6 November 2012. On Thursday, 18 October, the 20th World Congress of Psychiatric Genetics, held in Hamburg, Germany, ended with two plenary talks that left attendees with a lot to ponder. Trevor Robbins of the University of Cambridge, United Kingdom, argued that the field ought to pay careful attention to phenotypes, and their component parts (“intermediate phenotypes” or, if inherited, “endophenotypes”), because these may break psychiatric diseases into more tractable pieces, leading to clearer diagnoses, cleaner genetic studies and clinical trials, and earlier detection. Based on his own research linking impulsivity to compulsive drug use in rodents (Robbins et al., 2012), Robbins argued that it is helpful to develop different ways to measure the same phenotype (like impulsivity), that these different measures may detect different stages of disease, and that response to drugs is a good phenotype because it is close to biology. He suggested that schizophrenia’s clinical features may be deconstructed into discrete cognitive impairments, and that aberrant learning in general may underlie the positive and negative symptoms defined by psychiatry.
James Dee Higley of Brigham Young University in Provo, Utah, finished with a meditation on how the environment can modulate the effects of a genetic variant. Drawing from his studies of alcohol abuse in a population of monkeys monitored from birth onward, he noted how early childhood experience altered the influence of serotonin-related genotypes on various phenotypes, including alcohol intake.
Concluding the meeting, organizer Markus Nöthen of the University of Bonn noted the progress made since the meeting’s inception 20 years ago, and congratulated everyone, from speakers to poster presenters, for their scientific contributions this year. Next year’s meeting is slated for 17-21 October in Boston, Massachusetts.
Here, SRF brings you a sampling of other conference news, including hints of rare forms of genetic variation at work in schizophrenia and attempts to determine the function of some of the genes coming into focus.
A menagerie of variation
Even penetrant variants like copy number variations (CNVs) are associated with a spectrum of phenotypes, making it hard to connect the dots between genes and behavior. For example, deletions within the neurexin 1 gene (NRXN1) have been associated with schizophrenia, autism, and intellectual disability (Schaaf et al., 2012). But tracking CNV occurrence in families may help get a handle on this kind of pleiotropy, according to Linh Duong of the University of Oslo in Norway, who suggested in a talk on Monday, 15 October, that the outcome may depend on the context in which a mutation appears. She described a compound heterozygous mutation of NRXN1 found in a person that consisted of a 451 kb NRXN1 deletion inherited from his mother, and a protein-truncating point mutation within NRXN1, likely inherited from his father. This person had intellectual disability, autism, and epilepsy, whereas other family members who carried only one of these mutations had less severe symptoms: his mother had subclinical signs of autism, and a brother carrying the point mutation but not the deletion had a psychotic disorder.
Citing recent evidence, in a talk on Monday, for contributions by recessive alleles to schizophrenia (Keller et al., 2012), Douglas Ruderfer of the Mount Sinai School of Medicine in New York City described his efforts to locate instances of homozygous mutations relevant to schizophrenia. Using an exome chip to identify variants in a Swedish sample of schizophrenia cases and controls, he reported that homozygous loss-of-function mutations are very rare, at about 0.24 per person, which will make it hard to definitively pin these to the disorder. Such mutations did not occur more frequently in schizophrenia compared to controls in his sample, but the subset hitting synaptic genes did. Looking at a different kind of variation, Colm O’Dushlaine of the Broad Institute in Cambridge, Massachusetts, described his initial survey of tandem repeats as a source of schizophrenia-related variation. Consisting of repeating sequences of two to six nucleotides, these short tandem repeats are prone to mutation, which can alter transcription and translation when found within coding sequences. Scanning for this kind of variation in 5,000 exomes from the same Swedish sample, he did not, however, find evidence for an increase in frequency of tandem repeats in cases compared to controls.
In a talk on Wednesday, Ann Collins of the University of North Carolina in Chapel Hill described her work to localize the causative variant near the SNP giving the biggest GWAS signal near MIR137. MIR137 was the big surprise in the last report from the Psychiatric Genome Consortium (see SRF related news story), and suggested a role for regulatory processes in schizophrenia; consistent with this, other schizophrenia GWAS hits are also targets of MIR137. Using haplotype analysis to narrow in on the causal variant, Collins suggested that the most significant GWAS signal is close to the source of the signal, as the haplotype containing that signal was most strongly associated with schizophrenia. She also found evidence for two protective haplotypes in the region.
Figuring out function
Genetics is supposed to help biologists figure out which genes to go after to understand the roots of psychiatric disease, but biology is also lending a hand to geneticists as they contemplate the rare variants turning up in sequencing studies. In a talk on Wednesday, 17 October, Menachem Fromer of Mount Sinai described how considering whether a variant occurs within a domain of the encoded protein could help zero in on the variants relevant to disease. Though tools like PolyPhen-2 predict whether a single nucleotide substitution is likely to be damaging to the resulting protein, Fromer found that considering whether the mutation hits a protein domain, as annotated by a database called Pfam, added information. For example, of the missense variants found in the exomes of the aforementioned Swedish sample of schizophrenia cases and controls, those variants deemed damaging by PolyPhen and also landing in a domain revealed an enrichment of mutation in certain gene sets (e.g., postsynaptic density) for schizophrenia.
A symposium on Tuesday, 16 October, featured efforts to chase down the biology of some top GWAS hits in schizophrenia. Matthew Hill of King’s College London, United Kingdom, probed the effects of knocking down TCF4, a transcription factor, in different cell lines. Though it is not yet known what the risk alleles do to TCF4 expression, Hill performed a moderate 30 percent knockdown of TCF4 expression in neural progenitor cell lines and found altered expression of cell cycle genes. Consistent with this, he reported changes in cell proliferation, which varied according to the cell line used.
At the other end of the biological spectrum, brain imaging correlates of SNPs within ANK3 were presented by Danai Dima of King’s College London. Encoding a protein localized to the initial segment of the axon, ANK3 is thought to influence the discharge of action potentials, and, thus, information flow through the brain. Two different SNPs within the gene have been associated with schizophrenia and bipolar disorder, respectively. Dima reviewed how the risk allele of the “schizophrenia SNP” is associated with working memory deficits and increased activation of two regions in the prefrontal cortex (Roussos et al., 2012), and also reported a decrease in functional connectivity, as measured by fMRI among risk allele carriers, both healthy ones and those with schizophrenia.—Michele Solis.