Schizophrenia 3
For a phenotypic description and a discussion of genetic heterogeneity of schizophrenia, see 181500.
MappingAssociation with Major Histocompatibility Locus on Chromosome 6p21
In linkage disequilibrium mapping of the MHC region in 80 British parent-offspring trios, Wei and Hemmings (2000) found that NOTCH4 (164951) was highly associated with schizophrenia. The A-to-G substitution in the promoter region and the (CTG)n repeat in exon 1 of NOTCH4 were considered possible candidate sites conferring susceptibility to schizophrenia.
Using all markers of NOTCH4 previously shown to be associated with schizophrenia, Sklar et al. (2001) found no evidence for an association in 3 independent family-based samples totaling 519 parent-offspring trios, and in a case-control sample derived from the same ethnic background as the original observation. Similarly, McGinnis et al. (2001) failed to replicate the reported association in a large sample of unrelated Scottish schizophrenics and controls, finding instead that each putative schizophrenia-associated allele in the study of Wei and Hemmings (2000) had a somewhat lower frequency in schizophrenics than in controls.
Stefansson et al. (2009) combined SNP data from several large genomewide scans and followed up the most significant association signals. They found significant association with several markers spanning the major histocompatibility complex (MHC) region on chromosome 6p22.1-p21.3. Stefansson et al. (2009) concluded that their findings implicating the MHC region are consistent with an immune component to schizophrenia risk. The T allele of the SNP rs6932590 in the MHC region achieved a P value of 1.4 x 10(-12).
The International Schizophrenia Consortium (2009) performed a genomewide association study of 3,322 European individuals with schizophrenia and 3,587 controls. Using 2 analytic approaches, the consortium showed the extent to which common genetic variation underlies the risk of schizophrenia. The MHC complex was implicated, as noted by Stefansson et al. (2009), and the authors provided molecular genetic evidence for a substantial polygenic component to the risk of schizophrenia involving thousands of common alleles of very small effect. They showed that this component also contributes to the risk of bipolar disorder, but not to several nonpsychiatric diseases.
Shi et al. (2009) demonstrated that schizophrenia is significantly associated with SNPs in the extended MHC complex region on chromosome 6. Shi et al. (2009) carried out a genomewide association study of common SNPs in the Molecular Genetics of Schizophrenia (MGS) case-control sample of 2,681 cases and 2,653 controls of European ancestry, and then a metaanalysis of data from MGS, International Schizophrenia Consortium, and SGENE data sets. No MGS finding achieved genomewide statistical significance. In the metaanalysis of subjects of European ancestry (8,008 cases, 19,077 controls), significant association with schizophrenia was observed in a region of linkage disequilibrium on chromosome 6p22.1 at rs13194053 with a p = 9.54 x 10(-9). This region includes a histone gene cluster and several immunity-related genes.
For discussion of the contribution of the C4 locus on chromosome 6p21 to schizophrenia risk, see MOLECULAR GENETICS.
Association with Variation in the DTNBP1 gene on Chromosome 6p22.3
Straub et al. (2002) performed family-based association analysis of 36 simple sequence length polymorphism markers and 17 SNP markers implicating 2 regions, separated by approximately 7 Mb, in the 6p24-p21 region. Straub et al. (2002) focused on one of these regions, 6p22.3, and found that SNPs within the DTNBP1 gene (607145) were strongly associated with schizophrenia.
Kendler et al. (1996) gave a detailed description of the Irish Study of High Density Schizophrenia Families (ISHDSF). Straub et al. (1995, 2002) presented evidence in support of linkage of SCZD to the region 6p24-p21 in the ISHDSF. Straub et al. (2002) described the results of family-based association analysis of simple sequence-length polymorphism (SSLP) markers on chromosome 6p and additional analysis of SNP markers in 6p22.3 and of SNP haplotypes. They reported finding genetic variation in the DTNBP1 gene, encoding dysbindin, that was associated with schizophrenia and related phenotypes.
Van Den Bogaert et al. (2003) investigated the DTNBP1 gene in 3 samples of subjects with schizophrenia and unaffected control subjects of German, Polish, and Swedish descent. They identified significant evidence of association of a 5-marker haplotype (ACATT) in the Swedish sample but not in those from Germany or Poland. The results in the Swedish sample became even more significant after a separate analysis of those cases with a positive family history of schizophrenia (p = 0.00009).
