BRITISH JOURNAL OF PSYCHIATRY (2007), 190, 194^199. doi: 10.1192/bjp.bp.106.025585 SPECIAL ARTICLE AUTHOR’ S P ROOF

Schizophrenia: a common disease caused We argue that current observations from epidemiology and genetics of by multiple rare alleles{{ are consistent with the influ- ence of a large number of individually rare deleterious mutations, many of which have JON M. McCLELLAN, EZRA SUSSER and MARY-CLAIRE KING occurred in the present or recent genera- tions. Several features of schizophrenia sup- port this view: (a)(a)SchizophreniaSchizophrenia is familial, i.e. close rela- tives of affected persons are at increased risk of the illness. Occasional families are very severely affected(Gottesman Summary Schizophrenia is widely held modelling suggested that the observed & Shields, 1982), but most patients to stem fromthefrom the combined effects of decline in recurrence risk of disease with have no close affected relative. Taken increased genetic distance from affected together, these observations are con- multiple common polymorphisms, each individuals is inconsistent with monogenic sistent with a subset of families with a smallimpactsmallimpacton on disease risk.We inheritance of large-effect alleles (Risch, harbouring high-penetrance, recently suggest an alternative view: that 1990). As pointed out by Risch, these mod- occurring alleles predisposing to schizo- schizophrenia is highly heterogeneous els were based on the assumption that the phrenia, with different alleles present in different families. Unless caused by genetically and that manypredisposing illness was genetically homogeneous throughout the population for which the detectable chromosomal alterations (as mutationsmutationsarehighlypenetrantand are highly penetrant and recurrence risks were calculated (Risch, in the case of DISC1DISC1), such alleles individually rare, even specific to single 1990). Increasing evidence suggests that have heretofore been difficult to find cases or families.This‘common disease ^ schizophrenia is genetically heterogeneous because individual families are not sufficiently informative for single- ((FanousFanous & Kendler, 2005). If so, then re- rare alleles’ hypothesis is supported by family linkage analyses. currence risk data are also consistent with recent findings in human genomics and by monogenic inheritance of large-effect alleles (b)(b)PaternalPaternal age is consistently associated allelic and heterogeneity for other in a proportion of people with schizo- with increased risk of schizophrenia complex traits.We review the implications phrenia, with different alleles for different (Brown(Brown et al, 2002;Dalman & Allebeck, ofthis model for discoveryresearch families.families. 2002; Malaspina et al, 2002; Byrne et aletal,, 2003; El-SaadiEl-Saadi et al, 2004; Sipos et aletal,, in schizophrenia. We suggest that the ‘common disease – rare alleles’ model explains many cases of 2004; Tsuchiya et aletal, 2005). Paternal age is also associated with increased Declaration of interest None. schizophrenia.schizophrenia.Ourhypothesisisthatmany Our hypothesis is that many rates of several types of de novodenovo germ-germ- mutations predisposing to schizophrenia are Funding detailed in Acknowledgements. line mutations (Crow, 2003). highly penetrant and individually rare, even Current research in the genetics of schizo- specific to single patients or families. In this (c)(c)TheThe illness is associated with decreased phrenia is guided primarily by the ‘common model, different families harbour different fertility (fertility(Nimgaonkar,Nimgaonkar, 1998; Haukka etet disease – common alleles’ model (Chakra-Chakra-( mutations, either in the same gene or in differ- alal, 2003). If this had been the case over varti, 1999). This model originated from ent , but any one family carries only one long periods, then the frequencies of the hypothesis that the illness results from or two mutations. Many different disease- any ancient common alleles associated the cumulative impact of multiple common associated mutations may occur in the same with schizophrenia would be reduced. An ongoing contribution of new and small-effect, genetic variants, interacting gene.gene. therefore individually rare risk alleles with environmental exposures to exceed a The ‘common disease – common alleles’ could explain the persistence of the biological threshold (Gottesman & Shields, and ‘common disease – rare alleles’ models disorder.disorder. 1982). The ‘common disease – common are not mutually exclusive (Goldstein & alleles’ model for schizophrenia is heuris- Chikhi, 2002). Rare severe mutations may To explain our reasoning, we first tically appealing. The illness is relatively occur in genes that also harbour more describe recent findings in genomics and frequent and is found worldwide. Thus common variants with modest effects on genetics of other complex human disorders. common susceptibility allelesshared across disease risk. However, the two models have Then we show that the results of schizo- populations are plausible. The ‘common distinctly different implications for gene- phrenia research are consistent with the ex- disease – common alleles’ model has also finding strategies. Most current psychiatric istence of multiple individually rare alleles been posited to explain the variable and in- genetic research is designed to identify com- of large effect. Finally, we consider the im- consistent results of linkage studies devoted mon alleles or haplotypes associated with plications of this model for future schizo- to finding genes of large effect responsible with increased risk of disease and shared by phrenia research. for schizophrenia, and the weak associa- large numbers of patients compared with tions of various candidate genes with appropriate controls (Merikangas & Risch, THE schizophrenia. Furthermore, mathematical 2003). If many cases of schizophrenia stem from individually rare large-effect alleles, The dynamics of the human genome are current approaches – even if executed proving more complex than anticipated, {{See invited commentary,pp. 200^203, thisissue perfectly – will fail to identify critical genes. revealing more mechanisms by which

194

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use. SCHIZOPHRENIA: GENETICS

