August 9, 2013. Clinical microarray testing may be useful to identify rare copy number variants (CNVs) in schizophrenia, reports a new study published online July 14, 2013, in Human Molecular Genetics. Researchers led by Anne Bassett of Canada’s Centre for Addiction and Mental Health at the University of Toronto used microarrays to detect rare CNVs in a community-based sample of schizophrenia subjects and found that a substantial 8 percent of cases contained these structural variants.
The deleted or duplicated chromosomal regions known as CNVs are thought to supply some of the genetic risk for schizophrenia (see SRF related news story; SRF news story). Microdeletions at chromosome 22q11.2, for example, lead to abnormally high rates of schizophrenia, and up to 1 percent of the patient population may carry this particular deletion (see SRF related news story). Despite the evidence linking CNVs to schizophrenia, clinical microarray testing is rarely used for schizophrenia. In contrast, it is widely used for autism and intellectual disability (Miller et al., 2010).
To investigate the utility of these microarrays in the detection of CNVs in a real-world clinical setting, first authors Gregory Costain and Anath Lionel recruited a community-based sample of 459 subjects with schizophrenia or schizoaffective disorder who were compared to 416 controls. This differed from previous CNV studies in that the researchers recruited patients through four outpatient mental health clinics in the region, rather than gathering them through research center settings around the world. Also, the microarray results were read by cytogenetics experts, who normally interpret genetic testing results in a clinical setting.
Using high-resolution genomewide microarrays, the researchers detected CNVs in the control and schizophrenia populations. Two cytogenetic laboratory directors performed independent, blinded analyses of the rare CNVs (defined as present in less than 0.1 percent of a separate cohort of 2,357 controls) and rated them on their pathogenicity according to the American College of Medical Genetics guidelines, which consider things such as size, number of genes in the CNV, and gene identity.
The researchers identified a 16.4-fold enrichment in large (500 kb-6.5 Mb), rare autosomal CNVs deemed to be “clinically significant” by the cytogenetic experts in the schizophrenia subjects compared to the controls. They found a total of 16 of these CNVs at eight loci in the schizophrenia group and only one clinically significant CNV, a typical 22q11.2 duplication, in the control group. Four loci were linked to schizophrenia for the first time: 2q13, 3q13.31, 5p15.33-p15.32, and 10q11.22-q11.23. The researchers highlighted the 2q13 locus, which was hit by either a 1.7 Mb duplication or deletion in three cases of schizophrenia; CNVs at this locus were not found in their control group or in another control group of 23,838 people. To estimate the prevalence of the clinically significant CNVs detected by this type of real-world clinical microarray testing, the researchers used a subset of schizophrenia subjects (n = 248) who were drawn from a single region and who had been recruited through a single mental health clinic. In this group, they found that 8.1 percent of individuals with schizophrenia harbored CNVs 500 kb or larger. Put another way, one in 13 patients carried a clinically relevant CNV. Importantly, the majority of schizophrenia patients did not exhibit syndromic physical features, suggesting that microarrays would be necessary to identify such CNVs.
Currently, knowing that people carry one of these CNVs would not likely change how their illness is managed. But guidelines may be developed in the future, as in the case of the better-known 22q11.2 deletion. Similar to autism, developmental delay, and intellectual disability, “the availability of clinical CNV data for schizophrenia, coupled with an improved ability to interpret these data for the benefit of patients and families, could have far-reaching scientific and clinical implications,” concluded the authors, who noted that further studies were necessary to validate the use of this approach.
If schizophrenia patients and their families have access to clinical microarray testing, they will also need genetic counseling along with the results. While studies suggest that patients and families obtain some psychological comfort from the explanatory power of these genetic findings (Costain et al., 2012; Costain et al., 2012), such counseling will also likely raise new ethical issues, as the University of Chicago’s Elliot Gershon and Ney Alliey-Rodriguez discuss in a review published online July 30, 2013, in the American Journal of Psychiatry (Gershon and Alliey-Rodriguez, 2013). For example, the potential for abortion and pre-implantation selection of embryos based on the probability of psychiatric disease and a determination of the rights of family members to a relative’s genetic information will need to be carefully navigated. “A psychotherapeutic approach may be needed as a routine part of risk counseling, particularly for resolution of ethical issues and for within-family stigma and conflicts over genetic results,” suggest Gershon and Alliey-Rodriguez.
Bassett and colleagues also observed an enrichment of “variants of unknown significance”—CNVs whose pathogenicity could not be determined—in schizophrenia subjects compared to controls. They point in particular to three novel loci (6q11.1, 7p21.3, and 12q21.31) affected in more than one unrelated schizophrenia subject, as well as a case with a CNV at the 15q12-q13.1 locus already on the radar for schizophrenia. These regions contained several genes that may be of interest in schizophrenia, including GABA receptor and postsynaptic density genes.
To get a feel for the total burden of rare CNVs detected by the microarrays, the researchers combined the number of large CNVs (500 kb-6.5 Mb) judged either as clinically significant or of unknown significance. They left out 22q11.2 deletions or CNVs greater than 6.5 Mb in this tally, however, because these could conceivably be detected without microarrays. Still, their schizophrenia group showed a threefold higher incidence of the large CNVs compared to controls. Individuals with schizophrenia were more likely to carry multiple CNVs that hit exons compared to controls—something consistent with the idea that multiple hits to the genome increase risk for a disorder.
Analysis of very rare CNVs (found in none of the 2,357 controls) also held some clues: Several overlapped the same gene in at least two schizophrenia subjects. This led to 15 candidate genes, including RBFOX1, SOX5, DNM1L, and JAK2. Other very rare CNVs that were found in just a single schizophrenia subject yielded other new candidates as well as some old favorites including DISC1, RELN, and several glutamate receptor genes.—Allison A. Curley.
Corrections: This story originally misstated that cytogenetic review had detected a 22q11 deletion in one of the control subjects (it was a duplication) and misidentified author Anne Bassett's primary institution. Also, we have clarified that the authors pointed out novel CNVs of unknown pathogenecity only where they were detected in more than one unrelated schizophrenia subject.
Costain G, Lionel AC, Merico D, Forsythe P, Russell K, Lowther C, Yuen T, Husted J, Stavropoulos DJ, Speevak M, Chow EW, Marshall CR, Scherer SW, Bassett AS. Pathogenic rare copy number variants in community-based schizophrenia suggest a potential role for clinical microarrays. Hum Mol Genet. 2013 Jun 27. Abstract