Molecular Genetics of Schizophrenia from Approximately Dehydrogenase (PRODH2, 22Q11.21), G72 (13Q34) with the Middle of 2001
Total Page:16
File Type:pdf, Size:1020Kb
Moleculargenetics of schizophrenia:a review of the recent literature DouglasF. Levinson Purpose of review Abbreviations The recent literatureon the moleculargenetics of schizophrenia ASP affected sibling pair is reviewed, to familiarizethe reader with several important LD linkage disequilibrium NIMH National Institute of Mental Health developments as well as a broad range of research efforts in a NMDA N-methyl-o-asparlate rapidly progressing field. SNP single-nucleotidepolymorphism UTR untranslated region Recent findings VCFS velo-cardio facial svndrome New genome scan projects, seen in the light of previousscans, provide support for schizophreniacandidate regions on :a.,2003 LippincottWilliams & Wilkins -7367 chromosomes1q, 2q, 5q, 6p, 6q, 8p, 10p, 13q, 15q and 22q. 0951 Linkage disequilibriummapping studiesof severalof these regions have produced evidence from relativelylarge samples lntroduction supportingthe associationof schizophreniato neuregulin-1 This article will revicw new literature relevant ro rhe (NRG1, 8p21-p12), dysbindin(DfNBPt, 6p22.3),protine molecular genetics of schizophrenia from approximately dehydrogenase (PRODH2, 22q11.21), G72 (13q34) with the middle of 2001. Remarkable progress has been made weaker evidence implicating its interactinggene o-amino acid in the 15 years since serious investigation began in this oxidase (DAAO, 12q24), and catechol-O-methyltransferase field. The past year alone witnessed the publication of (COMT,22q11.21). Other reportshave described including the eight new genome scans [,"-6",7',8"] and a meta- application of microarray techniques to schizophreniapost- analysis of published genome scans [9"], rvirh evidence mortem tissue, candidate gene studies in diverse regions, that gcnome scan data are starting to converge on a sct of efforts to develop quantitativephenotypes (e.g. neuroimaging chromosomal regions; five reporusof substantial evidence and neuropsychologicalvariables) and proposed models of for the associationof schizophrenia with specific genes in schizophreniapathogenesis. positional candidate regions [10"-14"], and rwo reporrs Summary of thc use of microarray technology to screen for genes Schizophrenialinkage findings are beginningto convergeon a showing differential expression in the brains of schizo- number of chromosomalregions. Linkage disequilibrium phrenia patients [15',16"], one of which detectcd a gene mapping studies are beginningto produce findingsof great for which evidence for generic association was also interestin some of these regions,and additionalfindings should observed ll7"l.It appears likely that positional cloning, be expected. Enlarged linkage and associationsamples, microarray and other tcchnologies will soon producc combined with rapidlyevolving technologies, hold out the rcplicable findings that will begin ro elucidate the promise that in the next 5-10 years, the role of some specific pathophysiology of this severe comn-ron disease. schizophreniasusceptibility genes will be confirmedresulting in an initial understandingof the pathogenesisof schizophrenia. Progress in finding genes through positional cloning methods Keywords Positional cloning is the primary srraregyavailable for schizophrenia,genetics, genetic linkage,genetic association, finding susceptibilitygcnes for disorderswith no knor,vn candidate genes pathoprhysiology.The strategyis summarizedin Table 1. CurrOpin Psychlatry 16:157-170. rr 2OO3 Lippincott Williams & Wilkins. Multiply affected families are screencd with a DNA markermap of all chrornosomes(genome scan). Statistical Department of Psychiatry, Universityof PennsylvaniaSchool of Medicine, analyscsconsider whether ill relativcs Philadelphia,PA 19104-3309, USA have inherited the samemarker alleles, indicating the proximity of a disease Correspondenceto DouglasF. Levinson,MD, Departmentof Psychiatry,University of susceptibility gene. For complex disorders (such as PennsylvaniaSchool of Medicine,3535 Market Street,Room 4006, Philadelphia,pA schizophrenia) 19104-3309, USA that probably have multiple interacting Tel:+1 215746 5151;fax: +1 2157465155; e-mail:[email protected],upenn.edu susceptibilitygencs, genome scanscan only locategcnes within 10000 000-30 000 000 basepair (bp) regions. Current Opinion in Psychiatry 2003, 16:157-170 In thcsc candidatc regions, additional markers are gcnotyped 1-? cM apart,which somerinlcsnarrou,s the candidatcregion. DOI: 10. 