AUTHOR’SAUTHOR’ S PROOFP ROOF

genetic changes may lead to human disease. human population has resulted in many leading to hearing loss (PetitPetit( et aletal, 2001;,2001; Particularly relevant to our argument are new alleles, each individually rare and each Friedman & Griffith, 2003). All mutations the following observations (International specific to one population (or even one are recent and all but one are individually Human Genome Sequencing Consortium, family). Most alleles are of this sort. Thus rare. The one frequent mutation, 30delG 20012001aa,,bb, 2004).,2004). the paradox: most human variation is in connexin 26, is the exception that proves (a)(a)OnlyOnly about 2% of the genome consists ancient and shared; most alleles are recent the rule, in that the same mutation has of -coding genes. There are and individually rare. Given the size of occurred independently numerous times in approximately 20 000 protein-coding the present human population and the rate a mutational hot-spot. genes, far fewer than the 80 000– of occurrence of new mutations, all muta- 100 000 hypothesised a decade ago. tions compatible with life have probably EpilepsyEpilepsy However, the human proteome is enor- already occurred and will occur again. The inherited forms of epilepsy are charac- mously complex. At most genes, vari- However, mutations with deleterious terised by allelic and locus heterogeneity able transcription leads to multiple effects before or during the reproductive transcripts, and thus multiple , years will be less frequently transmitted to ((MeislerMeisler et aletal, 2001). Mutations in any of derived from the same locus but with subsequent generations owing to their several genes involved with neuronal signal- different amino acid sequences. Variable adverse impact on fertility or viability. ling can lead to broadly defined epilepsy. transcription is frequently tissue-specific. Therefore, mutations with large effects on Rare mutations in three different sodium As a result, the consequences of a muta- disease may be disproportionately of recent channel genes lead to one more narrowly tion may also be tissue-specific. origin and individually very rare. defined form of epilepsy (generalised epi- lepsy with febrile seizures plus). (b)(b)Germ-lineGerm-line mutations occur more To the extent that any class of mutation commonly than previously thought. – point mutations, copy number errors or Potentially deleterious new mutations abnormalities of number – Alzheimer’s disease may occur at a rate as high as three occur spontaneously, they appear at similar Alzheimer’s disease illustrates that the per zygote (Eyre-Walker & Keightley, rates in all human populations. All humans ‘common disease – common allele’ and 1999; Crow, 2000). Rates of occur- share the same basic genomic architecture, ‘common disease – rare allele’ models need rence of different classes of de novodenovo including the same genomic regions vulner- not be mutually exclusive. The common ee44 mutations are differently influenced by able to mutations. The incidence of schizo- allele ofalleleof APOEAPOE (apolipoprotein E) is asso- parent of origin and parental age phrenia does not appear to vary ciated with a threefold to fourfold in- (Crow, 2000, 2003). The increased substantially across populations by virtue creased risk in individuals of European mutation rate associated with greater of genetic ancestry. This pattern is consis- descent of developing common, late-onset paternal age is particularly relevant, tent with a disease due to multiple, inde- given that risk of schizophrenia is also Alzheimer’s disease (Bird, 2005). On the pendentpendent de novodenovo mutations that arise in associated with paternal age. other hand, multiple rare mutations in genes many different vulnerable genes and geno- encoding amyloid precursor protein (APP),), (c)(c)EpigeneticEpigenetic alterations – stable changes mic regions. Furthermore, environmental presenilin 1 and 2 (PS1 and PS2) and ubiqui- in that do not depend exposures with mutagenic consequences linlin1( 1 (UBQLN1) are responsible for familial on changes in DNA sequence (JaenischJaenisch( may lead to high rates of new mutations early-onset Alzheimer’s disease. Therefore, & Bird, 2003) – may play an important among exposed individuals. For example, both common modest-effect alleles and rare part in human disease, including psy- environmental factors such as maternal large-effect alleles have a role in Alzheimer’s chiatric disorders. Recent intriguing starvation that are associated with disease observations of possible epigenetic disease. The role of APOE4 is an excellent ((CannonCannon et aletal, 2003) may mediate their example of the ‘common disease – common effects related to development include effects through de novodenovo genetic or epi- phenotypic variability in monozygotic allele’ model. However, the effect of APOE4 genetic mutations. We explore this theme twins and imprinting effects on neuro- on Alzheimer’s disease risk is substantially in more detail below. developmental disorders (FragaFraga( et aletal,, larger than the effect sizes of 2 or less that Complex illnesses are almost univer- 2005; Wong et aletal, 2005).,2005). are typically estimated for schizophrenia sally characterised by allelic heterogeneity susceptibility genes. (multipledifferentmutationsinthesamegene(multiple different mutations in the same gene GENETIC HETEROGENEITY leading to disease) and locus heterogeneity IN HUMAN DISEASE (mutations in multiple different genes leading Inherited predisposition to cancer to the same disease) (Botstein(Botstein & Risch, 2003; In each of the two major genes for inherited The genetic heterogeneity of complex ill- GoldsteinGoldstein et al, 2003). We propose that both breast and ovarian cancer, BRCA1BRCA1 andand nesses is the natural result of the origins are characteristic of schizophrenia. To under- BRCA2BRCA2, more than a thousand different of human genetic variation. The oldest stand the potential implications of genetic pathogenic mutations have been found human alleles originated in Africa millions heterogeneity for schizophrenia, we briefly (Walsh(Walsh et aletal, 2006). Large genomic rearran- of years before people first migrated out consider other complex disorders for which gements account for about 10% of these of Africa some 50 000 years ago (Cavalli- genes have been identified. mutations. All inherited BRCA1BRCA1 andand SforzaSforza et aletal, 1994). These ancient variants BRCA2BRCA2 mutations are individually rare. are found in all human populations, are Both locus and allelic heterogeneity are also the most common worldwide, and account DeafnessDeafness characteristic of inherited colon cancer and for approximately 95% of human varia- To date, nearly a hundred genes have been the rarer cancer syndromes (Vogelstein & tion. Yet the exponential growth of the identified that harbour inherited mutations Kinzler, 2004).

