In the largest genetic studies yet of autism, a team from across the United States has pinpointed previously overlooked DNA that predisposes children to the disorder. The suspect DNA includes the first variant that might underlie a large percentage of autism cases. Although the work is unlikely to lead to new treatments anytime soon, it bolsters a leading theory about autism's biology and points to new avenues for research.
Autism's causes are elusive, and treatments are hit and miss. Most of the DNA linked to autism involves copy-number variation, in which segments of DNA are repeated or deleted in the genomes of affected children. But these copy-number variants (CNVs) tend to be extremely rare, showing up in less than 1% of children with autism. In addition, it's been difficult to repeat the findings from individual genetic studies--partly, researchers think, because autism is so variable and may really be many different diseases.
Hakon Hakonarson, a pediatrician and genetics specialist at Children's Hospital of Philadelphia in Pennsylvania, is leading a massive effort to identify the genes behind many chronic childhood diseases (Science, 16 June 2006, p. 1584). He and his colleagues relied on two overlapping cohorts of children, drawn mainly from ongoing studies of autism, who had autism or closely related conditions. One genetics study screened for CNVs in 2195 children with autism and 2519 unaffected ones. A second, genome-wide association study scanned the genome for common variants behind autism in more than 10,000 affected children, their families, and controls.
The second study yielded the first common variants for autism. They lie in a region of chromosome 5 that sits squarely between two genes that produce cell-adhesion molecules, which govern how neurons connect to each other. The variants don't raise risk all that much--about 20% for the strongest. But, Hakonarson says, the work is exciting for two reasons: 65% of kids in the study had that variant, compared with just under 60% in the control group--a difference that, across a population, is significant, he says. And even more important, the link with cell-adhesion molecules firms up suspicions that in autism, "nerves don't connect normally or break down."
The CNV study turned up nine repeated or deleted regions, some of which had been pinpointed previously. Four of the CNVs were in genes that belong to the ubiquitin gene family, which helps control nerve connections. Children's Hospital of Philadelphia has applied for patents on the variants and hopes to help develop a test that incorporates these and other variants, mainly for siblings of children with autism. Both papers appear today in Nature.
The ubiquitin finding is "very intriguing," says Joachim Hallmayer, an autism genetics researcher at Stanford University School of Medicine in Palo Alto, California. "We have to see how this ties in" to other biological and genetic theories about the origins of autism. And the common, cell-adhesion variant, he adds, suggests that cell adhesion is "implicated in disease." Even though these studies seem large, they're still too small to nail down the major genes behind autism, he says. Hakonarson agrees that many more subjects are needed, and he is recruiting even larger groups for more genetic analysis.