By Aaron Krol
January 24, 2014 | Two days ago, Bio-IT World reported on a pair of Nature papers that together constituted the largest genetic study of schizophrenia ever performed. Those papers tried to gauge genetic risks for schizophrenia considered as a discrete condition, but both raised the specter of schizophrenia existing on a spectrum, with large areas of overlap with other cognitive disorders, where small genetic effects additively contribute to a range of possible pathologies. As one of these papers speculated in its Discussion section:
Our findings of overlaps between the pathogenic mechanisms underlying schizophrenia and those in autism and intellectual disability lend support to recent, controversial suggestions that our understanding of these disorders might be advanced better by research that integrates findings across multiple disorders and places more emphasis on domains of psychopathology (for example, cognition) and their neurobiological substrates rather than current diagnostic categories. (De novo mutations in schizophrenia implicate synaptic networks, Nature, Jan 22, 2014.)
In a paper published in Nature last December, a team of researchers from Icelandic genetics company deCODE (now a subsidiary of Amgen) and the University of Heidelberg tackled that very question in an intriguing study of asymptomatic schizophrenia-linked mutations. DeCODE has been a powerhouse of large genome-wide studies, thanks to its remarkable database of genetic information – including tens of thousands of whole genomes – representing roughly a third of the entire Icelandic population. In the paper "CNVs conferring risk of autism or schizophrenia affect cognition in controls," lead authors Hrein Stefansson and Andreas Meyer-Lindenberg focused on copy number variations (CNVs) in this population that had been previously associated with either schizophrenia or autism spectrum disorder (ASD). (CNVs were also a particular focus of the two papers released this Wednesday, as large CNVs across the genome have been some of the genetic variants most reliably linked to a heightened risk of schizophrenia.)
Using deCODE's bank of Icelandic genotypes, the researchers identified over 1,000 individuals carrying one of 26 very rare CNVs related to schizophrenia or ASD. More than 150 people carrying these CNVs, but not diagnosed with schizophrenia, agreed to take part in a series of cognitive and psychiatric tests. For comparison, the researchers administered the same tests to three other groups: one of individuals with schizophrenia; one set of controls carrying CNVs, but not from the set of 26 CNVs being studied; and one set of controls with no CNVs. The tests included measures of cognitive functions like spatial memory, verbal IQ, and cognitive flexibility, and measures of psychological and social health like the Global Assessment of Function (GAF) and mini international neuropsychiatric interview (MINI) tests.
The schizophrenia group, as expected, deviated significantly from the controls in almost every category. The core finding of the paper, however, was that the group carrying schizophrenia-associated CNVs consistently fell in between the schizophrenia group and the controls. In particular, this group's GAF scores, and evaluations of depression and suicidal ideation from the MINIs, were significantly worse than those of the controls. Meanwhile, the two control groups differed very little from each other, and almost never to a statistically significant extent, indicating that this was not an effect of CNVs in general, but of the 26 CNVs under discussion in particular. (A few of those CNVs had large enough sample sizes to be analyzed individually.) Effect sizes were smaller after controlling for general IQ – which was also impaired in the schizophrenia-associated CNV group, and known to affect many of the same measures of social and psychiatric health – but many remained significant.
This research lends some important context to the papers Bio-IT World reported on Wednesday. It lends some support both to the notion that there is significant overlap between the pathologies of ASD and schizophrenia, and that small, additive, partially-penetrant mutations can contribute to a schizophrenia spectrum that does not always result in the full clinical disorder. Schizophrenia remains one of the most complex conditions to study genetically, but large-scale genomic studies are starting to converge on a few common themes.