195

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use. McCLELLAN ET AL

AUTHOR’ S P ROOF

Lipid metabolic pathways heterogeneity but not to locus heterogene- Candidate genes have been suggested in Rare variants in genes related to lipid meta- ity. For each of these designs, hundreds or several of these regions, including dysbin- bolism are associated with low levels of high- thousands of rigorously diagnosed cases of din on 6p22, neuregulin on 8p22, G72G72 onon density lipoprotein cholesterol (HDL–C) unrelated affected and unaffected individ- 13q34,13q34, COMTCOMT on 22q11,on22q11, RGS4RGS4 on 1q21on1q21 (Cohen(Cohen et aletal, 2004) and low-density lipo- uals are examined. If a substantial portion andand GRM3GRM3 on 7q21 (see reviews by Blouin protein cholesterol (LDL–C) (Cohen et aletal,, of schizophrenia stems from different indi- et aletal, 1998; Harrison & Weinberger, 2005). 2006). Although each variant is individu- vidually rare alleles, increasing the number Each of these genes is biologically plausible ally rare, in total these variants are found of cases also increases the number of differ- (Owen(Owen et aletal, 2004). However, for each in a substantial portion of individuals at ent disease risk mutations represented candidate gene, both positive and negative the far end of the spectrum in terms of among them. As a result, increasing sample associations have been reported with the levels of HDL–C or LDL–C respectively. size does not confer a corresponding in- same SNPs; strengths of effects are gener- These examples illustrate two ways in crease in statistical power. In the most ex- ally weak; the specific allele or haplotype which mutations of large effect are import- treme scenario, in which every case results associated with the illness varies across stu- ant for understanding human disease. First, from a different mutation, an increase in dies; and definitive causative mutations the collective effect of individually rare mu- sample size would not lead to any increase have not been identified. tations in the same gene may explain a con- in the power to detect any one mutation. To illustrate the implications of the siderable proportion of an illness. Second, Consequently, even very large studies may common allele v.v. rare allele models, we will rare mutations in genes of large effect can fail to detect individually rare disease risk review two promising susceptibility genes, reveal pathways critical to disease develop- mutations.mutations. dysbindin (dysbindin(DTNBP1DTNBP1) and)and DISC1DISC1. Very dif-dif-.Very ment.ment. In addition, genetic analyses that focus ferent study designs revealed these genes, only on single nucleotide polymorphisms with correspondingly different results to (SNPs), either individually or in haplotypes, date. Linkage, association and functional IMPLICATIONS OF GENETIC rather than fully sequenced DNA, will in- studies all support some role for dysbindin HETEROGENEITY FOR GENE evitably miss rare disease alleles and thus in schizophrenia. Several studies involving DISCOVERY fail to detect critical genes harbouring such different populations have found positive alleles. Association and linkage studies gen- associations of dysbindin alleles or haplo- All complex illnesses evaluated thus far are erally assume that individuals sharing the types with schizophrenia (StraubStraub( et aletal,, characterised by locus and allelic hetero- same SNP-defined haplotype share the en- 2002;2002;SchwabSchwab et aletal, 2003; Funke et aletal,, geneity. Disease genes for these illnesses tire region, including any hypothetical em- 2004;2004;KirovKirov et aletal, 2004; Kohn et aletal, 2004;,2004; have been identified primarily by positional bedded disease alleles. This assumption is Numakawa et aletal, 2004; Williams et aletal,, cloning in large kindreds. Although subse- reasonable for ancient alleles and nearly al- 2004; Bray et aletal, 2005; Gornick et aletal,, quent association studies confirmed their ways true for related individuals for whom 2005). Dysbindin is widely expressed in role, the original gene discoveries were de- there is direct inheritance of the haplotype. brain, and appears to play a part in cogni- pendent upon individual highly informative However, this assumption is not reasonable tive functioning and capacity (Owen et aletal,, families. Such large informative kindreds for a study of unrelated individuals who car- 2004). Post-mortem studies suggest that are extremely rare in schizophrenia. Link- ry disease alleles of recent origin. Rare recent brain levels of dysbindin may be reduced age studies of schizophrenia based on single mutations causing schizophrenia within the in individuals with schizophrenia (Weickert gene models have not been successful at same haplotype will differ among unrelated et aletal, 2004). However, no variant of dys- identifying causal mutations (Owen et aletal,, individuals, diluting any association. bindin has been specifically linked to 2004). Because pedigrees with schizo- schizophrenia. Across different studies, the phrenia have not been large enough to be risk conferred by any dysbindin variant is individually informative, studies generally TWO CANDIDATE GENES small, with effect sizes of about 2.0 or less. pool data from different families. If many FOR SCHIZOPHRENIA: Among positive association studies, the spe- different genes were responsible for the DYSBINDIN AND DISC1DISC1 cific alleles associated with the disease dif- illness in different families, pooling results fer. Moreover, an allele may be associated across families would preclude The familial nature of schizophrenia is well with increased disease risk in some studies identification of any of them. established. Large collaborative linkage and decreased risk in others (Owen et aletal,, Currently, most gene-discovery strategies studies have suggested multiple candidate 2004). In general, the variants of interest for schizophrenia research – case–parent chromosomal regions that may harbour (defined by SNPs) are common and without triad studies, candidate gene studies and genes associated with the illness. Regions known functional significance. An excep- haplotype association studies – are designed best supported by genome-wide scans in- tion is SNP rs1047631, which has been as- to identify alleles or haplotypes that appear clude 6p22–p24 (Straub et aletal, 1995),,1995), sociated with differences in the expression more frequently among affected individuals 1q21–q22 (BrustowiczBrustowicz( et aletal, 2000) and of dysbindin in brain (Funke et aletal, 2004).,2004). than among appropriate controls (Cannon 13q32–q34 (BlouinBlouin( et aletal, 1998). Other However, the frequency of the haplotype et aletal, 2003). These designs are not robust regions with positive linkage findings with this SNP was similar between cases to either allelic heterogeneity or locus include 1q42, 5q21–q33, 6q21–q25, (45.6%) and controls (40.4%). Thus far, heterogeneity. Sib-pair linkage analyses 8p21–p22, 10p15–p11 and 22q11–q12 resequencing efforts have not revealed any are designed to detect genomic regions ((OwenOwen et aletal, 2004). These regions com- coding sequence mutations in dysbindin consistently shared by affected siblings. bined represent a substantial portion of among individuals with schizophrenia Sib-pair analyses are robust to allelic the genome. ((LiaoLiao & Chen, 2004).