Schwab et al. (2003) tested 6 of the most positive DNA polymorphisms in the DTNBP1 gene identified in a sib-pair sample and in an independently ascertained sample of triads (both parents and 1 child) with schizophrenia comprising 203 families, including families for which linkage on chromosome 6p was detected. Evidence of association was observed in the 2 samples separately as well as in the combined sample. Multilocus haplotype analysis increased the significance further. Estimation of frequencies for 6-locus haplotypes revealed 1 common haplotype with a frequency of 73.4% in transmitted, and only 57.6% in nontransmitted, parental haplotypes. All other 6-locus haplotypes occurring at a frequency of more than 1% were less often transmitted than nontransmitted.
To investigate the association between DTNBP1 and schizophrenia, Williams et al. (2004) performed a case-control association study using a sample from Cardiff, Wales, comprising 708 subjects meeting DSM-IV criteria for schizophrenia who were individually matched for age, sex, and ethnicity to 711 controls and a second sample from Dublin, Ireland, comprising 219 subjects meeting DSM-III-R criteria for schizophrenia or schizoaffective disorder and 231 controls. In the Cardiff sample, Williams et al. (2004) found no association between schizophrenia and previously implicated haplotypes (Straub et al., 2002; Schwab et al., 2003), but strong evidence for association with multiple novel haplotypes with maximum evidence for a novel 3-marker haplotype (SNPs P1655, P1635, and 'A') (global p less than 0.001), composed of 1 risk haplotype (p = 0.01) and 2 protective haplotypes (p = 0.006 and p less than 0.001, respectively). These risk and protective haplotypes were replicated in the Dublin sample with p = 0.02, 0.047, and 0.006, respectively.
Bray et al. (2005) reanalyzed the data reported by Williams et al. (2004) and showed that a defined schizophrenia risk haplotype consisting of 1 or more cis-acting variants resulted in a relative reduction in DTNBP1 mRNA expression in human cerebral cortex. Subsidiary analyses suggested that risk haplotypes identified in other sample groups of white European ancestry also may index lower DTNBP1 expression, whereas putative 'protective' haplotypes index high DTNBP1 expression. The authors concluded that variation in the DTNBP1 gene may confer susceptibility to schizophrenia through reduced expression.
Funke et al. (2004) analyzed 7 DTNBP1 SNPs initially described by Straub et al. (2002) in a cohort of 524 individuals with schizophrenia or schizoaffective disorder and 573 control subjects. The minor alleles of 3 SNPs, P1578 (1578C/T, rs1018381), P1763 (1763A/C, rs2619522), and P1765 (1765G/A, rs2619528), were positively associated with the diagnosis of schizophrenia or schizoaffective disorder in the white subset of the study cohort, with P1578 showing the most significant association. The same 3 SNPs were also associated in a smaller Hispanic subset. No association was observed in the African American subset.
In a Japanese sample of 670 patients with schizophrenia and 588 controls, Numakawa et al. (2004) found strong evidence for association in a 2-marker haplotype in DTNBP1 (SNPs P1635 and P1325; P = 0.00028).
Li et al. (2005) studied 638 nuclear families from the Han Chinese population of Sichuan Province with at least 1 member with schizophrenia as well as 580 Scottish schizophrenia patients and 620 controls. The samples were genotyped for 10 SNPs in DTNBP1 plus rs2619538 (SNP 'A') located in the promoter region of the gene. In the Chinese trios, 2 SNPs, P1635 and P1765, were significantly overtransmitted, but these alleles were opposite of those reported in other studies. SNPs P1757 and P1765 formed a common haplotype which also showed significant overtransmission. In the Scottish patients and controls, no individual markers were significantly associated with schizophrenia. A single haplotype, which included rs2619538 and P1583, and 1 rare haplotype, composed of P1320 and P1757, were significantly associated with schizophrenia, but no previously reported haplotypes were associated. Thus, the data from the Chinese population support a role for DTNBP1 as a susceptibility gene for schizophrenia, but with haplotypes different from those previously observed. The lack of replication in the Scottish samples also indicates that caution is warranted when evaluating the robustness for evidence of DTNBP1 as a genetic risk factor for schizophrenia.
