In this final report from the 2011 World Congress on Psychiatric Genetics in Washington, DC, we provide summaries of a potpourri of talks related to schizophrenia from different sessions.
7 November 2011. In an epigenomics session on Sunday afternoon, 11 September 2011, Karolina Aberg of Virginia Commonwealth University in Richmond presented data from a whole-methylome study of patients with schizophrenia. This represents a big advance over previous work, which has focused on the regions—CpG "islands" or "shores"—where methylation sites are found in high concentrations. But Åberg said that more than 90 percent of CpG sites are located outside these regions. In a two-phase study, Åberg and her colleagues identified methylation differences in the introns of a number of genes (NF1A, FNDC3B, DCTN2, AKAP5, FTSJ2, GPHN, and RLBP1). Because they were by necessity using blood cells, they also did side projects in mice to see what difference there might be between blood and brain. Åberg reported that there is higher methylation in the hippocampus and cortex than in blood cells from mice, but there is greater than two-thirds concordance, and only 1 percent of sites methylated in blood are not methylated in brain.
In another Sunday afternoon session, this one on phenotypes and endophenotypes, Miles Hamshere of Cardiff University, United Kingdom, discussed an association study in bipolar patients that was not genomewide, but did sample immune system genes (using the Illumina ImmunoChip), as well as some of psychiatric research interest. Looking at the single nucleotide polymorphisms (SNPs) with highest scores from the Psychiatric Genetics Consortium schizophrenia sample, Hamshere and colleagues find preliminary evidence for bipolar association with CACNA1C, TCF4, as well as the genes PTPRG, PIK3C2A, and PLCB2. The direction of effect for these SNPs in the bipolar sample was generally similar to that seen in the schizophrenia study.
He was followed by Tristam Lett of the Centre for Addiction and Mental Health, Toronto, Canada, who presented a study of the effects of SNPs in the gene for neurexin-1 on brain morphology, sensorimotor function, and response to clozapine. In addition to evidence that disruption of the gene can contribute to schizophrenia, neurexin has the additional point of interest that it regulates NMDA receptor movement. Lett and colleagues find that one of the SNPs they looked at (rs1045881) is associated with differences in white matter volume between schizophrenia patients and controls, as well as differences in finger tapping speed and response to clozapine.
In the last talk of the session, Astrid Rauch presented data from National Institute of Mental Health, Bethesda, Maryland, on the gene for neurogranin, which has been implicated in schizophrenia risk by a genomewide association study (GWAS) and is also linked to the NMDA receptor by its molecular activity in neurons. The researchers found the disease-linked SNP altered the function of dorsolateral prefrontal cortex of control subjects during a working memory test, which they see as indicative of inefficient cortical information processing.
On Monday afternoon, 12 September, Elvira Bramon of the Institute of Psychiatry kicked off a session on candidate genes from GWAS and single-gene association studies by describing a second-round GWAS of the Wellcome Trust Case Control Consortium, specifically of psychosis and related endophenotypes. Drawn from a number of European countries, the case-control portion of their study (>1,200 cases with psychotic illness) did not reveal any hits with genomewide significance. However, Bramon mentioned an upcoming family-based analysis which, when combined with the case-control sample, will provide more data.
In the same session, Alan Sanders of the NorthShore University Health System, Chicago, Illinois, presented data from gene expression in lymphoblastoid cell lines (LCLs) derived from patients with schizophrenia. The transcripts detected in the cells were enriched for genes expressed in brain, Sanders said, and he noted that the LCLs have high-quality RNA and are relatively far removed from environmentally induced epigenetic effects. The group found a significant number of transcripts that were differentially expressed in cases versus controls, and in particular, some of these were in the MHC region of chromosome 6 found in previous GWAS, including histone genes. The differentially expressed transcripts were enriched for brain expression genes, and to a lesser degree, for immune genes.
Edwin van den Oord of Virginia Commonwealth University in Richmond concluded the session with a meta-analysis of 18 previous GWAS (of ~22,000 subjects), which included integration of the findings with "disease informative" databases (including SZGene) of association, expression, and other data. The "enriched" dataset of "most promising" SNPs determined in this way was then studied in a separate family-based replication sample of >6,300 subjects. The use of the database information was deemed to have been informative, van den Oord reported, by virtue of the fact that the SNPs identified were as likely to replicate as those that ranked highest in P values based on the meta-analysis alone. A handful of novel genes were tentatively implicated, and markers in previously associated genes TCF4 and NOTCH4 were replicated. Pathway analyses on the many small-effect hits pointed to immune system processes (antigen processing, cell adhesion molecules relevant to T and B cells) and neuronal function (cholinergic synapse, cell adhesion, MAPK signaling).—Hakon Heimer.