196

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use. SCHIZOPHRENIA: GENETICS

AUTHOR’ S P ROOF

Therefore, at present the evidence sup- (phosphodiesterase 4B), which interacts withwith them. It has been recognised for decades porting dysbindin is mixed. Variable asso- DISC1DISC1 in the neuronal cyclic adenosine that such cases would include any large ciations with different alleles have been monophosphase (AMP) pathway, was dis- kindred with a number of well-diagnosed attributed to allelic heterogeneity; yet alle- rupted by a balanced translocation in two individuals or cases with identifiable geno- lic heterogeneity refers to different disease- related individuals with chronic psychotic mic events of recent origin (e.g. balanced causing mutations in the same gene, not to illnesses (Millar(Millar et al, 2005). Finally, mice translocations). The emerging story of the same allele reducing risk in some cases with a deletion variant that disrupts the DISC1DISC1 highlights this strategy, since the and increasing risk in others. There are at DISC1 protein have deficits gene was originally identified by a balanced least three possible interpretations of these (Koike(Koike et al, 2006). These findings suggest a translocation on chromosome 1q42 co- data. The most favoured in the literature role of rare large-effect mutations of DISC1DISC1 inherited with schizophrenia (MillarMillar( et aletal,, is that dysbindin variants mediate disease and of genes involved in DISC1 pathways in 1998). Similar promising findings have risk as part of a complex interaction with the development of schizophrenia. been noted for specific genes in the 22q11 other genes and environmental factors. This Not surprisingly, DISC1DISC1 became the region (region(MaynardMaynard et aletal, 2002), stemming in is possible, in principle, although difficult subject of association studies. Haplotypes part from the recognised association between to test given the challenge of establishing of DISC1 were associated with schizo- deletions at 22q11 (i.e. the velo-cardio-facial the role of a mediating factor of small effect phrenia and other mental illness in European syndrome) and schizophrenia. on a complex illness of unknown cause. and North American populations (Ekelund Current genomic technology now en- A second possibility is that relatively et aletal, 2001;,2001; HennahHennah et aletal, 2003; Hodgkin- ables the identification of an increasingly rare, as yet unidentified variants in the sonson et aletal, 2004; Callicott et aletal, 2005; Can- large number of classes of mutations. Here- dysbindin locus are embedded in illness- nonnon et aletal, 2005; Hamshere et aletal, 2005),2005) tofore, identifiable genomic events have associated haplotypes in some, but not all, but not in Japanese or Scottish populations been limited to chromosomal abnormalities cases (including those potentially located (Devon(Devon et aletal, 2001; Kockelkorn et aletal, 2004;,2004; such as translocations or deletions. How- in non-coding regulatory regions). Such ZhangZhang et aletal, 2005). Within populations ever, as the resolution of genome-wide alleles could have substantial effects on with positive associations, DISC1DISC1 haplo-haplo- mutation screening technologies improves, the phenotype, but would be masked by types were also associated with putative smaller genomic events in informative cases studying only the common haplotype. The endophenotypes, including neuroanatomical or families can be detected. Once identified, third possibility is that many, most, or all and/or neurocognitive profiles (Hodgkinson a gene altered by a single chromosomal of the various positive associations with et aletal, 2004; Burdick et aletal, 2005; Cannon event becomes a candidate to be screened dysbindin are false positives. The number et aletal, 2005). However, disease-risk haplo- for other (typically smaller) mutations in of different positive association studies (al- types vary across populations and effect other cases. Advanced resequencing tech- beit with different variants) is taken as pri-pri- sizes are small. Therefore, although there nology allows for more rapid identification ma faciemafacie evidence that the gene must be is compelling evidence that rare large-effect of mutations in candidate genes. The occur- involved with the disorder. However, dys- mutations in DISC1DISC1 are associated with rence of multiple deleterious mutations bindin, like many genes involved with brain schizophrenia, it is not clear whetherwhether among unrelated cases provides both biolo- development, is large (44140 kb). The dys- common polymorphisms play a part. gical evidence and epidemiological support bindin locus includes at least 363 SNPs for the causal role of the gene, using gene- (Hinrichs(Hinrichs et aletal, 2006), from which various IMPLICATIONS based hypothesis testing strategies (Chen candidates are selected for association stu- FOR GENETIC RESEARCH et aletal, 2006).,2006). dies. Incorporating linkage disequilibrium IN SCHIZOPHRENIA Individuals who develop schizophrenia across the locus, many thousands of SNP following a known environmental exposure and haplotype combinations appear in dif- The ‘common disease – rare allele’ model are also potentially informative. Such expo- ferent populations. The potential for false has important implications for gene-finding sures may focus gene discovery in two positives is enormous. Unless negative and strategies. A current mantra in schizo- ways. First, genomic approaches (e.g. positive studies were published with equal phrenia genetic research is the need for resequencing efforts) can focus on candi- frequency, this possibility is also nearly ever-larger sample sizes in order to obtain date genetic pathways relevant to the sus- impossible to test. adequate statistical power to detect com- pected exposure, screening for otherwise In contrast, DISC1DISC1 and its associated mon small-effect variants (Devon et aletal,, benign variants that are deleterious given non-coding antisense RNA DISC2DISC2 werewere 2001). These designs are dependent upon the exposure. For example, associations be- originally identified by a balanced translo- the existence of disease-risk alleles that tween schizophrenia and in uteroinutero exposureexposure cation involving chromosome 1q42 which are shared across large numbers of un- to maternal starvation (SusserSusser( et aletal, 1996;,1996; segregated with schizophrenia (and other related cases. These strategies will be inade- St ClairStClair et aletal, 2005), and associations be- major psychiatric disorders) over four gen- quate if schizophrenia in large part stems tween schizophrenia and genes in the folate erations in a large Scottish kindred (St Clair from individually rare disease-risk muta- metabolic pathway (Lewis et aletal, 2005;,2005; et aletal, 1990; Millar et aletal, 1998, 2000). Sachs etet tions in a large number of different genes. Picker & Coyle, 2005) suggest that alal (2005) found that a frameshift mutation We propose an alternative strategy in mutations in genes in the folate metabolism that abnormally truncates DISC1 co-segre-co-segre- selecting cases for study. Rare cases of network could be linked to the illness. gated with schizophrenia in three siblings. schizophrenia with mutations that can be Second, the mutagenic effects of the envir- However, this mutation is also rarely found individually detected using current genomic onmental exposure can be evaluated. in controls with unknown diagnostic status technologies are extremely valuable. It is Following the same example, gestational (Green(Green et aletal, 2006). The gene PDE4B worthwhile devoting resources to finding folate deficiency may be mutagenic, in that

