15 April 2010. Two genomewide studies, both from the Wellcome Trust Case Control Consortium, provide complementary findings on the role of common versus rare copy number variants in disease. In the Nature study, published on April 1, Peter Donnelly, University of Oxford, United Kingdom, and colleagues find no evidence to suggest that common copy number variants play any big causal role in any of eight common human diseases, including bipolar disorder. In the April Archives of General Psychiatry, a team that includes Nick Craddock and George Kirov, both of Cardiff University, Wales, conclude that subjects with bipolar disorder carry no excess burden of rare copy number variants and, in fact, harbor considerably fewer than subjects with schizophrenia.
So far, common single nucleotide polymorphisms in genes have explained only a small part of the heritable risk of developing schizophrenia and bipolar disorder (see SRF related news story). To help explain the rest, some researchers have pinned their hopes on copy number variants (CNVs)—deletions or duplications of DNA segments (see SRF related news story). They have drawn encouragement from reports of an increased CNV load in schizophrenia (see SRF related news story; SRF related news story) and from studies identifying specific CNVs that may explain a minority of schizophrenia cases (see SRF related news story; SRF related news story). Such mutations include a deletion at 22q11 that is associated with velocardiofacial syndrome (see SRF related news story).
Researchers know less about the role of CNVs in bipolar disorder, but its overlapping genetic origins with schizophrenia (see SRF related news story; SRF news story; SRF news story; and SRF Live Discussion) certainly fuel hopes that such structural variants may offer clues to its etiology also. Research has found specific CNVs in individual subjects with bipolar disorder (Wilson et al., 2006). Furthermore, a genomewide study by Dandan Zhang and colleagues (Zhang et al., 2009) found that extremely rare deletions—those found just once in the dataset—occurred more often in bipolar subjects than in controls. However, whether the overall load of CNVs is heightened in bipolar disorder has remained unclear, prompting the Archives study.
Any role for rare mutations?
Research groups across the United Kingdom formed The Wellcome Trust Case Control Consortium (WTCCC) in 2005 to facilitate the search for genetic causes of human disease. Using WTCCC data, Kirov, Craddock, and others at Cardiff University found that large, rare CNVs, especially deletions, occurred more often in subjects with schizophrenia than in a control group (Kirov et al., 2009). To learn whether subjects with bipolar disorder carry a similar burden, they performed the Archives study.
First author Detelina Grozeva, of Cardiff University, and colleagues focused on variants that involved more than 100,000 base pairs and that occurred in less than 1 percent of the population. Like other studies of rare CNVs, they used SNP chips to indirectly assay CNVs. They used two independent arrays to type them and only accepted CNVs that both arrays identified independently. They tested DNA from white residents of the United Kingdom, including 1,697 subjects with bipolar disorder and 2,806 non-psychiatric control subjects.
Grozeva and colleagues found no significant difference between cases and controls in either the total number of rare CNVs or duplications alone. However, deletions did separate the cases from the controls. Actually, the bipolar group averaged fewer deletions per person, with 0.19 versus the 0.23 seen in the control group (P = .01). On the other hand, cases and controls did not differ in the frequency of specific CNVs that occurred only once in the sample, contradicting the findings of Zhang and colleagues. When they limited the analyses to mutations that disrupted genes, the results stayed the same for both total and unique CNVs.
An earlier study by many of the same researchers found a higher burden in subjects with schizophrenia only in regard to CNVs that exceeded 1 megabase in length (Kirov et al., 2009). This burden does not seem to carry over to bipolar disorder, because Grozeva and associates found no excess of these larger CNVs in bipolar subjects compared to healthy controls.
To compare the role of these larger CNVs in schizophrenia versus bipolar disorder, Grozeva and colleagues made use of data from 440 subjects with schizophrenia from an earlier study that used the same methods to identify the same class of CNVs. The researchers found more CNVs larger than 1 megabase in subjects with schizophrenia than in bipolar cases (P <.001). In fact, deletions of this size occurred five times more often in schizophrenia (P <.001). “Our data are consistent with the possibility that possession of large, rare deletions may modify the phenotype in those at risk of psychosis; those possessing such events are more likely to be diagnosed as having schizophrenia, and those without them are more likely to be diagnosed as having bipolar disorder,” Grozeva and colleagues write (see SRF related news story).
Even though people with bipolar disorder seem to carry no excess burden of copy number variants overall, they might still harbor a surplus of CNVs associated with schizophrenia. However, when the researchers checked, they found no more schizophrenia-related CNVs in bipolar cases than in controls. Despite reported ties between bipolar disorder and velocardiofacial syndrome (Papolos et al., 1996), the researchers saw no 22q11 deletions in bipolar cases. On the other hand, cases did show extra CNVs in two regions linked to schizophrenia, 16p11.2 and 15q13.3. Craddock told SRF that even though the study found no increased burden of rare CNVs overall in bipolar disorder, some may still contribute to the illness.
Common mutations wash out
Rather than focus on rare CNVs, the Nature study looked at whether common ones contribute to eight common diseases: bipolar disorder, type 1 diabetes, type 2 diabetes, Crohn’s disease, rheumatoid arthritis, breast cancer, coronary artery disease, and hypertension. Donnelly and colleagues analyzed DNA from 19,050 subjects, including about 2,000 in each disease group and 3,000 controls. They were able to use new technology that enabled them to directly assay common CNVs in a large number of people. They designed an array that, they estimate, identifies about half of all autosomal CNVs with a length of over 500 base pairs and a minor allele frequency of over 5 percent. They warn that conducting CNV association studies can be tricky and requires great attention to methodological issues, such as whether the reference DNA comes from blood or cell lines.
All their hard work yielded only three loci containing a common CNV that the study found to be associated with one or more of the diseases, but not with bipolar disorder. Even more disappointing, prior SNP association studies had already implicated these regions and many more. Furthermore, none of the regions serve any known functional role that could contribute to disease.
“We conclude that common CNVs typable on current platforms are unlikely to have a major role in the genetic basis of common diseases,” Donnelly and colleagues write. They acknowledge that they cannot know whether unidentified CNVs relate differently to disease. Craddock told SRF that based on these findings, “we can be reasonably confident that common copy number variation isn’t a big contributor to the genetic susceptibility to bipolar illness.” This study did not look at schizophrenia, but Craddock and his colleagues are currently undertaking a study of the role of common copy number variation in that disease.—Victoria L. Wilcox.
Grozeva D, Kirov G, Ivanov D, Jones IR, Jones L, Green EK, St. Clair DM, Young AH, Ferrier N, Farmer AE, McGuffin P, Holmans PA, Owen MJ, O'Donovan MC, Craddock N, for the Wellcome Trust Case Control Consortium. Rare copy number variants: a point of rarity in genetic risk for bipolar disorder and schizophrenia. Arch Gen Psychiatry. 2010 Apr;67(4):318-27. Abstract
The Wellcome Trust Case Control Consortium. Genomewide association study of CNVs in 16,000 cases of eight common diseases and 3,000 shared controls. Nature. 2010 Apr 1;464(7289):713-20. Abstract