Prospect of 'autism drug' raised after early tests
The prospect of a drug to treat autism has been raised after symptoms of the condition were reduced in experiments on mice that were performed by the US National Institutes of Health.
There is no cure for the condition.
The results published in Science Translational Medicine showed increased social skills and less repetitive behaviour in animals taking a drug.
However, treatments which work in mice frequently fail in humans and potential medication would be years away.
Autism spectrum disorder is thought to affect around 1% of children. It ranges from mild to severe and symptoms include social problems, delayed language and repetitive movements such as hand tapping.
Autism is mainly treated with specialist education, speech and behavioural therapies.
Researchers at the National Institutes of Health said autism had been thought to be untreatable by drugs. The theory was that any problems would be "hardwired" into the brain.
However, they said there was evidence that in some cases autism could be down to the way cells in the brain communicate with each other at synapses, the gaps between individual brain cells.
They tested a drug, GRN-529, which interferes with the chemical glutamate, which helps two brain cells talk to each other.
Mice with "autistic behaviours" - this is not the same as mice actually having autism - were used.
"Autistic mice" are less social and communicate less with other mice. They also spend huge periods of time repetitively grooming themselves.
After the injection the mice spent less time grooming and also showed improvements in social levels.
The researchers said their findings "raise the possibility" that a drug could be used in autism.
Dr Jacqueline Crawley, one of the researchers from the National Institute of Mental Health, said: "Given the high costs - monetary and emotional - to families, schools and health care systems, we are hopeful that this line of studies may help meet the need for medications that treat core symptoms."
Uta Frith, a professor of cognitive development at University College London, said: "Processes at the level of the synapse have long been suspected in the origin of autism.
"However, it will be a long time until these findings can be translated for human patients. Tampering with the synapse may well result in undesirable side effects.
"Despite hopeful signs for a future drug treatment of at least some autistic behaviours, it would be sad if too much pressure was now put on researchers to rush into applications."
Richard Mills, the director of research at the National Autistic Society, said: "The NAS welcomes all research that improves our understanding of the neurobiology of autism.
"Research using animal models is important but it is not always easily translated into our understanding of autism in humans."