitative

real time polymerase chain reaction. Identificationof aberrantly splicing transcripts will lead to identi-fication of potential contributing genetic variants atthe genomic DNA level and subsequent validationusing in vitro systems. The ‘reverse’ approach is animportant and necessary step in case neuronalmRNA information cannot be obtained throughlymphoblastoid cell lines. This approach will utilizethe assembly of partial genomic DNA into ‘mini-gene’ constructs driven by cytomegalovirus pro-moter, thus bypassing the limitations of certaintissue-specific transcriptional regulations. These‘mini-gene’ constructs will be expressed in appro-priate cell lines (neuronal or peripheral cell lines)and will assess alternative splicing transcripts in thepresence of genetic variants of interest as comparedwith its wild-type version. If functional splicinggenetic variants are identified by in vitro cell culturesystems, it will be necessary to confirm the exist-ence of these alternative splicing transcripts inhuman tissues.

2 Proteomic approach

This approach is to apply cellular and biochemicalsystems as well as electrophysiological techniquesto assess protein functions to understand potentialpathological consequences of genetic variants incandidate genes. We will explore protein functionsunder physiological conditions by studying proteinphosphorylation/de-phosphorylation (kinaseassays), protein–protein interactions (immunopre-cipitation, yeast two-hybrid assays and proteinpurification), protein localization and trafficking(imaging assays), as well as channel-gating proper-ties (electrophysiological recording). These assayswill provide a ‘baseline’ for wild-type protein func-tion, and subsequently, the implications of geneticvariants of interest can be analyzed. For example, SZsusceptible gene neuregulin-1 has been recentlyimplicated in glutamatergic function through theneuregulin-1/ErbB4 receptor signalling mecha-nisms.49 Therefore, identification of functionalmutations in NGR1 and subsequent functionalcharacterization with respect to their contributionto glutamatergic neurotransmission will not onlyprovide functional genetic markers for SZ and cog-nitive impairment but also reveal novel signallingmechanisms underlying the pathophysiology ofthese conditions.

3 Translational approach

Transgenic knock-in mice expressing functionalmutations or transgenic knockout mice for novelcandidate genes can be developed. Through the



combination of neurobehavioural, neuropharma-cological and neurochemical approaches, theseanimal models will provide powerful tools to dissectand understand the molecular mechanisms under-lying the pathophysiology of SZ and cognitiveimpairment as well as to develop novel pharmaco-logical treatments for this condition.

Longitudinal expression profiling

A novel approach in this paper is to study the geneexpression longitudinally. An initial comparison willbe made between UHR individuals and healthyindividuals at baseline, seeking to identify earlychanges distinguishing the UHR state. Second, forthe same individual, the gene expression profileswill be compared at three or more time points, thatis, before and after ‘conversion’ and when themental state shows progression. This will allow us toidentify expression patterns that are altered, impor-tantly in a progressive fashion, during the pathogen-esis of SZ. Time-course analysis will also help us toovercome the adverse impact of the well-expectedtemporal fluctuation of expression measurements.We will also attempt to correlate the expression pat-terns with any observed progressive neurocognitiveand neuroimaging parameters. Collecting expres-sion information from DNA, RNA and proteins willallow us to cross-validate the expression results atdifferent levels. This is one of the first longitudinalstudies where information on neurocognition, neu-roimaging and expression profiling will be collectedon the same UHR individuals during the course ofthe conversion from a prodromal status to psycho-sis including SZ, allowing us to investigate on thegenetic and molecular mechanisms underpinningany neurocognitive, neuroimaging aberrations andthe conversion to SZ.