DeRosse et al. (2006) studied 181 Caucasian patients with schizophrenia for association of negative symptoms with risk haplotypes of DTNBP1 (607145). They found a significant association between lifetime severity of negative symptoms of schizophrenia and the CTCTAC haplotype overrepresented in the sample described by Funke et al. (2004). The authors speculated that given evidence that the dysbindin genotype influences glutamatergic function, this genotype may exert its effects on negative symptoms through a glutamate receptor-related mechanism.
Kohn et al. (2004) used identity by descent haplotype sharing analysis in a study of 52 patients with major psychiatric disorders from a genetically isolated village in Israel. Analysis of 8 Y chromosome markers confirmed that this isolate had a common paternal origin. Analysis of 359 microsatellite markers on 9 candidate chromosomes identified 2 significant peaks of haplotype sharing (p less than 0.0001): one for psychotic patients with any diagnosis located in the region of the dysbindin gene (DTNBP1) and the other for patients with schizophrenia on chromosome 1p32.
Mutsuddi et al. (2006) noted that since the original association study in Irish pedigrees implicating DTNBP1 as a schizophrenia susceptibility gene (Straub et al., 2002), several replication studies reported confirmation of the association in independent European samples; however, reported risk alleles and haplotypes appeared to differ between studies, and comparison among studies was confounded because different marker sets were employed by each group. To facilitate evaluation of existing evidence of association and further work, Mutsuddi et al. (2006) supplemented the extensive genotype data, available through the International HapMap Project, about DTNBP1 by specifically typing all associated SNPs reported in each of the studies of the CEPH-derived HapMap sample (CEU). Using this high-density reference map, they compared the putative disease-associated haplotype from each study and found that the association studies are inconsistent with regard to the identity of the disease-associated haplotype at DTNBP1. Specifically, all 5 'replication' studies defined a positively associated haplotype that is different from the association originally reported. Mutsuddi et al. (2006) further demonstrated that, in all 6 studies, the European-derived populations studied have haplotype patterns and frequencies that are consistent with HapMap CEU samples (and each other). Thus, they considered it unlikely that population differences are creating the inconsistency of the association studies. They concluded that the evidence of association was equivocal and unsatisfactory.
Allen et al. (2008) performed a metaanalysis comparing 2,696 Caucasian patients with schizophrenia with 2,849 controls and found that the P1325 allele (rs1011313) in the DTNBP1 gene was associated with susceptibility to schizophrenia (OR, 1.23; 95% CI, 1.07-1.40; p = 0.003). According to the Venice guidelines for the assessment of cumulative evidence in genetic association studies (Ioannidis et al., 2008), the DTNBP1 association showed a 'strong' degree of epidemiologic credibility.
Association with Other Regions of Chromosome 6
Wang et al. (1995) presented evidence for a schizophrenia susceptibility locus (SSL) on 6pter-p22. They performed linkage analysis in 186 multiplex families. Assuming locus homogeneity and using a model with partially dominant inheritance and moderately broad disease definition, they obtained a lod score of 3.2 for D6S260 on 6p23. A multipoint lod score of 3.9 was obtained when the F13A1 and D6S260 loci were analyzed, allowing for locus heterogeneity. The nonparametric affected pedigree test provided results that also supported the 6p23 assignment. In general, Wang et al. (1995) concluded that the data supported a model of locus heterogeneity.
Straub et al. (1995) presented the results of linkage analysis in 265 pedigrees using 16 6p markers; the 186 pedigrees tested with 7 markers by Wang et al. (1995) were a subset of the pedigrees reported by Straub et al. (1995). The full data supported the presence of a vulnerability locus for schizophrenia on 6p24-p22. The greatest lod score, assuming locus heterogeneity, was 3.51 with D6S296. This locus appeared to influence vulnerability to schizophrenia in 15 to 30% of the pedigrees. Schwab et al. (1995) performed multipoint affected sib-pair linkage analysis scanning 6p in 54 families ascertained in 2 regions of Germany (43 families) and in Israel (11 Sephardic Jewish families). Positive lod scores were obtained over a wide region with a maximum lod score of 2.2 occurring near D6S274 where a positive lod score had been reported also by Straub et al. (1995). A combined total lod for the 2 studies of 3.6 to 4.0 supported the presence of a susceptibility locus in this region.
