31 July 2008. Three large, independent studies aimed at understanding the genetic causes of schizophrenia were published online in Nature journals on July 30—two studies looking at rare genetic deletions or duplications implicate the same two novel chromosomal locations in the disease, while a third study parsing the contribution of common genetic variants comes up with one new candidate risk gene. The studies, which together involve thousands of people with and without schizophrenia, are likely to stimulate much news coverage in the media, and debate within the research community about just what the results mean, and where to go from here.
The two studies of rare copy number variations (CNVs) appear back to back in Nature, where one reports that the number of rare deletions or duplications leading to aberrant gene copy number is increased modestly in patients with schizophrenia. Both studies identify the same two chromosomal regions (1q21.1, 15q13.3) where copy number variation is associated with the risk of schizophrenia. The third study, based on a genome-wide association study, finds that variants in a putative gene regulator may play a role in schizophrenia and bipolar disorder.
The additions and subtractions multiply
The risk for schizophrenia is widely believed to have a large genetic component, and the prevailing theory of genetic causation is that common variants of multiple susceptibility genes each have small effects that couple with environmental influences to cause the disease. However, that view is being challenged by researchers who are studying small deletions or amplifications of parts of chromosomes that cause changes in gene copy number. CNVs were recently recognized as a major source of genetic variation among people (see SRF related news story), and have been linked to autism (see SRF related news story). Recent studies suggest a link between CNVs and schizophrenia: people with schizophrenia are reported to have elevated incidence of CNVs (see SRF related news story), including CNVs that are not inherited but have originated in the individuals (de novo CNVs; see SRF related news story).
While previous studies looked at relatively few subjects, the two studies published today use much larger sample sizes to identify rare CNVs and then to measure their impact on the risk of schizophrenia. One study, from the International Schizophrenia Consortium (ISC), looked at the genome-wide occurrence of rare CNVs in 3,391 patients and 3,181 controls, and the association of specific CNVs with disease. The researchers, a large group from Europe, North America, and Australia, were managed by corresponding author Pamela Sklar and colleagues at the Broad Institute of Harvard and MIT in Cambridge, Massachusetts. The investigators report first of all that people with schizophrenia had more rare deletions and insertions (defined as present in less than 1 percent of people and more than 100 kilobases long) than the control population, and that the CNVs involved more genes. Control subjects had an average of 0.99 CNV per person. In people with schizophrenia, the number of CNVs was increased 15 percent, a difference that was “subtle but highly statistically significant,” according to the authors. The number of affected genes was 1.41 times higher in the disease group than in healthy controls. When the researchers looked only at CNVs that were seen only once in the sample, or were larger than 100 kilobases, or disrupted at least one known gene, all of those categories showed even stronger evidence of increased occurrence in schizophrenia. These data would appear to support and extend the recent findings of Walsh et al., 2008 and SRF related news story cited above) in a far larger sample. The authors write, “The etiology of schizophrenia has been vigorously debated. We now have strong and replicated evidence that individuals with schizophrenia have a greater burden of structural variation across their genomes.”
Because of the sample size, the investigators were able to test specific CNV regions for association with the disease. As expected, they found deletions at 22q11.2 (the region involved in velo-cardio-facial syndrome and previously associated with schizophrenia) in 13 cases but no controls. There were two additional regions where patients showed significantly more CNVs than controls. Chromosome 15q13.3 showed up in nine cases and no controls, while 1q21.1 was seen in 10 cases and one control. The 1q21.1 region, which contains 27 known genes, has been previously linked to schizophrenia (Brzustowics et al., 2000; Gurling et al., 2001), while 15q13.3 has not. That region, however, recently turned up in relation to a syndrome of mental retardation with seizures (Sharp et al., 2008). A schizophrenia candidate gene also lies in this region. CHRNA7 encodes the a7 subunit of the nicotinic acetylcholine receptor, and has been linked to the schizophrenia-associated phenotype of auditory evoked potential deficits (Freedman et al., 1997; Xu et al., 2001).
There are other CNVs that were found only in the schizophrenia subjects, and of these the authors write, “At least some of these rare CNVs seen in cases but not in controls are probably risk factors for schizophrenia, although like Walsh et al., we are unable to identify which.” An important future direction will be to characterize clinically and phenotypically deletion carriers with an eye to establishing a biologically plausible link to the affected genes.