197

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use. McCLELLAN ET AL

AUTHOR’ S P ROOF

it leads to an increase in the rate of muta- the Stanley Medical Research Foundation to J.M.M., Crow,J.F.(2000)Crow, J. F. (2000) The origins, patterns and implications tion genome-wide (McClellan et aletal, 2006).,2006). the Lieber Center for Schizophrenia Research to of human spontaneous mutation. Nature Reviews. GeneticsGenetics,, 11,40^47., 40^47. Among such cases, genome-wide muta- E.S. and a National Association for Research on Schizophrenia and Affective Disorders Distinguished tional screening may detectdetect de novodenovo mu-mu- Crow, J. F. (2003) Development: there’s something Investigator Award to M.-C.K. curious about paternal-age effects. Science,, 301301,, tations. If disease-associateddisease-associated alleles are 606^607. identified, other mutations within the same REFERENCES Dalman, C. & Allebeck, P. (2002) Paternal age and gene may confer some risk for other indi- schizophrenia: further support for an association. viduals with the same exposure. Further- American Journal of Psychiatry,, 159,1591^1592., 1591^1592. Bird, T. D. (2005) Genetic factors in Alzheimer’s more, severe mutations in the same gene disease.disease. New England Journal of Medicine,, 352, 862^864. Devon, R. S., Anderson, S.,Teague, P.W., et aletal (2001) may lead to the disorder without the Identification of polymorphisms within Disrupted in Blouin, J. L., Dombroski, B. A., Nath, S. K., et aletal Schizophrenia 1 and Disrupted in Schizophrenia 2, and exposure.exposure. (19 9 8)8)(19 Schizophrenia susceptibility loci on an investigation of their association with schizophrenia As informative cases are identified, 13 q 32 and 8 p 21. Nature Genetics,, 2020, 70^73. and bipolar affective disorder. Psychiatric Genetics,, 11,, 71^78. genomic technologies are needed for effi- Botstein, D. & Risch, N. (2003) Discovering genotypes cient screening for potential disease-asso- underlying human phenotypes: past successes for Ekelund, J., Hovatta, I., Parker, A., et aletal (2001) ciated mutations. Effective transcriptome- Mendelian disease, future approaches for complex loci in Finnish schizophrenia families. disease.disease. Nature Genetics,, 33, 228^237.,228^237. Human Molecular Genetics,, 10,1611^1617.,1611^1617. and proteome-based tools are needed to characterise such variants. These technolo- Bray,N.J.,Preece,A.,Williams,N.M.,Bray, N. J., Preece, A.,Williams, N. M., et aletal (2005)(2005) El-Saadi,El-Saadi,O., O., Pedersen, C. B., McNeil,T. F., et aletal Haplotypes at the dystrobrevin binding protein 1 (2004) Paternal and maternal age as risk factors for gies are under rapid development. It is ((DTNBP1DTNBP1) gene locus mediate risk for schizophrenia psychosis: findings from Denmark, Sweden and already possible to detect deletions, through reduced DTNBP1expression. Human Molecular Australia.Australia. Schizophrenia Research,, 6767, 227^236.,227^236. GeneticsGenetics,, 1414,1947^1954. duplications and other chromosomal aber- Eyre-Walker, A. & Keightley, P.P.D. D. (1999) HighHigh rations of multiple kilobases anywhere in Brown, A. S., Schaefer,C. A.,Wyatt, R. J., et aletal genomic deleterious mutation rates in hominids. Nature,, the genome, and the sensitivity of these (2002) Paternal age and risk of schizophrenia in adult 397, 344^347.,344^347. offspring.offspring. American Journal of Psychiatry,, 159159,1528^1533. methods for detecting smaller mutations is Fanous, A. H. & Kendler, K. S. (2005) Genetic improving (Sebat et aletal, 2004; Sharp et aletal,, Brustowicz, L. M., Hodgkinson, K. A., Chow, E.W., etet heterogeneity, modifier genes, and quantitative alal (2000)(2000) Location of a major susceptibility locus for phenotypes in psychiatric illness: searching for a 2005). Resequencing tools are increasingly familial schizophrenia on chromosomechromosome1q21-q22. 1q21-q22. Science,, framework. Molecular Psychiatry,, 10,6^13., 6^13. efficient, so large numbers can be screened 288, 678^682.,678^682. Fraga, M. F., Ballestar, E., Paz, M. F., et aletal (2005) for rare events in candidate genes first iden- Burdick, K. E., Hodgkinson, C. A., Szeszko, P. R., etet Epigenetic differences arise during the lifetime of tified by rare, individually detectable alal (2005)(2005) DISC1DISC1 and neurocognitive function in monozygotic twins. Proceedings of the National Academy events. User-friendly bioinformatics re- schizophrenia. Neuroreport,, 1616, 1399^1402. of Sciences of the United States of America,, 102102,10604^,10604^ 10609. sources now exist to help characterise the Byrne, M., Agerbo, E., Ewald, H., et aletal (2003)(2003) structure and function of potential candi- Parental