Studies Tackle Cognitive Endophenotypes in Big Ways
1 October 2010. Two studies in the September Archives of General Psychiatry took fresh approaches to using cognitive measures to explore the genetic roots of schizophrenia. Hannelore Ehrenreich, Nils Brose, and colleagues at the Max Planck Institute in Göttingen, Germany, conducted a large phenotype-based genetic association study that tied six single-nucleotide polymorphisms in complexin 2, a gene that encodes a synaptic protein, to cognition in schizophrenia. They also showed that one of the polymorphisms regulates gene expression and, in mice, does so by affecting microRNA binding. In a combined family and twin study, a team led by Timothea Toulopoulou at King's College London created what may be the largest international database of cognition in schizophrenia. Genetic modeling revealed that cognition and schizophrenia share many of the same genetic underpinnings, but genes unrelated to cognition also contribute a great deal to schizophrenia heritability.
These studies reflect concerns that the binary concept of schizophrenia glosses over so much heterogeneity that it hampers the search for relevant genes. To avoid this problem, both groups focused on endophenotypes, measurable, heritable disease-related traits thought to lie intermediate in the pathway from genes to disease (see SRF related news story; SRF news story; SRF news story; SRF news story; also see SRF Live Discussion). In particular, they examined the cognitive deficits that often accompany schizophrenia, that signal approaching psychosis, and that can even occur in nonpsychotic relatives of people with schizophrenia (Reichenberg and Harvey, 2007; and see SRF Interview With Philip Harvey).
Focus on the phenotype
Ehrenreich, Brose, and co-first authors Martin Begemann and Sabrina Grube took a different approach than genome-wide association studies (GWAS) do. Instead of seeking major schizophrenia genes, which might not even exist, they studied one gene’s relationship to specific cognitive traits. They focused on complexin 2 (CPLX2, see SZGene entry), one of a family of genes whose proteins govern the functioning of synapses and neural networks. Studies have found altered complexin expression in various psychiatric and neurological disorders, including schizophrenia (reviewed in Brose, 2008).
Begemann and colleagues conducted their study as part of their larger effort to thoroughly characterize the phenotypes of patients with schizophrenia and to assess the contributions of candidate genes to those phenotypes. The research team traveled to psychiatric hospitals throughout Germany, collecting data on 1071 patients with schizophrenia. The resulting Göttingen Research Association for Schizophrenia (GRAS) database contains over 3000 data points per subject, including information on disease history, treatment, comorbidities, environmental risk factors, family members, and neuropsychological functioning.
In case-control analyses, no single genetic marker distinguished patients from 1079 healthy subjects, but a haplotype in block 2 did. This haplotype, which included SNPs rs1366116 and rs3892909, occurred in 2.3 percent of cases but only 0.9 percent of controls (odds ratio = 2.47, 95 percent CI, 1.41 to 4.34, P <.001).
In contrast, phenotype-based genetic association analyses tied six out of 11 CPLX2 SNPs to performance on tests of executive functioning, reasoning, or verbal learning and memory. Of all the haplotype combinations, the CTC sequence at markers rs1366116/rs3892909/rs3822674 predicted the worst cognition. None of the SNPs themselves correlated with premorbid intelligence.
Another research group had found overt cognitive deficits in CPLX2-knockout mice only after a “second hit”—in that case, the stress of being separated from their mother (Yamauchi et al., 2005). This made Begemann and colleagues think that the cognition-relevant SNPs might impair cognition only in the presence of environmental stress during puberty. To model the stressor, they used a chemically cooled copper cone to cause a mild brain lesion. This worsened the spatial memory performance of Cplx2-null mutant mice, but not their wild-type littermates. Nonlesioned mutants performed normally, backing the two-hit theory.
Searching for a mechanism for their findings, Begemann and colleagues observed that one of the implicated SNPs, rs3822674, lies in a noncoding gene region that is thought to bind to microRNA 498. They further found that only the T allele permits this binding, thereby curbing CPLX2 expression. In human blood cells, they tied specific genotypes at this SNP to CPLX2 messenger RNA expression.
These findings hint that CPLX2 variants shape cognition in schizophrenia by regulating gene expression. However, they might also alter cognition in people with other neuropsychiatric disorders or those with good health, according to the researchers. Even slight up- or down-regulation of CPLX2 could disrupt its ability to hone synaptic function.
