Copy-Number Variations Associated With Schizophrenia Identified in Finnish Families

Jacquelyn K. Beals, PhD

November 09, 2009

November 9, 2009 (Honolulu, Hawaii) — Finnish researchers have identified several copy-number variations (CNVs) associated with schizophrenia that are enriched in isolated populations in a "high-risk region of Finland." Within these isolate populations, the CNVs occur more frequently in families with schizophrenic members than in the general isolate population, they reported here at the American Society of Human Genetics 59th Annual Meeting.

The investigators proposed "reverse genomics" or "inverse mapping" as an alternative approach for seeking genetic associations with phenotypes. "Going from genotype to phenotype would be unbiased, without assuming a common etiology," noted Olli P.H. Pietiläinen, BSc, from the Institute for Molecular Medicine Finland in Helsinki, in his presentation.

Research has recently demonstrated associations between CNVs and cognitive and neuropsychiatric disorders, including schizophrenia, but primarily in sporadic rather than familial cases. This study focused on a familial form of schizophrenia enriched in a founder population in the 2 northern provinces of Finland.

"The underlying assumption in studying founder populations is that less genetic variation underlies a certain phenotype or disorder within the isolate, compared with the general population," Mr. Pietiläinen explained to Medscape Pathology & Lab Medicine.

Harmful variants typically result in fewer offspring and decrease over time through negative selection, but population bottlenecks can "allow such harmful variants to escape . . .  selection pressure and become relatively enriched in the founder population," explained Mr. Pietiläinen. In such cases, the phenotype is enriched in the population. The site of this study is recognized as a "high-risk region of Finland" for the familial form of schizophrenia.

Data were obtained from individuals in the Northern Finnish Birth Cohort of 1966, which consisted of 12,058 live births in the 2 northern provinces of Finland. Cohort members were 31 years old at the time of this study. Genome-wide scans in nearly 5000 of these individuals looked for CNVs larger than 500 kilobases, and found 638 CNVs in 165 CNV regions. The regions were evenly distributed among chromosomes, but many CNVs can occupy 1 region.

The scans showed that 11% of individuals carried large CNVs, whereas 82% had a single CNV. Deletions were less frequent than expected, comprising only 29% of all CNVs found. Mr. Pietiläinen emphasized in his presentation that "developmental and neurological genes were more often overlapped by the CNVs than would be predicted by chance," leading investigators to look for central nervous system relationships among phenotypes, such as mental retardation and learning difficulties.

Mental retardation was more common in CNV-carriers than in noncarriers; subtle learning difficulties were also more common among carriers of large deletions. Of the common mental disabilities, approximately 12% co-occurred with large deletions. For example, 10% of deletion carriers were below their age-appropriate grade level in school, compared with 3.9% of the cohort as a whole (P = .00016).

Focusing on schizophrenia, the study identified 3 large CNV alleles at loci 9p24.3, 17p13.3, and 22q11.22, which occurred significantly (P < .05) more often among schizophrenia cases in the isolate population. These loci contain the developmental genes DOCK8, ABR, and TOP3B, respectively, not previously associated with schizophrenia.

Screening more than 5000 additional Finnish cases, 2061 European schizophrenia cases, and 3091 control subjects revealed that the 3 CNVs were enriched in the general Finnish population (P < 1.9 × 10 –9 ) and in the isolate population of northern Finland (P < 2.0 × 10–4).

The 3 CNVs occurred about 3 times more often in isolate families with schizophrenic members than in the general isolate population (P < .001). In addition, each CNV was roughly 50% more common in family members manifesting any type of mental disorder than in unaffected people.

The reverse genomics approach was used to screen 4932 members of the 1966 birth cohort for CNVs longer than 1 megabase. Of 118 people with a CNV of this size, 30% turned out to have a psychiatric or neurological disorder.

"Selecting individuals from a population-based sample according to allelic similarity within a certain locus enables a phenotype-independent approach to study the true physiological impact of a certain variant without any a priori assumption of shared genetic background," explained Mr. Pietiläinen." He acknowledged, however, that the reverse genomics approach can require large sample sizes, depending on the frequency of the variant being investigated.

Medscape Pathology & Lab Medicine asked session comoderator David H. Ledbetter, PhD, professor of human genetics and director of the Emory Division of Medical Genetics, Emory University School of Medicine in Atlanta, Georgia, for his perspective on these 2 approaches.

Starting with cases and finding CNVs that were significantly enriched "is obviously much easier and quicker to find candidates," said Dr. Ledbetter. Identifying CNVs in a population, then noting the phenotypes of individuals with these CNVs "requires much larger studies and is much more expensive," he said, "but less biased."

Overall, Mr. Pietiläinen's study found the inverse mapping or reverse genomics approach in an unbiased population sample to be a useful tool for characterizing phenotypic variability related to genetic variation. "I think the isolated population structure is also an advantage in the phenotype-independent inverse mapping. This way we can utilize the extended pedigrees and identify multiple carriers of 1 founder mutation," Mr. Pietiläinen concluded.

Dr. Pietiläinen and Dr. Ledbetter have disclosed no relevant financial relationships.

American Society of Human Genetics (ASHG) 59th Annual Meeting: Abstract 259. Presented October 24, 2009.