In a follow-up study of 10 of 26 loci previously found to be suggestive of linkage in an Icelandic population, Moises et al. (1995) found evidence of linkage to markers in 6p, 9, and 20 in a second group of schizophrenic patients drawn from Austria, Canada, Germany, Italy, Scotland, Sweden, Taiwan, and the US. Evidence of linkage to markers distal to the HLA region on 6p was found in a third population of Chinese schizophrenics.
In a follow-up of previously positive results at the Johns Hopkins Hospital and at the University of Virginia, a consortium of 14 research centers assembled a new sample of some 500 informative pedigrees in which they found supportive, but inconclusive evidence for a schizophrenia susceptibility locus on chromosome 6, with an affected sib-pair maximum lod score of 2.19 (nominal p = 0.001) (Schizophrenia Linkage Collaborative Group for Chromosomes 3, 6, and 8, 1996).
In a sample of 211 families ascertained on the basis of having an affected sib-pair diagnosed with schizophrenia or schizoaffective disorder, Garner et al. (1996) found no evidence for linkage to markers spanning 37 cM around 6p24.
Wang et al. (1996) reported evidence for linkage disequilibrium between schizophrenia and the ATXN1 (601556) CAG repeat. The study was undertaken because of the report by several groups of linkage between a schizophrenia-susceptibility gene and markers on 6p. The authors suggested that, if the evidence of linkage disequilibrium with the ATXN1 gene is valid, this narrows substantially the region containing the gene, since linkage disequilibrium may be detectable over regions much smaller than those implicated by linkage analyses. In the second place, given the known biologic function of the ATXN1 gene in the brain and the possible relevance of anticipation and CAG repeat expansion in some rare cases of schizophrenia, ATXN1 may become a promising candidate for schizophrenia.
Using 9 microsatellite markers spanning 40 cM around 6p24-p22, Daniels et al. (1997) found no evidence of allele sharing identity by descent in a sample of 102 sib pairs from 86 families.
In 1 of 18 pedigrees studied, Maziade et al. (1997) found lod scores of 2.49 and 2.15, respectively, at the D6S296 and D6S277 loci under a dominant model when schizophrenia, schizoaffective disorder, and bipolar disease were used as criteria of affected status. This provides further evidence that these 3 disorders may share some susceptibility loci. (See possibly related susceptibility loci for both schizophrenia and bipolar spectrum disorders on 18p (125480; 603206) and 22q11 (192430).) In a study of 10 Canadian kindreds of Celtic origin, Brzustowicz et al. (1997) found no evidence of linkage of 6p markers to an either narrowly or broadly defined schizophrenia phenotype under a dominant or a recessive model. However, they found significant association with D6S1960 (p = 1.2 x 10(-5) under 2-point and p = 5.4 x 10(-6) under multipoint analyses) and quantitative scores on a scale for positive (psychotic) symptoms. This suggested that a locus on 6p may be associated with severity of psychotic symptoms.
In a multicenter study of 42 sib pairs with either schizophrenia or schizoaffective disorder drawn from 30 nuclear African American nuclear families, Kaufmann et al. (1998) found that several regions, including chromosomes 6q16-q24, 8pter-q12, 9q32-q34, and 15p13-q12, showed evidence consistent with linkage (p = 0.01-0.05). However, neither these nor any of the other 459 short-tandem repeat markers in this genomewide scan showed evidence of significant linkage by the criteria of Lander and Kruglyak (1995).
To facilitate the identification of candidate genes for schizophrenia, celiac disease (see 212750), and orofacial clefting (see 119530) previously assigned to the 6p25-p23 region, Olavesen et al. (1997) sublocalized and ordered 39 ESTs that had been assigned to this interval by radiation hybrid mapping. Most of the ESTs (31 of 39) were positioned in the 6p24-p23 interval, and of these, 8 were located within a single PAC clone.