1 097/01 .vco.000005861 7.61 505.a5 157 158 Schizophrenia Table1. Stagesof positionalcloning studies Stage Clinicalsamole Molecularstrategy Outcome Genome scan Multiplyaffected pedigrees Genotype microsatelliteDNA markers Evidence for linkage in 1O-30 cM at 5000000-10000000 bo candidate region(s) (5-10 cM) spacing,or denser SNP markers Fine-mapping Multiply affected pedigrees Genotype microsatellitemarkers in Evidencefor linkagein 2-10 cM candidate regions aI 1-2 cM spacing candidate region(s) (or denser SNPs) LD mapping Cases versus controls or Genotype markers (primarilySNPs) at Evidence for associationof small sets case-parent trios 10000-50000 bp (10-50 kb) spacing of adjacent markers with disease, implicatingone or two specific genes 'knockout' Functionalgenomics Animalmodels (e.9. Physiologicalstudies to establishthe Plausibleevidence that the gene plays mice), cases, postmorlem tissue role of the gene and protein, effects of a role in disease susceptibility; mutationson physiologyand behavior, detection of other interactinggenes and response to treatmentsfor the and oroteins disease LD, Linkagedisequilibrium; SNP, singlenucleotide polymorphism. Linkagc disecluilibrium (LD) mapping is bascd on a New genome scans differcnt principlc: mosr of the gcnome's DNA Eight new scans and one partial scan \l/erc published scqucnce variations have arisen only oncc or a few this past ycar. Regarding the interpretation of scan 'unrclated' 'gcnome-wide times. Somc ill individuals may havc results, significance' usually refcrs to a inherited from a single ancient ancesror an identical result that is expected to occur by chancc once in 20 'suggcstive DNA scgment that incrcascs rhe risk of diseasc. Each scans, and significance' refers to a result paticnt's complcte DNA sequencc cannot yer be that would occur once per scan on average t261. scquenced cost-effectively, but one can study many However, for complex disorders, statistically significant variations (single-nuclcotide polymorphisms; SNPs) linkage is not reliably obscrved at feasiblc sample 10000-50000 bp (10-50 kb) apart in the candidatc sizes, and the pattern of results across studies may 'unrelate 'haplotypes' 'narrow' rcgion in d' cases ro look for prove to bc more important. In this discussion, (adjaccnt variations on rhe samc chromosome) that are diagnoses rcfcr generally to schizophrenia and schizo- 'broad' 'spectrum' more common in ill individuals. It is believed that affective disorder, and or diagnoses somc (not all) susceptibility variants underlying com- refer to schizophrenia-related psvchoscs and personality non complex disorders can be identified in this way disorders. with larger samples, improved SNP maps and cvolving molecular mcthods. Straub et o/. [1"] published scan results for 270 kish pedigrees. They prcviously reported evidence for Functional studies can rhen bc initiated to determinc linkage on chromosomc 6p2l-24, which approached thc genc's function, inrcractions and possiblc rclation- gcnome-wide significance dcpending on horv one ship to tl-rc disease. corrccts for multiplc tests [27], and suggesrive evidence on 5q21-31[28], Bp22-21t29] and 10p15-p11t301. Gencrally, two gcnome scan strategics have becn Resultssuggested at least two suscepribiliryloci on employed. One is to study srnall samples of familics chromosome6p, both of which have been supportedby with as many ill relatives as tr-rossiblc,in the hopc that other studies(see below), but peaks this closc rogerher one or a few genes are conferring rnost of the risk of are difficult to provc with currcnr methods. This diseasc [18",19"]. The other is to collect many farnilies relatively large samplc was drawn from one ethnic with at least an affccted sibling pair (ASP), assuming that population, which might improve the power to dctect multiple intcrzrcting gcnes each confer a small proportion linkage [31]. But the scanwas completcd someyears ago of risk [20], so that hundreds to thousands of families rvill using threc different, sparse (20-30 cN,,I) maps in bc nccded t?t1. Thc complexitics of this debate are subsamplesof 90 farnilics. A range of diagnostic and beyond thc scopc of this paper. Existing data abotrt genetic models werc analysedto cxplore rhe data fullv, fan-rilial pattcrns of schizophrenia tend to slrpporr rhe btrt this also makes thc findines more difficult ro second vicw [20]. Srnall samples tend to exaggcrarc rhe intcrprct. gcnetic cff-ectsof somc loci whilc missing orhcrs entirely [22,23]. Howevcr, somc of chc bettcr-su1-rportcdlinkzrge Delisi et a/. l2"l published rwo scans.The largest findings in schizophrenia wcre initially detectccl in srnall schizophreniascan to date was of 291 fan-rilies(333 samples124,251. indcpcnclcnt ASPs). Tu,o slrggesrive findings \\rcre Moleculargenetics of schizophreniaLevinson 159 obscrved, on chromosornes 10p and Z (centromcric reports (including some but not all of the above data) region),lr,ith more modest evidenceon 22q, The sccond reported P values < 0.001 in candidate regions of scanwas of 99 families from the Ccntral Vallcy region of clrrcrmosomes1q, Zcl, 8p, 13q and 22q 19"1. (A larger Costa