December 18, 2013. Rare copy number variants (CNVs) associated with schizophrenia or autism affect cognition and brain structure even in carriers who do not have the illnesses, according to new research published online in Nature on December 18, 2013. Led by Kari Stefansson, deCODE genetics/Amgen, Reykjavík, Iceland, and Andreas Meyer-Lindenberg, University of Heidelberg, Mannheim, Germany, the study finds that these carrier controls show cognitive functioning that is intermediate between subjects with schizophrenia and population controls. The paper also reports that deletion or duplication of the same region on the short arm of chromosome 15 produces opposite changes in gray and white matter structures, which the authors say provides the first evidence of an allele dose effect on human brain structure.
Findings that some CNVs—small deletions and duplications of chromosomal regions—increase the risk for neuropsychiatric illnesses such as schizophrenia (see SRF overview and SRF related news story) and autism (see SRF related news story) have been accumulating over the past several years. The CNVs are present only in a small minority of patients, but for that group, they may be a major force behind the illnesses. In addition, these DNA troublemakers are not fully penetrant—not everyone who has a particular CNV will develop schizophrenia or autism. Just how (and if) the CNVs affect these control carriers is largely unknown, but they afford researchers a unique chance to study the effects of CNVs without many of the complicating factors such as medication and chronicity that go hand in hand with neuropsychiatric illnesses.
In the current study, first authors Hreinn Stefansson and Meyer-Lindenberg examined the effect of 26 rare CNVs associated with schizophrenia or autism (chosen based on a literature search) on cognition and brain structure in an Icelandic cohort. Altogether, the CNVs were present in 1,178 (1.16 percent) of 101,655 subjects. Carriers between 18 and 65 years of age, including both controls and those with neuropsychiatric illnesses, were analyzed to determine the effects of the CNVs on reproductive outcome (fecundity). Three hits were found: Individuals with a 16p11.2 deletion or 22q11.21 duplication had fewer children by age 45 than controls did, and the effects were significantly greater in males than in females. Subjects with a 16p12.1 deletion, on the other hand, displayed increased fecundity. Schizophrenia subjects also exhibited reduced fecundity, similar to previous reports (Haukka et al., 2003), and the effect was also greater in males than females.
P’s, Q’s, and cognition
To examine the effect of the CNVs on cognition, the researchers administered a neuropsychological battery to 167 control carriers (no diagnosis of a psychotic illness, autism, intellectual disability, or developmental delay), 465 controls carrying other CNVs not associated with psychiatric illnesses, and 475 population controls lacking any large CNVs. In addition, 161 subjects with schizophrenia who were not carriers of the neuropsychiatric CNVs were also examined. The tests measured cognitive domains such as attention, working memory, and processing speed that are altered in schizophrenia.
Consistent with the literature, schizophrenia subjects performed worse than controls on all tests. Neuropsychiatric CNV carriers showed milder deficits, performing at a level in between the schizophrenia subjects and population controls, while controls carrying CNVs not associated with neuropsychiatric illnesses showed no such difficulties. Accounting for IQ diminished the effects, a finding not surprising given that cognitive tests measure abilities that contribute to IQ, write the authors.
Neuropsychiatric CNV control carriers displayed significantly lower scores on the General Assessment of Function scale (GAF), as well as more depression and suicidal ideation than population controls. They also appeared to have a greater history of reading and mathematical learning disabilities than population controls, while carriers of other CNVs did not show this effect.
CNVs singled out
Eleven of the 26 CNVs were present in enough study participants to permit an analysis of their individual effects on cognition. Six CNVs, including the 16p11.2 and 16p12.1 deletions associated with fecundity changes as well as a 16p11.2 duplication, were significantly associated with large verbal and/or performance IQ deficits in carrier controls. Several CNVs were also associated with performance on individual cognitive tests, a history of learning difficulties, and GAF score. Similar to the findings when the neuropsychiatric CNVs were lumped together, accounting for IQ also diminished these effects.
Four CNVs were associated with milder cognitive deficits in carrier controls. For example, a deletion at 15q11.2 between breakpoints 1 and 2 (15q11.2[BP1-BP2]; n = 47) that has previously been linked to schizophrenia and developmental delays was associated with only small deficits in the neuropsychological battery and a slightly lower IQ, although control carriers had significantly lower GAF scores and significant histories of reading and math learning difficulties. Carriers of the reciprocal duplication were not different from controls on any of the measures.
Concentrating on brain areas implicated in a meta-analysis of first-episode psychosis (Radua et al., 2012), Stefansson, Meyer-Lindenberg, and colleagues found that 15q11.2(BP1-BP2) deletion (n = 15) and duplication (n = 55) control carriers exhibited several structural MRI abnormalities compared to population controls (n = 201). Deletion carriers exhibited lower perigenual anterior cingulate cortex gray matter volume, as well as reduced temporal lobe and increased corpus callosum white matter volumes. The authors noted that these changes partially overlap with those observed during the onset of psychosis. Importantly, opposite changes in the same gray and white matter areas were observed in the duplication carriers, “the first demonstration of allele-dose-dependent effects of CNVs on the structure of the human brain,” they wrote.
However, Stefansson, Meyer-Lindenberg, and colleagues point out that although the 15q11.2(BP1-BP2) deletion appeared to have a negative impact on reading and math learning in control carriers, those with the duplication performed as well as controls, suggesting that the reciprocal changes in brain volume may not extend to learning.
The current findings that cognitive and structural brain abnormalities of schizophrenia are also found in control subjects carrying high-risk CNVs bolster “the idea that cognitive abnormalities are fundamental defects in schizophrenia,” conclude the authors.—Allison A. Curley.
Stefansson H, Meyer-Lindenberg A, Steinberg S, Magnusdottir B, Morgen K, Arnarsdottir S, Bjornsdottir G, Walters GB, Jonsdottir G, Doyle OM, Tost H, Grimm O, Kristjansdottir S, Snorrason H, Davidsdottir SR, Gudmundsson LJ, Jonsson GF, Stefansdottir B, Helgadottir H, Haraldsson M, Jonsdottir B, Thygesen JH, Schwarz AJ, Didriksen M, Stensbøl TB, Brammer M, Kapur S, Halldorsson JG, Hreidarsson S, Saemundsen E, Sigurdsson E, Stefansson K. CNVs conferring risk of autism or schizophrenia affect cognition in controls. Nature. 2013 Dec 18. Abstract