NEW YORK (GenomeWeb News) – The first common variant associated with autism has been identified and validated, according to a new study scheduled to appear online today in Nature. The paper, one of several autism genetic studies published online this week, supports the idea that autism involves altered connections between neurons or brain cells.
A team of researchers from across the US genotyped more than 10,000 individuals in an attempt to uncover genetic variants and copy number changes linked to autism and autism spectrum disorders. Their search turned up half a dozen common SNPs in a region on chromosome 5 between CDH9 and CDH10, cadherin genes coding for proteins that help glue cells together. One of these reached genome-wide significance and appears to account for roughly 15 percent of the population risk of autism.
The team's CNV analysis supported the idea that neuronal connections may be altered in autism. Although such CNVs were rare overall, the deletions and duplications detected tended to occur in pathways affecting either brain connectivity or protein degradation.
"Many of the genes we identified concentrate their effects in brain regions that develop abnormally in autistic children," Hakon Hakonarson, a researcher at the University of Pennsylvania School of Medicine and director of the Children's Hospital of Philadelphia's Center for Applied Genomics, said in a statement. "Our current findings, when coupled with anatomical and imaging studies, may suggest that ASDs are a problem of neuronal disconnection."
"Detailed analysis of the genes and how they affect brain development is likely to yield better strategies for diagnosing and treating children with autism," Acting National Institutes of Health Director Raynard Kington said in a statement today. The study was funded in part by NIH.
Autism spectrum disorders are neurodevelopmental disorders affecting up to one in every 150 American children, usually boys. In general, ASDs are characterized by unusual social interactions, poor verbal and non-verbal communication, and repetitive behaviors, though the specific symptoms and severity vary widely. Past genetic studies have identified numerous potential genetic variants and copy number changes, though most have been individually rare.
In an effort to identify and validate common variants associated with autism, Hakonarson and his team genotyped 3,101 individuals of European ancestry from 780 families affected by ASD at nearly half a million SNPs using the Illumina HumanHap550 BeadChip. Samples were obtained from the Autism Genetic Resource Exchange, a repository run by Autism Speaks. Subjects for other parts of the study were also recruited in Philadelphia, Miami, Los Angeles, and elsewhere.
This initial experiment didn't turn up genetic associations with genome-wide significance. But combined analysis of the AGRE data along with genotyping data for another 1,204 individuals of European ancestry with ASD and 6,491 controls from the Autism Case-Control cohort implicated six SNPs in one area on chromosome 5, between the cadherin genes CDH9 and CDH10.
Despite the large sample size, only one locus reached genome-wide statistical significance. In a conference call with reporters this morning, Hakonarson said that this is the first common variant associated with autism that has been replicated in independent studies with this rigid criteria. He said the variant was detected in about 65 percent of autism or ASD cases.
Most of the genetic associations detected in this discovery stage were confirmed in two replication studies — one involving 1,390 subjects from nearly 450 families genotyped at about a million SNPs with the Illumina HumanHap1M BeadChip and another involving a case-control dataset of 108 affected individuals and 540 controls genotyped at more than 300,000 SNPs with the Illumina HumanCNV370 BeadChip.
Hakonarson and his colleagues also detected or imputed associations between autism and several other loci, including variants in or near other genes involved in neuronal cell adhesion.
An independent study led by University of Miami Institute for Human Genomics Director Margaret Pericak-Vance and published in the Annals of Human Genetics appears to confirm the connection between the CDH9/10 region and autism, suggesting neural cell adhesion plays a role in autism.
"Until now, no common genetic variant has been identified with such overwhelming evidence to support its role in autism spectrum disorders," Pericak-Vance said in a statement. "The identification of a common variant for autism is a monumental achievement. Researchers have been looking for clues about the genetic architecture of autism for decades."
Consistent with the idea that cadherins are important to brain function, Hakonarson and his colleagues found that CDH10 is highly expressed in fetal brain tissue, particularly in an area thought to govern speech and the ability to interpret social interactions. On the other hand, CDH9 was expressed at low levels, though the researchers cautioned that this data only represents one development stage.
Because cadherins are found at low levels in blood samples, it remains to be seen if, when, and how the CDH9 and CDH10 genes contribute to brain development and/or overall function, researchers noted in the teleconference with reporters this morning. They speculated that the genes may contribute to a particular stage of brain development, though they noted that cadherins are also found in mature brain tissue.
In a second Nature paper, Hakonarson and his colleagues looked for copy number changes associated with autism using Illumina HumanHap550 BeadChip data from the GWAS. Based on data for 859 individuals with ASD and 1,409 controls, the team identified duplications and deletions affecting several genes in neuronal cell adhesion and ubiquitin pathways, as well as previously described copy number changes, including CNVs affecting the neurexin gene NRXN1.
These findings were validated by quantitative PCR and other approaches and replicated in an independent cohort including 1,336 ASD cases and 1,110 controls.
University of Pennsylvania researcher and study co-author Gerard Schellenberg told reporters this morning that the studies support the notion that autism is a multigenic condition caused by common and rare changes in several genes.
While having the high risk variant increases autism risk, Schellenberg explained, it does not cause autism on its own. Other genes need to contribute as well. Based on their calculations, the team suggested that the variant accounts for a population-attributable autism risk of roughly 15 percent.
"In most cases, it's likely that each gene contributes a small amount of risk, and interacts with other genes and environmental factors to trigger the onset of disease," Hakonarson said in a statement.
At the moment, researchers don't know enough about autism genetics to use the information diagnostically or to predict autism risk. But taken as a group, genetic studies of autism and ASD are at providing clues about the genetic risk of autism, Hakonarson said.
Hakonarson said his team is currently re-sequencing the CDH9/10 region to identify all the variants in this region and find key culprits in autism risk. The researchers also noted that larger studies are needed to identify additional autism-related genes and determine which of the weaker genetic associations detected so far actually contribute to autism. "There are other genes that need to be identified," Schellenberg said.
Meanwhile, in another autism genetics paper scheduled to appear in Molecular Psychiatry, a European research team did a high-density association analysis of regions on chromosomes 7 and 2 that were previously implicated in autism by the International Molecular Genetic Study of Autism Consortium in a few hundred families. That work suggests at least two genes on chromosome 7 — IMMP2L and DOCK4 — contribute to autism.