"This will lead to a paradigm shift in understanding causes of autism," study researcher Stephen Scherer, PHD, of Toronto's Hospital for Sick Children, said at a news conference held to announce the findings.
Finding the cause of autism is like trying to put together a huge jigsaw puzzle with no idea of what the final picture should look like, says study researcher Anthony Monaco, MD, PhD, of the U.K.'s University of Oxford.
"What we have found are the edges of the puzzle. We now can say some of the pieces are the sky or the sand of the picture," Monaco said at the news conference. "We now see that some of the genetic changes in autism are related to connections in the brain."
The study used recently developed technology to look for unusual DNA deletions or duplications -- known as copy number variants or CNVs -- in 996 people with autism and 1,287 matched people without autism.
People with autism didn't have more CNVs than people without autism, and their CNVs weren't any bigger than usual. But in autism, CNVs are much more likely to occur in gene-containing regions of the genome.
"This basically shows us that CNVs can account for a pretty appreciable percentage of autism spectrum disorder," study researcher John R. Gilbert, PhD, of the University of Miami's Institute for Human Genetics, tells WebMD.
The CNVs implicated in autism aren't all the same. In fact, the most common CNV identified in the study occurred in less than 1% of people with autism. Nearly every child studied has a unique CNV profile.
But in people with autism, CNVs cluster around networks of gene sets -- pathways -- that control brain-cell development and function. That's the paper's most important finding, says study researcher Louise Gallagher, MD, PhD, of Trinity College Dublin.
"So even if children have different genes that are influencing the development of their autism, we hope that if drug companies or biotech companies target these pathways, that the therapies might work for a broad number of the children who are affected," Gallagher said at a news conference. "Some of these therapies would still work because they act on pathways that are linked."
Autism Gene Studies: What's Next?
Studies under way are trying to link specific sets of CNVs to specific autism symptoms. Once this is done, genetic testing for CNVs could help parents identify children at high risk of autism, says study researcher Geri Dawson, PhD, of the advocacy group Autism Speaks.
"I can imagine a day when one can identify the specific genetic risk that led to an individual child's autism, and then using that genetic information to say what pathway has been affected and then choosing a medical intervention," Dawson said at the news conference.
There are no such medical treatments, although there are drugs that affect some of the pathways now linked to autism.
The new findings also shed light on the puzzling question of why children with severe autism may have brothers and sisters -- even identical twins sharing 100% of their genes -- who are unaffected.
Many of the CNVs identified in the study have what geneticists call "incomplete penetrance." This means that even if a person carries an autism-linked CNV in his or her genome, there's a chance it will have no effect.
This poses a problem for genetic testing, as children found to have autism-linked CNVs won't necessarily have autism. According to Scherer, the genes identified in the study would aid in early autism diagnosis in only about 10% of families.
Autism isn't the first disease to be linked to CNV's affecting brain function. CNV variants -- affecting some of the same pathways affected by autism-linked CNVs -- also play a role in intellectual disability (formerly called mental retardation) and schizophrenia.
Might CNVs be the long-sought factor that might make some children more susceptible to the risks posed by environmental toxins or even vaccinations?
"We do believe that environmental factors play a role in autism," Dawson said. "It is important as we continue with our science to learn how environmental factors interact with the genes we have identified here. But as of now, there is no evidence that vaccination is one of these factors."
The study findings appear in the June 9 online issue of the journal Nature.