By a novel phenotyping strategy in schizophrenia, targeting different neurocognitive domains, neurobehavioral features, and selected personality traits, Hallmayer et al. (2005) succeeded in identifying a homogeneous familial subtype of the disease characterized by pervasive neurocognitive deficit. Their genome scan data indicated that this subtype, which accounted for up to 50% of their sample, had a distinct genetic basis and explained linkage to 6p24 previously reported. If representative of other samples, the ratio of schizophrenia subtypes observed in their families could have a profound impact on sample heterogeneity and on the power of genetic studies to detect linkage and association. They proposed an abbreviated battery of tests to facilitate phenotype characterization for genetic analyses and allow a focus on a crisply defined schizophrenia subtype, thus promoting a more informed search for susceptibility genes. In the study of Hallmayer et al. (2005), 2 pure types, referred to as 'cognitive deficit' (CD) and 'cognitively spared' (CS), displayed markedly contrasting test profiles. The CD pure type was characterized by a high probability of poor performance on the majority of cognitive tasks, an increase of prevalence of nonlocalizing ('soft') neurologic signs, and non-right-handedness. The CS pure type exhibited high scores for psychometric schizotypy and for traits associated with psychosis proneness.
In studies of cultured fibroblasts from patients with schizophrenia, Gysin et al. (2007) found 26% decreased GCL activity, and 29% GCLC protein expression under stressed conditions compared to control cells. GSH content was not significantly different between the 2 groups. Two independent studies of 66 Swiss patients and 322 Danish patients showed an association between schizophrenia and a trinucleotide GAG repeat polymorphism, with 7, 8 or 9 repeats, located 10 bp upstream from the start codon. The most common genotype 7/7 was more frequent in controls, whereas the rarest genotype 8/8 was up to 3 times more frequent in patients. Patients with disease-associated genotypes, including 7/8, 8/8, 8/9, and 9/9, had decreased GCL activity, GCLC protein, and GSH content compared to individuals with the protective genotypes 7/7 and 7/9. There was no correlation between genotype and GCLC gene expression at the mRNA level, suggesting that the trinucleotide repeat polymorphism may affect mRNA transport or translation.
Molecular GeneticsThe functionally distinct C4A (120810) and C4B (120820) genes, located on chromosome 6p21.3 in the MHC locus, vary in structure and copy number, and segregate in both long (L) and short (S) genomic forms distinguished by the presence or absence of a HERV insertion in intron 9 that does not change the protein sequence. In postmortem human adult brain samples, Sekar et al. (2016) found that C4 RNA expression was directly proportional to C4 copy number; that expression levels of C4A were 2 to 3 times greater than levels for C4B, even after controlling for relative copy number; and that copy number of the C4-HERV sequence increased the ratio of C4A to C4B expression. Sekar et al. (2016) characterized the structural variation of human C4 genes (C4A and C4B) and evaluated association to over 7,000 SNPs across the extended MHC locus (chr6:25-34 Mb), C4 structural alleles, and HLA sequence polymorphisms in 28,799 schizophrenia cases and 35,986 controls. They found an association of schizophrenia with a large set of similarly associating SNPs in the distal 2 Mb of the MHC, the most strongly associated being rs13194504 (p = 5.5 X 10(-28)). The other peak of association centered at the C4 locus, where schizophrenia associated most strongly with the genetic predictor of C4A expression levels (p = 3.6 X 10(-24)). These peaks had little correlation with each other, suggesting they reflect distinct genetic influences. The more strongly a SNP correlated with predicted C4A expression, the more strongly it associated with schizophrenia. Sekar et al. (2016) compared C4A RNA expression levels in brain tissue from 35 patients with schizophrenia and 70 individuals without schizophrenia, and found that median C4A expression in patient tissue was 1.4-fold greater (p = 2 x 10(-5)) and was elevated in each of 5 brain regions assayed. The relationship remained significant after controlling for higher average C4A copy number among patients. Studies in mice demonstrated defects in synaptic remodeling in C4-deficient mice. Sekar et al. (2016) concluded that association of schizophrenia with the MHC locus on chromosome 6p21.3 involves many common, structurally distinct C4 alleles that affect expression of C4A and C4B in the brain, with each allele associated with schizophrenia risk in proportion to its effect on C4A expression.
HistoryMarshall (1995) suggested that the report by Wang et al. (1995) 'may say as much about problems of scientific collaboration as about the biology of the disease.' Scott R. Diehl, the senior author of Wang et al. (1995) (who in accordance with the tradition of the times was listed last), had a falling out with collaborators in Virginia who had previously initiated the studies of inherited schizophrenia in Ireland. Diehl had obtained DNA and diagnostic information from psychiatrists in Ireland who were not included in the authorship.