Lipidomic profiling

Lipids have been suggested to be involved in thepathogenesis of SZ, and there has been an increas-ing body of evidence to support this notion. Themembrane phospholipid hypothesis of SZ statesthat alterations in phospholipid metabolism in thebrain, as a direct result of increased rate of loss, willlead to changes in the functioning of membrane-associated proteins and of cell signalling sys-tems.50,51 Therefore, we plan to extract and profileperipheral membrane lipids in the UHR cohort andto elucidate lipid perturbations longitudinally. Stateof the art methods and equipment will be used forcomprehensive analysis of lipid extracts. Quantifi-cation of individual lipid molecular species will then

be performed using multiple reaction monitoringon triple-quadruple instruments. Each individualion-dissociation pathway for close to 450 lipidspecies has already been optimized with regard tocollision energy to minimize variations in relativeion abundance because of differences in rates of dis-sociation. Significant lipid pathway aberrations willbe identified from the lipid profiles and will providepotentially novel candidate genes for analysis.

CONCLUSION

This article has painted in broad-brush strokessome of the details involved in the integratedgenetic and genomic approach in understandingpsychotic disorders such as SZ within the this study.Identifying genes causing SZ has been an arduousprocess related in part to the complex nature of theillness including phenotypic heterogeneity, geneticheterogeneity, gene–gene and gene–environmentinteractions. To address some of these issues, a com-binatorial genomic approach and correlations withdisease and specific psychosis endophenotypessuch as neurocognition will allow a more compre-hensive investigation of the condition. Thisapproach also enables cross-validation betweendata from genetic association, lipidomic and pro-teomic studies as well as findings from examinationof animal models. The integrated genetic andgenomic approach has the potential to generate ahost of candidate genetic signals and biomarkerswith the possibility of further shedding light on thepathophysiological pathways underlying such acrippling illness.

ACKNOWLEDGEMENTS

The Singapore Translational Clinical Research inPsychosis is supported by the National ResearchFoundation Singapore under the National MedicalResearch Council Translational and ClinicalResearch Flagship Programme (Grant No: NMRC/TCR/003/2008).

REFERENCES

1. Heider D, Bernert S, König HH et al. Direct medical mentalhealth care costs of schizophrenia in France, Germany andthe United Kingdom – findings from the European Schizo-phrenia Cohort (EuroSC). Eur Psychiatry 2009; 24: 216–24.

2. Mojtabai R, Fochtmann L, Chang SW, Kotov R, CraigTJ, Bromet E. Unmet need for mental health care in

K. Sim et al.


schizophrenia: an overview of literature and new data froma first-admission study. Schizophr Bull 2009; 35: 679–95.

3. Sim K, Ng BY. Con brioso e rigore! – understanding theneurobiology of schizophrenia spectrum disorders. Ann AcadMed Singapore 2009; 38: 381–2.

4. Sullivan PF, Kendler KS, Neale MC. Schizophrenia as acomplex trait: evidence from a meta-analysis of twin studies.Arch Gen Psychiatry 2003; 60: 1187–92.

5. Lewis CM, Levinson DF, Wise LH et al. Genome scan meta-analysis of schizophrenia and bipolar disorder, part II: schizo-phrenia. Am J Hum Genet 2003; 73: 34–48.

6. Ng MY, Levinson DF, Faraone SV et al. Meta-analysis of 32genome-wide linkage studies of schizophrenia. Mol Psychia-try 2009; 14: 774–85.

7. Shifman S, Johannesson M, Bronstein M et al. Genome-wideassociation identifies a common variant in the reelin genethat increases the risk of schizophrenia only in women. PLoSGenet 2008; 4: e28; doi:10.1371/journal.pgen.0040028.

8. Stefansson H, Ophoff RA, Steinberg S et al. Common variantsconferring risk of schizophrenia. Nature 2009; 460: 744–7.

9. Burmeister M, McInnis MG, Zöllner S. Psychiatric genetics:progress amid controversy. Nat Rev Genet 2008; 9: 527–40.

10. Risch N, Merikangas K. The future of genetic studies ofcomplex human diseases. Science 1996; 273: 1516–17.

11. Lencz T, Morgan TV, Athanasiou M et al. Converging evi-dence for a pseudoautosomal cytokine receptor gene locusin schizophrenia. Mol Psychiatry 2007; 12: 572–80.

12. Kirov G, Zaharieva I, Georgieva L et al. A genome-wide asso-ciation study in 574 schizophrenia trios using DNA pooling.Mol Psychiatry 2009; 14: 796–803.

