By Julia Karow
Adding to similar observations for other neuropsychiatric disorders, researchers led by a team at the University of Montreal have found that de novo mutations are likely involved in schizophrenia, possibly accounting for as much as 50 percent of non-familial forms of the disease.
The researchers sequenced the exomes of 14 schizophrenia patients and their parents with no family history of the disease and found a higher-than-expected number of de novo mutations in eight of the patients. Of those, several interfere with protein function, which the authors said is "more than what is expected by chance."
The results, published this week in Nature Genetics, suggest that a large number of genes may be involved in schizophrenia, and could explain why previous linkage and genome-wide association studies have only been modestly successful in identifying genes implicated in the disease.
According to Guy Rouleau, director of the Centre Hospitalier Universitaire Sainte-Justine Research Center in Montreal and the senior author of the study, linkage and GWAS studies "can only explain a very small fraction of the heritability of schizophrenia," which some estimate to be as high as 80 percent.
Several observations are consistent with de novo mutations playing a role in the disease, he said: identical twins, for example, who share the same de novo mutations, both develop schizophrenia at a higher rate than fraternal twins or siblings; schizophrenia patients have fewer children than the general population, preventing de novo mutations from being passed on; and the rate of the disease increases with paternal age, along with the rate of de novo mutations in sperm cells.
Last year, Rouleau and his colleagues published a study in which they sequenced about 400 genes that are expressed in the synapse in several hundred schizophrenia and autism spectrum disorder patients and controls and found a number of de novo mutations that impact protein function. To follow up on this result, "we thought we should do an unbiased study, looking at all of the genes, and the development of the new [sequencing] technologies allowed us to do that," Rouleau said.
For their latest study, the scientists used Agilent Technologies' SureSelect Human All Exome Kit version1 to capture exonic DNA of 14 schizophrenia patients and their parents. They had already excluded potentially causative copy number variants using microarray technology. They sequenced the DNA using Illumina's Genome Analyzer IIx, covering, on average, 72 percent of the target regions with a read depth greater than 20-fold.
They found 73 candidate variants that appeared to be present in the probands but not their parents or controls, of which 15 — in eight of the 14 cases — were validated by Sanger sequencing as genuine de novo mutations. That number is about twice as high as expected, based on data from the 1000 Genomes Project. Also, four of the 15 mutations are nonsense mutations, affecting protein function, a rate several times higher than expected.
Based on their results, the researchers cannot say for sure that any of the mutations or genes they found actually cause schizophrenia. That, Rouleau said, requires additional sequencing in a larger number of cases as well as functional studies, both of which his team is pursuing.
[ pagebreak ]
The researchers have already sequenced the exomes of another 45 schizophrenia family trios, he said, which they are currently analyzing, and they are validating candidate genes in 400 to 500 unrelated schizophrenia cases. They are also conducting functional studies of candidate genes in zebrafish and cultured mouse neuronal cells.
If the results pan out, it could mean that "de novo mutations could be responsible for up to 50 percent of schizophrenia" that is not familial, Rouleau said. They could also shed light on some familial cases of the disease where the parents have passed on the mutation to their children.
The approach "opens the door to a strategy to really understand the genetics of schizophrenia," he said, ultimately leading to a list of several hundred genes that could be grouped into networks and pathways.
Within a few years, he said, there could be diagnostic tests that would sequence panels of genes found to be involved in schizophrenia. Eventually, the pathways discovered could become targets for drug treatments, although that is still "far in the future," he added.
The fact that the researchers did not find de novo mutations in six out of the 14 probands points to some of the technical limitations of exome sequencing, but could also have a biological explanation. For example, the exome capture kit they used in their study only targets 35 megabases of DNA, whereas the latest kit targets 50 megabases. In addition, only about two-thirds of the target DNA was covered with sufficient read depth to call variants with confidence, so some de novo mutations could have been missed.
There might be other causes in these patients, as well, such as common variants or environmental influences, Rouleau said. "I do not think that de novos explain all of schizophrenia. We are saying 50 percent — I'd be very happy if we were right, but that means there is 50 percent where it's something else."
The Canadian researchers' results mirror those of other groups that have found potentially causative de novo mutations in neuropsychiatric diseases.
This spring, for example, scientists at the University of Washington found such mutations in patients with autism spectrum disorder after sequencing the exomes of 20 family trios (CSN 5/17/2011), and Rouleau said his own team has also found a number of de novo mutations in a group of autism family trios they are studying.
Also, last year, a team at Radboud University in the Netherlands reported that some cases of mental retardation appear to be caused by de novo mutations as well, based on sequencing the exomes of 10 family trios (IS 11/16/2010).
Have topics you'd like to see covered in Clinical Sequencing News? Contact the editor at jkarow [at] genomeweb [.] com.