November 22, 2013. Duplications of the 22q11.2 chromosomal region seem to protect against schizophrenia, according to a study published online November 12 in Molecular Psychiatry. Led by Michael Owen and Michael O’Donovan of Cardiff University, United Kingdom, the study surveyed the genomes of 47,005 people and found that the frequency of large 22q11.2 duplications in people with schizophrenia was much lower than in people without the disorder. In fact, a person having the duplication had only about one-sixth the chance of developing schizophrenia that people without the duplication did. As the well-known counterpart deletion of 22q11.2 confers risk for schizophrenia, the results suggest that the dosage of genes within this region contributes to a sliding scale of risk or protection.
Previous meta-analyses have found common alleles with protective effects against schizophrenia (Allen et al., 2008), but this is the first time a large, structural variation has been identified. Though some losses or gains of chromosome regions, or copy number variations (CNVs), seem harmless enough (e.g., Sebat et al., 2004), rare and large CNVs greater than 100 kb turn up more frequently in several brain disorders, including intellectual disability (ID), autism, epilepsy, and schizophrenia (Malhotra et al., 2012). In the limited reports to date, duplications or deletions of the same chromosomal regions—called “reciprocal” CNVs, such as those found at 16p11.2—appear to increase risk for different disorders (see SRF related news story) rather than producing the bi-directional risk-protective effect found in the new study.
Large (1.5-3 Mb) deletions of 22q11.2 cause velocardiofacial syndrome (also known as DiGeorge syndrome), which has a wide variety of symptoms, and 22q11.2 deletions are also associated with autism and ID. About 25 percent of people carrying the 22q11.2 deletion—which contains at least 30 genes—develop psychosis and receive a diagnosis of schizophrenia, making it one of the most potent genetic risk factors for schizophrenia. Carriers of the reciprocal duplication also display a spectrum of phenotypes, including ID, autism, and even fairly normal (Ou et al., 2008). The new study, however, is the first to specifically survey people with schizophrenia for these duplications and found they were conspicuously missing.
Clues from 22q
First author Elliot Rees and colleagues began by genotyping 6,882 people with schizophrenia and 11,255 controls with single nucleotide polymorphism (SNP) arrays. They found no 22q11.2 duplications in the schizophrenia group, but 10 in the controls. In contrast, they detected 20 22q11.2 deletions in the schizophrenia group, but none in the controls, as expected. Looking in additional cohorts, the researchers found three people with schizophrenia who carried 22q11.2 duplications, but this frequency still remained lower than that in the control group. When combining these results with the original discovery cohort, they found that, in 47,005 people surveyed, 22q11.2 duplications turned up in 22 (0.085 percent) of 25,867 controls but only in three (0.014 percent) of 21,138 people with schizophrenia. Though rare in both cases, they were significantly more rare in schizophrenia (p = 0.00086).
Although the protective effect was not complete, the three people with schizophrenia who also carried 22q11.2 had relatively late ages of onset (28, 34, and 43). They did not carry other known risk CNVs. Of the 22 controls with the duplications, there was information available about psychiatric histories for only six, but these did not show anything resembling schizophrenia, schizoaffective disorder, or major affective disorder.
This apparent bi-directional effect for schizophrenia, with duplications decreasing risk and deletions increasing it, is not found for other disorders. For autism and ID, for example, the researchers noted that both 22q11.2 deletions and duplications were more frequent in cases than in controls. This disassociation suggests that genes within 22q11.2 underlie some schizophrenia-specific molecular effects.
Consistent with this notion, gene expression varied with gene dosage in a subset of samples tested. Lymphoblasts from a subset of duplication and deletion carriers, surveyed with RNA sequencing, showed the expected increases and decreases in expression of genes within the 22q11.2 region, but not in genes immediately flanking the region. This supports the idea that genes within the region confer the protective or risk effects, rather than that the CNV disrupts chromatin elsewhere in the genome.
The finding may intensify the hunt for the operative gene(s) for mental illness in the 22q11.2 region. Previous research focused on the candidate genes COMT, PRODH, GNB1L, and TBX1 (see SRF related news story), but the new protective effect could give researchers another phenotype to work with in their efforts to find the relevant genes. It may also provide a clue to understanding the varied penetrance and expressivity that comes with CNVs.—Michele Solis.
Rees E, Kirov G, Sanders A, Walters JT, Chambert KD, Shi J, Szatkiewicz J, O'Dushlaine C, Richards AL, Green EK, Jones I, Davies G, Legge SE, Moran JL, Pato C, Pato M, Genovese G, Levinson D, Duan J, Moy W, Göring HH, Morris D, Cormican P, Kendler KS, O'Neill FA, Riley B, Gill M, Corvin A; Wellcome Trust Case Control Consortium, Craddock N, Sklar P, Hultman C, Sullivan PF, Gejman PV, McCarroll SA, O'Donovan MC, Owen MJ. Evidence that duplications of 22q11.2 protect against schizophrenia. Mol Psychiatry. 2013 Nov 12. Abstract