age and risk of schizophrenia: a case-control Friedman, T. B. & Griffith, A. J. (2003) Human study. Archives of General Psychiatry,, 60, 673^678.,673^678. nonsyndromic sensorineural deafness. Annual Review of date genes. It is increasingly possible to Genomics and Human Genetics,, 44, 341^402.,341^402. characterise not only mutations in protein Callicott, J. H., Straub, R. E., Pezawas, L., et aletal Funke, B., Finn, C. T., Plocik, A. M., et al (2004) coding sequence but also mutations and (2005) Variation in DISC1 affects hippocampal structure and function and increases risk for schizophrenia. Association of the DTNBP1DTNBP1 locus with schizophrenia in a epigenetic changes in regulatory regions, Proceedings of the National Academy of Sciences of the U.S. population. American Journal of Human Genetics,, 75,, in non-coding RNA and in transposable United States of America,, 102102, 8627^8632. 891^898.891^898. Goldstein, D. B. & Chikhi, L. (2002) Human elements.elements. Cannon, M., Kendell, R., Susser, E., et aletal (2003)(2003) migrations and population structure: what we know and To summarise, we propose that indivi- Prenatal and perinatal risk factors for schizophrenia. In why it matters. Annual Review of Genomics and Human The Epidemiology of Schizophrenia (eds R. Murray, P. dually rare alleles with large effect, many GeneticsGenetics,, 33,129^152,129^152 Jones, E. Susser, et al), pp. 74^100.Cambridge74^100. Cambridge of which are recent in origin, have a sub- University Press. Goldstein, D. B., Cavalleri, G. L. & Ahmadi, K. R. stantial role in causing schizophrenia. Cur- (2003)(2003) The genetics of common diseases: 10 million Cannon,T. D., Hennah,W., van Erp,T.Erp,T.G., G., et aletal (2005)(2005) times as hard. Cold Spring Harbor Symposia on rent research designs that focus on Association of DISC1/TRAX haplotypes with Quantitative Biology,, 6868,,395^402. 395^402. collecting large samples of unrelated indi- schizophrenia, reduced prefrontal gray matter, and viduals for analysis of shared alleles or hap- impaired short- and long-term memory. Archives of Gornick,Gornick,M.C.,Addington,A.M.,Sporn,A., M. C., Addington, A. M., Sporn, A., et al General Psychiatry,, 6262,1205^1213.,1205^1213. (2005) Dysbindin (DTNBP1DTNBP1, 6p22.3) is associated with lotypes are not suitable for detecting such childhood-onset psychosis and endophenotypes disease alleles. In contrast, rare disease Cavalli-Sforza, L. L., Menozzi, P. & Piazza, A. (1994) measured by the Premorbid adjustment scale (PAS). mutations may be revealed by studies of History and Geography of Human Genes.Princeton Journal of and Developmental Disorders,, 1010,1^8. University Press. individuals and families that harbour in- Gottesman, I. I. & Shields, J. (1982) Schizophrenia:The formative genomic events, and by studies Chakravarti, A. (1999) Population genetics ^ making Epigenetic Puzzle.Cambridge. Cambridge University Press. sense outofsequence.outof sequence.Nature Genetics,, 2121 (suppl.),56^60. of exposed cohorts. A gene harbouring Green, E. K., Norton, N., Peirce,T.Peirce, T. et aletal (2006)(2006) one mutation predisposing to schizophrenia Chen, B. E., Sakoda, L. C., Hsing, A.W. et aletal (2006) Evidence that a DISC1 frame-shift deletion associated Resampling-based multiple hypothesis testing with psychosis in a single family may not be a pathogenic is likely to harbour more than one, with fre- procedures for genetic case-control association studies. mutation. Molecular Psychiatry,, 11, 798^799.,798^799. quencies ranging from relatively common Genetic Epidemiology,, 3030, 495^507.,495^507. Hamshere, M. L., Bennett, P.,Williams, N. et aletal to rare, and effects ranging from modest Cohen, J. C., Kiss, R. S., Pertsemlidis, A., et aletal (2004)(2004) (2005) Genomewide linkage scan in schizoaffective to severe. To see one is not to see them all. Multiple rare alleles contribute to low plasma levels of disorder: significant evidence for linkage at 1q42 close to HDL cholesterol. Science,, 305305, 869^872.,869^872. DISC1, and suggestive evidence at 22q11and 19p13. Archives of General Psychiatry,, 6262,1081^1088. ACKNOWLEDGEMENTS Cohen, J. C., Pertsemlidis, A., Fahmi, S., et aletal (2006)(2006) Multiple rare variants in NPC1L1associated with reduced Harrison, P.J. & Weinberger, D. R. (2005) sterol absorption and plasma low-density lipoprotein Schizophrenia genes, gene expression, and Support for this work was provided by grants from levels. Proceedings of the National Academy of Sciences of neuropathology: on the matter of their convergence. the National Institutes of Health (MH01120) and the United States of America,, 103103,1810^1815. Molecular Psychiatry,, 10, 40^68.,40^68.