In the other Archives study, Toulopoulou and colleagues put cognitive endophenotypes for schizophrenia to the test, combining data from several large data sets collected at the U.S. National Institute of Mental Health by Daniel Weinberger's group, at the Institute of Psychiatry in London, under the leadership of Robin Murray, and at Harvard University by Larry Seidman and colleagues. Endophenotypes should not only be heritable, they also must share genetic roots with the disorder, and the researchers wanted to see how well certain cognitive measures meet these criteria. An earlier twin study from Toulopoulou’s lab (see SRF news story) had pointed to intelligence and memory as particularly promising schizophrenia endophenotypes. This time, the researchers used a more powerful family-and-twin design to evaluate them.
Data came from 657 patients with schizophrenia, 674 first-degree nonpsychotic relatives, including co-twins, and 725 healthy controls. The research team tested subjects’ intelligence, visual memory, and verbal memory and learning. Memory tests included checks of both immediate and delayed recall.
Genetic modeling estimated how much each endophenotype reflected genetic effects, environmental effects shared by family members, and environmental effects that differed within the family. All 3 candidate endophenotypes—intelligence, immediate recall, and delayed recall—cleared the heritability hurdle. Genes accounted for 66 percent of the variance in both intelligence (95 percent CI, 0.62 to 0.85) and immediate recall (95 percent CI, 0.62 to 0.71); for delayed recall, they explained 48 percent (95 percent CI, 0.42 to 0.55). Most of the remaining individual differences in cognition came from personal, rather than familial, environmental influences.
Correlations between schizophrenia and each of the endophenotypes ranged from -0.35 to -0.38; the researchers wondered how much these associations reflect genetic overlap. Genetic modeling provided some answers, by divvying up the covariance between schizophrenia and the cognitive phenotypes into the parts associated with genes, unique environment, and family environment. It revealed that cognition and schizophrenia genetically have much in common.
Genes explained 89 percent of schizophrenia’s covariance with intelligence; they accounted for 72 and 86 percent, respectively, of its correlation with immediate and delayed recall. In contrast, the environmental factors that influence cognition seemed to differ from those involved in schizophrenia.
According to Toulopoulou and colleagues, their findings validate intelligence, immediate recall, and delayed recall as schizophrenia endophenotypes, but do not favor one over another. While genes appeared to strongly contribute to their associations with schizophrenia, Toulopoulou cautioned that the explained correlations are modest to begin with. She and her colleagues wrote, “More than 50 percent of the genes that affect liability to schizophrenia do not affect cognition, and therefore, the genetics of schizophrenia are more than the genetics of cognition.”
Unlike the Begemann study, this one did not look at gene-environment interactions that might influence the endophenotypes, although Toulopoulou plans to do so and would welcome collaborators. Ehrenreich and Brose told SRF that they have a list of genes to study in relation to disease traits using the GRAS database. They, too, see opportunities for collaboration, especially with researchers who do GWAS but want to know more about subjects’ phenotypes than just whether or not they have schizophrenia.—Victoria L. Wilcox.
Begemann M, Grube S, Papiol S, Malzahn D, Krampe H, Ribbe K, Friedrichs H, Radyushkin KA, El-Kordi A, Benseler F, Hannke K, Sperling S, Schwerdtfeger D, Thanhäuser I, Gerchen MF, Ghorbani M, Gutwinski S, Hilmes C, Leppert R, Ronnenberg A, Sowislo J, Stawicki S, Stödtke M, Szuszies C, Reim K, Riggert J, Eckstein F, Falkai P, Bickeböller H, Nave K-A, Brose N, Ehrenreich H. Modification of cognitive performance in schizophrenia by complexin 2 gene polymorphisms. Arch Gen Psychiatry. 2010 Sep;67(9):879-88. Abstract
Toulopoulou T, Goldberg TE, Mesa IR, Picchioni M, Rijsdijk F, Stahl D, Cherny SS, Sham P, Faraone SV, Tsuang M, Weinberger DR, Seidman LJ, Murray RM. Impaired intellect and memory: a missing link between genetic risk and schizophrenia? Arch Gen Psychiatry. 2010 Sep;67(9):905-13. Abstract