13. Athanasiu L, Mattingsdal M, Kähler AK et al. Gene variantsassociated with schizophrenia in a Norwegian genome-widestudy are replicated in a large European cohort. J PsychiatrRes 2010; 44: 748–53.

14. Sullivan PF, Lin D, Tzeng JY et al. Genome-wide associationfor schizophrenia in the CATIE study: results of stage 1. MolPsychiatry 2008; 13: 570–84.

15. Grozeva D, Kirov G, Ivanov D et al. Rare copy number vari-ants: a point of rarity in genetic risk for bipolar disorder andschizophrenia. Arch Gen Psychiatry 2010; 67: 318–27.

16. Le-Niculescu H, Balaraman Y, Patel S et al. Towards under-standing the schizophrenia code: an expanded convergentfunctional genomics approach. Am J Med Genet B Neurop-sychiatr Genet 2007; 144B: 129–58.

17. Waddington JL, Corvin AP, Donohoe G, O’Tuathaigh CM,Mitchell KJ, Gill M. Functional genomics and schizophrenia:endophenotypes and mutant models. Psychiatr Clin NorthAm 2007; 30: 365–99.

18. Giallourakis C, Henson C, Reich M, Xie X, Mootha VK. Diseasegene discovery through integrative genomics. Annu RevGenomics Hum Genet 2005; 6: 381–406.

19. Georgieva L, Moskvina V, Peirce T et al. Convergent evidencethat oligodendrocyte lineage transcription factor 2 (OLIG2)and interacting genes influence susceptibility to schizophre-nia. Proc Natl Acad Sci U S A 2006; 103: 12469–74.

20. Carlson CS, Eberle MA, Kruglyak L, Nickerson DA. Mappingcomplex disease loci in whole-genome association studies.Nature 2004; 429: 446–52.

21. Le-Niculescu H, Patel SD, Bhat M et al. Convergent functionalgenomics of genome-wide association data for bipolar dis-order: comprehensive identification of candidate genes,pathways and mechanisms. Am J Med Genet B Neuropsychi-atr Genet 2009; 150B: 155–81.

22. Gupta S, Bisht SS, Kukreti R, Jain S, Brahmachari SK. Booleannetwork analysis of a neurotransmitter signaling pathway. JTheor Biol 2007; 244: 463–9.

23. Ross CA, Margolis RL, Reading SA, Pletnikov M, Coyle JT.Neurobiology of schizophrenia. Neuron 2006; 52: 139–53.

24. Tsai G, Coyle JT. Glutamatergic mechanisms in schizophrenia.Annu Rev Pharmacol Toxicol 2002; 42: 165–79.

25. Carlsson A, Waters N, Holm-Waters S, Tedroff J, Nilsson M,Carlsson ML. Interactions between monoamines, glutamate,and GABA in schizophrenia: new evidence. Annu Rev Phar-macol Toxicol 2001; 41: 237–60.

26. Lieberman JA, Kane JM, Alvir J. Provocative tests with psy-chostimulant drugs in schizophrenia. Psychopharmacology(Berl) 1987; 91: 415–33.

27. Janowsky DS, el-Yousel MK, Davis JM, Sekerke HJ. Provoca-tion of schizophrenic symptoms by intravenous administra-tion of methylphenidate. Arch Gen Psychiatry 1973; 28: 185–91.

28. Abi-Dargham A, Rodenhiser J, Printz D et al. Increased base-line occupancy of D2 receptors by dopamine in schizophre-nia. Proc Natl Acad Sci U S A 2000; 97: 8104–9.

29. Luby ED, Gottlieb JS, Cohen BD, Rosenbaum G, Domino EF.Model psychoses and schizophrenia. Am J Psychiatry 1962;119: 61–7.

30. Pogorelov VM, Rodriguiz RM, Insco ML, Caron MG, WetselWC. Novelty seeking and stereotypic activation of behavior inmice with disruption of the Dat1 gene. Neuropsychopharma-cology 2005; 30: 1818–31.