198

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use. SCHIZOPHRENIA: GENETICS

AUTHOR’ S P ROOF

Haukka, J., Suvisaari, J. & Lonnqvist, J. (2003) Fertility of patients with schizophrenia, their siblings, and JONJON M. McMcCLELLANCLELLAN,, MD,Department of Psychiatry,University of Washington, Seattle;EZRA SUSSER,SUSSER,MD, MD, the general population: a cohort study from 1950 to DrPH, Department of Epidemiology,MailmanEpidemiology, Mailman School of Public Health, Department of Psychiatry,Columbia 1959 in Finland. American Journal of Psychiatry,, 160160,, University, and New York State Psychiatric Institute, New York;MARY-CLAIRE KING,KING,PhD,Departmentof PhD,Department of 460^463. Genome Sciences and Department of Medicine (Medical Genetics),University of Washington, Seattle, Hennah,W.,Varilo,Hennah, W.,Varilo, T., Kestila, M., et aletal (2003)(2003) Washington,Wa shington, USA Haplotype transmission analysis provides evidence of association for DISC1 to schizophrenia and suggests Correspondence: Dr Jon McClellan,Department of Psychiatry,Box 356560,University of Washington, sex-dependent effects. Human Molecular Genetics,, 12,, Seattle,WA 98195,USA.Email:98195,USA. Email: drjack@@u.washington.edu 31513151^3159. ^3159.

Hinrichs, A. S., Karolchik, D., Baertsch, R., et aletal (First received 21April 2006, final revision 18 August 2006, accepted 4 October 2006) (2006)(2006) TheUCSCGenomeBrowserDatabase:updateThe UCSC Genome Browser Database: update 2006. Nucleic Acids Research,, 3434, D590^598.

