16 September 2011. Over 500 participants from around the world gathered at the Omni Shoreham Hotel in Washington, DC, on September 11 for the first day of 19th World Congress of Psychiatric Genetics, marked by organizer Francis McMahon of the National Institute of Mental Health with a moment of silence for those who had died in the 2001 terror attacks. His co-organizer, Thomas Schulze of the University of Göttingen, Germany, provided a second, and more hopeful, image for the attendees: that of Bill Clinton playing his saxophone in the very same room at one of his inauguration balls. This set the stage for researchers to present their latest findings as they seek a way to move psychiatric genetics forward.
A plenary talk by Seth Grant of the Wellcome Trust Sanger Institute in Cambridge, United Kingdom, highlighted the potential for synapse dysfunction in brain disease (see SRF related news story). Specifically, he discussed how genes encoding components of a multi-protein complex consisting of N-methyl D-aspartate (NMDA) receptors, adhesion proteins, scaffolding molecules, and enzymes (called “MASC” for MAGUK-associated signaling complex) appear to be involved in 50 different diseases, including schizophrenia. Removing components of this so-called “information processor” influenced cognitive function in mice, and Grant noted that similar changes might play a role in schizophrenia.
In her plenary talk, Suzanne Leal of Baylor University, Houston, Texas, then offered guidance to researchers who are trying to make sense of the rare variants turning up in the gathering storm of data from next-generation sequencing, and soon, exome chips. She compared current methods for detecting associations with rare variants, finding each had its own advantages and drawbacks, and differed in computing efficiency.
The CNV club grows
A Sunday afternoon symposium devoted to copy number variations (CNVs)—the wholesale loss or gain of a segment of DNA—in schizophrenia unveiled some new loci, and replicated others. As evidence for CNVs in schizophrenia adds up (see SRF related news story), researchers have looked to these events as a way to get closer to the genes disrupted in the disorder. All presenters focused on de-novo CNVs—those not inherited from parents, but occurring anew in the germline. These offer one way to explain “sporadic” cases in families without a history of schizophrenia.
George Kirov of Cardiff University in the United Kingdom began by describing the CNVs found in a Bulgarian sample of 638 families, consisting of 662 individuals with schizophrenia and both parents. Excluding CNVs shared with a parent, they arrived at 34 de-novo events in 5.1 percent of individuals with schizophrenia. This was twice as many as detected in an Icelandic sample of controls (2.1 percent). These CNVs landed in loci already implicated in schizophrenia, including 3q29, 15q11.2, 15q13.3, and 16p11.2, but a new one also turned up in the form of a duplication at 7q11.23. Deletions of this region result in the hypersociable Williams syndrome, while duplications of the region have been found recently in autism (Sanders et al., 2011). Kirov also noted that de-novo CNVs are larger than inherited ones in both controls and in schizophrenia—something that suggests they are under high negative selection, and could reflect a pathogenic nature. Kirov's Cardiff colleague Michael Owen picked up the story here, delving into the genes disrupted by the CNVs in this study. Two newcomers included EHMT1, which encodes a histone methyltransferase, and DLG2, which encodes a protein involved in the aforementioned MASC complex; in Drosophila versions of these genes, EHMT1 regulates DLG2. Gene set analysis found 12 of 34 de-novo CNVs hit genes encoding synaptic proteins, particularly those belonging to MASC. This represented an enrichment compared to the de-novo events occurring in controls, and further implicates the synapse as a locus of disruption in schizophrenia.
Dheeraj Malhotra from the University of California in San Diego presented the first evidence for a role of CNVs in bipolar disorder, and added to the evidence for their role in schizophrenia. Taking a different tack from previous CNV studies of bipolar disorder, he looked for de-novo CNVs by examining the genomes of 185 individuals with bipolar disorder, 177 with schizophrenia, and 426 controls, plus both parents in each case. This turned up 23 de-novo CNVs, with an average size of 100 kb. The frequency in bipolar (4.3 percent) was similar to that found in the schizophrenia group (4.5 percent), but higher than controls (0.9 percent). The rate was especially high in early-onset cases of bipolar disorder, beginning before 18 years of age. Interestingly, one CNV found in a bipolar case hit LRRTM2, which encodes a receptor of neurexin (see SRF related news story). Together, these molecules mediate glutamatergic synaptic formation, so the finding further implicates the synapse in psychiatric disorders.
In closing the session, Dan Rujescu of Ludwig-Maximilians University, Munich, Germany, reminded the audience that CNVs implicated in schizophrenia tend to come with baggage, like intellectual disability, seizures, anatomical malformations, or developmental delay. He cautioned that these CNVs cause a more severe phenotype than that seen in the average schizophrenia patient. This is important to keep in mind when following up the genes hit by a CNV: are they inherent to the disorder, or involved in these other, secondary phenotypes?
Fumbling toward function
Other researchers are taking the next step with established risk alleles for schizophrenia to pinpoint their function in some context, be it at the level of cell biology or of brain imaging and behavior. Taking the latter approach, Gary Donohoe of Trinity College in Dublin gave a sobering account of the complexities in finding consistent associations between cognitive measures and a common variant in ZNF804A (rs1344706), which has been identified by several genomewide association studies (GWAS) of schizophrenia. Previous work associated the risk allele with preserved cognitive function in a subset of individuals with schizophrenia (Walters et al., 2010) and intact gray matter and white matter volume (Donohoe et al., 2011). In contrast, the risk allele was deleterious in an EEG study of P300, a brain response reflecting working memory and attention: carriers of the risk allele had normal P300 compared to those homozygous for the non-risk allele. And yet, an opposite pattern emerged in a different sample; these same samples had been used in the previous studies linking the risk allele with preserved function. Apparently, pinning down a consistent function for even well-established variants may not be straightforward.
Switching gears, Caroline Tinsley of Cardiff University presented a single-cell view of the function of TCF4, another GWAS-established risk factor for schizophrenia. As a transcription factor, TCF4 regulates gene expression during development. In an interesting point of convergence, Tinsley found that, in vitro, TCF4 activates expression of NRXN1B and CNTNAP2—two other genetic risk factors for schizophrenia and autism. With the same cell biology perspective, Matt Hill of King’s College London examined the function of ZNF804A in terms of transcription, as the risk variant is found in a non-coding region of the gene. He recently reported that the risk allele binds 46 percent less nuclear protein than the non-risk allele, which could interfere with its proper transcription (Hill and Bray, 2011). In postmortem brain, the risk allele had no effect on ZNF804A expression in adult samples, but it did reduce the peak of ZNF804A expression normally observed during the second trimester of brain development. Together with findings that knocking down ZNF804A altered expression levels of other genes, particularly those involved in cell adhesion, these data support the idea that vulnerability for schizophrenia has its roots in early brain development. Overall, the findings leave one with the sense that finding the genes relevant to schizophrenia marks only the end of the beginning for understanding the disease process.—Michele Solis.