31. Ralph RJ, Paulus MP, Fumagalli F, Caron MG, Geyer MA.Prepulse inhibition deficits and perseverative motor patternsin dopamine transporter knock-out mice: differential effectsof D1 and D2 receptor antagonists. J Neurosci 2001; 21:305–13.

32. Mohn AR, Gainetdinov RR, Caron MG, Koller BH. Mice withreduced NMDA receptor expression display behaviors relatedto schizophrenia. Cell 1999; 98: 427–36.

33. Duncan GE, Moy SS, Lieberman JA, Koller BH. Effects ofhaloperidol, clozapine, and quetiapine on sensorimotorgating in a genetic model of reduced NMDA receptor func-tion. Psychopharmacology (Berl) 2006; 184: 190–200.

34. Creese I, Iversen SD. The role of forebrain dopamine systemsin amphetamine induced stereotyped behavior in the rat.Psychopharmacologia 1974; 39: 345–57.

35. Schaid DJ. Evaluating associations of haplotypes with traits.Genet Epidemiol 2004; 27: 348–64.

36. Li Y, Sung WK, Liu JJ. Association mapping via regularizedregression analysis of single-nucleotide-polymorphism hap-lotypes in variable-sized sliding windows. Am J Hum Genet2007; 80: 705–15.

37. Schaid DJ, McDonnell SK, Hebbring SJ, Cunningham JM,Thibodeau SN. Nonparametric tests of association of mul-tiple genes with human disease. Am J Hum Genet 2005; 76:780–93.

38. Porreca GJ, Zhang K, Li JB et al. Multiplex amplification oflarge sets of human exons. Nat Methods 2007; 4: 931–6.

39. Mamanova L, Andrews RM, James KD et al. FRT-seq:amplification-free, strand-specific transcriptome sequencing.Nat Methods 2010; 7: 130–2.

40. Ng SB, Buckingham KJ, Lee C et al. Exome sequencing iden-tifies the cause of a Mendelian disorder. Nat Genet 2010; 42:30–5.

41. Redon R, Ishikawa S, Fitch KR et al. Global variation in copynumber in the human genome. Nature 2006; 444: 444–54.

42. Gothelf D, Schaer M, Eliez S. Genes, brain development andpsychiatric phenotypes in velo-cardio-facial syndrome. DevDisabil Res Rev 2008; 14: 59–68.

43. Stefansson H, Rujescu D, Cichon S et al. Large recurrentmicrodeletions associated with schizophrenia. Nature 2008;455: 232–6.

44. International Schizophrenia Consortium. Rare chromosomaldeletions and duplications increase risk of schizophrenia.Nature 2008; 455: 237–41.



45. Sebat J, Lakshmi B, Malhotra D et al. Strong association of denovo copy number mutations with autism. Science 2007;316: 445–9.

46. Johnson JM, Castle J, Garrett-Engele P et al. Genome-widesurvey of human alternative pre-mRNA splicing with exonjunction microarrays. Science 2003; 302: 2141–4.

47. Licatalosi DD, Darnell RB. Splicing regulation in neurologicdisease. Neuron 2006; 52: 93–101.

48. Cartegni L, Chew SL, Krainer AR. Listening to silence andunderstanding nonsense: exonic mutations that affect splic-ing. Nat Rev Genet 2002; 3: 285–98.

49. Li B, Woo RS, Lin M, Malinow R. The neuregulin-1 receptorErbB4 controls glutamatergic synapse maturation and plas-ticity. Neuron 2007; 54: 583–97.

50. du Bois TM, Deng C, Huang XF. Membrane phospholipidcomposition, alterations in neurotransmitter systems andschizophrenia. Prog Neuropsychopharmacol Biol Psychiatry2005; 29: 878–88.

51. Horrobin DF. The membrane phospholipid hypothesis as abiochemical basis for the neurodevelopmental concept ofschizophrenia. Schizophr Res 1998; 30: 193–208.

K. Sim et al.


integrated genetic and genomic approach in the singapore translational and clinical research in...

Documents