A research team led by Christopher Walsh, M.D., Ph.D., and Eric Morrow, M.D., Ph.D., of Harvard University, found two large sections missing on chromosomes in people with autism and traced them to likely inherited mutations in such genes regulated by neuronal activity. They report their findings in the July 11, 2008 issue of Science. The study was also supported in part by the NIH's National Center for Research Resources, National Human Genome Research Institute, Eunice Kennedy Shriver National Institute of Child and Human Development, and the National Institute on Neurological Disorders and Stroke.

The study breaks new ground for complex disorders like autism, taking advantage of a shortcut to genetic discovery by sampling families in which parents are cousins. The researchers found genes and mutations associated with autism in 88 families from the Middle East, Turkey and Pakistan in which cousins married and had children with the disorder.

"The emerging picture of the genetics of autism is quite surprising. There appear to be many separate mutations involved, with each family having a different genetic cause," explained NIMH Director Thomas R. Insel, M.D. "The one unifying observation from this new report is that all of the relevant mutations could disrupt the formation of vital neural connections during a critical period when experience is shaping the developing brain."

Earlier studies had suggested that the individually rare mutations are present in at least 10 percent of sporadic cases of autism, which is the most common form.

The researchers used a technique that pinpoints from a relatively small group of families genes responsible for disorders that can be amplified by parenthood among relatives, which can increase transmission of recessive diseases. Evidence had hinted at such transmission in autism, and the large amount of genetic information obtainable from such families reduced the need for a much larger sample including many families with multiple affected members.

The ratio of females to males with autism - normally one female to four males - was less lopsided in such families in which parents share a common recent ancestor. This ratio equalized even more in a subset of these families with more than one affected member, suggesting a doubling of the rate of autism, due to recessive causes on non-sex-linked chromosomes. Also, autism-linked spontaneous deletions and duplications of genetic material were relatively uncommon in these families, suggesting recessive inherited causes.

The researchers found multiple different genetic causes of autism in different individuals with little overlap between the families in which parents shared ancestry. Yet a few large inherited autism-linked deletions, likely mutations, in a minority of families stood out. The largest turned out to be in or near genes regulated, directly or indirectly, by neuronal activity.

"Autism symptoms emerge at an age when the developing brain is refining the connections between neurons in response to a child's experience," explained Walsh. "Whether or not certain important genes turn on is thus dependent on experience-triggered neural activity. Disruption of this refinement process may be a common mechanism of autism-associated mutations."

nimh.nih/