29 Oct 2014
October 29, 2014. Because the genetics of other disorders may provide beacons for psychiatric disease, Margaret Pericak-Vance of the University of Miami, Florida, told the story of genetic discoveries for Alzheimer's disease in a lecture on Tuesday, October 14, which marked her receipt of the Ming Tsuang Lifetime Achievement Award. The story includes a mixed bag of variants, including common variants of large effect (APOE4, which she discovered), newer common variants of small effect, as well as rare variants revealed by sequencing. The last type have compelled researchers to get “back to basics” by analyzing family pedigrees to see how rare variants segregate with disease, she said.
In a plenary that afternoon, Preben Bo Mortensen of the University of Aarhus, Denmark, highlighted the Danish civil registration system so beloved of epidemiologists and geneticists, which has turned the entire population into a cohort. These data provide the core ingredients for iPSYCH, a three-year initiative funded by Lundbeck to explore the genetic and environmental causes of mental disorders. One effort is a phenomics project, in which children of people with schizophrenia or bipolar disorder will be enrolled and monitored over time with detailed measures of cognition, brain imaging, environment, and genetics to get at the mix of factors that put these children at risk.
Mortensen also made the point about how more refined information can affect genetic signals, using as an example the Danish sample contributed to the Psychiatric Genomic Consortium’s (PGC) schizophrenia genomewide association study (GWAS) (see SRF related news report). The Danish sample showed the same pattern as other samples, with odds of having schizophrenia increasing with polygenic risk score, though it achieved smaller levels of risk. Mortensen showed how this risk signal could be boosted by restricting controls to those who lacked any family history of schizophrenia, and by including only people who were hospitalized more than once.
One important, and tragic, phenotypic difference among those with schizophrenia is whether a person responds to antipsychotic drugs. Those who do not get relief from their symptoms have “treatment-resistant schizophrenia” (TRS), accounting for one-third of all those with schizophrenia. These people are the most impaired, with fewer than 5 percent employed. On Wednesday afternoon, October 15, James Walters of Cardiff University, Wales, United Kingdom, presented his work in trying to determine whether TRS is a distinct genetic subtype of schizophrenia or a more severe form of the condition. As a TRS sample, he used the CLOZUK sample, which consists of over 6,000 anonymous blood samples from people who use clozapine, a last resort antipsychotic for those who do not respond to other drugs.
Compared to other samples within the PGC’s recent schizophrenia GWAS, the CLOZUK sample had a comparable burden of polygenic risk, suggesting that TRS is not just a severe form of the disorder. When Walters conducted a GWAS with CLOZUK providing these cases and the rest of the PGC cases and controls, some genomewide-significant signals fell out, including some already known to schizophrenia research, such as CACNA1C and even DRD2. Several other new signals also materialized, suggesting there may be specific genes for TRS. Looking at rare variation, Walters did not find TRS-specific signals among copy number variants (CNVs).
Tuesday afternoon, October 14, David Glahn of Yale University, New Haven, Connecticut, outlined how just a few pedigrees containing cases of schizophrenia can be sufficient to pull out endophenotypes—heritable traits that are correlated with a disorder (Glahn et al., 2014). Endophenotypes could help find genetic risk factors for a disorder because the traits could be closer to the action of genes than a categorical diagnosis based on varied behavioral symptoms. Glahn described how among randomly collected Mexican-American pedigrees, only six cases of schizophrenia were found, but these came from large enough families to give 233 related but unaffected people. Using an analysis that took into account each person’s risk for schizophrenia based on relatedness to the person with schizophrenia, Glahn reported pulling out three cognitive endophenotypes, including facial memory and emotion recognition, and neuroanatomical endophenotypes, consisting of the surface areas of six different cortical regions.
A different kind of family matter came up in the discussion of de novo variants—those genetic mishaps that are not inherited from parents but rather arise spontaneously, possibly in sperm or egg cells. De novo mutations rise with age (see SRF related news report), and de novo mutations increase risk for psychiatric disorders. Some have proposed that an increased rate of de novo mutations in sperm from older dads could explain the epidemiological findings that older fathers are more likely to have children with schizophrenia (see SRF Hypothesis). But Peter Visscher of the University of Queensland, Brisbane, Australia, did not find evidence for this by using population genetics models that took into account the rate of age-related de novo mutations and the relative risk to children with older fathers. Instead, he found the data could be best explained by inherited genetic risk factors—i.e., those shared by both dad and offspring. This liability could complicate a person’s social life and possibly prolong the time it takes for a father to reach the married-with-children stage of life.
Daniel Howrigan of Massachusetts General Hospital in Boston gave an update on de novo mutations in schizophrenia, as detected by exome sequencing a person with the disorder and each parent. Combining data from a Taiwanese cohort with previously published data to give over 2,000 schizophrenia trios and nearly 700 control trios, he reported no difference in the rate of de novo mutations in schizophrenia compared to controls. There was a slight enrichment in schizophrenia for loss-of-function mutations predicted to disable a protein’s function, but of those genes hit multiple times, none reached exome-wide significance. Still, 11 genes had two or more loss-of-function mutations, which was more than expected by chance. Howrigan suggested that de novo mutations might not have as strong a role in schizophrenia as they do in autism or intellectual disability, but that more sequencing would likely implicate specific genes.—Michele Solis.