Hodgkinson, C. A., Goldman, D., Jaeger, J., et aletal (2004)(2004) Disrupted in schizophrenia 1 (DISC1) association Meisler, M. H., Kearney, J., Ottman, R., et aletal (2001) Sharp, A. J., Locke, D. P., McGrath, S. D., et aletal (2005) with schizophrenia, schizoaffective disorder, and bipolar Identification of epilepsy genes in human and mouse. Segmental duplications and copy-number variation in disorder. American Journal of Human Genetics,, 7575,, Annual Review of Genetics,, 3535, 567-588.,567-588. the human genome. American Journal of Human Genetics,, 862^872.862^872. 7777,78^88. International Human Genome Sequencing Merikangas, K. R. & Risch, N. (2003) Will the Consortium (2001aa)) Initial sequencing and analysis of genomics revolution revolutionize psychiatry? American Sipos, A., Rasmussen, F., Harrison, G., et aletal (2004) the human genome. Nature,, 409409,860^921., 860^921. Journal of Psychiatry,, 160160, 625^635.,625^635. Paternal age and schizophrenia: a population based cohort study. BMJBMJ,, 329,1070.,1070. International Human Genome Sequencing Millar, J. K., Brown, J., Maule, J. C., et aletal (19 9 8)8)(19 AA Consortium (2001bb)) A physical map of the human long-range restriction map across 3 Mb of the St Clair, D., Blackwood, D., Muir,W., et al,(1990),(1990) genome. Nature,, 409409,,934^941. 934^941. chromosome 1111breakpoint breakpoint region of a translocation Association within a family of a balanced autosomal linked to schizophrenia: localization of the breakpoint translocation with major mentalillness. Lancet,, 336336,13^16. International Human Genome Sequencing and the search for neighbouring genes. PsychiatricPsychiatric Consortium (2004) Finishing the euchromatic Genetics,, 88,175^181.,175^181. St Clair, D., Xu, M.,Wang, P., et al (2005)(2005) Rates ofofRates sequence of the human genome. Nature,, 431431,931^945., 931^945. adult schizophrenia following prenatal exposure to the Millar J. K.,Wilson-Annan, J. C., Anderson S., et aletal Jaenisch, R. & Bird, A. (2003) Epigenetic regulation of Chinese famine of 1959^1961. JAMAJAMA,, 294294, 557^562.,557^562. (2000)(2000) Disruption of two novel genes by a translocation gene expression: how the genome integrates intrinsic co-segregating with schizophrenia. Human Molecular Straub, R. E., MacLean, C. J., O’Neill, F. A., et aletal and environmental signals. Nature Genetics,, 33 (suppl.), Genetics,, 99,1415^1423. (19 95) A potential vulnerability locus for schizophrenia 245^254.245^254. on chromosome 6p24-22: evidence for genetic Kirov, G., Ivanov, D., Williams, N. M., et aletal (2004)(2004) Millar, J. K., Pickard, B. S., Mackie, S., et aletal (2005)(2005) heterogeneity. Nature Genetics,, 1111, 287^293. Strong evidence for association between the DISC1 andand PDE4BPDE4B are interacting genetic factors in dystrobrevin binding protein 1 gene (DTNBP1)and schizophrenia that regulate cAMP signaling. Science,, 310310,, Straub, R. E., Jiang,Y., MacLean, C. J., et aletal (2002)(2002) schizophrenia in 488 parent-offspring trios from 1187^1191.118 7 ^ 1191. Genetic variation in the 6p22.3 gene DTNBP1,the,the Bulgaria. Biological Psychiatry,, 55, 971^975. human ortholog of the mouse dysbindin gene, is Nimgaonkar,V.L. (1998) Reduced fertility in associated with schizophrenia. American Journal of Kockelkorn,Kockelkorn,T.T., T.T., Arai,Arai,M.,Matsumoto,H., M., Matsumoto, H., et aletal schizophrenia: here to stay? Acta Psychiatrica Human Genetics,, 7171, 337^348. (2004)(2004) Association study of polymorphisms in the 5’ Scandinavica,, 9898, 348^353.,348^353. upstream region of human DISC1 gene with Susser, E., Neugebauer, R., Hoek, H.W., et aletal (19 9 6)6)(19 Numakawa,T.,Yagasaki,Y.,Ishimoto,T., et aletal (2004)(2004) schizophrenia. Neuroscience Letters,, 368368, 41^45.,41^45. Schizophrenia after prenatal famine. Further evidence. Evidence of novel neuronal functions of dysbindin, a Archives of General Psychiatry,, 53,25^31.,25^31. Kohn,Y., Danilovich, E., Filon, D., et aletal (2004)(2004) Linkage susceptibility gene for schizophrenia. Human Molecular disequilibrium in the DTNBP1 (dysbindin) gene region Genetics,, 13, 2699^2708. Tsuchiya,Tsuchiya,K.J.,Takagai,S.,Kawai,M., K. J., Takagai, S., Kawai, M., et aletal (2005)(2005) and on chromosome 1p36 among psychotic patients Advanced paternal age associated with an elevated risk Owen, M. J.,Williams, N. M. & O’Donovan, M. C. from a genetic isolate in Israel: findings from identity by for schizophrenia in offspring in a Japanese population. (2004)(2004) The molecular genetics of schizophrenia new descent haplotype sharing analysis. American Journal of Schizophrenia Research,, 76,,337^342. 337^342. Medical Genetics. Part B, Neuropsychiatric Genetics,, 128,, findings promise new insights. Molecular Psychiatry,, 99,, 65^70. 14^27.14^ 27. Vogelstein, B. & Kinzler, K.W. (2004) Cancer genes and the pathways they control. Nature Medicine,, ,, Koike, H., Arguello, P. A., Kvajo, M., et aletal (2006)(2006) Petit,C., Levilliers, J. & Hardelin, J. P. (2001) 1010 Disc1is mutated in thethe129S6/SvEv 129S6/SvEv strain and modulates Molecular genetics of hearing loss. Annual Review of 789^799.789^799. working memory in mice. Proceedings of the National Genetics,, , 589^646. 3535 Walsh, T., Casadei, S., Coats, K. H., et aletal (2006) Academy of Sciences of the United States of America,, 103103,, Spectrum of mutations in BRCA1BRCA1,, BRCA2BRCA2,,CHEK2,and 3693^3697.3693^3697. Picker, J. D. & Coyle, J. T. (2005) Do maternal folate and homocysteine levels play a role in TP53 in families at high risk of breast cancer. JAMAJAMA,, 295,, Lewis, S. J., Zammit, S., Gunnell, D., et aletal (2005)(2005) AA neurodevelopmental processes that increase risk for 1379^1388137 9 ^ 138 8 meta-analysis of the MTHFR C677T polymorphism and schizophrenia ? Harvard Review of Psychiatry,, 1313,197^205. schizophrenia risk. American Journal of Medical Genetics. Weickert,C. S., Straub, R. E., McClintock, B.W., et aletal Part B, Neuropsychiatric Genetics,, 135,2^4.,2^4. Risch, N. (1990) Linkage strategies for genetically (2004)(2004) Human dysbindin (DTNBP1) gene expression in complex traits. I. Multilocus models. American Journal of normal brain and in schizophrenic prefrontal cortex and Liao, H. M. & Chen, C. H. (2004) Mutation analysis of Human Genetics,, 4646, 222^228.,222^228. midbrain.midbrain. Archives of General Psychiatry,, 6161,544^555. the human dystrobrevin-binding protein 1 gene in schizophrenicpatients.SchizophreniaResearch,,7171,185^189.,185^189. Sachs, N. A., Sawa, A., Holmes, S. E., et aletal (2005)(2005) AA Williams,N.M.,Preece, A.,Morris,D.W., et aletal (2004)(2004) frameshift mutation in Disrupted in Schizophrenia 1in1 in an Identificationin 2 independent samplessamplesofanovelschizo- of a novelschizo- Malaspina, D., Corcoran, C. & Fahim, C., et aletal (2002)(2002) phreniariskhaplotypephrenia riskhaplotype ofthe dystrobrevinbindingproteindystrobrevin bindingprotein Paternal age and sporadic schizophrenia: evidence for dede American family with schizophrenia and schizoaffective gene (DTNBP1).ArchivesofGeneralPsychiatryArchivesof General Psychiatry,,61,336^344.,336^344. novo mutations. American Journal of Medical Genetics,, disorder.disorder. Molecular Psychiatry,, 1010, 758^764.,758^764. 114, 299^303.,299^303. Schwab, S. G., Knapp, M., Mondabon, S., et aletal (2003)(2003) Wong, A. H., Gottesman, I. I. & Petronis, A. (2005) Maynard,T. M., Haskell, G.T., Lieberman, J. A., et aletal Support for association of schizophrenia with genetic Phenotypic differences in genetically identical organisms: (2002)(2002) 22q11DS: genomicmechanismsandgenefunction variation in the 6p22.3 gene, dysbindin, in sib-pair the epigenetic perspective. Human Molecular Genetics,, in DiGeorge/velocardiofacialsyndrome. InternationalInternational families with linkage and in an additional sample of triad 1414,R11^R18.,R11^R18. JournalofJournal of Developmental Neuroscience,, 2020,407^419.,407^419. families.families. American Journal of Human Genetics,, 7272,185^190. Zhang, X., Tochigi, M., Ohashi, J., et aletal (2005)(2005) McClellan, J. M., Susser, E. & King, M. C. (2006) Sebat, J., Lakshmi, B., Troge, J., et aletal (2004)(2004) Large- Association study of the DISC1/TRAX locus with Maternal famine, de novo mutations, and schizophrenia. scale copy number polymorphism in the human genome. schizophrenia in a Japanese population. Schizophrenia JAMAJAMA,, 296,582^584. Science,, 305305,525^528. Research,, 79, 175^180.,175^180.

199199

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 03:58:25, subject to the Cambridge Core terms of use.