The second study, another worldwide collaborative effort, by the SGENE consortium, was led by senior author Kari Stefansson of deCODE Genetics, Reykjavik, Iceland. The trio of first authors were Hreinn Stefansson, also of deCODE, Dan Rujescu of Ludwig-Maximilians-University, Munich, Germany, and Sven Cichon of University of Bonn, also in Germany. Their working hypothesis was that the reduced reproductive success associated with severe mental disorders puts negative selective pressure on risk alleles, so that genes conferring considerable risk will be quite rare. By this reasoning, rare variants may contribute a larger percentage of risk for schizophrenia than assumed, but are accordingly harder to find. As the researchers themselves allow, it is very difficult to screen for rare nucleotide changes, but somewhat easier to look for CNVs. To look for rare large and recurrent CNVs that seemed to be under negative selective pressure, the group carried out a genome-wide search for de novo CNVs in a group of unaffected people made up of 2,160 trios of two parents and one child, plus 5,558 parent-child pairs. They identified 66 de novo CNVs, which they then examined in nearly 5,000 people with schizophrenia.
For eight of the 66 CNVs tested, at least one schizophrenia patient carried the CNV, and for three large deletions, nominal association with schizophrenia and related psychoses was detected. The three deletions nominally associating with schizophrenia in the first sample were followed up in up to six samples comprising a total of 3,285 cases and 7,951 controls. All three deletions, at 1q21.1, 15q11.2, and 15q13.3, significantly associate with schizophrenia and psychosis in the combined sample with high odds ratio (OR) (OR = 14.83, OR = 2.73, and OR = 11.54, respectively).
The results implicated the same two regions as the ISC study, with deletions at 1q21.1 detected in 11 out of 4,718 cases (0.23 percent) compared to eight of 4,119 controls (0.02 percent). Chromosome 15q13.3 was also significantly associated with schizophrenia and related psychoses in both studies. Only the SGENE study found an association with the 15q11.2 region. This part of the chromosome contains the CYFIP1 gene, which is interesting in that its deletion has been associated with Prader-Willi syndrome, which may present with perseverative behavior (lack of cognitive flexibility, also a feature of schizophrenia) and Angelman syndrome, which may present with autistic features.
The authors conclude, “The identification of these rare, recurrent risk variants, having occurred independently in multiple founders and being subject to negative selection, is important in itself. CNV analysis may also point the way to identification of additional and more prevalent risk variants in genes and pathways involved in schizophrenia.”
Don’t forget SNP associations
The new studies might indicate to some that the search for common variations with modest risk profiles has given way to the hunt for rare but catastrophic genetic changes. However, a third report, published today’s in Nature Genetics, suggests it is too early to give up the search for common genetic variants that contribute to schizophrenia. In that work, Michael O’Donovan and Michael Owen at Cardiff University, United Kingdom, and collaborators argue that studies on genome-wide associations of using single nucleotide polymorphism data (SNPs) have successfully identified genes with very small effects on other diseases, similar to what is expected for schizophrenia. They described a genome-wide association study using a multi-step approach and a total of more than 6,286 cases and 12,993 controls. Their initial analysis included 479 cases compared to 2,937 controls, genotyped at 362,532 single nucleotide polymorphisms (SNPs). They chose 12 SNPs with moderately strong evidence for association for follow-up in another set of 1,664 cases and 3,541 controls, which replicated six of the 12 SNPs. A second replication set of 4,143 cases and 6,515 controls supported the significance of the ZNF804A gene in 2q32.1 as a schizophrenia susceptibility variant, and also pinpointed two intergenic regions at 11p14.1 and 16p13.12.
In a secondary analysis, they added additional samples from people with bipolar disorder, and found the ZNF804A association was strengthened, suggesting that variants of the gene could be linked to a broader psychosis phenotype (see SRF live discussion). The protein encoded by the gene is uncharacterized, and while its function is unknown, it bears some resemblance to a possible transcription factor.
“Our study demonstrates that despite the lack of biological validating criteria for diagnosis, schizophrenia is amenable to the same genetic approaches as other common disorders, and like most other disorders, the effect sizes are small,” the authors conclude. “Our findings strongly suggest that further GWA analysis of larger samples will identify many additional specific genetic risk factors with the potential to shed light into the pathophysiology of one of the most enigmatic major causes of human morbidity. Collection and analysis of large enough samples to provide convincing association signals should now be a priority,” they write.—Pat McCaffrey.
The International Schizophrenia Consortium. Rare chromosomal deletions and duplications increase risk of schizophrenia. Nature. 2008 July 30, Advance online publication. Abstract
Stefansson H., Rujescu D., Cichon S. et al. Large recurrent microdeletions associated with schizophrenia. Nature. 2008 July 30, Advance online publication. Abstract
O’Donovan MC. et al. Identification of loci associated with schizophrenia by genome-wide association and follow-up. Nature Genet. 2008 July 30, Advance online publication. Abstract