A new study links autism to subtle changes in a wide range of genes. The finding, published in the journal Science, suggests that autism has many causes, and that whatever triggers autism usually occurs long before birth.
The study, led by Jonathan Sebat, a geneticist at Cold Spring Harbor Laboratory in New York, examined the genes of 264 families. Some families had members with autism, others did not.
Sebat says the goal was to compare the genes of autistic children with the genes of their parents.
"By comparing the two," he says, "we could find a mutation in the child that was not inherited from either parent."
It turned out that children with autism were much more likely than other kids to have these "spontaneous" mutations. The mutations affected bits of genetic code that tend to appear more than once, like duplicate copies of certain pages of a book.
In many children with autism, some of the duplicates were deleted.
For example, Sebat says one child in the study was missing a copy of the gene for oxytocin – a hormone that seems to influence social behavior.
"That child has one copy of oxytocin instead of the normal two copies," Sebat says, "and that may have resulted in the corresponding decrease in the levels of oxytocin."
It's a change that could have affected normal social development.
But Sebat says that in other children with autism, the deletions affected other genes.
"There may in fact be many genes — I would speculate 100 or more — that play some role in cognitive development, and when they are altered, could cause autism," Sebat says.
That's many more than researchers have found involved in inherited autism.
Sebat's research is causing a lot of excitement among autism researchers, including Ezra Susser, an epidemiologist at the Mailman School of Public Health at Columbia University and the New York State Psychiatric Institute.
"It changes our thinking about what kind of genetic causes are important," he says, "and knowing that changes our thinking about what kinds of environmental causes are important. Because I think everybody believes that there is an interplay of genes and environment in most diseases, and that would be true for autism, too."
The sort of mutations found in the study tend to occur in eggs or sperm before conception or in the earliest stages of an embryo's development, Susser says.
The finding suggests that things such as exposure to vaccines after birth probably aren't the main causes of autism.
Susser also says the study suggests a new direction for researchers seeking the causes of autism.
"It doesn't mean that we should stop looking for [early-life] exposures," he says. "It does not mean that we should stop looking for early life exposures. But it does mean we should start looking for exposures that might be preconceptional also."
Susser has already shown that older fathers are more likely to have genetic mutations in their sperm, and more likely to have a child who is autistic. He says scientists need to ask whether a parent's exposure to certain chemicals has a similar effect.
In the meantime, Sebat says, the new research offers at least one practical application. Parents who already have a child with autism could undergo genetic tests to see what sort of mutations are present. If the mutations are spontaneous, rather than inherited, future children would have no special risk for autism.