19 Oct 2015
October 20, 2015. The 23rd annual World Congress of Psychiatric Genetics got underway October 16 in refreshingly chilly Toronto, Canada. Conference Chair James Kennedy of the University of Toronto welcomed 650 attendees from 42 countries to the sumptuous Fairmont Royal York Hotel and offered up a hearty stew of psychiatric genetics, which held its own against the potent distractions of various sporting events and a Canadian election.
Aswin Sekar from Harvard Medical School in Boston, Massachusetts, kicked off a schizophrenia session on Saturday morning with more tantalizing news about the complement component 4 (C4) gene work coming from Steven McCarroll's lab (see SRF related conference report and SRF conference report). Located within the complicated major histocompatibility complex (MHC) region long associated with schizophrenia, the C4 gene, in a particular form, seems to drive risk for schizophrenia, and its expression is higher in postmortem brains of those with schizophrenia compared to controls. Sekar also reported that C4 appears in synapses, with C4 peak expression in mouse visual cortex early in postnatal development, a time when synapse elimination begins. This suggests that too much C4 drives pathological synapse loss in schizophrenia.
In the next talk, from Mads Engel Hauberg of Aarhus University in Denmark, attention turned to microRNAs (miRNAs), the small, unassuming, non-coding RNAs that powerfully modulate gene expression. Genes related to schizophrenia are no exception, with Hauberg reporting that combining a database of predicted targets of miRNAs with the results from the Psychiatric Genomics Consortium's (PGC) genomewide association study (GWAS) of schizophrenia highlighted two: miR-137 and miR-9-5p, both of which have turned up in GWAS themselves (see SRF related news report). Though these two both had 231 gene targets in common, this was not a significant overlap.
Data presented by Jennie Pouget of the University of Toronto suggests little genetic overlap between schizophrenia and autoimmune diseases, despite epidemiological findings linking the two (e.g., Benros et al., 2014; see also SRF related news report). Pouget compared single nucleotide polymorphisms (SNPs) associated with 18 different autoimmune diseases (and which are outside of the MHC region) with those from the PGC's latest schizophrenia GWAS. The resulting overlap was not greater than expected by chance in aggregate, but five individual autoimmune SNPs were significant, including one implicating PLCL1—a phospholipase C-like enzyme involved in calcium signaling. Pouget also did not find evidence for a subgroup of people with schizophrenia carrying a high burden for autoimmune disease risk alleles. In the question period, she said that including the MHC region SNPs did not change the gist of her results.
Peter Holmans of Cardiff University followed with a sequel to the pathway analysis he and colleagues published for the first round of PGC GWAS data for schizophrenia (see SRF related news report). Pathway analysis tries to see whether a collection of SNPs fall preferentially on places in the genome that are biologically related. When applied to the PGC2 dataset with 108 genomewide-significant loci, the analysis pointed to 28 different pathways of the 9,000 considered. Top pathways included dopamine synapses, the post-synaptic density, calcium signaling, and those having to do with seizures. The current results were similar to the earlier results, except that the histone methylation pathway dropped out of the new analysis.
Daniel Howrigan of Massachusetts General Hospital in Boston gave an update on the PGC's analysis of its GWAS dataset for copy number variants (CNVs), the deletions or duplications of chunks of DNA that confer risk for a range of psychiatric disorders. Putting together a centralized pipeline to identify CNVs in data from over 41,000 subjects consisting of schizophrenia cases and controls identified plenty of CNVs. For the most part, these fingered CNV loci previously identified as increasing risk for schizophrenia. Still, a significant burden of new CNVs, mainly consisting of extremely rare events, was noted. Among the new findings, Howrigan highlighted
a CNV hotspot on the X chromosome that had a mix of protective and risk alleles for schizophrenia.
The same CNVs often increase risk for schizophrenia and other disorders such as autism or intellectual disability. One reason for this variable diagnostic expressivity may lie on the intact counterpart chromosome; a functional variant there may make it hard for that piece of the genome to compensate for the CNV. This "double hit" scenario was explored for schizophrenia by Jacob Vorstman of the University of Utrecht, the Netherlands, in a Dutch sample of over 1,200 subjects. Though double hits were found, the rate of a mutation in the sequence of a corresponding CNV did not significantly differ between cases and controls. However, Vorstman did report that cases had an increased burden of deletion CNVs plus mutations deemed deleterious compared to controls.—Michele Solis.