1 May 2012. More than 1,600 researchers and clinicians gathered in a rainy Florence from 14-18 April 2012 to discuss new schizophrenia research at the Third Schizophrenia International Research Society (SIRS) conference. Despite the clouds, the combination of espresso and data kept things animated in talks ranging from psychosocial treatment strategies and immune system interactions to the latest genetic puzzle pieces pouring from the sequencing machines.
Monday’s plenary speaker, Bill Byerly of the University of California, San Francisco, outlined the landscape of schizophrenia genetics, beginning with the linkage studies of yore, traversing the common variants detected by genomewide association studies (GWAS), moving on to the suspected copy number variations, and ending with sequencing, which he pointed out will rely on family data to help make sense of the rare variants turning up. He and other speakers emphasized that an assortment of variants—both rare and common—would contribute to schizophrenia risk.
Maria Karayiorgou of Columbia University in New York focused on the contributions made by rare “de-novo” events—spontaneously arising genetic glitches not inherited from parents. Though these may not explain schizophrenia’s high heritability, they could account for the stable prevalence of schizophrenia, despite the fact that individuals with the disorder tend not to have children. Karayiorgou reviewed her evidence for a higher burden of de-novo CNVs (see SRF related news story) in schizophrenia. In her view, these rare variants contribute to brain disease in combination with secondary factors, such as other rare mutations, common variants, environment, and chance—which would color the outcomes of these de-novo events.
Spotlight on sequencing
With the stage set, a genetics symposium the same afternoon featured several presentations of sequencing work in progress. Despite its ability to resolve rare, single nucleotide anomalies, sequencing seems to be complicating more than clarifying the picture of psychiatric genetics so far. That’s because sequencing finds variants galore in cases and controls alike—even rare, protein-altering ones—leaving researchers to figure out which of the many needles in the haystack contribute to risk for disease (see SRF related news story). It was clear at this symposium that many researchers are looking for direction from family studies to determine how to weigh the evidence.
David Porteous of the University of Edinburgh, Scotland, used a family approach to sort potentially causal variants from bystanders in a targeted resequencing study of disrupted in schizophrenia 1 (DISC1). With colleagues at Cold Spring Harbor Laboratory, Porteous and his team have been resequencing a 528 kb region of the DISC1 gene disrupted by a translocation found in a Scottish family beset not only by schizophrenia, but also by major depressive disorder (MDD) and bipolar disorder (BD). Sequencing this region in over 1,500 individuals (240 with schizophrenia, 221 with BD, 192 with MDD, and 889 healthy controls) turned up 2,010 rare variants, mostly located in the exome. One of these was R37W, a rare, protein-altering variant recently found to disrupt DISC1’s action in the nucleus (Malavasi et al., 2012). Porteous reported finding R37W in recurrent MDD, and it segregated with this illness in three families. The group is still sorting out whether any of these rare variants are specifically associated with schizophrenia.
How best to use sequencing depends on the genetic architecture of the disease, which researchers are still grappling with, said Richard McCombie of Cold Spring Harbor Laboratory, Long Island, New York. He stressed the difficulty presented by the overabundant variants when describing an ongoing case-control study of BD in his lab, saying, “This is a real signal-to-noise problem.” For example, sequencing all the synaptic genes of the genome (the “synaptome”) of 186 individuals revealed that 4,000 genes had a non-synonymous (amino-acid altering) variant in each person. While focusing on gene function might whittle these down, McCombie outlined another approach that first targets linkage regions for resequencing in families, finds the recurrent variants that crop up in other families, then follows those up in large case-control studies of about 20,000 samples total. He is pursuing this strategy for a region of chromosome 4 fingered by linkage in families with BD and MDD.
Shane McCarthy of Cold Spring Harbor Laboratory also presented the family approach in his account of exome sequencing trios consisting of an individual with schizophrenia and both parents. In 33 trios, he reported 138 de-novo variants, including 79 non-synonymous ones and 10 misplaced start or stop codons. Intriguing genes suggested by the predicted-to-be-damaging variants included TRAPPC9, a trafficking protein associated with mental retardation, and MECP2, the master transcription regulator known for its role in Rett syndrome. In collaboration with Byerly, McCarthy is also pursuing whole-genome sequencing in an extended family pedigree. To start to make sense of the three million variants found, he is focusing on those landing in linkage peaks. Five individuals had 126 non-synonymous variants in common, and one gene highlighted by these is DCDC2, already associated with neuronal migration and dyslexia.
Wanting to get at the rare variants not captured by genomewide association studies (GWAS), Shaun Purcell of Mount Sinai School of Medicine in New York described an exome-wide, gene-based association study of 1,500 Swedes, consisting of cases of schizophrenia and controls. This produced 290,000 single nucleotide variants, and one gene hit more than expected was NRG1, a favorite schizophrenia candidate in some quarters. Five different non-synonymous variants, both rare and common, were found in NRG1, and these were overrepresented in cases (n = 12) compared to controls (n = 1). Looking beyond this gene-based perspective, he also tried a functional pathway analysis to see if certain sets of genes carried more of these variants than expected. This showed that targets of miR-137, a transcriptional controller highlighted in the largest schizophrenia GWAS to date (see SRF related news story), are enriched for a burden of rare, non-synonymous coding variants in schizophrenia. Purcell also described ongoing exome sequencing of schizophrenia trios from a Bulgarian sample, reporting 309 de-novo variants enriched for synaptic proteins in 354 trios.
In an extreme version of the family approach, some are looking to the rare cases of monozygotic twins discordant for schizophrenia in the hopes that any differences would quickly divulge the genetic culprits. In a student-delivered symposium Tuesday afternoon, Christina Castellani of the University of Western Ontario in London, Canada, described preliminary findings of the whole-genome sequencing of two such pairs of twins, and one set of parents. She estimates these twins differ in 3 percent of their genome—more than expected, and representing thousands of variants to follow up—but still substantially less than the 75 percent difference obtained for two unrelated people.—Michele Solis.